•
Symposium
Proceedings:
Process
Measurements for
Environmental
Assessment
(Atlanta,
February 1978)
Interagency
Energy/Environment
R&D Program Report
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments of. and development of, control technologies for energy
systems; and integrated assessments of a wide range of energy-related environ-
mental issues.
EPA REVIEW NOTICE
This report has been reviewed by the participating Federal Agencies, and approved
for publication. Approval does not signify that the contents necessarily reflect
the views and policies of the Government, 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-600/7-78-168
August 1978
Symposium Proceedings:
Process Measurements
for Environmental Assessment
(Atlanta, February 1978)
Eugene A. Burns, Compiler
TRW Systems Group
One Space Park
Redondo Beach, California 90278
Contract No. 68-02-2165
Task No. 24
Program Element No. EHE624
EPA Project Officer: James A. Dorsey
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
-------
PREFACE
These proceedings document the 26 presentations made at
the Process Measurements for Environmental Assessment Symposium
held February 13 - 15, 1978, at the Peachtree Plaza Hotel, Atlanta,
Georgia. This Symposium was sponsored by the Process Measurements
Branch of EPA's Industrial and Environmental Research Laboratory-
Research Triangle Park, as Task 24 of EPA Contract Number 68-02-2165
to TRW Defense and Space Systems Group, Redondo Beach, California.
The objective of this Symposium was to bring together people who
were responsible for planning and implementing sampling and analysis
programs for multi-media environmental assessment. The program
consisted of sessions defining the uses of environmental assess-
ment data, the techniques for acquiring information, and recent
user's field experiences with environmental assessment measurement
programs.
Mr. James A. DorBey, Chief, Process Measurements Branch,
was Symposium Chairman. The Welcoming Address was delivered by
John K. Burchard, Director of Industrial Environmental Research
Laboratory-RTP; the Keynote Paper was delivered by Dr. Stephen J.
Gage, Assistant Administrator for Research and Development for
the Environmental Protection Agency. The Symposium was organized
and planned by Dr. Eugene A. Burns, who was employed by TRW Defense
and Space Systems during the initial part of the project and
finally by Systems, Science and Software during the last four
months of the project activity. He was assisted by Mr. Charles T.
Weekley of TRW who in turn had support services at the Symposium
from M. A. McKay, L. Shober and M. W. Wong of TRW and B. Foil
and S. Sharpe of EPA's IERL.
11
-------
TABLE OF CONTENTS
PAGE
PREFACE ............................................. 11
PARTICIPATING EPA OFFICIALS AND SESSION CHAIRMEN ......................... V1
MONDAY, FEBRUARY 13
PROCESS MEASUREMENTS FOR ENVIRONMENTAL ASSESSMENT
Stephen J. Gage, Assistant Administrator for
Research and Development, EPA
ENVIRONMENTAL ASSESSMENT OVERVIEW, SESSION I
R. P. Hangebrauck, Director, Energy Assessment and
Control Division, IERL-RTP. EPA, Session Chairman
THE DOE INTEGRATED ASSESSMENT PROGRAMS .............................. 4
Ray Cooper, Division of Regional Assessments. E&S, DOE
RELATED EPRI PROGRAMS ....................................... 10
Ralph Perhac, Electric Power Research Institute
ENVIRONMENTAL ASSESSMENT OVERVIEW, SESSION II
Janes Dorsey, Chief, Process Measurement Branch. IERL-RTP,
EPA, Session Chairman
EPA AIR PROGRAMS' USE OF ENVIRONMENTAL ASSESSMENTS ........................ 14
Richard G. Rhoads, Director, Control Programs
Development Division, OAQPS, EPA
AN INTEGRATED APPROACH TO THE ASSESSMENT AND CONTROL
OF INDUSTRIAL POLLUTION PROBLEMS ......................... . ....... 17
Eugene E. Berkau, and A. B. Craig, Industrial Pollution
Control Dtvtston, IERL-C1, EPA
SOURCE ASSESSMENT METHODOLOGY ................................... 23
Thomas N. Hughes. Monsanto Research Corp.
HEALTH RELATED PROGRAMS ..................................... 28
Shahbeg Sandhu. and Michael D. Haters, HERL-RTP, EPA
BIOLOGICAL TESTING METHODOLOGY ............................. ..... 38
Kenneth M. Duke. Battelle Columbus Laboratories
111
-------
TABLE OF CONTENTS (CONTINUED)
TUESDAY, FEBRUARY M
MEASUREMENT TECHNOLOGIES, SESSION I
Larry Johnson, IERL-RTP, EPA, Session Chairman
SOURCE ASSESSMENT SAMPLING SYSTEM DESIGN DEVELOPMENT AND CALIBRATION 45
David Blake and J. M. Kennedy, Acurex/Aerotherm
FIELD EVALUATION OF THE SASS TRAIN AND LEVEL 1 PROCEDURES 59
Franklin Smith, Eva D. Estes and Denny E. Wagoner,
Research Triangle Institute
INORGANIC EMISSIONS MEASUREMENTS 72
Ray F. Maddalone and Lorraine E. Ryan, Applied
Technology Division, TRW DSSG
ORGANIC ANALYSIS FOR ENVIRONMENTAL ASSESSMENT 84
Philip L. Levins, Arthur D. Little. Inc.
A CRITQUE OF ORGANIC LEVEL 1 ANALYSIS 93
Peter W. Jones and Robert J. Jakobsen. Battelle
Columbus Laboratories
MEASUREMENT TECHNOLOGIES, SESSION II
Charles Lochmuller, Duke University. Session Chairman
ENVIRONMENTAL ASSESSMENT MEASUREMENT TECHNIQUES FOR FUGITIVE
EMISSIONS 98
Henry J. Kolnsberg, The Research Corporation of
New England
SAMPLING AND ANALYSIS PROCEDURES FOR SCREENING OF
INDUSTRIAL EFFLUENTS FOR PRIORITY POLLUTANTS 104
William A. TelHard, and Gall S. Goldberg.
Effluents Guidelines Division. ONPS. EPA
ALTERNATIVE LEVEL 1 ANALYSIS METHODS 108
Karl Bonbaugh, Radian Corporation
SYNTHETIC FUELS PRODUCTION: ANALYSIS OF PROCESS BY-PRODUCTS
FROM A LABORATORY SCALE COAL GASIFIER 121
Charles M. Sparaclno, R. A. Zweldlnger. S. Willis, and
D. Mlnlck, Research Triangle Institute
TRANSFORMATION OF POM IN POWER PLANT EMISSIONS 138
D. F. S. Natusch. W. A. Korftiacher, A. H. Miguel.
M. Schure, and B. A. Tonkins. Colorado State University
1v
-------
TABLE OF CONTENTS (CONTINUED)
HEDNESDAY, FEBRUARY 15
INDUSTRIAL PROCESS APPLICATIONS
Eugene A. Burns. Systems, Science and
Software. Session Chairman
ASSESSMENT OF ATMOSPHERIC EMISSIONS FROM PETROLEUM REFINING 147
D. D. Rosebrook, R. G. Wetherold, and G. E. Harris,
Radian Corporation
TOXICITY OF SECONDARY EFFLUENTS FROM TEXTILE PLANTS 153
Gary D. Rawllngs, Monsanto Research Corp. and Max Samfleld,
IERL-RTP, EPA
NONFERROUS METAL PROCESSING 170
D. Meek, G. Nichols, Southern Research Institute, and
J. 0. Burkle, IERL-C1, EPA
APPLICATION OF THE PHASED APPROACH TO ENVIRONMENTAL ASSESSMENT
TO THE EMISSION ASSESSMENT OF CONVENTIONAL COMBUSTION SERVICES 176
J. Warren Hamersma, Applied Technology Division, TRW DSSG
EMISSIONS FROM THE GLASS MANUFACTURING INDUSTRY 184
C. Darvln. IERL-C1, EPA. R. Barrett. Battene Columbus
Laboratories and W. Blakeslee, Scott Environmental
Services'.
ENERGY PROCESS APPLICATIONS
Karl Bonbaugh, Radian Corporation,
Session Chairman
COMPREHENSIVE ANALYSIS OF EMISSIONS FROM FLUIDIZED-BED
COMBUSTION PROCESSES 189
K. S. Murthy. J. E. Howes, and H. Nack, Battelle Columbus
Laboratories and R. D. Hoke, Exxon Research and Engineering Co.
ENVIRONMENTAL ASSESSMENT PROGRAM FOR THE HYGAS PROCESS 207
L. J. Anastasla, W. G. Ba1r, and D. P. Olson. Illinois
Institute of Gas Technology
ANALYSIS OF SYNTHANE/SYNTHOIL PRODUCTS 211
A. G. Sharkey, Jr.. and J. L. Shultz, Pittsburgh Energy
Research Center
CHARACTERIZATION OF OIL SHALE PROCESSES ' 216
Jack E. Cotter, Environmental Engineering Division, TRW
v
-------
PARTICIPATING EPA OFFICIALS AND SESSION CHAIRMEN
JAMES A. DORSEY
Symposium Chairman
Mr. Dorsey is Chief of the Process
Measurements Branch in the EPA's Industrial
Environmental Research Laboratory-RTP. He
received his B.S. in Chemistry from Florida
State University. He is responsible for
integrated in-house/contract programs for
the development and application of measure-
ment procedures for energy and industrial
processes. Prior to joining EPA in 1964,
Jim was employed by Shell Oil Company as a
research chemist. He has over 20 years of
experience developing sampling and analysis
techniques using advanced methods of instru-
mental analysis. Current studies include
the sampling and analysis of organic and
inorganic trace materials in gas, liquid
and solid feed, product and waste streams.
Jim is a member of the American Chemical
Society and the Air Pollution Control
Association.
JOHN K. BURCHARD
Welcoming Address
Dr. Burchard is the Director of the
Industrial Environmental Research Labora-
tory-RTP. He is responsible for the man-
agement of programs to develop and demon-
strate cost-effective technologies to
prevent, control, or abate pollution from
industrial operations involving energy
and mineral resources. He was recently
appointed as senior ORD Official for EPA's
Environmental Research Center in Research
Triangle Park, N.C. Since joining EPA in
1970, he has served in several capacities -
as Chief of the Laboratory's Technical
Analysis Section, Chief of the Development
Engineering Branch, and Assistant Director.
Before joining EPA, Dr. Burchard worked in
industry for 10 years. He holds B.S., M.S.,
and Ph.D. degrees in Chemical Engineering
from Carnegie Tech.
STEPHEN J. GAGE
Keynote Address
Dr. Gage is Assistant Administrator
for Research and Development for EPA. He
is responsible for planning, directing and
coordinating all Agency research activities
covering air, water, toxic substances,
radiation energy, pesticides and solid
wastes. He joined EPA in 1974 as Acting
Director of the Office of Energy Research;
in 1975 he became the Deputy Administrator
for Energy, Minerals and Industry. Prior
to joining EPA, he was with the Council on
Environmental Quality. During 1971-73,
Dr. Gage was a White House Fellow. He
joined the faculty at the University of
Texas in 1965 and became Director of the
Nuclear Reactor Laboratory in 1966. He
received a B.S. in Mechanical Engineering
from the University of Nebraska and M.S.
and Ph.D. degrees from Purdue University.
ROBERT P. HANGEBRAUCK
Session Chairman
Mr. Hangebrauck is the Director of the
Energy Assessment and Control Division at
EPA's Industrial Environmental Research
Laboratory. Mr. Hangebrauck received his
B.S. from Cal-Tech in 1959. He has served
with the EPA and its predecessor agencies
for 17 years in various capacities, includ-
ing Chief of the Clean Fuels and Energy
Branch, Chief of the Demonstration Projects
Branch, Control Systems Laboratory, and
Assistant to the Director of the Bureau of
Engineering and Physical Sciences. He is
currently in charge of conducting the
divisions' research and development pro-
grams to identify and control multimedia
pollutants discharged in the environment
from stationary sources.
LARRY JOHNSON
Session Chairman
Dr. Johnson is an Analytical Chemist
in the Process Measurements Branch for
EPA's Industrial Environmental Research
Laboratory-RTP. He received his B.S. in
Chemistry from Iowa State University and
his Ph.D. in Chemistry from the University
of Texas. Dr. Johnson manages contracts
to develop screening procedures for environ-
mental assessment programs and compound
specific procedures for detailed analysis
of complex samples from industrial process-
es. He is also responsible for evaluating
the applicability of biological screening
tests to complex effluent samples. He has
specialized in the areas of organic samp-
ling and analysis for trace components in
industrial feed stocks and energy process
effluents. Prior to joining EPA in 1972,
Dr. Johnson was employed as an analytical
chemist by PPG Industries and as a research
chemist by Monsanto.
CHARLES H. LOCHMULLER
Session Chairman
Dr. Lochmuller is an Associate Profes-
sor of Chemistry with Duke University. He
is a consultant to the EPA Industrial
Environmental Research Laboratory/RTP
Process Measurement Branch in the areas
related to the development of analytical
methodology for environmental assessment
and a member of its advisory panel on
vl
-------
organic analysis. Prof. Lochmuller was
one of the originators of the staged
sampling and analysis protocol for environ-
mental assessment process measurements. He
received his B.S. in Chemistry from Manhat-
tan College and his M.S. and Ph.D. degrees
in Analytical Chemistry from Fordam Univer-
sity. His research interests are in the
area of fundamental aspects of the molecu-
lar basis for selectivity in chemical
separation methods. His recent work in-
cludes areas of charged particle induced
x-ray emission analysis and Fourier Trans-
form Magnetic Resonance Spectroscopy.
EUGENE A. BURNS
Session Chairman
Dr. Burns is Manager of the Chemistry
and Chemical Engineering Program for
Systems, Science and Software (S3). He
received his B.A. in Chemistry from Pomona
College and his Ph.D. in Analytical Chem-
istry from Massachusetts Institute of
Technology. Over the past 25 years, he has
held a variety of research and management
positions. Prior to joining S9, Dr. Burns
was employed by TRW for 15 years. As
Manager, Chemistry Department, one of his
responsibilities was energy-related environ-
mental processes including the development
of sampling and analysis methodology for
characterizing process streams. Previous
experience includes Chief of the Chemistry
Section, Jet Propulsion Laboratory; and
Head of the Analytical Chemistry Section,
Stanford Research Institute's Propulsion
Services Division.
KARL J. BOMBAUGH
' Session Chairman
Mr. Bombaugh is a Principal Scientist
at Radian Corporation where he is respon-
sible for the development of strategies and
systems for environmental tests. He ob-
tained his B.S. in Chemistry from Juniata
College. Over the past thirty years, he
has held a variety of positions in both
research and management. His experience
covers a broad range of analytical and
process technology, including infrared
spectrometry, both gas and liquid chroma-
tography, and on-stream analysis. As
Vice-president for Research and Development
at Waters Associates, he directed a group
who pioneered in the development of modern
high pressure liquid chromatography. He
has authored more than fifty publications
including chapters in several books. Re
has served on the editorial advisory board
of the Journal o& ChJiomatogiaph-Lc. Science.
and the Chem-tcal Rub be* Handbook o£ Ckioma-
togiaphy. He is now Chairman Emeritus of
ASTM Committee E19 on Chromatography.
vll
-------
SESSION SPEAKERS
ENVIRONMENTAL, ASSESSMENT OVERVIEW
SESSION I
Prior to joining EPA, he conducted operations re-
search, systems analysis, and aeronautical design
studies for the Federal Government.
RAYMOND D. COOPER
Dr. Cooper is Acting AD for Integrated
Assessment in the Division of Regional Assess-
ment for the Department of Energy. He received
his B.S. in physics from the University of Illinois.
Dr. Cooper has advance degrees in physics (M.S.
and Ph. D.), from Iowa State Universtiy and Mas-
sachusetts Institute of Technology, respectively.
Prior to assuming his present position, Dr. Cooper
served as Assistant Director for Integrated
Assessment in the Division of Technology Over-
view, Energy Research and Development Admin-
istration, and Program Manager at the U.S.
Atomic Energy Commission for Radiological
Physics in the Division of Biomedical and Envi-
ronmental Research. Prior to this he did
research for the U.S. Army on the effects of
nuclear radiation, electron and gamma ray inter-
actions, dosimetry, and application of electron
accelerators to food irradiation. Before his work
for the Federal Government, he taught physics at
Tufts University.
RALPH M. PERHAC
Dr. Perhac is Program Manager, Environ-
mental Assessment Department at the Electric
Power Research Institute (EPRI). He received his
A.B. from Columbia, A.M. from Cornell Univer-
sity, and his Ph.D. from the University of Michi-
gan. Prior to joining EPRI, he was the Program
Director for Environmental Effects of Energy for
the National Science Foundation. Prior to that, he
was a Consulting Researcher, Environmental Sci-
ences Division at Oak Ridge National Laboratory.
He was a professor of Geochemistry at the Univer-
sity of Tennessee for seven years after he left the
Exxon Corporation where he was a. Senior Research
Geochemist.
EUGENE E. BERKAU
Dr. Gene Berkau is Director, Industrial
Pollution Control Division, Industrial Environ-
mental Research Laboratory (IERL) Cincinnati,
Ohio, U.S. Environmental Protection Agency.
Gene's Division is responsible for sponsoring pro-
grams in research, development and demonstra-
tion of cost-effective technologies to control pollu-
tion from the manufacturing of industrial products.
Dr. Berkau has advanced degrees in Chemical
Engineering (M.S. and Ph.D.), from Vanderbilt
University. He has been with EPA and its prede-
cessor Agency since 1970. Prior to the current
assignment, his EPA responsibilities were in re-
search development and demonstration of control
methods for noise and air pollution from combus-
tion systems. Before joining the Federal Govern-
ment, he was employed by the Monsanto Textile
Division at the Chemstrand Research Center, Re-
search Triangle Park, North Carolina.
THOMAS W. HUGHES
Mr. Hughes is a Research Group Leader at
Monsanto Research Corporation. Tom received
his B.S. in Chemical Engineering from the Univer-
sity of Cincinnati in 1973. He has been involved
in Environmental Assessment studies for EPA at
Monsanto Research Corporation since 1974. He
has been performing Source Assessment studies
for IERL in the organic chemicals and product
areas. Tom has also been involved in control
technology development for both criteria pollutants
and toxic substances in air and water media. The
Source Assessment studies that Tom discussed
constitute one of the first major EPA programs
involving comprehensive pollutant assessment of
source emissions.
ENVIRONMENTAL ASSESSMENT OVERVIEW
SESSION II
RICHARD G. RHOADS
Mr. Rhoads is Director of the Control Pro-
grams Development Division in the Environmental
Protection Agency's Office of Air Quality Plan-
ning and Standards at the Research Triangle Park,
North Carolina. He is responsible for developing
national policy and guidance, and providing evalu-
ations and assessing effectiveness, of the air pol-
lution control programs under Section 110 and
lll(d) of the Clean Air Act. Dick graduated from
Rensselaer Polytechnic Institute in I960. Prior to
assuming his present position, Mr. Rhoads served
in several other areas of the Environmental Pro-
tection Agency's air pollution control program.
SHAHBEG S. SANDHU
Dr. Sandhu is a Research Biologist for
Health Effects Research Laboratory, Environ-
mental Protection Agency, Research Triangle
Park, North Carolina. He received his B.S. and
M. S. from Punjab University, India. He received
his Ph.D. from Purdue University in 1968. Prior
to joining EPA, he was Professor of genetics at
North Carolina Central University. His major re-
search areas are mammalian cell culture, muta-
genetics and plant genetics. Dr. Sandhu is pres-
ently assisting the IERL on programs for the
application of bioassay tests to industrial samples.
V111
-------
KENNETH M. DUKE
PHILIP L. LEVINS
Or. Duke IB Associate Manager, Ecology and
Ecosystems Analysis Section, Battelle Columbus
Laboratory. He received his B.S. in Zoology and
Entomology (1966) at Brigham Young University;
his M.S. in Zoology and Entomology (1968), from
Brigham Young University, and his Ph.D. in Ento-
mology (1971) from the University of Georgia. He
is currently the leader of a program designed to
assist the U.S. Environmental Protection Agency
in developing a protocol of bioassays to screen
various gaseous, liquid, and solid industrial efflu-
ents for their toxicity.
Dr. Levins is Leader of the Analytical
Chemistry Unit at Arthur D. Little, Inc. He has
directed several experimental programs whose
focus has been on problem definition and method
development concerning organic species in the
environment. Among these is the present organics
measurement program for EPA, development of
sampling and analysis methods for industrial hy-
giene studies and determination of the chemical
odor species in diesel exhaust. He received his
B.S. in Chemistry from the University of Vermont
and Ph. D. in Physical Organic Chemistry from the
University of New Hampshire.
MEASUREMENT TECHNOLOGIES
SESSION I
DAVID E. BLAKE
Mr. Blake is the Manager, Industrial and
Environmental Engineering Department, Aerotherm
Division of Acurex Corporation. Mr. Blake ob-
tained bis B.S. in Chemical Engineering at Cali-
fornia State University at San Jose in 1965. He
worked for seven years at Stanford Research In-
stitute and three years at Electro Print, Inc. be-
fore joining Acurex in 1975. At Acurex, Mr. Blake
has managed a number of projects primarily in the
areas of environmental instrumentation and devel-
opment of coal conversion processes. He was the
Program Manager for the SASS design and develop-
ment program which was conducted for the Process
Measurements Branch.
FRANKLIN SMITH
Mr. Frank Smith has been an engineer on
the technical staff of the Research Triangle Insti-
tute since 1970. His primary areas of work have
been in the development and application of quality
assurance techniques to air pollution and measure-
ment systems and networks. He received a B.S.
in mathematics from Arkansas State University in
1959 and an M.S.T. in physical sciences from the
University of Missouri in 1964. Prior to 1970,
Mr. Smith was employed as a physicist at the
Naval Weapons Laboratory, Dahlgren, Virginia,
where he was concerned with the effects of elec-
tromagnetic radiation on various ordnance items.
RAYF. MADDALONE
Dr. Maddalone is Head of the Environmental
and Process Chemistry Section at TRW Defense
and Space Systems. He is Manager of the Reduced
and Oxidized Inorganic Emissions Term Form
Level-of-Effort Program sponsored by EPA-IERL/
RTP. He received bis B.S. in Chemistry from
Notre Dame in 1970 and his Ph.D. in Analytical
Chemistry from Louisiana State University in 1974.
Ray is a specialist in SO2/SO3 emission
measurements.
PETER W. JONES
Dr. Jones is Associate Manager of the
Organic, Analytical and Environmental Chemistry
Section, Battelle Columbus Laboratories. His
principal responsibilities have been to initiate and
direct hazardous ambient and stack gas pollutants.
Dr. Jones received his B.S. in Chemistry and
Physics in 1966 from Hatfield College of Technol-
ogy, Hatfordshire, England, and his Ph. D. in
Chemistry from the University of York, Yorkshire,
England in 1969. Prior to joining Battelle in 1971,
he was a post-doctoral fellow at the University of
Manitoba. Dr. Jones is a specialist in sampling
and analysis of trace organic emissions from in-
dustrial and energy processes.
MEASUREMENT TECHNOLOGIES
SESSION II
HENRY J. KOLNSBERG
Mr. Kolnsberg is a Senior Project Manager
with TRC — The Research Corporation of New
England in Wethersfield, Connecticut, where he is
responsible for contracts to develop fugitive emis-
sions measurement techniques and to provide envi-
ronmental consulting services to a variety of cli-
ents. A mechanical engineering graduate of The
Cooper Union, he received his M.B.A. in Manage-
ment from the University of Connecticut.
WILLIAM A. TELLIARD
Bill Telliard is Chief of the Energy and Min-
ing Branch within the Effluent Guidelines Division
of the Office of Water and Hazardous Materials,
the Environmental Protection Agency. He attended
Kent State University and Western Reserve where
he received his Masters Degree in chemistry.
Among his major responsibilities is the analytical
chemistry phase of the current BAT review proj-
ects. In this role. Bill handles the development
and updating of sampling and analytical protocols,
as well as validation procedures appropriate to
each of the 21 industrial studies.
1x
-------
CHARLES M. SPARACINO
Dr. Sparacino received his Ph.D. in Organic
Chemistry from Emory University in 1969. He
spent a year as a postdoctoral fellow at the
Worcester Foundation for Experimental Biology in
Massachusetts, studying biogenetic problems re-
lating to steroidal terpenes. After joining the Re-
search Triangle Institute in 1970, Dr. Sparacino
has been involved in drug synthesis and metabolism
studies, and more recently, in programs involving
environmental problems. These programs include
method development for coal gasification by-
products and air partierulate and vapor analysis.
DAVIDF.S. NATUSCH
Dr. Natusch is Professor of Chemistry and
Acting Chairman of the Chemistry Department at
Colorado State University. He obtained his B.S.
and M.S. degrees from the University of Canter-
bury in New Zealand and his Doctorate as a Rhodes
Scholar at Oxford University in Britain. He has
worked as a research scientist for the New Zealand
Department of Scientific and Industrial Research in
the areas of air pollution and geothermal power
production. Most recent, he was Associate Pro-
fessor of Environmental Chemistry at the Univer-
sity of Illinois. His research involves the develop-
ment and application of instrumentation for the
detailed characterization of environmental pollu-
tants and energy process stream materials and
the investigation of the physical and chemical be-
havior of such materials.
INDUSTRIAL PROCESS APPLICATIONS
DONALD D. ROSEBROOK
Don Rosebrook is a Program Manager at
Radian Corporation. He received hie B.S. in
Chemical Engineering from Purdue in 1958 and a
Ph. D. in Analytical Chemistry from Kansas State
in 1964. He began his professional career at the
Midwest Research Institute where he was primarily
involved in trace analysis, development of sam-
pling approaches for trace organics in all media,
gas chromatography-mass s pectrometry. He
joined Syracuse Research Corporation in 1972
where he was Manager, Analytical Services.
Since 1975, he has been with the Radian Corpora-
tion where he started the Organic Chemistry De-
partment, primarily involved in trace analysis and
programs dealing with refining and coal
conversion.
experience in assessing the potential environmental
impact of several industries in support of an EPA
contract entitled "Source Assessment. "
J. WARREN HAMERSMA
Dr. Hamersma is a staff engineer for envi-
ronmental projects in the Chemistry Department of
TRW Defense and Space Systems. He received his
B.S. in Chemistry from Calvin College and his
Ph. D. in Physical Organic Chemistry from the
University of Connecticut. He is presently in
charge of the laboratory operations for the Envi-
ronmental Assessment of Conventional Combustion
Systems Program. Dr. Hamersma was manager
for EPA-funded tasks to develop and integrate the
phased (Level 1 and Level 2) environmental assess-
ment sampling and analytical strategy and to docu-
ment procedures for Level 1 sampling and analysis.
Prior to joining TRW, he had eight years in quality
control and organic analysis experience at Interna-
tional Chemical and Nuclear Corporation and at
Arco Chemical Corporation.
RICHARD L. MEEK
Dr. Meek is Head, Chemical Process Section
of Southern Research Institute, Birmingham, Ala-
bama. He received his Ph.D. in Chemical Engi-
neering from Georgia Tech in 1952. Dr. Meek is
currently working on a program to evaluate control
systems in the nonferrous metals industry spon-
sored by EPA1 s Cincinnati Industrial Environmental
Research Laboratory. Prior to joining Southern
Research Institute, he was with the Cities Service
Company as Research Coordinator and Assistant
Director, Research and Development for Cities
Services Chemicals and Metals Division.
CHARLES H. DARVIN
Mr. Darvin is Program Manager for Metal
Finishing Industries and Miscellaneous Industries
Program for the EPA Industrial Environmental
Research Laboratory, Cincinnati, Ohio. His
assignments include die development of the labora-
tory research programs for these industries. His
education includes a B.S. degree in Mechanical
Engineering from the University of Evansville,
Evansville, Indiana, and graduate work in Applied
Mathematics and Business. His previous employ-
ment includes assignment to the Office of Air Qual-
ity Planning and Standards, Research Triangle
Park, North Carolina, where he was the standards
development engineer for the lead and zinc
industries.
GARY D. RAWLINGS
Dr. Rawlings is a Senior Research Engineer
at the Monsanto Research Corporation. He is cur-
rently the Project Leader of a special source
assessment project designed to evaluate the tox-
ic it y of textile mill waste waters. He received his
B.S. in Physics in 1970 and his M.S. in Nuclear
Physics in 1971, both at Southwest Texas State
University. He received his Ph.D. in Environ-
mental Engineering in 1974 at Texas AfcM Univer-
sity. Dr. Rawlings has had wide industrial
ENERGY PROCESS APPLICATIONS
K.S. MURTHY
Kesh Murthy is a Senior Engineer of the
Energy and Environmental Process Department of
Battelle Columbus Laboratories. He is currently
the Deputy Program Manager of the ongoing pro-
gram of environmental assessment of the fluidized-
bed combustion process. Mr. Murthy received his
-------
B.S. in Chemistry, Physics and Mathematics In
1958 from the University of Mysore, India; his
B.S. in Chemical Engineering in 1966 from the
Indian Institute of Chemical Engineers, Calcutta,
India; and his M.S. in Environmental Engineering
in 1973 from the University of Cincinnati. At
Battelle, Mr. Murthy has focused on many re-
search projects involving energy processes and
their environmental interaction.
L.J. ANASTASIA
Lou Anastasia joined the staff of the Institute
of Gas Technology in 1973. He is currently Man-
ager of Environmental Engineering at the HYGAS
Pilot Plant where he is responsible for the environ-
mental assessment of the HYGAS Process and
plant pollution control. He received his B.S. and
M.S. degrees in Chemical Engineering from Pur-
due University. Previously, Lou was associated
with Argonne National Laboratories where his 15
years of experience included studies on fluidized-
bed combustion of coal, pyrochemical reprocessing
of spent nuclear fuel, safety aspects for lithium-
sulfur batteries for automobile propulsion, and
fluidized-bed fluoride volatility reprocessing of
spent nuclear fuel.
A.G. SHARKEY, JR.
A.G. "Jack" Sharkey is Manager of the
Chemical and Instrumental Analysis Division of
the Pittsburgh Energy Research Center. He is
also Adjunct Associate Professor and Member of
the Graduate Faculty, Earth and Planetary Science
Department, University of Pittsburgh. He received
his A. B. in 1941 from the College of Wooster and
his M.S. from Case Institute of Technology in 1943.
After three years of employment at Westinghouse
Research Laboratories, he was employed by the
Bureau of Mines and now the Pittsburgh Energy
Research Center for the past 31 years. His major
research interests are numerous but he has focused
on the application of spectral analysis techniques
to coal and coal-derived fuels. He has over 200
papers in the area of analysis of coal, coal-derived
fuels and mass spectrometry instrumentation.
JACK E. COTTER
Dr. Cotter is currently an industrial pro-
grams manager for the Environmental Engineering
Division of TRW. He has managed a program for
IERL-Cincinnati, initiated in 1975, to evaluate pol-
lution control technologies needed for oil shale
processes, together with recent field sampling and
analysis work. Dr. Cotter received a B.S. in
Chemical Engineering from MIT, and a Ph. D. in
Engineering Science from the University of Cali-
fornia (Berkeley). Prior to joining TRW, Dr.
Cotter was a project engineer for air and water
pollution control systems design at Daniel, Mann,
Johnson fe Mendenhall in Los Angeles. In addition,
he has done industrial instrumentation systems
consulting with Bunker-Ramo Corporation, Stan-
dard Oil Company of California, and as co-founder
of Intersystems Associates.
X1
-------
xll
-------
KEYNOTE ADDRESS
PROCESS MEASUREMENTS
FOR
ENVIRONMENTAL ASSESSMENT
by
Stephen J. Gage
Assistant Administrator for
Research and Development
Environmental Protection Agency
In the beginning, environmental
assessment was a whole lot easier. When
the cave man stuck his head out of his
cave, he could tell immediately whether a
forest fire had polluted the air during
the night. His keen senses could also
detect the carrion that might be polluting
his water supply—the stream that ran down
the valley. But then again, he lost the
major share of his family to mysterious
and unknown causes.
Only in modern times have we been
able to determine the causes of most of
the afflictions which killed or debilitated
the cave man's family. We've made the most
progress in identifying and controlling
germs or bacteria with unbelievable success
in essentially stopping plagues, smallpox,
tuberculosis, and diptheria, to name a few.
We've made some important progress in
identifying viruses, less so in controlling
them. Unfortunately, we probably communi-
cate the common cold from one to another—-
and suffer the consequences—in the same
way the cave man did, the many commercial
palliatives notwithstanding.
Our progress in identifying and con-
trolling man-made chemicals in the environ-
ment has also been mixed. But then again,
the general dispersion of toxic chemicals
in the environment has only occurred
seriously since the Industrial Revolution.
In fact, most of the environmental
chemicals we fear today have come into
production only after World War II.
Certainly one of the earlier mani-
festations of man-made environmental
problems, probably predating the Bronze
Age, was lead poisoning from clay pottery.
The naturally occurring lead salts leached
from the clay pot had slow, cunmulative
effects on the early technologist and his
family as he used his wits to superimpose
his will on a harsh environment. But, in
that act of shaping the environment to
satisfy his needs, he had simultaneously
concentrated a dilute toxic chemical and
increased his exposure to that chemical.
A nice lesson from primitive technology,
isn't it.
Now, we recognize the threat of heavy
metals like lead, cadmium, arsenic, and
mercury, although we're not sure what con-
trol levels and methods to specify in all
cases. We are less sure about the nature
of the threat from toxic organic chemicals,
intentionally or accidentally introduced
into the environment. We are much less
sure of what to do in many instances.
The threat from synthetic chemicals
appears to be very general. The discovery
of many man-made toxic chemicals, some of
which were known cancer-causing agents, in
New Orleans' drinking water several years
ago was extremely sobering. Since then,
analysis of the drinking water supplies of
one hundred American cities indicated that
a number of the cities sampled might have
contamination problems with such exotic
chemicals as tetrachloroethylene, penta-
chlorophenol bis-(2-chloroisopropyl) ether,
and 3,4-benzofluoranthene. Fortunately,
treatment techniques, such as activated
carbon filtration, may remove most of these.
Of even greater concern are those
situations where the chemicals have become
widely dispersed in the environment before
we knew it was happening, where the
chemicals resist transformation into non-
toxic products or do transform into very
toxic materials, and where the chemicals
or their products will continue to expose
man or part of his food chain for years
to come. DDT throughout the world, PCB's
in the bottom of the Hudson River, kepone
in the bottom of the James River—these
are examples of where industrial activities
have made perhaps an irreversible impact
on the environment.
-------
This is where Environmental Assess-
ment comes in. We've learned a lot about
the threats of pollution coining out of
industrial air, water, and residual waste
streams in the last few years. But, we
want to do all we can to avoid being
surprised by the DDT-, PCB-, and kepone—
like disasters of the future.
If we are to keep the Nation's economy
growing, that growth will depend in large
measure on industrial expansion. To keep
industry from fouling its nest and ours as
well, we must know what is in industrial
products, by-products, and wastes and must
act to protect us and the environment, even
while allowing industry to grow.
But, Environmental Assessment is no
easy task. For the past three years we
have been systematically measuring, with
as much precision as possible, what and
how much are being emitted by many indus-
trial, especially energy processing,
technologies. He have found that the
major components of the products, by-
products and waste streams are generally
well-characterized. But with increasing
concern over trace level contaminants and
their subtle, cumulative effects on human
health, we have been faced with increas-
ingly difficult assessment problems.
The difficulty of obtaining repre-
sentative samples from process streams
which are at high temperature and pressure,
which involve two- or even three-phase
flow, and which often include other highly
reactive chemical species, is immense.
Chemically analyzing complex mixtures
of chemicals, especially to trace contam-
inant levels, is the next tremendous
challenge, even with today's powerful
analytical chemistry techniques.
Once the emissions are chemically
characterized, determination of the
relative degree of health or environmental
risk is the next painstaking step. This
activity is particularly vexing when the
long-term cumulative effects, such as those
which might result from exposure to carcin-
ogenic, mutagenic, or teratogenic agents
are considered.
The fourth step, required when the
preliminary screening results give strong
evidence of toxicological or other hazards,
is to conduct health effects studies of the
waste streams or at least of their most
hazardous components. This step is beyond
the scope of Environmental Assessment as
we usually define it. Similarly, the
fifth step, identification of methods to
control the substance, usually follows the
Environmental Assessment.
The emergence of the Environment
Assessment Program as a distinct and impor-
tant part of EPA1s environmental research
efforts over the past several years is an
excellent example of how the Agency's
efforts have turned from primarily research
in reaction to known environmental problems
to include research which anticipates and
tries to avoid future environmental problems.
This change of emphasis also predated the
rechartering of EPA's position toward toxics.
One of the best examples of these
efforts has been in the area of chemical
processing of coal. As we all know, the
United States is going to have to use much
more coal if our economy is to grow and our
national security is not to be seriously
compromised. Part of the coal will probably
have to be converted to clean liquids and
gases if the multiplicity of U.S. fuel
needs are to be met. During the conversion
process, coal is typically treated with
hydrogen in one form or another to increase
the ratio of hydrogen-to-carbon atoms in
the resulting products. Because of the
complex chemistry of coal and the economic
necessity to hydrogenate only as far as
required to make the desired product, the
conversion process produces hundreds of
different hydrocarbon compounds as by-
products. Some of these by-products end
up as trace contaminants in the products,
others are emitted in fugitive emissions,
and still others end up in residuum from
the process. Even though these compounds
may exist only at trace levels, they may
present a serious health risk to plant
workers or to the general public. Our
Environmental Assessment activities have
made considerable progress in identifying
and quantifying many of these by-product
materials. More will be made in the near
future.
During the past decade, there has been
a subtle but definite shift in our approach
to environmental protection. He have moved
steadily away from doing barely enough to
protect public health and safety and toward
doing as much as is practicable. In other
words, we are beginning to use as much
control technology as we can economically
tolerate in order to have some margin of
safety and flexibility for the future.
In 1970, the modified Clean Air Act
established the requirement for New Source
Performance Standards, necessitating the
best commercially demonstrated pollution
control equipment for certain air pollutants
on all major new industrial facilities.
In 1972, Amendments to the Federal Hater
Pollution Control Act established a system
of effluent guidelines for the Best
Practical Technology (BPT) and Best
-------
Available Technology (BAT) for control of
the most common water pollutants for
each major industrial category. Last
year the Environmental Protection Agency
was ordered by a U. S. district court to
apply, on an accelerated schedule, Best
Available Technology standards to the con-
trol of 65 toxic materials in the waste-
waters from 21 priority industries.
Finally, six months ago the House of
Representatives voted to require best
available control technology on all new
utility boilers in order to minimize the
atmospheric loading of sulfur oxides and
their by-products, to increase the use of
locally available coals, and to give some
room for industrial growth in most parts
of the country.
In short, we are beginning to control
air and water pollution wherever and when-
ever it can be controlled. And the
scientific information on health and eco-
logical effects of such pollution has
consistently supported the wisdom of this
trend.
Although we recognize the importance
of environmental protection afforded by
investments in control technology, there
still linger ambivalent feelings about
environmental control technologies. Do
we really need them? Aren't there other
ways of achieving the same objectives?
Aren't there breakthroughs waiting just
around the corner which will obviate the
need of our control technologies? Or, if
we shut our eyes tight and wish very hard,
won't the problems go away? My answer
is that control technologies are our
bridge, and our only bridge, into an
uncertain future.
Hence, there is no longer any question
whether, or how far, we will compromise
our environmental goals to achieve expanded
energy resources and industrial growth.
There need be no compromise. There shall
be no articifical dichotomy of energy and
industrial growth versus environmental
protection. We will have the growth we
need, and it will not be at the cost of our
property, our health, or our sense of
aesthetics. This is President Carter's
policy..
With this in mind, we have been
accelerating the development of controls
for four of the major problems, three of
them air pollutants—sulfur oxides,
nitrogen oxides, and fine particles—and
the fourth—mining disruption. While our
technology bridge to the future over each
of these problems is far from perfect,
at least we can now see the other shore.
As I mentioned earlier, we are also
looking into the future to assure that
advanced technologies, when they do
become commercially applied, do not pose
their own set of environmental hazards.
It would be a distorted perspective to
limit our attention to solving today's
problems while allowing tomorrow's to grow
unchecked. For example, in parallel with
efforts by the Department of Energy to
develop coal-fired fluidized bed combustors
for heat, steam, and power generation, EPA
is conducting a complete environmental
characterization of these processes. Our
goal is to help to avoid any potential
environmental problems that may be associ-
ated with this very promising technology.
On Wednesday, you will be given a descrip-
tion of the scope and preliminary findings
of this effort.
Synthetic fuels, as I have indicated,
also offer solutions to problems associated
with the use of coal, but may present some
new, and potentially serious, environmental
threat. Working with DOE, we will be moni-
toring the Country's early synthetic fuels
plants, as they become operational within
the next few years. Such efforts will
help assure that the best controls are
available at the lowest costs and at the
right time so we can meet our needs for
alternative supplies of liquid and gaseous
fuels. The details on these efforts will
also be presented in a session on Wednesday.
Since the initiation of EPA's
Environmental Assessment Program three
years ago, much progress has been made in
developing adequate sampling and analytical
techniques. Such technology is evolutionary
in nature, in that we improve on the tech-
niques as we get experience in their appli-
cation and as we attempt more complex
sampling and analysis. The people involved
in these efforts have a right to be proud
of their accomplishments. You who have
been involved in these efforts will hear
about them on Tuesday. There still,
however, are significant challenges for
improvements in methods, specifically
selectivity, sensitivity, accuracy, and
lower cost.
This symposium provides an important
formum for those applied researchers who
are developing and using process measure-
ments for environmental assessments to
exchange ideas. Your findings and
accomplishments will continue to guide the
Environmental Assessment Program so that
we can assure that the Nation's energy and
industrial processes will meet our present
and future environmental control require-
ments .
-------
THE DOE INTEGRATED ASSESSMENT PROGRAMS
R. D. Cooper
Division of Regional Assessments
U.S. Department of Energy
Washington, D.C. 20545
Abstract
The Integrated Environmental Assessment programs
of DOE are designed to support the Assistant
Secretary for the Environment in the conduct of
his responsibilities. In order to fulfill these
responsibilities, particularly those associated
with policy guidance for the Department, there are
three kinds of assessment activities which must be
carried out. Since the Department is responsible
for developing a national energy policy, national
environmental assessments of this policy are
required. Technology assessments must be done in
order to explore environmental Issues which affect
technology development decisions and the levels of
environmental control required. Finally, regional
environmental assessments must be carried out
because environmental and social Impacts of energy
policies and technological developments are very
dependent upon the specific characteristics of
the region Involved. Examples of all three types
of assessment were recently completed in connec-
tion with the National Energy Plan as submitted
to Congress. A national assessment of the environ-
mental impacts associated with the plan was made
using a comprehensive simulation model based on
the SEAS system calibrated to the HEP assumptions
for the period 1975 to 2000. Regional assessments
were made of the impact of the plan on New England
and on Region VI where conversion from oil and gas
to coal is expected. Finally, technology assess-
ments were made of the prospects for solar energy
and of the Impacts of the coal solid waste, and
local socloeconomlc well-being.
Introduction
The Department of Energy has broader responsibil-
ities than its predecessor agencies and the
Integrated Assessment Program under the Assistant
Secretary for Environment has also changed and
broadened its scope. The assessments supported
under this program are aimed at three types of
policy questions which basically determine the
different thrusts of the studies. First, there
is a need to review and assess national energy
policies and strategies for environmental impact.
This will assist in the development of a national
energy policy which is environmentally acceptable.
Secondly, there Is a need to assess the health,
environmental, and social Impacts of new tech-
nologies being developed by DOE. These technology
assessments will provide guidance on the rate and
direction that development should take and the
environmental controls which must be built in.
Finally, since most impacts on the environment as
well as on man's social well-being are local, It
is necessary to do regional assessments of the
impacts of alternative future energy strategies.
This will provide analyses of regional issues which
must be taken into account in national energy
planning as well as providing options for
mitigating regional impacts and removing
constraints to further energy development.
These three kinds of assessments are not mutually
exclusive but, in fact, often overlap in any
particular study. For example, a National Coal
Dtilization Assessment now being completed is a
regional study of the environmental effects of
Increased coal use which will be integrated into a
national picture. An assessment of the impacts of
geothermal development in the Imperial Valley of
California is both a study of the technology and
the region. Future studies, however, are expected
to fall more clearly in one of the three categories
because of the questions which are being asked of
the assessments.
The data and information needs of the three types
of assessments differ. The national assessments
will be done annually using a series of models
which calculate energy residuals by region after
a dlsaggregation based on economic activity. Since
the models do not yet calculate Impacts on eco-
systems or health effects, the level of
sophistication of the data need not be as high as
in the other two types of analyses. The technology
assessments require the best physical and chemical
characterization which can be made of the effluents
from each part of the fuel cycle. It is here that
much better Information is needed. The regional
assessments also require careful measurements of
the present environment so that calcuatlon of
impacts of future energy options can be made.
A National Assessment of the President's Energy
Plan
The first Annual Environmental Assessment Report
has been completed to determine the Impacts of the
National Energy Plan on the environment and to
compare this to a base case scenario without the
NEP initiatives. The analysis was made to deter-
mine if there are potential environmental problems
associated with the plan, to develop the method-
ology for an annual update of the analysis, and to
assist in setting environmental research priorities.
A large part of the Input information used in this
type of assessment consists of economic and demo-
graphic assumptions. The principal energy and
environmental data developed included residual data
on a national and regional level and cost data on
pollution abatement for both energy and nonenergy
supply technologies. Residuals and abatement costs
associated with Industrial, transportation, com-
mercial and residential demand activities were also
used. Data bases for energy supply sectors such as
those collected by Hlttman Associates, Teknekron,
EPA, etc., formed the basis for the residuals
analysis.
-------
The results of this assessment Included the Iden-
tification of a number of key environmental Issues
and the regions where these issues are of primary
concern. Figure 1 shows how SOX emissions vary in
several regions between 1975 and 2000 based on the
two scenarios. In general, the NEP scenarios is
seen to result in smaller SO* emissions because of
the Increased conservation component. Figure 2
shows the water consumption by energy technology
according to NEP and pre-HEP scenarios, 1975, 1985,
and 2000. This indicates clearly the very large
increase in the use of water for energy by the
year 2000.
The requirements for new measurements and much more
sophisticated data in this type of analysis are less
than in the other kinds of assessments to be des-
cribed here. The modeling techniques used and the
questions being asked of the analysis make it un-
necessary to know energy emissions in detail much
beyond the criteria pollutants. As experience Is
gained with the national assessments, more data of
the kind needed in the technology and regional
assessments will be required.
Technology Assessment
The second major type of assessment carried out
under the Assistant Secretary for Environment,
DOE, Is a technological assessment. These studies
are principally concerned with new technologies
being developed by the agency but include an
analysis of impacts of the present and possible
future Installed energy Industry.
There are three major uses within DOE for these
technology assessments. First, they are used to
identify present and future environmental control
needs. In order to help specify future environ-
mental controls, needed environmental and health
research and development Is identified by these
assessments. Finally, they are used as Input for
both technology development planning and environ-
mental development planning.
The data and information needs for a technology
assessment are much more sophisticated than for
the national assessment described above. Even
though the technology is still being developed,
it is necessary to characterize the fuel cycle.
The physical and chemical characteristics of the
residuals from each element of the fuel cycle must
be obtained either by measurement or from theory.
Finally, the transport and fate of the effluents
within the environment must be understood.
As an example of a technology assessment, the at-
mospheric fludized bed combustion of coal will be
described. This study was done approximately one
year ago, so the numbers shown should not be
considered representative of our present knowledge.
Figure 3 shows the different elements Involved in
atmospheric FBC and the streams connecting the
elements and the external environment. The effluent
streams are listed in Table 1 since these will be
responsible for the principal environmental impacts.
Resource use (land, water, etc.) must be considered
in addition to the effluents.
9.8.
19J5 1985 2000 1975 1985 2000 1975 1985 2000 1975 1985 2000
UCION 3 IECIOH 4 ICCION 5 UCIOM 6
ncou 1. cotffARisoN or som EMISSIONS
•E1WEN THE HEP AMD PU-NET SCENARIOS
(FOR SELECTED REGIONS)
-------
12000
10000 —
MOO —
6000 —
4000 —
2000 —
1*7}
ItU
2000
FICOB 2. B*m OJMUMPTIOB BY E«MCY WOWOIOCY ACCOtDDB TO HEP AJTO
FIE-BF SOIAKIOS. 1975, 1985. «ad 2000
FICDB 3. AWD8FIBUC IBC OF GOAL
0
-------
TABLE 1
Emonrr STREAM DESIGNATIONS
ATMOSPHERIC AMD PRESSURIZED FBC
STREAM HO. DESICHATION
1 Stack Caa
2 Particular Removal Discard
3 Bad Solids Discard
* Partlculate Removal Discard—Regeneration Operation!
5 Other Effluents from Regeneration and Sulfur Recovery Operations
6 Slowdowns from Steam Turbine Cycle
7 Slowdown from Water Treatment Operations
8 Product from Sulfur Recovery (Sulfur or SulfurIc Acid)
21 High Carbon Ash from Partlculate Removal (recycle or
alternate eombuator)
Table 2 lists the trace metal discharges from
fluldlzed bed combustion and from flue gas desul-
furizatlon. At the time that this assessment was
done, there was no data on trace elements from
FBC, so these are listed as unknown.
Table 3 lists the waterborne effluents discharged
per day by a 500 megawatt fluldlzed bed combustion
unit and by a similar conventional combustion plant
with flue gas desulfurization, each operating at
75 percent load factor. The FBC unit is seen to
have, In general, a lower release of effluents.
Table 4 shows the results of the calcuatlon of
public health impacts for a 1000 megawatt electric
plant using fluidized bed combustion. Three dif-
ferent analytical models were used: 1) assuming new
source performance standards in S02 and particu-
lates, 2) in partlculates alone and finally, 3) in
a conventional combustion plant with a FGD unit
attached.
From this kind of technology assessment, several
research needs become apparent. For example, the
following questions regarding trace elements must
be answered:
1) Are all elements present in the coal emitted
into.the effluent streams?
2) What are the portioning factors for the elements?
3) What is the chemical form of the emitted
element?
4) Is the deposition pattern different from that
In conventional combustion and with FGD?
Questions associated with the waste product from
FBC units can take the following forms:
1) What chemical form of the trace and major
elements end up in the fly ash, in the bottom
ash and in the slag?
2) What trace elements are abosrbed on the spent
sorbent?
3) How effective are the settling pool liners?
4) What reclamation procedures can be used for
the waste ponds?
5) What are the environmental effects of ash used
In roads and building materials?
Finally, looking at the calculated health effects
we see a number of further questions which are
raised by this assessment. These Include the
following:
1) What is the composition and chemistry of the
effluent stream from the stack?
2) What atmospheric chemistry takes place after
the release of the effluents?
3) Are there synergistic effects between the
effluents and other materials in the environ-
it?
4) Finally, what are the damage functions of the
principal effluents from FBC units?
All of these questions will result in further re-
search and development, the outcome of which will
feed into improved technological assessments. The
final results are used to set research priorities
for environmental studies and development priorities
for both the technology and the environmental
control system associated with it.
Regional Assessments
The final type of assessment activity carried out
under the environmental office of the Department of
Energy is the regional assessment. These are con-
ducted to characterize the environmental issues
associated with the future energy supply and demand
within a region. A regional entity will include
a river basin, an electric power pool, or a poli-
tical unit such as a state or county.
The principal purposes of these regional assess-
ments are to obtain a regional perspective on the
issues associated with national policy such as the
National Energy Plan. In addition, they are a
-------
TABLE 2
Tract M«t«l Dlachargea, FBC and FCD
Dlachargaa, Ib/day
Pollutant
41
Ho
Cd
S«
A*
B
Pb
B«
FBC
0
D
D
0
D
0
U
D
Conventional
1.2
0.05
0.01
0.05
0.01
0.05
0.01
0.07
0 - Unknown
TABLE 3
Uatarborne Effluent*. FBC and FCD
DlacharcM. IbAUy
Elaaent
TSS
OH and graaaa
*amnn1i nitrogen
•Itrata Blcro(«i
Chlorld*
f.». Chlorine
8ol£«t«
F«
Cu
Zo
Cr
F
8<
•1
Ml
sooow
rac.
75X Lo«d Factor
93 HX
28 avg
19 MX
14 r»f
0.92
0.17
91
9.2 ux
3.5 «vg
133
20
4.8
0.44
0.60
4.8
67
4.8
106
500-MI
Conventional.
75Z Load Factor
189 max
57 avg
38 max
28 avg
1.4
1.2
127
9.2 max
3.5 avx
270
U
4.8
0.47
0.62
4.8
143
4.8
116.0
-------
TABLE 4
Raaulta Fro* the Public Health lapact Analyst! for • 1000-Mfo Plane
Avcrag* Valuaa for
50-aila-radlua Circle
Highly lapaetad Subaraa
Expactatloa
of Ufa" (a.)
Daatha/10*
Daatha/10*
(aaallna Valuaa
(Barkahlra Co.. Maia.,
Whica Populacloo. 1970)
68. Ml 75.541 11.583
Modal
Aaauaptton
MSPS In 802. TSP
MSPS In TSP
68.831
68.847
75.529
7S.S28
10.705
6
6
Convantlonal
Coabuacion, FCD
1000 us/10* Itu POM* Mo daeaceabla dlffaranca from baaalloa
AFBC.
3000 u»/10' Ecu POX* 68.860 75.541 1 0
68.830 75.510 16
68.734 75.422 64
68.860 75.541 1
68.848 75.538 7
15
55
*Polycyellc Organic Hatarlal.
means to surface the state and regional environ-
mental concerns which can have an Impact on the
Implementation of national energy and environmental
policies.
The regional assessments compare or balance the
competing energy technologies and resources which
can be utilized in a given region. The various
technological options available in the near term
and in the future are carefully considered and dis-
tributed in order to minimize environmental Impacts
on the region.
These assessments are used primarily to assist
national energy planners to develop environmental
and energy policies which will be acceptable within
the different regions of the country. Regional
assessments are designed to compliment the tech-
nology assessments and thereby provide the analy-
tical support required for making policy choices
at all levels of governmental energy and environ-
mental planning.
The data and information required in a regional
assessment is generally of two types. First, It
Is necessary to know the supply and demand picture
of the region. This includes a knowledge of the
future energy options which are available. The
second major type of information required is a
baseline characterization of the environment of
the region. The kinds of ecosystems, the distribu-
tion of people, the distribution of natural re-
sources and water resources within a region are all
important in the analysis of future options and
the environmental concerns which will be generated.
The results of a regional assessment are often
given In terms of issues of concern to the region
which are raised by a certain policy or a parti-
cular scenario. For example, an analysis of the
Impacts of the National Energy Plan on Mew England
has identified a number of concerns centering on
questions of regional equity. These range from
concern over the degree to which coal conversion
initiatives addressed to the industrial sector will
accelerate the already serious outmigration of
Industry, issues associated with Interregional
transport of pollutants from fossil fuel, combs-
tlon, to Issues concerning the Imposition of taxes
triggered by conservation performance. This type
of analysis, therefore, provides a feedback
mechanism raising concerns of Importance in the
implementation of the policy being addressed.
In some cases, constraints are identified which
will result In an inability to meet future
energy or environmental goals. Such things as
shortages of water or class I air quality regions
will clearly constrain local development of
energy resources.
These and many other factors must be considered
in energy planning, and the regional assessments
provide the fine structure for such planning.
National assessments, technology assessments,
and regional assessments when taken together
give those responsible for developing energy
and environmental policies the Information
and analytical base needed in order to come up
with plans which can be implemented.
-------
RELATED EPRI PROGRAMS
Ralph M. Perhac
Program Manager, EPRI
INTRODUCTION
The Electric Power Research Institute
(EPRI), through its funding activities,
seeks to develop a coordinated electric
power research program. The Environmen-
tal Assessment Department of EPRI is
responsible for assessing the environ-
mental effects of electric energy pro-
duction. Actually, environmental studies
are an important part of the work of all
of EPRI's technical divisions. In fact,
nearly half of EPRI's budget is used
for environmentally-related projects.
The Environmental Assessment Department,
however, is the only EPRI group whose
primary function is assessing environ-
mental impacts. In order to fulfill its
charge, the Department is organized into
three programs: (1) Health Effects and
Biomedical Studies, (2) Ecology, and
(3) Physical Factors. The 1978 budget
for the department is approximately
$13 million, with an expected rise to
over $25 million in 1981 (Table 1). This
increase of nearly 100 percent in three
years exceeds that of EPRI as a whole,
the EPRI budget rising from about $190
million (1978) to nearly $275 million
in 1981, a 45 percent increase.
Obviously, concern over environmental
consequences plays an important role in
EPRI's thinking.
An important emphasis in EPRI's
funding is on problems related to the
use of fossil fuels (Table 2). Similarly,
the focus of the Environmental Assessment
is on atmospheric pollution arising from
coal-burning plants. The Department's
prime concern is with pollution effects
on biota, especially on man. The
Ecology and Health Effects programs
account for nearly two-thirds of the 1978
budget. The Physical Factors Program,
however, plays a significant role in that
it provides the information on the
identification of pollutants, their
physico-chemical nature and their fate,
i.e., it provides the measurements
needed for environmental assessment. The
importance of the Program is shown in the
budget allocations. The remainder of
this paper will deal entirely with the
activities of the Physical Factors
Program.
PHYSICAL FACTORS PROGRAM
OBJECTIVES
The concern of the Physical Factors
Program is with the physical and
chemical environmental consequences of
electric energy generation. More
specifically, the Program's objectives
are to define pollutant distribution
and the industry's contribution to
that distribution. It is the latter
aspect (defining the industry
contribution) which distinguishes the
Program from similar ones in
governmental organizations. To achieve
its two objectives, the Physical Factors
Program is divided into two sub-programs:
(1) Identification, Characterization and
Monitoring, and (2) Transport and
Interaction. Defining pollutant
distribution is principally the charge
of the first sub-program; defining the
industry contribution falls into the
second. Like the Environmental
Assessment Department as a whole, the
Physical Factors Program strongly
emphasizes studies related to air
pollution and coal burning. Although
present emphasis is on air pollution,
research is being directed increasingly
toward studies of pollutants in
terrestrial and aquatic environments
and on fuels other than coal.
In developing a research program,
attention has been directed to specific
pollutants and to some specific, broad
environmental problems (Table 3). At
present, the emphasis on specific
pollutants is on sulfur compounds,
nitrogen compounds, trace metals, and
organic pollutants (principally
polycyclic organics). The broad,
environmental problems receiving
attention are regional distribution of
pollutants, acid precipitation,
visibility, and disposal of solid waste.
The broad, environmental problems
cannot be studied independently; studies
of specific pollutants are essential to
an understanding of general problems.
For example, the program on acid
precipitation really has no meaning
10
-------
unless it is coupled to a program of
research on atmospheric sulfur species.
IDENTIFICATION, CHARACTERIZATION
AND MONITORING
This sub-program involves much more
than just identifying and measuring
pollutant levels. It comprises three
facets: (1) analytical validation,
(2) analytical development and
improvement, and (3) identification/
measuring. The analytical validation
focus is on evaluating existing
analytical techniques. One example is
a study by Radian Corporation (Austin,
Texas) aimed at evaluating the
magnitude of error in measuring
atmospheric sulfate concentrations with
the use of high-volume filters. The
specific error suspected was that
arising from the possibility that
sulfates might actually form (from
SO.) on the filter itself. Such
formation would, of course, result in
more sulfate being measured than
actually exists in the atmosphere. If
secondary sulfate formation is
significant, then existing data are
suspect, and the basis of many health-
effects studies would be without a firm
foundation. The EPRI-supported research
did show, indeed, that S02 on standard
glass-fiber filters will convert to
sulfates in amounts equivalent to 1.6
to 8.9 iig/m3. The problem is especially
severe for samples collected over short
periods of time (2-3 hours). Fortunately,
the problem can be minimized by
pre-treatment of filters with acid, by
using long sampling times, and
especially by use of teflon-coated
filters.
The sub-program has (and is) support-
ing a number of projects which deal with
developing new analytical techniques.
He have put over $400,000 into
developing means of identifying specific
POM isomers. A project at the University
of Tennessee has been successful in
using matrix isolation, coupled with
fluorescence and infrared spectroscopy,
for identification of a number of
isomers in simple, synthetic mixtures.
Work is continuing, but with natural
samples, including coal-conversion
liquids. He have also funded SRI
International for the development of
a truck-mounted lidar unit for
measuring atmospheric concentration of
SO,, NO2, and 03. Work is about to
start on fabricating an airborne lidar
for measuring particle distribution.
Long-range plans call for ultimately
developing a UV-IR unit which can be
airplane-mounted and which is capable
of analyzing the atmosphere for a wide
range of atmospheric gases and
particles.
Identification and measurement of
pollutants is an important part of the
sub-program. We are supporting
extensive studies on trace metals and
organics in the New York City
atmosphere (New York University
Institute of Environmental Medicine).
This study has already demonstrated
that the trace metal contribution to
atmospheric pollution from automobiles
has doubled since 1969, whereas that
from oil burning (homes, industry, and
power plants) has halved, now comprising
only 11 percent of the total trace metal
inventory in the atmosphere. Present
work in the project emphasizes organic
pollutants and their source attribution.
EPRI is also supporting, at the
University of Wisconsin-Milwaukee, some
very sophisticated (and pioneering)
research on speciation of particulate
sulfur compounds (phase identification)
through the use of such surface
techniques as SEM, Auger, ESCA, and SIMS.
Perhaps one of the most significant
voids in knowledge of the atmospheric
chemistry of sulfur is the lack of good
data on the chemical form of particulate
sulfur compounds. The EPRI-sponsored
work at Hisconsin should shed
considerable light on the subject and
should prove exceedingly valuable for
studies of SO2 oxidation to sulfates
and for inhalation toxicology research.
TRANSPORT AND INTERACTIONS
The sub-program on Transport and
Interactions is concerned with two
things: (1) reaction mechanism and
(2) fate of pollutants. Fate studies
are essential for assessing biotic
impacts. Elucidating reaction
mechanisms (and fate studies) is needed
for understanding such problems as acid
precipitation and for defining the
industry contribution to pollution
levels. Perhaps the most important
project in the Transport and Interactions
sub-program is the $6.25 million Sulfate
Regional Experiment (the SURE program).
SURE1s goal is to define the relation
between regional, ambient concentrations
of a secondary pollutant (e.g., sulfates)
and local emissions of its precursor
(e.g., SO?). Such information is vital
if an ambient sulfate standard is
11
-------
promulgated. To achieve its goal, SURE
involves an extensive monitoring program
of air quality at 54 ground stations
throughout northeastern United States
(Environmental Research & Technology).
The project also involves use of
airplanes for measuring air quality
(Meteorological Research Inc. and
Research Triangle Institute) and
development of a thorough emissions
inventory (GCA Corp.). Coordinating with
SURE are plume studies of SO2 conversion
(Battelle Northwest Laboratories), dry
deposition (ARAP), and measurement of
biogenic emissions of sulfur compounds
(Washington State University).
A newly developing focus for the
Transport and Interactions sub-program
is on acid precipitation. Some work is
being supported on measuring acid
precipitation (RPI) but research will
also be directed toward atmospheric
chemistry and the formation of acid
droplets in clouds. A new project
(Central Electricity Research Laboratory,
Great Britain) will attempt to study
changes in cloud water chemistry by
monitoring the British industrial plume
as it interacts with clouds over the
North Sea from England to a point of
rainfall in Norway. A second, newly
developing focus is on visibility
degradation. Specifically, we are
funding research on means of measuring
visibility degradation and on the
atmospheric chemistry which causes such
degradation.
is on individual species identification
and on reaction mechanisms. In the
area of broad environmental problems, we
already have a sizeable effort in
regional distribution studies and we
are now expanding our acid precipitation
and our visibility programs. In the
future, we anticipate considerably more
effort on solid waste disposal and
possibly on extensive studies of plume
model validation.
Many of the projects we support are
based on ideas or proposals which we
have received from the scientific
community. Without the assistance of
that community in program planning and
in evaluating projects, our achievements
would be considerably less than they are.
We are indebted to the enthusiastic
cooperation shown by our colleagues both
in the utility industry and in the
scientific community at large.
FUTURE DIRECTION
During the first few years of BPRI's
existence (1974-1977), much of the effort
in the Physical Factors Program was on
identification and characterization of
pollutants. Emphasis is now definitely
shifting to an understanding of the
physico-chemical reactions which occur
in the environment and the role of these
reactions in a number of broad
environmental problems. For example, we
are rapidly expanding our efforts in
atmospheric reactions of nitrogen oxides,
the goal being to use the information to
understand better the formation of acid
rain, similarly, our research on the
kinetics of SO, oxidation feeds into our
acid rain and visibility projects.
With regard to specific pollutants,
nearly all our effort has been on
sulfur compounds. We are now expanding
our nitrogen effort and we expect to
increase considerably our funding of
research on organic pollutants. Our
emphasis in studying specific pollutants
12
-------
TABLE 1
Budget - Environmental Assessment Department
($ millions)
Program 1978 1979 1980 1981
Health Effects 4.6 7.3 7.2 9.7
Ecology
3.2 3.0 3.8 3.8
TABLE 2
Budget - EPRI Technical Divisions - 1978
($ millions)
Energy
Fossil Fuel & Electrical Analysis &
Advanced Systems Nuclear Systems Environment
83.1
53.7 32.8
20.9
Physical
Factors
4.7 4.5 6.8 8.9
Total • 190.5 + 2.50 - 193
EPRI
193 202 232 274
* $2.5 million not designated to a specific
division
TABLE 3
Programmatic Interests - Physical Factors
Program
Sulfur Nitrogen Organic Trace
Compounds compounds compounds Metals
Environmental Problems
Pollutant Acid Solid
Distribution Visibility Precipitation Waste
13
-------
EPA Air Programs' Use of Environmental Assessments
Richard 6. Rhoads, Director
Control Programs Development Division
EPA Office of Air Quality Planning and Standards
Abstract
Under the Environmental Protection Agency's legislative authority, a variety of
strategies are available for controlling air pollution. Principal among these are direct
emission standards for specific new sources which represent the best demonstrated control
technology, direct emission standards for both new and existing sources to minimize
specific hazardous pollutants, ambient air quality standards for selected pollutants
which are necessary to protect the public health and welfare, mobile source emissions
controls and fuel additive standards which assist in protecting the public health and
welfare, and programs to prevent significant deterioration of air quality. This
presentation describes these various alternative strategies, and discusses how environ-
mental assessments can assist in (1) setting priorities for the pollutants and sources
to be controlled, (2) selecting the strategy or combination of strategies to be employed,
and (3) assessing the necessary or desired level of control.
Protection of the environment is a
very complex process, and environmental
assessment has a very important role in
that process. Pollution exists in various
media—air, water, solid waste, pesticides,
etc.—and is controlled through various
regulatory programs, either individually or
in combination.
For some forms of pollution, the
media and the appropriate regulatory pro-
grams are obvious. For others, some form
of environmental assessment is essential
to identify the significant media and the
most effective regulatory program or pro-
grams. For this purpose, the environment
assessment need not identify all the de-
tails of the problem—it is normally neces-
sary to identify only the broad aspects of
the problem.
In the case of pollutants for which
air is identified as the primary or at
least a significant medium, the appropriate
strategy within the air pollution control
program must then be identified. The cur-
rent air pollution control program consists
of two separate but interrelated broad
strategies. I will refer to these two
broad strategies as "National Emission
Standards" and "Air Quality Management,"
and will explain each separately even
though both strategies are mutually sup-
porting.
National Emission Standards
There are four major types of Nation-
al Emission Standardsi (1) Motor Vehicle
Emission Standards, (2) New Source Perfor-
mance Standards, (3) National Emission
Standards for Hazardous Pollutants, and
(4) Fuel Additive Standards. Each type of
National Emission Standard has advantages
and disadvantages, and environmental assess-
ments can assist in selecting the appro-
priate standard.
The Motor Vehicle Emission Standards
are essentially legislated by the Congress.
They are imposed directly upon the motor
vehicle manufacturer, and their primary
purpose is to reduce harmful vehicle ex-
haust to the lowest feasible level in order
to assist in managing air quality. In many
areas, particularly major cities, the cur-
rent motor vehicle emission standards are
not, by themselves, adequate to protect the
public health. In these areas, various
forms of air quality management must be
implemented to supplement the emission
standards.
The New Source Performance Standards
are developed by EPA on a national level to
ensure that the best demonstrated controls
(considering costs) are applied to all ma-
jor stationary sources of air pollution.
New Source Performance Standards represent
the maximum emissions that new stationary
sources are permitted to emit. A case-by-
case determination is also required to
determine if additional controls can be
applied to either protect the public health
and welfare, or to prevent significant
deterioration of air quality in clean areas
of the country.
The National Emission Standards for
Hazardous Pollutants (NESHAPS) are applied
to both new and existing stationary sources
of very specific pollutants which are
particularly hazardous to the public health.
NEBHAPS are applied to a relatively few
pollutants, generally emitted by a rela-
tively few sources, and represent very
stringent emission standards which are
necessary to solve very severe problems.
14
-------
The fourth major type of National
Emission Standard is the Fuel Additive
Standard. Up to this time, EPA has had to
restrict fuel additivies of only one
pollutant—lead—but it may become neces-
sary to exercise 'this authority for other
pollutants in the future.
National Emission Standards are
established by EPA as a national program
under which all sources, regardless of
location or air quality impact, are
treated essentially the same. The strategy
of National Emission Standards is very
effective in controlling air pollution,
but the strategy by itself is not always
adequate to protect the public health or
welfare, nor to prevent significant deteri-
oration of air quality in clean areas of
the country. Hence, in some cases, we
employ the second, interrelated, strategy
which I will refer to as Air Quality Man-
agement .
Air Quality Management
The Air Quality Management Strategy
also has four major facets: (1) Primary
National Ambient Air Quality Standards,
(2) Secondary National Ambient Air Quality
Standards, (3) Prevention of Significant
Deterioration of Air Quality, and (4) En-
hancement of Visibility. Air Quality
Management basically consists of defining
a maximum acceptable level of pollution in
the ambient air, and then controlling both
the emissions from sources and the location
of sources to ensure that the acceptable
level of pollution is not exceeded.
Primary National Ambient Air Quality
Standards are established to protect the
public health. They are based upon the
concept that there is a threshold level of
pollution below which the public health is
not jeopardized. That level is identified
from health effect studies, and the pri-
mary standard is then established with an
adequate margin of safety. The original
primary standards were to have been attain-
ed by mid-1975 in most areas. Although we
made substantial progress toward attainment,
many areas still have air quality worse
than the primary standards. The Clean Air
Act of 19-77 now requires attainment by
1982 with, in some cases, the possibility
of extensions to 1987.
Secondary National Ambient Air Qual-
ity Standards are established to protect
the public welfare. They are based upon
the same threshold concept as primary
standards, but the effect being protected
against is welfare related such as mater-
ials damage, soiling, crop damage, etc.
Although for some pollutants the primary
and secondary standards are identical,
secondary standards are generally more
stringent than primary standards. Also,
whereas primary standards are tofbe
attained by a specific date, secondary
standards have no statutory deadline for
attainment so long as they are attained
within a "reasonable" time.
Prevention of Significant Deteriora-
tion is the third major Air Quality Manage-
ment program. It was established to pro-
tect the relatively clean areas of the
country from becoming dirtier. Unlike the
ambient standards which are based upon a
threshold concept, significant deteriora-
tion is based upon an incremental deteri-
oration above some pre-existing level of
air quality.
What constitutes "significant" de-
terioration, however, may be different in
some areas of the country than in other
areas. Therefore, the local people may
classify their areas into one of three
classes. A Class I area is intended to
have virtually no deterioration. A Class
II area (which is the most common classifi-
cation) may permit moderate deterioration
without that deterioration being "signifi-
cant." The Class III classification is
reserved for areas in which air quality may
not deteriorate to the level of the the
health or welfare related ambient standards.
At this time, no areas have been classified
as Class III.
The fourth major Air Quality Manage-
ment program is brand new. It was estab-
lished by the Clean Air Act of 1977 and is
intended to protect and enhance visibility.
This program is still a year and a half
from initial implementation, because much
research and regulatory development work
must still be done. It will probably not
have much impact in the heavily industrial-
ized areas of the nation, but I anticipate
a significant impact on the large isolated
sources of air pollution in the West.
Summary
As you can see, the air pollution
control program is very complicated. It
has two basic strategies—National Emission
Standards and Air Quality Management. Each
of these strategies has four basic com-
ponents providing a total of eight major
regulatory options through which, either
individually or in combination, air pol-
lution can be controlled.
The choice of options can be diffi-
cult, yet making the correct choice is
essential to the success of the air program.
The other environmental programs—water,
solid waste, radiation, toxic substances,
etc.—also have a variety of options avail-
able for achieving their goals. From my
perspective, the main benefit (although by
no means the only benefit) of environmental
assessment is to assist in determining and
prioritizing the goals, selecting the
appropriate regulatory program or programs,
IS
-------
and then selecting fr-qn> the. yar-ous
the most effective methods for achieving
the goals.
This is a very important task. When
we speak of environmental programs we are
speaking of billions of dollars annually,
and we are speaking of the health, welfare,
and quality of life of everyone in the
nation. I encourage you in your endeavors
during this symposium, and I wish you great
success.
-------
AN INTEGRATED APPROACH TO ASSESSMENT
AND CONTROL OF INDUSTRIAL POLLUTION PROBLEMS
by Eugene E. Berkau* and Alfred B. Craig. Jr.,
Industrial Environmental Research Laboratory,
U.S. Environmental Protection Agency,
Cincinnati, Ohio 45268
Abstract
EPA now has comprehensive legislative and
court mandates for regulating all pollution result-
Ing from the production and use of Industrial chem-
icals and products. Emphasis 1s being placed on
toxic and hazardous pollutants. To aid EPA 1n
focusing Its attention and future efforts on those
pollution problems which have the greatest health
and ecological Impacts, EPA has developed 1n one
organizational unit an approach for assessing
simultaneously the environmental Impacts of Indus-
trial pollution discharged to the air, water, land
and municipal systems. This multimedia approach
necessarily Involves the active participation of
the appropriate EPA regulatory components and
Industry. Field sampling and RD&D programs are
Initiated based upon the outcome of the Integrated
Industrial assessment approach.
Introduction
In the last several years, there has been a
major change 1n emphasis 1n the Agency's environ-
mental pollution control programs. This has
resulted principally from recent legislative and
court mandates. This change began with the sign-
Ing of the Resource Conservation and Recovery Act
of 1976 (PL 94-580} and the Toxic Substances
Control Act also of 1976 (PL 94-461). These
legislative Acts, along with the Clean A1r Act
Amendments of 1970 (PL 91-604) and the Hater
Pollution Control Act Amendments of 1972 (PL 95-
500), provided EPA with comprehensive legislation
to address all aspects (I.e., air, water, land and
exposure through consumer products) of environ-
mental pollution. In addition to this comprehen-
sive legislation, there has been, 1n the last two
years, Increased emphasis on specific toxic and
hazardous pollutants. This, of course, has been
brought about by the signing of TSCA and, more
Immediately Impacting, the court settlement 1n
1976 between EPA and the National Resources
Defense Council (NRDC) requiring EPA to develop
effluent guidelines for 65 pollutant categories
suspected of toxic and hazardous properties which
may be discharged from 21 Industry categories over
the next three years. In addition to these man-
dates, the Clean A1r Act Amendments of 1977 also
placed more emphasis on specific hazardous pollu-
tants such as arsenic, lead, cadmium, polycycllc
organic materials and required an Increased rate
of regulation for the conventional criteria pollu-
tants. The Clean Water Act Amendments of 1977
supported the specific pollutant emphasis and the
requirements of the NRDC/EPA Consent Decree of
1976. The net result was to place greater empha-
sis on Industrial sources of pollution and the
specific processes which generate known or poten-
tially hazardous and toxic pollutants, regardless
of whether they pollute the air, water, land, or
are discharged to municipal systems.
^Speaker
Previous activities within the Agency have
focused on generic or surrogate pollutant control
of the criteria pollutants, NOX. SOX total particu-
late, total 'hydrocarbons and, similarly, BOD, COD,
pH, suspended solids. Little or no data base has
been developed by EPA or Its precedessor agencies
on the characterization and control of industrial
discharges considering specific pollutants or
chemicals thought to be toxic or hazardous. Con-
sequently, to meet the mandates of the recent
legislation and court settlements, EPA's regulatory
and R&D offices have Initiated extensive national
efforts to define and assess the industrial pollu-
tion problems created by specific toxic and haz-
ardous pollutants. These efforts by EPA's regula-
tory offices have been undertaken using the tradi-
tional approaches, that 1s, along single media
lines with the appropriate air, water, and solid
waste office addressing its own limited portion of
the total Industrial problem. As a result of an
organizational change 1n EPA's Office of Research
and Development 1n 1975, research's approach to
addressing Industrial pollution problems deviated
from this conventional approach.
In 1975, the reorganization of the Office of
Research and Development resulted in the formation
of two laboratories, the Industrial Environmental
Research Laboratory 1n Cincinnati, Ohio, and the
Industrial Environmental Research Laboratory in
Research Triangle- Park, North Carolina. These
laboratories were provided unique charters to
address Industrial pollution problems. Specific-
ally, they were charged with the responsibility to
assess and to develop control methods for the total
gaseous, liquid, and solid waste pollution problems
of Industry. This represented the premiere, within
the Agency, of two organizational units with the
responsibility to address the total Industrial
pollution problem and to place their emphasis on
an Integrated or "multimedia" solution to the air,
water, and solid wastes pollution problems of
Industry. To respond to this unique and broad
responsibility, 1t was necessary for these labora-
tories to develop systematic and comprehensive
approaches for examining Industries, their pro-
cesses and practices, and concomitant pollution
problems.
This paper 1s a brief description of the
approach which evolved at the Industrial labora-
tory 1n Cincinnati from consideration of the
multimedia pollution problems of the nonferrous
metals production and Inorganic and organic chemi-
cals Industries. A specific discussion of the
nonferrous metals assessment activities 1s sched-
uled for later in the program. Consequently, this
presentation will be limited principally to the
general approach which has been entitled "An
Integrated Multimedia Approach to Assessment and
Control of Industrial Pollution Problems."
17
-------
The Multimedia Approach to Environmental
Assessments - An Overview
The multimedia approach was developed to
assure full utilization of available Information
sources, principally the open literature, govern-
ment reports, and other Information generated In
the Government's enforcement and regulatory
activities. It relies heavily upon site visits
and the use of nongovernment expertise Including
consultants and Industrial technologists. Field
sampling and analysis of Industrial process waste
streams 1s not a part of the Initial Industrial
assessment. Rather, the results of the assessment
provide a comprehensive and scientific basis upon
which appropriate research and development and
field sampling and analysis programs can be planned
and Implemented.
Goals of the Assessment
The following are major products or results
desired through Implementation of the "multimedia"
approach: (1) to Identify the most serious Indus-
trial pollution problems based upon potential
health and ecological Impacts regardless of the
type of discharge. I.e., to the air, water, land,
or municipal systems; (2) to help the Agency con-
centrate Its limited resources, particularly 1n
the future, on the more serious Industrial pollu-
tion problems; (3) to establish an EPA understand-
ing, from a pollution control perspective, of
Industrial processes, operations, and practices to
facilitate the development and Implementation by
Industry of pollution control methods; (4) to
establish government-Industry working relationships
to solve these Industrial pollution problems
through technological applications; (5) to develop
pollution control technologies that minimize
secondary pollution effects, that 1s, avoid trans-
fer of the pollution problem to another media, and
finally (6) to provide a basis upon which a more
efficient and scientific approach can be taken to
developing experimental data on the characteriza-
tion and control of Industrial pollution problems
for (a) regulatory purposes or (b) for better
definition of pollution problems and their control.
Products of the Environmental Assessment
The Initial multimedia Industry assessment
results 1n three major products: (1) a preliminary
data base on the Industry Including companies
Involved, production sites, products, processes,
discharges, pollution control technologies, asso-
ciated health effects; (2) the verified data base,
derived from a detailed analysis and extensive
expert review supported by site visits of the pre-
liminary data base; and (3) the Initial assessment
report Identifying the most serious Industrial
pollution problems that can be established from
the available literature and the major Information
gaps that exist 1n defining the total Industrial
pollution problems.
Development of a Preliminary Data Base
The development of a preliminary data base Is
composed of four major tasks, shown In Figure 1.
These are (1) the definition of the Industry,
(2) Identification of the processes and associated
discharges to the environment, (3) the assembling
of available health and ecological effects data
associated with the Industry, Its discharges or
components of the discharges, and (4) a review of
the preliminary data base by Industry and health
and ecological effects experts. The overall pur-
pose of this Initial effort 1s to be complete and
comprehensive 1n obtaining, from the open litera-
ture and other available sources, all Information
that relates to the specific Industrial products,
processes, and associated pollution problems.
The first three of these tasks are often conducted
simultaneously.
Industry Definition
The Industry definition task establishes the
Identity of an Industrial category, Unking
together competitive companies producing similar
products. The approach attempts to consider
actual Industrial conditions and operations.
Information 1s assembled, however, 1n such a way
that 1t can be cross-referenced to standard Indus-
trial classification codes (SIC) for use 1n con-
junction with economic analyses and related to the
Agency's regulatory and enforcement activities.
Key factors 1n defining an Industry Include the
companies Involved, raw materials consumed, pro-
ducts and their uses, overall process flow sheets,
capacities and actual production, trends and fore-
casts with emphasis on growth, new production pro-
cesses, economic status, and environmental Impacts
of the overall operations.
These basic factors are analyzed and devel-
oped so as to emphasize their relationship to
environmental control and to facilitate assessment
of the economic Impact of potential regulations.
For example, analyses of raw materials and pro-
cesses provide Information that can aid 1n deter-
mining the effects changing feedstock composition
and processing conditions may have on environmental
pollution. A complete product list 1s particularly
useful 1n that 1t defines a limitation on the
boundaries of an Industry. Its technologies and
the related environmental Impacts. These lists,
developed during the Industry definition phase,
lead to a family of processes by which the product
materials can be made. Information concerning
Industrial trends 1s also helpful especially when
the use of alternative feedstocks may alter the
environmental Impact of an Industry's discharges.
Identification of Processes and Discharges
Process flow sheets depicting the Industry as
a series of Interrelated modules Is an Important
tool. It 1s used to Interrelate all process
operations from raw materials, Including their
production for Integrated Industrial operations,
to the final production of goods and to Indicate
all discharges to the air, water, land, or munici-
pal systems. Each production process Is described
1n the following terms:
Function of each unit operation
Feed materials
Operating conditions
Utility requirements
Waste streams
Pollution control technologies
Occupational and environmental
health effects
Reference materials
Related SIC codes
18
-------
IDENTIFICATION
OF PROCESSES
AND DISCHARGES
ANALYSIS OF
AVAILABLE HEALTH
AKO ECOLOGICAL EFFECTS DATA
SITE
VISITS
ENGINEERING
ANALYSIS
»
HEALTH AND
ECOLOGICAL EFFECTS OF
INDUSTRIAL DISCHARGES
CONSULTANTS
PROGRAM DEvaOPMENT:
•DATA BASE GAPS
-RID NEEDS
-DEMONSTRATION NEEDS
Figure 1. Industry Assessment and Program Development Protocol
This task Identifies potentially hazardous
discharges warranting detailed study and puts the
Industry's multimedia or total pollution problems
Into the proper perspective. When data are non-
existent or limited to conventional pollutants.
preliminary estimation of specific pollutants can
be made through limited engineering analysis.
Studies by Industry, EPA. Department of Interior.
Department of Commerce, and the Department of
Energy can provide much of the data on Important
processes. Trade journals. Department of
Commerce patents, key word literature searches,
and standard reference works such as the Encyclo-
pedia of Chemical Technology (K1rk Othmer), Chemi-
cal Process Industries (Shreve) and Textbook of
Industrial Chemistry (Rlegel) are used to obtain
Information on lesser known processes. It 1s also
Important 1n this step to make a preliminary eval-
uation of foreign process and control technologies
under development or 1n actual use. Such Informa-
tion 1s useful 1n assessing trends and potentially
significant process changes.
Review of Available Health and Ecological Effects
Data
Assembling existing mortality/morbidity data
1s done principally to aid in establishing the
priorities for Industrial processes or discharges
requiring further work. Consideration of the ratio
of death to population (regardless of cause) and
relative Incidence of disease 1s made with the
Intent of Identifying correlations between specific
Industrial activities or processes and health
effects. This 1s conducted 1n cases where Infor-
mation 1s available and the proliferation of plant
sites does not prevent the Isolation of the effects
of the plant or process 1n question. Toxldty
data on materials, additives, products and emis-
sions from production processes are searched
through computerized data bases such as the
National Library of Medicines (NLH), Toxllne and
Medllne and related National Institute of Occupa-
tional Safety and Health and EPA health files.
Any Industry-associated compounds known or sus-
pected to be carcinogenic, teratogenlc, or muta-
genlc are given special attention. Substances
known to have high potential for producing chronic
toxlclty are also Included. Industry-associated
19
-------
compounds appearing on other lists of toxic sub-
stances such as those maintained by NIOSH or
structure activity relationships that Imply sig-
nificant toxic potential are also Included 1n this
survey. The Information generated 1n this task
can also serve as a preliminary definition of the
scope of required chemical analysis of emission
streams 1n future Industry field sampling and
analysis programs.
The literature review may be followed by a
retrospective ep1dem1o1og1ca1 analysis to Identify
actual health Impacts associated with the opera-
tions of an Industry. The literature reviews
serve to Indicate the type of disease-specific
mortality that should be examined for a represen-
tative number of production processes 1n the In-
dustry. Using Health, Education and Welfare
Department (HEW) data, mortality profiles for
counties exposed to a given Industry discharges
can be compared to the mortality profiles of sur-
rounding counties. The selection of diseases used
1n mortality profiles 1s' based upon available data
concerning known target organs of toxic materials.
Plants must have operated for a sufficient period
of time to allow for the latency periods for
chronic diseases such as cancer. The most speci-
fic mortality data are available In the HEW
publication. "U.S. Cancer Mortality by County,
1953 through 1969." Another principal source of
data, "Vital Statistics of the United States,
1971. Volume II - Mortality," has the advantage
of not being limited to cancer mortality.
Although this type of retrospective epidemic-
logical analysis 1s an extremely useful tool for
establishing priorities and Identifying the most
potentially serious Industrial pollution problems,
1t cannot conclusively establish cause and effect
relationships. However, occupational diseases are
often found to be an early warning of related
environmental health problems which may not surface
until many years later. On the other hand, the
difficulty 1n separating health effects directly
related to Industrial pollution from general urban
pollution will often preclude any utilization of
this type of analysis. These studies simply serve
as a mechanism for focusing the emphasis on sus-
pected problem areas and for Identifying where
field ep1dem1olog1cal studies may become a neces-
sary part of a recommended research program 1n the
future.
Data collected in the first three tasks are
assembled 1n a format, which allows convenient and
consistent display of pertinent process and dis-
charge data for Individual and comparative review
and analysis. One convenient format 1s that used
in the EPA publication entitled, "Industrial
Profiles for Environmental Use" (EPA-600/2-77-023).
The assembled Information 1s then submitted for
review by Industry and health experts to Insure
that no significant omissions of processes, dis-
charges, or related health and ecological data have
been made and that the Industry 1s correctly
defined. Upon completion of the review, the pre-
liminary Industrial data base 1s considered com-
plete. The Importance of this phase and the need
for completeness 1n assembling available Informa-
tion 1s stressed as all future work 1s dependent
upon the validity and completeness of the prelim-
inary data base.
The Importance of building and encouraging a
working partnership between EPA and Industry early
in the assessment activity cannot be overempha-
sized. This relationship 1s essential if the pre-
liminary data base 1s to represent real world
industrial boundaries, processes, practices and
pollution problems.
Development of a Verified Data Base
The next phase In the initial industry assess-
ment leads to the verified data base and consists
of two main tasks: (1) engineering analysis of the
preliminary data base on production processes sup-
ported by site visits to specific plants and more
directed analysis of health and ecological effects
data; (2) review of the product of the engineering
analysis by practicing experts 1n the industry.
Engineering Analysis
The detailed engineering analysis of the pro-
duction processes of an industry is performed for
three principal reasons: (1) to Identify obscure
discharges; (2) to determine 1f closer attention
should be given to a unit operation traditionally
Ignored 1n the past or considered only to have
limited environmental impact; and (3) to predict
the quantity and character of the discharges
associated with each unit operation based upon
materials processed, production capacity and pro-
cess conditions. The engineering analysis can
often be focused by health and ecological data
that can be associated with the production pro-
ducts, plants and/or their processes. When process
discharges are poorly defined, however, and little
health data exists, a step-by-step analysis of the
process Is done to ferret out unidentified but
potentially hazardous discharges. This analysis
must be supplemented with site visits by qualified
personnel to fill gaps in process descriptions and
to gather data on operating conditions and factors
that affect pollutant discharges.
Expert Review
The results from the engineering analysis and
the expert review provide a verified data base upon
which the air, water and solid waste pollution
problems can be assessed to Identify those that are
the most serious. In addition, the data base 1s
structured in such a way that 1t can now provide a
site specific basis upon which Industrial processes
and operations can be examined comprehensively and
scientifically. It can now provide the basis upon
which field sampling and analysis programs can be
designed to generate technically sound data suit-
able for developing regulations and control tech-
nology programs. The verified data base Is
assembled 1n a manner that 1s conveniently handled
manually or stored for computer manipulation but 1n
either case can be updated continuously. The need
for computer applications 1s based primarily on
the complexity of the Industry.
Preparation of an Initial Assessment Report
Impact Analysis
The final step before development of appro-
priate R&D and field sampling and analysis programs
20
-------
1s the Impact analysis or assessment of the total
environmental pollution problems created by an
Industry's discharges to the air, water, land or
municipal systems. This analysis 1s principally
an In-house EPA effort Involving the active par-
ticipation and support of EPA health, ecological,
technology, regulatory and enforcement personnel.
The purpose of the Impact analysis 1s to determine
the most serious public health hazards caused by
Industrial pollution considering proximity of the
discharge sources to population centers, transport
of pollutants, health hazards presented by the
pollutants versus the available control technolo-
gies. Commercial and prototype control technolo-
gies are examined for their technical and economi-
cal feasibility to control both conventional and
specific hazardous pollutant discharges. Particu-
lar emphasis 1s directed toward Insuring that
pollution problems are not transferred from one
media to another. Consequently, ultimate disposal
of collected wastes with hazardous properties 1s
given priority consideration. Priority 1s also
given to control of those processes and discharges
for which hazardous properties are conclusively
Identified or strongly suspected. Experience
Indicates that this Initial Industry assessment
will normally Identify those control technologies
that have been applied to many waste streams
before their effectiveness to control specific
pollutants has been demonstrated. Quanltative
characterization of many waste streams will also
be limited to surrogate pollutants. Because of
such data gaps, the Initial assessment can be
expected to Identify more unknowns regarding
sources of particularly harmful pollutants and the
applicability of control methods than demonstrated
solutions to problems. In spite of this limita-
tion, however, 1t provides a comprehensive data
base for Identifying the most serious Industrial
pollution sources upon which future field sampling
and analysis and RD&D programs can be focused.
Critical Expert Review
Before the assessment results are used 1n
guiding future program development activities, the
results of the Impact analysis are subjected to a
final, highly critical review. Reviewers Include
representatives of the Industry, related Industry
associations, EPA's regulatory and enforcement
offices, EPA research and health effects labora-
tories, consulting engineers, public health orga-
nizations, and universities. The purpose of this
review 1s to Insure that the program development
activities to follow are based upon an accurate
and complete a basis as possible.
Resources Required for
Environmental Asessments
The resources required for conducting an
Initial Industry assessment are shown 1n Figure 2
and normally range from 500 to 2,000 professional
manhours, depending upon the number, complexity
and obscurity of the processes used by the Indus-
try. This range 1s based upon experience to data
which has shown that approximately 20-50 profes-
sional manhours are required to examine a single
production process. The major expenses are associ-
ated with the engineering and retrospective epi-
demlologlcal analyses and development of Individual
process descriptions. Typically, Initial Industry
assessments conducted to date have been completed
1n six months to one year.
Applications of the Integrated Approach
The Initial Industry assessment and verified
data base are just the beginning of the efforts
necessary to define and control industrial pollu-
tion problems. However, they can form a basis
upon which research and regulatory programs can be
coordinated, planned, and Implemented to address
the total Industrial pollution problem. While the
"multimedia" approach has been applied only to
research activities to date, there has been some
utilization for structuring and coordinating
related research and regulatory sampling and
analysis programs. Specifically, the verified
data base for the industrial organic chemicals
Industry, termed the Organic Chemical Producers
Data Base (OCPDB) has been used to develop field
sampling programs for both air and water regula-
tions development for this industry. The OCPDB
was used to Identify where specific products and
processes to be regulated were employed and to
specify the minimum number of production sites
necessary to be sampled to cover all combinations
of products and processes. The computerized data
base also provided alternative locations should
the original sites not be convenient for sampling.
The multimedia assessment approach and the
same Organic Chemical Producers Data Base are also
being used to structure a proposed program to sup-
port EPA's Office of Toxic Substances in deter-
mining the total human exposure potential, through
process discharges and direct contact from the
production and downstream use, of the major 400
U.S. Industrial organic chemicals. In addition,
the approach will be used to assess the electro-
nics and mechanical products Industries. Based
upon the data bases derived from the Integrated
assessment of these highly complex and diverse
Industries, a statistically sound field sampling
and analysis program will be developed and Imple-
mented. The results of this effort will provide
the technical basis for establishing pretreatment
and effluent guidelines for the most serious
sources of the 65 pollutant categories defined in
the 1976 NRDC Consent Decree.
All of these programs will be used to validate
and refine the Integrated assessment approach. If
successful, it could provide guidelines upon which
the data necessary for research purposes and for
air, water, solid waste, and toxic substances
regulations can be obtained simultaneously in
cooperative EPA field sampling and analysis pro-
grams. An ongoing study supporting this concept 1s
an attempt to Identify all of the Agency's known
data requirements for assessing and regulating
Industrial pollution discharges. The purpose 1s
to establish a compendium of all common Industrial
and process waste stream data required for each
activity.
Conclusions
These efforts represent but a few of those
ongoing which utilize the Integrated approach to
assessing Industrial pollution problems. The
results will determine to a large extent the
acceptance and general utility of this approach
by the Agency in its future activities to develop
21
-------
Activity
Industry Definition and
Identification of
Industrial Processes
Review of Health and
Ecological Effects Data
Development of Process
Descriptions
Engineering Analysis and
Retrospective Epldemlo-
loglcal Analysis
Project Management
Definition of Data Needs,
Preliminary Definition
of R&D Needs, Report
Preparation
Total
Professional
Manhours
10S
10%
20%
25%
15%
20%
Professional Skill Required
Jr. engineering personnel under
supervision of Sr. staff,
chemical engineers
Industrial hyg1en1sts, toxlcologlsts,
b1ostat1st1dans
Jr. engineering personnel under
supervision of Sr. staff,
chemical engineers
Experienced chemical engineers,
Industrial hyglenlsts, toxlcolo-
glsts, b1ostat1st1c1ans
Sr. engineering staff
Engineering staff
Figure 2. Description of Personnel Needs During
Industrial Environmental Assessments
a data base on specific toxic and hazardous pollu-
tants produced by Industrial processes. The Agency
has already Initiated major data gathering efforts
along traditional media lines. Me hope that the
multimedia approach will provide a vehicle by
which this Information can be gathered more effi-
ciently with a minimum Impact on Industry and will
result 1n technology and regulations that address
the most serious Industrial pollution problems.
22
-------
SOURCE ASSESSMENT METHODOLOGY
Thomas W. Hughes
Monsanto Research Corporation
Dayton, Ohio
Abstract. The Industrial Environmental
Research Laboratory (IERL) of EPA has the
responsibility for insuring that pollution
control technology is available for sta-
tionary sources to meet the goals of
environmental legislation. IERL performs
Source Assessments for determining the need
to reduce emissions and discharges from
pollution sources. This paper presents a
discussion of the steps involved in pre-
paring a Source Assessment, decisionmaking
information used by IERL and an example of
the effect of data uncertainty on IERL
decisionmaking.
Introduction
The objective of a Source Assessment is to
provide IERL with sufficient information to
determine the need to reduce pollution from
stationary sources. A source is an entire
industrial, commercial, or municipal opera-
tion which is national in scope. An
assessment is the determination and extent
of pollution based on all available process
emissions, discharges and pollution control
information. The product of a Source
Assessment is a Source Assessment Document
(SAD). The result is an EPA decision
regarding the need to reduce pollution.
In the following discussion, there will be
two items addressed. These are: 1) activ-
ities involved in preparing the Source
Assessment Document; and 2) information
which EPA uses as an aid in determining the
need to reduce pollution. To give an over-
view, expert information from scientists
(analytical chemists, sampling personnel,
health and ecological experts, meteorolo-
gists, etc.) and engineers (chemical,
civil, environmental, etc.) is used to pre-
pare a complete assessment of a source.
The IERL uses this information in making
environmental decisions.
Preparation of a Source
Assessment Document
Figure 1 is a diagram of the steps involved
in preparing a Source Assessment Document.
Work Plan
A work plan is prepared at the start of
each Source Assessment. The objectives of
the work plan are two-fold: 1) to describe
the proposed study; and 2) to provide IERL
with a management control point. The work
plan contains the objective and scope of
the assessment, technical approach, source
definition, benefits expected, timing,
anticipated problems and proposed solu-
tions, estimated cost, manpower and per-
formance, project schedule and milestones.
Work Plan
Preliminary Source
Assessment Document
(PSAD)
Reid Sampling
Source Assessment
Document
(SAD)
number of
lly hazardous
- of pe^-
Emissions Reduction
Figure 1. Steps in performing a
Source Assessment.
Preliminary Source Assessment Document
IPSAD)
Preparation of the Preliminary Source
Assessment Document (PSAD) is initiated
upon EPA approval of the work plan. The
Preliminary Source Assessment Document
looks exactly like the Source Assessment
Document. The PSAD is a baseline of in-
formation about a source. It is prepared
by using available process, emissions, dis-
charges, pollution control and other
industry information. It is prepared with-
out extensive field sampling; a presurvey
sample is collected and analyzed using
methods resembling the Level I Environ-
mental Assessment procedures. The purpose
of the presurvey sampling is to identify
the types and quantities of previously
unknown hazardous or potentially toxic
pollutants.
The PSAD is a unique document in contract
research efforts. The purpose of the PSAD
is to determine if there is sufficient
information for IERL decisionmaking. This
purpose is expanded as follows: 1) to
provide an IERL management control point
for development of the SAD; 2) to provide
a basis for determining the need to acquire
additional information through sampling;
3) to identify the types and quantities of
previously unknown hazardous or potentially
23
-------
toxic emissions; and 4) to provide a basis
for obtaining additional information from
industry.
The PSAD contains the types of information
given in Table 1.
uncertainties and potential problems, and
options for test schedules and levels of
effort. The sampling and analysis proced-
ures used in the sampling program resemble
those of the Level II Environmental Assess-
ment protocols.
TABLE 1. INFORMATION IN A PRELIMINARY
SOURCE ASSESSMENT DOCUMENT
Process information
• General process and product flow
schematics
• Variations of operations within the
industry
• General operating parameters, espe-
cially those relating to pollution
• Plant capacities and locations
Pollutant data
Emission/discharge points
Materials emitted/discharged
Health and ecological effects
Pollutant inventories
Control techno! «->gy
ne-Art in control technology
application of controls
fficiencies
jnsiderations
Growth and nature of industry
• Process technologies
• Product market areas
• Production rate trends
Reference material
The PSAD is based on information from
available information sources. These
sources include personnel, published liter-
ature, local, state and Federal EPA offices
and files, industry surveys, trade assoc-
iations, presurvey testing at plants, and
equipment vendors.
If EPA determines that the PSAD contains
sufficient information for decisionmaking
purposes, the PSAD is published as a
Source Assessment Document. If it does
not, a field sampling program is performed
to collect the needed, but missing,
information.
Field Sampling
A sampling plan is prepared to obtain the
information needed for IERL decision-
making but which is not available in the
PSAD. The sampling plan is designed:
1) to allow IERL to exercise options on a
cost/effective basis; 2) to provide the
protocol for obtaining missing information
needed for the Source Assessment Document;
and 3) to allow IERL to judge the reliabi-
lity and acceptability of the sampling and
analysis methodology.
It contains a discussion of the additional
data required, test site(s) preparations
required, sampling and analysis methods,
Source Assessment Document (SAD)
The product from a source assessment is a
Source Assessment Document and it can be
prepared with or without field sampling.
The SAD contains the same types of inform-
ation as the PSAD (Table 1). It contains a
summary of the study (a concise presenta-
tion of all relevant information for IERL
to determine the need to reduce pollution);
a discussion of the industry, the severity
of the pollution from the industry,
summary of existing and anticipated control
technology, industry trends, unusual
results, and supporting data. Each Source
Assessment Document must be capable of
withstanding scrutiny, credible, factual,
accurate, self supporting, complete, con-
cise and lucid.
IERL Decisionmaking Procedures
IERL uses the Source Assessment Document to
determine the need to reduce pollution.
When it comes to pollution from a source,
EPA has only two choices: 1) to do some-
thing about the pollution, or 2) to do
nothing.
IERL uses a set of criteria as an aid in
determining the need to reduce pollution.
These criteria are shown in Table 2.
TABLE 2. IERL EVALUATION CRITERIA
Major Criteria
Source severity
Nation emissions burden
State's emissions burdens
Minor Criteria
Affected population
Pollution growth trends
The major criteria are used to determine
the need to reduce pollution. The minor
criteria are used to modify or prioritize
the decision. A description of each of the
criteria follows.
Source Severity(s)
Source Severity is defined as the ratio of
the concentration to which the population
is exposed to the concentration which
represents a potentially hazardous con-
centration.
The following discussion regarding the
criteria will be limited to air emissions
for simplicity. The same concepts and
approaches apply for other pollutant
media.
24
-------
The exposure concentration is the time-
averaged maximum ground level concentration
as determined by Gaussian plume dispersion
methodology. This determination requires
expert input from engineers, sampling
personnel, analytical chemists and
meteorologists.
The potentially hazardous concentration is
estimated in two ways. It is the Primary
Ambient Air Quality Standard for criteria
pollutants. It is a surrogate Primary
Ambient Air Quality Standard for noncrite-
ria pollutants. The potentially hazardous
concentration requires expert input from
health effects personnel.
The values of Source Severity which IERL
uses for decisionmaking are shown in
Table 3. There are only two decisions to
be made: 1) IERL will reduce pollution,
or 2) IERL will not reduce pollution.
TABLE 3. SOURCE SEVERITY MEASURE/DECISION
s
s
Measure
= § >. 0.05
= C < 0.05
Decision
There is sufficient cause
to reduce pollution.
There is not sufficient
cause to reduce pollution
The value of S = 0.05 as the cut-point was
obtained by evaluating the uncertainties
involved in engineering data, sampling and
analysis results, atmospheric dispersion
models, and health effects data. The
details of how this cut-point was developed
is described in a report entitled: "Source
Assessment: Analysis of Uncertainty" (1).
Nation Emissions Burden
The National Emissions Burden is the mass
of criteria pollutant emissions from a
source divided by the national mass of
criteria pollutant emissions. This
criterion uses engineering and emissions
input for a source to develop an emissions
inventory for the source. The values
which IERL uses for decisionmaking are
shown in Table 4. In short, if emissions
from a source amount to more than 0.1% of
the U.S. total criteria pollutant emis-
sions, then a source is considered as a
candidate for emissions reduction.
TABLE 4. NATIONAL EMISSIONS BURDEN*
Measure
Decision
x 100 * 0.1 There is sufficient cause
to reduce emissions
B =
N
x 100 < 0.1 There is insufficient
cause to reduce emissions
"calculated for each criteria pollutant, i.«.,
Particulates, Sulfur Dioxide, Nitrogen Dioxide
Carbon Monoxide, and Hydrocarbons.
State's Emissions Burden (B,J
^ • 5—
The State's Emissions Burden is identical
in concept to the National Emissions Burden.
It differs only in that it is calculated
for each of the 50 states instead of the
nation. Table 5 shows that if the State's
Emission Burden exceeds 1% of the state's
total criteria pollutant emissions, then it
is considered as a candidate for emissions
reduction.
TABLE 5. STATE'S EMISSIONS BURDENS (Be)
o
Measure
Decision
BS =
ZMP
-
. SMP
x 100 2 1 There is sufficient
cause to reduce
emissions
Bg = -fif- * 100 < 1 There is insufficient
S cause to reduce
emissions
Affected Population
The Affected Population is the number of
persons exposed to a potentially hazardous
environment. It is the number of persons
exposed to a Source Severity greater than
1.0. The affected population is not used
to determine the need to reduce emissions.
It is used in establishing a priority on
those sources which require emissions
reduction.
Emissions Growth Trends
The emissions growth trends is a measure of
predicted future emission rates relative to
current emission rates. If a source will
go out of business within 5 years, then
IERL will not attempt to reduce emissions
as this will happen anyway. If a source
will significantly reduce its emissions as
a result of local or state EPA regulations,
then IERL may not consider additional
emissions reduction.
IERL Decisionmaking Overview
IERL uses a Source Assessment Document to
determine the need to reduce pollution
using criteria discussed above. A Source
Assessment Document incorporates expert
information from engineers, sampling
personnel, analytical chemists, meteorolo-
gists, and health effects personnel. IERL
has a few concerns about the correctness of
its decisions. These concerns are:
(1) Is the decision correct?
(2) What impact does data quality have on
the correctness of the decision?
(3) Which information areas have the
greatest impact on correct
decisions?
(4) What can be done to improve the
decisionmaking process?
25
-------
Whether or not any given decision is cor-
rect cannot be determined at the time the
decision is made. All that can be done at
the time the decision is made is to assess
the likelihood that the decision is correct.
This likelihood is based on the quality of
the information on hand at the time the
decision is made.
The impact of poor data quality on
decisionmaking is that it forces the
decision maker to be conservative. This
conservation is used to make up for the
increased likelihood of incorrect decisions
due to poor data quality. An example of
how this applies is in the cut-point for
Source Severity.
Ideally, IERL would like to base its
decisions on a true value of source
severity. If a true value of Source
Severity could be obtained, lERL's decision-
making would be based on a cut-point of 1.0.
However, the true value of Source Severity
is unknown. It can only be estimated.
Monsanto Research Corporation's project on
the Analysis of Uncertainty showed that the
calculated value of Source Severity is
within a factor of 20 of the true value of
Source Severity. This range of possible
true values of Source Severity for a cal-
culated value requires that the IERL cut-
point be 0.05 (1/20).
This range of uncertainty in Source
Severity is caused by the uncertainties in
the data used to calculate Source Severity.
These data inputs are:
(1) Emissions data
(2) Dispersion models
(3) Health effects data
The contribution to the total uncertainty
is shown in Table 6.
TABLE 6. CONTRIBUTIONS TO UNCERTAINTY IN
SOURCE SEVERITY VALUES (1)
TABLE 7. EFFECT OF UNCERTAINTY REDUCTION &
ON CUT-POINT FOR SOURCE SEVERITY
Complete reduction
of uncertainty here
Sampling and analysis
Health effects data
Dispersion sodeling
of uncertainty in
yields source severity
» 35%
» 80%
1. 90%
The values shown in Table 6 do not add to
100% because the contributions to total
uncertainty for each input are not linear.
If the uncertainty in all three areas were
completely removed, the uncertainties in
Source Severity would be completely removed.
Another way of viewing the effect of the
uncertainty in each input area is to deter-
mine the cut-point which would be used in
the decisionmaking process. This is shown
in Table 7.
Table 7 shows that as the data quality
improves, the degree of conservatism in
decisionmaking decreases.
Cosplete reduction
of uncertainty her*
Sampling and analyst*
Health effect* data
Dispersion Modeling
Corresponding cut-point
Yields for source severity
»• O.OB
•• 0.25
The specific contributors to Source
Severity uncertainty are shown in Table 8.
This list of the elements of uncertainty
can be used to prioritize the data quality
improvement for Source Assessment.
TABLE 8. ELEMENTS OF UNCERTAINTY IN
SOURCE SEVERITY
Element of uncertainty
Reduction in total
uncertainty if uncertainty
in element is eliminated, «
Sampling and analysis
Random uncertainty
Bias
Dispersion model
Dispersion equation
Plume rise
Potentially hazardous
concentration
Primary Ambient Air
Quality Standards
Safety factor on
health effects data
35
2
33
90
80.
74'
'b
80
^Uncertainties not linear.
The simplest area to reduce the uncertainty
in Source Severity is in the plume rise
calculation in the dispersion model. The
dispersion model, as currently used in
Source Severity, ignores plume rise.
Recent EPA studies have shown that the
Brigg's equation for plume rise yields a
better correlation of predicted versus
actual downwind concentrations. Use of the
Brigg's equations for plume rise in the
dispersion model would reduce source sever-
ity uncertainty by 74% or increase the cut-
point from 0.05 to 0.19. This is quite
simple to do since it only requires a
computational change.
The most difficult area to reduce the
uncertainty in Source Severity is in the
estimation of the potentially hazardous
concentration. This concentration requires
a great deal of laboratory experimentation
to generate the needed health and ecologi-
cal effects data. Once these data are
generated, a value judgement is required to
estimate a "safe" or "acceptable" concen-
tration of exposure.
The area where a reduction in data uncer-
tainty has the least impact in reducing
Source Severity uncertainty is sampling and
analysis. This statement applies when the
quality assurance procedures specified by
the IERL-RTP Data Quality Manual are fol-
lowed. Any deviation from those procedures
or the procedures specified by the Quality
Assurance Branch will increase the uncer-
tainty in sampling and analysis data;
hence, Source Severity.
26
-------
Example
An example is given below to illustrate the
results of a rather complex statistical
treatment of data used in Source Assess-
ment. Emissions from the main process
vent at carbon black plants are used in
the example. Table 9 shows the materials
emitted as determined through field
sampling. The Source Severities for carbon
monoxide, participates and hydrogen sulfide
will be used.
TABLE 9. CARBON BLACK EMISSION FACTORS3
TABLE 10. CARBON BLACK SOURCE SEVERITIES
Emission factor,
Material emitted g/kg product
Criteria pollutants
Particulates
Nitrogen oxides
Nonmethane hydro-
carbons
Carbon monoxide
Chemical substances
Hydrogen
Hydrogen sulfide
Carbon disulfide
Carbonyl sulfide
Carbon black
Methane
Acetylene
Ethylene
Propane
Isobutane
Polycyclic organic
material
Trace elements
aMain process vent.
Particulate matter is
0.11 ± 70%
0.28 ± 15%
50 ± 48%
1,400 ± 19%
120 ± 39%
30 ± 82%
30 ± 76%
10 i 99%
0.11 ± 70%
25 ± 47%
45 ± 48%
1.6 ± 85%
0.23 ± 100%
0.1 1 80%
0.002 ± 52%
<0.25
carbon black.
Material
emitted
Carbon monoxide
Particulates
Hydrogen sulfide
Case IIC
Carbon monoxide
Particulates
Hydrogen sulfide
Source
severity
(S)d
1.8
0.02
20
0 . 36
0.004
4.0
Uncertainty in
source severity
(Jbrll
0.09
0.001
1.0 <
< S < 36
< Si < 0.4
400
0.07 < S < 1.9
0.0008 < SI, < 0.02
0.8
20.8
Slain process vent.
Assuming no plume rise.
°Using a plume rise of 11.5 meters.
Sc - calculated value of source severity.
eS_ = true value of source severity.
Table 10 shows the Source Severities for
each of these emissions, and the uncertainty
in Source Severity. Two cases are shown.
Case I represents the Source Severity cal-
culations assuming no plume rise from the
main process vent. The effective stack
height was assumed to be the only physical
stack height of 25 meters. Case II is
based on an effective stack of 36.5 meters.
This includes the physical stack height of
25 meters and a plume rise of 11.5 meters
based on Brigg's equation.
Note that when plume rise is used, the
magnitude of Source Severity and the
uncertainty in Source Severity decrease.
Not using plume rise in Source Severity is
a worst case analysis in that it over-
estimates Source Severity. However, it does
this at the expense of increasing
uncertainty.
Summary
This paper presented a discussion of the
steps involved in performing a Source
Assessment, decisionmaking information used
by IERL and an example of the effect of
data uncertainty on IERL decisionmaking.
Input from engineers, sampling and analysis
personnel, health effects experts and
meteorologists have significant impact on
IERL decisionmaking. ' The relationship of
these inputs to IERL decisionmaking were
presented.
27
-------
HEALTH RELATED PROGRAMS
Shahbeg S. Sandhu, Ph.D.
and
Michael D. Waters, Ph.D.
Health Effects Research Laboratory
U. S. Environmental Protection Agency
Research Triangle Park, NC 27711
Abstract
The growing concern for human health safety
due to the ever Increasing presence of a wide
variety of chemicals In our environment has been
amplified by studies showing strong associations
between chemical carcinogenesis and mutagenesls.
The diversity of genotoxic insults inflicted by
environmental chemicals and the need to evaluate a
large number of chemicals and complex mixtures has
resulted in the use by this laboratory of a multi-
level approach employing a battery of tests at
each level. The emphasis In the Level I battery
is on the detection of acute toxicity using mamma-
lian cells in culture and intact animals and
primary DNA damage and point mutation using micro-
fa lal species. The Level II battery la designed to
verify the results from Level I tests by employing
bioassays involving mammalian cells in culture,
plants, Insects, and short-term tests in Intact
animals. The emphasis in Level III tests is on
quantitative risk assessment using conventional
toxicological methods.
Introduction
There has been a widespread concern for the
last two decades about the progressively deterior-
ating quality of our-environment. A recent
article in Science estimates about 63,000
chemicals in common use. Research over the past
10 years has established that a significant
proportion of these chemicals is capable of caus-
ing genetic abnormalities including birth defects,
carcinogenesis, and mutagenesls. In addition,
environmental chemicals may be adversely affecting
more subtle attributes of life such as aging,
cardiovascular functions, immunity to diseases and
behavior.
These toxicants occur in various products and
media including foods, drugs, cosmetics, pesti-
cides, household and industrial chemicals, as well
as pollutants of air and water. The Food and Drug
Administration estimates approximately 4,000
active ingredients in drugs, 2,500 additives to
increase nutritional value in food, and about
3,000 chemicals used to "enhance the quality of
life". The Environmental Protection Agency esti-
mates .about 1,500 active Ingredients in pesti-
cides . At present, only a very small propor-
tion of these chemicals have been evaluated in
terms of their potential long-term effects on the
human population.
It is recognized that the toxlcological
effects which can be produced by environmental
chemicals are numerous and yet the specificity of
chemical structure and conformation make it possi-
ble for one effect to be produced exclusively. In
the past, toxicologists have relied on animal
bioassays to evaluate the safety of the chemicals
for human health. This conventional methodology
must be relied upon for purposes of risk assess-
ment. However, comprehensive animal tests are
expensive, costing $100,000 to $300,000; and time
consuming, lasting from one to two years. In
addition, the animal systems have their own limita-
tions, most Importantly experimental population
size.
In response to the need for testing a wide
variety of chemicals and complex mixtures for
toxic and genotoxic effects, the Health Effects
Research Laboratory, Research Triangle Park, N.C.
(HERL-RTP) of the Environmental Protection Agency
has adopted a matrix of bioassays based on the use
of mifrn—nrfttmlgmnt plants, mammalian and human
cells in culture, and whole animals. This program
is briefly outlined In Fig. 1.
FIGURE 1. LEVEL APPROACH TO SCREENING FOR ENVIMMHENTAL HEALTH EFFECTS
LEVEL I
DETECTION
TOXICITY
Cytotoxlclty
Acute Toxicity
absorption
distribution
•eubollsa
excretion
HUTASENES15
CARC1H06EHES1S
Mlcroblal Mutagenesls
i •Hwltin netabollc
activation
LEVEL II
VERIFICATION
LEVEL III
RISK ASSESSMENT
Cellular HeUbollM
Subacute Toxicology
MoacciMJlatlon
cellular toxicity
organ toxicity
teratology bloafsay
t
Chronic Toxicology
cellular pathology
organ pathology
acceptable dally Intake
illan Cell
Hit*genesis
ChroBosoae and
Gene Mutations
In Plants and Insects
Heritable Translocatlon
Hultlgeneratlon Reproduction
Neoplastlc Transformation
In Hawaiian Cells
Hutu 11 an Carctnogenesls
carcinogen pathology
tuaorlgen pathology
28
-------
The testing procedures adopted by HERL-RTP
for use with environmental emissions and effluents
could be grouped into two major catagories: (1)
those dealing with genetic toxicology and (2)
those related to general toxicology. For the
purpose of this report, the bioassays dealing only
with cytotoxiclty and genotoxicity are discussed.
The screening program for genotoxic effects
is based on a stepwlse or multi-level approach.
This scheme involves the evaluation of suspect
substances by beginning with the least expensive
and most rapid assays, mostly qualitative or semi-
quantitative in nature, followed by more expensive
and time-consuming assays, mostly quantitative in
nature.
The emphasis in Level I testing for mutagens
and potential carcinogens is on detection. False
negatives are of great concern, hence the require-
ment for a battery of tests. False positives are
of lesser concern since they will likely be elimi-
nated in higher levels of testing.
At Level I, simple in vitro assays for
mutagenesis and potential carclnogenesis employ
microbial indicator strains with and without
exogenous mammalian microsomal activation systems.
It has become apparent that a majority of the
genotoxins are procarcinogens or promutagens which
must be converted in in vitro into their reactive
forms before their effects can be evaluated. The
metabolic conversion is believed to be mediated by
oxidative enzymes and Involve the formation of
reactive electrophllic metabolites which interact
with DMA. Although there has been controversy as
to the best mode of coupling the •"•'""I metabolic
activation with the in vitro testing systems, due
mainly to differences in species, sex, age, and
organ specificity for different classes of chemi-
cals, a post-mitchondrial mammalian liver mlcro-
somal preparation (S-9) is routinely added to the
test system.
Level II testing emphasizes verification of
Level I results in appropriate test cells or
organisms. The end points of mutagenesis and
potential carcinogenesis are now separated for
greater end point definition. Mutagenesis tests
in Level II involve use of mammalian cells In
culture (e.g. L5178Y - mouse lymphoma cells, CHO -
Chinese hamster ovary cells, V79 - Chinese hamster
lung cells), plants (e.g. Tradescantia pallasada).
or use of insects (e.g. Drosophila melanogaster).
Carclnogenesfs testing in vitro Involves morpho-
logic cellular neoplastic transformation studies
(e.g. Syrian hamster embryo cells, and BALB/c 3T3
or C3H10T1/2 mouse fibroblasts). Other relevant
end points such as tests for DMA repair (unsched-
uled DMA synthesis) are also used in carclnogenesis
screening at Level II.
The other major area of investigation under
the program is that of general toxlcity. It is
not' possible to replicate in vitro all of the
manifestations of toxic response observed in vivo.
For this reason we must rely on the use of the
Intact animal in toxlcity screening. Certain in
vitro systems are, however, useful in cytotoxiclty
screening. In studies on airborne particulate
material, the rabbit alveolar macrophage system
developed In this laboratory appears to have
unique applicability. This cell type is largely
responsible for the removal of particulate material
from the lung. Hence, in theory, any substance
which alters its functional integrity could influ-
ence respiratory pathophysiology. The strain WI-
38 human lung fibroblast has been employed in
general toxicity evaluations under this program
and has been used elsewhere to study damage and
repair of DMA. A cellular toxicity test system
now under development involves cultivation of
primary rat liver cells. This system offers
valuable screening possibilities since the liver
is the principal metabolic organ In the body and
is vital in the detoxification of hazardous
chemicals. Such a cell system might be used at
Level II for toxicity evaluations or to provide
metabolic activation for other indicator organisms
(e.g. when co-cultivated with Salmonella typhimu-
rlum).
Level III testing involves the use of conven-
tional whole an-imai methods. Emphasis Is placed
on quantitative risk assessment in all toxicolog-
ical procedures. In mutagenesis studies the
emphasis is on genetic alterations induced in
mammalian germinal tissues (e.g. dominant lethal
test or heritable translocatlon test in mice) . In
carclnogenesis studies lifetime exposures (e.g. 2-
year feeding studies in rats) are used to detect
and quantify neoplasia in mammals. Conventional
chronic toxicity studies in mammals are also
employed at Level III.
These conventional whole an-tnui tests should
be used in conjunction with the simpler in vitro
and submammalian tests to provide the required
validation of positive results in screening tests.
However, because many substances will have been
eliminated 'in screening tests , the number of
substances which must be evaluated by conventional
methods will have been drastically reduced.
Likewise, the total cost of toxicological evalu-
ation will have been greatly reduced. The Imple-
mentation of screening tests by industry early In
the development of new chemical substances or in
the pilot plant stages of new process technology
may be expected to reduce the production of
hazardous substances and thereby reduce or prevent
their release to the environment.
Level I Bioassays
A battery of tests for this level is illus-
trated in Fig. 2. Again, the emphasis in Level I
testing is on detection of mutagens, potential
carcinogens and acutely toxic chemicals. This
level consists of a battery of in vitro and in
vivo tests. The in vitro end points which are
considered based upon the current level of devel-
opment of bloasaay systems are point mutations,
primary DMA damage, and cytotoxiclty. All bioas-
says are performed with and without the presence
of mammalian metabolic activation enzymes where
feasible. The results obtained from Level I tests
are used to assign priorities for further testing.
Conventional rodent acute toxlcity tests are also
considered essential at Level I in view of the
limitations of cytotoxlcity screening tests discus-
sed below.
29
-------
FIGURE 2. LEVEL 1 8IOASSAYS FOR EVALUATING HEALTH EFFECTS
POINT EVALUATIONS
0 SALMONELLA NICROSOME (AMES) REVERSE MUTATION.
PROTOTROPNY TO HISTIOINE
o ESCHERICHIA COU-WZ-MICROSOME REVERSE MUTATION,
PROTOTROPHT TO TRYPTOPHANE
o SACCHAROMYCES CEREVISCIAE REVERSE AND FORWARD
MUTATION
PRIMARY ONA DAMAGE
0 ESCHERICHIA COLI POL A" REPAIR DEFICIENT STRAINS
0 BACILLUS SU8TILLIS. REC* REPAIR DEFICIENT
STRAINS
o SACCHAROMYCES CEREVISCIAE GENE CONVERSION AND
MtTOTIC RECOMBINATION
CYTOTOXICITY
o RABBIT ALVEOLAR MACROPHAGE (FOR PARTICULATES)
CHINESE HAMSTER OVARY CELLS
0 HI-38 HUMAN FIBROBLASTS
ACUTE TOXICITY
0 14 DAY STUDY IN RODENTS
Bioassays to Detect Point Mutation
Salmonella typhimurium (Ames test). Developed
by Bruce Ames , this bioassay is the moat exten-
sively used as a mutagenesls prescreen to assign
priorities for further testing. The Sal "•""•I la
mutagenesis/mlcrosome teat Is a reverse mutation
assay measuring reversion to hlstidlne Independ-
ence. The experimental procedures normally employ-
ed include the spot test, the plate Incorporation
test, and the liquid suspension test. The five
strains commonly used in the bioaasay have been
extensively engineered to increase their sensitiv-
ity for measuring point mutations over a wide
range of test substances.
Specific Advantages;
1. Useful as a tool in rapidly obtaining
information about the potential mutagenlc/carcln-
ogenlc activity of uncharacterixed compounds In
complex mixtures. Has been applied to cigarette
smoke condensate and fractions, hair dyes, soot
from city air, etc.
2. Valuable as a bloassay to direct the
chemical fractlonation and identification of
mutagenlc components in complex mixtures.
3. Known human pure chemical carcinogens
are positive. These include 0-naphthylamlne,
benzidine, bls-chloromethylether, aflatoxin-B.,
vinyl chloride, 4-amino-blphenyl.
4. Reliable: (in testing known carcinogens
and noncarcinogens) false positives, £lOZ; false
negatives _<10Z.
5. Has extensive data base.
6. Tester strains extensively engineered
to enhance sensitivity.
7. Results in different tester strains
provide information about the mechanism of mutation
(e.g. frameshlft vs base pair substitution).
Specific Limitations;
1. Test is considered qualitative In
nature.
2. Test is not responsive to metals,
chlorinated hydrocarbons (long chain), asbestos
and like particles.
3. Sample must be sterile.
sterilization of samples.)
(Problems in
4. Highly toxic components of complex
samples may mask mutagenlc properties of other
chemicals .
5. Organic solvents used in chemical
fractlonation may be mutagenlc (e.g. methylene
chloride) and therefore may require removal (e.g.
by solvent exchange) prior to bloassay.
(A)
Escherichia coll WP2. Developed by Bridges v
and his associates in the United Kingdom, this is
also a reverse mutation assay which measures
prototrophy to the amino acid tryptophan. DMA
repair deficiency has been Included in the tester
strains to enhance their sensitivity. The bloassay
has been adopted for spot test, plate incorporation
test, and liquid suspension assay.
Specific Advantages;
1. Can be employed to detect certain
classes (e.g. nltrofurans) of chemical compounds
missed by non-plasmid containing strains of
Salmonella.
2. Has a better resolving ability than
Salmonella In distinguishing between specific
revert ant s and suppressor mutants.
3. Adequate data base is present.
Specific Limitations;
1. The bioassay is less sensitive than
due to the lack of cell wall mutants
and incorporation of R-plasmids in the latter.
2. It detects only base-pair substitutions.
It is deficient in the detection of frame-shift
mutagens.
The following advantages and limitations are
common to both the S. typhlaurlum and K. coll as
mutagenlclty test systems.
Advantages ;
1. Genetically well defined.
2. Inexpensive, simple and rapid.
3. Analysis of large populations over many
generations.
4. Genetic damage readily detected.
Limitations:
1. Organization of genetic material dif-
ferent from higher life forms.
30
-------
2. No mitosis or melosls.
3. Chromosomal damage not detected.
4. Results mostly qualitative in nature.
5. No endogenous metabolic activation.
Sacchan
Zimmerman
cerecislae. Developed by P. K.
(5)
with further development by Brualck
and Mayer , Saccharomyces may be used both for
forward and reverse mutation assays. Forward
mutation can be detected by the loss of function
as a change in color of cell colonies from red to
white or pink. The advantage of forward mutatlonal
assays is that mutational events at several loci
can be detected. Reverse mutation experiments are
easy to perform. However, the major drawback with
reverse mutation is that it requires a specific
genetic alteration for the restoration of function.
Not all mutagens induce all possible types of
alterations. As « result, reverse mutational
assays may be more limited in applicability.
Advantages:
1. Eukaryotic organism, with genetic organ-
ization similar to that of higher life forms.
2. Point mutations and small chromosomal
deletions may be detected.
3. Rapid, simple, and inexpensive.
4. Meiosis observable under well defined
conditions.
Limitations:
1. Problems with coupling of mammalian
metabolic activation systems.
2. Lack of genetic engineering to enhance
sensitivity.
Bioaasaya to Detect Primary DNA Damage
These tests provide an indirect measure of
damage to DNA from environmental toxicants.
The rationale for adopting these methods as
indicators of genotoxlcity is based on the recent
findings that many manmade and naturally-occurring
substances which have capacity to Interact with
DNA, through electrophlllc attack, have potential
to Induce mutation
For lack of better understanding of mechanisms
and resolution of end points; the following bio-
assays are Included in this catagory.
Escherichijucoli. Pol A". Developed by Herbert
BnagnVT-pnr1 ', the assay measures the differential
killing between DNA repair proficient JJ. coll
strain (W3110, Pol A*) and DNA repair deficient
strain (P3478, Pol A~) as affected by environmental
toxicants. Spot test and liquid suspension test
procedures are employed.
Advantages;
1. The assay is well suited for detecting
chemicals causing frame shift mutations.
2. Adequate data base Is present.
Limitations;
Tester strains are not extensively engine-
ered to Increase sensitivity.
Bacillus sul
ibtjlus rec assay. Developed by T.
i V** / •»!. J *. V j — -. ^~. — _ 1 *.*
Kada in Japan , this bioassay also measures the
differential killing between recombinatlonal DNA
repair proficient strain GSY 1035 (rec ) and
repair deficient strain 45 (rec~). The most
commonly used procedure is the spot test. Usually
both bacterial strains are streaked out along
intersecting lines. The test chemical Is spotted
at the intersection. This affords a direct
comparison of the zone of inhibition between the
repair proficient and deficient strains on one
plate. The basic difference between j>. subtllus
rec" assay and Z. coll A~ lies in the defects in
the respective repair systems involved.
Advantages;
1. Very rapid and simple bioassay.
2. Highly versatile.
Limitations;
1. Requires fairly large amount of test
substance.
2. Not suitable for substances that do not
diffuse rapidly in agar.
Saccharomyces cerevisiae - Gene Conversion and
Mitotic Recombination. Mltotic recombination of
the reciprocal type, mltotic recombination, and
the non-reciprocal type, mltotic gene conversion,
provide convenient markers for assessing the DNA
damaging potential of environmental chemicals.
The degree of mltotic crossing over is evaluated
by the frequency of twin spot sectors and that of
mltotic gene conversion by the differential growth
in a selective medium.
Advantages:
1. Diplold cells with eukaryotic chromo-
somal organization.
2. Rapid and inexpensive assay.
Limitations;
1. Less versatile due to problems associ-
ated with cell wall permeability.
2. Problems associated with coupling of
metabolic activation with the assay.
3. Inadequate data base showing the reli-
ability of this assay system.
Bioaasays For Assessing Cytotoxicity
The HERL-RTP has employed three assays to
evaluate the cytotoxicity of parent chemicals,
metabolites and complex environmental mixtures.
31
-------
These Include (1) the rabbit alveolar macrophage
(RAM), (2) proliferating cultures of Strain WI-38
human lung flbroblasts and (3) clonal assay*
employing continuous rodent cell lines such as the
Chinese hamster ovary (CHO) cell line.
Rabbit Alveolar Macrophage (RAM) Assay
dividing cells. Growth inhibition is commonly
used as a measure of cytotoxicity in this system.
Clonal Toxicity (e.g. CHO system). The impair-
ment in the clonal growth of continuous rodent
cell lines after treatment with a test substance
is used as a criterion of cytotoxicity.
The basic methodology for this assay is
illustrated in Fig. 3.
FIGURE 1 RAB8IT AVEOLAR MACROPMAGE (RAM) TEST SYSTEM
IAVAGE
POOL
•ASH
RESUSPEND
COUNT
DILUTE
ALIQUOT
ATTACH
(3HRS)
DISCARD ADO
UNATTACHED •• TEST
CELLS COMPOUNDS
(20HRS)
COUNT VIABILITY TOT. PROT. ENZYMES
\
PHAGOCYTOSIS
Alveolar macrophages obtained by saline
lavage from lungs of rabbits are maintained in
tissue culture and exposed to particulate materi-
als, soluble chemicals or gases. The end points
measured Include cell numbers, cell viability,
adenosine trlphosphate (ATP) levels, phagocytlc
activity, etc.
Advantages and Limitations of Cytotoxicity Tests.
These assays are less expensive, more rapid and
require less samples than conventional whole
an-im«l bloassays. The assays are used to provide
preliminary information on the relative cellular
toxlclty of unknown samples. The basic disadvan-
tage of these assays is that they do not represent
intact animals and therefore provide only prelim-
inary information about the potential health
hazards of the test chemicals.
Level II Bloassays (In Vitro and In Vivo)
The Level II test battery, Illustrated in
Fig. 4, Is designed to verify the results obtained
FIGURE 4. LEVEL 2 BIOASSAYS FOR HEALTH EFFECTS
PRIMARY UNA DAMAGE
o UNSCHEDULED DMA SYNTHESIS-U1-38
o SISTER-CHROMATED EXCHANGE FORMATION (IN VITRO
AND IN VIVO)
POINT MUTATION
0 MAMMALIAN CELLS IN CULTURE
o INSECTS-OROSOPHILA
o PlANT-TRAOeSCANTIA AND MAIZE
CHROMOSOMAL ABERRATIONS
0 IN VITRO CYT06ENETICS - MAMMALIAN CELLS IN
CULTURE
0 IN VIVO CYT06ENETICS - LYMPHOCYTIC CULTURE AND
BWHXRROU CELLS
Advantages;
1. Test indicates potential for patho-
physlologlcal damage to pulmonary cell systems and
Impaired lung defense against inhaled particulate
materials.
2. Participates such as silica and asbes-
tos, soluble and Insoluble metallic compounds, and
oxldant gases have demonstrated cytotoxlc activity
in vitro and la vivo.
3. It has been possible to ''rank" crude
particulate materials according to relative toxlc-
lty based upon concentration-response data.
Limitations:
NEOPLASTIC TRANSFORMATION
o SYRLIAN HAMSTER EMBRYO SYSTEM
o MOUSE FIBROBLAST CELL LINES
TERATOGENESIS STUDIES
SUBACUTE STUDIES IN RODENTS
In Level I bloassays. The test systems are
selected to provide Information on primary DMA
damage, point mutations, chromosomal aberrations
and cellular neoplastic transformation. The test
organisms are mammalian cells in culture supple-
mented with exogenous metabolic activation, plants,
insects, and whole animals, to provide greater
relevance to the human situation.
1. Protective Influences of the intact
animal may not be reflected in vitro.
2. Cells are non-dividing and cannot be
maintained in continuous culture.
Human Lung Flbroblaat (WI-38) Assay. These dip-
loid cells of human origin have bean used effec-
tively for determining the cytotoxicity of envir-
onmental toxicants. These cells exhibit major
pathways of macromolecular synthesis common to all
Bioassays to Detect Primary DMA Damage
WI-38 - unscheduled DHA synthesis. Developed by
Stlcht*&), this assay evaluates the test compounds
for their ability to induce unscheduled DNA
synthesis (DDS) In human dlplold WI-38 fibroblasts
blocked in G, phase.
Advantages;
DNA repair can be measured in human cells
32
-------
in culture. Similar studies can be performed
using peripheral leucocytes in animals permitting
comparison between in vitro and in vivo exposures
to carcinogens or mutagens.
Limitations;
1. The precise type of molecular binding
between carcinogens and DMA which triggers excision
repair is unknown.
2. DMA repair synthesis does not measure
residual damage to DMA.
Sister-Chromatid Exchange Formation (SCE).
Developed by Sam Latt1 , the SCE Involves a
reciprocal exchange between sister-chromatids
which does not result in change in overall chromo-
somal morphology. The technique depends upon a
differential stainabllity of sister-chromatids by
exposing cells to 5-bromodeoxyuridine (BudR)
through two successive cell generations. Subse-
quent staining with fluorochrome and/or Giemaa
readily demonstrates the number of times per cell
where darkly-staining chromatlds are exchanged
with lightly-staining chromatids.
The assay may be used either In the in
,(20,21,29) or jn „«„„ *ir,,mHHn(*-) _
vitro
Advantages:
The test is an inexpensive, sensitive and
a rapid method of measuring the effect of a
substance on chromosomes.
Limitations:
1. The mechanism and significance of SCE
is not well understood.
2. No clear relationship between SCE and
chromosomal breaks has been established.
3. BudR which forms the basis of differ-
ential staining in this bloassay is itself a
mutagen. Therefore, it is difficult to evaluate
the synergistic, antagonistic or interactive
effects of BudR with the test compounds.
Bloassays to Detect Point Mutation
Mammalian Cell Mutagenesls In Vitro. There are
three »nmman«n cell mutagenesis bloasaays which
are commonly used:
xanthine-guanine-phosphorlbosyltransferase (HGPRT)
and ouabaln loci. These loci are assumed to be X-
1inked. The HGPRT locus la better characterized
and considered to be more sensitive than the
ouabaln locus.
Advantages:
1. The full range of mutation responses of
somatic cells observable, not just point mutations.
2. Faster, less expensive than in vivo
assays.
3. Can be coupled with in vitro metabolic
activation system from various sources.
4. Mutations measured in cultured mammalian
cells are more relevant to the Intact mammal than
in bacterial systems.
5. Responds to particulates (asbestos).
6. May offer better quantitative correla-
tion with carcinogenicity than does mutagenlcity
data of Salmonella, though this is very prelimi-
nary.
Limitations;
1. The cell line used for experimentation
are transformed lines and therefore do not repre-
sent "normal" cells.
2. These cells do not undergo meiosis.
3. Only one or two loci from the entire
genome are monitored.
4. Cells have limited endogenous metabolic
activation capability.
5. There are some doubts about the genetic
nature of the events observed at certain loci.
Trad
itiaSt
•hair. Developed by A. H.
L5178Y Bouse Lymphoma
,(7)
Developed by Donald Clivev", this Is a
forward mutation assay which measures the mutation-
al induction of thymidine klnase (TK) deficiency
from TK competent heterozygotes using a selective
medium. The TK locus Is believed to be autosomal.
Chinese Hamster Cells
There are two assay systems derived from
Chinese hamsters: Chinese hamster ov*F{8cf^}8
(CHO) developed by Bale and associates ' and
Chinese hamster lung cells (V79) developed by
Chuw and Krahn and Heldelburgerv '. Both
assays measure resistance to drugs at the hypo-
Sparrowv"", this plant provides a unique screening
system for mutagenesis. The clones heterozygous
for stamen hair color are exposed to test substan-
ces and a change of stamen hair cells from blue to
pink coloration is recorded as a mutatlonal event.
Advantages;
1. The Tradescantla stamen hair system is
very sensitive to physical and chemical mutagens.
2. Easy to handle.
3. Has eukaryotic chromosomal organization.
4. Suitable for In situ testing.
5. Suitable for testing liquid and volatile
compounds.
Limitations;
1. Stamen-hair cells are somatic cells and
therefore do not provide the Information on the
Influence of test compounds on the germinal
cells.
33
-------
syste
2. Nature of mutational events in this
is not well understood.
3. To date, the data base on this system
is very meager.
Sex-linked Recessive Lethal in Drosophila. The
sex-linked recessive lethal test in Drosophila is
a relatively economical and sensitive test for the
evaluation of environmental chemicals for potential
health effects in higher organisms. This system
is capable of detecting mutations at approximately
1,000 loci on the X chromosome.
Advantages;
1. A large number of the test organisms
can be reared easily and economically.
2. Provides the advantages of endogenous
metabolic activation.
3. Genotoxic effects are evaluated in the
germ cells.
4. Many loci in the genome can be monitored
at the same time.
Limitations;
1. Drosophila does not activate some
polycycllc hydrocarbons.
2. Several classes of chemicals Including
certain insecticides cannot be evaluated in the
Drosophila system.
Assays to Detect Chromosomal Aberrations.
Test systems to evaluate chromosomal abnormal-
ities should be an integral component of a scheme
designed to evaluate the genotoxic effects of
toxicants. Many types of cancer and several
categories of birth defects are related to the
structural and/or numerical changes in the chromo-
somal complements of animals.
Both ^n vitro and in vivo bloassay are cur-
rently employed to determine the ability of envir-
onmental chemicals to cause chromosomal abnormali-
ties.
In vitro cytogenetlc studies to detect numer-
ical and structural chromosomal alterations can be
performed with the .mammalian tumor lines, Muntjack
or Chinese hamster . In vivo studies Include
short.term (48 to 72 hoursT culture of lympho-
cytes L Land micronucleus formation in bone marrow
cells derived from exposed animals.
The micronucleus assay is based on the obser-
vations that small places of chromatin material
without centromeres are produced as a result of
chromosomal breaks and can be easily identified in
the interphase. The assay Is relatively simple,
rapid, and inexpensive.
Limitations;
Chromosomal alterations are relatively less
sensitive indicators of genotoxic effects than
point mutations.
Oncogenic Transformation
Oncogenlc transformation is the process
whereby normal cells grown in culture are converted
into malignant cells after treatment with an
oncogen. The demonstration of malignancy (tumor
formation) can be observed by injecting the trans-
formed cells into whole animals. A number of
mammalian oncogenic transformation bioassays
utilizing cells derived from different rodent
species are currently available. Some of these
cell systems have the endogenous capability to
activate procarcinogens while with others exogenous
metabolic activation has been used successfully.
Two types of assay systems in use for onco-
genic transformation are: (1) Syrian hamster
embryo system and (2) mouse flbroblast cell lines.
The Syrian hamster embryo system , utilizes
primary cells which are diploid and retain high
metabolic activation capability, thereby providing
more relevance to human conditions.
Mouse embryo cell lines are continuous cell
lines with aneuplold chromosomal complements. The
tester strains have limited metabolic activation
ability. These systems are easily adaptable for
testing agents as Initiators and promoters, cocar-
clnogens and carcinogens.
Advantages;
1. Bioassays using mammalian cells are
considered to be directly relevant to the process
of oncogenesls in experimental animals.
2. These tests are highly predictive. Few
if any false positive are detected by this method-
ology.
Limitations;
1. The assays are relatively expensive to
perform.
2. Highly trained personnel are required
to perform the tests.
Level III Bioassays
Level III testing involves the use of conven-
tional whole animal methods. The emphasis Is
placed on quantitative risk assessment. Experi-
mentation with Intact mammals are needed to provide
information on the presence of mutagenlc concentra-
tion of toxins in the target cells. In addition,
information on pharmacoklnetlcs, such as absorp-
tion, distribution, metabolic transformation, and
excretion,cannot be obtained without studies on
Heritable Translocation Test. The heritable
translocation assay is designed to measure recipro-
cal translocation in the germ cells of treated
animals, usually the male mice. The presence of
Induced translocatlons can be observed by mating
the F. male progeny of the treated males with the
untreited females*24'11'.
Advantages;
1. Chromosomal effects detected in the
34
-------
progeny of treated animals demonstrating a herit-
able effect.
2. Effects can be confirmed by examination
of germinal tissues.
3. Metabolism more similar to that in man.
Limitations:
1. Poorly defined genetically.
2. Genetic damage not readily detectable.
3. Delayed mutant expression tine.
4. Small populations and few generations
amenable to analysis.
5. Expensive.
Dominant Lethal Test
The genetic basis for dominant lethal test is
the induction of chromosomal damage and.rearrange-
ments resulting In nonviable zygotes - The
genetic damage is detected as pre-Implant loss of
non-viable blastocysts and early embryonic death.
Advantages:
Produces general Information about effects
of a substance on the germ cells of the teat
animals at a relatively low cost.
Limitations;
The bioassay Is incapable of detecting
certain weak mutagens.
Carcinogenesls and Tumorlgenesis in Intact
Animals
These tests are essential in evaluating risk
to human health. However, the very high cost and
long study period required for these tests, limit
their utility.
Resource Implications
The resource implications of carclnogenlcity/-
mutagenicity testing programs are shown in Table
1. It is evident that most of the short-term
bioassays assigned to detect primary DMA damage
and point mutation are relatively rapid, inexpen-
sive, and require small amounts of test material.
Table 2 presents « summary of resource re-
quirements for each level of testing. Levels I
and II test batteries could be applied at a cost
of $35,000 and tests completed within three
months, whereas, a battery of Level III bioaasaya
would be approximately 8 to 10 times more costly
In terms of dollars and time.
cals and complex mixtures. This fact supports the
adoption of a multi-level approach with batteries
of tests at each level. Such an approach Is cost
and resource effective.
Table 1
Resource Implications of Carc1nogen1c1ty-Nuugen1c1ty
Program)
, Study
Taat $ Coat tlM
Cana (Point) Mutation*
Bacteria 3 SO - 600 2-4 waaka
(ABM Flat* Taae)
Bactarta 1.000 - 2.000 2-4 waaka
(Liquid Suapanalon)
Eukaryotie micro- 200 - 500 2-* waaka
organlama (yaaat)
loaccta (Droaophila. 6.000 - 7,500 4-6 month.
receaalva lathal)
Mimm.lUn aomatlc ealla 2.300 - «,800 1-2 month.
in eultura (mouaa
lymphoma)
Mouaa Spaciflc Locua 20.000+ 1 yaar
Chromosomal Mutation*
In vivo cytoganatlea 3,000 - 6,500 6-8 waaka
Inaacta, haritabla 3.000 - 6.500 4-6 waaka
chromoaomal (((act a
(Droaophila) non- 3.000 1-3 aontha
disjunction
Dominant lathal in 6.000 - 10.000 3 month.
rodanta
Baritabla tranaloca- 40.000 - 67,000 12-18 mootha
clon in rodanta
Primary Dm Da»a«a
DM rapalr in bactaria 200-500 2-4 waaka
Onachadulad DMA aynthaaia 350 - 2.000 4-6 waaka
Mlcotlc recombination 200-500 4-6 waaka
and/or (ana converaion
la yaaat
Slatar ehromltid 1,000 - 1.200 4-6 waaka
axchanaa
In Titro Tirana- 6.500-7.500 10-12 waaka
formation
Quantity of
matarlal
raqulrad
2 t
2 t
2 8
10 g
2 S
25 (
20 g
10 g
20-25 g
25 g
2 1
2-5 g
2-5 g
2-5 g
2-5 g
1. Coat of thaaa taata haa variad and can be expacted to vary
until taat raqnlrmmmnes ara atabilizad.
2. Thia tima period covert tba experimental tlaia and raport
praparatlon.
Table 2
Relative Cost. Study Tlae and Quantity of Material Required
for the lapHaentatlon of Testing for
Mitagenldty Carclnogenlclty Progrw
Laval
•umbar
of taata
Approximata
coat
Study
tima
Quantity of
matarlal
raquirad
I
II
III
S 3.500-5.000
$27.000-30.000
S200.000-300.000
4-6 waaka 2-5 g
6-12 waaka 5-20 g
1-2 yaara 25-30 g
Considering the advantages and limitations of
the assays discussed, it is evident that no single
method la entirely adequate in assessing the
toxiclty and genotoxlcity of environmental chemi-
35
-------
References
1. Allen, J. W. and S. A. Latt (1976). Analysis
of sister chromatid exchange formation
in vivo in mouse spermatogonia. A new
test system for environmental mutagene-
sls. Nature 260;449-451.
2. Ames, B., J. McCann, and E. Yamasakl (1975).
Methods for detecting carcinogens and
mutagens with the Salmonella/mammalian
microsome mutagenicity test. Mutation
Res. 3J_: 347-364.
3. Bateman, A. J. and S. S. Epstein (1971).
Chemical Mutagenesis, Principles and
Methods for Their Detection. Vol. 2. A.
Hollaender, ed. Plenum Press, New York.
p. 541.
4. Bridges, B. A. (1972). Simple bacterial
systems for detecting mutagenic agents.
Lab. Practice 21_:413-416.
5. Bruslck. D. J. and V. W. Mayer (1973). New
developments in mutagenicity screening
with yeast. Environ. Health Prospect.,
i:83-96.
6. Chu, E. H. Y. (1971). Induction and analysis
of gene mutations in mammalian calls in
culture. In; Chemical Mutagenasis,
Principles and Methods for Their Detec-
tion, Vol. 2. A. Hollaender, (ed.).
Plenum Press, New York, pp 441-444.
7. Clive, D. and J. F. S. Spector (1975).
Laboratory procedure for assessing
specific locus mutations at TK locus in
cultured L5178Y mouse lymphoma cells.
Mutation Res. JH:17-29.
8. Cohen, M. and K. Hirschhorn (1971). Cyto-
genetic studies in animals In; Chemical
Mutagens, Principles and Methods for
Their Detection, Vol. 2. A. Hollaender
(ed.). Plenum Press, New York. pp.
515-534.
9. DiPaolo, J. A., P. J. Donovan, and R. I.
Nelson, (1969). Quantitative studies in
vitro transformation by chemical carcin-
ogens. J. Hat. Cancer Inst. 42;867-874.
10. Evans, H. J. (1976). Cytological methods
for detecting chemical mutagens. In;
Chemical Mutagens, Principles and Methods
for Their Detection, Vol. 4. A. Hol-
laender (ed). Plenum Press, New York.
p. 1.
11. Generoso, W. M., K. T. Cain. S. W. Huff, and
D. G. Gosslee (1977). Heritable trans-
location test in mice. In; Chemical
Mutagens, Principles and Methods for
Their Detection, Vol. 5. A. Hollaender,
(ed.). Plenum Press, New York. pp. 21-
56.
12. Kada, T., K. Tutikawa, and Y. Sadale (1972).
In vitro and host mediated "rec-assay"
procedures for screening chemical auta-
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
gens, and phloxine, a mutagenic red dye
detected. Mutation Res. 16; 165-174.
Krahn, D. F. and C. Heidelburger (1977).
Liver homogenate mediated mutagenesis in
Chinese hamster V79 cells by polycyclic
hydrocarbons. Proc. Natl. Acad. Sci.,
USA. 73:188-192.
Latt, S. A. (1976). Analysis of human chromo-
some structure, replication and repair
using BudR. V-33258 Hoechst techniques.
£. Reprod. Med. ^7:41-52.
Maugh, Thomas H. (1978). Chemicals: How
Many Are They? Science 199:162.
McCann, J. and B. N. Ames (1976). Detection
of carcinogens as mutagens in the Sal-
monella/microsomes test: Assay of 300
chemicals: Discussion. Proc. Natl.
Acad. Sci. USA 73:950-954.
Miller, J. A. (1970). Carcinogenesls by
Chemicals: An Overview. Cancer Res.
30:559-576.
O'Neill, J., Patrick, P. A. Brlmer, R. Mach-
anoff, 6. P. Hirsch, and A. W. Hsie
(1977). A quantitative assay of mutation
induction at the hypoxanthlne-guanine
phosphorlbosyl transferase locus in
Chinese hamster ovary cells (CHO/HGPRT
system): development and definition of
the system. Mutation Res. 45:91-101.
O'Neill, J. Patrick, B. D. Couch, R. Machan-
off, J. R. Sebastian, P. A. Brimer, and
A. W. Hsle (1977). A quantitative assay
of mutation induction at the hypo-
xanthlne-guanine phosphoribosyl trans-
ferase locus in Chinese hamster ovary
cells (CHO/HGPRT system): utilization
with a variety of mutagenic agents.
Mutation Res. 45(1):103-109.
Perry, P. and H. J. Evans (1975). Cytologi-
cal detection of mutagen-carclnogen
exposure by sister chromatid exchanges.
Nature 258;121-125.
Popescu, H. C., D. Turnbull, and J. A.
DiPaolo (1977). Sister chromatid ex-
change and chromosome aberration analy-
sis with the use of several carcinogens
and noncarcinogens. J. Natl. Cancer
Inst. 59:289-293.
Russell, L. B. (1951). X-ray Induced muta-
tions in mice. Cold Spring Harb. Symp.
Quant. Biol. 16:327-336.
Schmld, W. (1975).
Mutation Res.
The micronucleus test.
31:9.
Searle, A. G. (1975). The specific locus
test in mouse. Mutation Res. 31;277-
290.
Slater, E. S., M. D. Anderson, and H. S.
Rosenkranz (1971). Rapid detection of
mutagens and carcinogens. Cancer Res.
36
-------
J»_:970-973.
26. Sobels, F. H. (1977). Some problems assoc-
iated with the testing for environmental
mutagens and perspective for studies In
"comparative mutageneais." Mutation
Res. 46:245-260.
27. Sparrow, A. H., L. A. Schalrer, and R.
Villalobos-Pietrini (1974). Comparison
of somatic mutation rates induced in
Tradescantia by chemical and physical
mutagens. Mutation Res. 26:265-276.
28. Stich. H. and S. Laighes (1973). DHA repair
and chemical carcinogenesis. Patho-
biology Annual 3tU2-376.
29. Syuiti, A. and S. Motomichl (1977). Chro-
mosome aberrations and sister chromatld
exchanges in Chinese hamster cells
exposed to various chemicals. J. Hatl.
Cancer Inst. 58:1635-1640.
30. Waters, M. D., D. E. Gardner, C. Aranyl, and
D. L. Coffin (1975). Metal toxicity for
rabbit alveolar macrophages in vitro.
Environ. Res. j): 32-47.
31. Zimmermann, F. K. (1975). Procedures used in
the induction of mi totic recombination
and mutation in the yeast Saccharomyces
cerevlsiae. Mutation Res. 31:71-86.
37
-------
BIOLOGICAL TESTING METHODOLOGY
Kenneth H. Duke
Associate Manager
Ecology and Ecosystems Analysis Section
Battelle Columbus Laboratories
Columbus, Ohio
Abstract
Biological procedures for testing feedstocks
and waste streams of Individual processes have
been developed as part of EPA's phased approach to
Environmental Source Assessment. The objectives
of the first phase or level of biological testing
are to (1) provide preliminary data on the biolog-
ical effects of feedstocks and waste streams,
(2) Identify problems In Implementing the tests on
the streams, and (3) prioritize the streams accord-
Ing to their relative hazard. The Bioassay Sub-
committee of lERL-RTP's Environmental Assessment
Steering Committee has developed a Level 1 biolog-
ical test protocol to meet these objectives. The
protocol includes four health tests and eight eco-
logical tests to be implemented on liquid, solid,
and gaseous streams. Pilot studies are underway
to test and refine this protocol. Results indi-
cate the phased approach is effective 1n directing
environmental assessment activities.
Introduction
Environmental source assessment, as conceived
by the Industrial and Environmental Research Labor-
atory of the Environmental Protection Agency at
Research Triangle Park, North Carolina (IERL-RTP),
involves the testing of feedstocks and waste
streams associated with energy or Industrial pro-
cesses in order to define control technology needs.
Four goals are recognized for environmental source
assessment Programs: (0
• Evaluation of physical, chemical, and
biological characteristics of all pro-
cess streams
• Prediction of environmental effects
• PrioHtizatlon of streams relative
to hazard
• Identification of control technology
programs.
Biological testing, along with physical and chemi-
cal analyses, provides the Information needed to
meet the first three goals. The unique contribu-
tions made by biology to environmental assessment
Include the direct measurement of toxicity of pro-
cess streams to the test organisms and the detec-
tion of synerglstic or antagonistic effects often
characteristic of complex chemical mixtures. The
objective of this paper 1s to provide an overview
of the bloassay protocol used in environmental
assessment and give an example demonstrating the
application of this protocol.
Phased Approach
The biological, chemical, and physical data
requirements for environmental assessment are con-
siderable. IERL-RTP has developed a phased ap-
proach to environmental assessment to efficiently
organize the data collection and analysis efforts.
The phased approach may be characterized as both
comprehensive and cost-effective.O) It 1s com-
prehensive in that all feed and waste streams are
tested. It 1s cost-effective because the analyti-
cal procedures are implemented sequentially with
each successive step or phase building on informa-
tion gained in previous phases.
The phased approach has three levels with each
level designed to achieve specific objectives
(Table 1).(2) Level 1 provides a preliminary screen-
Ing of all process feed and waste streams using
simple, inexpensive sampling and analysis proce-
dures. It 1s meant to provide qualitative or semi-
quantltatlve Information on these streams so that
their relative hazard may be estimated. Level 2
validates and expands on the data obtained in
Level 1. The test procedures are generally more
complex and costly than those used In Level 1.
Level 2 efforts are focused first on those streams
Identified in Level 1 as being the most hazardous.
Streams with low toxicity or hazard have second
priority. Streams definitively classed as non-
toxic (using biological, chemical, and physical
analyses) in Level 1 need not be tested in Level 2.
The third level deals with process variations and
chronic toxicity. It 1s facility-specific with
the test protocol tailor made for the particular
facility being assessed. Currently, a sampling and
analysis protocol exists for Level 1 biology.
Level 2 procedures are under development but a com-
plete protocol 1s not yet available. Level 3 tests
are facility-specific and a test protocol can be
developed only when the particular facility 1s Iden-
tified and Level 1 and 2 testing Is completed. The
remainder of the discussion concerning the phased
approach focuses on Level 1 biological testing.
TABLE 1. OBJECTIVES OF THE PHASED APPROACH
Level 1 -
0 To provide preliminary EA data
• To Identify problem areas
• To generate data for pHoritization of
process feedstocks and waste streams
Level 2 -
• To validate and expand Information from
Level 1
Level 3 -
• To monitor process and time variations 1n
toxicity
• To monitor sublethal, chronic effects
38
-------
The Level 1 bloassay protocol was developed
by the Bloassay Subcommittee, a subset of EPA's
Environmental Assessment Steering Committee. The
protocol consists of 12 health and ecological
tests which are briefly described In Table 2.
Liquids are sampled using heat exchange, tap sam-
pling, and dipper sampling methods. The biological
tests do require considerably larger volumes of
liquids (~240l) than do the chemical tests. Solid
samples are collected using grab sampling techiques.
TABLE 2. LEVEL 1 HEALTH AND ECOLOGICAL EFFECTS
Test
Ames Mlcroblal
Mutagenesls
Rabbit Alevolar
Macrophage
WI-38
Rodent Acute
Toxiclty
Freshwater Algal
Assay
Daphnla Static
Bloassay
Fathead Minnow
Static Bloassay
Marine Algal
Assay
Grass Shrimp
Static Bloassay
Sheepsnead Minnow
Static Bloassay
Soil Microcosm
Stress Ethyl ene
Sample
Test Objective Quantity
Mutagenlc activity 1 g/10 ml
Cytotoxlclty 1/2 g/50 ml
Cytotoxlclty 1/2 g/SO ml
Whole animal 100 g/1 1
toxoclty
Toxiclty "
Acute toxlclty
Acute toxlclty
Toxlclty
Acute toxlclty
Acute toxlclty
200 1
40 1
Toxlclty 1 9/"l
Toxlclty 1500 1 gas
Test
Organism
Salmonella typhlmurlum
Rabbit lung macrophage
cells
Human lung flbroblasts
Laboratory rat
Selenastrum
caprlcornutum
Daphnla pulex
Plmephales promelus
Skeletonena
costatum
PaUeaonetes puglo
Cyprlnodon varlegatus
Intact soil system
Soybean
Test
Time
2-4 days
2-4 days
2-4 days
14 days
14 days
4 days
4 days
14 days
4 days
4 days
3 weeks
28 hours
Test Results
•*•/- (mutageneclty)
EC50
EC50
LD50. necropsy
observations
EC50
LC50
LC50
EC50
LC50
LC50
DOC, Ca transport.
total Ca. ATP
Ethyl ene production
These tests are organized Into protocols that per-
mit solid, liquid, gaseous, and suspended partleu-
late samples to be properly analyzed (Figures 1 and
2). Level 1 samples for biological testing are
collected using the procedures developed to collect
Level 1 chemistry samples.(3) The Source Assess-
ment Sampling System (SASS) is used for the partic-
ulate and sorbent (XAD-2) samples (see Figure 2).
A large volume of gas (IBOOl) Is needed for Level 1
biology. Occasionally the quantity of sample avail-
able for Level 1 testing 1s limited, and the full
biological test protocol cannot be implemented.
Under such circumstances the tests have been prior-
itized and are Implemented in the following order:
i
Isoll
Microcosm
Ames
Test
1 |
[ HI-38 |
LIQUIDS
AND
SOLIDS
1
Rodent 1 Martw or Fre
Toxlclty 1 Ecology
1 |
I
Freshwater -
Algae Bottle Test
Fathead Minnow Toxlclty
Daphnla Toxlclty
shwater
late
T
Marine -
Algae
Sheepsnead Minnow
Grass Shrimp
Figure 1. Bloassay Protocol for Liquid and Solid Streams
39
-------
, 1
Gaseous Grab
Sample
GASES AND SUSPENDED
PARTICULATE MATTER
(DUCTED AND FUGITIVE)
>
Parti cul ate Matter
1 1
Plant Stress
Ethyl ene
Soil Microcosm Ames Test
Extract from Sorbent
J
1 1 1
RAM
Ames Test WI-38
Figure 2. Bioassay Protocol for Gases and Suspended Participate Matter Streams
1. Ames, WI-38. and RAM
2. Rodent Acute Toxicity
3. Marine or Freshwater Tests
4. Soil Microcosm
The stress ethylene test does not appear 1n this
scheme since it 1s the only biological test for
gaseous samples.
Many of the samples collected for the Level 1
bioassays can be tested without any processing or
treatment. Others require some type of prepara-
tion before the biological tests can be performed
on them (Figure 3, Preparation). Preparation
process. Liquids with suspended solid may require
filtering before testing. Solid samples are ground
to resplrable size (<5 um) before being used 1n the
health tests. Solid streams that may be subjected
to leaching by surface or groundwater when they
are stored or disposed are tested using the aquatic
tests. Appropriate quant1tes of leachate are pre-
pared using the 30-m1nute shake technique. CO
When the samples have been properly prepared, they
are analyzed using the test protocol Indicated 1n
Figures 1 and 2.
Once developed, 1t was necessary to validate
the bloassay protocol to determine 1f Level 1
SAMPLE FOR BIOLOGICAL AHALTSIS
"ASES AKI> SUSrUNDKUl
PARTICULATE HATTER
LIQUIDS
Caieoua
Crab Sample*
Participate*
ISorbciul
I Extract I
Aquaou*
(<0.2Z
organic]
(OLIOS
IOrganic I
With
impended
Sol Ida
Solvent
Exchange
filter
(2 •• »Uva)
Aqueous
Extr.icc
Grind to .
" J i
i
Figure 3. Biological Sampling and Preparation Overview
procedures include extraction, solvent exchange,
filtering, grinding and sizing, and aqueous extract
or leachate production. Extraction 1s used for the
SASS XAD-2 column to remove the adsorbed organics.-
Solvent exchange 1s used wherever a sample 1s
deemed Incompatible with the test procedure. Di-
methyl sulfoxide (DMSO) is often used in the
objectives could 1n fact be met with this series of
tests. Four pilot studies were selected for tes-
ting the bloassay protocol. The Flu1d1zed Bed Com-
bustor (FBC) environmental assessment program sup-
ported by IERL-RTP was chosen to conduct one of
the pilot studies and the results of that study are
reported here.
40
-------
Fluidized Bed Bloassay Pilot Study
The FBC bioassay pilot study was performed on
samples collected from the Exxon pressurized FBC
mlnlplant located in Linden, New Jersey and was
part of the Comprehensive Analysis of this facil-
ity. The pilot study had two objectives:
1. Test the applicability of the Level 1
bioassay protocol
2. Provide Level 1 environmental data
for the Exxon FBC.
Pilot study work was initiated 1n April, 1977 and
completed in February, 1978.
A schematic of the Exxon FBC mlnlplant 1s
given in Figure 4 showing the feed and waste
The bioassay test protocol was implemented on
the 13 samples from the Exxon FBC mini pi ant (Table
4). A total of 67 tests were required to complete
the pilot study. Although only marine or fresh-
water tests (depending on the type of water body
receiving the effluents) would normally be used for
most environmental assessments, both sets of tests
were employed for the pilot study in order to more
fully test the validity of the bioassay protocol.
The results of the tests are essentially complete
and reported in Table 5.
Interpretation of the results and the priori-
tization of the streams according to toxicity is
currently being performed by both the Bioassay Sub-
committee and Battelle. Preliminary analysis indi-
cates that the scrubber slurry and fine particu-
lates are among the more toxic streams. The bed
discard leachate was relatively toxic to
ORIFICE
r—•--
COOLING
WATER
PIC
C j Circled numbers denote
^^ streams sampled
14
AUXILIARY fX,
AIR I I
COMPRESSOR tJ
O Circled numbers denote streams sampled, uncircled numbers were not sampled.
Samples Include la, Ip, 16, 19, 2, 3, and 35.
Figure 4. Exxon Flu1d1zed-Bed Combustion Mlnlplant
streams. Six of these streams—the flue gas (la
and Ip), the fly ash from the second cyclone (2),
the bed discard (3), the coal feed (16), the dolo-
mite feed (19), and the slurry from the flue gas
scrubber (35)--were collected using Level 1 proce-
dures and returned to the laboratory for appropri-
ate sample preparation. After preparation a total.
of 13 samples were available for Level 1 biologi-
cal analysis (Table 3).
terrestrial and freshwater organisms but not to
saltwater species. Other samples such as the bed
discard and dolomite showed very little or no toxi-
city. It must be stressed that biological test
results and the ranking of streams are relative and
determination of absolute toxicity 1s beyond the
scope of Level 1 testing. The bioassay results have
not yet been compared with those from Level 1 chem-
istry so no conclusions are possible concerning the
possible correlation between the two types of Level
1 procedures.
41
-------
TABLE 3. THE EXXON FBC STREAMS COLLECTED FOR THE PILOT STUDY AND
THE THIRTEEN SAMPLES. RESULTING FROM PREPARATION PROCEDURES.
WHICH WERE TESTED USING BIOASSAYS
Stream
Gas
~T. Flue Gas (1 a and 1 p)
Liquids
2. Scrubber Slurry (35)
Sol Ids
3. 2nd Cyclone Discard (2)
4. Bed Discard (3)
5. Coal (16)
6. Dolomite (19)
Collection Saople
Procedure Preparation
SASS (1
SASS 11
SASS 11
Grab (1
Grab
Grab
Grab
Grab
Grab
Grab
Grab
Grab
Grab
a)
a I
a I Sorbent Extraction
p
—
Grind
Aqueous Leach
Grind
Aqueous Leach
Grind
Aqueous Leach
Grind
Aqueous Leach
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
Sample
Tested
Coarse Partlculates, > 3um
Fine Partlculates, < T urn
XAD-2 Extract In Nethylene Chloride
Gas
Scrubber Slurry, Aqueous
2nd Cyclone Discard, < 5 um
Leachate of 2nd Cyclone
Bed Discard. < 5 u«
Leachate of Bed Discard
Coal. < 5 u»
Leachale of Coal
Dolomite. < 5 urn
Leachate o? Dolomite
TABLE 4. BIOASSAY TEST MATRIX FOR EXXON FBC PILOT STUDY
BIOASSAY
Health
1. Ames
Cytotoxlclty
2. RAM
3. WI-38
4. Rodent Toxlclty
Ecological
Aquatic
Freshwater
5. Algal
6. Daphnla
7. F1sh
Saltwater
8. Algal
9. Shrimp
10. Fish
Terrestrial
11. Soil Microcosm
12. Stress Ethyl ene
SAMPLE
in
CO
id
3
U
J_
£
01
M
t.
a
3
X
X
X
M
S
0
u
(.
£
X
X
X
*»
|
X
Ul
CM
1
X
X
M
S
X
^
k
3
i.
S
J3
3
L.
X
X
X
X
X
X
X
X
X
X
IB
0
a
01
o
£>
o
CM
X
X
X
X
w
a
^^
u
«
at
c
o
5*
o
•o
c
CM
X
X
X
X
X
X
X
X
X
X
•g
S
I/I
5
1
X
X
X
X
a*
0
^E
y
at
T
S
M
5
i
X
X
X
X
X
X
X
X
X
X
1
X
X
X
X
w
5
u
Id
01
«d
3
X
X
X
X
X
X
01
£
£
8
X
X
X
X
Ol
Id
.c
u
M
Ol
_j
a*
^i
I
X
X
X
X
X
X
42
-------
TABLE 5. BIOASSAY RESULTS FOR FBC STUDY
Test
Ames
Cyto toxlclty
RAH
Rodent Toxlclty
Aquatic
Freshwater
Algal
Daphnla
Fish
Saltwater
Algal
Shrlnp
Fish
Terrestrial
Soil Hlcrocosa
Stress
Ethyl ene
Test
Parameter
V-
EC
50
LOM
LC50
LCSO
ECSO
LC50
LC50
Ranked In
order of
toxlclty
Percent of
Increase
over cor
trol
3
+
NT
5
«J
3
u
£
=
*
596-
3000
6
C
X
UJ
CNJ
1
s
HA
HA
s
NT
u"
3
fc
"?
-
> 146
NT
91
31.61
67.91
151
181
70.51
NT
«
u
c
o
«
CM
-
NT
NT
2
£
U
S
«u
c
0
s
c
CJ
NA
NT
NT
VLT
91.61
VLT
VLT
72.51
NT
3
U
VI
O
cS
-
NT
NT
7
2
2
S
0
2
-
NT
NT
451
40.91
25.31
NT
NT
HT
1
I
'
HT
NT
4
u
-------
Allen, J.M., J.E. Howes, Jr., S.E. Miller, and
K.M. Duke. 1977. Draft Report on Comprehen-
sive Analysis of Emissions from Exxon Fluid-
ized-Bed Combustion Mlnlplant Unit. Battelle
Columbus Laboratories, Columbus, Ohio.
44
-------
SOURCE ASSESSMENT SAMPLING SYSTEM-DESIGN
DEVELOPMENT AND CALIBRATION
D. E. Blake and J. M. Kennedy
Acurex Corporation
485 Clyde Avenue
Mountain View, CA 94042
Intrpductlon
The Process Measurements Branch, IERL/RTP,
has developed strategy for sampling and analysis
1n Environmental Assessment Programs. Three levels
of sampling/analysis detail are specified. Level 1
measures organic and Inorganic mass emissions
semiquantitatively (within a factor of 2 to 3).
Levels 2 and 3 provide quantitative and/or
continuous monitoring of specific pollutant species.
The Source Assessment Sampling System (SASS)
Is the primary sampling tool for Level 1 gaseous
and partlculate emissions. The SASS train performs
the following functions:
i Extractive sampling of gaseous streams
from ducts or stacks,
• Measurement of partlculate mass loading
and size distribution,
• Collection of organic species for subsequent
analysis, and
• Collection of vaporous trace elements for
subsequent analysis
In addition to these functional requirements,
the SASS train must be portable, corrosion resis-
tant, easily cleanable, reliable, and accurate.
Included 1n this paper will be a review of the
SASS design philosophy, followed by a brief descrip-
tion of the Individual components of the SASS train,
a review of tradeoffs and alternate configurations
considered during conceptual design, a description
of cyclone calibration methods and their results,
and a 11st of suggested modifications and Improve-
ments to the SASS. These topics, and others, are
covered 1n greater detail 1n Reference 3.
The SASS design and development program was
sponsored by the Industrial Environmental Research
Laboratory (RTF) of the U. S. Environmental
Protection Agency. Project Officer for the program
was Mr. William B. Kuykendal of the Process
Measurements Branch; his support and guidance are
gratefully acknowledged.
Design Philosophy
At the start of the SASS development program,
Level 1 procedures had not yet been established,
and the function of the SASS not fully defined.
The basic SASS design philosophy was established
at a meeting at Acurex 1n March, 1976. Attending
were representatives of Acurex, EPA, and TRW, Inc.
Figure 1 shows the design schematic determined at
that meeting for the complete SASS train. This
basic design concept has been adhered to throughout
the course of the SASS development program.
Figure 1. Design schematic - SASS train.
In addition to establishing the design philoso-
phy, at the beginning of the development program the
EPA suggested guidelines for the detailed design
phase. These Included: '
1. The cost of the completed SASS train was to
be kept as low as possible. A target figure of
$17,000 for a barebones, but complete, SASS train
was suggested. An additional cost of not more
than $3,000 was allowable for an automatic control
feature in the event that such a feature was
judged desirable.
2. The SASS design was to be made as inter-
changeable as possible with the High Volume Stack
Sampler (HVSS). This was felt to be desirable
because many of the potential users of the SASS
train already owned an HVSS. In some cases it
might be possible to upgrade an HVSS to a SASS,
cutting costs considerably to users.
3. Cyclones, rather than a stage impactor, were
specified as the method for determining partlculate
size distribution. The primary reasons for choosing
cyclones was the desire to collect large partlculate
samples (~1 gram) for subsequent chemical and
biological analysis, and the requirement for
trouble-free field use.
It was considered necessary to maintain a
constant sampling flowrate through the cyclone
assembly, since cyclone collection efficiency varies
with gas flowrate.
4. The probe and oven were to be heated to
eliminate the posslbllty of water, acid, or organics
condensing.
5. The organic sorbent material in the organic
module was to be held at 20°C.
6. A flowrate of 0.00189 m'/sec (4.0 scfm) was
to be maintained at the cyclones.
7. The only acceptable materials of construction
for the SASS parts that would contact the sample
stream were to be Type 316 stainless steel, fully
fluorinated Teflon, or Pyrex glass.
8. The SASS was to be designed for ease of sample
recovery and post-test cleanup.
45
-------
The design of the SASS as ultimately construct-
ed reflects our best solution to these conflicting
requi rements.
SASS Design and Construction
The complete SASS train is shown in Figure 2.
A few basic components of the train such as the
heated probe, the vacuum pumps, and the control
module are the same ones used previously for the
High Volume Stack Sampler (HVSS). The heated
cyclone-filter particulate collection system,
the organic module, and the impinger/trace element
collector are original designs of the SASS program.
Figure 3 shows a detailed schematic diagram of
the Source Assessment Sampling System as constructed.
Onn With Cyclone*
Cornprcucrt
Figure 2. Source Assessment Sampling
System (SASS).
Figure 3. Schematic of SASS.
Heated Probe
The SASS probe extracts gas/particulate samples
from the source being tested, monitors the tempera-
ture and gas velocity of the source, and maintains
sample temperatures above the condensation point
of water/SOs mixtures.
Some important features of the probe (shown
disassembled in Figure 4) are the Type 316 stain-
less steel sampling tube; the fiberglass-insulated
strip heater wrapped around the sampling tube; a
round probe body to allow sealing of the sampling
port and rotation of the probe as necessary; strain
relief for all electrical, thermocouple, and pilot
line connections; a calibrated S-type pitot tube;
and easily interchangeable probe tips with avail-
able diameters from 0.64 cm (1/4 in) to 1.92 cm
(3/4 in) in 0.16-cm (1/16-in) increments. The
standard probes come in lengths of 3, 5 and 10
feet and are designed to withstand duct temperatures
of 330°C.
No,,..
Heating Tap*
S1M3 Tempeictur*
'. '
Figure 4. Probe
A stainless steel liner is used in the standard
probe rather than a glass liner for two reasons.
First, the durability of the metal liner is much
greater than a glass liner. And second, even though
glass is much more resistant to corrosion, there is
so much stainless steel in the cyclones, filter,
and organic module that the small additional amount
in the probe was felt to be unimportant.
Particulate Collection System
The purpose of the particulate collection
system is to maintain the sample gas stream at 205°C
while collecting the particulate in three cyclones
and a backup absolute filter. Figure 5 shows the
three cyclones and the filter holder inside the
cyclone oven.
The nominal cut sizes of the three cyclones
are 10, 3, and 1 urn respectively. Together with
the fiber glass backup filter, the system separates
particulate fraction of the sample into four nominal
size ranges:
• > 10 urn
• 3 to 10 urn
t 1 to 3 urn
i < 1 urn
The design of the small and middle cyclones
(1 and 3 urn) were based on Southern Research
Institute designs that predated the SASS development
program. Applying standard cyclone design equations
to the 10-um cyclone results 1n a collector 75 cm
high by 20 cm In diameter, which 1s far too large
for portable, lightweight sampling equipment.
Professor Andrew McFarlane and one of his
students vU, had developed a design procedure for
smaller "stub" cyclones; his methods were adapted
to reduce the cyclone size. This type of cyclone
46
-------
Figure 5. Cyclones and oven.
can be much smaller because it uses internal vanes
to destroy interior vortices. Figure 6 shows an
exploded view of the 10-um cyclones. Note the
"vortex breakers" ~ the crossed metal pieces --
in the collection CUD and in the outlet tube.
Efficiency is reduced with the stub cyclone design.
For the SASS application the reduced efficiency was
felt to be tolerable, since the large cyclone serves
only as a scalping device.
The bodies of all three cyclones are spun 316
stainless steel to accommodate both the lightweight
and strength requirements. Top and bottom flanges
and inlet and outlet tubes are machined and welded
to the spun bodies.
Close packing of the cyclones and filter system
is essential for small size and light weight. How-
ever, the particulate collection system must also
Figure 6. Large (10 um) cyclone breakdown.
47
-------
be easy to assemble and disassemble since the system
is completely broken down for sample recovery after
each test. Short interconnecting tubing and no
sharp bends are also necessary as tubing can act
as a collector and sharp bends can cause erosion,
damaging the train and/or contaminating the sample.
The cyclone-filter system can be completely
assembled outside the oven as a unit as shown in
Figure 7.
R.lutn Flow To
Implng.f BMh
Cold wn.r From
Implng.r BMh
Hot G.i _
From O»«n
Liquid Pntagt
3-W., Sol.nold V.I..
To H««( Eich«ng»t
In Implngtf CM*
Figure 8. Organic module schematic.
The two primary functions of the organic module
are to cool the gas to 20°C and collect organics on
the sorbent. From a design viewpoint, gas cooling
is the more difficult task. The cooler must remove
about 1 kW of heat from the gas, requiring a
considerable amount of heat transfer surface. The
cooler must also be lightweight, durable, free of
plugging, and easy to clean and recover the sample.
Figure 7. Cyclones — top view.
The oven provides a constant temperature
environment and mechanical protection for the
cyclones and filter. It also supports the probe
by means of a collar attached to its side.
~~ne sample gas leaves the filter holder at
2:05 ''C. cleaned of particulate but still containing
organic and trace element vapors. The organic
nodule (Figure 8) cools the gas stream and directs
the cooled gas and any condensate through an
adsorbent bed. The bed will collect most organic
species of molecular weight greater than about Cfi,
as well as some fraction of the metallic trace
elements present. The condensate is collected and
the cool gas passes to the impingers.
Three separate conceptual cooler designs were
considered: an externally-cooled coil, a parallel
tube heat exchanger (with the sample gas passing
down the inside of the tubes and cooling fluid
outside), and a thin film heat exchanger in which
a thin film of sample gas passes between cooled
walls. As each design was considered in detail,
it became clear that the two most difficult
constraints were the requirement to cool the gas
from 205°C to 20°C and still retain easy disassembly,
cleaning, and reassembly. These constraints favored
a design with a large surface area composed of
smooth, easily accessed surfaces. The thin film
heat exchanger concept was judged clearly superior
in these areas. This cooler could be made from
metal or glass, although metal was chosen for its
durability. Figure 9 schematically shows each of
the gas cooler concepts, and Table 1 lists the
advantages and disadvantages of each.
48
-------
Ho* Oml
W«*r NT
.Cooing
WitarOut
ICootod
Figure 9a. Colled-tube cooler.
Figure 9b. Straight-tube cooler
Cooing i
**r Out I
W«tor Out
Figure 9c. Thin film cooler.
49
-------
TABLE 1. COMPARISON OF GAS COOLER CONCEPTS
Parallel
Tubes
Thin Film
Advantages
Easy to construct. Relatively
inexpensive. Low gas pressure
drop. Capable of adequate
cooling with metallic
construction, but not glass.
Low gas pressure drop.
Capable of adequate cooling
with metallic construction,
but not with glass. Compact.
Cleaning very easy. All surfaces
accessible. Low pressure drop.
Small, compact. Capable of
adequate cooling with either
metallic or glass construction.
Disadvantages
Very difficult to clean.
Must be large and bulky
to achieve necessary heat
transistor.
Moderately difficult to
clean. Complex design.
Relatively costly.
Somewhat more expensive
than the coil. Cheaper
than the parallel tube
design.
Precise temperature control of the sorbent bed
is essential for reproducible organic species col-
lections. Accordingly, the organic module temperature
control system is designed to maintain the sample
gas temperature just upstream of the sorbent
cartridge at any temperature between 15°C and 80°C.
(For Level 1 sampling, 20°C is the required setpoint.)
As the sample gas is cooled, water, acid, and
heavy organics will usually condense. The organic
module is designed so that the cooled gas leaving
the gas cooler section, along with any liquid
condensation formed, passes through the sorbent
bed. The sorbent is typically a porous polymer gas
chromatographic bed packing material. The sorbent
is held within the sorbent cartridge (Figure 10) by
80-mesh Type 316 stainless steel screens. The
cartridge is designed to be easily removed and
replaced with an identical clean cartridge when
several SASS tests are to be made sequentially.
KJUCSH J1KS
UPPER CSIUP BINO
AND 8EAUHO FLAMOI
THREADED CAP NUT
Figure 10. Sorbent cartridge.
50
-------
After the gas and condensate exit the sorbent
cartridge, they pass into the condensate reservoir
section, where the condensate collects and is
periodically pumped to a storage bottle. The gas
exits a tube at the top of the condensate reservoir
section and passes to the impinger/trace element
assembly.
Impinger Assembly
The impinger assembly (Figure 11) collects any
remaining trace elements for subsequent analysis,
and dries the sample gas stream to avoid damaging
the gas pumps and flow monitoring instrumentation.
Four heavy-wall glass bottles contain chemical
solutions or moisture sorbent. The first impinger
bottle contains an oxidation solution of hydrogen
peroxide to collect sulfur oxides. The next two
bottles contain a solution of 0.2 molar armonium
persulfate with 0.2 molar silver nitrate to collect
trace elements. In each of these three liquid-
containing bottles, a straight section of tubing
ducts the sample gas below the liquid levels. The
sample gas bubbles through the liquid, allowing the
various pollutant species to be scrubbed out. The
fourth impinger bottle contains granular silica gel
to dry out the gas.
The standard HVSS four-bottle impinger train
provides the basis for the SASS impinger design,
with one important modification. Because of the
relatively high-pressure drop in several of the
SASS components, the impinger assembly operates at
a substantial vacuum (10 to 20 inches Hg). The
actual volumetric flowrate in the impingers is as
high as 0.0057 m3/sec (12 acfm), leading to exces-
sive splashing of the impinger solutions and
possible solution carryover. In order to eliminate
this problem, special oversized glass impinger
bottles are used.
Vacuum Pumps
Two vacuum pumps connected in parallel are
used with the SASS. These carbon vane-type pumps
are modified with a special shaft seal to reduce
the leak rate to better than Method 5 standards.
Only one of the 3/4-hp pumps is necessary to draw
the 4.0 scfm of sample gas through a clean train,
but once the filter begins to accumulate particulate,
a second pump is needed to make up for the addition-
al pressure drop.
Control Unit
Impingcr train out ol c»ie
Figure 11. Impinger assembly.
The control unit contains all of the instru-
ments for measuring stack velocity, sampling
flowrate and cumulative flow, and temperatures of
various points in the sampling system. The unit
is identical to that of the Method 5 train and
its controls and data readouts are identified in
Figure 12.
Figure 12. Control unit.
SASS Cyclone Calibration
The characterization of the SASS cyclones has
been underway almost continuously since the comple-
tion of the first SASS prototype. Initial efforts
were conducted by Southern Research Institute using
a Vibrating Orifice Aerosol Generator. Later cali-
bration tests were performed at Acurex using a
different method involving dispersion in air of
51
-------
polydisperse aluminum spheres. All tests were
done at normal cyclone operating conditions
(0.00189 m'/sec [4.0 scfm] and 205°C [400°F]). At
the time of writing, results have been obtained
with both methods that are reasonably consistent
and are believed to represent the actual performance
of the cyclones.
The object of the various cyclone calibration
tasks ultimately 1s to determine the cyclone
efficiency curve; from the curve can be obtained
a commonly used f1gure-of-mer1t for the cyclone
called the DSQ cut diameter. Figure 13 Illustrates
these concepts. The efficiency of particle collec-
tion is plotted against the particle diameter. For
each particle diameter, therefore, the effectiveness
of the cyclone 1s determined. For example, Figure
13 shows that for this particular (fictitious)
device, 1f a large number of 2.5-um diameter
particles are introduced, 17.5 percent will be
collected and 82.5 percent will pass through
uncoilected. The particle diameter at which half
of the particles are collected 1s the DSQ cut
diameter; Figure 13 shows the DSQ cut diameter of
that device to be 3.0 urn. The DJJQ cut diameter,
often abbreviated to "cut size" is commonly used as
a rough indication of the collection cutoff of a
cyclone.
uo
0.0
oa
0.7
0.6
as
0.4
OS
02
0.1
0
2 3 4 5 6 7 8 • 10
PARTICLE AERODYNAMIC DIAMETER wn
Figure 13. Typical cyclone fractional
efficiency curve.
Note that Figure 13 expresses particle diameters
as aerodynamic particle diameters. It 1s Important
to distinguish aerodynamic diameters from physical
diameters. The physical diameter 1s the dimension
of the particle obtained by physical measurement,
for example with a microscope and reticle. For
nonsymmetrlcal particles, the physical diameter
of a given particle may have several different
values, depending on the measurement axis chosen.
The aerodynamic diameter (sometimes called the
Stokes diameter) is defined as the diameter of the
equivalent spherical particle of unit specific
gravity having the same terminal settling velocity
as the particle 1n question. The advantages of
using the aerodynamic diameter to characterize the
particles used for cyclone calibration are twofold.
First, each particle is uniquely characterized,
Independent of any choice of physical dimension.
Second, and most Important, since the basic cyclone
separation mechanism depends on Stoke's law,
measuring particle diameter in terms of Stoke's law,
behavior assures that calibration data will be
valid over wide ranges of particle size, shape,
and density.
Cyclone Calibration Tests With Honodisperse Aerosol
The Southern Research Institute calibration
efforts which used monodisperse particles, occurred
in two phases. Initial calibrations were made at
room temperature, wtlh the Intent of calculating
cyclone Djfl cut diameters at 400°F by use of
accepted design equations. Based on these intial
calculated values, modifications were made to the
SASS cyclones to shift their cut points closer to
the 1, 3, and 10 urn values desired. It was discov-
ered subsequently that the design equation used
to adjust the cyclone DSQ'S to 400°F was inapplicable
to small cyclones such as the SASS. The second
phase of calibrations at SoRI Involved actual
calibrations at 400°F. These efforts resulted in
calibration data (for the middle cyclone only)
that are believed to be accurate.
The same basic procedure has been used in all
of the SoRI calibration work. The following
description of their experimental procedure 1s
drawn from Reference 2.
The SASS train cyclones were calibrated using
a Vibrating Orifice Aerosol Generator (VOAG). The
VOAG generates monodisperse (all particles the same
size) dye particles of ammonium fluorescein or
Fast Turquoise 8 GLP dye. The VOAG used in this
study was designed and built at Southern Research
Institute. However, similar devices have been
reported by several authors previously, and a
commercial unit 1s available from Thermo Systems,
Inc.
Using VOAG, cyclone efficiency can be determined
as follows. Monodisperse aerosol of a certain
particle size, generated by the VOAG, 1s fed Into
the cyclone to be calibrated. Some of the aerosol
will be collected 1n the cyclone and some will pass
through and be collected on a backup filter. The
mass ratio of partlculate collected by the cyclone
divided by the total aerosol fed to the cyclone
defines the efficiency of the cyclone at that
particle size. The size of the particles generated
by the VOAG 1s then adjusted and the experiment
repeated. After a number of iterations, a complete
fractional efficiency curve results.
52
-------
Figure 14 1s a schematic diagram showing the
operating principle of the VOAG. The dye 1s fed
through a vibrating orifice were 1t 1s atomized
and mixed with air in the drying chamber. The
particle size 1s controlled by the frequency at
which the orifice Is being vibrated and the
concentration of the dye solution. As the particle
stream leaves the drying chamber, it passes through
a charge neutral1zer to reduce agglomeration and
loss of particles due to electrostatic forces.
The advantage of the monodisperse calibration
method 1s the direct measure of collection efficiency
of the cyclone versus particle size. The chief
drawback is that the cyclones experience very low
mass loadings, many orders of magnitude lower than
field conditions.
Table 2 summarizes Southern Research Institute's
calibration of the middle cyclone. Data are shown
both with and without the vortex breakers (Figure 6)
in place. Note that the aerodynamic cut diameter
of the middle cyclone is closer to the desired
value of 3 um when the vortex breaker is absent.
The collection efficiency curves for these tests
are shown 1n Figure 15. Note that in Figure 15
the data are plotted using physical diameter, rather
than aerodynamic diameter.
Dry Air
o Without Vortex Breaker
Vortex Breaker
2 3 4 5 4789-iO
Physical Particle Diameter, Micrometers
Figure 14. Schematic representation of the
Vibrating Orifice Aerosol Generator.
Figure 15. SASS middle cyclone
efficiency curves.
TABLE 2. SUMMARY OF SORI CALIBRATIONS OF MIDDLE CYCLONE
Material
Turquoise
Dye
Turquoise
Dye
Vortex
Breaker
OUT
IN
Temperature
400°F
400°F
Flowrate
ft'/mln
Actual/Standard
6.50/4.00
6.50/4.00
DSO Physical
Micrometers
2.5
3.4
50 Aerodynamic
Micrometers
3.5
4.9
53
-------
Cyclone Calibration Tests With Polydisperse Aerosol
The Acurex SASS cyclone calibration test series
was conducted by an entirely different method than
was used by SoRI. The Acurex method involves the
dispersion of polydisperse (particles of varying
diameters) particles, at concentrations more
representative of actual field conditions. It was
desirable to do a second series of calibration
tests for several reasons:
• A separate series of calibrations using a
different method would — if the results
agreed ~ greatly increase confidence in
the correctness of the calibrations
• The extremely low particle mass concentra-
tions with the monodisperse (SoRI)
calibration method required confirmation
at more realistic loadings
• The physical state of the SoRI dye particles
was unknown. There was a feeling that
any stickiness caused by hydroscopic
absorption or incomplete solvent evapora-
tion might bias the results.
• At the time the polydisperse tests began,
problems with the SoRI method at elevated
temperature made ultimate success with
that method uncertain
Figure 16 shows the test apparatus In schematic
form. Hetered amounts of air and test dust are
combined in a powder feeder. The powder feeder is
so designed that the dust particles are deagglorn-
era ted and suspended in the air. The dust cloud
then enters a heater, where its temperature 1s
raised to the desired level. The hot dust cloud
then enters the cyclone being evaluated. Each
particle is either captured by the cyclone (ending
up in the cyclone cup) or exits the cyclone and
is captured on an absolute backup filter. Clean air
is exhausted from the filter holder to the room.
Itotared Ak-
Durt-
Cycton*
Figure 17 shows the apparatus in more detail.
Compressed air is filtered, regulated, and then
passed through a square-edged critical flow orifice.
The flowrate through this type of orifice depends
only on upstream pressure, so long as the temperature
remains constant and the downstream absolute pressure
is less than about half of the upstream pressure.
In the calibration apparatus these conditions are
easily met, so a constant volumetric flowrate of
0.00189 m'/sec (4.0 scfm) was held simply by main-
taining a constant reading on the upstream pressure
gauge.
Figure 16.
Polydisperse cyclone calibration
apparatus schematic.
Figure 17. Detail of cyclone
calibration apparatus.
The carrier air — clean, dry, and at constant
flowrate — now enters the dust feeder. The dust
feeder is of the grooved-disc type, in which the
test dust 1s metered by means of pneumatic unloading
of a groove cut in a rotating disc. The size of
the (powder-filled) groove and the speed of
rotation of the disc determine the rate at which
test dust is fed Into the calibration system.
The test dust 1s mixed with the metered air
stream in the dust feeder outlet tube. The dust
cloud velocity in the outlet tube is deliberately
held at near sonic conditions in order to assure
maximum dispersion of the dust particles. At the
point the outlet tube enters the heater, a step
enlargement in the diameter of the tube reduces
the gas velocity through the heater to about
120 m/sec. The heater Itself is a stainless steel
tube wrapped with a heating tape. The wall tempera-
ture of the tube 1s maintained at about 330°C,
allowing the dust cloud to reach 205°C by the time
it exits the 0.5-meter long heater. Feedback
temperature control of the exit cloud temperature
1s maintained by a thermocouple and temperature
controller.
The cyclone being calibrated is attached to the
outlet of the heater, and 1s wrapped with thermal
Insulation during calibration tests so the 205°C
operating temperature 1s maintained. A standard
SASS filter holder supports a glass fiber absolute
filter.
For the polydisperse aerosol calibration tests,
an aluminum powder with a size range from 1 urn to
25 urn and a mass median diameter of 6 urn was used.
The data reduction method requires measurement of
the size distribution and quantity of dust collected
in both the cyclone dust cup and on the filter.
Coulter counter measurement was chosen for deter-
mining the distribution, since for the particle
54
-------
sizes of interest it was known to give reliable
and reproducible results.
From this information a simple material balance
on each differential element of particle size allows
reconstruction of the distribution of the test dust
entering the cyclone, no matter how it may have
been changed during passage through the dust feeder
and heater. This method of analysis was used for
all of the cyclone calibrations reported here.
Details of the data reduction method are included
in Reference 3.
Two complete sets of SASS cyclones were cali-
brated using the polydisperse method. Each set
consisted of three cyclones — one small, one
medium, and one large. The first set calibrated
was a part of an EPA-owned SASS train. The large
and medium cyclones were calibrated both with and
without their swirl breakers; thus a total of five
complete cyclone calibrations were obtained.
Figures 18 and 19 show scanning electron photomicro-
graphs of the test dust and representative samples
of dust collected by the cyclones. The spheroidal
nature of the particles can be seen.
Figure 19b. Large cyclone catch.
Figure 20 shows the calibration results for
the first set of SASS cyclones. Table 3 shows the
DSQ cut diameters for each of the five cyclone
configurations expressed as both physical and
aerodynamic diameters. The SoRI DSQ values for
medium cyclone are shown for comparison. The
agreement between the two methods is good.
Figure 18. Aluminum test dust.
Physical particle diameter.
Figure 19a. Small cyclone catch.
Figure 20.
Physical particle
diameter, urn.
55
-------
TABLE 3. SUMMARY OF CALIBRATION RESULTS
Cyclone DSQ Cut Diameter, pm
Cyclone
Large
(with SB)
Large
(w/o SB)
Medium
(with SB)
Medium
(w/o SB)
Small
Aerotherm
Physical
6.6
6.2
3.65
2.18
1.05
Aerodynamic
10.8
10.2
6.0
3.6
1.55
SoRI
Aerodynamic
4.9
3.5
o.*
o.;
o.t
05
T«t *at - t\ml*m
t«it *at ttmm • I.I tit*'
Ulltratta
Callfcfttlw
Ptrtlc
tr teller
.J .< .5 .§ .7.H.?1 ? J
n^llul Mrttclt 4t«wt*r. «•
5 I 7 9 110
A second set of SASS cyclones has recently
been calibrated by this method. The cyclones tested
are a part of a SASS train owned by KVB, Inc.
These cyclones were calibrated with the swirl
breakers removed from the large and middle cyclones.
Figure 21 compares the calibration results for the
KVB and EPA SASS cyclones. The two cyclone sets
compare quite well for the large and medium cyclones,
and reasonably well for the small cyclone. This
1s to be expected, as all of the uncertainty factors
In the calibration method ~ particle deagglonera-
tlon, Coulter counter accuracy, particle shape
uniformity -- are more significant for the smaller
particle sizes.
Figure 21. SASS cyclone calibration data.
Table 4 shows the physical and aerodynamic DSQ
cut diameters of the KVB and EPA cyclones. The
DSQ aerodynamic cut diameter of the two SASS
cyclone sets calibrated to date are reasonably
close — averaging 9.7, 3.7, and 1.5 pm — to the
desired cut diameters of 10, 3, and 1 urn that
were established at the start of the SASS develop-
ment program.
TABLE 4. COMPARISON - KVB AND EPA SASS CYCLONES,
AEROTHERM CALIBRATION METHOD
Cyclone
Large*
Med1una
Small
DSQ Cut Diameters, urn
KVB EPA
Physical
5.61
2.30
0.81
Aerodynamic
9.2
3.8
1.3
Physical
6.20
2.18
1.05
Aerodynamic
10.2
3.6
1.7
aSw1r1 busters removed
56
-------
SASS Modifications and Improvements
As with any complex Instrument, field use has
demonstrated the need for changes to the SASS.
Some changes have been accomplished, and some have
been recommended and are being studied. Some of
these recommended changes being studied are
discussed here.
Table 5 lists SASS Improvements and modifica-
tions that have been suggested by a number of SASS
users as a result of testing occurring during the
past year. Most of the suggested modifications
are self explanatory; an exception 1s the concept
of adding 1sok1net1c sampling capability to the
SASS.
TABLE 5. SUGGESTED SASS MODIFICATIONS
Type of
Improvement
Corrosion
Resistance
Convenience
Durability
Improved
Accuracy
Suggested Change
• Precious metal plating of the
sample-contacted parts of the
organic module
t Fully concentric organic module
gas cooler
• All glass/Teflon organic module
• J1g for out-of-oven cyclone
assembly
• Index marks on cyclones
• Ball-and-socket fitting for
filter holder
• Increase 1mp1nger case size.
Insulate.
• Add drain cock to 1mp1nger case
• Check valves 1n 1mp1nger tubes
• Replace 1ce with mechanical
refrigeration system
• Increase cyclone wall thickness
• Tapering Inlet to filter holder
to eliminate filter erosion by
reducing gas velocity
• Heavier cyclone collection cups
• Less fragile condensate collec-
tion bottle and adaptor
• Add 1sok1net1c sampling capabil-
ity to the SASS
It may be desirable for some (as yet undefined)
Level 2 or Level 3 procedures to Improve the accuracy
of particulate size distribution measurements.
Isok1net1c sampling at the nozzle would provide for
this Improved accuracy; however, isoklnetlc sampling
1s not possible with the present SASS train
because of the requirements for constant gas flow-
rate through the cyclones. Figure 22 shows one
way of adding Isoklnetlc sampling capability to the
SASS, while continuing to operate the cyclones at
a constant flowrate.
The SASS train would be basically unchanged
from Its original configuration, except for the
addition of the recycle loop. Isokinetic sampling
would be achieved by first choosing a nozzle size
such that at 0.00189 m'/sec (4.0 scfm) through the
nozzle, and for the maximum stack gas velocity
expected during the test, stack gas and nozzle
velocities are matched. The probe would then be
Inserted Into the stack, and the pi tot reading
used to set the control-module flowrate at a value
that gives Isoklnetlc sampling at the nozzle. This
flowrate will in general be less than 4.0 scfm,
and will vary as the stack gas velocity varies.
Throughout the duration of the test, the flowrate
through the train (as measured at the control
module) will be changed frequently to reflect
changes 1n stack gas velocity, temperature, or
pressure.
The purpose of the recycle loop shown on
Figure 22 1s to automatically maintain a constant
4.0-scfm flowrate at the cyclones while the train
flowrate varies. The small cyclone is used as
differential pressure flowmeter. A gas pump moves
the recycle gas stream from the low-pressure point
after the filter to the high-pressure point upstream
of the first cyclone. The recycle loop is fully
automatic in that any deviation from a flowrate of
4.0 scfm at the small cyclone immediately actuates
the recycle stream valve to correct the flowrate
to that value.
Figure 22. Isokinetic SASS.
57
-------
REFERENCES
1. Ancel, J. E., "Development of a Cyclone for In-Stack Particle Sampling." M.S. Thesis. University of
Notre Dane. Notre Dame, Indiana. August 1973.
2. Gushing. K. N. et. al.. "Partlculate Sampling Support: 1977 Annual Report." EPA-600/7-78-009.
January 1978.
3. Blake. D. E., "Source Assessment Sampling System-Design and Development," EPA-600/7-78-018.
February 1978.
58
-------
FIELD EVALUATION OF THE SASS TRAIN AND LEVEL-1 PROCEDURES
Franklin Smith
Eva D. Estes
Denny E. Wagoner
Research Triangle Institute
Research Triangle Park, North Carolina
This paper presents the results of a two-phased
evaluation of Level-1 environmental assessment pro-
cedures. Phase I was a field evaluation of the SASS
train. Three sample runs were made with two SASS
trains sampling simultaneously and from approxi-
mately the same sampling point. A Hethod-S train
was used to estimate the "true" particulate loading.
Comparisons of the SASS trains are made for total
particulate, particle size distribution, organic
classes, and trace elements. Phase II consisted
of providing three participating organizations with
control samples to challenge the spectrum of Level-1
analytical procedures. Estimates of intra- and in-
terlaboratory precision are made.
Introduction
An experimental program designed to evaluate
the source assessment sampling system (SASS) and the
associated Level-1 analytical procedures has been
completed. The project was conducted in two phases.
Phase I consisted of a field evaluation of the SASS
involving simultaneous sampling with two SASS trains
and a Method-5 train. Results of Phase I are used
to estimate within and between train precisions for
particulate, organic, and inorganic sampling, and to
estimate the biases of the SASS trains with respect
to Method 5 for total particulate determinations.
Phase II consisted of an interlaboratory evaluation
of the analytical methods involving the analysis of
split samples by participating laboratories.
The Research Triangle Institute (RTI) coordi-
nated the experimental program with Arthur D. Little
(ADL), Southern Research Institute (SoRI), TRW, and
Radian Corporation. ADL analyzed all the field sam-
ples collected in Phase I of the program and pre-
pared and analyzed the control samples used in Phase
II of the program. SoRI and TRW each provided a
field crew and a SASS train for Phase I and partici-
pated in Phase II by analyzing the control samples
provided by ADL. Radian Corporation provided a
field crew and a Method-5 train for Phase I and par-
ticipated in Phase II by analyzing the control sam-
ples.
As stated above, the objectives of this project
were; (a) to evaluate the SASS, and (b) to evaluate
the analytical procedures. Assessment of field crew
and/or analyst performance was not a program objec-
tive. Actions taken to eliminate or minimize ex-
traneous sources of variability in the field evalua-
tion of the SASS included the following:
a. Each participating organization was re-
quested to provide a crew experienced in
the operation and field use of SASS.
b. Field crews were briefed on and directed
to use the approved and documented Level-1
sampling procedures.1
c. RTI provided on-site coordination of the
field sampling activities.
d. Calibration checks were made on the vol-
ume measurement systems (dry gas meters)
of the SASS trains and on the gas veloc-
ity measurement systems of the two SASS
trains and the Method-S train.
e. All field samples were analyzed by one
organization (i.e., ADL), eliminating the
between laboratory component of variabil-
ity of the analytical methods.
In an effort to minimize analyst/laboratory
biases in Phase II of the program, RTI personnel
visited each organization to discuss the analytical
procedures and to review the laboratory facilities
and apparatus to be used in the analysis of the con-
trol samples. Also, as RTI analyzed the data for
the final report, outliers or suspicious data were
brought to the attention of the reporting organiza-
tion for verification and/or correction as appro-
priate.
A description of the test plan for the field
evaluation of the SASS and for the interlaboratory
evaluation of the analytical procedures is given in
Section 2. Results of the field evaluation of the
SASS are presented and discussed in Section 3. The
interlaboratory evaluation of the analytical proce-
dures is described in Section A. A brief summary
and interpretation of the results of both phases of
the program are contained in Section 5 of this pa-
per.
Discussion of Experimental Test Plan
Procedures for Level-1 environmental assess-
ments for both sample collection and sample analyses
have been specified by the Process Measurements
Branch (PMB) of the Industrial Environmental Re-
search Laboratory (IBRL).1 In order for the Level-1
procedures to be effective, the precision and accu-
racy of both the sample collection and sample analy-
sis phases of the measurement process must be suffi-
cient to satisfy Level-1 data quality requirements.
The primary procedure for characterizing gase-
ous process streams in environmental assessments is
to use the SASS for sample collection and specified
analytical methods for subsequent sample analysis.
The SASS and some of the analytical methods, at
least for this application, are still in the devel-
opmental stage to the extent that prior to this
study they had not been subjected to collaborative
(or interlaboratory) tests. The purposes of this
project were to evaluate the SASS under field con-
ditions (Phase I) and to conduct an interlaboratory
evaluation of the associated analytical methods.
59
-------
Phase I. SASS Train Evaluation
The SASS train evaluation test plan—starting
with source selection criteria, continuing through
sanpling requirements, and ending with directions
for sample analyses—is delineated in the following
paragraphs.
Source Selection Criteria
Criteria used in the source selection process
were:
1. The process stream should be sufficiently
high in organics and particulate to provide a
stiff challenge of the SASS train.
2. The process stream should be sufficiently
stable to allow for comparison of data between
days or runs.
3. The process stream must be amenable to
this test in terms of: space for simultaneous
operation of two SASS trains and a Method-5
train, available electrical power to operate
the trains and two mobile laboratories, and a
physical stack or duct configuration such that
sampling port locations for Method 5 are con-
sistent with criteria set forth in EPA Refer-
ence Method I.2
Field Sampling
Samples were collected with the two SASS trains
and the Method-5 train running simultaneously.
Three complete sample runs were made. The relative
positions of the trains were fixed with the probes
of the two SASS trains positioned at a point of
average duct velocity and within a few inches of
each other. The Method-5 train was positioned down-
stream from the SASS trains and operated according
to the Federal Register method, i.e., the duct was
traversed and iaokinetic sampling conditions were
maintained.3 The test site configuration is shown
in Figure 1.
Analysis Scheme for Field Samples
To insure consistency, all analytical work for
SASSj
METHODS
TRAVERSE POINTS
IJ'1"
T
IT'
1
FHwra 1. TEST SITE CONFIGURATION FOR FIELD
EVALUATION OF THE SASS.
60
-------
Phase I was done by one organization. Table 1 sum-
marizes the analyses performed on one set of SASS
runs. For the other two runs, the only analyses
performed were gravimetric analyses of particulate
for the cyclones, filter, and rinse. The analysis
scheme of Table 1 is described in the following
listing.
Table 1. PROCEDURES FOR ANALYSIS OF
A SELECTED PAIR OF SASS RUNS
1.
Particulate
a. For each SASS train run, the partic-
ulate on the filter and in each cy-
clone was dryed, then weighed, and
the total weight of particulate de-
termined. The particulate for each
Method-5 run was also dryed and
weighed. Weighings were done in the
field by one person. This allowed
for a comparison of the SASS trains
to each other and to the Method 5 for
each run (sane day) and on a day-to-
day basis.
b. For one test (two SASS trains and a
Method-5 run simultaneously), the or-
ganics were extracted (soxhlet extrac-
tion) from the particulates and the
particulates reweighed. The SASS
particulate extracts were then ana-
lyzed for volatile (TCO) and nonvola-
tile (Grav) organic material, then
subjected to a full Level-1 organics
analysis, including LC-IR-LRMS.
2. XAD-2 Module
a. For each SASS run, the total weight
(TCO + Grav) was determined for the
condensate and for the combined XAD-2
extract and module rinse.
b. On the same pair of runs selected for
the particulate organic analysis, the
above TCO + Grav determinations were
followed by eight class separations
with a TCO + Grav determination on
each of the fractions. The fractions
were also analyzed by the IR-LRMS
scheme.
3. Impingers
For the same pair of runs selected for
particulate organics analysis and XAD-2 eight
class separation, Eg, As, and Sb were deter-
mined on the combined second and third impin-
gers by current Level-1 methods. No analyses
were done on the impinger solutions for the
runs.
SAMPLE
10 tan CYCLONE
3 urn CYCLONE
1 urn CYCLONE
RLTER
ORGANIC RINSE
(SORBENT MODULE)
AQUEOUS
CONDENSATE
2ND AND 3RD
IMPINQERS
8
x
g
S
i
•^
•X
•"-
*
_
X
Lu
!j«
ii
s^
?^
T
O
c
i
^
^
2
o
Q
a
^
or
o
§
3
-J
C
IB ^5
Ip
8s
i§
1
#
!
Phase II. Verification of the Analytical Scheme
Three sample types were supplied to each of the
three participating organizations for analysis by
current Level-1 procedures. The three sample types
were:
1. A known artificial, liquid sample contain-
ing 8 to 10 components.
2. A real particulate sample obtained from a
source significantly different from the one
selected for Phase-I sampling.
3. The combined XAD-2 extracts from the SASS
runs in Phase I.
Each participating organization was sent three
aliquots each of the three above sample types for a
total of nine samples per participant., The samples
were coded and specific instructions for the analyt-
ical work to be done on each were provided. Each
participant did a full Level-1 analysis on one ali-
quot of each of the above three sample types. For
the other two aliquots of each type, there was a
reduced analysis scheme.
Analysis Scheme for Control Samples
Procedures for analysis of the control samples
for Phase II of the evaluation are summarized in
Table 2 and discussed in the following paragraphs.
1. Sample 1
a. Aliquot 1. This aliquot was taken
through a complete Level-1 organic
analysis beginning with a TCO + Grav.
The sample was then separated into 8
fractions by LC with a TCO + Grav and
IR-LRMS on each fraction.
b. Aliquots 2 and 3. The analysis of
these aliquots involved a TCO + Grav,
8 class separations by LC, and TCO +
Grav on each of the 8 fractions.
2. Sample 2
a. Aliquot 1. The analysis of this
sample type followed the Level-1
scheme for particulates. One por-
tion of the sample was extracted and
a TCO + Grav performed on the ex-
tract. The extract was then sepa-
rated into 8 fractions by LC and a
TCO + Grav and IR-LRMS performed on
each fraction. The remaining par-
ticulate was Parr-bomb combusted and
analyzed by SSMS and by approved
Level-1 procedures for AS/Hg/Sb.
b. Aliquots 2 and 3. These two aliquots
61
-------
were extracted and a TCO + Grav per-
formed on the extract.
3. Sample 3
a. Aliquot 1. Analysis of this combined
extract sample started with a TCO +
Grav followed by the 8 class LC sepa-
ration and TCO + Grav IR-LRMS on each
of the 8 fractions.
b. Aliquots 2 and 3. These two aliquots
involved only a TCO + Grav analysis.
Results of SASS Evaluation
The purpose of an Intel-laboratory test such as
this is to, within the project constraints, deter-
mine:
1. Where possible, the comparability of the
experimental system results with reference
methods or standard material (accuracy),
2. Comparison of results between similar sets
of equipment operated by different laboratories
(reproducibility or interlaboratory precision),
and
3. Comparison of duplicate results from the
same system operated by the same laboratory
(repeatability or intralaboratory precision).
At this time, all of the analytical data from
the SASS evaluation are not available for this pa-
per. Measurements for which data are available and
the order in which they will be discussed are as
follows:
1. Participate concentration determinations
allowing comparison of the SASS with Method 5
and comparison between SASS's for three runs.
2. Particle size fractionation between SASS's
for three runs.
3. Organic material collected between SASS's
for one run.
4. Total, volatile, and nonvolatile organics
by LC fractions between SASS's for one run.
5. Organic categories in sample between
SASS's for one run.
6. IR results (functional groups) for one
sample between SASS's.
7. Number of subcategories and specific com-
pounds identified by LRMS for one sample be-
tween SASS's.
Table 2. PROCEDURES FOR ANALYSIS OF
PHASE II SAMPLES
SAMPLE 1
ALIUUUI 1
ALIQUOTS 2 AND 3
SAMPLE 2
ALIQUOT 1
ALIQUOTS 2 AND 3
SAMPLE 3
ALIQUOT 1
ALIQUOTS 2 AND 3
HLET
RACTION
X K
82
+ GRAV
O
• _
•_
J
•_
+ GRAV
•. ,_
CO
DC
_J
C
RBOMB
BUSTION
OC2
-------
8. Arsenic, mercury, and antimoney determi-
nations for one sample between SASS's.
Particulate Concentration Determinations
Three complete sample runs were Bade with the
two SASS's and the Hethod-5 train sampling simul-
taneously as described in the test plan discussion.
Particulate concentration determined by each SASS
and by Method 5 are given by run in Table 3.
Table 3. PARTICULATE CONCENTRATION ma/m?
X TRAIN
RUN#\
RUN-1
RUN-2
RUN-3
SASS-1
408
399
353
SASS-2
337
349
315
M-5
342
322
371
o - 28 mg/m3 (8%) WITHIN TRAIN PRECISION.
38 mo/m3 (10%) BETWEEN TRAIN PRECISION.
An analysis of variance (ANOVA) was performed
to test at the 0.05 level of significance, the fol-
lowing hypotheses:
1. Hypothesis 1: there are no train effects.
2. Hypothesis 2: there are no run effects.
The ANOVA table is given in Table 4.
From the experimental data, the calculated F
values are 1.48 and 0.13 for trains and runs, re-
spectively. The tabulated value for Fo.96 (2,4) is
6.94. Therefore, neither of the previously stated
hypotheses can be rejected. That is, based on this
set of data we cannot say that the SASS's differ
from each other or that either SASS differs from
Method 5.
From the ANOVA, the best estimate of the pre-
cision of a single observation, regardless of train
(intertrain precision), is (o* + 0*)1'1 = 36 mg/m3.
The coefficient of variation (or relative standard
deviation) is 0.10, or 10 percent.
Intratrain precision is estimated by 0 and is
equal to 28 mg/m'.
The results of this evaluation show that for
this one source the SASS's precision and accuracy
were not significantly different from the precision
and accuracy of the Method-5 determinations.
Table 4. ANALYSIS OF VARIANCE TABLE
Film
Tralm
Total
4.674
3.164
df
1t7
2^37
Particle Size Fractionation
Particulate matter is divided into four size
fractions by the SASS using three cyclones and a
filter in series. From Table 3 in the previous
section, comparison of particulate concentrations
measurements can be made. The particulate concen-
tration, as determined from each cyclone and the
filter, is given as a percent of the total concen-
tration determined by the train in Figure 2.
As seen fro* Figure 2, results from correspond-
ing components of the SASS's compare very well for
Runs 1 and 2. The significance of the differences,
if they are significant, as shown by the lOp cy-
clones and filters in Run 3 will have to be evalu-
ated by individuals knowledgeable on particle size
measurements.
SAS1Z
D
s *
a »
i:
•V
e
^ if
i "
K ,
1
I
11
Cyri
•
wt
1
a
Cyd
u
•M
t»
CydM
TTT[
I
W
Y'
<
n
IP
•>
COMPARttOi OF PARTICLE SUE FHACTI01ATION (Rn II
PERCENT OF TOTAL CATCH
. . s s 8 a «
— » —
I
^™
-))— .
I
__M rV/ll-ix-
i
It,
COMPARBOR0V PARTICLE SIZE FNACTIOIATIOH (Rw 1)
«•
g "
a »
i »
1 a
mV
fi "
S 11
i i
_ii—
I
—«—
I
S
COMTARKO* OF PARTICLE SUE FRACTIOHATION (Ru 1)
Figure 2. COMPARISON OF PARTICLE SIZE
FRACTION OF TWO SASS's
63
-------
Organic Extractables
Extracting organic matter from SASS samples is
an important procedure in the analysis process.
Table 5 compares the volatile (TCO) and nonvolatile
(Grav) organic contents of the samples taken from
corresponding components of the SASS's. The data
show that organic material collected by the corres-
ponding components was comparable in quantity (to-
tal) and in composition (volatile and nonvolatile).
Table 5. ORGANIC EXTRACTABLES (mo/m?)
TCO
GRAV
TOTAl
CYCLONE
SASS-1
am
1.66
1.7
SASS-2
am
1.58
1.6
XAD-2 (EXTRACT)
SASS-1
141
10.2
116
SASS-2
168
8.99
12.6
XAD-2 (MODULE)
SASS-1 | SASS-2
(RINSE)
69
69
81
81
Organics in LC Fractions
In the Level-1 analysis procedures, the sample
extract is separated by silica gel liquid chroma-
tography and a solvent gradient series into 8 frac-
tions of varying polarity. TCO and gravimetric
analyses of each fraction are done to determine the
distribution of the sample by the various class
types.
Comparison of the distribution of the sample in
terms of volatile and nonvolatile organics by LC
fractions for an XAD-2 extract can be seen in Table
6.
As seen from the table, the totals (Z) across
all fractions for TCO and Grav agree very well. The
comparison for individual fractions with few excep-
tions is good.
Organic Categories Identification
Identifications of organic categories in pro-
cess streams are important functions of environmen-
tal assessments. Table 7 lists the categories and
their concentrations as determined from one set of
SASS runs. As seen in the table, the categories
compare well across trains, both qualitatively and
quantitatively.
Qualitatively, only inorganics at 0.1 mg/m3
and silicones at 0.1 mg/m3 were identified in the
SASS-1 sample and not the SASS-2 sample. Quanti-
tatively, when the concentration levels are con-
sidered, only the difference in the heterocyclic 0
concentrations appears to be much larger than de-
sired.
Functional Groups Identified by IR
IR spectnoscopy is used in the Level-1 environ-
mental assessment procedures to determine the types
of functional groups present in a sample. Table 8
shows the results of an IR analysis of the XAD-2
extract for SASS-2. A similar analysis for SASS-1
yielded almost identical results.
The wave number v(cm 1), the intensity level
[weak (w), medium (m), or strong (s)], and func-
tional group symbol or name are given in Table 8.
Subcategories and Compound Identification by LRMS
A low resolution mass spectrum (LRMS) is ob-
tained on all LC fractions that exceed the concen-
tration threshold in order to determine the princi-
ple compound types present in each fraction. A
comparison of the number of compound types identi-
fied by both SASS's and by each SASS separately is
given in Table 9 by LC fraction.
As seen in the totals of Table 9, out of a
total of 41 compounds identified by one or both of
the SASS's, 25 were identified by both SASS's.
Thirteen compounds were identified in the SASS-1
sample only. Before a judgment is made on whether
this is critical or not, the total SASS samples will
have to be compared to determine if these 13 com-
pounds were identified in samples from other parts
of the SASS train. This point will be further stud-
ied and documented in the report to be issued at a
later date.
Arsenic, Mercury, and Antimony Comparisons
Antimony and mercury are determined by atomic
absorption, and arsenic is determined by the silver
diethyldithiocarbonate (SDDC) method in the Level-1
procedures. Table 10 compares the levels of these
elements found in the samples from the two SASS's.
The estimates of precision of analysis were provided
by Arthur D. Little, Inc. The agreement appears
reasonable, based on the precision estimates for As
and Sb. However, the difference in the Hg concen-
trations is larger than would be expected from anal-
ysis imprecision alone. The difference is less than
a factor of two.
Table 6. XAD-2 EXTRACT SUMMARY
TOTAL ORGANICS,
HTQ/nr*
TCO, mg
GRAV. mg
(SASS-1/SASS-2)
LCI
.54
.33
&2
2.6
13
7.5
LC2
.71
2.4
19
35
3.3
38
LC3
ai
as
73
58
182
216
LC4
.95
.59
a?
1.7
23
17
LC5
.36
.25
3.7
1.1
7.3
a7
LC6
1.5
.93
5.3
5.7
41
23
LC7
.47
.24
0.1
0.1
15
13
LC8
.01
.07
0.2
2.1
0
0
2
12.6
13.6
113
106
284
315
64
-------
Table 7. COMPARISON OF CATEGORIES FROM
ORGANIC EXTRACTS
Table 9. SUBCATEGORIES AND SPECIFIC COMPOUNDS
IDENTIFIED BY LRMS (XAD-2 EXTRACT)
CONCENTRATION (mg/m3)
CATEGORIES
ALIPHATIC HYDROCARBONS
HALOGENATED AROMATIC
HC"S
AROMATIC HC*S- BENZENE
<216
>216
HETEROCYCLIC N
HETEROCYCLIC S
HETEROCYCLIC O
PHENOLS
ESTERS
ETHERS
AMINES
AMIDES
CARBOXYLIC ACIDS
SULFONIC ACIDS,
SULFOXIDES
SULFUR
INORGANICS
UNCLASSIFIED
SILICONES
SASS-1
1.1
0.6
28.7
25.6
20.1
Z4
2.2
0.2
0.5
0.6
0.2
0.1
0.3
0.1
SASS-2
1.1
0.1
28.2
28.7
24.0
2.5
6.7
0.3
02
0.7
0.7
1
Table 8. FUNCTIONAL GROUPS IN SASS SAMPLE
IDENTIFIED BY IR
v, cm*1
3400
3050
2950-30
1700
1600
1500
1460-1420
1180
850-700
INTENSITY, ASSIGNMENT
Vw (br)
M
M
W
M
W
M
M
S(MULTIPLE)
NH.OH
AROMATIC. C-H
ALIPHATIC, C-H
ESTER, IMIDE, KETONE
UNSUBSTITUTED AMIDINE
HCI, CARBONATE
C-N-0, N-ON;
RING VIBRATIONS
RING VIBRATIONS
Si-AROMATIC. SCH2-,
SUBSTITUTED PYRIDINE,
ALIPHATIC AND AROMATIC
C-H. SiCH2-
ESTER, SiO-CHa, CaP-0
SUBSTITUTED AROMATIC
OR FUSED RINGS
Results of Analytical Methods Evaluation
Evaluation of Level-1 environmental assessment
methodologies for analysis of SASS samples was per-
formed by providing control samples of three types
to the participating laboratories. The control sam-
ples were prepared and analyzed by Arthur D. Little,
Inc. Arthur D. Little's results are used as •
FRAC-
TION
LC1
LC2
LC3
LC4
LC5
LC6
LC7
LC8
SASS-1 AND
SASS-2
SASS-1
ONLY
3
10 j 3
11
13
6
5
5 1
9
I
LC1-LC8 25 13
SASS-2
ONLY
1
4
2
2
3
TOTAL
3
14
21
20
8
9
\75
41\
Table 10. ARSENIC, MERCURY, AND ANTIMONY
DETERMINATIONS
SASS-1
SASS-2
As
Oig/m3)
0.71
a 83
Hfl,
Wit3)
0.24
0.40
Sb
(M9/m3)
0.06
0.10
ESTIMATED PRECISION OF ANALYSIS
CV(As) - 5%, CV(Hfl) = 10%, CV(Sb) = 25%
fourth set of data for interlaboratory comparisons.
Laboratories or participants are coded as A, B, C,
and D, and are not further identified in this paper.
Data from all participants have not been avail-
able long enough to allow for a rigerous analysis of
the results. Also, we are still in the process of
developing means for presenting these large quanti-
ties of data in a meaningful and effective manner.
Therefore, to provide an overview of the results of
Phase II, and at the same time to hold this paper to
a presentable length, only the data from the organic
analysis of the real sample (sample 3, Table 2) and
the SSMS results on aliquot 1 of sample 2 are pre-
sented here.
Organic Analysis Results
The order of discussion for organic analyses
follows the analytical scheme presented in Table 2.
That order is:
pie
1. TCO + Grav analyses of the composite sam-
2. TCO + Grav analyses of the LC fractions
3. IR analyses of LC fractions
65
-------
4. LRMS analyses of LC fractions
TCO and Grav Analyses of the LC Fractions
TCP + Grav Analyses of Composite Sample
The first step in analyzing the XAO-2 extract
is the determination of the volatile (TCO) and non-
volatile (Grav) organic contents in the sample.
Table 11 compares the TCO and Grav values deter-
mined by the four participants for three aliquots
or replications.
Table 11. ORGANIC CONTENTS IN XAD-2 EXTRACT,
INITIAL VALUES
PARTICIPANT
A
B
C
D
A
B
C
D
A
B
C
D
TCO
106 mg
78
88
ISO
144
64
110
86
142
46
118
134
GRAV
386 mg
380
359
340
343
432
366
360
364
428
380
360
TOTAL
492 mg
468
447
490
487
496
476
446
496
474
498
494
Total organic (TCO + Grav) determinations show
good agreement. The average and coefficient of var-
iation (CV) for the 16 determinations are 480 mg and
4 percent, respectively. The range for the 16 val-
ues is only 52 mg or 11 percent of the average.
Determinations of nonvolatile organics showed
good Intel-laboratory agreement. The average and CV
for the 16 values is 374 mg and 8 percent, respec-
tively.
TCO determinations show within laboratory CV's
of 16, 26, 15, and 27 percent for participants A, B,
C, and D, respectively. The agreement between par-
ticipants A, C, and 0 are good. Participant B re-
ports an average value of 63 mg, almost half of what
the other participants reported.
Results of TCO and Grav analyses of the LC
fractions (XAD-2 extract sample) are given in Table
12. The data indicate an overlap or "smearing" of
fractions, resulting in differences in distribution
among the four contractors. For example, the totals
(TCO + Grav) for fraction 2 and for fraction 3 show
large differences among the four contractors, where-
as the sums of fractions 2 and 3 are fairly consis-
tent.
Also worthy of note are the unexpected differ-
ences in %TCO across the four laboratories, ranging
from 8 percent for laboratory B to 30 percent for
laboratory C.
Identification of Functional Groups by IR
Two sets of IR data are presented here. One
set is presented in Table 13 and represents a list-
ing of the functional groups identified by three
participants in the XAD-2 extract prior to LC sepa-
ration. The other set of data is a graphical pres-
entation of IR results by LC fraction and partici-
pant. The latter set of data is given in Figures 3
and 4.
From Table 13 it is seen that there is not a
one-to-one agreement between participants. However,
we feel that the comparisons are generally accepta-
ble.
Examination of the data presented in Figures 3
and 4 reveals the same types of fractional overlaps
as were noted in the TCO and Grav analyses. In
addition, laboratories C and D report a number of
bands that are not reported by laboratory A or B.
A possible explanation is sample contamination or
artifacts from the column, since laboratories C and
D report recoveries greater than 100 percent for the
TCO and Grav analyses of the LC fractions.
Categories Identification by LRMS
Categories identified in the XAD-2 extract by
LRMS are shown in Table 14 by LC fractions for two
participants. (Participants C and D did not run
LRMS's on several LC fractions.) The LC fraction,
the category name, and the relative intensity are
given in the table. Intensities in decreasing order
of intensity are recorded as 100, 10, or 1.
Table 12. TCO AND GRAV ANALYSES
RESULTS BY LC FRACTION (mg)
A
LC-1
LC-2
LC-3
LC-4
LC-5
LC-6
LC-7
LC-8
TOTAL
1.4
15.2
54.7
0.5
0.9
3.2
0.5
0.9
77
B
<0.4
<0.4
31
2.3
0.84
1.0
<0.4
<0.4
37
TCO
C
12.0
60.3
40.3
9.2
0.4
18.4
21.6
N.R.
162
D
2.1
120
2.6
1.4
7.4
13
1.9
as
157
A
11.5
4.6
259
10.1
31.3
16.6
2.8
3.7
340
GRAV
B C
4.0
6.4
284
22
as
38
23
37
423
4.8
84.6
177.6
12.8
7.6
27.1
23.1
45.1
383
D
__
200
19
—
—
110
—
50
379
A
13
20
314
11
32
20
3
5
418
TOTAL
B
4
• 6
315
24
10
39
23
37
458
C 0
17 2.1
145 320
218 22
22 1.4
8 7.4
46 123
45 1.9
46 58
546 536
66
-------
Table 13. FUNCTIONAL GROUPS IDENTIFIED BY
IR ANALYSES OF XAD-2 EXTRACT
A
3400
3060
2960.2930.
2860
1730
1600
1500
W
M
M
W
M
W
OH.NH
UNSATURATEDC-H
SATURATED C-H
KETONE. oCI
KETONE. ESTERS
RING VIBRATIONS,
N-ON
RING VIBRATIONS
B
3600-3300
3100-3000
3000-2800
2000-1860
1710
1696-1486
W OH AND NH/TRACE
S AROMATIC C-H
M ALIPHATIC C-H
W COMBINATION BANDS,
AROMATIC
M 00, ACIDS
M OC, PHENYL RINGS
D
3*20
3060
29180. 2920.
2860
1710
1800
W AMINE, POSSIBLY
CARBOXYLIC
ACID
S AROMATIC C-H
M ALIPHATIC C-H
W KETONES, ALDE-
HYDES. ESTERS
M ESTER,
AMINE.
AROMATIC
OVERTONES
1380
1200
CHjCHyO
W CH3
W(BROAD) C-0, ESTER. ETHER.
PHENOL
840.700 ^MULTIPLE)
740
SUBSTITUTED
AROMATIC AND
FUSED RINGS,
PYRIDINE
C-d
900-700
•Y(CH). AROMATIC RINGS
1466
1440
1240
1180
885.840
816.780.
746.735
M
M
W
W
M
ALIPHAT1CS
CARBOXYLIC
ACIDS.
ALIPHATICS
ESTER. KETONE
ESTER, PHENOL
AMINE. ARO-
MATIC SUBSTI-
TUTION
AROMATIC
SUBS
Several of the categories correspond between
participants as seen in the table. The most obvious
difference is that participant B reported benzene in
fractions 2 through 8 while participant A did not
find benzene in any of the fractions. (Similarity
of the IR spectrums of A and B for this sample in-
dicates that this is an interpretation problem.)
Inorganic Analysis Results
The Phase-II flyash sample aliquots were Parr
boabed in accordance with Level-1 procedures and
sent to independent laboratories for analysis by
spark source mass spectometry (SSMS). Results for
selected elements are given in Table 15. In all
but two of the cases shown (copper and chromium),
the high and low values for a given element differ
by a factor greater than 3. Nickel and beryllium,
which have the lowest MATE values of the elements
shown, were found to range from 13 to 380 ppm and
from 0.5 to 14 ppm, respectively.
An intralaboratory comparison can be made using
the data for contractors C and D since both SSMS
analyses were performed by the same outside labora-
tory. Differences, however, could be partially
attributed to sample preparation since the sample
aliquots were Parr bombed by the individual con-
tractors before being sent for SSMS analysis.
Included in the table are values for arsenic
and antimony, two of the elements for which alter-
nate procedures are specified in Level 1. A com-
parison of these SSMS values to the values obtained
by the recommended silver diethyldithiocarbonate
(SDDC) method for arsenic and the atomic absorption
method for antimony is given in Table 16. For Sb,
the values obtained by AA are slightly less than
those by SSMS in both cases. For arsenic, however,
the differences are much greater with the SDDC value
being higher than the SSMS value in one case and
lower in the other.
A brief summary of what we believe this pre-
liminary analysis of the SASS evaluation indicates
is as follows:
1. Participate concentrations determined by
the SASS1s compared very well with Method 5
and with each other;
2. Particle sizing compared very well between
SASS's;
3. Organic material collected by the SASS's
agreed well in quantity and composition (i.e.,
volatile, nonvolatile, and categories), and
was collected proportionally in corresponding
SASS components.
The results of the analytical methods evalua-
tion are interpreted as follows:
1. Certain methods employed in the organic
67
-------
IR RESULTS: XAD-Z EXTRACT, HELD SAMPLE
v 3500 3000 2500 2000 1800 1600 MOO 1200 WOO 800 600
I | I III)'111
Al + II 11 +
Bl NOT REPORTED
Cl
I I
Dl
I I
» 3500 3000 2500 2000 1800 1600 MOO 1200 1000 800 600
I | I I I I I I I I I
A2 + | || I I + I +
B2 NOT REPORTED
C2 II I I I + I +1 11+ I * +
3900 3000 2500
2OOO BOO
1 1
1
1
1 1
1600 MOO
II ill
1 .1 III ,.
, Jl
1200
1 1 1
II II II*M
T> "i0
H-H
I IMH-lllll-H
ll 1
"
03
35pO 3000 2900 2000 1800 WOO MOO 1200 KXX) 800
A4 ' I II ' II II I
C4 III I I I I II II I I
04 ' 'IM I I I ,1 II,
Figure 3. IR ANALYSES OF LC FRACTIONS 1-4
68
-------
IR RESULTS: XAD-E EXTRACT, HELD SAMPLE
v 3900 3OOO 2900 2000 1800 1600 1400 1200 IOOO 800 600
I I I I I I l I I I t
AS
B5
C5
05
V
A6
B6
C6
06
•
A7
B7
C7
07
*
-«- 1 1 1 1
_ t 1
II III,
1 II,
39pO 3000 2900 2OOO 1800 I6OO
+ , ,, 1
i i 1 , 1 , ,
i i | i | , ,
+ , 1 ,
+ 111 ll
390O 3000 3900 2OOO 1800 1600
+ , 1
4000 (t) - 1
NO IR BANDS
3900 3000 2900 2OOO 1800 WOO
1 •*• H- I 1 II 1
1 1 1 III
MOO 1200 IOOO 800 600
1 1 1 1 1
1 +
1 +
1 III
MOO 1200 IOOO 800 600
1 H-
1 1 1 +
MOO I2p0 IOOO 800 600
B8 NOT REPORTED
08 I I I I I —^ —H- •+-!!! I 4-
08 +
Figure 4. IR ANALYSES OF LC FRACTIONS 54
69
-------
Table 14. CATEGORIES IDENTIFIED BY LRMS
A
LC-1
• SULFUR (100)
. ALIPHATIC HYDROCARBONS (1)
B
. SULFUR (100)
LC-2 . FUSED ALT/NON-ALT HYDRO-
CARBONS, nt/e 216 (10)
• HETEROCYCLIC SULFUR
COMPOUNDS (10)
FUSED AROMATICS, MW<216 (100)
FUSED AROMATICS, MW>216 (10)
BENZENE. SUBSTITUTED BENZENE
HYDROCARBONS (100)
HETEROCYCLIC NITROGEN
COMPOUNDS (10)
LC-3 . FUSED ALT/NON-ALT HYDRO-
CARBONS, m/e 216 (100)
. HETEROCYCLIC SULFUR
COMPOUNDS (10)
• FUSED AROMATICS, MW<216 (100)
. FUSED AROMATICS. MW>216 (100)
. BENZENE. SUBSTITUTED BENZENE
HYDROCARBONS (100)
LC-4 . HETEROCYCLIC NITROGEN
COMPOUNDS (100)
• FUSED ALT/NON-ALT HYDRO-
CARBONS. m/a 216 (10)
FUSED AROMATICS, MW<216 (100)
FUSED AROMATICS, MW>216 (100)
BENZENE, SUBSTITUTED BENZENE
HYDROCARBONS (100)
HETEROCYCLIC NITROGEN
COMPOUNDS (10)
LC-5 • HETEROCYCLIC OXYGEN
COMPOUNDS (100)
• HETEROCYCLIC NITROGEN
COMPOUNDS (100)
. FUSED ALT/NON-ALT HYDRO-
CARBONS. 216 (100)
BENZENE, SUBSTITUTED BENZENE
HYDROCARBONS (100)
NITRILES (100)
HETEROCYCLIC NITROGEN
COMPOUNDS (100)
LC-6 • HETEROCYCLIC OXYGEN
COMPOUNDS (100)
• HETEROCYCLIC NITROGEN
COMPOUNDS (100)
• ESTERS (10)
• CARBOXYLIC ACIDS (10)
• PHENOLS (10)
FUSED AROMATICS, MW<216 (100)
BENZENE, SUBSTITUTED BENZENE
HYDROCARBONS (100)
NITRILES (100)
HETEROCYCLIC NITROGEN
COMPOUNDS (100)
LC-7 . HETEROCYCLIC NITROGEN
COMPOUNDS (100)
. ESTERS (10)
• CARBOXYLIC ACIDS (10)
• HETEROCYCLIC OXYGEN
COMPOUNDS (1)
FUSED AROMATICS, MW<216 (100)
BENZENE, SUBSTITUTED BENZENE
HYDROCARBONS (100)
NITRILES (100)
HETEROCYCLIC NITROGEN
COMPOUNDS (100)
CARBOXYLIC ACIDS AND
DERIVATIVES (100)
LC-8
PHENOLS (10)
CARBOXYLIC ACIDS (1)
HETEROCYCLIC NITROGEN
COMPOUNDS (1)
ETHERS (1)
ESTERS (1)
FUSED AROMATICS. MW<216 (100)
FUSED AROMATICS, MW>216 (100)
BENZENE. SUBSTITUTED BENZENE
HYDROCARBONS (100)
IRON (10)
analysis scheme are still being refined, and
interpretation of the organic data from com-
plex sources can be an involved process re-
quiring great attention to detail. However,
fro* this preliminary analysis of Phase-II
data, it appears that the organic analysis
scheme can yield results of adequate quality
to satisfy Level-1 requirements, if judi-
cious care is exercised by the analyst(s)
to utilize all the analytical data generated
by the scheme in interpreting individual
blocks of data.
2. Results of the inorganic sample prepara-
tion and SSMS analysis scheme indicate that
variability in the analytical phase alone
may be exceeding the allowable factor of 2
or 3 in the Level-1 procedures.
70
-------
Table 15. COMPARISON OF ANALYSIS OF SELECTED
ELEMENTS BY SSMS (ppm, by weight)
ELEMENT
ARSENIC
ZINC
COPPER
NICKEL
COBALT
CHROMIUM
VANADIUM
CHLORINE
BERYLLIUM
URANIUM
THORIUM
LEAD
DYSPROSIUM
CERIUM
LANTHANUM
ANTIMONY
A
B
41
21
42
68
2.7
6B
68
210
0.6
Z6
6.1
33
8.4
23
18
6.9
C
38
67
41
380
23
160
28
66
OS
3
8
8
2
12
18
1
D
140
20
120
13
10
100
130
290
14
80
76
24
26
170
180
OJ
Table 1& COMPARISON OF As AND Sb BY SSMS
AND BY SDDC AND AA, RESPECTIVELY*
C
D
At (ppm)
SDDC
776
36
SSMS
38
140
Sb(ppm)
AA
<1
-------
INORGANIC EMISSIONS MEASUREMENTS
By
Ray F. Maddalone
Lorraine E. Ryan
Applied Technology Division
TRW Defense & Space Systems Group
Abstract. The analysis of inorganic compounds
requires the coordinated use of a variety of
analytical techniques. This paper describes'an
inorganic analysis scheme consisting of an initial
sample characterization (stability, elemental
composition, and morphology), bulk composition
characterization (anion composition, surface
characterization, and x-ray diffraction data) and
individual particle characterization (single par-
ticle elemental composition, x-ray diffraction
pattern, and morphology). The use of Multimedia
Environmental Goal (MEG) compounds and their Mini-
mum Acute Toxicity Effluent (MATE) values to focus
analysis activities will be described. Data from
a recent field test using this approach will be
used to illustrate the information derived from
these methods.
Introduction. With the increasing awareness of
the government and scientific community to the
possible hazards from the output of various indus-
tries, the Environmental Protection Agency has
developed an approach to assess the environmental
impact of any type of industrial process. This
approach consists of a two phase attack.
The first phase surveys the site to determine
whether or not a given pollutant is being emitted.
This so called Level 1 approach uses sampling and
analysis methods to obtain results accurate to a
factor of two to three. A set of criteria are
used to prioratize the streams so that those
streams which are found to be a problem are identi-
fied for further study. This next phase, Level 2,
is designed to be specific for a given stream and
perhaps for even a given pollutant. Compared to
this phased approach, a direct environmental
assessment of a site would use comprehensive sam-
pling and analytical methods to determine all pol-
lutants that are present with high accuracy.
Figure 1 summarizes the difference between taking
a phased approach or the direct approach environ-
mental assessment. TRW has studied the two ap-
proaches and found significant cost savings by
using the phased approach. ?
Level 1 Sampling and Analysis. The types of sam-
ples that will be obtained during a Level 1
sampling program can be broken down into four
areas:
• Gaseous Samples - Process or fugitive
emissions
• Liquid or Slurry Samples - Liquids or
liquids with suspended solids
• Solid Samples — Solid holding or transfer
systems
• Particulate Samples - Process or fugitive
emissions
FIGURE 1
ENVIRONMENTAL ASSESSMENT
MKCT VUSUS PHASED APPtOACH
COMFKHENSIVE
ONCETWOUGH
CHAMCTEHZATION
COMPKHENSIVE
SAMPLING FO» ALL
COMPONENTS IN
All STREAMS
COMMEHENSIVE
OtGANIC, INOtGAMC
AND •OLOraCAL
ANALYSIS FOI
ALL SAMPLES
STKAMS IN WHICH
PtOKEMSAK
FOUND TO EXIST
AK IDENTIFIED FOI
FUtTHEt STUDY
LEVEL 2
SAMPLING
MATBX
QUANTITATIVE
ANALYSIS
Gaseous samples are taken using a grab bag and
analyzed on site for SO-2, H£S, COS, CO, C02, 02,
NH3, HCN and (CN)2- A chemiluminescence detector is
used for NOX on-line measurement. Liquid or slurry
streams are sampled using standard ASTM procedures.
For particulate samples taken from flue gas streams,
a specific train has been designed to collect the
organic and Inorganic components of the stream. The
Source Assessment Sampling System (SASS), shown in
Figure 2, 1s a comprehensive Inorganic/organic
sampling train consisting of a particulate and gaseous
sampling sections. The particulate sampling section
consists of a probe, 10, 3 and 1 u micron cut-off
cyclones. These cyclones are followed by a glass
fiber filter which traps most particles down to a
0.3 u with 99.9% efficiency. The entire particulate
section 1s heated to 200°C to minimize H?SOA
condensation.
The gases from the particulate collection section
are passed through a condensation module containing
a cooling section and XAD-2 resin, which removes
volatile organics above CG. Inorganic materials
which are removed from the gas phase by the con-
densing moisture are collected 1n the condensate
trap. After the gases pass through the condensation
module they encounter an oxidative 1mp1nger train
This oxidative 1mp1nger train 1s designed to trap
all volatile Inorganic materials such as Hg, Cd,
or As whichconcelveablycould pass through the cooling
module. The oxidative implnger systems consists of
72
-------
FIGURE 2
SASS TRAIN SCHEMATIC
200 >C MAX.
1 1
I 10(1 3(1 )„ 1 COOUK
j CYaONE CYCLONE CYCLONE FILTER | C»«q
H h- -1
n
1
I
|
1
1
1
J
M
G
XAD-2
SORIENT
J
,u2
IMPINGE *S
peroxide impinger followed by two amnoniura persul-
fate, silver catalyzed Impingers. This Impinger
system has been tested thoroughly and Is essentially
100 percent efficient for volatile Inorganic
materials.
Once the samples have been obtained from the field,
they are either analyzed on site or returned to the
laboratory for further analysis. Figure 3 shows
the overview of the Inorganic and organic analyses
performed on gases, liquids, or solids. Primarily,
gaseous analysis of Inorganic materials relies
on an on-slte, GC for the combustion gas analysis.
Spark Source Mass Spectrometery (SSMS) 1s the key
Inorganic analytical technique. All the liquids
and solids are analyzed by SSMA for their elemental
content. The SSMS results are supplemented by wet
chemical techniques for Hg. As, and Sb, for improved
accuracy on these elements. Liquids obtained on
site are analyzed for selected anion's (N0§, Cl, F",
CN~ and S0|) using field test kits.
Transition to Level 2. Once all the Level 1 samples
have been analyzed, some method of correlating and
evaluating these data is necessary. Figure 4 shows
the decision procedure for proceeding from Level 1
to Level 2. In this procedure, analysis of the
FIGURE 3
LEVEL 1 MULTIMEDIA ANALYSIS OVERVIEW
1
1 LEVEL 1 SAMPLE 1
1
GASES 1 1 LIQUIDS J | SOLIDS |
1
INORGANIC
• GC-SO-, H,S, COS. CO,
COj/Oj, NHj, HCN,
(CM).
•NOy- CHEMI-
* LUMINESCENCE
• IMPINGERS
- SSMS
. WET CHEMICAL
ORGANIC
• GCFORC,-C6
• XAD-2 EXTRACT
-GC FOR C,-^
- m
- LC/KARMS
1
1 1
INORGANIC
• ELEMENTS
- SSMS
- WET CHEMICAL
• LEACHAILE MATERIAL
REGULATED BY EPA -
REAGENT TEST KITS
ORGANIC EXTRACTS
• GCFORCj-C,,
• K
• LC/IR/UMS
INORGANIC
ELEMENTS
- SSMS
- WET CHEMICAL
SELECTED ANIONS
AQUEOUS - SELECTED
WATER TESTS
ORGANIC
• EXTRACT AQUEOUS
SAMPLES WITH CHjCt,
. GCFORC7-C,6
• R
• LC/K/LRMS
73
-------
FIGURE 4
DECISION PROCEDURE FOR LEVEL 1 — LEVEL 2
'
•
ANALYSIS
c
LEVEL 1 A
SAMPLES J
1
*
EVALUATION |
I
/LEVEL 1 7
/CONCENTRATIONS/
LIST OF /
MEG CATEGORIES/
LIST VALUES BY
REPORTING POINT;
MATE
CONCENTRATIONS
BY SOURCE:
AIR. WATER,
SOLID WASTES
ASSIGN MEG
CATEGORIES TO
LEVEL 1 REPORT-
ING POINT
EACH REPORTING
POINT CAN SPAN
SEVERAL MEG
CATEGORIES
IS
RATIO
OF SAMPLE
CONCENTRATION
TO APPROPRIATE
MATE VALUE LESS
OR GREATER
THAN
0.57.
.LESS
LIST OF MEG
'CATEGORIES NOT
REQUIRING LEVEL 2
GREATER
LIST OF MEG CATEGORIES L
REQUIRING LEVEL 2 /
FOCUS POINT FOR
LEVEL 2 ANALYSIS
samples is completed and the Level 1 Inorganic
concentrations are determined. These values are
then listed by their elemental concentration In
ug/m3 (gaseous), ug/1 (liquids) or ug/g (solids).
At the same time a list of Multimedia Environ-
mental Goals (MEG) compounds are compiled with
their Minimum Acute Toxidty Effluent (MATE) con-
centrations listed by source,(air, water or solid).3
MEG compounds resulted from a study of fossil fuel
conversion processes and represents a listing of
compounds associated with coal and oil that could,
based on free energies and conversion conditions,
be released to the environment.
MATE'S describe very approximate concentrations for
contaminants 1n source emissions to air, water, or
land which will not evoke significant harmful or
Irreversible responses 1n exposed humans or ecology,
when those exposures are limited to short duration
(less than 8 hours per day).
The sample concentration 1s the divided by the
appropriate MATE value. If this ratio 1s greater
than 0.5, then that element In the stream deserves
further Level 2 attention. The value of 0.5 was
selected because the Level 1 uncertainty 1s a fac-
tor of 2 to 3.
Level 2 analysis 1s focused research because we
are able to limit the number of elements and poten-
tially the number of streams which need to be
re-examined In Level 2 tests. Level 2 will seek to
quantify the element more exactly (±25X) and
secondly determine the compound 1n which 1t 1s
found. Level 2 will require more sophisticated
analysis methods and an analyst to assess the data
and direct the research.
74
-------
Level 2 Sampling. In some cases it might be possi-
ble to analyze Level 1 samples further using
Level 2 type techniques. In many cases because
the SASS train is an all stainless steel train,
these samples will be contaminated with Ni, Cr and
Fe. Consequently a Level 2 inorganic particulate
train can be used to sample streams which have been
targeted for further research. This train, shown
in Figure 5, is an all glass construction train.
This train has been used in field tests and
because of its all glass design Ni, Cr and Fe can
be monitored. The train itself consists of a
particulate section which has a 3y cutoff cyclone,
and a glass fiber filter. The particulate sec-
tion is then backed up by a series of oxidative
impingers which use the same chemistry as the SASS
train. The main drawback with these glass trains is
that they currently are designed to sample at 1 cfm.
This low sampling rate compared to the SASS train
is partially offset by the high sensitivity of
most inorganic sampling techniques for elemental
analysis. '
For other streams the improvement in sampling is not
so much In the sampling system as it is in the
design of the sampling program. Level 2 seeks to
improve the representativeness of the sample taken
by taking composited samples or timing the sampling
to a specific phase of the process. The option is
always available to design out of the ordinary
equipment and procedures to obtain specific samples
or to meet difficult sampling conditions. Level 2
should be free to meet the needs of specific situa-
tions and, like the Level 2 analysis, will require
expert personnel to design the equipment and
procedures.
Level 2 Analysis. The analysis of inorganic
compounds requires the coordinated use of a variety
of analytical techniques. Some techniques, such as
XRD, TEM-SAED and ESCA, have the potential for
direct compound identification, but only for
selected compounds. The methods are of increas-
ing analytical complexity, designed to be cost and
time effective. The identification scheme con-
sists of:
0 Initial Sample Characterization - elemental
composition, sample stability, and bulk mor-
phological structure are determined.
• Bulk Composition Characterizations - qualita-
tive and quantitative anion, oxidation state,
and X-ray diffraction information are derived.
• Individual Particle Characterization - single
particle elemental composition. X-ray diffrac-
tion pattern and morphology are measured.
FIGURE
LEVEL 2 INORGANIC SAMPLING TRAIN (GLASS)
FILTER
HOLDER
HEATED
CONTAINER
(2WQ
CYCLONE
THERMOMETER
CHECK
VALVE
ORIFICE
THERMOMETERS BY-PASS
7VALVE VACUUM
(**\ GAUGE
DRIER1TE
0.2M(NH4)2S208
•K).02M AgNO3
I-
MAIN
VALVE
VACUUM
LINE
DRY TEST METER
AIR-TIGHT
PUMP
75
-------
The degree to which each method can be applied
varies considerably with the experience, sample
quantity and equipment available to the analyst.
It is recommended that continuing use of any one
method be evaluated in light of the information
derived. In general, it is far better to use a
variety of instruments operated in the most effi-
cient manner rather than pushing a single instru-
ment or technique to the limit of its capabilities.
In the Level 2 approach, emphasis is placed on
reaching an accurate closure to the MEG compounds
which exceed HATE values after a method or series
of methods has been applied, a comparison of lists
of Identified to potential MEG compounds for ele-
ments which exceed their MATE values is made. A
satisfactory analysis will depend upon a variety
of factors:
• Number of MEG compounds Identified
exceeding MATE values
• Interest in identifying the remaining com-
pounds for those elements that have exceeded
MATE values
• Cost/availability of necessary equipment
The analyst must decide as to what method will be
applied and how much more information can be
obtained by each further analysis. In many cases
some methods can be bypassed because of results
from previous tests, e.g., quantitative anion ana-
lysis may provide sufficient Information and FTIR
would be only repetitious. In other cases efforts
may direct the analyst to a specific method since
it would be best suited to analyze for a given com-
pound. The following sections provide a discussion
of the proposed methodology and information derived.
By understanding the outputs from each technique,
the analyst will be better able to select the
appropriate combination of techniques to determine
the compounds present in the environmental sample
of Interest.
Initial Sample Characterization. Initially 1nfor-
mation from all sources (Level 1 field and analy-
tical data) concerning the composition of the
sample is pooled, assessed, and used for reference.
This Information provides the first insight Into
the composition of the sample. The elemental con-
centrations are compared to MATE values to deter-
mine which compounds must be sought. Once the MEG
compounds exceeding MATE values are Identified,
then the elemental composition data will be used
by the analyst to determine whether or not specific
compounds are present as well as Indicating whether
a given method has sufficient sensitivity to detect
the potential compound. For example, if the SSMS
data shows that a given element Is present at
levels above 0.5 percent then bulk analysis by XRD
might be successful 1n determining the compound
form of this element. Elemental Information 1s
especially Important to XRD because diffraction
patterns of environmental samples are complex and
Information reducing the potential possibilities
is necessary for a cost effective analysis.
Besides the SSMS data, there are several other
sources of information that should be assembled.
This information Includes:
• Source of the sample: process type, e.g.,
oxidizing or reducing conditions.
• Composition of feed source: Input raw
materials, e.g., coal, limestone.
• Previous history of sample: age, storage
conditions, collection method.
• Results of previous analyses at this source:
elemental and compound Information.
As soon as the sample 1s in the laboratory, it
should be viewed under a polarized light microscope
(PLM) and a photomicrograph taken in color to act
as quality control. If any changes In the general
appearance of the sample occur, during the duration
of the analytical activities, these should be
noted. The PLM can also provide a measure of the
complexity of the sample simply by noting the num-
ber of different types of particles.
Polarized light microscopy 1s the first direct
compound analysis method. Particles can be iden-
tified by the determination of such properties as
the refractive Index, isotropy or anisotropy, bire-
fringence, pleochroism, fracture, color, and crys-
tal habit. Hicrospot tests for common anions and
tests of the solubility of the particles in water,
acid, and base can be performed directly on the
sample as It 1s being examined under the microscope.
These microtests will alert the analyst to perform
quantitative analyses for the anions detected and
they will also provide Information about the poten-
tial success of full scale dissolutions and
separation.
At this stage in the analysis, quantitative analysis
of anions Identified In the mlcrospot tests will
most probably be performed using classical wet test
methods, e.g., tltrimetric, coloHmetrlc, or spe-
cific 1on electrode tests. Level 2 anion methods
can be chosen by the analyst from Standard Methods
(Water and Wastewater), ASTM, or EPA procedures.
Also, during this initial sample characterization,
the analyst may choose to supplement the SSMS seml-
quantitatlve cation data by analyzing fractions of
the samples using such quantitative techniques as
atomic absorption spectrometry (either flame or
if Tameless), Induction Coupled Plasma Optical -
Emission Spectroscopy (ICPOES), Proton Induced X-Ray
Emission (PIXE), or X-ray fluorometry.
In conjunction with the PLM work a TGA/DSC scan of
the sample should be made. This test Is used pri-
marily to determine 1) the stability of the sample,
and 2) an appropriate temperature at which to dry
samples to be used in later tests. In a few cases
it 1s possible to determine the compounds present
by the weight loss at specific temperatures. Ele-
mental Information from SSMS and anion Information
from PLM (and later IR) can be combined to give a
11st of potential compounds that exhibit decomposi-
tion points at the weight loss points in the TGA or
the exotherms and endotherms of the DSC.
76
-------
At the end of the Initial effort, Information will
have been obtained In the following areas:
1. General appearance of a sample
2. Number of different particles present
3. Index of refraction and crystal structure
4. Individual particle anion composition
5. Individual particle solubilities
6. Weight loss with respect to temperature
7. Bulk elemental distribution.
This exercise 1n logic 1s summarized in Figure 6.
Bulk Composition Characterization. Supposing that
the need of the analyst to identify a specific
pollutant has not been met, then the next phase of
analysis, bulk characterization is started. The
methods used in this approach are:
• X-Ray Diffraction (XRO)
t Fourier Transform IR (FTIR)
• Electron Spectroscopy for Chemical Analysis
(ESCA).
To guide the analyst, a logic network for their
application is shown in Figure 7.
It is expected that the samples will have to be
dried to a constant water content to Improve both
IR and XRD spectra. Information from the TGA/DSC
step will be used to select a drying temperature
that removes water without decomposing the sample.
Further sample preparation will vary with the
requirements of the specific analysis method.
For IR analysis, the KBr pellet technique for qual-
itative analysis 1s not recommended due to 1on
exchange possible during the pelleting process. It
1s recommended that a Nujol null of the sample and
AgCl (1333-400 cur") and polyethylene (600-45 cnr')
windows be used. Interpretation of the Infrared
spectra on the basis of characteristic frequencies
can provide the Identity of specifications and some
of the Individual compounds.
Extensive compilations of Inorganic compounds are
available (Ref 14). There are definite analytical
frequencies which can be used to identify com-
pounds, particularly when supporting elemental
analysis Information 1s available.
Electron spectroscopy for chemical analysis (ESCA)
can be performed on both loose partlculates and
particles collected on filters. Loose partlculate
samples can be attached to a sample holder using
an approach called the "sticky gold" technique.
This technique was devised to overcome the conduc-
tivity problem and securely mount the sample. It
sandwiches the sample between a layer of sputtered
gold and carbon. The gold first layer applied to
a double sided scotch tape does not change the
tackiness of tape, which allows loose particles to
be stuck to the surface. A layer of carbon 1s
deposited on the surface assuring that all the
particles are near a conductive surface. Filter
pieces can be clamped directly onto the sample
holder after the bottom layers of the filter have
been peeled off.
In a recent sampling Level 2 sampling program taken
at a coal-fired boiler with a flue gas desulfur-
ization unit, and ESCA analysis was performed on a
filter sample taken from the outlet of the FGD.
During the analysis of this sample it was etched
with Ar Ions and the sulfur content, after each
layer of the particle was removed, was analyzed
using ESCA. Figure 8 shows the depth profile of the
sulfur content. The intensity is a ratio of the
sulfur 2s electron to the aluminum 2s electron. As
one can see in the figure there is a definite depth
profile with increasing sulfur content on the sur-
face of the particle. In another case with some
samples from a Fluidized Bed Combustor (FBC), a
layer of sulfate was found to cover a sulfide parti-
cle. This information illustrates the capability
of ESCA to determine the oxidation state of elements.
By knowing the oxidation state of the element, the
species that the element exists as can be determined.
The performance of specific anion tests, IR analysis,
and ESCA establishes substantial information on the
variety and depth profile of anions in the sample.
This information simplifies interpretation of the
XRD spectra, and provides an independent quantifi-
cation of the species present. In X-Ray diffraction
analysis, approximately 100 mg of material are
ground in an agate mortar, ultrasonically dispersed
with a 1:4 mixture of collodion with alcohol and
then evenly spread over a glass support. Mounting
in this fashion will produce the highest sensitivity
at low 2e values. The major disadvantage of XRD as
an analytical tool is its inability to detect non-
crystalline materials. In many environmental sam-
ples, the crystal structure of a compound could be
grossly affected by the conditions at the source or
those during sampling. For example, As2C>3 can be
amorphous or crystalline depending on its temper-
ature history. Furthermore, the sensitivity of XRD
1s normally limited to 1 percent or higher, although
new computer averaging techniques enable materials
to be detected in concentrations as low as 0.05
percent. (5)
Having completed these analyses, information will
have been obtained on the following:
1. Anions present
2. Valence state of elements present
3. Elemental depth profile
4. Major compounds present.
At this point the analyst must correlate all data
and determine if a reasonable (based on the ana-
lyst's judgement) agreement has been reached with
MEG elements exceeding their MATE values. If there
1s reasonable agreement between the elemental data
obtained from quantitative techniques and the com-
pounds determined in this characterization, further
work should be carefully evaluated in terms of
potential needs and end use.
77
-------
FIGURE 6
LOGIC FLOW CHART FOR INITIAL SAMPLE CHARACTERIZATION
(SOUD A
SAMPLE J
*
/ELEMENTAL LEVEL 1 /
DATA /
< MATE X«UT \^
. . . VALUES/ ELEMSNTSX
/ELEMENTS t
NOT EXCEEDING/
MATE VALUES /
AVAILABLE
MEG _
COMPOUNDS ^
—_/ EXCEED y
^^^ MATE /
\VALUES /
J^MATE VALUES
/UST ELEMENTS /
EXCEEDING MATE /
VALUES /
A
"?
LIST POTENTIAL
•• COMPOUNDS
PRESENT
/TOBE\
/UST UNSTABLE >
/COMPOUNDS AND/
/ CONDITIONS /
t
1 - J
I^REftrCTTvE-01
*
f
MiatOSOtlMUTY MICRI
TESTS ON
sawr «*°
*
/SOLUBIUTY Of /
SPECIFIC GROUPS /
Of PARTICLES /
cV
T
/ LISTS Of 7
/ ANION VS /
f SOLUBIUTY /
|
' _/tUWFOUNDSX
/ STABLE X
•— VUNDEI PROCESSOR/
MO X^SAMPUNG /
/UP-DATE POTENTIAL/
COMPOUND UST /
t
STUDY GENERAL
CHARACTERISTICS
Of PARTICLES
*
n~TGA/DSC "I
t
WEIGHT GAIN/ LOSS.
REACTION TEMPERAllUS.
PHASE CHANGES
1
0-SPOTTEST
SPEQRC
MS/CATIONS i
/ STABLE DRYING /
/ TEMPERATURES. VAPOR-/
/IZATION TEMPERATURES, /
/DECOMPOSITION POINTS/
/AND AIR STA8IUTY /
I
UP-DATE UST OF
POTENTIAL
COMPOUNDS
i
SELECTS SPECIFIC ANIOH
CATION TESTS FOR
TOTENTIAL ELEMENTS
\
WET
AAS OR
AM
\
"1
CHEMICAL
INSTRUMENTAL
ON TESTS
*
/* 15% ELEMENT / / *I5% ANION /
COMPOSITION / / COMPOSITION /
1
RATIO CATION/
ANION VALUES
*
/ UPDATE POTENTIAL /
/ COMPOUND UST /
/ WITH WEIGHT /
/ INFORMATION /
A
./MASS x. r
/CLOSURE Of >v /
\^tCBEDINGMAIE/^^7
^f
/ UST POSSIBLE /
/ ASSIGNED COM - /
/POUNDS AT SUSPECTED/
/ LEVELS /
1
UST IDENTIFIED /
COMPOUNDS /
WITH ESTIMATED /
CONCENTRATIONS/
78
-------
FIGURE 7
LOGIC FLOWCHART FOR BULK COMPOSITION CHARACTERIZATION
1
STUDY SULK
CHEMICAL
COMPOSITION
ASSIGN PRO1AMUTY TO I
SEE POTENTIAL COM- '
POUNDS WITH SPECIFIC
METHOD
ALLOWS
MATCH UP OF
METHOD WITH COMPOUNDS
BASED ON CONCENTRATION
[L
FTH
ESCA
PERFORM FAI It SCAN KM
TRANSITION ELEMENT
ANIONS IN IULK Of SAMPLE
suntAcr INSOLMUS
SPECTRA ROM
ORIGINAL SAMPLE
STUDY SURFACE TRACE
ELEMENT COMPOSITION
OXIDATION STATES.
CHEMICAL
ENVIRONMENT
X«D
1
DIHCT ID
OF CRYSTALLINE
COMPOUNDS
AT 0.1% Ot GREATER
CONCENTRATION
/TtANU™™ / / "**NCE OT / / UST OF SPECIFIC
ELEMENT ANIONS / / *KOMED SPECIES / / COMPOUNDS
1
*
QUANTJTATE
SPEQRC
ANIONS
i
WET CHEMICAL Ot
INSTRUMENTAL
AMON TESTS
\
/UST POSSIILE NEW
COMPOUNDS /
FOUND /
1
J
1
UST IDENTIFIED
COMPOUNDS WITH
ESTIMATED
CONCENTRATION
HAVE
ALL MEG
COMPOUNDS
EXCEEDING MA
VALUES KEN
FOUND
UST ASSIGNED
ELEMENTS EXCEEDING
MATE VALUES
79
-------
FIGURE 8
SULFUR DEPENDENCE WITH ETCHING DEPTH
FOR OUTLET COAL FIRED FILTER SAMPLE
100 200 300 400 500
ETCHING DEPTH (A)
600
700
800
Individual Particle Characterization. It should be
emphasized that this phase of the analysis should be
carried out at the analyst's discretion. The ana-
lyst should consider the sample, its source, the
information already available, the type of informa-
tion which is lacking, the instrumental techniques
available, and analysis cost before proceeding.
Analytical techniques which are suggested for iden-
tification of individual particles include:
• Scanning Electron Microscopy with Energy
Dispersive X-Ray Spectrometry (SEM-EDX)
• Electron Probe Microanalysis (EPMA)
• Transmission Electron Microscopy with
Selected Area Electron Diffraction (TEM-SAED)
The logic network for their application is sum-
marized in Figure 9.
In SEM, the sample specimen is swept by an electron
beam and the variation of the secondary electron
emission intensity is recorded. This signal
modulates the brightness of an oscilloscope
beam, producing an image of the sample surface
on the oscilloscope screen. Since the secondary
electron beam is localized in the area impacted by
the incident radiation, images of relatively high
resolution are achieved which can provide morpholo-
gical characteristics of individual particles. When
SEM is used in conjunction with an energy dispersive
X-ray spectrometer (EDX), the secondary X-rays pro-
duced can be monitored, thereby allowing identifica-
tion and quantification of individual elements
present in the sample. Determining the elemental
distribution of a particle is particularly useful
for those particles composed of various occluded
materials, the high resolution and magnification of
the SEM can produce images distinctive enough to
identify the particle. As such, the SEM information
is a valuable adjunct to the PLM, especially for
particles smaller than 0.5u.
Figure 10 shows a SEM photograph of a typical fly
ash sample and as can be seen in the photograph
there are areas for particles which appear to be
growing on the surface of a particle that have a
crystalline shape associated with them. Because
we know that many of the trace elements found in
fly ash can be found to a greater extent on the
surface, analysis of these small raicrocrystalline
structures on the surface of a large particle is
very Important.
80
-------
FIGURE 9
PERFOIM SINGLE
PARTICLE ANALYSIS
SEM-EDX
OBTAIN DETAILED MORPHOLOGICAL
INFORMATION, SINGLE
PARTICLE ELEMENTAL
SCAN AND ELEMENTAL RATIOS
TEM-SAED
f ESTABLISHED ELEMENT)
'RATIOS FOR SINGLE
PARTICLE
LIST COMPOUNDS
WITH CONCENTRATION
ESTIMATE RASED ON
ELEMENTAL VALUES
VE
GNED
G
CEEDING MAT
ALUESBEE
FOUND
7
IS
FURTHER
ANALYSIS COS
JUSTIFIED FOR
UNASSIGNED
ELEMENTS
ELEMENTAL
RATIOS ESTABLISHED
FOR SINGLE PARTICLES
FOR ELEMENTS EC
LIST COMPOUNDS
WITH CONCENTRATION
ESTIMATE BASED ON
ELEMENTAL VALUES
CAN
UNAS
SIGNED MATE
DENTIRED Wl
THESE
RATIOS
7,
NO
DETERMINE XRD SPECTRA
OF SINGLE PARTICLE OR
AREA IN PARTICLE
ELEMENTAL
DATA USED TO
QUANTIFY
VE ALL
UNASSIGNED
G COMPOUN
CEEDMGMA
ALUESBEE
FOUND
IS
FURTHER
CTEBZATI
JUS TIRED
7
LIST UNASSIGNED
FRACTION OF
KNOWN ELEMENTAL
COMPOSITION
BASED ON SOLUBILITY
AND ELEMENTAL DATA
SELECT SEPARATION
SCHEME
SELECTIVE
DISSOLUTION
SEPARATION
DENSITY
GRADIENT
SEPARATION
MAGNETIC
SEPARATION
LESS COMPLEX
MATRIX
SSMS
OF FRACTIONS
-------
FIGURE 10
SEM PHOTOGRAPH OF COAL FLYASH SHOWING
CRYSTALLINE MATERIAL ON SURFACE
OF PARTICLES
An alternative to scanning electron microscopy is
Scanning Auger Microanalysis which might prove in the
future to be a better method. Because of the power
of the electron beam used in SEM, surface analysis can
be difficult. In many cases, surface structures on a
particle cannot be analyzed for their elemental com-
position because the beam penetrates to the surface
of the particle, and consequently the elemental com-
position of the surface structures cannot be deter-
mined. However, with SAM the surface capability of
Auger can be used to specifically determine the ele-
mental composition of microstructures on the surface
of the particle. At the present time because the
lowest resolution is on the order of 0.2 microns, many
of the more interesting microcrystalline structures on
a particle cannot be seen to be analyzed thoroughly,
but this limitation is expected to be removed in the
near future.
In order to reduce the mounting time for both SEM and
EPMA (electron probe microanalysis), particles can be
mounted on a sticky gold to provide a conductive sur-
face. Normally, a carbon film would be deposited on
the sample to ensure its conductivity. If the sample
is reasonably conductive and long analysis times are
not necessary, then the carbon film may be omitted.
Mounting samples in this fashion will not interfere
with later EPMA analysis.
In EPMA, a small energetic electron beam impinges the
surface of the particulate specimen and produces
characteristic X-ray emissions. EPMA can be used to
qualitatively and quantitatively determine the ele-
mental composition of particles ranging in size from
20y down to about 0.2w, for most of the elements of
atomic numbers above that of carbon. Instruments
using wavelength dispersive X-ray spectrometers can
resolve spectra of elements sulfur through nickel in
atomic number. Peak heights, or intensity ratios,
are measured on samples and standards to provide a
quantitative analysis. To achieve the best accuracy,
it is necessary to do a considerable amount of sam-
ple preparation. In most cases it is necessary to
have standards similar in particle size and compo-
sition to the sample being analyzed. Further, iden-
tification is possible only for particles containing
discrete compounds rather than a homogeneous mixture.
Transmission Electron Microscopy with Selective Area
Diffraction (TEM-SAED) also involves the impingement
of an electron beam on a thin film (1500 A) of sample.
The resulting single particle X-ray diffraction
pattern permits identification of crystallin com-
pounds. The qualitative and quantitative data
obtained is excellent because individual particles
and fibers can be observed and identified. This
attribute of TEM-SAED has been used to provide
dependable identifications of such chemical species
as asbestos and silica.
Combining the information derived from TEM-SAED and
EPMA can aid the analyst in assembling the total
nature of the various species present. Many sub-
stances which appear essentially identical in ele-
mental composition as measured with the electron
probe, will be determined by TEM-SAED to have a
unique morphology and, therefore, their emitted
nature and source clearly indicated.
At this point, if all the compounds for MEG elements
exceeding their MATE values have not been found,
then the analyst might choose to reduce the sample
matrix into simple mixtures. He can either run
magnetic density gradient, or selective dissolution
studies. In magnetic separation, magnets are used
to remove the magnetic fraction from the sample.
In density gradient separation, particles are
floated in organic solvents of known density. Con-
siderable care must be used in selecting solvents,
because compounds could be soluble in the solvents.
This procedure can be used to obtain gross separa-
tions by density or can be used to determine indivi-
dual particle densities for identification purpose.
Selective dissolution uses a variety of solvents to
remove more and more of the sample and in the pro-
cess simplifying the composition of the residue.
FIGURE 11
GAS CHROMATOGRAM OF HYDRIDE
PRODUCED BY REDUCTION PROCEDURE
GAS CMOMATOGftAM
: 0.1 ML Of GAS
r ASPAS
CAUL »"• CTI
SONS I 10900
H,0
3.
190 XB
J«J 440
190
• JO
i . N-
82
-------
In all these separation techniques care must be taken
to avoid contamination and scrambling of compounds.
Also, reasonably large quantities of sample are nec-
essary. The end result of these separations 1s to
provide less complex fractions which can be studied
starting at the bulk characterization level.
Liquid Samples. These previous techniques which have
been talked about are primarily used for solid sam-
ples and are noramlly not applied to liquid samples.
In the case of the liquid samples we are not normally
dealing with true compounds. In all cases we can
derive Information on the cations and anlons present
with a high level of accuracy; determine the oxida-
tion states such as Fe+2/Fe+3, or As+3/As*5; or ana-
lyze for organlmetallics. Figure 10 shows a technique
that was developed 1n the labs at TRW to analyze for
various organometalllcs that might be present in a
sample. The procedure combines chemical reduction
with the operations, and Identification, and quanti-
tatives capabilities of GC/MS. The samples were
treated to evolve the hydrides, which were trapped in
LNg. The trapped hydrides are then Injected Into
the GC/MS for analysis. This technique has shown
great utility and has been applied to samples from
an oil shale gasifier with good success.
Summary. The Level 2 analysis procedures do have
several problems associated with them. The main pro-
blem Level 2 analysis 1s the analysis gap between the
capability to measure major versus minor constituents
of the sample. X-ray diffraction which Is the only
true compound identification technique for inorganic
analysis, yet it requires samples to contain on the
order of 0.1 percent or better of the compound to be
identified. Trace compounds at the ppm level are
difficult if not Impossible to Identify unambiguously
with the present capabilities. Presented are a series
of logic charts to guide the analyst through a Level
2 procedure. However, since Level 2 1s a more sophis-
ticated approach to the analysis of an Inorganic sam-
ple, it will require a trained analyst to Interpret
the data and to direct the research. It Is not
envisioned to provide a prescribed set of specific
procedures but to provide an overall general approach
to the analyst. Since there is no one analytical
technique that can Identify all inorganic compounds,
most samples will require an Integrated approach
using a variety of methods.
In the future, it will be necessary to develop record-
ing formats for both Level 1 and Level 2 inorganic
analysis results and to develop more definitive cri-
teria that will control the decision to proceed from
Level 1 to Level 2.
More work 1s necessary'to apply these techniques to
environmental samples which tend to be more complex
and difficult than routine samples normally seen by
the techniques. When these problems are met, a
specific set of procedures can then be developed to
cover the potential problems which an analyst might
find for Level 2 analysis.
References.
^Hamersma, J.W., Reynolds, S.L. and R.F.
Maddalone, "IERL-RTP Procedures Manual: Level
1 Environmental Assessment," EPA-600/2-76-160a,
U.S. Environmental Protection Agency, Research
Triangle Park, N.C., June 1976.
2Hamersma, J.W. and Reynolds, S.L., "Field Test
Sampling/Analytical Strategies and Implementa-
tion Cost Estimates: Coal Gasification and
Flue Gas Desulfurlzation," EPA-600/2-76-093b,
U.S. Environmental Protection Agency, Research
Triangle Park, N.C., April 1976
3Cleland, J.G., and Klngsbury, G.L., "Multimedia
Environmental Goals for Environmental Assessment
Volumes 1 and 2", EPA 600/7-77-136a and b, U.S.
Environmental Protection Agency, Research
Triangle Park, November 1977.
4Nygu1st, R.A. and Kogel, R.O., "Infrared Spetra
of Inorganic Compounds", Academic Press, New
York, 1971
5Jenk1ns, R., Haas, D.J. and Paoline, F.R.
Norelco Reporter. 18 (2), 1 (1971).
83
-------
Organic Analysis for Environmental Assessment
Philip L. Levins
Arthur D. Little, Inc., Acorn Park, Cambridge, Massachusetts 02140
Abstract
Systematic measurement methods are being developed
for the determination of organic species in emission
and process streams. The methods are structured on
the Phased approach developed by the Process Measure-
ments Branch of IERL/RTP. Level 1 analysis methods
have been tested and a reporting format developed
which is consistent with a variety of assessment
objectives. Level 2 procedures are focused on
general broad spectrum analysis protocols and speci-
fic analyte procedures such as for PCB's, PAH, etc.
Introduction
Chemical analysis for environmental assessment is
currently based upon the Phased Approach established
by the Process Measurements Branch of EPA/IERL/RTP. 0)
The purpose of this paper is a review of the status
and procedures for Level 1 and Level 2 organic analy-
sis, with a focus on changes in the Level 1 pro-
cedures originally proposed, the development of a
consistent reporting format and various activities
in the development of Level 2 organic analysis pro-
cedures .
Level 1 Organic Analysis Procedures
The Level 1 sampling and analysis procedures are de-
scribed in detail In the EPA Level 1 Procedures
Manual' ' and will not be discussed in detail in
this paper. Some highlights of the procedures and
areas where changes have been made to Improve the
accuracy of the data will be reviewed.
Streams to be sampled for environmental assessment
include process and effluent streams which may be
gaseous, liquid or solid and which are sampled by
a variety of methods. Grab sampling procedures are
used for most of the liquid and solid samples. The
most complex sampling is that of the gaseous stream
which Involves the use of gas sampling for on-slte
analysis and vapor and partlculate sampling with
the SASS (Source Assessment Sampling System) train.
The Grab and SASS samples are returned to the labo-
ratory for analysis. A summary of the types of
samples generated for analysis is shown in the over-
view given in Figure 1.
Care is taken in handling of these liquids and solu-
tions to recognize that many of the materials of
interest are volatile (boiling point 100 - 300*C)
and will be lost if samples are taken to dryness.
The overall Level 1 analysis scheme for the organic
components is shown in Figure 2. The procedure
covers the quantitative analysis of the major vola-
tility classes by the procedures indicated below:
Volatility
gases
volatile
non-volatile
Definition
boiling point
boiling point
boiling point
< 100°C
100-300'C
> 300"C
Procedure
field GC
TCO
GRAV
The TCO procedures Is a laboratory gas chromatographlc
(GC) method for the volatile components. The GRAV
procedure Is a gravimetric method for the non-vola-
tile components.
The major chemical categories in each sample are de-
termined by first separating the sample according
to polarity on a silica gel liquid chromatography
(LC) column. Each of the fractions are then iden-
tified by obtaining Infrared (IR) and low resolution
mass spectral (LBMS) data.
Reporting Level 1 Analysis Results
The first complete set of data to be developed in
the Level 1 organic analysis are the LC results
which are reported in the form shown in Table 1.
The volatile (TCO) and non-volatile (GRAV) compo-
nents of each LC fraction are reported after com-
puting back to equivalent quantities in the entire
sample. The total quantity of each fraction Is then
derived from the sum of TCO + GRAV and the concen-
tration in the sampled stream is calculated from
the sample quantity and volume sampled.
Results of the IR analysis of each sample and LC
fraction are tabulated In terms of the frequency of
peak maxima, intensity ( weak, medium or strong)
and probable assignment. A typical example is given
In Table 2.
This overview shows that the organic analyses are
either done directly on the gas samples (materials
with a boiling point less than 110'C) and neat or-
ganic liquids, such as fuels, or on methylene chlor-
ide extracts of the samples or portions of the SASS
train. In the case of rinses from the SASS train,
such as the sorbent module rinse, it is necessary to
evaporate the rinses to dryness before proceeding
with the analysis, because the alcohol used in the
rinsing will Interfere with the liquid chromatography
step of the analysis.
The LRMS results are reported in terms of the cri-
teria outlined in Table 3. A typical report using
this format is shown in Table 4. Most of the sub-
sequent initial Level 1 Interpretation of data is
done using only the major compound category data.
Compound categories for LRMS interpretation are
chosen primarily from a list of about twenty-four
general compound groupings.
84
-------
Fi«ur« li
noun i uvti i OMUHK Mutnn FIOB ouaum
85
-------
Report ef LC Fractlonatton Result!
Saaole: Sortoent tit net - SASS 4
Equivalent ToUl Sa^le Quantities
Quantity actually taken far LC: 42 *g TO); 101 i
toUl saaple eitract.
Table 2
IR Report
Co*I Extract - IX Fraction 6
1. Major peaks and assignments
BUY. 3/10 of
tibia *
L1HS taporc * Coha 0»an QuancB tfaata
1. Cataterl.. Praaant
latanaltT
100 Phanala
1 Amlaea
100 latarocycllc Bltroiaa compounde
1 •ittllaa
2 . iubeataaeriaa. SpaclMe Compounda
lacaaaltT
100 a. Phanoli, alkyl aariaa
KU 94-1J2
Coapoaltlon CfH(0 - C|B)oO
1 b. Amlaaa
Toluamlne. HU 107, C?Ral
c. Hatarocycllc Nlcro|aa Compouada
100 1. Qulnolloai, alkyl aariaa
NW 129-157
Coapoiltloo C,U,» -
1 2. Aa lald.iolt. MU 17*
1 3. lanaothloaola. KW 135
4. lltrilaa
1 Toluonicrtla. HU 117, Ct»7»
3. OTKI»
1 laotopa eluatar at »5, 16. »7. f«, 9»
10 >/a lit
U/co
3300 •
3080-3000 , •
2950. 2925. 2850 t
1690. 16SO. 1600 •
1450 •
1200 »
7SO •
2. Unaulonrt *eak bands: 1270. 127S. 870. 810. UO c«'
3. Other marks: Cannot confirm phthalatts fouid by LRNS
OK or Ill/broad
aroMtlc CM
aliphatic CM
acid, ketone or a«
-------
Example Organic Extract S
Table 5
ry Table for a Sorbent ttodule Extract
Total organlcs, mg/m3
TCO, mg
GRAV, mg
Categories
Sulfur
Aliphatic HC's
Aromatlcs-Benzenes
Fused Arom <216
Fused Arom >216
Heterocycllc S
Heterocycllc N
Heterocycllc 0
Carboxyllc Adds
Phenols
Esters
LCI
0.61
5.2
13.
INT/mg/m3
100/0.6
10/0.06
LC2
0.74
19.
3.3
10/0.06
100/0.6
10/0.06
10/0.06
LC3
8.4
73.
180.
100/4
100/4
10/0.4
LC4
1.0
6.7
23.
100/0.5
100/0.5
10/0.05
10/0.05
LC5
0.33
3.7
7.3
/O.I*
/0.01*
/o.r
/0.01*
/0.01*
LC6
1.5
5.3
41.
100/0.7
10/0.07
100/0.7
10/0.07
10/0.07
LC7
0.50
0.1
15.
10/0.02
100/0.2
100/0.2
10/0.02
E
13.
110
280
mg/m3
0.6
0.06
0.06
5.
5.
0.5
1.
0.3
1.0
0.1
0.08
* Estimated assuming same relative Intensities as LC6, since IR spectra of LC5 and LC6 very similar.
Level 1
^mmntm*
ovnpm
Level 1
Analysis of Each
Sample
of Compound
Categories
Each Category
Finished
Exceeds Criterion
level 2 Analysis
Figure 3; DECISION LOGIC FOR PHASED LEVEL 1 - LEVEL 2 ANALYSIS
87
-------
Level 2 Sampling and Analysis Activities
Level 2 Analysis
The Level 2 methods for organic analysis are cur-
rently In the developnent stage and cover a wide
range of possibilities. Level 2 procedures include
the analysis of specific compounds tentatively iden-
tified in the Level 1 study, more comprehensive sam-
pling and analysis schemes and the development of
procedures for specific purposes such as PCB and
PAH analysis and cm-site extractive water sampling.
Level 2 Sampling
Most of the sampling procedures chosen for Level 1
were the best available and are still appropriate
for Level 2 studies. Some procedures will need
Improvement, while some might be made simpler or
more specific. Examples of additional sampling
procedures which could be used in Level 2 are listed
below:
Method
- Bags, inpingers
- Sorbents, Modified 5, Impingers
Participates - Method 5, Special
Liquids - Continuous (XAD-2), Purge & Trap
Solids - Stop Belt, Cutter
An interim Level 2 Procedures Manual has been
drafted and will be published in Feburary, 1978.
This manual explains each of these methods in
greater detail and gives specific recommendations
for certain categories of compounds as is shown
in Table 6.
Table 6
LEVEL 2 SAMPLING
BY CHEMICAL CATEGORY
Chemical Categories
Aldehydes
b.p. < 100°C « Cg)
b.p. > 100°C
Azo Compounds,
Hydrazine. Etc.
Fined Polycydic
Hydrocarbons Fused
Non-Alternant Polycydic
Hydrocarbons
Gaseous Streams
Sampling Treatment
Bisulfite
Impingars
SASS
Special
Reagent
Impingers
SASS
None
Resin Adsorption
None
Resin Adsorption
Chemical analysis following Level 1 studies will
focus largely on more accurate specific compound
identification. To this end several procedures
should be considered in addition to those used in
the Level 1 studies. Some key techniques are:
Thermal Analysis (TGA)
Gel Permeation Chromatography (GPC)
High Performance Liquid Chromatography (HPLC)
High Resolution Mass Spectrometry (HUMS)
Gas Chromatography/Mass Spectrometry (GC/MS)
Nuclear Magnetic Resonance (NMR)
Ultraviolet Spectroscopy (UV)
The objective In adding these procedures is to deal
better with the high molecular weight species (TGA,
GPC), further resolve complex fractions (HPLC) from
the LC separation and provide more complete and/or
specific analysis capabilities (HRHS, GC/MS, NMR,
UV).
Examples of recommended analysis methods for speci-
fic compound categories given in the interim Level
2 manual are shown In Table 7.
Tibl< 7
LEVEL 2 ANALYSIS
BY CHEMICAL CATEGORY
Akyf Hrikta
b4>.< 100°C
OC/MS or OC/ECO (totMrNl) on
GC/MS 01 OC/ECO ItoflOTMl) on
Sf-2260 (or OV-17)
1. lodon»lllc THmion of BtalHM
i OC/MS on SP-1000
1. RMIM PIm* HPLC or
1 OC/MS on Tonu (Dtnct Aqumi Inaction) or
X OC/MS on SP-1000 AfUr DvtntiM
FiMd Non-Atamm MycycUc
1. OC/MS on O*aU 400 or
2. Rcxrat or Norral PhM HPLC
In some cases, the chemical analysis data alone
from the Level 1 study will not have been sufficient
to direct a more specific Level 2 study. However,
positive bioteat results or other reasons may dic-
tate a more complete Level 2 study. To cover these
cases, a general comprehensive Level 2 analysis
scheme is envisaged as shown in Figure 4. This
scheme utilizes the additional methods mentioned
earlier to search for high molecular weight materi-
als and also recognizes the need to resolve the
complex LC fractions by HPLC prior to analysis.
The best available specific analysis techniques
are used to determine the composition of the sample,
including those components present at low concen-
tration levels.
88
-------
GENERAL LEVEL 2 SCHEME FOR SOLUTIONS OF SAMPLE EXTRACTIONS
Other Level 2 Related Activities
Several other studies are in progress In the Arthur
D. Little, Inc. laboratories related to more speci-
fic Level 2 studies and the development of alter-
nate procedures. Some of these studies are:
• The development of specific PCB and
PAH procedures using GC/HS
• Studies of the application of HUMS
• Evaluation of gas sampling using multi-
layer bags and
• Design of a continuous on-site extrac-
tive sampler for water based upon com-
bined resin systems.
The PCB and HRMS studies are described briefly in
this paper.
Polychlorinated Biphenyl (PCB) Procedure
Moat of the current procedures for PCB analysis
rely upon recognition of the typical chrooato-
graphic profile to identify the specific PCB (or
Ardor) and quantltatlon by some combination of
peak height measurement and calibration. This
procedure breaks down for combustion sources due
to preferential combustion of the lower chlorinated
biphenyls, resulting in distorted GC profiles.
This problem has been overcome by using GC/MS pro-
cedures and selected mass scanning. The basic
procedure is to measure separately the monochloro,
dichloro, etc. chlorinated biphenyls. Hypo-
thetical mass spectra for each of the chlorinated
biphenyls is shown in Figure 5. This figure shows
that a specific mass may be chosen to measure each
n chlorinated biphenyl group with minimum inter-
ference from the others. There is no interference
between groups separated by one chlorine atom. For
those separated by two chlorine atoms there is mass
spectral interference, but there is no chromato-
graphlc overlap between the groups. These two
factors can be combined as shown in Figure 6 to
provide relative retention time (RRT) windows at
selected masses for the specific measurement of
a particular chlorinated biphenyl group.
Using this procedure, the results shown below were
obtained on a sample whose composition was known
from previous detailed GC studies.
PCB Group
di
C12
C13
Clw
C15
Found by
GC/MS
0
2
26
55
17
0
Known
0
1.2
24.7
57.8
19.8
0.4
High Resolution Mass Spectrometry (HRHS) Matrix
Analysis
The complete analysis of a sample is done best when
there is some initial knowledge of the sample com-
position and, therefore, an appreciation of special
precautions or procedures that should be used. In
environmental assessment studies, the analyst is
all too often hampered by lack of any knowledge of
the sample chemistry.
High resolution mass Spectrometry (HRMS) provides
a basis from which one can obtain a complete over-
view of the probable sample composition. The HRMS
matrix approach is Intended to present a quick,
relatively Inexpensive overview of a sample's com-
position to guide further analysis.
The HRMS matrix method consists of a series of
computer programs which sort and file the data from
a high resolution mass spectrum. A typical HRMS
may consist of 400 - 1000 spectral lines and associ-
ated composition data. The type of data available
is shown in Table 8 for the three different com-
position peaks observed at m/e 196. Listed are
the peak Intensity (HGT), exact mass (DET MASS),
difference in exact mass (ERROR) between the obser-
ved mass and that computed for the composition given,
the "rings plus double bonds" (R+DB) value and the
composition in terms of the number of atoms of the
element listed.
89
-------
224
326
1
1
1
, 1
li
h
1
1
1 1
II
1 .
. i
1 1
il
1.
i. 1
il
ll
1.
,. il
ii i
i .
1
h,
«0 190 220 230 2S2 2B2M 2W322 332 3M 388300 400
a, a, a, a4 c% q, a,
Figure 5! HYPOTHETICAL MASS SPECTRA OF PCB-i
100
100
MflSS 258
50
100 150 50 100 150
Fl«ur« 6; ANALYTICAL HAB CHMOMATOMUHi ran AM AKOCLOR IM«
90
-------
Table 8
Example of HUMS Data
— Observed — Calculated
HGT PET. MASS ERROR R+DB £ H N £ £
19 196.08886 0.04 9.0 14 12 010
95 196.12577 0.57 8.0 15 16 000
35 196.21963 0.53 1.0 14 28 000
The R+DB value is a measure of hydrogen (H2) unsatu-
ratlon In the molecule and is a convenient index of
structural type. The R+DB value is computed from
the formula
R+DB • 1.0 + nc - 0.5 Hg +
where k values for some key elements are
N
0
S
Cl
0.5
0.0
0.0
- 0.5
The R+DB value computed for some polynuclear aro-
matic hydrocarbons is given in Table 9.
Table 9
R+DB Values for Polynuclear Aromatic
Hydrocarbons
R+DB
9
10
11
12
13
15
MW*
166
178
204
202
228
252
Composition
C13H10
Cli»H10
C16H12
C16H10
C18H12
C20H12
Compound
Fluorene
Anthracene
Aceanthrylene
Pyrene
Chrysene
Benzopyrene
'lowest possible value in R+DB group
**and related isomers
Through a combination of the R+DB value and the
chemical composition assignments, Che entire HRMS
output may be described as « simplified matrix of
data, such as shown in Table 10 for a solvent ex-
tract of an API separator waste.
Table 10
R+DB
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
R+DB Matrix:
Relative
CH (only)
5.9
13.6
8.4
5.4
8.9
4.7
4.8
19.4
8.2
4.6
5.7
1.1
0.1
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
API Waste Extract
Abundance
CH+N
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
by Composition
OHO
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
(Q)
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
CB+S
0.0
0.0
0.0
0.0
©
0.0
o
0.0
0.0
©
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
numerical values in matrix given as Z of total
Examination of the HRMS R+DB matrix of this sample
reveals that there are no nitrogen-containing
species (R+DB 9) and three types of sulfur species
(R+DB 4, 6 and 9) all at low levels. The sample
is mostly hydrocarbons, predominantly aromatic,
with no detectable species above pyrene (R+DB 12).
Further examination of the subset data used to pre-
pare the final matrix reveals that the R+DB 9 oxygen
and sulfur species are benzothiophene (R+DB 6) and
a series of thiophenols (R+DB 4).
Information derived from this type of matrix analy-
sis can be of great value in establishing criteria
for the full analysis.
91
-------
Further information concerning the material presen-
ted in this paper may be obtained by contacting the
author or the EPA Project Officer, Dr. Larry D.
Johnson, Process Measurements Branch, lERL/EPA/RTP.
The work described in this paper has been conducted
under EPA Contract No. 68-02-2150.
References
1. J. A. Dorsey, C. H. Lochmuller, L. D. Johnson
and R. M. Statnlck, "Guidelines for Environ-
mental Assessment Sampling and Analysis Pro-
grams; Historical Development and Strategy of
a Phased Approach", Draft Revision March 9,
1976.
2. J. W. Hamersma, S. L. Reynolds and R. F.
Maddalone, "IERL-RTP Procedures Manual:
Level 1 Environmental Assessment, Report No.
EPA-600/2-76-160a, June 1976.
92
-------
A CRITIQUE OF ORGANIC LEVEL-! ANALYSIS
by
Peter W. Jones and Robert J. Jakobsen
Battelle, Columbus Laboratories
Columbus, Ohio 43201
Abstract
This paper provides an objective review of
organic Level-1 analysis. It attempts to point
out present strengths and weaknesses, and to make
recommendations concerning how Improvements may be
made. Areas which are addressed include the
adequacy of sample size, the optimum LC separation
scheme, problems associated with contamination
both in the field and as a result of adsorbent
degradation, the utility of FT-IE, and LRMS
analysis.
Introduction
It is important to remember that the evolu-
tion of Level-1 was Intended, It was supposed to
happen. There was never any pretence that the
first drafts, prepared two years ago, would be
used without modification. The most recent up-
date is presently (February 1978) being assembled
by Research Triangle Institute.
This brief review will not attempt to cover
all aspects of organic Level-1 analysis, rather,
It will discuss selected aspects of analysis of
the organic Level-1 liquid chromatographlc frac-
tions.
Analysis of LC Fractions
The most straightforward way to address the
analysis of the LC fractions In organic Level-1
analysis is to consider each fraction in turn, and
discuss typical strengths and weaknesses as appro-
priate. We have thus chosen to examine a typical
sample from FBC coal combustion effluent, and will
add other examples by way of Illustration as
necessary.
Fraction 1
We do not normally expect to see anything in
Fraction 1, above normal background, and Figure 1
shows a typical Fraction 1 spectrum. All peaks
in this case were attributed to background contam-
ination.
However, appreciable contamination from sill-
cone grease is occasionally evident in Fraction 1,
as Illustrated in Figure 2. It is apparent that
such spectra arise when slllcone grease is used in
any part of the heated sampling train. Since such
strong absorbance would render further IR inter-
pretation either difficult or impossible. It is
recommended that the use of sillcone grease should
be discontinued. Besides interferring with qual-
itative interpretation, the gravimetric measure-
ments would also be rendered meaningless by the
presence of such contaminants.
FRACTION *l —TYPICAL PCTSI9TOIT BACKGROUND
— NOTHING DETECTED
3400
two
two
cm"1
I MO
1000
FIGURE 1. EXXON 1601-1
(Level-1 IR, m-10)
(m=40) Si-O-Si
STRETCH
SILICOME 6EEAQE
-MAIM COMPONENT
1800 I 1400 I iOOO I
1600 1200 800
cm'1
FIGURE 2. FLUIDIZED BED COAL COMBUSTION
MERC/SAMPLE 49, FRACTION 11
93
-------
Since IR analysis of Fraction 1 has never
provided any useful data, it is recommended that
qualitative analysis of this fraction should be
restricted to low resolution mass spectrometry
(LRMS), which may give useful details pertaining
to the molecular weight range of hydrocarbons
present.
Fraction 2
As in the case of Fraction 1, little useful
information is typically obtained by IR analysis
of this fraction, and again, LRMS would be a
better choice since polycycllc aromatic compounds
present give characteristic molecular ions by
this technique. Figure 3 shows a typical IR
spectrum of a Fraction 2 obtained by Fourier
transform infrared (FT-IR) analysis.
FBCTON*S TRACE OF XAD-Z
3400 2800
WOO
cnf
FIGURE 4. EXXON 1601-3
(Level-1, IR, m-10)
FRACTION^ NOHONB DCTECTED
3400 2800 2200 IbOO
FIGURE 3. EXXON 1601-2
(Level-1 IR, m-10)
1000
3200 I 2800
3000
i i i i i i w
1800 I (400 I 1000 I WO
IfcOO 1200 800
on
On observing data such as Figure 3, it is
easy to reach the subjective conclusion that the
sample size was too small; however, provided that
the analytical sensitivity was that defined by the
Level-1 protocol, the correct conclusion would be
that the fraction contains no compounds of con-
cern.
Fraction 3
Figure 4 shows the FT-IR of a Fraction 3 from
FBC coal combustion effluent. This spectrum re-
veals another common source of contamination,
XAD-2 resin used in the sorbent module of the SASS
train. Although XAD-2 does cause some background
difficulties, these have been found to be appre-
ciably less than those caused by the use of Tenax
as the organic vapor sorbent.
Referenced spectra of both sorbents are shown
In Figure 5. In a study of background problems
associated with organic Level-1, Fractions 3 and 4
most commonly exhibit interference from the sor-
bent used. Figure 6 shows representative IR spec-
tra of contamination originating from XAD-2 and
Tenax. Contamination due to the use of XAD-2 Is
typically not due to the presence of the resin
itself, but appears to be either decomposition
TiMx/KBrd«k
320G 2*00
3000
1800 1400 1000
ItOO 1200 800
cm
FIGURE 5. IR SPECTRA OF XAD-2 AND TENAX
products or other organic species which may be re-
leased from the body of the resin beads. On the
other hand, contamination caused by Tenax is most
typically solublllratlon of the resin Itself, and
as can be seen from Figure 6, Is typically about
an order of magnitude more severe than with XAD-2.
The continued use of XAD-2 In preference to Tenax
in this application Is recommended, although a
lower background would be preferable.
94
-------
XAD-2 BUnK.FrAcrion**
xl
s .~~"""
j
FIGURE 6. IR SPECTRA OF SORBENT BLANKS
Fraction 4
Figure 7 exhibits the presence of a phthalate,
which may or may not have been present as an Im-
purity. The concern regarding contamination orig-
inating from solvents used is very real, If it Is
remembered that evaporation of 1 litre of solvent
to 10 pi would concentrate a relatively involatlle
solvent contaminant by 10 .
FRACTION #4 —SOME XAD-2
— PHTHLATC OR OTHER
esreR (IMPURITY*)
1700 I WOO I 1000 I bOO
IfeOO 1200 800
cm'1
FIGURE 8. YORK RESEARCH-QUENCH
TOWER EFFLUENT
(Sample 13a, Fraction 4)
Fraction 5
Figure 9 shows the FT-IR analysis of a Fraction
5 from FBC coal combustion effluent, exhibiting the
presence of an ester together with a trace of phenol
or amlne. In this Instance, the weight of Fraction
5 was too small to warrant Level-2 analysis, and
thus, the application of the Level-1 protocol obvi-
ated the requirement for more extensive analysis.
FRACTION** — ESTER
— TRACE
3400
MOO
euo
cm
1400
WOO
FIGURE 7. EXXON 1601-4
(Level-1 IR, m-lO)
In one Fraction 4, shown in Figure 8, IR
analysis revealed the presence of large quantities
of two different phthalates. We suspect that both
may be real in this particular case.
3400
WOO
FIGURE 9.
ttOO
cm'1
IfcOO
1000
EXXON 1601-5
(Level-1 IR. m-25)
95
-------
Fraction 6
Figure 10 shows extremely intense spectra of
an ester and phthalates. In this instance, the
phthalates were most likely substituted compounds
present in the original sample, since the common
plasticier phthalates normally appear in Fraction
4.
FRACTION *fc—ESTER
— PHTHALATES
(IMPURITY?)
FRACTION *6 —WATER OF HYDRATION
— PROBABLY SOME SALTS
3400
2800
2200
cm'1
IfcOO
1000
3400
2800
2100
IbOO
1000
FIGURE 10. EXXON 1601-6
(Level-1 IR, m-100)
Fraction 7
Figure 11 shows a typical Fraction 7 spectrum,
containing highly polar species which are probably
polyfunctional.
reAcnoM*7—AMIDE?
—CARBONYL COMPOONO.POLAR
3400
2600
2100
»00
1000
FIGURE 11.
EXXON 1601-7
(Level-1 IR, m-100)
Fraction 8
Figure 12 shows an FT-IR spectrum which is
rendered virtually useless by an earlier modifica-
tion to Level-1 In which water was added to the
eluate for Fraction 8. This spectrum shows water
of hydration, salts, and evidence of solublllzatlon
of the silica gel column; the broad water band
would mask much available Information In the
spectrum
FIGURE 12. EXXON 1601-8
(Level-1 IR, m-50)
While an expected update of Level-1 protocol
is expected to delete the analysis of Fraction 8,
this may simply be avoiding artificial problems
created by the addition of an aqueous eluate to
the original non-aqueous LC fractionation scheme.
Examining the utility of non-aqueous Fraction 8
eluate may be worthwhile, there does not appear to
be any data available in this regard at present.
We would agree, however, that Fraction 8 as pres-
ently used is of doubtful value.
Mote on Use of FT-IR
Although the organic Level-1 protocol requires
the use of a spectrometer equivalent to a Ferkin-
Elmer 521 or 621, many samples which have been
analyzed to date could not readily be evaluated
without the use of FT-IR. Samples have normally
been run successfully, only with the equivalent
of 10-20X scale expansion.
However, the real power of FT-IR is that of
spectral subtraction. By way of illustration,
consider the spectra in Figure 13.
Spectrum A is a Fraction 5,
Spectrum C is a corresponding Fraction 4, and
Spectrum B is the result of subtracting C
from A
From preliminary examination of Spectrum A, It Is
not clear whether this represents a single aromatic
compound, or a mixture of several. Spectrum C is
readily identifiable as a ketone. Subtraction of
C from A yields a spectrum which is readily identi-
fiable as a quinone, which leads to the conclusion
that A was in fact a mixture of a quinone and a
ketone. Spectra do not always subtract as cleanly
as this, but the facile elimination of the blank
spectrum from every fraction is a very powerful aid
towards Interpretation, and highlights the utility
of FT-IR.
96
-------
NO USEFUL DATA
FIGURE 14. EXXON 1602, FRACTION 8
30OO
FIGURE 13.
2000
Crn1
1000
EXAMPLE OF THE UTILITY OF
SPECTRAL SUBTRACTION BY FT-IR
DICHLOROBENZOIC ACID?
^ COOH
(IR INDICATED CARBOMVl)
9 M
im ne iv
Note on Use of LRMS
Our experience with LRMS has been that it does
not always prove useful, but that when it does do
so, the information provided can be very useful.
We recommend the continued selective use of LRMS
in organic Level-1. LRMS is especially useful for
Fractions 1 and 2, it would be useful to delete IR
analysis in favor of LRMS for these fractions.
Figures 14 and 15 show ion chromatograms from
LRMS analyses where no useful data and a specific
conpound identification were respectively obtained.
Many ion chromatograms which we have observed are
similar to Figure 14, which may be due to lack of
adequate sample, or volatile compounds which are
lost before analysis can be affected. However,
strong Ion chromatograms such as Figure 15 are
observed on sufficient numbers of occasions to
merit the use of LRMS whenever the Level-1 fraction
weight threshold is exceeded.
FIGURE 15. EXXON 1023-22, FRACTION 6
CONCLUSION
In conclusion, we wish to reiterate some of the
more Important points raised earlier:
Avoid use of silicone grease,
Use highest purity solvents, recognize
that some phthalates may be "real",
Use of a non-aqueous Fraction 8,
Utility of FT-IR for spectral subtraction,
Discontinue IR analysis in Fractions 1
and 2 in favor of LRMS.
Finally, we would reiterate that it would, In-
deed, be surprising if changes to the Level-1 organ-
ic protocol did not continue to be made. Whenever
Improvements in the analytical stratedy become evi-
dent, It la Important that these should be communi-
cated to the analytical community as soon as
possible.
97
-------
ENVIRONMENTAL ASSESSMENT MEASUREMENT
TECHNIQUES FOR FUGITIVE EMISSIONS
Henry J. Kolnsberg
Senior Project Engineer
TRC - THE RESEARCH CORPORATION of New England
Wethersfield, Connecticut
Abstract
The paper describes the sampling and measure-
ment techniques currently being employed or de-
veloped to determine the Impact of industrial
fugitive emissions on the environment. Three
general sampling techniques for airborne fugitive
emissions and one for waterborne fugitive emissions
as stormwater runoff are presented and evaluated
with respect to their inherent accuracies and
limitations. Site-specific modifications of the
general techniques used in recent studies at a
variety of industrial locations are described and
the results of the measurement programs reviewed.
Efforts toward the development of a fugitive
ambient sampling train for the measurement of air-
borne particulate and organic emissions are
summarized.
Introduction
Fugitive emissions are those pollutants
that are transmitted into the ambient atmosphere or
into ground or surface waters without first passing
through some stack, duct, pipe or channel designed
to direct or control their flow. Known also as
non-point source emissions, they are generated by
such a large variety of industrial processes and
operations that almost every industrial site must
be presumed to include some degree of environmental
pollution attributable to such emissions. No
environmental assessment at an Industrial location
can, therefore, be considered complete unless an
accounting of the impact of fugitive emissions is
included.
It is impossible to generalize as to the
Impact or magnitude of fugitive emissions at in-
dustrial sites. The almost endless variety of
combinations of site-specific parameters affecting
the generation and transmission of fugitive emis-
sions; such as the number, size and locations of
sources; site topography; and local meteorological
conditions; requires that each site be considered
individually. The relative magnitude or impact of
such emissions as compared to that of the more
traditional point source emissions can, however,
be assumed to be generally increasing. Improve-
ments in programs to control many point source
emissions, fostered by both technological advances
in control equipment and more rigorous enforcement
of emissions regulations, have effectively reduced
the impact of such emissions. The remaining un-
controlled fugitive emissions have thus been in-
creased In relative Impact, in some instances even
to the degree of becoming the prevalent source of
pollution at a site.
The measurement of industrial fugitive emis-
sions poses some unique problems. Standard stack
or similar sampling techniques are of little use
in typical airborne fugitive emission situations
where the pollutants exist In a poorly defined
plume or cloud in generally low concentrations.
Grab samples from the atmosphere will usually be
of such low concentrations that meaningful measure-
ments are Impossible, or they will contain a
preponderance of point-source or background
pollution so that Identification of the fugitives
cannot be made. Waterborne fugitive emissions
transported to receiving water bodies will be
similarly diluted or masked by other pollutants so
as to preclude their identification.
The Environmental Protection Agency, through
a number of its contractors, and other research
and industrial organizations have addressed the
problems and developed a number of techniques to
measure both airborne and waterborne fugitive
emissions. This paper describes those techniques
that are generally applicable to the requirements
of Industrial environmental assessments.
Fugitive Emissions Measurement Techniques
Airborne Emissions
Airborne Industrial fugitive emissions may,
in general, be measured at their source, before
the pollutants begin to diffuse into the ambient
air; in the air Immediately surrounding their
source, where the diffusion is limited to a
relatively small volume of air; or In the ambient,
where the diffusion is extensive. The measurement
techniques currently in use or under development
are, respectively, the quasl-stack, roof monitor
and upwind-downwind sampling methods.
The techniques, described below, may all be
used with suitable samplers and analysis pro-
cedures to measure both particulate matter and
gaseous pollutants. Modifications or adaptations
to the equipment or procedures may be required to
meet special needs for sampling a few specific
pollutants, but the techniques as described may be
generally considered to be applicable to all
classes of airborne pollutants.
Quasl-stack Sampling*1^
This method captures fugitive emissions at
their source in a temporarily Installed hood and
transmits them through a duct of regular cross-
sectional area where standard stack sampling
98
-------
techniques are used to measure the pollutant
concentrations and the flow rate of the emission-
carrying air stream. Pollutant source strengths
are then determined as the product of these two
measured values. A simplified quasi-stack sampling
system Is shown in Figure 1.
Figure 1. Simplified quasi-stack sampling system.
The quasi-stack sampling method is the most
accurate of the fugitive emissions measurement
techniques since it captures virtually all of the
emissions from a given source and conveys them
with a minimum of dilution by transport air to
their sampling devices. Accuracies in the range
of ±25Z can be achieved with a carefully designed
system. This method is also the least applicable
of the techniques since its use must be restricted
to those emission sources that can be physically
and operationally isolated and are arranged to
permit the installation of the system in a manner
that will not interfere with normal plant opera-
tions or alter the character of the emissions or
their generating process.
(2)
Roof Monitor Sampling^ '
This method is used to sample the emissions
from processes or operations taking place within
buildings or enclosures with only a small number
of openings to the ambient atmosphere. The struc-
ture serves as a large hood, confining the emis-
sions to a finite volume of air before trans-
mitting them through an opening, such as a roof
monitor, wall vent, door or window, to the outside
air.
Samples of the emissions are taken at the
opening, using hi-vol or similar filter-type
samplers for particulate emissions and either grab
sampling or portable gas analyzer trains for
gaseous emissions, to determine their concen-
tration in the transport air. The transport air
flow rate through the opening is also measured,
using the standard techniques. The combined
source strengths of all sources producing emis-
sions inside the enclosure is then determined as
the product of the measured concentration and
flow. Figure 2 shows an arrangement for obtaining
samples and flow data by traversing a set of
instruments across a roof monitor. A number of
such arrangements or a network of fixed instru-
ments may be utilised for large area openings.
Figure 2. Traversing roof monitor sampling system.
The roof monitor sampling method is not as
accurate as the quasi-stack method since a sig-
nificant portion of the emissions generated
within the enclosure may escape measurement
through other vents and since a much higher degree
of dilution by transport air occurs before samp-
ling. Accuracies in the range of ±50Z can be
expected with a well designed system. This method
is generally more applicable than the quasi-stack
method since it does not require the Isolation of
sources and will not Interfere with normal plant
operations. It does, however, require instru-
mentation and trained personnel capable of making
measurements of usually low air velocities through
relatively large openings and involves mass
balances of small quantities of materials.
Upwind-Downwind Sampling
This method may be used to measure the emis-
sions from almost any source after they have been
transmitted into the ambient air. Emission
concentrations are determined in samples taken
from the air approaching (upwind) and leaving
(downwind) a source or an entire industrial site,
and the source contribution determined as the
difference between the measured values. Hi-vol
samplers are usually employed for particulate
matter and grab samples for gaseous emissions.
The calculated source contribution is then used in
proven diffusion equations usually embodied in
computer programs, along with the measured wind
speed and direction and topographic data, to back-
calculate the emission source strength.
The upwind-downwind method is the least
accurate of the fugitive emissions sampling tech-
niques since only a small portion of the emissions
can be sampled for analysis and the flowrate of
the large volume of transporting air cannot be
directly measured. Measured values of pollutants
can generally be expected to be within a factor of
99
-------
2 or 3 of the actual values. The method is the
most universally applicable of all techniques,
capable of measuring emissions from large and
small sources located Indoors or outdoors under
almost any operating conditions or schedules. It
is, however, sensitive to external influences such
as weather conditions and changing wind directions
or speeds.
Waterborne Emissions
The principal mechanism for the transportation
of waterborne fugitive emissions to receiving
water bodies near industrial sites is the runoff
of storm water from relatively heavy rainfall
events and melting snow. Runoff will typically be
encountered from such sources as material and
waste storage piles, large open areas such as
material transfer yards and parking areas, and
flat building roofs that do not drain into storm
sewers or drainage systems. The runoff may con-
tain suspended particulatc matter from ground-
based materials or settled airborne emissions,
dissolved solids from similar sources, and sus-
pended or dissolved liquids and oils from process
operations, spills, and leaks. Such runoff may be
sampled for analysis either as overland runoff as
it flows on the ground surface near its source or
in open channels where such runoff collects in its
flow path to the receiving body. The contribution
of pollutants from a specific source or combin-
ation of sources is determined by analyzing runoff
samples taken at short time intervals to establish
the concentration of the pollutants as a function
of the simultaneously measured rainfall or runoff
flow. Integration of the area under a curve
plotted of measured concentration versus total
flow or rainfall then provides, by extrapolation,
the amount of pollution that can be expected for
any rainfall or snow melt. In combination with
historical precipitation data, this procedure pro-
vides an estimate of seasonal or annual pollution.
Runoff sampling is, in general, most effec-
tively conducted during rainfall events when the
rainfall rate is high enough to produce visible
runoff within about a half hour of its onset and
which continues for several hours thereafter.
Samples of overland runoff are obtained in collec-
tion plugs similar to that shown in Figure 3. A
number of plugs are driven Into the ground at
locations near the source before the onset of the
rain so that their top faces are just below the
surface of the surrounding ground material. The
runoff then flows across the top screen which
restricts the flow of entrained particulate matter
and permits suspended particles and dissolved
materials to pass into the plug. The plugs are
manually collected and replaced at appropriate
Intervals during the storm event to provide data
on the decrease in pollutant concentration from
the usually high concentration "first flush" as
the total rainfall and runoff flow Increases.
Rainfall measurements are made at. about the same
intervals to establish the total flow and flow
rate as a function of time.
Figure 3. Overland runoff sampling plug.
Open channel runoff Is sampled by dipping
collection bottles directly into the runoff
stream, either manually or with an automatic
sampler that can be preprogrammed to collect up
to four separate samples in rapid sequence at
selected Intervals. Runoff flow in the channel is
measured using temporarily installed weirs or
flumes. All samplings and measurements are made
at appropriate Intervals to provide data on the
concentration - flow - rainfall relationships.
These measurement techniques are generally
applicable to all relatively large sources of
waterborne fugitive emissions. While their accura-
cies are difficult to quantify, since no method to
accurately determine the absolute 'levels of pollu-
tion attributable to a source exists, it is rea-
sonable to assume that the values obtained will be
within a factor of 2 or 3 of the actual values.
Measurement Method Applications
The basic techniques described above
for the measurement of fugitive emissions have
been used, with modifications to meet the require-
ments for the sampling of specific pollutants or
sources, by a number of organizations In recent
years with varying degrees of success. This
section reviews a few of the more successful
applications of the techniques to illustrate how
they might be used in industrial environmental
assessments.
100
-------
Quaai-Stack Measurements at a Gray Iron Foundry
As one phase of an Environmental Protection
Agency sponsored contract to develop fugitive
emissions measurement techniques, an environmental
consultant performed a quasl-stack sampling pro-
gram to identify and quantify the fugitive emis-
sions generated by the pouring and subsequent
cooling of a medium-sized gray iron casting in «
sand mold.
A capture hood that enclosed the mold on
three sides was fabricated from sheet metal and
connected to a horizontal duct of circular cross
section. An exhaust fan was utilized at the end
of the duct to draw the emissions through the
duct to the sampling ports. An Ikor continuous
particle monitor was connected at one sampling
port to measure dry, filterable partlculate matter;
an EPA Method 5 Participate Train was connected
to another port to measure partlculate matter,
organic and inorganic condensibles, particle size,
and particle size distribution; and a Cascade
Impactor with an EPA Method 5 Condensible Train
connected at a third port to provide additional
measurements of particulate matter, organic and
inorganic condensibles and particle size distri-
bution. These three redundant samplers were used
as part of the developmental study to evaluate the
sampler effectiveness. In a normal environmental
assessment measurement program, only one would be
used. A fourth sampling port was used to draw
gaseous samples through a flame ionization detec-
tor for total hydrocarbon analysis and a non-
dispersive infra-red analyzer for carbon monoxide
analysis. The test set-up as installed at the
foundry is shown in Figure 4.
Figure 4. Gray iron foundry quasi-atack sampling
arrangement.
The hood opening and fan were designed to
provide a face velocity of 150 feet (46 meters)
per minute at the opening, ensuring the capture
of the exhaust gases from the pouring and cooling
operation. The duct diameter was sized to provide
an exhaust gas velocity of 2500 to 3000 feet (770
to 925 meters) per minute, sufficient to ensure
that particle deposition in the duct due to
settling would not occur.
The results of a series of 23 separate pour-
Ings of castings ranging In size from 250 to 1000
indicated that the emissions were, in total, not
significantly different from the background emis-
sions in the foundry. The continuous particle
monitor traces showed an Instantaneous peak value
of partlculate emissions about an order of magni-
tude higher than the average reflected by the other
samplers, but the peak was of such short duration
that the total particulate loading was almost
unaffected.
Roof Monitor Measurement at a Graphitlzlng Plant
(6)
In a program designed to prove compliance with
air pollution regulations, a manufacturer of car-
bon and graphite products performed a study of Its
total plant particulate emissions using the roof
monitor technique. The plant contains a large
variety of process operations, each producing
particulate emissions at different rates and on
varying schedules, with the total of all process
emissions being transmitted to the atmosphere
through a large roof monitor. A careful study of
the flow patterns at the monitor determined that
samples taken at three locations In the monitor
would provide data representative of the total roof
monitor flow.
The emissions were sampled simultaneously
and continuously at the three selected locations
using standard hl-vol filters modified to permit
directional sampling from below. Air flow rates
through the filters were maintained at about 1.3
to 1.4 cubic meters per minute as Indicated by
rotameter measurements at each location. Filters
were changed when the lower flow rate was indi-
cated. This flow range provided sampling at
slightly less than the isoklnetlc velocity and
biased the measured values of concentration
slightly toward the high side. Actual roof moni-
tor air velocities at each sampling location were
•monitored with propeller electronic anemometers
connected to strip chart recorders to provide a
continuous record. Volumetric flow rates were
then calculated using integrated averages of the
measured velocities.
The results of a week of continuous sampling,
with filter changes providing average values for
periods of from 1 to. 15 hours, showed average
particle emission rates for the plant of from 1
to 9.5 pounds per hour. Correlation of the mea-
surements with a log of plant activities enabled
.the estimation of emission rates for specific
processes, which is necessary information for
planning a control program If the need for con-
trols is established.
Upwind-Downwind Measurements at Integrated Iron
and Steel Plants UJ
As part of an EPA sponsored study to deter-
mine fugitive emission factors for a variety of
open sources at integrated iron and steel plants,
a research organization performed a series of
measurements of the particulate emissions gen-
erated by slag loading operations using a modi-
fied upwind-downwind technique.
101
-------
This loading operation, which utilizes a high
loader to transfer the slag from a storage pile
into trucks or rail cars, takes place in an open
area and produces an intermittent cloud or a
series of puffs of high concentration suspended
particulate matter. In order to effectively
isolate the emissions from the loading operation,
samples had to be taken within the cloud as close
to the source as possible. To accomplish this,
the contractor designed and fabricated a portable
sampling system mounted on a trailer that could be
positioned within a few meters of the vehicle
being loaded. The sampling system consisted of a
tower, 6 meters high, with a 5 meter crossarm
located about 4.5 meters high. Exposure sampling
heads, containing a settling chamber for large
particles and an 8 x 10 inch glass fiber filter,
were mounted at heights of about 3 and 6 meters on
the tower and about 0.7 meter from the tower
centerline on each side of the crossarm. Smaller
samplers, containing 2 inch diameter glass fiber
filters, were mounted on each side of the cross
arm about 2.4 meters from the centerline. A high
volume cascade impactor and a standard hi-vol
filtration unit were mounted as close to the tower
as possible at cross arm height to provide par-
ticle size distribution data and back-up exposure
data. The arrangement of the sampling heads was
designed to encompass about 90Z of the mass flux
of particulate matter in the cloud with the assump-
tion that the particle distribution was normal.
In operation, the inlets of all sampling
heads were pointed into the wind and the sampling
velocities were adjusted to match the local wind
speed as measured by anemometers mounted near the
top of the tower and at the base of the cascade
impactor. Nominal sampling rates of about IS
cubic feet (0.42 cubic meters) per minute for the
four main sampling heads and about 0.75 cubic feet
(0.02 cubic meters) per minute for the two auxili-
ary samplers were varied as required to provide
nearly isoklnetic sampling.
A series of three samplings was made for
loading operations of each of two different slag
types containing high (7.3Z) and low (3.0Z) silt
concentrations. Each test sampled, In a 30 to 40
minute period, the loading of from 140 to 175 tons
of slag by a 10 cubic yard front end loader. The
samplers were run only during actual loading
operations. Background particulate concentrations
were measured upwind of the source with standard
hi-vol filtration units.
The filters were collected and analyzed after
each sampling and their exposures plotted versus
sampler location. The resulting curves were then
graphically Integrated to determine the total
exposure or mass flux. An emission factor was
then determined by dividing the exposure by the
mass of the material loaded. This, factor was then
adjusted by the application of correction factors
determined from the measurements to include only
particles less than 30 micrometers in diameter and
to correct the measured or calculated values to
isokinetic conditions.
The emission factors thus determined were in
almost perfect agreement in two of the three
samplings for each slag type and were consistent
with expected influences of silt content. Particle
size distributions and concentrations determined
with the sampling system were in good agreement
with those determined with the hi-vol sampler and
the cascade Impact or.
Stormwater Runoff Measurements from Utility Plant
Coal
As part of an EPA sponsored study to evaluate
the nature and extent of non-point source water
pollution from industrial operations, a contractor
performed a program to measure the Stormwater run-
off from coal storage piles at electric generating
stations located on rivers and the effect of the
runoff on the rivers. The program utilized samp-
ling plugs to obtain overland runoff samples at
selected locations between the coal piles and the
rivers and dip-sampling bottles to obtain open-
channel samples in a water discharge canal and at
the outlets of covered drains. Flow rates were
determined from rainfall data for the overland
runoff and from velocity measurements in the
channels. Samples of river water were taken at
different depths at sampling stations upstream and
downstream of the plants and the river flow rates
estimated from velocity measurements at a point
where the river's depth profile was determined.
The runoff and river water samples were
analyzed to determine the concentrations of total
suspended and dissolved solids, Iron, aluminum,
manganese, sulfate ion, and total alkalinity or
acidity. The concentrations In the runoff were
generally of the magnitudes anticipated and dis-
played the expected high "first flush" effect, then
declined as the rainfall increased to dilute them.
The measured concentrations and measured or esti-
mated runoff flows provided reasonable estimates
of the total pollutants that would enter the river
as the result of runoff from a storm event, but
were generally too low to have any measurable
effect on the concentrations in the river.
Applications to Environmental Assessments
None of the measurement programs described
above was specifically intended to provide data
applicable to an environmental assessment. Each,
however, In the fulfillment of their various
overall objectives did produce data that would be
useful la establishing the Impact of fugitive
emissions from the sources measured in the programs.
The quasl-stack measurements of the foundry casting
operation, for example, could be used to calculate
a representative emission factor relating emis-
sions to the amount of metal poured and/or the
number of pourings. Such a factor would be useful
in estimating the total emissions that would be
generated in any given period based on production
schedules .
The roof monitor measurements made at the
graphite plant would be directly applicable to an
environmental assessment study, either by utilizing
the measurements as representative of the total
plant emissions or by using the measurements to
develop emission factors for the Individual plant
operations and applying the emission factors to
production schedules.
102
-------
Ac the integrated Iron and steel plant, the
upwind-downwind measurements resulting in the
establishment of emission factors for a single
loading operation could be used to predict the
amount of suspended partlculate matter from all
such operations within the plant. The emission
factor data obtained could be used with diffusion
equations to predict how much of the suspended
particulate matter would be carried beyond the
plant boundaries under a variety of meteorolo-
gical conditions.
The measurements made of the stormwater run-
off from the utility plant coal piles could be
used to establish factors to relate pollutant
generation rates to rainfall intensity, duration
and frequency. Used in combination with historical
data on past rainfalls, they can predict, with a
reasonable degree of accuracy, the pollutant
levels that can be expected for any period. The
measured values can also be used as input data for
modeling the impact of runoff on receiving bodies
in programs currently under development.
Measurement Method Developments
The measurement methods and applied modifi-
cations described thus far provide the means to
measure fugitive emissions from almost any indus-
trial source for environmental assessment purposes.
In many instances, however, these methods provide
more information than is actually required or are
not practical from the standpoints of time or
equipment requirements or excessive costs. There
is a need for simpler, less expensive measurement
methods that may be generally applied to a large
variety of fugitive emission sources to provide
quicker, if somewhat less accurate results.
The EPA, has sponsored, through its con-
tractors, the development of one such method, the
Source Assessment Sampling System (SASS); and is
currently sponsoring the development of a second
method, the fugitive ambient sampling train (FAST).
The SASS, though not developed for fugitive
emissions sampling, has some practical application
In the assessment of some sources of fugitive
emissions.(8) The SASS is designed to collect
particulate matter, organic species and volatile
trace elements from point sources. It consists of
« stainless steel probe that captures the emis-
sions and transfers them into a convection oven
containing-three cyclones and a filter assembly
that collects and separates particulate matter
into four size classifications. The particle-free
sample is then passed through a canister of XAD-2
adsorbent resin which removes most of the organics,
and into a series of impingers that collect the
condensed volatile inorganics. The sample is
drawn through the train by a 10 cubic feet (0.28
cubic meter) per minute vacuum pump. A control
module provides variable pressure,' temperature,
power and flow conditions as required for each
sampling situations. The SASS Is commercially
available and is readily adaptable to use any
combination of its components.
The SASS has been recommended for fugitive
emissions sampling where quick results are desired
from sources that enable the estimation or simul-
taneous measurement of the fugitive emission
carrying air stream volume or flow rate. Such con-
ditions exist In many roof monitor type locations
and even in a few open plume arrangements. In such
situations, the SASS can provide good estimates of
particulate matter and organic species emissions
with almost no set-up time.
The FAST, designed to provide large samples
of partlculate matter separated into two size
classifications simultaneously with a smaller
sample of organics, is currently in the preliminary
design phase. The system will consist of two
separate portable modules. One module will contain
the sampling train consisting of a 200 cubic feet
(5.6 cubic meters) per minute cyclone and glass
fiber filter for particle collection and a 5
cubic feet (0.14 cubic meters) per minute XAD-2
resin canister for organics. The second module
will contain the blower and vacuum pump for the
sampling train. The FAST will provide a 500
milligram sample of partlculate matter in an 8
hour sampling period at most industrial locations
where sampling of pollutants after they have
reached the ambient atmosphere is required. The
successful development of the FAST will provide
larger samples in shorter time periods than is
possible with any standard hl-vol sampler now in
use, and will greatly simplify and shorten
upwind-downwind measurement programs.
Additional developmental effort is required
for both generally applicable measurement methods
and methods to identify and quantify specific pol-
lutants. An advance in the state of the art to
the level now existing for point sources will
facilitate the Inclusion of fugitive emissions
measurements in future environmental assessments,
ensuring that the total impact of the emissions
from industrial sites is considered.
References
(1) Technical Manual for the Measurement of
Fugitive Emissions: Quasi-Stack Sampling
Method for Industrial Fugitive Emissions EPA-
600/2-76-089c May 1976.
(2) Technical Manual for the Measurement of
Fugitive Emissions: Roof Monitor Sampling
Method for Industrial Fugitive Emissions EPA-
600/2-76-089b May 1976.
(3) Technical Manual for the Measurement of
Fugitive Emissions: Upwind/Downwind Sampling
Method for Industrial Emissions EPA-600/2-76-
089a April 1976.
(4) Sampling and Modeling of Non-Point Sources at
a Coal-Fired Utility EPA-600/2-77-199 Sep-
tember 1977.
(5) Development of Procedures for the Measurement
of Fugitive Emissions EPA-600/2-76-284
December 1976.
(6) Souka, Abbas F., Continuous Roof Monitor
Emission Tests, in Symposium on Fugitive
Emissions Measurement and Control EPA-600/2-
76-246 September 1976.
(7) A Study of Fugitive Emissions from Metallur-
gical Process, unpublished draft final report
of EPA Contract 68-02-2120.
(8) IERL-RTP Procedures Manual: Level 1 En-
vironmental Assessment EPA-600/2-76-160a June
1976.
103
-------
SAMPLING AND ANALYSIS PROCEDURES
FOR SCREENING INDUSTRIAL EFFLUENTS
FOR PRIORITY POLLUTANTS
William A. Tel Hard
Gail S. Goldberg
Effluent Guidelines Division
Office of Water 6 Hazardous Materials
Environmental Protection Agency
Abstract
The Environmental Protection Agency agreed to review and revise regulations
based on the Best Available Technology Economically Achievable for effluents in 21
industrial categories, as a settlement of several cases in the District Court for the
District of Columbia. An integral part of this process is a technical study to deter-
mine whether or not compounds from a list of 65 materials (single compounds and unde-
fined classes of compounds) exist in industrial waste waters. The Effluent Guidelines
Division in EPA has established an unambiguous list of 129 compounds, referred to as
priority pollutants, which it believes fulfills the requirements of the court order and
can be determined analytically. To maintain consistent sampling and analytical pro-
cedures for 21 industrial studies, EPA has developed a protocol of the sampling and
analytical methods to be used for the screening of priority pollutants in waste water.
This protocol represents the incorporation of the current state-of-the-art procedures
for the sampling and analysis of the priority pollutants. These analytical procedures
include: purge and trap gas chromatography-mass spectrometry (GC/MS) for volatile
organic compounds; semi-volatile organics are done by a liquid-liquid extraction with
GC/MS and metals are determined using an inductively coupled argon plasma emission
spectrometer.
The Effluent Guidelines Division
within the Environmental Protection Agency
(EPA) is responsible for the preparation
of industrial effluent limitations guide-
lines. Part of the development of these
regulations, under the Federal Water Pol-
lution Control Act, involves a technical
study including treatment technology and
waste water analysis. In June 1976, EPA
agreed to review and revise regulations
based on Best Available Technology Econom-
ically Achievable (BAT) for effluents in
21 industrial categories, as a settlement
of several cases in the District Court for
the District of Columbia.
In order to carry out these respons-
ibilities, EPA developed an analytical
protocol of sampling and analytical meth-
ods to be used in the screening of prior-
ity pollutants in industrial waste waters.
The Effluent Guidelines Division, in order
to maintain consistent sampling and ana-
lytical procedures, minimize duplication
of effort and enhance comparability of
data for 21 industrial projects, needed
this protocol to serve as a reference
for the various labs involved. The proto-
col is specifically titled, Sa.mptJ.ng and
Ano£yA.iA P/ioceduAea jo* Sc4.ee>u.ng oj In-
du*t>U.at E66t.ue.nti, fai PtUolity Pottu.ta.ntl>.
The protocol, revised in April 1977, rep-
resents the incorporation of current
state-of-the-art procedures in analytical
chemistry. These analytical procedures
include: for volatile organic compounds,
a purge and trap gas chromatography-mass
spectrometry (GC/MS) procedure; semi-
volatile organic compounds are determined
by a liquid-liquid extraction and GC/MS
and metals analysis by atomic absorption
or inductively coupled argon plasma emis-
sion spectrometry. However, metals anal-
ysis is only referenced in the protocol.
In as much as possible, the protocol rec-
ommends preparation of samples, special
apparatus, materials and data reporting
formats. Most of all, it should be kept
in mind that the protocol does not repre-
sent formally approved standard methods at
this point in time. The protocol shall be
revised in the future, pending the com-
pletion of contracts attempting to vali-
date these procedures. Moreover, these
analytical methods were developed for use
in analyzing all 114 organic compounds in
all 21 industries. This means that the
protocol has limited applications and in
some cases, a more specific procedure may
be more accurate.
Before these GC/MS methods were
finally selected, several others, such as
specific methods by groups of substances,
were considered. Various priorities dic-
tated the selection process. One of these
requirements was the short time period
allowed for each industry's technical
104
-------
study. In order to meet court deadlines,
each project had one year in which to
survey, sample, analyze and categorize
the waste water and treatment systems
involved. Not only was time a commodity
in short supply, so were the funds.
While analysis by groups may have
certain advantages, such as the use of
additional, specific GC columns, this
procedure takes longer and costs more.
The rationalization for the use of more
general procedures, was that the protocol
is designed for the initial screening
phase of a technical study. The technical
studies were divided into a screening
phase, verification phase and standard
setting. For the screening phase, the
main focus is to determine the presence
or absence and order of magnitude of the
129 priority pollutants. Therefore,
GC/MS was selected for screening because
it was fast, less expensive than single
compounds analysis, and reliable for semi-
quantitative work.
Court Settlement Agreement
With the signing of the Settlement
Agreement, a broad new program was init-
iated to control the discharge of danger-
ous substances into U.S. waterways. The
strategy, as articulated in the Settlement
Agreement, will focus on regulating dis-
charges on an industry-by-industry basis.
Appendix B of the court order listed those
point source categories to be studied.
They are: Timber Products, Steam Electric
Power Plants, Leather Tanning and Finish-
ing, Iron and Steel Manufacturing, Petro-
leum Refining, Soap and Detergent Manu-
facturing, Autos and Other Laundries,
Plastic and Synthetic Materials Manufac-
turing, Pulp and Paper Mills, Rubber Pro-
cessing, Miscellaneous Chemicals, Ma-
chinery and Mechanical Products, Electro-
plating, Ore Mining and Dressing, Coal
Mining, Zinorganic Chemicals Manufactur-
ing, Textile Mills, Organic Chemicals
Manufacturing, Nonferrous Metals Manu-
facturing, Paving and Roofing Materials
and Paint and Ink Formulation and Print-
ing.
-The Settlement Agreement also man-
dates a schedule for the accomplishment
of the program. Industries are to be
studied in stages, over a period of years.
Technical studies for the several indus-
tries in the first group, or stage, were
initiated October 13, 1977, with a con-
tractor being required to the maximum
extent feasible, to complete their per-
formance within 12 months. Not later
than six months after contract completion,
a proposed regulation must be published
in the Federal Register. Final rules for
all 21 industries must be published not
later than December 31, 1979.
Appendix A of the Settlement Agree-
ment, lists 65 chemicals and chemical
classes generally believed to include the
most potentially hazardous substances com-
monly released into the environment. This
list of 65 chemicals and classes of chemi-
cals was refined by EPA into a list of 129
unambiguous specific "priority pollutants"
for the purpose of technical studies. The
Agency believes that the specific list ful-
fills both the requirements of the court
order and that these compounds, in most
cases, can be analytically measured. The
protocol then, applies to the organic com-
pounds included in the 129 unambiguous
priority pollutant list.
Sampling
The initial characterization (screen-
ing) of the varied industrial discharges
covered by this program will be made on an
analysis of a composite sample. Any scheme
for collecting a composite sample is, in
effect, a method for mechanically integrat-
ing to obtain average characteristics of a
discharge. During the screening phase, the
sample composite can be used to determine
the average characteristics which would be
representative of that discharge. Compos-
ite samples are those that are made up of
a series of aliquots of constant volume,
collected at regular time intervals in a
single container. The duration of compos-
iting will depend on the type of sample
being collected, the type of facility being
sampled and the time varying characteris-
tics of the discharge. For the analysis of
semi-volatile compounds by liquid-liquid
extraction GC/MS, a composite sample is
taken using an automatic, peristaltic pump
with a timer and a single glass compositing
jug. A minimum composite volume of 2-1/2
gallons is required. The sample must be
maintained at 4*C during collection, ship-
ment and storage. A teflon-lined cap is
used to seal the jug. From this composite
sample the following portions are obtained:
an acid, base/neutral, and pesticide frac-
tion along with associated blanks plus an
aliquot for metals analysis. In addition,
a field blank for the automatic sampler is
required, using blank water as free from
organic interferences as possible, through
the sampling system used for each sampling
point.
Field sampling for purgeable organ-
ics requires special consideration and
equipment. The sample container should be
a 45-ml screw cap vial fitted with a teflon-
faced silicon septum. The sample bottles
are filled to overflowing (by grabbing the
sample) and sealed teflon side down. Dup-
licate samples are required along with a
blank water sample, because of leakage
and because the measurement process is
sample destructive. All samples are label-
ed with waterproof labels.
105
-------
At each sampling point, additional
grab samples are taken for total cyanides
and total phenols. In both cases, a new
one-liter plastic bottle is used, sealed
and maintained at 4°C during transport and
storage.
For the purposes of this program, a
sample is defined as consisting of: a
composite field sample, a composite samp-
ler blank, duplicate grabs for purgeable
organics (VGA's), a VGA blank, a cyanide
and a phenol grab.
Organics by Purge and Trap - GC/MS
Here again, it should be stressed
that the protocol used for the Effluent
Guidelines Division was intended only for
those organic compounds of the 129 unambig-
uous priority pollutants, and only for a
qualitative and semi-quantitative determin-
ation of these compounds during the survey
phase of the industrial effluent study.
Of the volatile organic compounds,
acrolein and acryloritile are not done by
purge and trap. These two substances are
determined by direct aqueous injection
GC/MS. Direct aqueous injection GC/MS is
recommended for all compounds that exceed
1000 ug/1.
In general, the purge and trap method
is not suitable for the determination of
compounds eluting later than chlorobenzene.
The Tekmar Liquid Sample Concentrator
is recommended in this section. Using the
equipment settings set forth in the proto-
col, a sample is injected and purged for 12
minutes and the organics are sorbed on the
Tenax-silica gel trap. The sample is
heated and desorbed into the gas chromato-
graph. GC/MS data is collected as soon as
the GC/MS vacuum system has stabilized.
Quality control is obtained by the
analysis of blank samples. These are
carried through the field trips by the
sampling contractors and contained in
bottles just like the sample is. Preci-
sion is determined by dosing blank water
with the compounds selected as internal
standards—bromochloromethane, 2-bromo-l-
chloropropane and 1, 4-chlorobutane and
running replicate analyses. A table
listing the elution order of volatile
priority pollutants is given, along with
the characteristic ions.
\
Organics by Liquid-Liquid Extraction-GC/MS
The method described in this section
of the protocol applies to compounds which
are solvent extractable. After extraction,
a gas chromotographic-mass spectrometric
procedure is used except for pesticides.
The pesticide fraction is determined
initially by electron capture-gas chroma-
tography and qualitatively confirmed by
mass spectrometry.
A two-liter sample is taken from the
composite sample. The base-neutral extrac-
tion is done first at a pH of 11 or greater.
Methylene chloride is the extracting sol-
vent. A sample is dried and filtered
through sodium sulfate. The next step is
to concentrate the solvent by Kuderna-
Danish (K-D) evaporation (distillation)
fitted with a three-ball macro-snyder (then
micro-snyder) column down to a 1.0-ml vol-
umn. The internal standard (10 pi of
2 pg//Jl DIO-anthracene) is added.
To measure phenols, there is an acid
extraction done at pH two or less using the
base-neutral extracted water. The same
procedure is followed as in the base-
neutral fraction.
In any sample where an emulsion
forms, thus preventing an 85 percent re-
covery of solvent, the laboratory should
use a continuous liquid-liquid extractor.
A description of this procedure may be
found in the protocol.
The pesticide fraction is determined
using a standard method published in the
Federal Register, June 29, 1973. A one-
liter sample is used and determined by
EC/GC, as previously mentioned. Blank
extractions for all of these fractions are
also prepared and analyzed.
Each fraction is then separated and
eluted into the MS under given chromato-
graphic conditions. Relative retention
times (compared to hexachlorobenzene) are
listed in the protocol.
A performance criteria is stipulated.
In order to begin a run, an operator must
demonstrate the ability to chromotograph
benzidine at 40 ng for the base-neutral
fraction. One hundred ng of pentachloro-
phenol must be detected before the acid
extract run.
For the purposes of our program,
three conditions must be met in order to
indicate the presence of a compound by
GC/MS. First, the characteristic ions for
the compounds (given) must be found to
maximize in the same spectrum. Second, the
time at which the peak occurs must be with-
in a window of ± 1 minute for the retention
time of this compound. Finally, the ratios
of the three-peak heights must agree with
the relative intensities given in tables
in the protocol, within ±20 percent.
Quantification of compounds ident-
ified, is done by the internal standard
method using deuterated anthracene.
Asbestos
One of the substances included in
the original list of 65 from the Court
Settlement Agreement is asbestos. As it
turns out, this is a very strange material
to deal with. Several attempts have been
166
-------
made by EPA, with the cooperation of other
agencies, among them the United States
Geological Survey and the Bureau of Mines,
to define the term asbestos. Short of a
major research effort, this substance is
still a difficult one to define much less
analyze in industrial waste waters. It is
this Agency's position to recommend, for
the purposes of developing effluent guide-
lines, the use of the fniLLm4.na.nt) Jnte.tL4.rn
Pioce.du.ie. jo* Fibtiout /(tbtAtoA, using
Transmission Electron Microscopy, by C. H.
Anderson and J. MacArthur Long, U.S. EPA,
Environmental Research Laboratory, Athens,
Georgia.
General Overview and Status
The protocol, SampLLng and
P*ocedu*e4 ton Sc4.eeru.rtg oj
Ejjluen** jo* PH4.oi4.ty PoLtut&nt*, in its
present form, has been in use for almost
one year and since the draft version, this
protocol has been used for over a year and
a half. Many analytical contract labs have
worked with the protocol for EPA. More-
over, several industries under study have
their own laboratories which are also
scrutinizing the applicability of this
protocol, along with our EPA laboratories.
In short, those methods of analysis recom-
mended by this Agency are being tested in
actual field situations and passing the
tests!
On November 9-10, 1977, a seminar
was conducted by EPA on the subject of
analytical methods for priority pollutants.
Participants at the seminar included the
many contractors doing sampling and analy-
sis for the Effluent Guidelines Division,
as well as industrial trade representatives
and other interested scientists. The pur-
pose of the meeting was to discuss issues
of concern and in general, to compare notes
on the various experiences of laboratories
working with the methods. The overall at-
mosphere at the seminar was one of refin-
ing, or fine-tuning the analytical pro-
cedures. Basically, the protocol has been
accepted as a reasonable, logical, workable
way of uniformly analyzing waste waters
from 21 industries for 129 substances.
Many of the major issues discussed
at that meeting in Denver, were resolved
or acknowledged for the record. One of the
topics of discussion concerned the use of
internal GC/MS standards provided by EPA.
Some very volatile compounds were being
lost. Therefore, it was recommended that
individual laboratories prepare their own
internal standards for vinyl chloride and
methylene chloride. Another useful pro-
cedure to alleviate the problem of a
water build-up in the tenax trap, is to
increase the number of times the silica
gel is changed.
On the use of liquid-liquid extrac-
tions for semi-volatile compounds, emphasis
was placed on the need to use continuous
extractors when emulsions are formed.
Another procedure to aid in breaking emul-
sions would be to do the filtration through
glass wool packs. In general, it should be
kept in mind that an 85 percent solvent re-
covery should be obtained.
In the testing process, a flaw has
been found. It now appears that there is
a problem in separating benz(a) anthracene
+ chrysene and anthracene + phenanthrene
and benzo(b) fluoranthene + benzo(k) fluor-
anthene on the recommended GC column.
While these substances might be separated
on different columns, the columns selected
for use are still considered the best when
one is limited to three GC columns for all
industrial waste water types during the
screening phase. GC/UV or liquid chroma-
tography may resolve the problem during
the verification phase.
Further study to improve analytical
methods is ongoing. Several contracts have
been let through EPA, Cincinnati, on the
development and application of test pro-
cedures for the specific compounds of con-
cern. The Agency's list of priority pol-
lutants have been divided into groups for
this research effort. Not only will these
contracts aid in the development and re-
finement of analytical procedures, they
will also use actual field samples in order
to evaluate their recommended methods. Any
changes which develop out of these con-
tracts shall be used in future revisions of
the protocol.
At the present time, it is the posi-
tion of the Effluent Guidelines Division
of EPA to endorse this protocol for use in
preparing regulations. The tests of time
and use appear to be passed. It now re-
mains for these current state-of-the-art
methods to become formally accepted in the
near future, by publication in the Federal
Register.
107
-------
ALTERNATIVE LEVEL I ANALYSIS METHODS
by
Karl J. Bombaugh
Radian Corporation
Austin, Texas
Abstract
This presentation includes a descrip-
tion of some modifications and additions
to the Level I Analysis Scheme that were
used to characterize the emissions from
coal gasification plants. While some of
these modifications go beyond the intent
of Level I analyses, the experience gained
from their use should be of value to
others performing environmental analyses.
Introduction
The word alternative may convey the
concept of replacement; i.e. of using
Method B instead of A. Such is not the
intent of this presentation. Its purpose
is to present some recommendations based
on our experience with process measurement
analyses as they apply to commercial coal
gasification facilities. These recommen-
dations deal with the following:
1. Proposed additions to the EPA
Level I Scheme for environmental
assessment.(1)
2. Proposed corrections of the Level
I Scheme.
3. Considerations for more detailed
analyses.
The recommendations described here
are based on the analytical results that
will be discussed in this presentation.
These results were obtained primarily from
an analytical screening study of selected
samples from operating gasification
units. "' Some results were obtained
from a Level I pilot study which is still
in progress.
Consideration will be given to the
extraction procedures for organics, to
IR interpretation methods, and to the
potential benefits of UV analyses.
Additionally, an alternative scheme for
organic analyses will be discussed. First
consideration will be given to proposing
modifications to the Level I procedure.
Proposed Modifications to
the Level I Procedure
Before entering into a discussion of
the proposed modifications to the Level I
procedures, some background information
will be presented to 'acquaint the reader
with the problems that led to these
proposals.
Background Information
The work on which this presentation
is based was done as a part of an EPA spon-
sored program (EPA Contract No. 68-02-2147)
whose overall objective is to provide a
comprehensive environmental assessment of
Low BTU coal gasification and utilization
processes.
Grab samples from several different
commercial scale gasifiers and one pilot
scale gasifier were obtained for this
study. The purpose of this study was to
gain insight into the nature of the samples
that will be encountered in the assessment
program. A second objective was to gain
some concept of the differences between
various plants and processes. The samples
were obtained during inspection trips
which were made to develop an environmental
test program. The analyses performed were
intended to identify, in a very preliminary
manner, the classes of compounds present
in each sample which may warrant further
attention. A significant part of the
analyses followed, in principal, the
methods defined by the EPA procedures
manual for a Level I Environmental
Assessment. (*•'
The screening study provided a signif-
icant amount of information that was useful
to the test program(2) and a number of
conclusions were drawn from the results.
The conclusions related to:
(a) The classes of compounds and con-
centrations of components found
in specific samples,
(b) The effects of feed stock and
process operating condition
changes upon levels of key com-
ponents detected in these
samples, and
(c) Modifications/additions to pro-
cedures which will help insure
that a maximum amount of environ-
mental information is obtained
from the program.
This presentation shall deal primarily
with matters relating to topic (c).
108
-------
Extraction of Organics
The extraction procedure defined by
the procedures manual for a Level I
assessment'*•' should be expanded to include
more polar extracting solvents. The basis
for this need is illustrated by the data in
Table 1.
TABLE 1. ORGANICS EXTRACTED FROM A
GASIFICATION PROCESS LIQUOR
Solvent
MeCl
Ether
Acid-Ether
Sample m .„,.
Extractant
g/101
28
38
9
(n) no.
Z
37
51
12
of extractions • 3
From these results it is evident that
less than half of the organics are being
extracted by the methylene chloride at a
10:1 sample to solvent volume ratio. One
might conclude that this error is within
the ±2 to 3 accuracy specification sought
by the Level I objectives. However, it
must be recognized that the materials ex-
tracted by the ether from the acidified
sample might be totally different from
those extracted from the neutral sample
with methylene chloride. These conditions
can be recognized by comparing the infrared
spectra of the extract-residue as illu-
strated in Figures 1 and 2.
The top spectrum shown in Figure 1 was
obtained from a methylene chloride extrac-
tion of quench liquor from one test site
(Site A); the midale was obtained from an
ether extraction and'the bottom from an
extraction with ether after the sample had
been acidified to pH - 1. In this
experiment, the extractions were carried
out sequentially on the same sample. The
numerous differences in these spectra con-
firm that the three extracts contain widely
different materials. For example, the
carbonyl band in the 1700-1740 cm-1 region
is absent from the first two extracts and
the three bands in the 800 cm~l region
indicative of sulfones are absent from
the methylene chloride extract.
EXTRACT OF GASIFIER LIQUOR
SITE A
MeCt2
2. ETHER
3. ACIDIFIED + ETHER
AWNUMM* ioir
Figure
ijoo ' i«'oo i4rfo w»viMOM»imcy:h»Ao «io
1. Infrared spectra of three successive
extracts of gasifier liquor from Site A.
(Sample: Extractant - 1:1. n - 4)
JET
»oo
109
-------
The extracts of a comparable stream
from another site were compared. Their
spectra are shown in Figure 2. Although
the spectra of these extracts were quite
different from the Site A extracts, the
result was similar; i.e., the ether and
acid-ether extracts were indeed different
from the methylene chloride extract and
different from each other. These differ-
ences indicate that methylene chloride
does not extract all organic substances
from water.
Additional attention should be direc-
ted to the sample to solvent ratio. The
Level I procedure recommends a 20:1 sample-
solvent ratio as a means of concentrating
organics into the extracting solvent. This
approach is valid only for materials having
a distribution coefficient of perhaps five
or greater; that is, a substance which
has five times the preference for methylene
chloride that it has for water. Many
polar organics fall outside this limit.
The extractions performed for the
screening study were made with a 1:1
sample-solvent ratio and a 4-step
extraction. The results are shown in
Table 2.
TABLE 2. ORGANICS EXTRACTED FROM SITE A
QUENCH LIQUOR (1:1)
Step ng/ft
1 Methylene Chloride 2231
2 Ether Extractable* 1457
at pH 7
3 Ether Extractableb 177
at pH 1
TOTAL EXTRACTED 3865
. Following Step 1
Following Steps 1 and 2 n - 4
Z of Total
Extracted
57.7
37.6
4.7
100.0
Equal volumes were used because the
sample was rich in organics and only about
500 ml of sample was available. However,
it is evident that by using a 1:1 volume
ratio, a larger percentage of the organics
were extracted by the methylene chloride.
Attention is called to the fact that the
spectra in Figures 1 and 2 are from extracts
that were obtained with a 1:1 volume ratio
EXTRACT OF GASIRER LIQUOR - SITE C
M»CI2
2. ETHER
3. ACIDIFIED-I-ETHER
4000 NAVINUMMJI ICtT'l
IMO
I4bo NAVMMMH ICtT*» too
•00
400
Figure 2.
Infrared spectra of three successive
extracts of gasifier liquor from Site C.
(Sample: Extractant • 1:1, n - 4)
110
-------
using a 4-step extraction and that these
conditions favor a more complete extraction
by the methylene chloride. The differences
between the methylene chloride extract and
the ether extracts could be even greater
at the 20 to 1 ratio.
To gain insight into the effect of
volume ratios, use can be made of some
basic relationships of extraction theory.
Equation (1) defines the fraction of sol-
vent transferred (F) in terms of relative
extractant volume and distribution
coefficient
F -
(1)
where:
V - volume of extractant. and
fc ' volume of sample.
In order to relate the fraction trans-
ferred in a single extraction to the total
mass transferred by a sequence (Zn) use
can be made of the Remainder Theorem de-
fined in Equation (2):
- F)n
(2)
Figures 3 and 4 depict these relationships
graphically. Figure 3 shows the relation-
ship between E,j and F for a 3-step extrac-
tion specified by the Level I procedure.
Figure 4 shows the relationship between
KD and F at three different volume ratios:
20:1 Specified by the Level I
Manual.
10:1 Used in the Pilot Study.
1:1 Used in the Screening Study.
Note that in order to extract 90Z of
a component in three steps (the point is
designated by an arrow in Figure 3),
approximately 50Z must be extracted per
step. In order to achieve a 507L extraction.
the solute must have K values as follows:
20:1
10:1
1:1
20
10
1
In order to comply with the Level I
objective of ±2, the 3-fold extraction must
remove 507.. requiring an F of 0.2 and KD
values as follows:
20:1
10:1
1:1
5
1
0.5
In order to be assured that polar
species are detected in screening studies
and in subsequent analyses, serious con-
sideration should be given to these funda-
mentals. This work indicates that more
favorable volume ratios and a sequence of
solvents are needed to characterize the
effluents from Low BTU coal gasification
processes. However, these principles are
applicable to all environmental tests
where extractions are used to isolate
pollutants .
Infrared
In this section, attention is focused
on a problem in the quantitative assessment
of infrared spectra for Level I organic
analyses. The "Suggested Format for Level
I Organic Analyses Data" (3) recommends that
spectral intensity be reported relative to
the strongest band on a transmittance
basis. The practice is undesirable because
the results can be highly misleading. The
problem is illustrated in Figure 5.
The diagram in Figure 5 shows two
hypothetical bands of apparently equal
intensity; i.e., they both show 457. trans-
mission relative to their base line. When
the transmission is converted to absorbance ,
the lower band is 3.5 times as strong as
the upper band. The reason for this is as
follows :
0.45 - (0.95 - 0.50) * Log - - Log
0.45 - (0.50 - 0.05) j« Log
- Log
- 2.8
- 1.0
The fallacy is further illustrated
by Figure 6 which shows two spectra of a
single sample run at two base line levels.
The instrument span was set with no sample
in the beam. The sample was inserted and
the spectrum run. Then the optical attenu-
ation was adjusted to raise the base line
and the spectrum was rerun. In this
figure the absorbances (A) and the trans-
mittance ratios of the top and bottom
curve are each equal to each other and
their ratios are therefore equal to unity,
i.e. :
(*o to? A
I /
top
but:
T
r~
o bottom
- V top
bottom
- 1
- 1.7
(T - T,,)
o' bottom
This result illustrates that it is
fallacious to use T as a measure of band
intensity. To avoid this fallacy, bands
must be determined as either absorbances
or as intensity ratios relative to the
base line.
Clarification is also needed for the
measurement of shoulder bands. The prac-
tice used for this work was to measure all
111
-------
10
.1 .2 .3 .4 .5 .6 .7 .8 .9
Figure 3. Plot of remainder theorem - relationship between
fraction of mass transferred per extraction (F)
and total mass transferred on three extractions (E3).
20/1
10/1
1/1
0.1
0.2
0.3
0.4
0.6
0.6
Figure 4. Relationships between fraction of mass
transferred (F) and distribution coefficient
(KD) at three different phase ratios.
112
-------
tee.
%T
45
0.28
. _ RATIO - 3.5
45
1.00
Figure 5. Effect of transmission level on band intensity
Theoretical band of equal transmittance.
Note: 3.5 times as much absorbance is required
to produce the band at the lower transmission level.
Figure 6. Effect of transmission level on band intensity -
Actual spectra of equal absorbance (Atop - Abottom'
Both spectra were obtained with the same sample,
only the reference beam energy was attenuated.
113
-------
bands from the base line rather than
tangentially. This practice is "fail safe1'
since it gives maximum values; however,
it may produce false levels and thereby
contribute to misdirected effort.
Benefits Offered by UV Spectrometry
for Level I Analyses
Ultraviolet absorption spectrometry
is a highly sensitive technique for poly-
cyclic hydrocarbons and it is a simple
technique to perform. The LC fractions can
be scanned in minimal time. The UV absor-
bance of polycyclic aromatic components is
very strong. For example, materials such
as anthracene, shown in Figure 7. exhibit
molar absorptivities of about 1CP which
enables their detection at the nanogram
level. An additional advantage is afforded
by the principle of the Bathochromic shift
that accompanies the increase in the num-
ber of fused rings. This property permits
polynuclear components to be detected in
the presence of other aromatic compounds.
It should be practical therefore to use UV
spectrometry to screen samples and LC
effluents (particularly the early fractions)
for evidence of polycyclics in the presence
of other aromatic compounds.
The spectra in Figures 8, 9 and 10
illustrate this potential. Fraction LC-2
of tar in Figure 8 shows distinct peaks at
about 320 and 340 nm which is a clear indi-
cation of polynuclear substances.
Similarly, LC-3 of tar in Figure 9 shows
a peak near 330 nm plus a sharp maximum of
250 nm, with a significantly lower absor-
bance at 220 nm. This is clear evidence of
polynuclear aromatics. In contrast, LC-5
from gasifier condensate in Figure 10 shows
minimal absorbance above 300 nm. The curve
shape is characteristic of substituted
aromatic compounds such as phenols etc.
UV spectrometry therefore provides a po-
tentially useful, cost effective method of
screening effluents or effluent extracts
for polynuclear hydrocarbons. Its use is
recommended.
Extensions to the Level I
Procedure for Organic Analysis
Many streams examined by the Level I
procedure during this study were observed
to contain large amounts of organics.
However, the presence of organics could be
established by various means other than
the Level I analyses. For example, use
could be made of:
• Process Flow Sheets,
• Engineering Analysis of the Process,
• Visual Observation, and/or
• Olfactory Sampling (Smelling).
Samples such as tar were, by their
very nature, polyaromatic and therefore
required no screening. (During the study,
tar was analyzed by the Level I method to
gain information about its characteristics
as defined by the LC-IR analysis and to
gain information about trace elements.)
By definition , Level I analyses could be
waived in favor of a more detailed analyses.
Since one major objective of the Low BTU
program is to obtain data for a control
technology assessment, more specific type-
analyses were needed. The decision was
made, therefore, to conduct a species-
specific analysis-program in conjunction
with the Level I pilot study, using essen-
tially the same samples. This work is not
yet completed. However, the methods used
in this study were developed and tested
previously .and will therefore be
described. (5-7)
The essence of this method is a
preseparation-derivatization scheme that
is designed to separate materials into
classes and make them compatible with the
GC Mass Spectrometer. This approach is
well founded, in view of the resolving
power of the GC-MS and the searching power
of the data banks. However, even without
the support of the GC-MS, the LC separation
scheme is a useful tool for characterizing
organics .
Preseoaratign by Liquid-Liquid
Extraction (5- 7)
The preseparation scheme illustrated
in Figure 11 is based on liquid-liquid
extraction. It separates the sample into
four principal fractions :
• Neutral Lypophylic substances ,
• Organic acids whose salts partition
into water at high pH,
• Organic bases whose salts partition
into water at low pH, and
• Neutral Hydrophylic substances .
The neutral substances are further
separated by column chromatography on
into four fractions :
• Paraffins ,
• Aromatics ,
• Polar Neutrals , and
• V Polar Neutrals.
The acid fraction is methylated in two
steps to convert phenols to ethers and the
acids to esters.
A total of eight fractions is obtained,
each containing a distinct class of com-
pounds defined by their solubility and
polarity. Each fraction is easily sepa-
rated by gas chromatography since optimal
columns can be selected for each type of
material. Tailing is minimized. Similarly,
114
-------
5.0
4.0-
o
o
3.0
2.0
ANTHRACENE
I
300
200 250 300 350 400
WAVELENGTH
-------
LC-3 OF TAR
200
aso
300
360
Figure 9. Ultraviolet spectrum of chromatographic
fraction 3 from gasifier tar.
1.0
LC-5 OF GASIFIER CONDENSATE - SITE A
200
260
300
360
Figure 10. Ultraviolet spectrum of chromatographic
fraction 5 of gasifier condensate.
116
-------
non-volatile substances can be taken as
needed for separation by HPLC using an
optimal column.
Recovery Studies
This analysis scheme was used to
characterize a condensate from a fuel syn-
thesis process. (') Recovery studies were
completed as a part of that investigation.
Some typical results are included here.
Figure 11 shows the recovery of aro-
matics. A 78% recovery was realized.
Significantly only 1.57, of the material
was found outside of the intended fraction.
Figure 12 shows the recovery of acids and
phenols which were methylated by dimethyl
sulfate followed by methylation with
diazomethane. A 78% recovery was also
realized with the phenols. Only 3% was
found outside of the intended fraction.
The acids were less "selective" as is
seen in Figure 13. In this case. 247. were
methylated with dimethyl sulfate and the
remaining 50% were methylated with
diazomethane. Total recovery was 84%.
No material was found outside of these
fractions.
Methylation of phenols by DMS is by
no means stoichiometric. However, yields
are reasonable and improvement is believed
to be achievable. The data in Table 3
shows some results with known compounds.
Early work reported in the literature shows
that much higher conversions can be
achieved with optimized conditions.
Summary
The ABN separation provides a
versatile method for characterizing
organics. It can be used for both bulk
characterization and detailed analyses by
GC-MS and/or other ancillary techniques.
The merit of a liquid-liquid separa-
tion cannot be overstated. As a bulk to
bulk separation between discrete compounds
(e.g., water and solvent) it is highly
reproducible. Uncertainties associated
with surface properties (adsorption) are
eliminated. . As a bulk separation, it
enables large quantities of sample and of
solute to be handled easily. Acids and
bases of known concentration can be used
to influence the solubility (i.e., the
partition coefficients) of the sample,
making it possible to handle a wide range
of substances with the same system.
The principle disadvantage is that
only a single stage of separation is
practicable, however, this limitation is
readily compensated by the resolving power
of the GC-MS.
References
(1) Hamersma, J. W., S. L. Reynolds and
R. F. Maddalone, IERL-RTP Procedures
Manual: Level 1 Environmental
Assessment"! EPA-600/2-76-160a, EPA
Contract No. 68-02-1412, Task 18.
Re dondo Beach, CA, TRW Systems Group,
June 1976.
(2) Bombaugh, Karl J., Analyses of Grab
Samples from Fixed-Bed Coal Gasifica-
tion Processes.EPA Contract No.
68-02-2147, EPA 600/7-77-141. Dec.
1977. Exhibit A, Program Element
No. EHE623A.
(3) Harris, Judith C., Suggested Report
Format for Level 1 Organic Analysis
Pita, ADL Report No. 79347-16-4. EPA
Contract No. 68-02-2150. Oct. 21,1977.
(4) Friedal, R. A., and M. Orchin. Ultra-
violet Spectra of Organic Compounds.
John Wiley & Sons, Inc., New York
(1953).
(5)
(6)
(7)
Bombaugh, K., E. Cavanaugh,
J. Dickerman, S. Keil, T. Nelson,
M. Cowen, and D. Rosebrook, Sampling
and Analytical Strategies for Com-
ounds in Petroleum Refinery Streams,
bl. I and II. EPA Contract No.
68-02-1882, U.S. Environmental
Protection Agency, Research Triangle
Park. N.C., 1975.
E!
V\
Bombaugh, Karl J., Atly Jefcoat, and
E. Cavanaugh, A Systematic Approach
to the Problem of Characterizing
the Emission Potential of Energy Con-
version Processes .A.I.Ch.E. SvmpoT-
ium Series No. 165 73 267 (1977)
American Institute oT Chemical
Engineers. New York, N.Y.
Oldham, Ronald G., and Robert G.
Wetherold, Assessment, Selection and
Development of Procedures for Deter~
mining the Environmental Acceptability
of Synthetic Fuel Plants Based on Coal.
Volume III. Results of Investigation.
ERDA Contract No. E(49-18)-17957
Radian Project No. 200-108. Austin,
TX, Radian Corp., August 1976.
117
-------
TABLE 3. METHYLATION OF PHENOL MIXTURE USING THE DIMETHYL
SULFATE PROCEDURE AS DESCRIBED BY WEBB (WE-158)(7)
Compound
Phenol
o-Cresol
m-Cresol
p-Cresol
2.6-Xylenol
2.4-Xylenol
2,5-Xylenol
3,4-Xylenol
3,5-Xylenol
Hydroquinone
1-Naphthol
2-Naphthol
Pyrogallol
TOTAL
Amount Added, mg
204
257
258 4?9
221
122
102 207
105
126
102
100
106 206
100
100
2008
Amount Found,
37
199
412
91
124
103
96
47
264
0
1286
mg 7, Recovery
18
77
86
75
60
81
94
47
129
0
Average 64
COLUMN
CHROUATOORAPHV
ON SILICA Of I
METHTLATC HUN
M*t SO*. EXTRACT
WITH ETHER
—j MOM POLAR NEUTRALS I ND
H MOO. POLAR NEUTRALS | 78%
\ J
L — J POLAR NEUTRALS I 1 « 5 A
I I .
I ^~ ' Mn
— — ) VERT POLAR NEUTRALS I
.J
r 1
1 ETHERS Of |
I
I CARSOXVUC ACID* I- _
I J
ACKNFV.
EXTRACT
USTHVLATE
WTTHCHtM,
I T
I UtTHVL ESTERS .
1 or CARBoxruc
I DISCARD |
U I
| SAStC AND WATER
I SOLURLE
Figure 11. Acid-base-neutral (ABN)
separation scheme for
organlcs - Numbers at
each step indicate the
Z recovery of a known
aromatic compound.
•AStfV.
EXTRACT WITH
ETHER
•ASIC COMPOUNDS I
I
I
| WATER SOLUSLE |_
I COMPOUNDS I
ACtOWV AND
EXTRACT WITH
ETHER
I
POLAR WATER ,
•OLUEtLC COMPOUNDS
VERT POLAR WATER
118
-------
ETHYL ESTERS
1 Of CAXOOITLIC
ACKHTT WITH Ht*O4,
HTHACT WITH f THtM
. -<»
J'«
| *AWC AHO WAT1R ,
| •OLUBLI COMPOUND* '
Figure 12. ANB separation scheme - Numbers at each
step Indicate the recovery of a known
phenolic compound.
119
-------
COLUMN
CHNOMATOaKAPHV
ON *IUCA OEL
NON POLAR NEUTRAL!
I 1
H MOO. POLAR NEUTRAL* |
I J POLAR NEUTRAL*
I I
MSTHYLAIE WITH
M»t*O4. EXTRACT
tMTH ETHER
— —I »EHY POLAR NCUTHAL* I
I J
I 1
_, f THCMS Of | 24%
, mtNOL* I
I
I CAMOXVUC ACIOI I- •
I _______ J
ACKMTY.
EITMC1
UiTHYLATI
•rrm CMtMf
I
| METHYL
1 Of CAIWOXVLIC
!__!"!!_ J
160%
I
I OWCARD |
U j
SAWC AMD WATEM
totmit COMPOUND*
I
I
EXTRACT WITH
ETHER
1 *A*IC COMPOUND*
I WATEN (OLUVLS (_
I COMPOUND* |
EITMACT WITH
ETHEK
I
I
J POLM WATM i
;_ lOUI«tE_CO«IPOIINni_,
_{~ VERY POLAR WATER .
| IOUMU.E COMPOUND* J
Figure 13. ABN separation scheme - Numbers at each
step indicate the recovery of a known
organic acid.
120
-------
SYNTHETIC FUELS PRODUCTION: ANALYSIS OF PROCESS BY-PRODUCTS FROM A LABORATORY SCALE COAL GASIFIER
C. M. Sparactno, R. A. Zweidlnger, S. Willis, and D. Minick
Chemistry and Life Sciences Division, Research Triangle Institute
P. 0. Box 12194, Research Triangle Park, North Carolina 27709
Introduction
A research program is currently in progress
at the Research Triangle Institute to investigate
the particular pollution problems associated with
coal gasification through the construction and
operation of a laboratory scale gaslfier. This
program is funded by EPA/RTP and is moving into
its second year. The major goal of the project
is the assessment and analysis of trace pollu-
tants possibly associated with the coal gasifica-
tion process. Such efforts are needed because of
the known hazard potential associated with cer-
tain materials that are likely to be introduced
into the environment from commercial gasifiers.
Coal screening studies are being conducted to
determine the relationship between various U.S.
coals and their pollutant potential. Another
Important feature of the program is the utiliza-
tion of the reactor for pollutant control through
parametric variation, ,l..e., the manipulation of
operational variables to effect environmental
control. A third phase of the project entails a
study of gasification kinetics to discern pos-
sible relationships between product formation
probabilities and process parameters.
The gasifier system has been operational for
several months after having undergone several
modifications to both the reactor and the samp-
ling train. A schematic of the entire system is
depicted in Figure 1. The overall direction of
the gasification program, and the operation of
the reactor is the responsibility of the Process
Engineering Department of the Research Triangle
Institute; the analysis of fixed gases and Ci to
C« hydrocarbons is accomplished through the En-
vironmental Measurements Department. The analy-
sis of the product gas stream for volatile com-
ponents, and the analysis of the non-volatile
tars and waters are carried out by the Chemistry
and Life Sciences Department and these studies
form the basis for this presentation.
Analysis
The materials constituting the gasifier by-
products are' exceedingly complex and represent a
formidable task for analysis both qualitatively
and quantitatively. Because of the very large
numbers of components associated with each sample
type, many of which are chemically quite similar,
the methodology utilized must possess high resol-
ving power, exhibit a high degree of specificity
and be reasonably sensitive. The only tool cur-
rently available that provides this combination
of features is gc-ms-comp. Use of capillary
column technology permits the direct analysis of
many samples without the need for extensive
prefractlonation. This provides an advantage in
terms of cost effectiveness that is unrivaled in
current analytical practice.
The laboratory gasifier produces two types
of samples which are collected and treated by
somewhat different procedures. Materials that are
carried along with the gas stream after passage
through the tar/water trap are termed "volatlles",
and are collected by entrapment on polymeric
sorbents. Those substances that are knocked out
by the tar/ water trap are termed "non-volatile",
and consist of the coal tars and condensate water.
Each of these samples will be discussed in terms
of qualitative analysis followed by a separate
discussion regarding quantitative aspects of their
analysis.
Volatiles
As can be seen from Figure 1, the volatile
materials are removed from the gas stream from two
types of sorbents XAD-2 and Tenax GC. The former
Is used as a means of collecting components during
the full course of a gasification run; the latter
is utilized as a "grab" device so that samples may
be selected at discrete times. Methodology for
the analysis of the components adsorbed onto the
trap material was borrowed from previous work car-
ried out in our laboratories involving the char-
acterization of gaseous contaminants from air.(l)
The method involves the thermal desorptlon of the
adsorbed components from the polymer (Tenax GC)
onto o cooled (LNz) capillary trap. The trap is
then rapidly heated to sweep the contents onto a
temperature programmed capillary GC system which
is Interfaced to a continuously scanning mass
spectrometer. Mass spectra are thus acquired for
each component which, after interpretation pro-
vides the Identity of each compound. The same
characterization process is applied to compounds
adsorbed onto the XAD-2 resin, except here the
polymer is too unstable for thermal desorptlon.
Instead, solvent extraction (Soxhlet) is used
followed by concentration of the extracts and
injection onto the gc-ms.
The results of a typical gasification run are
depicted in Tables 1 and 2 for volatlles collected
onto Tenax and XAD-2. Total ion current (TIC)
plots (Figures 2 and 3) depict the elution pattern
from the capillary GC system. The identifications
of individual components corresponding to the
various peaks on the TIC plot are shown on the
appropriate list (Tables 1 and 2). Some two hun-
dred compounds are thus collected from the gas
stream by both the Tenax and XAD-2. For the most
part the materials collected by each sorbent are
the same although the amounts collected may dif-
fer. This is a reflection of the different
selectlvities of the two polymers towards various
organic species, as well as the effects of varying
breakthrough volumes due to the presence of other
organics, some of which, for example benzene, are
present in relatively large amounts. The major
materials Identified Include saturated and unsatu-
rated hydrocarbons, alkylated aromatics, alkylated
phenols, and furan and thiophene derivatives.
Alkanes from C? to Cn are present as are lower
condensed ring aromatics upto anthracene. These
are the limits on materials found in the gas
121
-------
stream (the gc conditions used for this work are
known to elute much higher molecular weight
species).
Non-volatiles
The non-volatile materials, _!•.£•, those
materials collected in the tar/water trap, repre-
sent an exceedingly complex sample for which
total methodology is by no means fully developed.
Treatment of the condensed waters is reasonably
straightforward and consists of solvent extrac-
tion, chemical derivatization, and analysis by
gc-ms. Using methylene chloride as the extrac-
tant, approximately 0.5Z (wt./wt.) of the water
is removed as organic extractables. This ma-
terial contains a wide range of compound types
with phenolics comprising a large majority. The
derivatization is incorporated as a means of con-
verting carboxylic acids (found in very low pro-
portions) and aromatic hydroxyl species to their
corresponding methyl derivatives. This vastly
improves their gc characteristics and thus eases
the mass spectrometric identification and quan-
tification process. Diazomethane is used to
convert acids to methyl esters while dimethyl
sulfate is preferred for methylatlon of pheno-
lics. The latter process is fraught with some
difficulty in obtaining good yields of deriva-
tive; recent efforts in our laboratories have
shown that phenol can be converted quantitatively
to anlsole using high concentrations of base
(potassium hydroxide) during this process. The
concentrated solution of derivatized material is
then Injected onto the capillary gc-ms system
(Carbowax 20M) and the effluents scanned as
described above. Results of this procedure are
shown in Table 3 and Figure 4. This data was
collected before the completion of the above men-
tioned dimethylsulfate optimization procedure and
hence shows underlvatized phenols. As expected,
the fraction is relatively rich in those com-
ponents that possess greater water solubility,
.!..§., phenols, with lesser amounts of other ma-
terials that have been scrubbed from the tars.
The tar fraction represents the most chal-
lenging of the samples in terms of analytical
treatment. The sheer number of molecular types
ranging from highly polar oxygenated species to
paraffinic hydrocarbons, and from simple low
molecular weight molecules to polymeric material
dictates the need for some form of pretreatment
before final analysis. We utilize a solvent
partition scheme which reduces the tar sample to
5 fractions each of which is enriched in com-
pounds of a similar chemical nature. Each frac-
tion is then either analyzed by the gc-ms or
further fractionated using hplc. The solvent
partition procedure is a modification of Novotny's
method^2' and is depicted schematically in Figure
5. This procedure has been validated through the
testing of standard mixtures and by the use of
radiolabeled material.
The acid fraction as isolated from the tar
sample is very similar in composition to the
water extracts described above. Treatment is
therefore identical to the latter sample, .!..§_.,
derivatization and gc-ms. The organic base
fraction is ammenable to direct gc-ms analysis.
This fraction is almost entirely composed of
heterocycllc amines derived from three aromatic
classes; pyrldines, Indoles, and quinolines.
Some higher condensed ring systems are represen-
ted as shown in Table 4.
The neutral fraction Is conveniently separa-
ted by partition into three additional classes,
non-polar, (mostly paraffins), PNA's and polar.
The latter has not yet been actively investi-
gated. The paraffins have likewise remained
unexamined because they are, for the most part,
of low environmental hazard potential. Of most
interest is the PNA fraction both as a class and
individually. Their analysis at this point
consists of a direct gc-ms approach. With effi-
cient capillary systems,.sufficient resolution of
this complex fraction can be achieved with up to
five ring molecules eluting within a reasonable
time. Results of a typical run are shown in
Figure 7 and Table 5. Many of these compounds
are of course cancer suspect agents and as such
warrant careful examination with regard to the
amounts generated during a gasification run. For
this reason the quantitative aspects of our work
have been directed initially toward the PNA's.
Quantitatlon
Once the identification process has been
completed for a given sample type, the determina-
tion of actual quantities of specific compounds
or groups of compounds can be undertaken. This
process i.e_. the quantitation by gc-ms of a rela-
tively few components from a highly complex syn-
fuels sample has not been heavily reported in the
open literature. It is not possible therefore to
draw on pertinent data produced and published by
other workers. The generation of the information
necessary for these general quantitation pro-
cedures is a consequence only slowly obtained via
a step-by-step process that is carried out for
each gc-ms system being utilized.
The quantitation process has been described
in a recent publication.O) The method basically
Involves selected ion monitoring of specific ions
of the compounds In question, and comparing the
ion Intensities of those ion's with a charac-
teristic ion from a carefully chosen internal
standard. (The use of peak areas from a gc trace
is not feasible due to the complexity of the
sample). Using a predetermined response factor,
the amount of various components can then be
ascertained. The response factor is determined
on a molar basis and is reported as a relative
molar response (RMR) which is defined as follows:
RMR
Aun/Boleaun
'unknown/standard A ./moles .
std std
(1)
where A - peak area obtained from selected ion
plots. Thus to determine the RMR, a concentra-
tion of unknown and standard must be accurately
known and the system response must be measured.
Values are determined in replicate. From equa-
tion (1) then:
A -GMW «g .
on un std
gun " A
std ""'std'^un/std
(2)
where GMW - gram molecular weight of compound.
The choice of Internal standard will depend on
the type of compound being analyzed and should
display suitable retention and fragmentation
-------
characteristics as well as favorable ionlzation
cross-section. We customarily use perfluorinated
or perdeuterated materials, .e.jj. decadeuteroanth-
racene for PNA samples.
Several PNA hydrocarbons were chosen as re-
presentative of compounds that have been found In
coal tar mixtures previously. These materials
were then used to calculate RMR's based on the
use of decadeuteroanthracene as internal standard.
Both a quadrupole and a magnetic sector instru-
ment were used for this study so as to provide a
basis for comparing data from two gc-ms systems.
Both Instruments employed 2SM SCOT columns con-
taining OV-101. Data from each system was ob-
tained in a multiple ion detection (MID) mode.
Thus a mixture containing known amounts of the
chosen FNA's and internal standard was injected
onto the gc column and the effluent was moni-
tored by continuous ms scan. Peak areas cor-
responding to the parent ions of all components
were then generated from MID plots and ratioed to
the internal standard area. From equation (1)
RMR's were then calculated and are shown below
in Table 6.
It would seem that the response factors are
highly dependent on the particular gc-ms system
utilized as there is no discernable correlation
between the two sets of data. These factors were
then used to calculate the amounts of those
chosen PNA'a present in the PNA fraction col-
lected from a recent coal gasification run. The
MID plots thus generated included several compo-
nents with the same parent ion as the standard
mixture possessed. Selection of the peak cor-
responding to each of the standards used was made
on the basis of relative retention time. With the
peak areas of the compounds selected and knowing
the molecular weights and area of each compound
and the Internal standard, the amount of each
substance was calculated from equation (2). The
results are shown in Table 7.
The agreement between figures from the two
systems, while not completely gratifying, is with-
in expectations for the initial applications of
the procedure. A crucial experiment yet to be
carried out involves the analysis of a sample
mixture containing known amounts of materials.
This will provide an assessment of the validity
of the method and yield information on the magni-
tude of the errors Involved.
The use of RMR's is not restricted to direct
application o'f the type outlined above. More
expeditious use of the data would be desirable
when large numbers of compounds are to be quantl-
tated. From Table 6 it appears that, for 3 ring
systems, the RMR values are reasonably constant.
It is possible then that a single, "average" value
could be applied to all 3-rlng aromatic systems
with an acceptable limit of error. Other ring
system RMR's might also be obtained for further
applications. If, instead of using the parent
ion for monitoring and quantitation purposes, an
Ion that Is common to a given ring system were
chosen, then a vast amount of quantitative data
could be generated from a minimal amount of mass
spectral output. An examination of a compound
index based on mass spectral fragmentation pat-
terns for a number of PNA's indicates that a
common ion approach is not without some potential.
Although only PNA's have been examined thus
far, the same procedures will be applied to the
other non-volatile fractions as time permits.
Studies are already underway for the quantitation
of the volatile components, again using the RMR
approach.
References
1. Pelllzzarl. E. D., Carpenter, B. H., Bunch,
J. E., Sawickl, E., Environ. Sci. Tech., 9,
556 (1975).
2. Novotny, M., Lee, M. L., Bartle, K. D., J.
Chrom. Sci., 12, 606 (1974).
3. Bursey, J. T., Smith, D., Bunch, J. E.,
Williams, R. N., Berkley, R. E., Pellizzari,
E. D., American Laboratory, Dec. 1977, p. 35.
123
-------
Figure 1. Guifitr & Sampling Train
PARTICULATETRAP
rv>
6 AS
IN
TEIAX TRAP
71TS~1
/ Tn T
r 1
' »
I TO
CRYOGENIC
8BAI CHARCOAL™'
SAMPLE TR»'
PORT
TO DRV
TEST METER
ft CONTINUOUS
6ASANALYZER
TENAX TRAP
TO CRYOGENIC
TRAP
-------
i
uj
TEMPERATURE
Figure 2. TIC plot. Tenax volatlles.
TEMPERATURE
Figure 3. TIC plot. XAD-2 volatlles.
-------
TEMPERATURE
Figure 4. TIC plot. Condensate water.
-------
TAR
•V.3X | INSOLUBLE
| NON-POLAR I
•\-21X
Figure 5. Solvent partition scheme. Tars.
127
-------
TEMPERATURE
Figure 6. TIC plot. Non-volatlles-organlc base fraction.
-------
CO
s
I i r
TEMPERATURE
Figure 7. TIC plot. Non-volat1les-PNA fraction.
-------
Table 1. TENAX VOLATILES
Chrea>to- Uurlo
CO
CO
2
3
4
5
5.
5b
4
7
I
10
11
11.
12
13
13*
15
15*
15k
1*
17
17.
II
1*
20
21
22
22*
23
24
24.
23
25*
24
2*.
27
27.
21
30
51
33
5*
57
S»
tl
*3
*5
47
M
M
70
72
72
7*
71
•0
U
13
*0
•2
*4
(3
M
91
102
103
103
10*
101
110
111
11:
113
114
113
114
thlnphm
tTlcklara*thrl42
144
143
14*
147
141
U*
151
154
155
15*
157
151
13*
1*1
1*3
1*3
1*4
1*3
1**
14*
1*7
141
2.3.4-trlMtkrlchlaplMM
130
-------
Table 1 (cont'd)
ChroMCo- Uutloa
graphic Trap.
«**» •»• f"
Chroaato- Clue ion
(T.phic Taw.
LSL fSl
Jl
31a
32
52.
33
53.
M
54.
55
36
37
31
3«a
3*
M
We
61
U
62.
63
63a
64c
63
66
67
M
6*
16*
170
170
171
171
171
172
173
173
173
176
177
17*
17*
17*
Ul
Ul
Ul
Ul
1(2
1(4
1(4
1(6
1(7
1(7
1(7
IM
IM
IM
IM
1*0
1*1
1*1
Cj-pbanol
•acayl tcuofuraa l«
(2
(3
(4
(3
13.
(7
220
224
223
226
22*
232
236
237
23*
240
a^catrai
atkylaaifecbalaoa
diaackjlaapbtbalaB. iaoMr
dlaatkrUapbcbalaoa iaeaar
dlaacbjlaafbthalaaa UoMr
CI(|M iaovr * blpbaorlaM
C^-n.ibtbalrai
C}-aiabfblla»a
- accaapbi
2.2.4-trlatkrl •aaca-1.3-4iel
dl-laokocrrata
131
-------
Table 2. XAD-2 VOLATILES
airaauta- Clutloa
fraphlc Tavp.
Faak Ho. CO
Dajaco- Clucloa
ftaphlc t*mf.
Paak Ho. CO
la
Ik
1C
I
2a
la
3k
5.
5k
*k
*e
7
la
tk
»a
ft
lOa
10k
lla
Ilk
12*
12k
12c
lla
IJk
lie
1M
14
15.
It
17a
17k
II
Ita
70
70
70
70
70
70
70
70
70
70
70
70
70
70
70
70
70
70
70
70
70
70
70
7t
70
70
70
70
70
70
70
70
70
70
70
70
70
70
70
70
70
70
70
70
70
Vl4 U—r *
C_B iaaaar
T"IO "••"
toll
•atari ckl
crclooccaaa or
Vl4
Vl4
Vl4
Vl4
VM * VM
Vit * V*
Vl4
Vu * VM u— "
V» * Vu * Vi* 1~"
aqrraa* * C,-ekiia>aii iaoavr
Vu
Vu
Vi*
VM
e»"u
VM * cio"» u-ra
Vu
VM '-"
Vi* "~r
VM '—'
eio":o l"m"
=10".- * CIO"M "
21a
21k
22.
22k
22c
21
24.
24k
24c
25.
25k
2ta
Itk
2*c
27.
27k
27c
274
27*
27f
Me
2M
2*
X
lla
Ilk
J2.
12k
11
Ma
14k
lia
J5k
W
17k
(cratlniwa')
70
"70
70
70
TO
TO
TO
70
70
TO
Tl
n
72
71
74
7*
77
77
Tl
Tl
Tl
T»
N
n
u
12
u
17
17
17
M
II
M
fl
n
n
n
•4
9.
•5
91
C.-alhyl kaniaaa laoaar
C3-UK*-« « C^fa 1*
C.-alkyl kanxaaa laoawr
Cj-thlnphaaa
C10*20
C,-.lkyl bmana UaMr
Sfto
C10HM
C.-tlkyl
C10»20l«
trlMthrlchlophaoa
•.••curatcti bytfrocarbeu
1.2.1
C^-alkyl kauaaa tepaar
C4-alkyl keauna lacajar (caac.)
lain*
Uatkrl kaviaaa Uaavr
C(-alk?l kaauaa laaear
crnol llaaKC * C4-thlopha*a
l»a*ar (taac.)
C(-alkyl kaaiana liawr
Cit»x <>oMr (taat.)
Cj-chtopbaM laacar
uaaaturattfd n«droc.rkott»
C&-alkyl kaairae l*oa«r
CfU. kaat«a« Uoncr
-alkrl knia
creaol laoaar
132
-------
Table 2 (cont'd)
Chru.~-ico- ElutlM
crapMc T«op.
rVik 5j. CO
rtMto- Clutloft
>Mc To*.
*° CO
374
Ma
ISb
M
40a
40b
4CC
404
40.
41
42a
42b
42c
424
4Ia
43k
44a
44b
44c
444
44*
45.
43k
43c
46.
46k
47.
47k
47c
474
47.
4*b
4lc
30.
30!>
31j
3 Ik
ilc
i2«
»7
*t
*1
100
101
102
102
102
102
10)
104
103
106
107
10*
10*
110
111
111
111
112
112
111
114
113
116
lit
11*
11*
120
120
120
121
122
121
1 i
1 t
•aturattd by4rocarkona
C.Hj b«»«i
cr«*ol laoi
uftaacuracad by4rocarbo«a
C.-alkyl b«ni«n« UOMr
b*nx«M lcoM.tr
C.-alkyl kanzaii. laoaar
•acbyllndana laoaar
et-.u,i
C.-alkjrl
uphchalaaa
kattzocblopnhaiM Itoaar
12) Cj-alkrl
33i
32c
3)
33
34
55.
33b
36
57.
57k
J«a
SSb
3M
5*
60
tla
61c
614
62k
6)
64
43.
63k
67.
67k
67d
67e
3 C,,!.., !..«.»
1)1
1))
134
1M
142
143
143
146
146
147
14*
130
132
132
133
131
158
160
162
16)
164
163
166
It*
174
177
17*
IK)
112
It)
It)
117
117
114
191
1*7
Cl)"26
C.-*lkyl b«os«n« liomr
vnaatur.ctd hydroeatb«o.
ei4B21 U—r * C14B2t
C13H2» '
C14M26 l"
Cl)*20 l*
cu«2, i»
klphanyl
(ctmt.)
un**curat«4 hydrocarboo
C12«10 U—' .
C2-klphnyl Uoaar
41b*«a(ufaa
uaaacur.cad hydrocarkon.
uaaacuraead bydroeacbooa
dl»cetyl-6-hydrozirb.-nto(urin or
cu«:i
Cj-klaknyl
C17«)4 l—
ali«l lx<
aothr.icin
eisu:s
133
-------
Table 2 (cont'd)
Chrucuc.9- tlucioa
cr.iohu T«e». Co«pou«d
H-jk •;». <"c>
72. lit kj^a '"°«r
7Ik 217 uofauwn * el3«-. t*..k Ho. CO
Uvitloo
;r.phlc T««p.
Mk to. CO
5.
3b
6
I
12
11
13
14
17
II.
lib
1*.
20
2U
21b
24.
23
27
2*
30
31
32
35
37
70
72.1
73.2
73.1
12
13.6
M.4
M
If. 2
*2.4
M.4
102.4
102.1
103.2
103.6
1O9.6
110
117.2
111.I
124.0
12*. 2
133.2
114
111.2
142.1
143.6
130.4
15*. 2
164.4
163.2
fbmol
•-•ttgrlFknol
or £-(1.1.3. Vt.tr». thy !*«,!)-
1- or l-*t*jl**tbthml*»
•uuphclMn*
bytfroxyf laor«M
}.r~di».iliri-4.4'-kirrrUri
or 2.2'-dlMtlr»l-4,4>-blprrld]rl
or 2'-hydroxydlk«uafgr«
•athrMM.
pbunthna.
urkuol*
pbnntkrtdoo.
2.3-dllr/dTB-2-«tll)rl-4-(or|)-
'7-Mtlr/lphnulM
2-«.thjlUilo-3-—inothl.zolo-
43e 166.4 2-dl»b«7lflaor«M
2.4,7-trl».tl.yl-l.ft-
43d 166.1
43. 167.2
46. 161
46b 161.4
47. 171.6
47b 174.4
47c 173.2
41. 177.2
•Bk 171.1
4lc 17).6
4ld 1(2.0
41. 112.4
4* 113.2
30* 1(4.4
50k 114.1
51. 115.6
51k 1M.6
51e 1*0.1
51d 1»*.2
32 200
33. 201.2
5Jb 201.6
53c 206
34. 206.1
34k 212.1
34c 214.1
34d 211
34. 227.2
34f 213.6
34( 244.4
•.riprlirhininrk
2-pbMylljMjol.
1-Th*-T'lJ"*"'*'<—
prrn.
flooruttun.
Mtralr/drofluoruchn.
trlHtkrlponnthrn*
4,9-ilm* tlrjrlMphtbB- (2. 1>-
kithlaybn.
trlpawrUn*
134
-------
Table 4. ORGANIC BASE FRACTION
OiroMCO-
mthic
r«k»o.
1
I
3
4
J
»
7
1
9
10
11
11
12
13
U
14
15
It
U
17
11
11
1*
20
20
21
22
22 '
llutla*
T«».
CO
103.2
113. »
117.2
121.1
12J.«
131.*-
139.*
140.0
141.2
1«3.6
144.1
147.2-
153.4
147.2-
153.1
154.4
154
157
157
1M.4
1*2
162
1*4.4
166.2
1M.2
16*
170
170
171. 1
175
175
Co^ouuii
PTTldlM
•-wthTl-o-toluldlm or
4-aeAtyl pyrldlM
4-«e*t7l pyrldiM
qulaollM tnitlodld*
2.6-dl— ttrl-4-.tbrl pTrtdlm
6-Mthjrl qolnolln.
3- or 4-Mtbyl qidullM
2-«tao-5-chlora-4 , 6-41K thyl
ffTtmUimt
*-ni*i-|nuyyl qalaollaa
•thyl qulaollD** or 2-iBiao-5-
chloro-4.*-dl»tl>7l nrtmltSa*
2,6-dlKtlqrl qntoollM
•tkrl qulaallau
3.3.5-trlmtliTl-l-lxuaol
4-«tbrl qalaaUK
2.6-dlMtlqrl qataolln.
3-nBr-9np7l quiaDlla*
4-vliaTl-2-p7rl*)M
4-phBTl-2-prrl*>M
J.i-dlMthTl qulnolla.
1 , 2 , 3, 4- c* cnbrdracaibuola
4- «4 * >h«Byl-l-pyrUoM
•ckyl ttdaoliB*
Htkrl-VUlyllBdaima*
4-«i»i-»ioyyl qimlallM
3-wchTl-3-«llrlUdallM
3-Mtb7l-3-«ll7ll»4ollni
3-«thTl-3-«llrll^allM
4-pkmrl-2-yTrl*»«
QlTOVACO-
Sr.^hlc
Fuk la.
2)
24
24
25
26
27
21
n
X
30
31
32
33
34
34
34
35
36
'3?
31
3*
3*
3t
40
40
41
41
41
42
43
44
45
Clutlon
T«r-
CC)
177.1
171.1
171.1
179.1
191
193.1
197.4
191
200.4
200.4
207.2
210
210.1
216.4
21«.4
21t.4
221.2
214
226.4
221.*
234.4
234.4
234.4
231.4
231.4
239.*
239.6
2N.4
246
2*1.*
2(5
2*5
Co^xxmd
dlphcnyl ••!£•
1 ^Blno 4 nL4ii>l-C-mhiil
pyrlxldln*
phrayl-2-^TldOM
2.2'-dl«.th7l-4,»'-dlp7rld7l
l-«tbTl urfcualt
2-«tlqrl urkuaU
baiaOi)qiilaallM
•crldtD* or bMiBoOOqalAollM
btt&lD (h ) qtrlao lliu
1 •Inn 1 nli)l c*rWx>l>
V««luu 9 «Ui>l urbuoU
3-HChTlkMB9quiaalliu
J-«loo-9-.th7l eubuoU
3-MCbrUnnqalael 1m
>-iBlao-9*«db7l e«rbaxol«
ba»oO>)quli>olli» or acrldla*
3- nd 5-MDIT1-2 plmpltadeU
3- Bd 5-wibTl-2-vhnTlluloU
3-«thrlk«h«Tll»dol.
4-*CTTylqulAollaa
3- — d 5-MCkrl-2-pbrarlladol*
3-b»«iyH»
-------
Table 5. NON-VOLATILES - PNA FRACTION
Chroauto-
tr.pt. Ic
raak No.
1
2
3
4
4
5
3
4
7
(
9
9
10
11
12
13
1*
14
15
15
It
19
20
22
23
24
23
27
28
29
30
32
33
34
35
34
38
Llutlon
Trap.
CO
lit. 8
123. t
129.8
113.6
134.4
1J6.8
134.8
138.8
150.4
152.0
164.8
164.8
161.4
168.8
172.0
173.4
177*6
177.4
1(0.8
1(0.0
181.6
1(7.2
1(9.4
192. (
1*4.4
197.4
198.4
202.4
209.4
212.8
215.2
224.0
229.4
231.2
236.0
239.2
242.4
Co-pound
•athyl phmyl .catylana?
4->athylbanie(b)furan
•athyl Indanaa
naphthalana
2, 3-dlhydro-2-a»taylbanEoforaa
l-hydroxy-2-a.thyl-4-.tbyl-
baoxanaT
o-hydroxyacatophanonat
1-aa thy 1-4-ao r-bayl-1, 2.3,4-
tatrahydroaaphthalana
2-aathylnaphthalana
l-a*thylnaphthalana
•tb-lnaphthalan*
2. 3-dlvathy Ibanzo (b) thlophaoa
•atnylnaphthal.il* Uonara
1.5- and 2.3-d-aathyl
naphthalana
1 . 3-dia«thylnapkthalana
1 . 2-dlB*thylnaphthalana
2-athyl-3(or 7)-a*thylbanao (b)
thlophana
2-1-propylaaphthalana
dlban-of-na
propylnaphth.lana
flaoraaa
1. J-«lhydro-4,6-dl-«tbyUhlar«>
(3.4-c)enlophana
2-MChylblph-nyl
2-hydlu-yf luoraoa
2-tart-katrlmaphthalaaa
l-n.tbyl-7-lao-propyln.phth.lm.T
1-aatbylfuaraaa
•athoxyfUoran. laonar
•acboxyfloora*** or ortho- and
para-phax-lanlaola
dlaathylf luoren.
3-wthyldlbaniothlophan.
phanyl K-xylyl katona?
•athylphananthrcn*
2-vcthylcarb.xol.
t.tr.hydroanthr.quinon.7
ir.phlc
r-.k No.
19
40
42
43
44
45
46
47
48
49
30
50
51
51
52
32
52
53
33
54
54
54
35
55
56
57
58
59
to
41
42
62
63
44
45
44
47
6«
Elul lun
Imp.
CC)
244.0
248-
252.0
255.2
2)6.0
257.4
240.8
243.2
244.0
245
245
245
245
245
245
245
245
245
245
245
245
245
245
245
245
245
243
245
245
245
245
243
245
243
245
245
245
245
265
Conpound
4,5-dlaathyl-9.10-dlhydro-
phvnanthrana?
dln*thylphananthr.nai
8-nor-butyl-pnan.othren*?
athylanthracane
pyr.n.
l-aathylbanu>(1.2-b:4.3-b)-
dlthlophana
haxada capy r ana 7
trinathylphananthranaT
l-Mth-lpyrana
tri-athylphananthraaa
trlMthylpbanaathrana
1.4-diavthylanthracaoa
1.2.3.4-tatrattydrotrlphanylan.
i . 4-dlhydro-2 . 3-bantcarbatola
catraBydrotrlphanylan.
dlh,dn*«ucarba»l.
4,4'-dlchloroblphaoylT
tatratr/drotrlpbaiiylana
dihydroxyanthraqulnoaa
butyl phthalyl botyl phthalata
3-nor-haiylp.ryl«n.T
3,4-dl-athyoxyphananthrana
3.3'-bl-lndolyl
4.4'-dlchloroblphanyl
1.2-dlphanylbauana
1 . 4-4iphanylDaniana
1 . 3-dlpbanylbaniana
naiahrdrobani(a)anthracana
dlphaoylbancana
triphaoylanaT
di-nor-occylphtbalata
dl-2-athylhaxylphthalata
•athylbanio(a)anthracana or
3-Bathylchryaana or
2-MtaTltrlphanylana
3-nathTlbanto(a)anthTaca-at
9*. 10-, or ll-nathylbani(a>-
anthracan*
parylanaT or banxpyrma?
banipyrana or parylanc
3-«athylacaaaphthylan*
136
-------
Table 6. RELATIVE MOLAR RESPONSE
Compound
Naphthalene
2-Me thyl naphthal ene
Fl uorene
Phenanthrene
Anthracene
9-Methyl anthracene
Fl uoranthene
Pyrene
Chrysene
RMR (Mag. Sector)
42.51
4.31
2.81
1.28
1.41
1.32
1.87
2.35
15.64
RMR (Quadrupole)
1.48
1.00
1.16
1.26
1.80
.91
1.61
1.78
.74
Table 7. PNA QUANTITATION
Compound Found (Mag. Sector) Found (Quadrupole)
Naphthalene 25.5 ng
2-Methylnaphthalene 45.4 85.2
Fl uorene 13.9 33.7
Phenanthrene 61.4 36.9
Anthracene 57.4 29.7
9-Methylanthracene 76.0 48.4
Fluoranthene 12.4 5.2
Pyrene 14.7 8.0
Chrysene 11.3
Perylene 10.4
137
-------
TRANSFORMATION OF POM IN POWER PLANT EMISSIONS
by
D.F.S. Natusch, W. A. Korfmacher,* A. H. Miguel,**
M. R. Schure and B. A. Tomklns
Department of Chemistry
Colorado State University
Fort Collins, Colorado 80523
Abstract
With the advent of Increased usage of coal as
an energy source, environmental considerations must
be examined. It has been known that fossil fuels
produce organic emissions commonly known as
Polycycllc Organic Matter (POM) which constitute a
health threat 1n that some POM show a high degree
of carcinogenic activity. In this paper some
aspects of POM transformations will be examined
which Include the mode of transformation, reactions
which are believed to occur 1n the environment, the
environmental ramifications of POM transport, and
the associated considerations of sample collection,
theoretical models, laboratory experiments, and
emission controls.
Introduction
During the last several years there has been
increasing recognition of the importance of
polycyclic organic matter (POM) as an air pollu-
tant. rn This 1s largely due to the Increasing
weight of circumstantial evidence which relates the
Incidence of certain types of cancers (notably
bronchial carcinoma) to atoraspherlc levels of POM-
espedally benzo(a)pyrene.ri"^ Recently, an
amendment to the 1970 Clean Air Act'5"1 has
recognized POM as a potential hazard to health.
Whether this hazard 1s substantiated or not it is
reasonable to expect an increase in POM sampling
and analysis activities in the future. Conse-
quently, it 1s appropriate at this time to consider
some of the factors which determine the authen-
ticity of POM measurements.
Specifically, 1t 1s the purpose of this paper
to point out that many of the compounds which are
classified as POM can undergo both physical and
chemical transformations. These transformations
can, it 1s suggested, profoundly Influence the
meaning and validity of POM measurements, the
environmental Impact of POM, and the applicable
control technology.
There 1s now a considerable body of evidence1"11
which shows that POM formation occurs as a result
of combustion of carbonaceous material, that
formation is promoted by reducing conditions, and
that similar relative amounts of individual
compounds are produced Irrespective of the nature
of the fuel. Most of the Information available
concerns the polycyclic aromatic hydrocarbons (PAH),
though 1t 1s recognized that polycyclic aromatic
compounds containing both hetero-atom rings and
ring substituents are produced.m •
The detailed mechanism(s) of POM formation are
not well understood; however, it 1s widely accepted
that POM Is formed via a free radical mechanism'61
which occurs in the gas phase. As a result POM
originates as a vapor. On the other hand, there Is
a large body of data which attests to the fact that
*Present address: Dept. of Chemistry, University
of Tennessee, Knoxvllle, Tennessee 37916.
"Present address: Un1vers1dade Campinas. Sao
Paulo, Brazil
POM present 1n the atmosphere Is almost invariably
found in parti cul ate form. rn It is apparent,
therefore, that vapor-to-par tide conversion takes
place between the points of formation of POM In a
combustion source and its determination 1n the
atmosphere. It 1s the mechanism of this vapor-
particle transformation and the subsequent
oxi dative transformation of parti cul ate POM which
constitute the subject of this paper.
Vapor-Particle Transformation
A number of workers have reported the presence
of vapor phase POM in combustion sources where
elevated temperatures are encountered. '7-i
-------
and desorptlon, respectively. If one assumes, a
puoiu., that kj and k-j 1n equation (1) represent
first-order processes, then the rate of adsorption
of P can be written
ki[P](i-e)A - k.
(2).
In equation (2), 6 Is the fraction of the total
available adsorption sites which are occupied and
A 1s the surface area of the paniculate material.
The rate constant for desorptlon, k.lt 1s
given by
exp [-Ed/RT]
(3)
where E . 1s the activation energy for desorptlon,
T 1s the absolute temperature, and k, h and R are
Boltzmann's constant, Planck's constant, and the
universal gas constant, respectively. The rate
constant for adsorption, kt. 1s
where c 1s the so-called "sticking coefficient"
(the probability that the orientation of a molecule
with the spherical particle surface will result 1n
adsorption), Ma 1s the molecular weight of the
adsorbate species, and Ea 1s the activation energy
for adsorption. A number of assumptions are
Inherent 1n equations (3) and (4). These are
presented briefly 1n reference (14) and discussed
1n detail 1n a forthcoming publication. ns'>
In order to evaluate the rate and extent of
adsorption as a function of temperature 1t 1s
convenient to compute the mole fraction, X, of
total POM adsorbed and the time taken to achieve
one half of the equilibrium adsorption at a given
temperature, tv. The equations giving these
quantities are: flu«15'>
(2nMak)!» d* P T* exp (E -EJ/RT
I. = i + . a jn a a
'* ocw_n j 2 1-8 N *s
and
V
where
-i
In
- Xk2
(5)
(6)
l-e)J — •
U2I
and N0 = 6.02 x 1023.
The quantities P and Wp are, respectively, the
density and mass per unit volume of the adsorbing
particles and d_ and d are the mass and surface
median diameters, respectively, of the particle
size distribution. In the case of a log Gaussian
distribution of particle sizes dm and ds are
related by
In d, » lnd_ - In2o (7)
S Bl
where a 1s the geometric standard deviation of the
distribution. 'l*^
The evaluation of equations (5) and (6)
requires that realistic values be chosen for the
quantities c, Ma, WD, p, d-^s2' Ea- Ed» and 9-
In fact the fractional surface coverage, 6, 1s
almost certainly dependent on both temperature and
the values of Ea and Ed; however, it is reasonable
to assume that the adsorption process 1s zeroth
order 1n 6 so that a constant value can be chosen.
Using available literature values for the above
parameters (given 1n figure captions) one can
construct the temperature dependent plots of -X
and tj, presented in Figures 1 and 2.
Consideration of the data presented in
Figure 1 shows that, over a wide range of
conditions, POM present at combustion source
temperatures (>150°C) 1s predicted to occur mainly
in the vapor phase, whereas at the ambient temper-
atures encountered following emission (<40°C)
essentially quantitative adsorption 1s predicted.
Variations from this general behavior are, however,
predicted for Individual compounds and widely
different particle size distributions and mass
loadings.
The temperature dependencies shown in Figure 1
are generated with the assumption that adsorption
equilibrium is achieved at all temperatures. This
requires that the rate of attainment of equilibrium
1s fast compared with the rate of change of temper-
ature experienced by a given vapor-particle
combination. The rate of attainment of equilibrium
is Indicated in terms of the half-time for reaction
in Figure 2. These data show that reaction times
depend primarily on the activation energies Ea,E
-------
identification was established using GC retention
time data and in many cases was confirmed using
either a single-ion mass chromatogram or a mass
spectrum from a GC peak. Only crude vapor traps
were employed during sample collection so no
quantitative measure of vapor phase POM was
obtained. Fluorescence measurement of condensation
trap residues did, however, indicate that a
considerable quantity of POM was present from
in-stack sampling but none was in residues from
plume sampling.
These results establish quite firmly that, in
this power plant, considerably more POM is associ-
ated with fly ash collected from the plume at a
temperature of 5°C than with that collected from
the same stream at a temperature of 290°C. This
behavior is in exact accord with that predicted in
Figure 1. Furthermore, since the two collection
points were only -\-100 ft apart,quite rapid vapor-
to-particle conversion is indicated. Unfortunately
while the full range of aerodynamic equivalent
particle sizes accessible to Anderson Stack and
Hi Vol samplers was collected, this only
represented a small range of specific surface area
due to the considerable particle irregularity
encountered. Nevertheless correspondence
between specific concentration of POM and specific
surface area of fly ash fractions was noted. This
further suggests the operation of a surface
adsorption mechanism.
Laboratory Experiments
In an attempt to obtain direct measurements
of the rate and extent of POM adsorption and to
evaluate the quantities c, Ea, Ed in equations
(3) and (4), a series of laboratory simulation
experiments were set up. Fresh coal fly ash which
had previously been shown to contain no detectable
POM was presented in an expanded bed through which
a stream of air containing pyrene vapor(l7' was
passed. The objective was to expose all particles
to the same constant concentration of pyrene for
different times and to determine the specific
concentrations of pyrene as a function of time at
different temperatures.
The experiments showed that the uptake of
pyrene was so rapid that a uniform vapor phase
concentration could not be achieved. The amount
of pyrene required to saturate the fly ash was,
however, shown to increase significantly with
decreasing temperature. Futhermore,
attempts to remove adsorbed pyrene by heating in
a stream of clean air were unsuccessful.f18^
While these experiments were essentially
qualitative they do establish the facts that coal
fly ash will strongly adsorb pyrene (and probably
other POM) and that the saturation capacity is a
strong inverse function of temperature.
Overall, therefore, the results of these
three studies point strongly towards the Idea that
POM, formed initially as vapor, 1s adsorbed onto
co-entrained particulate material as the temper-
ature falls. There is some doubt about the rate
at which this process takes place but the evidence
1s in favor of rapid (on the order of seconds)
adsorption, even at ambient temperatures, under
most conditions encountered in or near combustion
sources.
Oxidative Transformation
It is widely accepted that POM present In
atmospheric aerosols is highly susceptible to
photochemical transformation in the presence of
sunlight. Indeed, several studies of POM, or of
individual polycycllc compounds, present on
substrates such as airborne particles, soot, and
aluminam have shown that conversion to an
oxidized form occurs quite readily. Yet, if one
subjects the available information to critical
examination it is apparent that, while photochemical
transformations undoubtedly occur, there Is
considerable disagreement regarding the behavior of
different compounds on different substrates.
Our own studies have been designed to
establish the factors which determine the rate and
extent of photochemical transformation of individual
polycyclic aromatic hydrocarbons present on the
surface of coal fly ash. Samples have been
prepared by adsorbing Individual species from the
vapor phasen7"191 onto POM-free fly ash which is
then subjected to irradiation in the form of
sunlight and of several sunlight-simulating ultra-
violet sources. Surprisingly, these experiments
showed no evidence of photochemical transformation
of the compounds employed (Table 3) even though
long irradiation times and unrealistically high
irradiation intensities were employed. Since
these compounds were rapidly transformed in solution
one concludes that association with the fly ash
surface stabilizes them against photochemical
transformation.
Even though photochemical transformations
were not observed, several of the compounds
studied did undergo partial oxidation simply as a
result of being adsorbed onto fly ash (Table 3).
In passing, most, but not all, of these reactive
compounds possessed at least one benzyl1c or
doubly-benzylic methylene, which is highly
susceptible to oxidation. This non-photochemical
oxidation process was found to depend greatly on
the nature of the adsorbing substrate. Thus, fly
ash and activated carbon were found to promote
oxidation whereas silica, alumina and glass did
not. «*
While highly preliminary, these results do
suggest that certain polycycllc aromatic species
undergo oxidation as a result of adsorption onto
active substrates. In all cases the extent of
conversion was less than 100 percent and the half
times for conversion varied from a few minutes to
several hours. The tentative conclusion, there-
fore, is that a number of compounds present in
so-called POM can be oxidized both in the
presence and absence of light and that this behavior
is greatly influenced by the nature of the
substrate with which a compound is associated.
Overall, it is apparent that both photochemical
and non-photochemical conversion of particulate
associated POM 1s a complex and 111-defined process
and merits considerable futher Investigation.
Ramifications of POM Transformation
The types of transformation discussed above,
while far from definitive, do point out several
Important considerations which should be borne In
mind in assessing the environmental/health Impact
and methodology for the measurement and control of
POM. These may be categorized as follows:
140
-------
Environmental Considerations
From the standpoint of the environmental and
potential human health Impact of POM two points
are Important. The first 1s that adsorption of
vapor phase POM onto participate matter will
result 1n the predominance of POM on small particles
which provide the largest available surface area
per unit mass. Thus, POM will be preferentially
concentrated 1n particles whose aerodynamic size
falls 1n the range which can remain airborne for
several days and which 1s capable of being deposited
1n the pulmonary region of the human respiratory
system when Inhaled. (20,2^ Tn1s prediction 1s
1n accord with the results of measurements of the
atmospheric aerosol size distribution of POM.f2n
The second point to be noted 1s that, due to
the demonstrated photochemical and nonphotochemlcal
transformation of POM, several of the compounds
which are present In Inhaled aerosols may be
significantly different from those actually measured
1n, and emitted from, a given combustion source.
Present Indications are that the Initial stable
conversion products of PAH's are their correspond-
ing ketones and qulnones which apparently exhibit
lower carcinogenic activity than their parent
compounds. However, it 1s appropriate to
recognize that the enhancement or reduction of
carcinogenic activity associated with POM 1s still
an entirely open question.
Sample Collection
If POM 1s capable of converting rapidly from
vapor to particulate form then the relative amounts
of each would be expected to vary with the position
(I.e. temperature and particle surface density) 1n
a combustion system. Consequently, measurements
of separate vapor and partlculate POM will apply
only to a specific point 1n a specific plant
operated under specific conditions and cannot be
extended to other power plants 1n general.
Furthermore, the predictions of Figures 1 and 2
suggest that considerable vapor-to-part1cle
conversion may actually occur within the sampling
device -especially when 1t 1s maintained at a
temperature which 1s different from that of the
stream being sampled. As a result measurements of
the particle size distribution of POM and of the
ratio of partlculate to vapor phase POM will not
represent true quantities but rather sampling
artefacts.
Of more practical Importance, however, 1s
the very clear need to collect both vapor and
partlculate POM 1n order to establish total POM
emission factors. Current methodology for sampling
emission sources does employ vapor collection
devices; however, some doubt has been expressed
concerning their collection efficiency,'22^ and
1t seems likely that present POM emission
estimates may be low where measurements are made at
elevated temperatures. Certainly emission
estimates based on analyses of partlculate material
alone, when 1t 1s collected from within an emission
source, are likely to be grossly 1n error. '23^
Finally, It will be noted that the
non-photochemical oxidation observed for POM
adsorbed on fly ash 1s unlikely to occur to the
same degree 1n adsorbent vapor phase POM collectors.
Consequently the compounds Identified 1n vapor
collection devices may differ considerably from
those emitted from a source —certainly the
relative concentrations may differ. In short, the
material collected may differ qualitatively and
quantitatively from that emitted.
Overall, therefore, it seems that the
transformation processes discussed herein may
result 1n extremely poor representation of emitted
POM unless special sampling precautions are taken.
Indeed, the only bona. fade, way of characterizing
POM emissions would be to collect material directly
from the emitted plume. Since this is unrealistic
for routine measurements, it is recommended that
emphasis be placed on the determination of total
POM under conditions which promote particle
association prior to sample collection. For this
purpose it 1s suggested that the sampling train be
operated at or below ambient temperatures.r2^
Emission Control
The vapor-to-particle transformation
behavior of POM has two Important consequences in
terms of emission control. First, the fraction of
the total POM which is in the vapor phase during
passage through particle control devices will not
be collected. Secondly, the effect of adsorption
1s to move the aerodynamic equivalent mass median
diameter of the adsorbed species to a value which
Is significantly smaller than that of the
substrate particle mass as described by equation
(7). Just as this enhances the ability of POM to
penetrate the human respiratory system so it will
also Increase the difficulty of controlling POM
emissions since small particles in the range of
interest (0.1 to 5.0 um aerodynamic diameter) are
collected with reduced efficiency by most particle
control devices.''25'*
One positive point which can be made is that
collection of POM vapor by scrubbers may be quite
efficient Insofar as the low temperatures
encountered in liquid systems may promote
adsorption and at least partial collection of
vapor species.
Conclusion
It 1s apparent from the foregoing discussions
that transformation of POM from vapor to
partlculate form and to different molecular forms
can take place within a combustion source, during
sampling, and following sample collection.
Indeed a number of chemical transformation processes
In addition to those discussed have been
reported.m It Is also apparent that such
transformations can profoundly influence the
meaning and validity of POM measurements, the
environmental Impact of POM, and the applicable
control technology.
Finally, It 1s important to recognize that the
processes described herein, although presented in
terms of a coal-fired power plant, will occur
quite generally. They must, therefore, be taken
Into account when considering any combustion source.
141
-------
Acknowledgeroents
This work was supported 1n part by research
grants ERT-74-24276 from the National Science
Foundation, EE-77-S-02-4347A from the Energy
Research and Development Administration, and by
R-803950 from the U.S. Environmental Protection
Agency, Environmental Research Laboratory-Duluth.
References
1. Committee on Biologic Effects of Atmospheric
Pol1utants: Partlculate Polycycllc Organic
Hatter. National Academy of Sciences,
Washington, O.C., 1972.
2. Proceedings of General A1r Pollution and
Human Health with Special Reference to Long
Term Effects Meeting, Stockholm, March, 1977,
EnviHomeMtal Health Peupective* (1n press).
3. Chrlsp, C.E., Fisher, G.L. and Lammert, J.E.
Mutagenicity of Filtrates from Resplrable
Coal Fly Ash. Science, 199, 73 (1978).
4. Pitts, J.N. Atmospheric Transformations of
Selected Carcinogens. In: Conference on Chemi-
cal Carcinogens 1n the Environment Proceedings,
California Institute of Technology, Pasadena,
California, 1978.
5. The Clean A1r Act As Amended August 1977.
Serial No. 95-11. U.S. Government Printing
Office, Washington, D.C.
6. Badger, G.M. Mode of Formation of Carcinogens
1n Human Environment, taut. Cttnce* Jntt. Monog*.
9, 1-16.
7. Emissions from Residential and Small
Commercial Stoker-Coal-Fired Boilers Under
Smokeless Operation. Publication EPA-600/7-76-
029. U.S. Environmental Protection Agency,
Office of Research and Development,
Washington, D.C., 1976.
8. Preliminary Emissions Assessment of Conven-
tional Stationary Combustion Systems; Volume
I - Executive Summary. Publication EPA-600/2-
76-046a. U.S. Environmental Protection Agency,
Office of Research and Development, Washing-
ton, D.C., 1976.
9. Field Testing: Trace Element and Organic
Emissions from Industrial Boilers. Publication
EPA-600/2-76-086b. U.S. Environmental
Protection Agency, Washington, D.C., 1976.
10. Bertsch, W., Chang, R.C. and Zlatkls, A.
The Determination of Organic Volatlles 1n A1r
Pollution Studies: Characterization of
Profiles. 3. Chaowttog*. Sci., 72(4), 175
(1974).
11. Pupp, C., Lao, R.C., Murray. J.J. and Pottle,
R.F. Equilibrium Vapour Concentrations of Some
Polycycllc Aromatic Hydrocarbons, Asi»06 and
Se02 and the Collection Efficiencies of these
Air Pollutants. A&HOA. Env., 8(9), 915 (1974).
12. Hangebrauck. R.P., von Lehmden, D.J. and
Meeker, J.E. Sources of Polynuclear Hydro-
carbons 1n the Atmosphere. Public. Health.
Se*vice Publication 999-AP-33, U.S. Dept. of
Health. Education and Welfare (1967).
13. Castellan, G.W. Physical Chemistry. Second
Edition. Add1son-Wesley Publishing Co.,
Reading, Mass., 1971.
14. Natusch, D.F.S. and Tomklns, B.A. Theoretical
Consideration of the Adsorption of PAH Vapor
onto Fly Ash in a Coal Fired Power Plant.
In: Cardnooenesis. Vol. 3. Polynuclear
Aromatic Hydrocarbons. P.W. Jones and R.I.
Freudenthal, eds., Raven Press, New York,
1978, pp 145-153.
15. Natusch, D.F.S. and Schure, M.R. Manuscript 1n
preparation.
16. Butcher, S.S. and Charlson, R.J. An Introduc-
tion to A1r Chemistry. Academic Press, New
York, 1972.
17. Natusch. D.F.S. and Miguel, A.H. A Diffusion
Cell for the Preparation of Dilute Vapor
Concentrations. Anal. Chen., 47, 1705 (1975).
18. Miguel, A.H. Studies of Gas Solid Reactions
of Environmental Significance: I. Pyrene
Adsorption Onto Fly ash; II. Thlol Oxidation
and Adsorption by Activated Charcoal. Ph.D.
Thesis, University of Illinois, 1976.
19. Korfmacher, W.A., Wehry, E.L. (Dept. of
Chemistry, University of Tennessee) and
Natusch, D.F.S. (Dept. of Chemistry, Colorado
State University), unpublished results (1978).
20. Davlson, R.L., Natusch, D.F.S., Evans, C.A.
Jr., and Wallace, J.R. Trace Elements In Fly
Ash: Dependence of Concentration on Particle
Size. Env. Sfri.. Techno*., 8, 1107 (1974).
21. Natusch, D.F.S. and Wallace, J.R. Urban
Aerosol Toxldty: The Influence of Particle
Size. Science, 1S6, 695 (1974).
22. Jones, P.W.. 61ammar, R.D., Strup, P.E. and
Stanford, T.B. Efficient Collection of
Polycycllc Organic Compounds. Env. Sci..
Techno*., NM«), 806 (1976).
23. Natusch, D.F.S. Potentially Carcinogenic
Species Emitted from Fossil Fueled Power
Plants. Envixon. Health Pe/iApective* (1n
press).
24. Environmental Assessment Sampling and
Analysis: Phased Approach and Techniques for
Level 1. Publication EPA-600/2-77-115. U.S.
Environmental Protection Agency, Office of
Research and Development, Washington. D.C.,
1977.
25. White. H.J. Industrial Electrostatic
Precipitation. Addlson-Wesley. Reading,
Mass.. 1963.
TABLE 1. INDIVIDUAL POLYCYCLIC AROMATIC
COMPOUNDS IDENTIFIED* IN EMITTED
COAL FLY ASH
Fl uorene**
Phenanthrene**
Anthracene**
9 . I0-d1methyl anthracene**
Fluoranthene**
Trlphenylene**
Pyrene**
Chrysene**
Benzofl uorene
1-metr
Benzoi
Benzo
Perylc
Benzo
Benzo
Benzo
Anthar
Corom
lylpyrene
ihenanthrene**
a)anthracene
me**
a)pyrene**
ejpyrene**
ghOperylene
ithrene
>ne
* By gas chromatography or literature data (MAS
1972, Hangebrauck et al., 1967).
**Ident1ty confirmed by GC/MS using reconstructed
Ion chromatograms.
142
-------
TABLE 2. MEASUREMENT OF POM EMITTED FROM
COAL FIRED POWER PLANT STACK
Compound Inside Stack
Fluorene
Phenanthrene
Fluoranthene
Pyrene
Benzofluorene
1-methylpyrene
Benzophenanthrene
Benzo(a)pyrene
Total , ,1M,
fluorescence/g J'01l«
Compound
AcHdlne
ND
ND
ND
ND
ND
ND
ND
ND
r3) units
TABLE 3.
Outside Stack
0.5 pg/g
12
17
12
2
0.6
3
8
3.68 units
OXIDATION OF POM ADSORBED ONTO FLY ASH SURFACES
Qii**f nf*o
•JUT 1 Ql*C D*»/\/Jii/» +
Oxidation
(^\Y^-\^-\\ No
Anthracene
Benzo(a)fluorene
Benzo(b)fluorene
Benzo(a)pyrene
Carbazole
9,I0-d1hydroanthracene
9.I0-d1hydrophenanthrene
l0.ll-d1hydro-5H-d1benzo-
[a,d]-cycloheptane
Yes
Yes
Yes
No
No
Yes
Yes
No
Anthraqulnone
143
-------
9,I0-d1methylanthracene
Fluorene
Fluoranthene
9-nltroanthracene
Yes
Yes
No
Probably
"Qulnone"
Fluorenone
Phenazlne
Phenanthrene
Pyrene
No
No
No
144
-------
!.00
Q90
0.80
0.70
0.60
X050
0.40
aso
OL20
0.10
0.0.
UOO
Q90
CL80
0.10
0.60
X050
Q40
0.30
020
0.10
250 300 350 400 450 500 550 600
T.K
°250 3C
300 350 400 450 500 550 600
T,K
X0.50
250 300 350 400 450 500 550 6OO
T.K
250 300 350 400 450 500 550 600
T.K
Figure 1. Dependence of Mole Fraction of PAH Adsorbed on Temperature
Reference values: c - 1; m, - 3.360 (10'") g/molecule; wp - W6 wg/«3 - W« 9/cm3;
,T= 3 9/cm3; d^/ds2 * 10'« cm; Ea - -10 kcal/mole; Ed = -30 kcal/mole; e » *.
(a) Variation of mole fraction of PAH adsorbed as a function of temperature and C =
(b) VaHatlOT°of Sole fraction of PAH adsorbed as a function of temperature and
d_3/ds2 = 0.1, 1.0, 10.0 um. £
(c) Variation of mole fraction of PAH adsorbed as a function of temperature and wp =
(d) Variation of mole fraction of PAH adsorbed as a function of temperature and Ea-Ed
-15, -20, -25 kcal/mole; Ea « -10 kcal/mole.
145
-------
10
\
Qoa
\
S,
250 300 350 400 450 500 550 600
T,K
300 350 400 450 500 550 600
T.K
250 300 350 400 ISO 500 550 600
300 350 400 430 500 550 600
T,K
t
eference values: c - 1; n. - 3.360 (l
-------
THE ASSESSMENT OF ATMOSPHERIC EMISSIONS FROM PETROLEUM REFINING
By
D. D. Rosebrook, R. G. Wetherold and G. E. Harris
Radian Corporation
Austin, Texas 78766
Abstract
A study, funded by the U. S. EPA, is
currently being conducted in order to
assess the atmospheric emissions from
petroleum refining operations.
To accomplish this assessment mea-
surements of fugitive hydrocarbon and
stack emissions are being made at a number
of refineries throughout the country.
Sources being sampled include valves.
flanges, pumps and compressor seals,
process drains, pressure relief devices,
process vents, heater and process stacks,
cooling towers,'API separators, dissolved-
air flotation units, open ditches, baro-
metric sumps and holding ponds.
This paper describes the methods
being employed for the selection and
screening of the above sources and the
criteria used for making the sample - no
sample decision.
Introduction
The program that Radian undertook for
the EPA has three objectives:
• to provide data which can be used
for "offset calculations" (This
involves the development of emis-
sion factors for the various
sources in a refinery.);
• to provide a data base for health
effects use (This involves the
qualitative and quantitative mea-
surement of potentially hazardous
components of refinery emissions.);
and
• to assess the availability and
effectiveness of control technology
(This involves the collection of
data regarding both controlled and
fugitive emissions.).
Pollutants emitted into the air from
a refinery fall into one of two categories:
• controlled emissions, or
• fugitive emissions.
The first type is released into the
air at a controlled rate from a point
source such as a stack or a vent. The
second is released without control of
rate or direction. Many types of fugitive
emissions cannot be measured by existing
standard sampling and analytical tech-
niques. Therefore, development of
reliable measurement procedures is an
essential prerequisite to the development
of strategies for the control of fugitive
emissions.
The decision to emphasize fugitive
emissions in our current study is the
result of discussions with the API and the
EPA. The interest of the API stems from a
need within the refining industry to have
access to updated emission factors. The
interest of EPA stems from its charge to
develop adequate control technology for
potentially hazardous pollutants and from
a need for emission factors accurate and
current enough to allow offset changes to
be made in existing refining operations.
The most cost-effective program combines
elements of both fugitive and nonfugitive
source sampling.
Previous studies (1958) have dealt
with total hydrocarbon emissions and with
emissions of other criteria pollutants.
This study considers total and nonmethane
hydrocarbons but does not include other
criteria pollutants except from nonfugitive
sources where current data are not avail-
able. The present study also seeks to
characterize the individual compounds being
emitted which may be potentially hazardous
even at trace levels.
The makeup of a "refinery1' is diffi-
cult to define since the process units
employed are dependent on the demands for
the refinery's products. Therefore, a test
plan was developed that is not definition
limited.
Approach
The type and arrangement of process
units within refineries may vary consider-
ably from one location to another; however,
the individual unit fittings and processes
contained in each refinery should show close
similarities.
The unit fittings, ranging from valves,
pumps, etc., to various seals, are available
to industry in a smaller number of varieties
than the varieties which may result from the
various arrangements of the unit fittings
to give unit processes. These fittings
largely control the amount of fugitive emis-
sions potentially available from a given
unit process.
Thus, the approach is to experimentally
determine the emissions from unit fittings.
These data will allow calculation of the
emission factors as well as calculation of
the fugitive emission potential of individ-
ual unit processes.
In the case of operations involving
fugitive emissions other than those from
147
-------
unit fittings, special sampling approaches
have been devised to allow measurements or
estimation of the total emissions from
those operations.
Sources of Hydrocarbon Emissions
There are several known sources of
hydrocarbon emissions in a refinery; they
are listed in Table 1.
TABLE 1. REFINERY HYDROCARBON
EMISSION SOURCES
Process Emissions Sources;
• Compressor Engines
• Catalytic Cracker Regeneration
Stacks
• Air Blowing
• Boilers and Process Heaters
• Flares and Blowdown Systems*
• Vacuum Jets (Vented to the Atmos-
phere)
• Sulfur Recovery or Tailgas-Treating
Unit Stacks
Fugitive Emissions Sources:
• Pipeline Valves
• Miscellaneous Joints
• Pressure-Relief Devices
• Pump and Compressor Seals
• Process Drains
• Blending Operations
• Cooling Towers
• Intermediate Storage Tanks*
• Blind Changing
• Maintenance Turnarounds (When
Vessels are Vented to the Atmos-
phere)
• Wastewater Systems
• Barometric Condensers on Vacuum Jets
• Loading Operations*
• Sampling
• Decoking Operations*
*Not within the scope of this work.
Hydrocarbon emission sources other than
stacks or vents were considered to be
fugitive sources. Stacks or vents which
can be identified as the principal hydro-
carbon emission points were considered to
be process sources. Some of the emission
sources can qualify as either (or both).
Process emissions have received a
great deal of study. Control technology
development has been concentrated on the
reduction of the emissions of hydrocarbons
(and other criteria pollutants) from these
sources.
Fugitive emissions, however, have
received only intermittent study since
1958. Fugitive emissions sources allow
hydrocarbons to escape to the atmosphere
principally by accident, inadequate mainte-
nance, or poor planning. Some hydrocarbon
leakage normally occurs even in the absence
of such conditions. The sources are
diverse both in their physical character-
istics and in the means by which they
allow hydrocarbon escape.
The next step is to define the pro-
cedures by which individual fugitive
sources are to be selected for sampling.
Criteria were selected for choosing the
refineries to be visited and units to be
sampled.
Choice of Refineries and Process Units
The first step in designating specific
sources for sampling was the choice of
refineries. A rigorously accurate sampling
procedure would include most, if not all,
of the 256 refineries operating in the U.S.
Such a large sampling plan would be virtu-
ally unmanageable as well as being cost-
prohibitive. Therefore, a number of
representative refineries were selected
for sampling.
The criteria for refinery selection
were age, size and location.
Age was a particularly difficult
variable to utilize. After a consideration
of the complexities introduced by using
either the age of the entire refinery, the
age of individual process units or the age
of individual fittings, the choice was
arbitrarily made to examine refineries
newer than 10 years and older than 20 years
(new and old respectively).
Size was used as a selection criteria
in the anticipation that it would have an
effect on the number and type of products
manufactured, the number and type of
hazardous species formed, the types of
units available for sampling, the amount
of effort put into maintenance programs,
and the quality of equipment purchased.
An arbitrary decision was made to use
8,000 m'/day as the dividing line. There
are 56 refineries processing more than
8,000 m'/day of crude and 200 refineries
of less capacity.
The effect of location is primarily
one of availability of different crude oils
and consequently the opportunity to form a
different slate of potentially hazardous
compounds.
The final choice of refineries was
thus to be four in each geographical area
of the country, i.e.. East Coast, West
Coast, Gulf Coast and Mid-Continent. Each
set of four would be comprised of one new/
large, one old/small and two old/large.
148
-------
Process Units to be
Temperature and pressure were thought
to have major effects upon fugitive emis-
sions from a refinery. They are, at least,
available and useful variables and are
currently treated as choice parameters.
In a refinery, there can be found as many
combinations of temperature and pressure
as one wishes to define. For the purposes
of this sampling plan, four pressure/
temperature classifications were employed:
• high pressure/high temperature ,
• low pressure/high temperature,
• high pressure/low temperature, and
• low pressure/low temperature.
These terms are defined as follows:
• pressure -
. high - >1000 kPa,
. low - <1000 kPa, and
• temperature -
. high - >100»C, and
. low - <100"C.
Several process units fall within each
category of pressure/temperature. A
single unit rarely has each size, type
and service that has been defined for each
piece of hardware. In most cases, several
units must be sampled to completely fill
the required variable categories. In the
event that all of the desired equipment
categories do exist on one process unit,
samples are still taken from as many other
units as possible. This serves two pur-
poses . Bias toward a particular unit is
eliminted, generating data more represen-
tative of the pressure/ temperature class
as a whole. Also, differences between
various units can be noted when the data
are analyzed. There may not be enough
data to draw firm conclusions about charac-
teristic process emissions, but it will
give indications of the factors that may
be significant.
It should be noted that the four
pressure/temperature classifications are
simply guidelines by which units most
likely to have hardware in service at the
desired conditions can be identified for
sampling. High pressure and temperature
units also employ valves, fittings, etc.
that fall into the lower pressure and
temperature ranges. These are not ignored,
but are also sampled to fill appropriate
choice variable categories.
t
Statistical Approach
The Value of a Statistical Approach
Seven variables have been selected as
potentially important for describing the
leakage of baggable devices in refineries.
The variables include device type and size,
temperature and pressure of operation, gas
or liquid service, and refinery location,
size and age. Even after dividing the
range of each variable into a minimal
number of subcategories, about 8,000
samples would be required to measure the
leakage from each combination of the
variables just once. Obtaining a statis-
tically significant sampling for each
combination of the variables would require
several times as many measurements.
Fortunately, sampling plans utilizing
fewer measurements can be used to determine
leakage rates within the desired accura-
cies. By assuming that complicated inter-
actions between the variables are
unimportant, the number of necessary
measurements is reduced considerably.
Factorial experimental design procedures
are used to select test combinations of
the variables so that the effect of each
variable on the leakage rate can be deter-
mined. Analysis of variance is used to
determine which variables significantly
affect the leakage rate. In our case,
both the magnitude of leakage rates and
confidence intervals for the estimates are
calculated. When appropriate, regression
analysis will be used to estimate trends.
Tests of the validity of the assumption
that higher order interactions are negli-
gible will indicate areas where more
measurements are required.
The value of statistical experimental
design is that it provides a systematic
and orderly procedure for selecting a
specific set of measurements for a sampling
program. The design is based on assumption
about the probability distributions of
errors, independence of effects of differ-
ent variables, and the insignificance of
multivariable interactions. As the data
begin to come in, the validity of our
assumptions will be tested and adjustments
made to the sampling plan.
Structured flexibility forms the tone
of the sampling plan. The structure
assures that all preconcemed measurement
and analysis requirements are efficiently
covered. Flexibility is maintained within
a procedural framework to apply what is
learned toward subsequent sampling and
analysis.
Factorial Experiment Design
Standard statistical techniques have
been developed for handling large numbers
of variables, or "factors". The methods
are especially useful when only a few
values of each variable need to be consid-
ered. Some important properties of the
factorial design for this program are
described below.
First, a list of the variables (fac-
tors) which were expected to most strongly
influence the dependent variable was
prepared. In this case, the dependent
variable was the leakage rate for each
type of equipment. A number of levels, or
149
-------
ranges of values, were selected for each
factor. The decision was based on a trade-
off between precision of results (many
levels) and economy (few levels per
variable). The preliminary selections
for independent variables and their
numbers of levels were:
f 1. four geographical
locations
2. three age/size
categories:
new/big
. old/little
. old/big
3. four temperature/pressure
combinations
4. two types of service (gas/
liquid)
.5. three sizes
Choice of
Refinery
Operating
Conditions
Device
Types
6. eight device types:
in-line block valves
in-line control valves
open-end valves
flanges
pressure-relief devices
pump seals
compressor seals
unit drains
Each sample was identified.by a set
of levels, one for each factor. For
example, a measurement could be obtained
for the third location, largest size, old
category, second temperature/pressure
category, etc.
Factorial experiment design is a
systematic procedure for selecting a
balanced set of experiments. All levels
for each variable are tested an equal
number of times, with evenly distributed
values for the other variables. The
result is maximum efficiency and the same
data set is equally suitable for deter-
mining the effect of all factors.
Fractional replication refers to
balanced subsets of the complete design
which can be analyzed for the effect of
each variable. The ability to analyze
for higher order interactions between the
variables is lost. In this program,
fractional replications are used to obtain
an overview of the results after a limited
number of samples have been obtained. An
estimate of the effect of each factor on
the leakage rate is available, however,
the accuracy of the estimates may still be
low because of the small number of samples.
Radian felt that a sampling plan
which adapted according to the early
results would be the most efficient. The
proposed procedure covers all possibilities
by first performing a fractional replica-
tion and analyzing for the effects of all
factors on the leakage rate. At this
point, some types of devices or variables
will be found to be adequately character-
ized. This could be because their contri-
bution to total emitted hydrocarbons has
been found to be small, or possibly
because some factor has a small effect on
the amount of emissions.
Subsequent sampling plans incorporate
early data and are designed to focus on
those remaining factors which are most
important for understanding the total
hydrocarbon emissions situation for a
"typical" process unit. The order of
experiments is selected so that fractional
replications of variable sets will be
completed at regular intervals. The
results are analyzed at these times.
Currently the sampling program has reached
a decision point, and is in the process of
undergoing a sampling plan modification so
that the project goal of identifying the
emission properties of equipment types and
typical process units is most efficiently
and thoroughly accomplished.
Number of Required Samples
As discussed above, the experimental
design procedure which most efficiently
utilizes the data is adaptive in nature.
Therefore, before sampling begins, the
number of samples necessary to achieve a
given hydrocarbon emissions accuracy cannot
be precisely determined. The procedures
for selecting samples and analyzing the
results will now be described.
The device-type variables for baggable
sources are:
• 48 in-line valve categories -
two gas/liquid,
two block/control,
three sizes,
four temperature/pressure
conditions,
• two open-end valve categories -
. sampling/drain,
• 16 flange categories -
two gas/liquid,
two sizes,
four temperature/pressure,
• six pressure-relief device cate-
gories -
two gas/liquid
three temperature/pressure (low
temperature/low pressure
excluded),
• 99 pump-seal categories -
three sizes
150
-------
four temperature/pressure,
, three shaft and packing classes,
three Reid vapor-pressure classes
(for three temperature/pressure
classes),
, two Reid vapor-pressure classes
(for the other temperature/
pressure class),
. two unit drain types, and
. compressor seals (all compressors
will be sampled).
The sampling plan for each of the
first four refineries is listed below.
alone has given sufficient basis for
modifying the sampling plan. We are
obtaining leakage rate estimates which are
also changing the focus of the program.
For example, we can establish an average
leak rate for valves and compressors but
the confidence interval for the compressors
is an order of magnitude larger for the
compressor number and this is primarily due
to the small data set for compressors. If
equivalent confidence intervals are
desired, more compressors will have to be
sampled.
Selection and Screening
The process units to be studied at any
TABLE 2. SAMPLING PLAN
Device Type
In-Line Valves
Open-End Valves
Flanges
Pressure-Relief Devices
Pump Seals
Unit Drains
Compressor Seals
Total Samples
No. of
Replications
4
4
2
8
1*
4
1
Samples Per
Replication
48
2
16
6
189
2
(all)
Total
Samples
192
8
32
48
189
8
(all)
477 + (all
compressors)
*The replications for pump seals have been selected according to the RVP category of the
fluid being pumped. Thus:
• one relicate - RVP<1.5
• two replicates - 1.526
The number of replications per device type
is selected according to expected leakage
rates and measurement accuracies.
For subsequent refineries, approxi-
mately the same number of measurements
will be obtained. The distribution of
samples among the device types at each
stage will depend on what has been learned
about:
• leakage rate for each device type,
• precision and accuracy of measure-
ments,
• prevalence of device types in
refineries (this relates leakage to
total emissions), and
• relative toxicity of leakages.
The total number of baggable source
samples obtained will be about 8,000 (500
each for 16 refineries).
To date, we have obtained enough data
to begin to predict the precision and
accuracy of many of the measurements. This
given refinery are chosen such that each
unit will be studied several times during
the program with more frequently used
processes being studied more often than
rarely-found or outdated processes.
Generally, eight process units are chosen
at each refinery (when a sufficient number
exists). The number of fittings to be
screened is then divided equally among the
chosen units.
The specific fittings to be screened
are then chosen from the Piping and Instru-
mentation Diagrams (P&ID's). The fittings
are chosen so as to fit the device-type
variables required by the experimental
design. Only after all fittings from a
given unit have been selected, does the
"screener" physically enter the process
unit. The fittings are screened with a
hydrocarbon sniffer with a sensitivity in
the ppm range.
The "screener" takes hydrocarbon level
readings at four equidistant points on the
circumference of the fitting in question.
If any of the points gives a reading of 200
ppm hydrocarbon or greater, the fitting is
tagged for sampling. The 200 ppm level
151
-------
corresponds approximately to a leak rate contributions of Mr. C. E. Rlese and the
of 150 ml/hr of hexane vapor. This cut- Radian staff are gratefully acknowledged.
off was arbitrarily chosen but it is
approximately 10~' Ib/hr of hydrocarbon
which would constitute a small fraction of
the average leak found in the 1958 Cali-
fornia study.
All the selected fittings are
screened, provided they still exit (P&ID's
can be very much out of date). All fit-
tings leaking at a rate above the cut-off
point are sampled regardless of their
location. Only in rare cases is a sampling
point considered inaccessible. Points as
high as 50 meters have been successfully
sampled as well as fittings over 300°C and
1.7 meters in diameter.
The point of screening is at the point
of entry of a valve or pump shaft into the
packing or seal; for flanges the entire
circumference is "sniffed ; for drains the
sniffer probe is held just above the rim
and at every two inches on the circumfer-
ence of the drain opening; and, for relief
fittings, the sniffer probe is held Just
inside the horn or against the valve seat
(depending upon design).
The "screener" records as much infor-
mation about each fitting as he can obtain.
This includes exact size, temperature,
pressure, make and model, function, type
of hydrocarbon service, etc. He then
attempts to learn from the refinery person-
nel as much as he can about the nature of
packings, frequency of maintenance, type
of maintenance, age of fitting and of the
packing or seal.
Within eight hours of the time that a
fitting is sampled, it is rescreened. At
that time, the screener obtains a second
set of readings taken at the surface and
at a distance of 5 cm from the surface.
To date neither method of screening has
shown itself to be superior in being
correlated to the subsequently measured
leak rate. The surface contact method,
however, is at least an order of magnitude
more sensitive than the other approach.
Radian is confident that usable
correlation factors will result from both
screening approaches. The importance of
obtaining a high correlation cannot be
overstated because of the necessity of
finding an economical method of estimating
the total amount of fugitive hydrocarbon.
The use of a portable hydrocarbon sniffer
may be compared, in this case, to a Level
I approach.
Data from this program will begin to
be available in April or May of 1978 and
the final data should be released about
April or May 1979.
Acknowledgements
This work is being supported by the
U. S. EPA under Contract Numbers 68-02-
2147, Exhibit B and 68-02-2665. The
152
-------
TOXICITY OF SECONDARY EFFLUENTS FROM TEXTILE PLANTS
Gary D. Rawlings, Ph.D.
Senior Research Engineer
Monsanto Research Corporation
Dayton, Ohio
and
Max Samfield, Ph.D.
Chemical Processes Branch
U.S. Environmental Protection Agency
Industrial Environmental Research Laboratory
Research Triangle Park, North Carolina
Abstract
The purpose of this study was to provide
chemical and toxicological baseline data
on wastewater samples collected from 32
textile plants in the United States. Raw
waste and secondary effluent wastewater
samples were analyzed for 129 consent
decree priority pollutants, effluent
guidelines criteria pollutants, and nutri-
ents; Level 1 chemical analyses were also
performed. Secondary effluent samples
from the 23 plants were submitted for the
following bioassays: mutagenicity, cyto-
toxicity, clonal assay, freshwater ecology
series (fathead minnows, Daphnia, and
algae), marine ecology series (sheepshead
minnows, grass shrimp, and algae), 14-day
rat acute toxicity, and soil microcosm.
Based on the bioassay results, 10 of the
23 textile plants were found to have
secondary effluents sufficiently toxic to
proceed to a second phase of the study.
In the second phase, samples will be
collected from these 10 plants to deter-
mine the level of toxicity removal attain-
ed by selected tertiary treatment
technologies.
Scope of Work
Background
To understand the nature and purpose of
the textile wastewater toxics program it
is first necessary to briefly review the
events which formed the study's foundation.
The principal event occurred on 1 October
1974 when the American Textile Manufac-
turers Institute (ATMI) filed a petition
with the U.S. Fourth Circuit Court of
Appeals asking for review of the 1983
effluent guidelines proposed for the
textile industry. ATMI's grounds for the
suit were that the best available technol-
ogy economically achievable (BATEA) had
not been demonstrated for the textile
industry. As a result, ATMI and EPA filed
a joint motion for delay of the petition,
stating that additional information would
be developed through a cooperative study
by ATMI and EPA (IERL/RTP).
The objective of this ATMI/EPA Grant Study
was to gather enough technical and econom-
ic data to determine what is the BATEA for
removing criteria pollutants from textile
wastewaters. Criteria pollutants for the
textile industry include 5-day biochemical
oxygen demand (BOD5), chemical oxygen
demand (COD), color, sulfide, pH, chromium,
phenol, and total suspended solids
(TSS). On 3 January 1975 the court in-
structed ATMI and EPA to proceed as prompt-
ly as feasible to a completion and review
of the study.
The ATMI/EPA Grant Study was divided into
two phases: Phase I, to determine the
best available technology, and Phase II,
to determine which technology(s) was
economically achievable. A generalized
program outline of Phase I is shown in
Figure 1. To evaluate the best available
technology, two mobile pilot plants were
constructed by ATMI. This strategy allow-
ed for real-time, flowthrough treatment
studies. Each pilot plant contained six
tertiary wastewater treatment unit opera-
tions; one was scheduled to visit 12
textile plants and the other to visit 11.
c
PIMSCI
COMSTWCI no nun PUNIS
OMIAIMWtTBmMV
TKATMBfl WIT OfOUDWS
COUKT ROD MTA FOt 4 KSXS
-------
Treatment operations in each mobile unit
include a reactor/clarifier (using combi-
nations of alum, lime, ferric chloride,
and anionic and cationic polyelectrolytes),
two multimedia filters, three granular
activated carbon columns, dissolved air
flotation, and ozonation. Powdered acti-
vated carbon treatability tests were
performed in the laboratory instead of in
the field with the pilot plant. Using
these six unit operations ATMI and EPA
selected seven treatment systems for
evaluation (Figure 2).
MODE A!
MOOEB:
MOOEC:
MOOED:
MODEEi
REACTOR /cum not -
MULTIMEDIA FILTER -
MULTIMEDIA FILTER -
OZONATOR
REACTOR'CURIFIER-
• MULTIMEDIA FILTER
• GRANULAR ACTIVATE) CARBON COLUMNS
•OZONATOR
MULTIMEDIA FILTER—GRANULAR ACTIVATED
CARBON—OZONATOR
(OPTIONAL)
MODEF: COAGULATION —MULTIMEDIA FILTER
MOOEG: DISSOLVED AIR FLOTATION
Figure 2. Seven tertiary treatment modes
for "best available technology"
evaluation.
TABLE 1. CHEMICAL COMPOUNDS AS
LISTED IN THE CONSENT
DECREE
Acenaphthene
Acrolein
Acrylonitrile
Aldrin/Dieldrin
Antimony and compound*
Arsenic and coepound*
Aabeetos
Bentene
Beniidine
Beryllium and compound*
Cadmium and compounds
Carbon tetrachlorlde
Chlordana (technical mixture
and metabolite*)
Chlorinated benzene* (other than
dichlorobensene*)
Chlorinated ethane* (including
1,2-dichloroethane, 1,1,1-trl-
chloroethene, and hexachloro—
ethane)
Chloroalkyl ether* (chloromitnyl,
chloroethyl, and mixed ether*)
Chlorinated naphthalene
Chlorinated phenol* (other than
those listed elsewhere) in-
clude* trichlorophenol* and
chlorinated cresols)
Chloroform
2-Chlorophenol
Chromium and coepound*
Copper and compounds
Cyanide•
DOT and metabolites
Dlchlorobeniene* (1,2-,1,3-,
and 1,4-di.chlorobenienes)
Dlchlorobeniidine
Polychlorinated biphanyl* (PCS)
Polynuclear aromatic hydrocar-
bon* (including banianthra-
oenes, beaiopyrenei, benio-
fluoranthene, chrysenes,
dlbensanthrscenes, and
indenopyrenes)
Selenium and compounds
Silver and compounds
2,3,7,6-Tetrachlorodibenio-p-
dioxin (TCDD)
Tetrachloroethylene
Thallium and compound*
Toluene
Toxapbene
Trlchloroethylene
Vinyl chloride
lino and compounds
Each of the seven systems was to be set
up, and operational and pollutant data
were to be collected over a 2-day to 3-day
period. Based on that data, the "best"
system was to be selected and set up for 2
weeks of operational evaluation. These
data were then to be forwarded to Phase II
for economic evaluation.
The second event that formed the founda-
tion for this project occurred when a
group of environmental action organiza-
tions filed a class action suit against
EPA (Natural Resources Defense Council et
al. v. Train, U.S. District Court of
Washington, D.C.) for not developing and
promulgating regulations establishing
effluent limitations and guidelines and
new source performance standards for 21
industrial point sources, including the
textile industry. As a result, on 7 June
1976 the court issued a consent decree
requiring EPA to enhance development of
effluent standards.
The most significant result from the court
mandate was that it focused federal water
pollution control efforts on potentially
toxic and hazardous pollutants. The
original consent decree required that 38
classes of chemical compounds (Table 1) be
analyzed in wastewater samples. Recogniz-
ing the difficulty of analyzing for all
chemical species present in each category
of compounds, EPA developed a list of 129
specific compounds (Appendix A) represen-
tative of the classes of compounds listed
in the consent decree. These compounds
are referred to as the consent decree
priority pollutants, or priority pollut-
ants for short.
EPA also developed a sampling and analyti-
cal procedures manual to be used as a
laboratory guide for the analysis of
priority pollutants (*). The analytical
methods recommended by EPA are still in
the developmental phase and require fur-
ther verification and validation.
Therefore, in addition to evaluating the
removal of criteria pollutants by tertiary
treatment technologies, EPA was charged
with the task of evaluating the removals
of toxicity and priority pollutants by the
treatment systems.
The final event which influenced the
formation of the present program was the
three-phase sampling and analytical strat-
egy for environmental assessment developed
by EPA, Process Measurements Branch,
IERL/RTP. The purpose of the assessment
procedure was to determine in a stepwise
and cost-effective manner all chemical
species being discharged to the environ-
ment from a point source. Level 1, the
first part of the three-phase approach, is
designed to focus available resources on
emissions that have a high potential for
causing measurable health or ecological
effects.
The second phase, Level 2, has as its goal
the identification and quantification of
specific compounds. Level 3 is designed
to continuously monitor indicator com-
pounds as surrogates for a large number of
specific pollutants. At the start of this
textile project, only Level 1 analytical
and biological procedures were available!2)
154
-------
In addition to chemical analyses, the
Level 1 recommended protocol included
bioassay testing procedures for evaluating
toxicity removal by control technolo-
gies (3). Bioassays are required to
provide direct evidence of complex bio-
logical effects such as synergism, antag-
onism, and bioavailability.
Program Objective
The fundamental objective of the textile
wastewaters program conducted by MRC in
conjunction with the EPA is to determine
the reduction in toxicity and priority
pollutant concentrations achieved by the
tertiary treatment technologies under
investigation in the ATMI/EPA Grant Study.
The latter study focuses directly on the
treatability of criteria pollutants. Thus
the overall textile program consists of
two separate projects, each with different
activities, running parallel in time, but
converging towards the same goal: deter-
mination of the best available technology
economically achievable for removing
pollutants from textile wastewaters
(Figure 3).
Figure 3.
Overall program
approach to
determine BATEA.
Figure 4. Program outlined for Phase I of the
MRC/EPA wastewater toxicity study.
c
moot wsiBuia KBICITY
PHtKlli
mem aeva §» in srnaa
eoufcr SMVU m nc nun HAJH
MIM« 1W HER lAt ttMUAINI fOMO
won nuui naut MUIAIXJ*
Figure 5.
Program outline for Phase II
of the MRC/EPA wastewater
toxicity study.
To evaluate the reduction in toxicity in a
cost-effective manner for the MRC/EPA
project, a two-phase approach was devel-
oped. Phase I was designed to collect
baseline toxicity data on secondary efflu-
ents from 23 selected textile plants and
to rank the plants in descending order of
toxicity (Figure 4). Phase II was design-
ed to determine the level of toxicity
removal attained by the tertiary treatment
systems in the ATMI/EPA Grant Study at
only those plants with relatively high
secondary effluent toxicity (Figure 5).
Sampling locations for Phase II of the
study are shown in Figure 6, and the
strategy used for evaluating control
technologies in terms of toxicity removal
is illustrated in Figure 7.
Figure 6. Sailing location* Cor PhaM II of th«
MHC/KPA mtmatvr toxicity itudy.
BIOASSAY RESULTS
INLET OUTLET
TOXIC SUBSTANCE INTERPRETATION
* CONTROL TECHNOLOGY IS NOT EFFECTIVE
CONTROL TECHNOLOGY IS EFFECTIVE
+ CONTROL TECHNOLOGY IS DETRIMENTAL
CONTROL TECHNOLOGY IS NOT DETRIMENTAL
flour* 7. Interpretation of bioaaiay ta«t rcaulta.
155
-------
Project Organization
The major effort in Phase I of the MRC/EPA
study was devoted to the collection,
chemical analysis, and biological toxicity
testing of wastewater samples from the 23
textile plants scheduled for testing in
the ATMI/EPA Grant Study. In addition,
samples were collected from nine other
textile plants for chemical analyses only.
Wastewater characterization data were
therefore assembled for a total of 32
plants.
The scope of work for Phase I was divided
into three separate task areas, each based
on different EPA data requirements, as
shown in Figure 8. CPB (Chemical Proces-
ses Branch, IERL/RTP, Project Officer,
M. Samfield) requested chemical and bio-
assay data on secondary effluent samples
from the 23 textile plants scheduled to be
studied in the ATMI/EPA Grant Study.
These data were used to characterize and
compare the relative toxicities of the
plant effluents tested. EGD (Effluent
Guidelines Division, EPA, Washington, D.C.,
J. D. Gallup) requested chemical analyses
of the raw waste streams entering the 23
wastewater treatment plants, as well as
chemical analyses of the raw waste and
secondary effluent streams at 9 additional
textile plants. Raw waste and effluent
data were needed to evaluate the pollutant
removal efficiencies of current state-of-
the-art secondary treatment systems. In
order to implement the entire Level 1
Environmental Assessment Protocol, PMB
(Process Measurements Branch, IERL/RTP,
L. D. Johnson) requested that Level 1
chemical characterization also be perform-
ed on the effluent samples at the basic 23
textile plants (*). Since these data were
requested after the program began, only 15
textile plants were sampled for Level 1
chemical characterization.
Chemical characterization of wastewater
samples involves four categories of
analysis:
• 129 consent decree priority pollut-
ants analysis (*)
• nutrient analysis C*» 5)
• effluent guidelines criteria
pollutants analysis ("•» 5)
• Level 1 chemical characteriza-
tion (2)
The 129 consent decree priority pollutants
as listed in Appendix A are divided into
volatile compounds, nonvolatile compounds,
and metals. The nutrient series required
to support algal tests includes analysis
of nitrite, nitrate, ammonia, total
Kjeldahl nitrogen (TKN), o-phosphate,
phosphorus, and total organic carbon
(TOC). Effluent guidelines criteria
pollutants include 5-day biochemical
oxygen demand (BOD5), chemical oxygen
demand (COD), sulfide, color, pH, total
suspended solids (TSS), total dissolved
solids (TDS), and total phenol.
The bioassay scheme established by EPA for
evaluating the reduction in toxicity of
water samples is shown in Figure 9. All
the tests shown were used in this project.
The marine ecology series and the soil
microcosm tests were requested after the
project began. These data were obtained
from 15 textile plants as opposed to the
basic 23 plants.
Figure 9.
EPA-reconnended bioassay testing scheme
for toxicity analysis of water samples.
UVU OF IFFOIT MQOISTII IV:
-anaoa CUIMUW nvnm e»i
USMSCKM. » C
OBUCM FMCOSO suKMicni
neusiHM. tocos Kvnm. ra.
BKMOI nusBi««. asm CMOUW
uomtUL moss Mvnm. n
SBUKM IMMCU MB. «m CMCUM
MMrzi urn man rot .
• in tnm
MMm onim «•:
• anvju* fouuusn
mucjn *»,
• utnmnrcuuuim
«w.m oiuion *»,
. UVB.I omim mum
00000
cnonacm
miisuiB ions
m Ftmiv rouvttin
ami* MUIIWSH
Figure 8. Scope of work for the analysis of textile plant wastewaters.
156
-------
Figure 10 illustrates the distribution of
samples among the eight EPA and private
laboratories that performed the chemical
analyses and bioassay tests.
MRC collected raw wastes and effluent
samples at 23 of the ATMI/EPA-designated
plants. Wastewater samples were collected
by EPA-Environmental Research Laboratory
(ERL) (Athens, Georgia) at two of the
additional textile plants and sent to MRC
for chemical analysis. Sverdrup and
Parcel and Associates, Inc., (St. Louis,
Missouri) collected the remaining samples
at the additional seven plants and sent
them to MRC for chemical analysis.
Figure 10. Laboratories and persons
involved in sample analysis
of textile plant effluents.
Grab samples and 8-hr continuous samples
were collected both before and after the
wastewater treatment system at each of the
32 plants (Figure 11). Samples were
stored in ice at 4°C and shipped by air
freight to the laboratories for analysis.
Chemical analyses and bioassays were
performed at eight EPA and commercial
laboratories.
TEXTILE PUNT
RAN 1
WASTEWATER©-1
SAWU
t
O
O
0 0
0 0
AERATION LAGOON
SECONDARY
ERUEMT
SAMPLE
Figure 11. Phase
EFRUBff
I sampling locations.
Summary of Results
Chemical Analyses
Analysis for the 129 priority pollutants
in raw waste and secondary effluent sam-
ples (totaling 64 samples) was performed
by Monsanto Research Corporation (MRC).
Analytical procedures followed those
recommended by EPA. However, the recom-
mended analytical protocol for priority
pollutant analysis is still in the devel-
opmental stage and requires further veri-
fication and validation. Consequently,
the analytical results of textile waste-
water samples must be looked upon as good
estimates of which priority pollutants are
present, with concentrations accurate to
within ±100%.
The 129 priority pollutants were divided
into 5 fractions for analysis: volatile
compounds, base/neutral compounds, acid
compounds, pesticides and polychlorinated
biphenyls (PCB), and metals. The EPA
protocol recommended that laboratories not
acquire analytical standards for 2,3,7,8-
tetrachlorodibenzo-p-dioxin (TCDD) because
of its extreme toxicity ('). Asbestos was
omitted due to the presence of other
fibrous materials in textile wastewaters.
A summary of the organic compounds found
in the 32 raw waste and 32 secondary
effluent samples is given in Tables 2 and
3. Of the 114 organic compounds on the
priority pollutant list, a total of 56
different compounds were found, 49 in raw
waste samples and 46 in secondary effluent
samples. The number of compounds found at
each plant is summarized in Table 4. On
an individual plant basis the greatest
number of organic compounds found in a raw
waste and in a secondary effluent sample
were 20 and 14, respectively, with an
average number per plant of 11 in the raw
waste and 9 in the secondary effluent.
The predominant compounds were bis(2-
ethylhexyl) phthalate in 57 samples
(0.5 mg/m3 to 300 mg/m3), toluene in 46
samples (0.4 mg/m3 to 1,400 mg/m3), and
a-hexachlorocyclohexane (a-BHC) in 43
samples (0.02 mg/m3 to 6 mg/m3).
Note that pesticides were not analyzed in
raw waste and secondary effluent samples
from the last seven plants listed in
Table 5. Also, plant R was inadvertently
sampled between the aeration lagoon and
settling pond and is not a typical sec-
ondary effluent sample.
The frequency of occurrence of 56 organic
species in 64 wastewater samples is given
in Table 5. Dominant compounds were
bis(2-ethylhexyl) phthalate found in 56
samples, toluene found in 46 samples, and
a-hexachlorocyclohexane (a-BHC) found in
43 samples.
A summary of the 13 priority pollutant
metals and cyanide concentrations in raw
waste and secondary effluent samples is
157
-------
TABLE 2. SUMMARY OF PRIORITY POLLUTANTS FOUND IN RAW WASTE SAMPLES,
FROM 129 TOTAL, SHOWING CONCENTRATION RANGES AND NUMBER OF
PLANTS WHERE THE SPECIES WERE IDENTIFIED
Viola til* o
c™,
TDlUMtt
B«nz*M
Chloroform
•tbyUmuM
Trau-l , 3 , -diehloroctliylMM
1 . 2-DlchloropropwM
Cil-1 . J-dlchlaroprop«M
rganlc
Of tlMS
faaod
22
4
12
6
20
2
£
B
1
2
1
1
CaacntxttloB
rang.. m,/m*
2 to 30O
t to 200
2 to 500
1 to 300
0.7 to 3,800
30 to SO
2 to 300
6 to 2 100
2
O.C to 4
2
2
•taiMy^Miiiti^Bl orflBf
ft^M
Of til
CD^VDUBQ Count
•hpfathftltUM 11
ttMthyl pfathAlat* !
Dl*tfayl phthalat* U
Bi* (2-«thylb«xyl) 21
phthaUt*
PymM
V?vnr^t}MiM
Di-n-batyl phthalat*
FlnowM
HauchlorobmMM
2 ,6-Dlnitrotolara*
ma*no(1.2.3-cd>pymM
ir
IM Concentration
I ma.. M/ta>
0.03 to 300
3 to 110
0.2 to 70
> O.S to 300
1 to 210
30 to 440
0.1
0.9
9 to 270
2 to 60
5 to 15
0 5 to 2
u
I 50
I 2
Acid organic
PoBOOl
Mntachloropbuol
2-Utrophanol
p-Chlutu • ciool'
4HUtxopbraol
2,4. 6-*r tchloropb«nol
2-Chlorophuol
19
8
1
2
1
2
1
o.s to 100
2 to 70
70
s to e
70
0.7 to 20
130
0-BBC
ft-nc
Y-BBC
S-BBC
Baptachlor
a-xndonlfan
B-BDdocalfan
Aldrin
Dlcldrln
tndrln
4,4'-DOT
4.4'HXO
Biptechlar •poxlo*
17
e
11
e
7
11
7
9
14
6
2
7
4
0.02 to
0.03 to
0.03 to
0.07 to
0.03 to
0.1 to
0.1 to
0.1 to
0.2 to
1 to
0.07 to
0.1 to
0.2 to I
.2
t
).4
TABLE 3. SUMMARY OF PRIORITY POLLUTANTS FOUND IN EFFLUENT SAMPLES,
FROM 129 TOTAL, SHOWING CONCENTRATION RANGES AND NUMBER
OF PLANTS WHERE THE SPECIES WERE IDENTIFIED
Baaa/hanitral organic
Concentration
range, mq/m3
Chloroform
CtilOTobtnsdM
BtbyllMOMM
Trichloro*thyl«iM
1 , 1 , 2 . 2-TBtrBchloroatby IMO»
, 3-dichloroprop«M
24
2
S
2
9
6
2
3
1
2
1
0.4 to 1,400
O.S to 60
7 to 60
4 to 30
0.7 to 3,000
2 to 2,100
5 to 80
0.4 to 40
6
0.9 to 4
2
Diarthyl phthalata
DUthyl pbthalata
Bi« (2-«thylh«eyl)phth«l«t«
1,4-DidilorobunwM
1,2,4-TridilorobenMn*
1.2-Diehlc
30
Anthraocn*
»-Mitroao-di-n-propyla«in«
Pyxvn*
W.-n-batyl pfathalata
Butylboncyl phthalat*
0.5 to 250
0.2 to 1
0.5 to 10
2 to 230
0.05 to 2
2 to 920
0.2 to 6
4
2 to 20
0.1 to 0.3
0.5 to 2
4 to 60
0.3 to 0.8
70
Acid organic Pesticides
Phenol
Pentachlorophenol
2 , 4-Diavthy Ipbanol
D Jjltllllll • l latHalll
Chloro creaol
2-Chlorophanol
2
1
2
«
,
1
1
2 to 3
60
8 to 9
2
20
30
10
0-BBC
&-BBC
Y-BBC
6— BBC
a-Bnacwttlfan
6-Kndoaulfan
Aldrin
DUldrin
Bxlrin
4,4'-COT
4,4'-DOO
4,4>-DGB
Haptachlor •poxide
26
14
18
18
8
11
14
5
9
3
4
2
0.07 to 6
0.1 to 3
0.1 to 2
On£ +A n <
• UO CO U* 3
0.04 to 2
0.2 to 1
0.1 to 2
0.2 to 1
0.2 to 1
0.1 to 2
0.2 to 1
0.1 to 0.4
0.3
158
-------
TABLE 4. NUMBER OP PRIORITY ORGANIC POLLUTANTS FOUND IN
THE RAH HASTE AND SECONDARY EFFLUENT STREAMS
Plant
A
B
C
D
B
P
G
B
J
K
L
M
N
P
R
S
T
U
V
H
X
Y
Z
C-001
Y-001
jjd
KK^
T.T.d
MMd
(jjjd
oo
-------
TABLE 5. OCCURRENCE OF PRIORITY ORGANIC POLLUTANTS COMBINED
FROM RAW WASTE AND SECONDARY EFFLUENT SAMPLES
Priority pollutant
Bla(2-athylhaxyl) phthalata
Toluene
o-BHC
Y-BHC
Bthylbaniana
a-Bndo«ulfan
Dialdrin
Naphthalana
6-BHC
Diathyl phthalata
Phanol
Aldrin
Chloroform
Haptachlor
B-Endoaulfan
1,2, 4-Trichloroban»ana
1 , 2-Dichlorobanzana
1.1,2, 2-Tatrachloroathylana
Endrin
4-BHC
4, 4 '-DOT
4, 4 '-ODD
Acanaphthana
Oi-n-batyl phthalata
Pantachlorophanol
Tr ichloroathy lana
Dimethyl phthalata
1 , 4 -Dichlorobansana
Chlorobanzana
Trichlorof luoromathana
Banzana
1,1, 1-Trlchloroathana
Haptachlor apoxlda
Pyrena
Haxachlorobanzana
4, 4 '-DDE
p-Chloro-»-craaol
2,4, 6-Trichlorophanol
Anthracana
H-nitroeo-di-n-propy lamina
Pluorana
1 , 1-Dichloroathana
Ci » - 1 , 3 -d ichloropropana
fran«-l,3-dichloropropana
2 , 4 -Dimathy Iphanol
2-Chlorophanol
M-nitroaodiphany lamina
2 , 6-Dinitrotoluana
Indano ( 1 , 2 , 3-cd ) pyrana
Butylbanzyl phthalata
T rant-1 , 2-dichloroathylana
1 , 2 -Dichloropropana
2-Hitropbanol
4-Hitrophanol
Chloro craaol
Broaodicbloromathana
*0ut of • total of 64 aaaplaa
toti
51
4(
42
29
2!
2S
21
24
22
21
21
2C
n
K
1!
14
13
11
11
11
1]
H
.
Mabar of aaai
pollutant wai
•aw waata
il aamplaa
27
22
17
11
20
11
14
11
•
12
19
>
12
7
. 7
a
i •
L (
L 6
L •
L 2
) 7
7
6
I
7
5
5
t
2
4
6
4
1
2
0
2
2
1
0
2
2
1
0
0
1
L 1
L 1
L 1
L 0
L 1
L 1
L 1
1 1
L 0
1 0
>!•• in which
i datactad*
Saoondary
affiant (ampla*
10
24
26
ia
9
ia
14
6
14
9
2
11
S
9
a
6
5
3
5
3
9
3
2
3
1
2
3
3
2
6
2
0
2
4
3
4
1
1
1
2
0
0
i
2
2
1
0
0
0
1
0
0
0
0
1
1
Obiarvad
concantration
ranoa," •»/»'
0.5 to 300
0.4 to 1,400
0.02 to 6
0.03 to 4
0.7 to 3,000
0.04 to a
0.2 to 6
0.03 to 300
0.03 to 3
0.2 to 70
0.5 to 100
0.1 to 3
2 to 500
0.03 to 6
0.1 to 4
2 to 900
0.1 to 300
0.4 to 2,100
0.2 to 3
0.06 to 2
0.1 to 2
0.1 to 2
0.5 to 270
2 to 60
2 to 70
2 to 200
0.2 to 110
0.05 to 200
1 to 300
2 to 2,100
0.5 to 200
2 to 300
0.2 to 0.4
0.1 to 0.9
0.3 to 2
0.1 to 0.«
2 to a
0.7 to 20
0.1 to 4
2 to 20
5 to 15
O.C to 4
2 to 6
0.9 to 4
a to 9
10 to 130
11
SO
2
70
2
2
70
70
30
2
Roundad to ona algalfleant figura.
160
-------
given in Table 6, which also summarizes
the criteria pollutant and nutrient con-
centrations for secondary effluent sam-
ples. Nutrient analyses were performed to
support freshwater algae bioassays.
On an individual plant basis it was fre-
quently observed, especially for the
metals data, that the concentration of a
specific pollutant was greater in the
secondary effluent sample than in the raw
waste sample. This phenomenon is due, in
part, to the hydraulic retention time of
the wastewater treatment facility. Since
raw waste and secondary effluent samples
were collected simultaneously, concentra-
tions in the secondary effluent were due
to raw waste loads that entered the treat-
ment system 1 day to 30 days prior to
sampling. The average retention time for
the 32 plants was about 5 days.
Level 1 chemical analyses were performed
on secondary effluent samples from 15 of
the 23 basic textile plants. Level 1
protocol identifies classes of compounds
present in environmental samples and
measures the general concentration range.
Results indicate that total concentration
of methylene chloride extractable organics
ranges from 3 g/m3 to 64 g/m3. This value
is 5 to 10 times lower than the range for
total organic carbon (Table 6).
In the Level 1 procedure each sample was
fractionated by a liquid chromatography
column into eight fractions based on
polarity. Infrared analysis of each
fraction indicated the presence of ali-
phatic hydrocarbons, esters and acids,
aromatic compounds, phthalate esters, and
fatty acid groups. Low resolution mass
spectrophotometric analysis of the eight
fractions of each sample detected the
following types of compounds: paraffinic/
olefinic, alkyl benzenes, alcoholic
ethers, di-n-octyl phthalate, bis(hydroxy-
t-butyl phenol) propane, tri-t-butyl
benzene, alkyl phenols, dichloroaniline,
toluene-sulfonyl groups, vinyl stearate
and azo compounds.
Bioasaay of Secondary Effluents
The primary objective of the entire waste-
water toxicity study is to determine the
level of toxicity removal from secondary
wastewater achieved by the tertiary
treatment technologies selected
for the ATMI/EPA BATEA study. To this
end, the purpose of this screening study
was to provide chemical and toxicological
baseline data on secondary effluents from
the 23 textile plants and to select plants
for the toxicity removal study.
Bioassays used were selected by EPA and
included tests for assessment of both
health and ecological effects (3). Health
effects tests estimated the potential
mutagenicity, potential or presumptive
carcinogenicity, and potential toxicity of
the secondary effluent wastewater samples
to mammalian organisms. Ecological
effects tests focused on the potential
toxicity of samples to vertebrates (fish),
invertebrates (daphnids and shrimp), and
plants (algae) in freshwater, marine, and
terrestrial ecosystems.
Biological testing, as well as chemical
and physical parameters, must be consid-
ered when assessing the potential impact
of industrial or municipal/industrial
wastewaters on the aquatic environment.
Biological testing involves determination
of toxicity for samples of treated ef-
fluents. In a toxicity test, aquatic
organisms will integrate the synergistic
and antagonistic effects of all the ef-
fluent components over the duration of
exposure.
Although toxicity tests with aquatic
organisms can be conducted by applying
wastewater samples directly to the test
organisms, or by injection or feeding,
most tests are conducted by exposing the
test organisms to test solutions contain-
ing various concentrations of effluent
samples. One or more controls are used to
provide a measure of test acceptability by
giving some indication of test organism
health and the suitability of dilution
water, test conditions, handling proce-
dures, etc. A control test is an exposure
of the organisms to dilution water with no
effluent sample added. Bioassay tests are
exposures of test organisms to dilution
water with effluent samples added. Gen-
erally the most important data obtained
from a toxicity test are the percentages
of test organisms that are affected in a
specified way by each concentration of
wastewater sample added. The result
derived from these data is a measure of
the toxicity of the effluent sample to the
test organisms under the test conditions.
Acute toxicity tests are used to determine
the level of toxic agent that produces an
adverse effect on a specified percentage
of test organisms in a short period of
time. The most common acute toxicity test
is the acute mortality test. Experimen-
tally, 50% effect is the most reproducible
measure of the toxicity of a toxic agent
to a group of test organisms, and 96-hr is
often a convenient, reasonably useful
exposure duration. The 96-hr median
lethal concentration (96-hr LCSO) is most
often used with fish and macroinverte-
brates. Thus the acute mortality test is
a statistical estimate of the LC50, which
is the concentration of toxicant in dilu-
tion water that is lethal to 50% of the
test organisms during continuous exposure
for a specified period of time. However,
the 48-hr median effective concentration
(48-hr EC50), based on immobilization, is
most often used with daphnids. The terms
median lethal concentration (LCSO) and
median effective concentration (EC50) are
consistent with the widely used terms
median lethal dose (LD50) and median
effective dose (ED50)» respectively.
161
-------
TABLE 6. SUMMARY OF METAL, CRITERIA POLLUTANT, AND NUTRIENT ANALYSES
Natal
Concentration range. g/»>
Clamant
antimony
AreenIc
Beryllium
Cadmium
Chromium
Copper
cyanide
Lead
Mercury
•ickel
Selenium
Silver
Thalliom
Unc
•av Mate
•ample
0.0005 to 0.06
0.005 to 0.2
<0.0001
0.0005 to 0.05
O.O002 to 0.9
0.0002 to 2.4
0.004 to 0.2
0.001 to 0.2
0.0005 to 0.004
0.01 to 0.2
<0.005
0.005 to 0.1
Figure 12. Laboratoriaa and EPA technical, adviaora involved
in biotaating of effluent aaaplea.
162
-------
A summary of the bioassay results is pre-
sented in Table 8. Toxicity is expressed
as the percent of a secondary effluent
solution that will cause the effect spec-
ified for each bioassay over the testing
period. For the cytotoxicity, Daphnia and
algal bioassays an Effective Concentration
20 or 50 (EC20 or EC50) was calculated.
EC2o for the cytotoxicity test means the
concentration of secondary effluent which
impairs metabolic processes in 20% of the
test cells.
The viability test is a measure of the
cells' ability to survive exposure to the
sample, and the adenosine triphosphate
(ATP) test measures the quantity of the
coenzyme ATP produced, indirectly measur-
ing cellular metabolic activity.
ECso for the algal tests means the con-
centration of secondary effluent which
causes a 50% reduction in algal growth
as compared to a control sample. The
freshwater algae test was performed over
a 14-day period and the marine algae test
over a 96-hr period.
For the fathead minnow, sheepshead minnow,
and grass shrimp bioassays, death was used
to measure toxicity, which was expressed
as Lethal Concentration 50 (LC50). LCso
indicates the calculated concentration of
secondary effluent that is expected to
cause the death of 50% of the test
species. Since rats were given a. spe-
cific quantity of secondary effluent,
toxicity was expressed as Lethal Dose 50
(LOSO). LOSO indicates the quantity of
material fed to the rats that resulted in
the death of 50% of the test animals.
The measure of toxicity to a soil micro-
cosm was the quantity of carbon dioxide
(CO2) produced after sample exposure as
compared to a control sample. The
quantity of CO2 produced over a 3-wk
period after subtracting the quantity
produced by the control was plotted on
graph paper. The slope of the curve then
represented the rate of increase or
decrease in CO2 production due to addition
to the sample.
Plant Ranking by Relative Wastewater
Toxicity
The primary objective of the Phase I
screening study was to rank textile plants
according to the toxicity of their second-
ary wastewater and to select plants for
detailed toxicity evaluation in Phase II.
To accomplish this objective, members of
the EPA Bioassay Subcommittee met to
evaluate the bioassay data. Members of
the Subcommittee are illustrated as EPA
Technical Advisors in Figure 12. A sum-
mary of all the bioassay results is given
in Table 7.
Data evaluation began with ranking of the
plants in each set of bioassays. Results
are discussed in the following sections.
Freshwater Ecology Series
Results from these tests showed sufficient
variation to permit relative ranking of
the toxicity of effluent samples. A com-
posite ranking based on the responses of
fathead minnows and Daphnia is shown in
Table 9. No general rule can be made
concerning the relative response between
fathead minnows and Daphnia. For example.
Plant E's effluent was significantly toxic
to Daphnia but not toxic at all to fathead
minnows; at Plant T, the reverse was true.
TABLE 8. SUMMARY OF BIOTEST DATA FOR SECONDARY EFFLUENT HASTEHATER SAMPLES*'b
VUMllty iff fdmtm
mm* mumm*
ma*
To*
ma
ma
ma
mf
ma
11. J
1 *"
•3
mS
ma
mmjg
00»MO~
on
ma
0.4
UJ
^
4.0
!••
•5
•r
11.7
4.1
ma
— •**•**.
i».o
is'
•t
V
•»
U.I
40.0
00*
40-1
M.O
».l
019
™*
"»oo»o
0.0
41.0
7.4
01.7
41.4
40« OM* 00 1000
'"
gg
10. 0
00.0
1004 OooO «t ill
•.O
U.1
0.4
0.1
^y
^g
41.0
74
\
oj
O1
M
1
0
^1
2
«J
ji
**i
Q*
4|*
u
u.o
•.T
•9
Tf
_f
B^t
Tf
47. >
_f
40.0
Tf
17. »
^tf
.(
"
U.I
2
ma
"3
ji
ma
*}
M.3
f
*f
ma
*s
It. 4
ma.
«J
'
r
~t
5
U to M
M
!>
_f
T7
J'?
1.1
.V
3
M
•0
Jt
_f
"
-0.011
-0.000
-o.oot
-4.040
-O.OM
0.017
^.on
^.10]
-O.004
-O.OM
-O.OSO
O.OM
o.oa
-0.001
-O.M7
0.020
O.OM
-0.044
0.011
0.047
-0.171
-•.111
nto rat •ililllj ofto* 14 Oo?o t
to* -no** aunt. •inn. oi
mmt^Ufl Mr MloO* oOoMO. I
I «f 10-0 mt/t,
O. C, F, U O. «4 II «MO ^ .,.!„,,
I -•«• > i ii ill 11 10 ot» unuu_. •*
4S04 !• 1OM r-'f^'— o(
ODx nlo»lTTI-ni
tto oM.Urmi«i cto tmml* *i« MM
163
-------
TABLE 9. RELATIVE TOXICITY RANKING
BY BIOASSAY TEST
.*• CflalaaKIKy
Marine Ecology Series
Based on toxicity data for sheepshead
minnows, grass shrimp, and marine algae,
ranking of effluents by toxicity was
accomplished and is shown in Table 9. In
all samples, grass shrimp were more sensi-
tive than sheepshead minnows. Also, the
fathead minnows were more sensitive in all
samples than sheepshead minnows. No
general correlation was seen between the
response of Daphnia and grass shrimp.
Cytotoxicity
Rabbit alveolar macrophage tests indicated
that none of the samples was highly toxic.
Two samples, N and C, were moderately
toxic and the following seven samples were
slightly toxic: L, W, T, X, A, F, and J.
Only eight samples were tested by MRC
using the clonal toxicity test: D, H, J,
M, P, R, Y, and Z. Of the eight samples,
four showed significant toxicity: D, M,
H, and J.
Mutagenicity
None of the 23 effluent samples produced a
positive response in any of the bacterial
tester strains. The Bioassay Subcommittee
expressed concern that the detection
limits for this bioassay series were not
sensitive enough to detect the presence of
significant concentrations (0.001 to 0.1
g/m3) of chemical mutagens.
Rat Acute Toxicity Tests
No acute toxicity was observed from the
maximum dose (10~5 m3/kg) of rat body
weight) ingested by the rats. However,
six effluent samples showed potential body
weight effects: F, N, C, L, S, and B.
The subcommittee expressed concern about
the detection limits of this test also.
Plant Ranking
Based on all of the above analyses, the
subcommittee ranked the 23 textile plants
in descending order of secondary effluent
toxicity, and results are shown in
Table 10.
TABLE 10. PRIORITIZATION OF TEXTILE
PLANTS BY TOXICITY OF
SECONDARY EFFLUENT
Toxicity ranking
Most toxic
N,A
L,T
C
P,S
Plant
Least toxic
Nontoxic
V,W,R
B,O,E,F,G,H,J,K
M,U,X,Y,Z
A,
From the above list, the subcommittee
recommended that the following nine
textile plants be tested to determine the
removal of toxicity achieved by the
tertiary treatment technologies being
tested in the ATMI/EPA Grant Study: N,
L, T, C, P, S, W, and V. (Plant R was
also recommended for study under Phase II
because its secondary effluent samples
were inadvertently collected prior to the
settling pond.) In addition, they recom-
mended that the freshwater ecology series
be used to measure reduction in wastewater
toxicity by the treatment technologies.
The marine ecology series was not selected
because none of the textile plants dis-
charge wastewater into a marine
environment.
Program Outline for Phase II Study
The objective of the second part of the
textile wastewater toxicity study is to
determine reduction in priority pollutant
concentrations and in acute toxicity as a
result of applying the ATMI/EPA BATEA
tertiary treatment technologies to the
secondary effluent at the 10 textile
plants.
Pilot plants are scheduled to be at each
(10) textile plant for from 6 wk to 8 wk.
For the first 4 wk, seven tertiary treat-
ment systems will be tested to determine
which one provides the best removal of
criteria pollutants. A treatment system
consists of one or more of the six
tertiary treatment technologies. From the
data collected, the "best" system will be
identified. This system will then be set
up and operated at steady-state conditions
for a final period of 2 wk.
For toxicity and priority pollutant
analyses at each plant, 24-hr composited
samples will be collected during the 2 wk
of steady-state operations from the one
system identified as the "best available
technology." Since the tertiary treatment
system will be composed of several of the
six treatment technologies, samples will
be collected before and after each unit
164
-------
operation in the system, resulting in
approximately four samples.
In order to evaluate the reduction in
toxicity and priority pollutant concentra-
tions, 24-hr composited secondary effluent
samples will also be collected. Due to
hydraulic retention time through the pilot
plant, secondary effluent sampling will
lead the tertiary treatment sampling by
the appropriate time for the tertiary
treatment system selected.
A 24-hr composited sample of the intake
water to the textile plant will be col-
lected at each of the 10 plants to eval-
uate the source of priority pollutants in
wastewater samples. Either continuous or
grab samples will be collected depending
upon the sampling conditions around the
intake water facilities. Samples will be
collected for volatile organics, non-
volatile organics, and metals analyses.
Therefore, a total of approximately 6
samples will be collected at each of the
10 plants as illustrated in Table 11.
TABLE 11. SAMPLE SCHEDULE AT
EACH OP THE 10 TEXTILE
PLANTS
Saaple site
Bo. or
pies
Analyse for
Plant intake water
Secondary effluent
Beet tertiary
treatment system
1 129 priority pollutant*
1 129 priority pollutants
and freshwater ecology
•series
4 129 priority pollutant*
and freshwater ecology
series
References
Draft Final Report: Sampling and
Analysis Procedures for Screening of
Industrial Effluents for Priority
Pollutants. U.S. Environmental
Protection Agency, Cincinnati, Ohio,
March 1977. 145 pp.
Hamersma, J. W., S. L. Reynolds, and
R. F. Maddalone. IERL-RTP Procedures
Manual: Level 1 Environmental
Assessment. EPA-600/2-76-160a, U.S.
Environmental Protection Agency,
Research Triangle Park, North Caro-
lina, June 1976. 131 pp.
Draft Final Report: IERL-RTP Pro-
cedures Manual: Level 1 Environmental
Assessment Biological Tests. K. M.
Duke, M. E. Davis, and A. J. Dennis,
eds. Contract 68-02-2138, U.S.
Environmental Protection Agency,
Research Triangle Park, North Caro-
lina, January 1977. 94 pp.
Methods for Chemical Analysis of
Water and Wastes. EPA-625/6-74-
003a, U.S. Environmental Protection
Agency, Cincinnati, Ohio, 1976.
298 pp.
Standard Methods for the Examination
of Water and Wastewater, Fourteenth
Edition. American Public Health
Association, Washington, D.C., 1976.
874 pp.
The freshwater ecology series consists of
bioassay tests on the following three test
organisms: fathead minnows, Daphnia, and
freshwater algae. Five sample fractions
are collected for priority pollutant
analysis: volatile organics, nonvolatile
organics, metals, cyanide, and phenol.
Criteria pollutant analyses will not be
performed since these analyses will be
routinely performed under the ATMI/EPA
Grant Study.
165
-------
Appendix A
Recommended List of Priority Pollutants
TABLE A-l. RECOMMENDED LIST OF PRIORITY POLLUTANTS
Compound name
Acenaphthene
Acrolein
Acrylonitrile
Benzene
Benzidine
Carbon tetrachloride (tetrachloromethane)
Chlorinated benzenes (other than dichlorobenzenes)
Chlorobenzene
1,2,4-Trichlorobenzene
Hexachlorobenzene
Chlorinated ethanes (including 1,2-dichloroethane,
1,1,1-trichloroethane and hexachloroethane)
1,2-Dichloroethane
1,1,1-Trichloroethane
Hexachloroethane
1,1-Dichloroethane
1,1,2-Trichloroethane
1,1,2,2-Tetrachloroethane
Chloroethane
Chloroalkyl ethers (chloromethyl, chloroethyl and
mixed ethers)
Bis(chloromethyl) ether
Bis(2-chloroethyl) ether
2-Chloroethyl vinyl ether (mixed)
Chlorianted naphthalene
2-Chloronaphthalene
Chlorinated phenols (other than those listed elsewhere;
includes trichlorophenols and chlorinated cresols)
2,4,6-Trichlorophenol
p-Chloro-m-cresol (4-chloro-3-methylphenol)
Chloroform (trichloromethane)
2-Chlorophenol
Dichlorobenzenes
1,2-Dichlorobenzene
1,3-Dichlorobenzene
1,4-Dichlorobenzene
D ichlorobenzid ine
3,3'-Dichlorobenzidine
Dichloroethylenes (1,1-dichloroethylene and
1,2-dichloroethylene)
1,1-Dichloroethylene (vinylidine chloride)
l,2-Tran»-dichloroethylene
2,4-Dichloropheno1
Dichloropropane and dichloropropene
(continued)
166
-------
TABLE A-l (continued).
Compound name
1,2-Dichloropropane
1,3-Dichloropropylene
(aia and £rana-l,3-dichloropropene)
2,4-Oimethylphenol
Dlnitrotoluene
2,4-Dinitrotoluene
2,6-Dinitrotoluene
1,2-Diphenylhydrazine
Ethylbenzene
Fluoranthene
Haloethers (other than those listed elsewhere)
4-Chlorophenyl phenyl ether
4-Bromophenyl phenyl ether
Bis(2-chloroisopropyl) ether
Bis(2-chloroethoxy) methane
Halomethanes (other than those listed elsewhere)
Methylene chloride (dichloromethane)
Methyl chloride (chloromethane)
Methyl bromide (bromomethane)
Bromoform (tribromomethane)
0 ichlorobromomethane
Trichlorofluoromethane
Dichlorodifluoromethane
Chlorodibromomethane
Hexachlorobutadiene
Hexachlorocyclopentadiene
Isophorone (3,5,5-trimethyl-2-cyclohexen-l-one)
Naphthalene
Nitrobenzene
Nitrophenols (including 2,4-dinitrophenol
and dinitrocresol)
2-Nitrophenol
4-Nitrophenol
2,4-Dinitrophenol
4,6-Dinitro-o-cresol
Nitrosoamines
N-nitrosodimethylamine
N-nitrosodiphenylaraine
N-nitroso-di-n-propylamine
Penta chlorophenol
Phenol
Phthalate esters
Bis(2-ethylhexyl) phtnalate
Butyl benzyl phthalate
Di-n-butyl phthalate
Diethyl phthalate (continued)
167
-------
TABLE A-l (continued).
Compound name
Dimethyl phthalate
Di-n-octyl phthalate
Polynuclear aromatic hydrocarbons
Benz(a)anthracene (1,2-benzanthracene)
Benzo(a)pyrene (3/4-benzopyrene)
3,4-Benzofluoranthene
Benzo(k)fluoranthene
(11,12-benzofluoranthene)
Chrysene
Acenaphthy1ene
Anthracene
Benzo(ghi)perylene (1,12-benzoperylene)
Fluorene
Phenanthrene
Dibenz(ah)anthracene
(1,2,5,6-dibenzanthracene)
Indeno(1,2,3-cd)pyrene
(2,3-o-phenylenepyrene)
Pyrene
Tetrachloroethylene
Toluene
Tr ichloroethylene
Vinyl chloride (chloroethylene)
Pesticides and metabolites
Aldrin
Dieldrin
Chlorodane (technical mixture and metabolites)
DDT and metabolites
4,4'-DDT
4,4'-DDE (p,p'-DDX)
4,4'-DDD (p,p'-TDE)
Endosulfan and metabolites
a-Endosulfan
8-Endosulfan
Endosulfan sulfate
Endrin and metabolites
Endrin
Endrin aldehyde
Heptachlor and metabolites
Heptachlor
Heptachlor epoxide
Hexachlorocyclohexane
o-BHC
6-BHC
X-BBC (lindane)
6-BHC
(continued)
168
-------
TABLE A-l (continued).
Compound name
Polychlorinated biphenyls (PCB)
PCB-1242 (Arochlor 1242)
PCB-1254 (Arochlor 1254)
PCB-1221 (Arochlor 1221)
PCB-1232 (Arochlor 1232)
PCB-1248 (Arochlor 1248)
PCB-1260 (Arochlor 1260)
PCB-1016 (Arochlor 1016)
Toxaphene
Elements
Antimony (Total)
Arsenic (Total)
Asbestos (Fibrous)
Beryllium (Total)
Cadmium (Total)
Chromium (Total)
Copper (Total)
Cyanide (Total)
Lead (Total)
Mercury (Total)
Nickel (Total)
Selenium (Total)
Silver (Total)
Thallium (Total)
Zinc (Total)
2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD)
169
-------
NONFERROUS METALS PROCESSING
(COPPER REVERBERATORY FURNACES)
by
Richard L. Meek and Grady B. Nichols
Southern Research Institute
2000 Ninth Avenue, South
Birmingham, Alabama 35205
and
EPA Project Officer: John O. Burckle
Environmental Protection Agency
Industrial Environmental Research Laboratory
Cincinnati, Ohio 45268
Abstract
This paper provides information on an application of environmental assessment
testing for a selected number of copper smelters as an example of the nonferrous metal
processing industry. Details are provided regarding the sampling methodology and subse-
quent analysis techniques used to characterize the effluents for hazardous toxic emis-
sions. A discussion is provided regarding the results of the findings in terms of the
quantities of emissions expected and the potential errors of measurement. Some specific
toxic emissions were found to be present in substantial quantities. Also presented is
information regarding the difficulties attendant to sampling smelter effluent streams.
Introduction
The nonferrous metals industry pro-
duces metals, such as aluminum, copper,
lead, zinc, and nickel, by various hydro-
metallurgical and pyrometallurgical tech-
niques. The prime objective is recovery
of the base metal and elimination of
extraneous by-products said wastes. In the
pyrometallurgical systems, the raw mater-
ials are subjected to high-temperature
processes which may involve selective
oxidation, reduction, volatization, melt-
ing and slagging to separate the waste
materials from the valuable nonferrous
metals. The high processing temperatures
result in evolution of large volumes of
off-gases containing large amounts of
particulate matter, much of which is metal-
lic compounds in the "fine particulate"
range. Since many of the nonferrous
minerals are associated with sulfur, the
off-gases frequently contain SO?, SO*, and
sulfuric acid mist along with the particu-
late matter. To further complicate the
problem, at the high processing tempera-
tures involved, some metals such as
arsenic, cadmium, zinc, mercury, and selen-
ium are volatilized and become part of the
pyrometallurgical off-gases. Other
materials such as fluorine and phosphorous
may also be volatilized.
For those off-gases containing large
amounts of SO2, production of by-product
sulfuric acid, liquid SOj, or elemental
sulfur may be technically and economically
feasible, and extensive measures are taken
to remove particulate matter before proces-
sing the gas stream. However, whether the
gas stream is to be processed for by-
product recovery or not, particulate mat-
ter is normally removed by electrostatic
precipitators, fabric filters, and various
types of scrubbers. Depending on the com-
position of the recovered particulate, it
may be either recycled to the process or
disposed of as solid waste.
Within the nonferrous industry, one
cannot safely generalize on the type of
particulate control equipment which should
be used, and many combinations of electro-
static precipitators, fabric filters, and
scrubbers are used depending on the par-
ticular raw material or process require-
ments, the secondary pollution problems
(such as scrubber purges causing waste-
water problems), the type pollutant to be
controlled, and comparative capital and
operating costs. Recognizing the danger in
generalizations, wet electrostatic precipi-
tators are widely used in the aluminum
industry to control emissions from reduc-
tion cells; dry electrostatic precipitators
are used in primary copper smelters on
roasters, reverberatory and electric fur-
naces, and converters; fabric filters are
used on primary lead updraft sinter ma-
chines and blast furnaces; and electro-
static precipitators are used on zinc
roasters and downdraft sinter machines.
There are numerous exceptions to all of
these generalizations; however.
170
-------
Since there are many unknowns re-
garding performance of control devices
on metallurgical off-gases, EPA's Metals
and Inorganic Chemicals Branch initiated
a program to obtain the necessary infor-
mation to evaluate performance of control
devices such as electrostatic precipitators,
Southern Research Institute has been in-
volved in research on electrostatic precip-
itators for a number of years under EPA
sponsorship, and one of our current pro-
grams being sponsored by EPA's Cincinnati
Industrial Environmental Research Labora-
tories is concerned with the evaluation
and control of particulates from nonferrous
smelting operations, with particular empha-
sis on ESP's.
Copper Smelter Tests
The first industry selected for study
under this program was the primary copper
smelting industry. Limited mass and par-
ticle size distribution measurements were
conducted across electrostatic precipi-
tators collecting particulates from the
effluent gas streams from two primary
copper reverberatory furnaces.
The first field test served as an
opportunity to evaluate the particulate
test methods that had been developed and
used extensively on coal-fired utility
applications for use in the nonferrous
metals industry. Since the effluent gas
stream from a copper reverberatory furnace
differs markedly from that of a coal-fired
utility, this evaluation was thought to be
necessary.
The second field test was expanded
significantly from the first test in an
effort to obtain a larger data base for use
in this research program. This test was
also conducted on an electrostatic precipi-
tator operating on a reverberatory furnace
offi-gas.
A simplified flow diagram of a typi-
cal copper smelter is shown in Figure 1.
The raw ore from the mine is beneficiated
to obtain a concentrate which typically
contains 20 to 25 percent copper with
associated iron, sulfur, and minor elements.
The concentrate is usually received at the
smelter as a slurry or wet cake. In a
typical process scheme, the concentrate may
be partially roasted to eliminate some sul-
fur and to provide a dry calcine as feed to
the reverberatory furnace. Autogenous
roasting is usually carried out at about
1150-1200*F, and the off-gases usually con-
tain enough SO2 for efficient conversion to
sulfuric acid. The reverberatory furnace
requires an external fuel source, such as
natural gas, fuel oil, or coal, and is
operated at about 2200-2500"?. The pri-
mary functions of the reverb are to provide
a balanced Cu-Fe-S matte for feed to the
converter while removing extraneous Fe,
SiOj, etc. as slag. Normally, off-gases
from the reverberatory furnace are low in
SOj and are most frequently vented to the
atmosphere after removal of particulate
matter. The converters are also operated
at about 2200-2500*F on a cyclic basis to
obtain a product blister copper and an iron
slag which is normally recycled to the
CONCENTRATES
PRECIPITATES
FLUX
STACK
AIR-
1
FLUOSOLIDS
ROASTER
i
OUSTS
r_l
-»-H*
AIR AND FUEL
ELECTROSTATIC
PRECIPITATORS
WASTE-HEAT
BOILERS
REVEHBEHATORV
FURNACE
SILICA
FLUX '
REVERB SLAQ
DISPOSAL
MATTE
FIERCE-SMITH
CONVERTERS
BLISTER COPPER
TO ANODE FURNACE '
COOLERS
CONVERTER SLAO
.IFOR RECYCLE TO REVERB)
OUST
OUST
ELECTROSTATIC
PRECIPITATORS
y
SO2-fllCH
• OAKS TO
ACID PLANT
Figure 1. Simplified Flow Diagram of Copper Smelter.
171
-------
ROASTED
CHARGE
COMBUSTION
AIR
FUEL
OIL
MATTE SLAG
TAP TAP
HOLES HOLES
ELECTROSTATIC
PRECIPITATOR
(4 CHAMBERS)
Figure 2. Reverberator? Furnace Schematic
reverberatory furnace. Converter off-gases
have highly cyclic flow patterns, but the .
SO2 content is usually high enough for pro-
cessing to acid.
Since off-gases from the reverber-
atory furnaces are most often vented to the
atmosphere, our first field tests were log-
ically focused on ESP's associated with
reverbs. In one case, the feed to the re-
verberatory furnace was a partially roasted
calcine; in the other, the feed was a dried
but unroasted concentrate. Otherwise,
operation of the two units was quite simi-
lar and is illustrated in Figure 2. Both
units were equipped to use either natural
gas or fuel oil, and both have waste-heat
boilers preceding the ESP's.
The mass tests across the ESP's were
conducted with an ASME-type mass train
inserted into the flue and maintained at
near in-line temperatures. This is some-
what different from the EPA Method 5 test,
in which the filter temperature is main-
tained at 250°F. The in-stack filter
method was used to assure that the particu<-
late captured in the mass train actually
entered or passed through the precipitator
as a particulate rather than a gas or con-
densate.
The particle-size distribution was
measured at the inlet and outlet of the
precipitators using cascade inertial im-
pactors, five-stage cyclones, and two real-
time measurement systems. The inertial
systems provide time-integrated size dis-
tributions with time. These test methods
are described in a report prepared for the
Industrial Environmental Research Labora-
tory entitled P/iocedu/iei Manual jo* Ett.c.tx.0-
4£a£tc Pie.c.'ip-ita.to'i Eva.taa.tZon.
Gas analyses were made at the precip-
itator outlet at various intervals during
the test period. Since it was anticipated
that the sulfur oxide content would vary
during reverb operation, samples were taken
immediately before and after charging
periods as well as during quiescent stages
of the operation.
The descriptive parameters for one of
the precipitator installations is given in
Table 1, and the test results are described
in Tables 2 and 3 and Figure 3.
ID'
...I
10° 10*
rumeLi DiAMtnn. xn
Figure 3. Average cumulative inlet end outlet mass loading vs.
particle size. Plant A copper reverberatory furnace.
172
-------
TABLE 1. ELECTROSTATIC PRECIPITATOR DESCRIPTIVE
PARAMETERS, REVERBERATOR* FURNACE FOR PLANT A
Item
Collection electrode area (A) (total-2 ESP)
Inlet set area (power set C)
Outlet set area (power set A)
Outlet set area (power set B)
Collection electrode spacing
Corona electrode diameter (round wire)
Collection electrode dimension
Number of gas passages (total - 2 ESP)
Gas passage length (active)
Volume flow rate design (V)
Design temperature
Design efficiency
Design precipitation rate parameter (w)
Specific collection electrode area (A/V)
English
39744 ft2
19872 ft2
9936 ft2
9936 ft2
9 in.
0.1055 in.
9 ft x
24 ft
46
18 ft
150,000 acfm
600-700"F
96.83%
0.21 ft/sec
265 ft*/
1000 cfm
Metric
3692.4 m*
1846.2 m2
923.0 m2
923.0 m2
0.229 m
2 .7 mm
2.74 m x
7.32 m
5.49 m
70.8 m'/sec
315-371°C
6.5 cm/sec
52 mVm' sec
TABLE 2. MASS CONCENTRATIONS AND EFFICIENCY, PLANT A
Mass Concentration
inlet
mg/DSCM
Impactor Mass Train
1146 1407
641 1304
outlet
mg/DSCM
Impactor Mass Train
41 48
21 41
Efficiency,
%
Impactor
96.4
96.7
Mass Train
96.6
96.8
TABLE 3. SULFUR OXIDE CONCENTRATIONS, PLANT A
Sampling Rate,
1/min
3.2
2.9
2.4
1.9
1.0
Furnace Charge
Cycle
after
before
after
before
after
% By Volume
SO
2
1.0
0.42
0.73
0.63
1.7
SO
i
0.024
0.019
0.018
0.025
0.067
173
-------
The operation of the precipitator was
within design specification, and no major
operating problems were encountered during
the test period. ?l»e overall efficiency
based on impactor and mass-train data
(Table 2) ranged from 96.4 percent to
96.8 percent versus the design efficiency
of 96.8 percent. The fractional effi-
ciency for "fine particulate" showed an
expected characteristic decrease in the
0.5 to 1.0 um range.
As a part of our field study of the
two precipitators operating on copper
reverberatory furnaces, we also investi-
gated the fate of some of the minor ele-
ments of environmental concern. Samples
were collected from the gas streams with a
wet-electrostatic-precipitator train, a
series of external impingers, cyclone
separators, and glass-fiber thimbles. Most
elemental analyses were made using a spark-
source mass spectrometer and atomic absorp-
tion. Fluorine content was determined with
an ion-selective electrode, and selenium
was determined by a fluorometric procedure.
The screening analyses identified the
presence of over fifty elemental species;
however, most of the evaluation was con-
cerned with the 19 elements shown in
Table 4. For the minor elements listed,
only arsenic, fluorine, mercury, and
1 Ib/hr as shown in Table 5. Since fluo-
rine from the reverb exists in some gaseous
form, probably as SiFi,, it would not be
caught by the ESP or a fabric filter, and
most of the fluorine entering the smelter
would probably leave in the off-gas from
the reverb ESP.
TABLE 5: QUALITATIVE DISTRIBUTION OF
ELEMENTS ACROSS REVERB AND
PRECIPITATOR
Inlat F««d
Gnatar than 1000 lb»/hr - Cu, ft. Si
100 to 1000 lb»/hr - Al. Aa, Ca
10 to 100 lh>/hr - Ba, Cd, Mo, Pb, Se, Zn
1 to 10 lb«/hr - e, Sb
Leaa than 1 Ib/hr - B«, Cr, Bg, Hi, V
Outlat Off-Gaaaa
Graatar than 10 lba/hr - As
1 to 10 lba/hr - Cu, P, Si
0.1 to 1 Ib/hr - Al, Ba, Fa, Mo, Pb, Sa, Zn
than 0.1 Ib/hr - Ba, Ca, Cd, Cr, Bg, Hi, Sb, V
TABLE 4: ELEMENTS
FURNACE
IN REVERBERATORY
Aluminum (Al)
Arsenic (As)
Barium (Ba)
Beryllium (Be)
Calcium (Ca)
Cadmium (Cd)
Chromium (Cr)
Copper (Cu)
Fluorine (F)
Iron (Fe)
Mercury (Hg)
Molybdenum (Mo)
Nickel (Ni)
Lead (Pb)
Antimony (Sb)
Selenium (Se)
Silicon (Si)
Vanadium (V)
Zinc (Zn)
perhaps chromium, selenium, antimony, and
molybdenum are present as vapors or ultra-
fine particulate at the ESP outlet for
this particular smelter. The remainder are
either caught by the ESP or leave the pro-
cess as a scrubber-water purge or as slag
waste. For the tests in question, only
arsenic and fluorine were present in the
ESP outlet in significant quantities above
Arsenic, which was the predominant
volatile metal in the feed to the copper
reverberatory furnaces, was partially
caught in the ESP operating above 500°F;
however, a large part of the arsenic re-
mained in the ESP outlet. An approximate
arsenic balance over the entire smelter
system is given in Table 6. For this
particular smelter, a small amount of the
arsenic remains in the copper product and
will eventually be removed in electrolytic
refining of the copper. About half of the
arsenic ends up in the reverberatory-
furnace slag and the acid-plant-purge water,
about equally divided between the two. The
remainder of the arsenic, nearly 50 percent,
is not caught by the hot ESP and is vented
to the atmosphere.
TABLE 6: APPROXIMATE ARSENIC
DISTRIBUTION
Reverb Slag 25 - 30%
Acid Plant Purge Water 25 - 30%
(Converter off-gases)
Blister (or Anode) Copper 5%
Reverb ESP Outlet 40 - 45%
174
-------
Referring back to Figure 1, we can
make a few qualitative observations on
arsenic and some of the other volatile
minor elements that may be found in the
copper-smelting system. At the operating
temperatures of most roasters (1150°-
1200°P), very few of the elements or their
oxides or sulfides are volatile. Some
arsenic, mercury or fluorine could be
evolved but probably in insignificant
amounts. At the operating temperatures
in the reverberatory furnace and converter
(2200-2500°F), a number of metals and
oxides may be volatilized including arse-
nic, antimony, barium, cadmium, lead,
mercury, molybdenum, rhenium, and selenium
in addition to fluorine and phosphorus.
Some of the minor elements such as molyb-
denum end up in the reverberatory furnace
slag, and some may end up in either the
slag or in the acid-plant-purge water.
However, since both the slag from the
reverberatory furnace and the acid-plant-
purge water are waste streams from the
smelter and could constitute environmental
problems if the arsenic and other toxic
metals are not bound in an innocuous form.
The arsenic, fluorine, and other minor
elements leaving the stack may also consti-
tute a potential environmental problem and
health hazard.
Additional research is needed to
develop a better understanding of the minor-
element emissions from nonferrous smelters
and to improve the control systems if ad-
verse environmental and health effects are
found. We do not represent our limited
findings as identifying a widely recurrent
circumstance, and it should be recognized
that the minor-element emissions will vary
widely from one smelter to another. How-
ever, we believe that further investigation
of these minor-element emissions and their
control is warranted.
Acknowledgments
We would like to acknowledge Radian
Corporation for their collaboration with
Southern Research Institute on these
studies, especially Dr. Klaus Switzgebel
and Mr. Robert V. Collins. We also appre-
ciate the cooperation of the management
and personnel at the two copper smelters,
and the support of EPA through Mr. George
S. Thompson, Jr. and our project officers,
Ms. Margaret Stasikowski and Mr. John O.
Burckle.
175
-------
APPLICATION OF THE PHASED APPROACH TO ENVIRONMENTAL ASSESSMENT TO
THE EMISSION ASSESSMENT OF CONVENTIONAL COMBUSTION SERVICES
By
J. Warren Hamersma
Abstract. The EPA program concerning the Emissions
Assessment of Conventional Combustion systems 1s
the first large scale application of the phased ap-
proach for environmental assessment. This program
Includes the evaluation of 170 sites In 50 source
categories grouped in 5 general areas. These are
residential, Internal combustion, electrical genera-
tion, commercial/Institutional and Industrial
sources. The major emphasis 1s on the application
of Level 1 sampling and analysis techniques to this
program and the Integrated decision criteria that
have been developed for the cost effective attain-
ment of the program's objectives. Also Included 1s
a discussion of the program, Its goals, some unique
problems, and a general discussion of some of the
results to date.
• Finally, additional data on source characteris-
tics, Including size and age distribution, mar-
ket trends, current and future comsumptlon, are
being compiled.
The source type sampling plan 1s 1n a constant state
of evolution. In total, 50 source sub-categories
are being considered which can be grouped Into 5
general categories shown in Table 1. The factors
that were considered in preparing this plan Included
grouping of certain source sub-categories, the
latest fuel usage trends, equipment obsolescence,
and general data base quality. For each of the sub-
categories 1n the general groupings, 5 test sites
are considered adequate for determining if addition-
al Level 1 or Level 2 testing will be required.
Introduction. The major objective of this program
1s to assess air and water pollutants generated by
conventional stationary combustion systems Includ-
ing pollutants from related solid waste disposal.
This Includes the Identification and quantification
of both criteria and noncrlterla pollutants using
Level 1 methods as well as acquiring some Level 2
or quasi-Level 2 data as required in the statement
of work. This objective 1s being Implemented 1n
four general steps. First, criteria were developed
to determine the adequacy of existing emissions
data. In doing this, the available data were
screened for definition of process and fuel param-
eters as well as for the validity and accuracy of
the sampling and analysis methods. The acceptable
emissions data were than subjected to an engineer-
ing and statistical analysis. Then after discard-
Ing the outliers, the mean and variability in the
emission factors are calculated. Second, this In-
formation 1s then used in a continuing process to
Identify parts of or whole emission source catego-
ries that have been adequately assessed, and to
specify those categories that require additional
investigation. Finally, a test program has been
developed to complete the emission assessment. In
a sense this process 1s the heart of the program in
that it provides the focus for both the sampling
and analysis effort.
Present Status. The overall status of the planning
and assessment task can be summarized as follows:
• Criteria for assessing the adequacy of the
emissions data base have been developed,
• A source type sampling plan has been prepared,
• A preliminary evaluation of existing emissions
data base has been completed,
• Additional emissions data from recently com-
pleted and on-going projects are being
evaluated,
• Draft reports on emissions from residential oil
and Internal combustion units have been submit-
ted to EPA, and
Sampling. In order to carry out the sampling por-
tion of this program, two complete sampling vans
have been constructed. One operates on the West
Coast and the other on the East Coast. The vans are
divided into two general areas. The forward area
contains an icemaker to partially supply ice for the
SASS train, gas bottles for the gas chromatographs,
and a water purification unit. In addition there 1s
space to assemble and disassemble the SASS train
components. This area is segregated from the rear
part of the van so that dust and contamination car-
ried into the van with the SASS train and other
equipment will not contaminate the sample when 1t 1s
recovered. The rear part of the van 1s maintained
as a clean area and contains bench space for SASS
train- sample recovery, on site analysis, and miscel-
laneous work-up of .the other samples prior to ship-
ment. Two gas chromatographs are used. One has a
thermal conductivity detector and is used for analy-
sis of the Inorganic gases 02, CO, C02, and N2- The
second has a flame lonization detector and is used
to analyze the CI-CA hydrocarbons. In addition,
there is a- laminar flow bench and a fume hood to aid
in on site analysis and sample preparation. A sche-
matic of the van 1s given 1n Figure 1.
In terms of field sampling, there are two general
areas where additional procedural guidelines are
needed. Although operating conditions for the SASS
train are fairly well established, there are no firm
guidelines as to what 1s an acceptable Level 1 sam-
pling port. On a program such as this where a large
number of sources are to be tested, there are not an
unlimited number of sites from which to chose and
some compromises must be made. In addition, sam-
pling after a control device gives a better indica-
tion of actual emissions, while sampling before a
control device probably gives a better Indication of
boiler operating conditions which can affect these
emissions. In addition to this problem, a general
protocol or program specific protocol must be
developed 1n terms of what streams should be sampled
and where they should be sampled. For instance, if
several different units are interconnected, where do
you sample cooling water, ash streams, and other
miscellaneous waste streams? The actual decision
can make an enormous difference 1n the number of
samples and subsequent analysis costs. Presently
these decisions are being made on a case by
case basis.
176
-------
General Decision Criteria. In terms of overall
decision criteria, the planning and assessment anal-
ysis plan 1s the most Important. As a result of
this plan specific sample types or parts of the sam-
pling train will be deleted 1f sufficient data
exists or trill be available.
For instance,
• TVA 1s studying coal pile drainage
• There are separate programs for studying fugi-
tive emissions and ash pound drainage from
utility plants
• Certain water data 1s available and other
water discharges can be calculated.
• Cooling tower emissions have been defined as
Independent of boiler type and are being as-
sessed as a separate task.
Field Sampling and Analysis Decision Criteria. The
general EPA Level 1 guidelines have been supple-
mented based on the general sampling requirements
of the combustion source assessment programs. Many
of the combustion sources are known or projected to
have low gain loadings and the expected cyclone
catchers will be very low. Based on available data,
1t has been possible to estimate 1n a very conserva-
tive manner the >3y and <3y catches as a function
of grain loadings for coal, residual oil, distil-
late oil, gasoline, dlesel fuel, and natural gas
fired units. For coal fired units, the effect of
control devices at lower grain loadings was taken
Into effect. The general curve shown in Figure 2
takes Into account all fuel and control device
variations so that the greater than 3u catch 1s al-
ways conservatively estimated on the high side.
Based on this curve, a series of cyclone deletion
guidelines 1n Table 2 have been developed that al-
lows the deletion of one or more cyclones. This
allows labor saving 1n terms of set-up and sample
recovery time when little or no sample will be col-
lected. This also has the technical advantage of
allowing the collection of samll amounts of sample
that would be lost 1f spread over one or more cy-
clones as well as preventing the concomitant higher
levels of contamination that occur in the recovery
of samll amounts (less than 6 mg) of material 1n
the field. Table 3 shows the expected application
of these guidelines as a function of fuel type. It
should be emphasized that these projections are 1n
the process of being verified but that present data
Indicates that the Information presented 1s applic-
able and sufficiently conservative.
Fugitive emissions assessment 1s normally part of a
Level 1 assessment. However, in the case of combus-
tion sources several studies are now underway to
assess the problem. For this reason, this effort
was dropped because the expected data will supply
the needs of the program. The field crews do carry
detector tube kits for use when obvious problems
exist.
criteria have been supplemented 1n order to limit
the analyses in these samples that will provide new
and useful data. These criteria Involve both inor-
ganic and organic analysis.
The first set of decision criteria involves small
participate catches. If after combining the filter
and cyclone catches into greater than and less than
3y functions, the sample is less than 10 mg no anal-
yses are performed; if it is between 10 and 100 mg,
SSMS analysis only is performed; if it is between
100 and 150 mg, SSMS as well as Hg, Sb, and As anal-
ysis are performed; and if it is greater than 150 mg
organic analysis is added. In many cases, solids
are collected in the SASS train probe rinse. These
are subjected to separate analysis by the above cri-
teria only if this quantity is greater than 10* of
the total catch.
When distillate fuels are used in units tested on
this program, it can be expected that the inorganic
element content can be defined within rather narrow
limits. A careful review of the literature and
actual fuel analysis Indicate that deviations from
site to site are statistically insignificant within
the limits of Level 1 accuracy, thus further analyses
will no longer be performed on the remaining sites
using these fuels.
During the course of the program, the Level 1 organ-
ic analysis scheme was modified in order to allow
for a more nearly accurate analysis of Cy-Cie mate-
rial. Much of this material is lost in the original
Level 1 organic analysis, so the analysis was modi-
fied by lERL's Process Measurements Branch to in-
clude a total chromatographable organic analysis
(TCO) on the bulk samples and all liquid chromato-
graphy (LC) fractions. Because this addition could
add up to eighty additional gas chromatographic
analysis per site, a set of decision criteria
were set up in order to perform only those analysis
that could be expected to yield useful data. Thus
1f the TCO is greater than 10X of the total organics,
the analysis scheme (Method 1, Fig. 3) using sol-
vent exchange and TCO on all LC fractions is follow-
ed. If the TCO 1s less than lOt of the total
organics, the original Level 1 analysis scheme
(Method 2, F1g. 3) is followed. If the total
organics is less than 500 mg/M3, the organic analysis
1s terminated per Level 1 criteria. Because of
the large number of infrared analysis (IR) that
must be performed on this program, lower limits
for IR analysis were adopted to limit the number
of IR analysis that provide no useful information.
Thus IR spectra are not obtained on samples less
than 0.5 mg for the bulk residues and 1 mg for
LC fractions. The above and other basic organic
analysis decision points are summarized in Table 4.
Laboratory Analysis Decision Criteria. If all
Level 1 analyses were attempted on all samples re-
gardless of the expected data output, the program
cannot be accomplished in a cost effective manner.
For this reason, the Level 1 analysis decision
177
-------
Table 1. Source Type Sampling
General Source Category No. of Sites
Residential sources 16
Internal combustion sources 11
Electricity generation external combustion sources 51
Commercial/Institutional sources 53
Industrial sources 30
Total committed sites 161
Total unasslgned sites 1n reserve pool 9
Total Level 1 sites 170
178
-------
TABLE 2. SASS Train Cyclone Use Criteria*
Grain Loading
>o.n
0.051-0.10
0.021-0.050
0.0001-0.020
<0.0001
T
Projected Utch g
>3\i <3n
>.006. >6.9
.004-. 006 3.1-6.9
nil 1.4-3.1
nil 0.007-1.4
nil <.007
Cyclones Required
In 3(i 10(i
Yes Yes Yes
Yes Yes No
Yes No No
No No No
Run XAD-2 module
and 1mp1ngers only
*A filter 1s used 1n all cases.
TABLE 3. Expected Application of SASS Train Cyclone
as a Function of Fuel Type*
Fuel Type
Coal
Residual 011
Distillate 011
Natural Gas
Gasoline
Diesel Fuel
Projected Grain Loading
>0.1
0.051-0.10
0.021-0.05
< 0.020
<0.02
<0.02
<0.02
<0.02
Cyclones Required
l(i 3|i 10)i
Yes
Yes
Yes
No
No
No*
No*
No*
Yes Yes
Yes Yes
No* No
No* No
No* No
No No
No No
No No
a) A filter Is to be used 1n all cases.
*TMs cyclone will be used until field tests show that 1t 1s
not necessary.
179
-------
TABLE 4. Organic Analysis Decision Points
Decision Point
Description
Initial Extraction
Gravimetric Analysis
Total Chroraatographable
Organlcs, Cy-Cie (TCO)
Infrared Analysis
jam Resolution Mass
pectroscopy
Analysis
Organic analysis omitted 1f quantity
of particulate catch 1s limited and/
or meets certain criteria (see
below).
a) Analysis terminated If organlcs
concentration 1s <0.5 ntg/in3
b) Analysis on the XAD-2 condensate
1s discontinued 1f nonvolatile
matter <1M of total organlcs or
<0.5 mg/m3.
Simplified liquid chromatographlc (LC)
separation used 1f TCO <10X total
organlcs.
Analysis not run on less than 0.5 mg
for basic gravimetric residues, or
1 mg for LC fractions.
Run only on total residues or LC
fractions. 1f present 1n concentrations
of>0.5 mg/m3.
Performed only on fractions showing
>0.5 mg/m3 organlcs.
180
-------
00
nm
EXnNOUSNB
ISOLATION
AiiooMnBsot/RMfia TIANSKMMEI RDANOTC
*V 1 INTEOMflNO / OAS CMOMATOOIAM
roots
AN^, (XnNOUSMiS /Z&w
STORAOC
CAHNn
WOtK
KNCH
VCNT
HOOO
ROW
HOOD
STOUOI
13'•
H r
30-
FIGURE 1. Schematic of Environmental Assessment Sampling Van
-------
% CATCH AS
A FUNCTION
OF CHAIN
LOADING
TOTAL WEIGHT OF
CATCH > 3 * IN
GRAMS
TOTAL WEIGHT OF
CATCH < 3 M IN
GRAMS
•—
^™
—
GRAIN
LOADING
^ 100.0
— 99* GRAIh
— 98 LOAD
— 97
— 96
—.95
— 94
—93
— 92
— 91
—90
— 80 1.0 —
— 70 0.9—
—60 0.8 —
— 50 0.7—
—40 0.6—
—30 0.3-
X
-r» 0.4-
-rlO 0.3—
—.5 0.2—
2 0.1—
-.1 0.04 —
CATCH GRAIN
WEIGHT LOADING
I0.04—
1 CATCf* j * *"• I
EXPANDED
.01- -.69 SCALE
.02. -1.4 ^S
.03- -2.1
0.3—
.04J.2.8
-55.1 0.4 —
-43.3 0.5-,
/
— 33.0 j£6—
— 24.1 X 0.7—
—U.s' 0.8—
— 10.3 0.9—
—3.5 1.0—
-2.1
— 0.7
—0.0006
_ 0.003
CATCH
WEIGHT
"6*9
—13.1
-18.5
-22.0
^21.1
-24.8
-24.1
-22.0
-18.6
— 13.8
FI6URE 2. Partlculate Analysis Prediction Nomograph
182
-------
LIQUID
CHROMATOGRAPHY
^
r
I I I I I I I
1234 367
T
TCO «• GRAV + IR
I I I I I I I
1 234567
T
GRAV. + IR
LOW RESOLUTION
MASS SPECTROSCOPY
*GC - GAS CHROMATOGRAPHY; TCO - TOTAL CHROMATOGRAPHAIU
0«GANIC MATERIAL; LC - LIQUID CHROMATOGRAPHY; MS - MASS
SPICTROSCOPY.
FIGURE 3. Organic Analysis Scheme
183
-------
EMISSIONS FROM THE GLASS MANUFACTURING INDUSTRY
C. Darvin, USEPA, Cincinnati, OH
B. Blakeslee, Scott Environmental Services
R. Barrett, Battelle, Columbus, OH
Abstract
The introduction of pollution control technology
into the glass industry has been very slow. This
was primarily due to the lack of understanding of
the nature of the gas streams coming from the
glass manufacturing process. Past assessment
studies of this Industry have indicated a serious
lack of emissions data on this Industry; therefore,
to provide an understanding of these gas streams,
a test program was initiated by lERL-Cincinnati to
characterize emissions from the glass manufacturing
process and particularly from the glass furnaces.
Ten different installations were selected con-
sisting of furnaces in the 200 tons per day range.
The test results indicated that although low in-
opacity, the glass furnace emission stream is of
complex nature and should not be considered free
of potentially dangerous pollutants. SO , HO
and toxic heavy metals such as lead, arsenic,xse-
lenlum and cadmium were found to be present in
the gas stream. Significant portions of the
metals were found in the back half of the Method
5 train Indicating either their vaporous or fine
particulate nature as they passed the filter.
Finally, particulate sice analysis indicated a
significant portion of the potential particulate
matter may be below the minimum size range at
which classical control technology is efficient.
Preliminary conclusions of the test program in-
dicate that additional testing is required to
further define the characteristics of the fine
particulate and condenslble matter. This would
require testing near the furnace exit and In the
regenerators. These areas represent more harsh
testing environments than in the original test
program.
Introduction
The domestic glass manufacturing Industry Is made
up of approximately 1000 plant* producing a vari-
ety of glass products. This Industry was assigned
to the Industrial Environmental Research Labora-
tory to conduct assessment R&D efforts In 1976.
The primary objective of this assessment was to
characterize emissions from the production pro-
cesses and thus, provide a data base to support
future EPA regulatory and control technology
development programs. Initial assessment activ-
ities conducted primarily as paper studies indi-
cated data gapes in available literature and re-
search Information. Therefore, as a continuation
of the Industry assessment program, an emissions
testing program was Initiated to quantify emis-
sions and fill the data gaps identified by assess-
ment studies.
The Glass Industry
The glass Industry is divided Into three major
segments by EPA: flat glass, pressed and blown
glass and container glass. The major production
of glass In the domestic industry Is in container
segment of the industry. Over 70Z of the total
U.S. production Is container glass and Is primar-
ily soda lime glass. Flat glass and pressed and
blown glass make up approximately 19Z and.llZ, re-
spectively, of the total U.S. production. The
glass Industry has approximately 140 container
plants, 260 pressed and blown plants and 75 flat
glass plants. The remainder is made up of small
specialty glass facilities.
Glass types are usually defined by their major
oxide constituent. They include: sodium oxide
and calcium carbonate In soda-lime glasses and
boric oxide in borate glasses. During the manu-
facturing process, raw materials are batched,
mixed and Injected, usually into 4 fossil-fired
furnace. The batch, typically, consists of silica
In the form of sand, cullet (recycled glass), lime-
stone, soda ash, and alumina-bearing materials.
There are, however, numerous other materials which
are used either alone or in combination to produce
the desired glass product. Figure 1 outlines a
typical glass manufacturing process. In terms of
emissions, melting can be considered the most Im-
portant process operation. The furnaces are
typically elevated to temperatures In excess of
1500 C to produce the homogeneous molten glass
mass. It is this process that the major emphasis
of the test program was directed.
The Glass Manufacturing Assessment Test Program
The major objective of any laboratory assessment
program is to develop a qualitative and quantita-
tive data base on pollutants from the affected
Industry. Thus, the glass manufacturing assess-
ment has as a goal the identification and deter-
mination of sources and characteristics of pollu-
tants from glass process operations: this Includes
identification of chemical species as well as quan-
tifying pollutant emission rates and determining
physical and chemical characteristics of the
pollutants.
Nine plant sites involving eleven different fur-
naces were selected as sources of the data. These
sites were selected based upon type of glass pro-
duced, quantity of glass produced and their simi-
larity to typical glass manufacturing plants.
184
-------
Gullet
Batch
Weighing
Mixing
Raw
Materials
Batch
Charging
Batch
Melting
Forming
Annealing
Cooling
Product
Packaging &
Storing
Figure 1 - Process Flow Diagram for Glass
Manufac turIng
The division among plant glass types Include six
producing container glass, two producing pressed
and blown, and one flat glass. Wherever possible,
the tests were conducted using both oil firing and
gas firing. Similarly, wherever possible, varia-
tions in batch composition were tested.
The test program strategy was designed to utilize
those procedures which were most likely to be
used by the EPA regulatory function or by Industry
when evaluating glass emissions in future pro-
grams. Thus, wherever possible, standard pro-
cedures were mandated to be used. In this manner,
it was believed that close correlation between
past and future testing programs would be main-
tained or at least inconsistencies would be
apparent. This latter aspect of the test program
was Important since two different contractors
were selected to perform the various tests. Scott
Environmental Services was selected to conduct
testing at five of the selected sites and Battelle
at four sites.
Emission Sampl^-fj ""> Analysis
The emission sampling and analysis plan was selec-
ted to provide data to fill in gaps in the data
available from the literature and from past emis-
sion measurement programs. Because much was al-
ready known about glass plants, and because of
limitations on the effort available to this pro-
gram, it was determined that the comprehensive
emission measurement program characterized by the
Level 1 SASS train measurements could not be em-
ployed. Hence, specific procedures were selected
to measure emissions thought to be of Importance.
Although all emissions were not measured at each
plant, each of the following emissions was meas-
ured at one or more glass plants:
o Partlculate Mass
o Fluorides
o Trace Metals
o Particle Size
o SOj and SO,
o HO
o CO, HjS, COS, CS., and low-molecular-weight
hydrocarbons
o Organics
Sampling and Analysis Procedures
Partlculate Sampling
A
The EPA Method 5 particulate sampling train (in-
cluding back half or Impinger portion) was used as
a guide for particulate sampling at the glass
plants. A minimum filter temperature of 225 F was
maintained. To reduce background values for trace
metals analysis, Battelle selected tissue quartz
filters and Scott selected acid-washed flberglas
filters.
Time Integrated gas samples were taken during the
particulate runs to determine the CO. and 0. con-
centrations and thus, molecular weigfit of tfie
stack gas needed for particulate emissions calcu-
lations. Analysis by Fyrite was completed at the
end of each run.
The required Method 5 cleanup and analytical pro-
cedures were followed to determine mass particu-
late emissions. Particulate samples were desic-
cated for 24 hours and weights determined by
measurements conducted In a constant temperature/
constant humidity room.
Fluoride Sampling and Analysis
For glass plants using fluoride flux in the glass-
making process, separate runs were made to deter-
mine fluoride emissions. The fluoride sampling
procedure used was that of EPA Method 13b.
Method 13b utilizes the Method 5 sampling train
and procedures, except that CaO Is added to the
water and wash solutions before solvent evaporation
to prevent loss of fluoride during evaporation.
Instead of using the specific ion electrode of
Method 13b to determine fluoride content of the
samples, Battelle utilized the distillation
method of Willard Winkle and determined the
fluoride by ion chromatography.
185
-------
Trace Metals Analysis
Various samples from the EPA Method S particulate
sampling train were analyzed for trace metals by
optical emission apectroscopy. Separate analyses
were made for front-half wash, filter, and back-
half wash.
Particle Site Measurements
The Anderson in-stack impactor was used for parti-
cle size measurements. To facilitate the gravi
metric determinations, a light-weight flberglas
Impaction substrate was placed on each Impaction
plate. The substrate material was preweighed to
the nearest mlcrogram in a constant temperature/
constant humidity environment. A backup filter
was placed at the outlet of the impactor to catch
any material passing through the last Impactor
stage.
The standard EPA Method 5 operating nomograph was
used to determine the appropriate nozzle diameters
and A p* s for isokinetic sampling with the impactor.
After assembling the impactor, the sampling tip
was plugged and the Impactor was placed in the
stack for a sufficient time to equilibrate to
stack temperature. After approximately 10 to 15
minutes, the Impactor was removed from the stack,
the nozzle plug was removed from the nozzle, the
impactor nozzle was reinserted into the stack and
placed at a preselected saaple point and sampling
commenced. Isokinetic sampling was conducted at
one or more sampling points. The sampling rate
through the nozzle and Impactor was maintained at
a constant flow rate, as determined from the
nomograph.
A sampling time was selected to collect a suffi-
ciently representative mass* for gravimetric anal-
ysis. At the end of a predetermined time, the
sample flow was stopped, the Impactor was removed
from the stack and the impactor flow terminated
simultaneously. The Impactor was taken to an on-
slte mobile laboratory and disaassembled. The
special light-weight substrates from each stage
and the backup filter were placed in individual
containers for storage until mass determinations
could be made at the laboratory. The Impactor
was allowed to cool and then the nozzle and all
surfaces upstream of the first impaction plate
were washed with reagent grade acetone to remove
any relatively large material impacted in the
nozzle or adhered to the Impactor surface. The
Impactor was then reassembled and testing was re-
sumed. All gravimetric determinations for gross
and tare weights were performed in a constant
temperature/constant humidity laboratory.
SO- and SO- Measurements
The EPA Method 8 sampling train was used for SO.
and SO- determinations.
Scott chose to conduct separate runs for SO- de-
terminations and particle size measurements.
Battelle chose to assemble the Anderson impactor
upstream of the Method 8 train (as shown in Figure
2) and conduct simultaneous sampling. The particle
size sampling period was shorter than the SO- sam-
pling period; thus, when a sufficient particle
size sample had been collected, Battelle removed
the Anderson impactor and continued the SO. run to
completion.
NO Measurements
The EPA Method 7 procedure was used to measure
NO concentrations in the exhaust from the glass
plants. Battelle collected integrated bag samples
and collected Method 7 samples from the bags;
Scott collected Method 7 samples directly from the
stack.
Gas Analysis
Grab samples of gaseous emissions were collected
in 3-1 glass flasks and the gases were analyzed
by gas chroma tography and mass spectrophotometry
for the following compounds:
o CO. o SO.
o CH
,
o 02
o AR
o Nn
o COS o C.H.
o CS.
o Iso-C.H.
o CO o H
Organic Sampling and Analysis
Where emissions from the glass forming area are
vented to the atmosphere via large hoods, it was
determined that organic emissions in the hood ex-
haust should be determined. To determine the or-
ganic constituents produced in the forming area,
a Tenax column was used to collect samples for
organic analyses. A schematic drawing of the or-
ganic sample train (with appropriate pumps and
metering apparatus) Is given in Figure 3.
Organic samples were fractionated using the liquid
chromatography/infrared spectroscopy procedure, as
incorporated in Level 1 Source Assessment Analysis
procedures. Gravimetric-determinations of each
fraction were made, and the primary compounds or
type of compounds in each fraction were identified.
Problems Encountered in Sampling and Analysis
Because nearly all the sampling and analysis pro-
cedures utilized in this program were established
methods, few unique problems were encountered.
The only troublesome problem encountered was ir-
regular flow profiles when sampling just upstream
of the Morgan ejectors utilized at some plants.
The problem of obtaining a representative sample
at the poor sampling location available was over-
come by using « large number (48) of sampling
points.
186
-------
FILTER HOLDER
IMPACTOR
NOZZLE
ICE BATH IMPINGERS
BY-PASS VALVE
THERMOMETER
CHECK
VALVE
VACUUM
LINE
VACUUM
GAUGE
Nsf\^
DRY TEST METER
MAIN VALVE
AIR-TIGHT
PUMP
FIGURE 2. MODIFIED EPA METHOD 8 SAMPLING TRAIN CONFIGURATION FOR USE
IN COMBINED PARTICLE SIZE AND SOX SAMPLING
FORMING AREA
HOOD VENT
ROOF,
FORMING AREA
EMISSIONS
TEMPERATURE CONTROLLED
WATER BATH
FLEXIBLE HOSE
THERMOMETER
VACUUM PUMP
DESICCANT
DRY TEST METER
FIGURE 3. SAMPLE TRAIN CONFIGURATION FOR SAMPLING ORGANIC EMISSIONS
FROM GLASS FORMING AREA
The process emissions measured at one glass plant
were sampled from a natural draft stack having gas
temperatures of approximately 850 F. Because of
the relatively high gas temperatures, a stainless
steel probe liner was used In place of the glass
liner. For these runs It was observed that the
Viton seals (which are used to prevent air leakage
where the nozzle connects to the probe) became
brittle due to the heat; thus, they were replaced
after each run.
For sampling runs where it was determined to use
more than one sampling point for obtaining par-
ticle sice data, it was necessary to change
nossles between traverse points. The flow through
the impactor bad to remain constant to maintain
the cutoff sixes for the various Impactor stages.
Thus, when maintaining isoklnetic sampling while
moving from one sampling point to another with a
different gas velocity. It was required that the
sampling nozzle site be changed rather than alter-
ing the sampling rate.
187
-------
References
1. Industrial Energy Study of the Glass Indus-
try to PEA and DOC, Battelle-Columbus,
Dec 1974, P-10.
2. Preliminary Source Assessment on Glass Con-
tainers to EPA, Battelle-Columbus,
March 8, 1976, P-3.
3. IEKL-RTP Procedures Manual: Level 1 Environ-
mental Assessment, J.W. Hamersma, S.L. Rey-
nolds, and R.R. Maddalone, EPA Publication
No. EPA-600/2-76-160a, June, 1976.
4. Federal Register, Vol.36, Mo.159, Part II,
August 17, 1971, P-15713-15717.
5. Code of Federal Regulations, 40, Protection
of Environment, Parts 60-99, Revised as of
July 1, 1976, P-110-117.
6. Standard Methods for the Examination of Water
and Wastewater, American Public Health
Association, 13th Edition, 1971, P-171-172.
7. Op. cit. (5), P-85-88.
8. Op. cit. (5), P-82-85.
188
-------
COMPREHENSIVE ANALYSIS OF EMISSIONS FROM
FLUIDIZED-BED COMBUSTION PROCESSES
by
K. S. Murthy, J. E. Howes, H. Nack and R. C. Hoke*
Battelle-Columbus Laboratories
Columbus, Ohio 43201
Abstract
Results of the comprehensive analysis
of emissions from a pressurized fluidized-
bed combustion unit (the Exxon Miniplant)
are described as an illustration of the
methodology for comprehensive analysis.
The results are discussed in the context
of the overall environmental assessment of
the process being conducted by the United
States Environmental Protection Agency.
The comprehensive analysis of the
fluidized-bed combustion emissions and
process streams involved approximately
740 measurements on about 90 samples, us-
ing more than 40 different inorganic,
organic, and physical analytical methods.
A discussion is presented covering the
methods used for sample preparation, in-
organic analysis, organic analysis, and
physical measurements. The quality control
procedures, and the accuracy of the esti-
mates derived from the data are discussed.
Introduction
Environmental data acquisition is one
of the seven major steps (Figure 1) in con-
ducting a complete environmental assess-
ment of new energy technologies such as
fluidized-bed combustion of coal (Refer-
ence 1). The primary means for this data
acquisition is comprehensive analysis of
emissions. Precommercial stage comprehen-
sive analysis (CA) of emissions from FBC
units provides the opportunity to detect
potential environmental problems early in
the development of the process. The en-
vironmental assessment of the process
based on the CA data should assist in the
identification and/or development of most
cost-effective control technologies that
may be required.
A phased approach in three levels is
the currently accepted technique for
sampling and analyzing emissions. The
three levels of analysis are defined as
follows:
• Level 1 analysis is devised for compre-
hensive screening of a wide variety of
organic and inorganic components.
Level 1 sampling and analysis is de-
signed to rapidly identify the potential
pollutants from a source, and to mea-
sure them with a target accuracy factor
of ±3. It also identifies all process
streams that may contain four types of
pollutants: gaseous, particulates,
liquids/slurry, and solids. Level 1
YES
Current
Process
Technology
Background
is
Better
Control
Needed
Control
Technology
Assessment
liJiAcguisitionjij
Current
Environmental
Background
Environmental
Objectives
Development
Environmental
Alternatives
Analysis
Quantified Control
Alternatives
Quantified Media
Degradation Alternatives
Standards of Practice
Manuals
Standards Development
Research Data Base
Reports
Figure 1. Simplified block diagram of environmental assessment steps.
*Exxon Research & Engineering Company,
Linden NJ 07036.
189
-------
strategy also includes bioassay test-
ing of several effluent streams to ob-
tain a direct index or estimate of
their toxicity potential.
• Level 2 analysis is based on Level 1
results. More accurate, compound-
specific analytical techniques are
used to pinpoint problem pollutants
and effluent streams. The Level 1
data together with bioassay data and
multimedia environmental goals will be
used to identify Level 2 and 3 analyti-
cal needs.
• Level 3 analysis (not yet defined com-
pletely) , would include routine con-
tinuous monitoring of those pollutants
identified as specific problems in
Level 2.
A set of twelve biological tests was used
in Level 1 testing. These tests and the
samples on which they were used are given
in Table 1. They are designed to test the
possible toxicity of a waste stream to
mammalian, marine, freshwater, plant, and
soil systems. This test protocol provides
a fairly good representation of the various
biological constituents of the environment
that might be exposed to a waste stream.
The bioassays are designed to be imple-
mented quickly and inexpensively, in keep-
ing with the screening nature of Level 1
testing. Their output will permit a rela-
tive ranking of waste streams according to
biological hazard, and, together with the
chemical and physical data, will provide an
overall hazard characterization of the
waste streams.
The measurement techniques and results
presented in this paper are based largely
on Level 1 analysis. The sampling matrix
for Level 1 (and some Level 2 analysis of
substances already known as problem pol-
lutants) is shown in Table 2. Data pre-
sented in this paper were obtained from the
pressurized FBC facility at Exxon; sampling
was conducted in accordance with Table 2.
This facility has a 0.32-m-diameter reactor
which was operated at 890 C, 900 kPa,
1.2 m/sec superficial velocity, 40 percent
excess air, 75 kg/hr coal feed and 11.0 kg/
hr dolomite sorbent feed at a Ca/S molar
ratio of 1.25 for the tests reported in
this paper.
Sampling Methods, Analytical
Techniques, and Quality
Control Procedures
Sampling Methods
The comprehensive analysis program for
the EXXON unit consisted of sampling seven
of the nine streams shown in Figure 2. The
nine streams are (1) coal feed, (2) dolo-
mite feed, (3) 2nd stage cyclone discard,
(4) bed reject material, (5) cyclone dis-
card leachates, (6) bed reject material
leachates, (7) undiluted stack gas,
(8) diluted stack gas, and (9) dilution
and combustion air. Sampling of the sor-
bent regenerator unit was not performed
since this unit was not operated during
the tests. Streams (5) and (6) were simu-
lated in the laboratory since no leachate
streams were actually present at the mini-
plant site.
Five tests were conducted at EXXON
over the period of March 28 to April 1,
1977, during which the Miniplant was
operated continuously for about 80 hours.
Sampling was performed on an around-the-
clock basis by two 7-man teams working
12 hour shifts. During each test, lasting
five hours, the various FBC streams were
sampled by the techniques given in Table 2
and elaborated in Table 3.
Grab samples of solids from the FBC
process streams were taken periodically
throughout each test. The individual
samples were composited to obtained one
sample per test. The gases (CO, CO2, O2,
SO2, etc.) in the undiluted flue gas at
(7) were sampled continuously for analyses
by the EXXON on-line instrumentation. In-
tegrated bag samples were also taken for
CO2 and 02 analyses by Orsat. S02 and NOX
were sampled by EPA Methods 6 and 7, re-
spectively, to provide backup data. The
controlled condensation method was used to
sample for 803/82804, and special impinger
trains were used for sampling NR3, HCN,
HC1, and HF.
Continuous sampling of the undiluted
flue gas was performed for total hydro-
carbon measurements. Grab samples in glass
bulbs were taken several times during each
test for gas chromatographic analysis of
C^-Cg hydrocarbons and sulfur compounds.
TABLE 1. LEVEL 1 BIOASSAY TEST MATRIX
Sample Type
Hater and
Liquids
Solids
Gases
Particulates
Sorbent
Health Effects Tests
Microbial
Hutagenesis
•
•
•
Rodent Acute
Toxicity
•
Cytotoxicity
•
M
II
Ecology Effects Tests
Algal
Bioassays
«
Static
Bioassays
•
Soil
Microcosm
M
Plant Stress
Ethylene
Soil
Microcosm
190
-------
TABLE 2. SAMPLING AND ANALYSES TO BE PERFORMED IN COMPREHENSIVE ANALYSIS OF FBC UNITS
Sojiln. PXularm To
CONTINUOUS OAS MEASUREMENTS
CO,
CO
NO
NO,
SO,
0,
INTEGRATED OAS MEASUREMENTS
°°»
CO
NO.
*°3
0,
"7
M,0
"a*
COS
°MjiM ~5»H1J^._
C« — CM tiydioovttont
C*j — C|j hytfrocvtoont
C| — Cfi cMoroM^ont
NH,
HCN
Cyonooon
•O_|M-1O«
^'Jlri2**J4
HO
Nuortdt
Sompto
roBonikiii
M«y.W
CM
CM
CM
CM
CM
CM
IO
IO
IQ * M7
10 » M«
10
IO
IO
IG
10
IG
IG
10
10
10
10
IO
OR/St
St
St
*»f
MM
Andy* MMhodlbl >»M
NOIRlcl
IR or UVlcl
CL'd
CL<«>
NDIR|C>
PM<"
FOC/TC
FOC/TC
FOC/TC » M7
FOC/TC * M8
FOC/TC
FOC/TC
FGC/TC
FOC/FPO
FOC/FPO
FGC/FPO
FOC/FID
FOC/FID
FOC/EC
FOC/TC
FOC/TC
FGC/TC
Ion cnrOffWtOQrvpti
TlliBUon
SIE
«v-.t«— -M^i
BCM!I «tolM WMM
u«Ma» CaKicUen Laoehata
Flo. Dortoa Bod Cod Sortxnt from Solid WaHa
<^l Oo> Dhaord Ro|M Faod Food CorHctton Dorlaa *ad
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
x
X
X
INTEGRATED SPCCIMENS FOR SUBSEQUENT GROUf ANALYSIS
71 olnnonu (U tJutx^h U)
•tonlnwM IrtMW
UltlnvK
Suttur lonrn I'udil
RadkMudlo*! (Gran a ft A
Oromtc Omnlroll
Oromta by do>
Orojnle compound!
POM
Orojnle - nduoid iuHiw oompoundi
Cj-Cf, li>iliotorbij»ii
Orionk imm
SASS/OS
Ol
Ol
Ol
SASS/Ol
SASS/Gl
SASS/Ol
SASsyai
SASS/Ol
tASS/Gl
SASS/Ol
SSMS X
ASTM 03172-71
ASTM Oil 76-74 X
ASTM 01492-ee
LBPC X
LC/IR Itaul •rapt* and R Iracriom) X
LC/LRMS (MUcud fnctlonil X
GC/MS X
GC/FPO IS fraction CDfrtHnad) X
OC/FIO X
Mfcrobalanca IS fraotkxiil X
X XXXX XX
X
x x
X
x xxxx
XXX X
xxx x
xxx x
xxx x
xxx x
XXX X
-------
INTEORATED SPECIMENS FOR SUBSEQUENT SPECIFIC ANALYSIS
(•». CD. HI. Pb. S*. (to. T.I
At
CT
a
MI
co,;
NOj-
NO,-
C Inon-eirterrai
fvtkto tin
BIOLOGICAL ASSAYS
Cytotoxldtv
CytonxMty
Acut» toxldtv
Cooloiilo»l
SASS/Gl
SAM/Ol
(AM/0*
SASS/Oi
Ol
Of
Ol
SASS/Gl
SASS/O.
SASS/Gl
SAM/Oi
SASS/Ql
SASS/Gl
Ol
Gi
Ol
Oi
SAM/MS
SASt/Ol
SASS/Gl
SASS/Ol
SASS/Ol
AAS
OotorlmMrie
Dtalllnlon/eolorlnMtrte
AAS
AAS/tltritlon
AAS
CM wolutlon
Tltntlon/lan ehromnogripln
SOj wohrtlon/eolorlmitrle
HS •Mhitlon/trtratlora
Cotofvnotnc
Combunlon
ASTM D 201S-«8
LM/SEM
SltM ASTM D 410-38
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
MntOflrita/AinM
rtorrwi lung fteratolMt IWI-JS)
FUbblt tlMOtor nweraptugi (RAM)
In vt
FrWtMU
FrMhMt*
FrwtnMtv
Silnwur
Mtwrar
Stlfntat
TirraMTW
TtcnnrW
r •*!
r mliral (dtphnli)
r •nhml (IWi)
•iBil
•nliml (gran riwlnv)
•nhral III*)
toll
ptont
SASS/Oi
SASS/Ol
SASS/d
SASS/Oi
SASS/Ol
SASS/Gl
SASS/Ol
SASS/Gl
Algrt bonto
Slttfc
Suite
Untedlulv mwlnt itgn
Sonic
Stalk
Soil mtooewm
Strm ttflylWM
X
X
X
X
X
X
XX X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
U)
(bl
Cw
IO
OR/St
Si
SASS
Gl
MS
M6
M7
NOIR
IR
UV
CL
- Oontlnuoui NUMrnwl through nonrwetrw Una with nwotunlal filtration
- InngreMd grab wivta of OM In glm bulb
- Ookiovr-«oii coll/«McM •mpllng train
- Stpwm MM dwmloil train to collect gu (wdi u Mnhod el
- Souro* Amnnwit Sonpllng SvMm (train uMd foe iinptndtd pvtlculMM.
orgtnlo. *nd voUtlll tm> fttmnin)
- Orab multipl* MmplM rllttad to raduoi to 100 9 raprMtnutha nmpli
- EPA MMhod 6
- EPAMtthod6
- EPA Mtttiod T.
- nk>ndl«»nlv* Irrfrarad
- Uttravtoltt
- Ch«nllumln«oi«oi
FGC/TC - FMd ChromMograph/ttwrmil oonductMtv dtnctor
FGC/FPO - FMd diromtograph/fltm* photomttrk dttwtor
FGC/FIO - FltM ChronMtoeraph/llimt tonlatlon d«Mtor
FGC/EC - FMd ChromMograpriM*ctn>n oipturt dttMtor
SIE - Mtcthn-lon ElKtrodt
SSMS - Spvk Souroi Mm Sptctroicopy
ASTM - Anwlcan Sodny for Ttrtng Minrlili Sundird Mtthod
LBPC - Low background g» proportloral oontrollir
LC - Liquid Chrormtography
LRMS - Low Rnolutlon MIH Spuuonntry
GC — On CrtromMography
OC/MS - On Oirormtognphy with Mm SpKtrogriphv
AAS - Atomic ABiorpotlon/Spictraicopv
LM/SEM - Light MIoroKopv/Sonnlna Electron Mlaonopt.
Id Or (colpubll InftrurnmtMlon llrMdy IrnalM il IrMtloni Included.
-------
([) COAL
©LIMESTONE
FEED
SUPPLY
AUXILIARY |\.
AIR
COMPRESSOR
CYCLONE
SEPARATOR
Circled numbers denote
streams sampled
Figure 2. The fluidized-bed combustion miniplant at EXXON
during comprehensive analysis tests.
TABLE 3. SAMPLING TECHNIQUES USED FOR EXXON CA
Stream
Substance
Sampling Method(s)
Coal feed (l)
Dolomite feed (2)
2nd stage cyclone solids @
Bed reject (2)
Leachate from cyclone solids
Bed reject leachates (V)
Flue gas, undiluted (?)
Flue gas, diluted (s)
Combustion and dilution air
• Solid samples
Inorganics, organics
Inorganics, organics
CO
C02
02
NO
NO2
SO2
Hydrocarbons
Sulfur compounds
S03/H2S04
NH3
HCN
HC1
HF
Particulates
Particulates
Organics
Multiple grab
720-hr shake test
720-hr shake test
Continuous
Continuous and integrated bag
Continuous and integrated bag
Continuous "1
Continuous J «* "•«"* 7 (NOx>
Continuous and Method 6
Continuous and grab, glass bulb
Grab, glass bulb
Condensation coil
Impingers, H2SO4
Impingers, KOH
Impingers, NaOH
Impingers, NaOH
Balston filter (heated)
SASS and Method 5
Tenax trap
193
-------
The primary particulate sampling was
performed in the flue gas stream which was
reduced to near atmospheric pressure by
dilution with air. The Source Assessment
Sampling System (SASS) was used to collect
samples for chemical and physical analysis.
In three tests, the stainless steel con-
denser module normally supplied with the
SASS unit was replaced with a glass module
of similar dimensions. The glass module
modification was included in these tests
since a preliminary sampling experiment
had indicated that excessive corrosion of
the stainless module occurred during the
sampling of the FBC emission. Two tests
were performed using the stainless steel
module.
Method 5 sampling was performed to
obtain compliance-related mass emission
data.
The Balston filter sample of particu-
lates from the undiluted flue gas stream
was obtained to study changes in the par-
ticulate characteristics or composition
that might be caused by the dilution of
the flue gas stream. The dilution air
was sampled with a Tenax trap to identify
any organics that might contaminate the
flue gas stream.
Sample Preparation. Before laboratory
analyses commenced, several operations were
needed to properly prepare the samples.
These operations included (1) obtaining
sample weight and volumes, (2) compositing,
(3) riffling, (4) pulverizing, (5) homog-
enizing, (6) aliquotting, (7) extractions,
and (8) Parr bomb combustion. All of
these steps were carried out so as to main-
tain sample integrity and representativity.
Sample size limitations frequently re-
quired careful aliquotting to assure suf-
ficient sample for all intended analyses.
Inorganic Analysis Methods. The tech-
niques used for the inorganic analyses of
the SASS, FBC process and leachates from
the bed reject and 2nd cyclone samples are
shown in Figure 3. The Level 1 analyses
included determination of 71 elements by
SSMS, Hg by flameless atomic absorption,
Sb by atomic absorption with graphite fur-
nace atomization, and As by the diethyl
dithiocarbamate colorimetric method. A
rather extensive group of Level 2 cation
and anion analyses were performed using
atomic absorption, ion chromatography, and
a variety of wet chemical methods. Coal
samples were analyzed by ASTM standard
methods.
Analytical Procedures
Analyses were performed on samples
from three of the EXXON sampling tests.
Approximately 90 samples from the three
tests were analyzed using more than 40 dif-
ferent inorganic, organic, and physical
measurement methods. The analytical work
was initiated about 1-1/2 weeks after the
sampling, and was completed in about 2-1/2
months. The initial 1-1/2 week period was
required to prepare samples for analysis.
Organic Analyses. The organic analy-
ses included the Level 1 and 2 determina-
tions shown in Figure 4. Initially
samples were Soxhlet extracted with methy-
lene chloride; pentane was used for the
extraction from the XAD-2 sorbent. The
subsequent Level 1 analyses included:
determination of the weight of organics in
the sample, identification of organic
classes by Fourier Transform Infrared
Spectroscopy (FTIR) in both the unsepa-
rated extract sample and the eight
|SASS, FBC PROCESS SAMPLES, AND LEACHATES |
Level 1
SSMS - 71 elements
Hg - FAAS
Sb - AAS/GF
As - Colorimetric
Level 2
SASS and Process Samples
Be - AAS/GF
Cd - AAS/GF
Pb - AAS/GF
Se - AAS/GF and hydride
Te - AAS/GF
Ca - AAS/flame
Na - AAS/flame
Mg - Gravimetric
C - Combustion
Cl~ - Gravimetric
- Ion chromatography
- C(>2 evolution
804= - Gravimetric
SC>3= - Colorimetric
evolution
- Colorimetric
- Ion chromatography
Coal feed
Proximate - ASTM
Ultimate - ASTM
Sulfur forms - ASTM
S= -
N(>2
NC>3
Figure 3. Inorganic analysis performed in EXXON CA Program.
194
-------
fractions derived by liquid chromato-
graphic separation, determination of
C?~C12 hydrocarbons by gas chromatography
with flame ionization detection (GC/FID),
and examination of selected liquid
chromatographic fractions by low resolu-
tion mass spectrometry (LRMS).
The Level 2 effort consisted of analy-
ses for specific polycyclic organic mat-
ter (POM) compounds by combined gas
chromatograph/mass spectroscopy (GC/MS)
and determination of specific organic sul-
fur compounds in the samples by gas
chromatography with flame photometric
detection (GC/FPD).
Physical Measurements. The physical
measurements outlined in Figure 5 were
performed on all SASS and FBC stream solid
samples. Normal light and polarized light
microscopy at magnifications of 10 to 2SOX
were used to obtain data on the shape,
structure, cleavage, color, and particle
size distribution in the various samples.
Scanning electron microscopy (SEH) was
used to examine the finer particles col-
lected in the SASS cyclones and on the
SASS filter.
Radioactivity measurements for gross
alpha and beta emissions were performed
by counting a portion of each sample in a
Beckman low-background proportional
counter.
Quality Control Procedures
Several techniques were used to assess
the quality of the EXXON inorganic analyti-
cal data. These included:
1. The analysis of NBS standard Refer-
ence Materials (SRM)
2. Dynamic spiking techniques
3. Comparison of results obtained by
two different analytical methods,
spark source mass spectroscopy
(SSMS) and atomic absorption spec-
trometry (AAS).
The NBS samples used in this work were
SRM 1631B (coal), SRM 1632 (coal), SRM 1633
(coal fly ash) and SRM 88a (dolomite). The
matrices of these standards are quite simi-
lar to the FBC feedstocks and emission
stream samples. Reference values are
available for over 40 elements in the
SRM 1632 coal and SRM 1633 fly ash, and
certified values for 11 elements in the
SRM 88a dolomite are given. SRM 1631B
coal is certified by NBS for ash and sulfur
content.
SSMS analyses were performed on SRM's
1632, 1633, and 88a. In these and other
analyses involving the standards, the mate-
rials were submitted as blind samples with-
out the analyst being informed that stan-
dards were included along with the series
I SASS AND FBC PROCESS SAMPLEsJ
Level 1
Weight
Organic classes - FTIR
C7-C12 HC - GC/FID
Major Compounds 6 Classes - LRMS
Level 2
POM - GC/MS
Sulfur Compounds - GC/FPD
Figure 4. Organic analysis performed in EXXON CA Program.
SASS AND FBC PROCESS SAMPLES
I
Level 1
Shape
Structure
Cleavage
Color
Size distribution
Light Micro-
' scope and SEM
Level 2
Radioactivity - a and 6
Figure 5. Physical analysis performed in EXXON CA Program.
195
-------
TABLE 4. SPARK SOURCE MASS SPECTROGRAPHIC ANALYSIS OF NBS STANDARD REFERENCE MATERIALS
Results in Wg/g, except as indicated
• NBS SRM 1632 (Coal)
Element
Al
As
Ba
Br
Ca
Cd
Ce
Cl
Co
Cr
Cs
Cu
Eu
Fe
Ga
Hf
In
Hg
K
La
Hg
Mn
Mo
Na
Mi
P
Pb
Rb
Sb
Sc
Se
Si
Sm
8r
Ta
Th
Ti
D
V
Zn
Zr
Ref. Value< •>
1.9Z
(5.9)
405
14.2
0.44Z
(0.19)
18.5
1000
(6)
(20.2)
1.4
(18)
0.21
(0.87X)
8.5
0.95
0.07
(0.12)
0.29Z
10.5
0.25Z
(40)
3.4
390
(15)
—
(30)
24
3.9
4.5
(2.9)
(3.2Z)
1.7
144
0.17
(3.0)
(800)
(1.4)
(35)
(37)
45
BCL
2.7Z
7
670
16
1Z
0.1
14
640
13
32
1
64
0.2
2Z
5
1
<0.1
<0.2
0.24Z
5
0.38Z
32
3
280
14
—
32
38
7
1
6
6.9Z
1
160
<2(0
~2
1000
1
16
70
6
NBS SRM 1633 (Fly Ash)
Ref. Value <«>
12. 5Z
(61)
2780
6.0
4.34Z
(1.45)
146
42
(38)
(131)
8.6
(128)
2.86
6.37Z
49
10.8
—
(0.14)
1.72Z
82
(1.98Z)
(493)
—
(3070)
(98)
—
(70)
(112)
6.9
32
(9.4)
21Z
15
(1380)
1.6
(24)
6920
(11.6)
(214)
(210) -
301
BCL
15Z
40
2000
15
3Z
5
100
40
30
100
20
150
3
10Z
40
6
—
0.4
1.5Z
40
2Z
500
—
1000
100
—
40
200
2
30
5
15Z
5
700
<0.6
15
6000
8
200
200
200
NBS SRM 88a
(Dolomite)(b)
Ref. Value
0.19
—
—
—
30.1
—
—
__
—
—
—
—
—
0.28
—
—
—
—
0.12
—
21.3
0.03
—
0.01
_
0.01
—
—
—
__
__
1.2
—
0.01
—
0.02
—
—
—
—
BCL
0.3
—
—
—
42
—
—
__
—
—
—
—
—
0.6
—
—
—
—
0.6
—
25
0.04
—
0.004
—
0.004
—
—
—
_
__
1.1
__
0.006
—
0.01
—
—
—
—
(a) Reference values in parenthese* are given by NBS. The other values were taken
from work by ORNL and LLL. Reference: Environmental Science and Technology,
Volume 9, Number 10, page 975 (October, 1975).
(b) Values in percent by weight as oxides. Reference values given by NBS.
(c) Possible contamination from source holder used in SSMS analysis.
196
-------
of FBC samples. The results of the SSMS
analyses are given in Table 4 along with
the corresponding reference values given
by NBS or derived from work by Lawrence
Livermore Laboratory (LLL) and Oak Ridge
National Laboratory (OKNL). A summary of
the comparison of the SSMS values with the
reference values is presented in Table 5,
including the factor of 2 to 3 accuracy
requirement for Level 1 analysis.
TABLE 5. SUMMARY OF COMPARISON OF SSMS
VALUES WITH NBS STANDARD
REFERENCE VALUES
SRM 1632 Coal (37 comparisons) ^a^
27 elements differ by less than a fac-
tor of 2
7 elements differ by less than a factor
of 3
1 element (Cu) differs by a factor of
3.6
1 element (Sc) differs by a factor of
4.5
1 element (Zr) differs by a factor of
7.5
SRM 1633 Fly Ash (35 comparisons) (a)
28 elements differ by less than a fac-
tor of 2
6 elements differ by a factor of 3 or
less
1 element (Cd) differs by a factor of
3.5
SRM 88a Dolomite (11 comparisons)
8 elements differ by a factor of 2 or
less
3 elements differ by less than a factor
of 3
(a) Comparisons do not include As, Hg,
or Sb.
Based on comparisons of SSMS results
and reference values for 37 elements in the
NBS coal samples, 73 percent of the mea-
surements agree within a factor of 2, and
92 percent agree within a factor of 3.
SSMS results for Cu, Sc, and Zr differed
from the reference values by more than a
factor of 3. Comparisons of 35 elements
in the NBS fly ash show that 80 percent of
the SSMS and reference values agree within
a factor of 2, and 97 percent agree within
a factor of 3. Only Cd differed slightly
more than a factor of 3 with the reference
value. For the dolomite (11 comparisons),
73 percent agree within a factor of 2 and
SSMS values for all elements are within a
factor of 3 of the reference values.
The NBS coal and fly ash samples were
also analyzed for As, Be, Cd, Hg, Pb, Se,
Sb, and Te by the Level 1 and Level 2
methods used for the FBC samples. The
results of the analyses and the reference
values are presented in Table 6. With ex-
ception of Se in the fly ash, all results
are within the estimated uncertainties
given by NBS and LLL.
The analyses of the NBS dolomite
standard for Ca, Mg, and CX>2 and the 1631B
coal standard for ash and total sulfur are
given in Table 7. The analyses were per-
formed by ASTM standard methods. Satis-
factory agreement with certified values
was obtained for each measurement.
Many samples in the EXXON work were
analyzed both by SSMS and by AAS for Be,
Cd, Pb, Se, and Te. Since AAS is an in-
herently accurate analytical method, com-
parison of the values obtained by the two
methods provides another means to estimate
the quality of the SSMS data. Table 8
presents a comparison of the SSMS and AAS
values obtained for various sample types
analyzed in the EXXON program. Based on
a comparison of 21 samples for the 5 ele-
ments, 80 percent of the SSMS values were
within a factor of 3 of the AAS results.
The disagreement observed in the compari-
sons is mostly due to the low SSMS values
for Be and the high values obtained for all
elements except Pb in the acid condensate
from the SASS condenser module.
In the analysis of the SASS impinger
solutions and selected FBC samples by AAS,
dynamic spiking was used to estimate the
analytical accuracy. A summary of the
spike recoveries for the various analyses
is shown in Table 9.
Satisfactory spike recoveries were ob-
tained for all element sample combinations
which were analyzed by the AAS/graphite
furnace and flameless AAS methods. Un-
acceptably low spike recoveries were ob-
tained by analysis of the SASS impinger
solutions using the AAS/hydride generation
method. Therefore, Se analysis by the AAS/
graphite furnace techniques was selected
as the preferred method for these samples.
Summary and Conclusions on Samplinc
Methods, Analytical Technique's^
Quality Control Procedures
anc
Based on experience gained to date in
applying CA sampling and analysis pro-
cedures to fluidized bed combustors, the
following comments and conclusion can be
made on the efficacy of the methods.
Sampling Methods. In general, the
sampling methods as used for the EXXON pro-
gram performed satisfactorily. However,
some problems were encountered in the use
of the SASS train. Leakage around the
filter was observed during the initial
tests because the two filter body halves
did not properly match to seal the filter
197
-------
TABLE 6. ANALYSIS OF STANDARD REFERENCE MATERIALS
Results in wg/g
Element
Be
Cd
Hg
As
Pb
Se
Sb
Te
BCL Analytical
Method ta>
AAS/6F
AAS/GF
FAAS
Color
AAS/GF
AAS/HG
AAS/GF
AAS/GF
NBS/SRM 1632
Ref. Value (b)
(1.5)
(0.19 ± 0.03)
(0.12 ± 0.02)
(5.9 ± 0.6)
(30 ± 9)
(2.9 ± 0.3)
3.9 ± 1.3
(<0.1)
(Coal)
BCL
2.5
0.19
0.10
5.5
31
2.6
3.5
<0.2
NBS/SRM 1633
Ref. Value (b)
(12)
(1.45 ± 0.06)
(0.14 ± 0.01)
(61 ± 6)
(70 ± 4)
(9.4 ± 0.5)
6.9 ± 0.6
(c)
(Fly Ash)
BCL
15
1.4
0.10
56
68
7.5
7.0
<0.5
(a) AAS/GF - Atomic absorption spectrometry/graphite furnace atomization
AAS/BG - Atomic absorption spectrometry/hydride generation
FAAS - Flameless atomic absorption spectrometry
Color - Colorimetric (silver diethyldithiocarbamate).
(b) Reference values in parentheses are given by NBS. The other values
were taken from work by LLL. Reference: Environmental Science and
Technology, Volume 9, Number 10, page 975 (October 1975).
(c) Reference value not given.
TABLE 7. CHEMICAL ANALYSIS OF NBS TABLE 8.
DOLOMITE AND COAL STANDARDS(b)
COMPARISON OF SSMS WITH AAS FOR
FOR SELECTED ELEMENTS
Analysis
Ref. Value, BCL Value,
percent percent
Dolomite - NBS/SRM 88a
co2
Ca as CaO
Mg as MgO
Coal
Ash
Sulfur
46.6
30.1
21.3
- NBS/SRM 1631 B
14.59
2.016
46.8
30.4
21.7
14.2
1.92
(a) Analyses performed by ASTM methods.
Comparison of
with AAS (a)
Sample Type
Coal
Sorbent
Bed Reject
2nd Cyclone
SASS, 10,3 v
cyclone catch
SASS, 1 p cyclone
+ filter catch
SASS, acid
condensate
Be Cd
L
L
L
L
L
H H
Pb Se
H H
L
L
H
SSMS
Te
H
(a) H - SSMS higner than AAS by more
than a factor of 3
L - SSMS lower than AAS by more
than a factor of 3
No entry - SSMS within ± a factor
of 3 of AAS.
Summary: 105 comparisons
80% within ± a factor
of 3
10% SSMS more than fac-
tor of 3 above AAS
10% SSMS more than fac-
tor of 3 below AAS.
198
-------
TABLE 9. SPIKE RECOVERIES FROM SASS IMPINGER SOLUTIONS AND
SELECTED FBC PROCESS SAMPLES
Spike Recovery , percent
Element
Be
Cd
Hg
Pb
Sb
Se
Te
Analytical
Method (a)
AAS/GF
AAS/GF
FAAS
AAS/GF
AAS/GF
AAS/GF
AAS/GF
H202
Sample No. 1074
96
104
100
95
96
95(70) (c)
95
(NH4)2S208/AgN03
Sample No. 1075
95
106
96
90
92
100 W
95
Dolomite Feed
Sample No. 1303
104
99
ND
112
ND
93
90
Coal Feed
Sample No. 1308
ND
99
100
116
ND
ND
95
(b)
(a) AAS/GF - Atomic absorption spectrometry/graphite furnace atomization
FAAS - Flameless atomic absorption spectrometry.
(b) ND - not determined.
(c) Recovery in parentheses obtained by hydride generation method using 2 ml of H202
impinger solution.
(d) Very low recovery obtained by hydride generation method.
perimeter. This problem was remedied in
subsequent tests by using two filter discs.
Checks prior to tests frequently showed
leakage around the Teflon cyclone gaskets.
Evidently the problem was caused because
the Teflon gasket became distorted in the
previous test and would not reseal prop-
erly. The use of new Teflon gaskets for
each test minimized this problem. Design
changes should be made to eliminate the
filter and cyclone leakage problems.
An unresolved problem with regard to
the SASS train is whether a glass or stain-
less steel condenser module should be used.
In the EXXON sampling, rather severe cor-
rosion of the stainless steel module was
observed as a greenish deposit at the inlet
from the filter. The deposit could not be
effectively removed either by brushing or
scraping during the cleanup operations and
could not be collected quantitatively as
desired.
Of more significance are the differ-
ences observed in the concentration of
certain elements in the condensate col-
lected in the glass and stainless modules.
Based on AAS analysis, selenium concentra-
tions were significantly higher (over 20X)
in the glass module condensate. On the
other hand, cadmium concentrations were
significantly higher (about 15X) in the
stainless module condensate. Significant
differences were not observed in the con-
centrations of Be, Hg, As, Pb, Sb, and Te
in the glass and stainless condensate
solutions.
Additional studies may be conducted to
determine the reasons for the differences
in Se and Cd concentrations in the con-
densates from the glass and stainless con-
denser modules, and to decide which type
should be used in future comprehensive
analyses to obtain reliable trace element
data. Till then, for FBC units, the glass
module should be preferred because of its
significantly reduced potential for corro-
sion and for contamination of samples.
Analytical Techniques. The quality
control data obtained in the EXXON study
demonstrates that SSMS can provide data
within the required Level 1 accuracy. How-
ever, in order to minimize sample handling
and related errors, and to reduce costs, it
is suggested that the Parr bomb combustion
step be eliminated when possible, i.e.,
when organics are not present in the
samples. In the case of FBC sample analy-
ses, this would include the SASS particu-
lates, 2nd cyclone discard, bed reject,
and sorbent.
Atomic absorption spectrometry has
been shown to be a suitable method for
Level 1 and Level 2 Hg and Sb analyses.
The colorimetric method for As is accept-
able for Level 1 and 2 analysis, although
the use of AAS for this element is known
to provide a more sensitive, lower cost
method of analysis. AAS procedures are
also shown to be acceptable for Level 2
analysis of Be, Cd, Pb, Se, and Te.
However, Se analysis of the SASS im-
pinger solutions by the hydride generation
method presents some problems. The sele-
nium hydride cannot be quantitatively
evolved from the impinger solutions. These
problems may be overcome by first destroy-
ing the peroxide and precipitating the
silver with hydrochloric acid. It is also
anticipated that a similar problem would be
encountered in the arsenic analysis of the
SASS impinger solutions by the arsine
generation method for either AAS or
colorimetric determination.
Quality Control Procedures. In the
EXXON work, procedures to assess data
quality were applied primarily to the in-
organic analysis. In future work, emphasis
199
-------
will also be placed on incorporation of
quality assessment measures into the
organic analysis scheme.
Inter-laboratory analyses of stan-
dards and process samples are also strongly
suggested as a means to better evaluate
and improve CA analytical methodology.
Comprehensive Analysis Results
Some qualifications to the data pre-
sented below should be noted:
(1) This is the first extensive analysis
of emissions on a fluidized-bed unit.
(2) Steady-state conditions may not have
been achieved during the CA tests
since the sulfur content of the coal
varied. This resulted in drastic
tapering of the SO2 concentrations
in the flue gas.
(3) The regenerator was not operated
during CA runs at the Miniplant.
However, since the data collection process
was well monitored and controlled, the
data are considered reliable.
Results and Evaluation
Comparisons to MATEs. To evaluate the
significance of the measured concentration
of substances in the effluent streams, the
measured concentrations were compared to
the MATE (Minimum Acute Toxicity Effluent)
values (Reference 2).
MATEs are indicators of allowable
concentrations of contaminants in the
effluent stream, and provide a point of
reference for control technology goals.
MATEs cannot be used as absolute indica-
tors of minimum toxicity since they are
still in the developmental stage.
The procedures involved in developing
MATE values can be found in Reference 2.
MATEs are approximate concentrations of
contaminants in air, water, or land
effluents which may evoke minimal signifi-
cant harmful responses to humans or the
ecology, within 8 hours. In general,
types of data chosen to provide the basis
for MATEs include threshold limit values
(TLV), NIOSH recommendations, lethal dose
concentrations and other toxicity data,
drinking water regulations, and water
quality criteria. For a single substance,
five specific MATE concentrations can be
defined: two air MATEs and two water
MATEs (one each based on health and ecology
effects) and one land MATE (based on the
lower water MATE).
Emissions in Flue Gas. Samples of the
flue gas were collected at about 900 kPa
pressure before air dilution, and analyzed.
Table 10 shows flue gas composition mea-
sured with on-line instruments and wet
chemistry tests.
TABLE 10. ANALYSIS OF FLUE GAS
Substances
CO
C02
O,
a8
H2SO4 Mist + 503
S02
NH3
CN
P
Cl
NOx as NO2
As
Be
Cd
Hg
Pb
Sb
Se
Te
Concentration
ug/n>3 (ppm)
61,734(53)
24 x 10^
5.5%
2196(3.3)
2079(5)
74,813(28)
501(0.6)
1.2
10,120(13)
54,824(33)
148,442(70)
<2
<0.4
0.1
0.85
<1.2
<1.7
<1.4
<1.7
Air
HATE
uq/m3
55,000
9 x 106
None
None
1,000
13,000
18,000
5,000
3, COO
9,000
2
2
8.2
50
150
500
10.8
100
Measured by
On-line
instruments
Wet chemistry
Atomic
absorption
The lowest concentrations of SC>2 and
NOX in the flue gas are respectively
28 ppm (0.18 lb/106 Btu) and 70 ppm
(0.09 lb/10<> Btu). These are very low
in comparison with existing New Source
Performance Standards for coal-fired
steam generators: 1.2 lb/106 Btu for SO?,
and 0.7 Ib/lO^ Btu for NO.
However, in comparison with the MATEs
shown in Table 10, both SO2 and NOx are
much higher in concentration than the
MATEs allow. In fact, the SO2 concentra-
tion in the flue gas is about 6 times the
MATEs value, and the NOX concentration is
16 times. This fact does not necessarily
mean that SO2 and NOX are a problem; but it
indicates that the MATEs are deliberately
conservative .
The concentrations of all the eight
volatile toxic elements were well below the
MATEs. Incidentally, arsenic concentra-
tion measured by spark source mass spec-
trography was higher than the MATEs; but
the colorimetric analysis, classed as a
Level 2 technique, showed it to be less
than the MATE value.
Concentrations of polycyclic organic
matter in flue gas given in Table 11 are
all lower than the air MATEs.
TABLE 11. ORGANIC COMPOUNDS IN FLUE GAS
Concentrations, Air MATE,
Substance* ng/afl ng/m^
Anthracene/phenanthrene
Methyl anthracenes
Pluoranthene
Pyrene
Methylpyrene/fluoranthene
Benzo(c) pbenanthrene
Chrysene/benz (a) anthracene
Benzo fluoranthenes
Benz (a) pyrene
HC > C6-C12, ng/m3
HC > C12, M9/»3
53
5
26
9
1.0
0.2
3.8
1.0
O.S
1740
58
483,000
483,000
90 X 106
233 x 106
No data
26.9 x 106
44,800
897,000
20
No data
No data
200
-------
Suspended Particulates in Flue Gas.
The particulate concentration in the flue
gas, after passing through two conventional
cyclones, but without the use of fine
particulate control equipment, was about
1.2 gr/scf (1.9 lb/10o Btu) as compared
to the EPA standard of 0.1 lb/10* Btu.
However, if an appropriate third-stage
particulate control device that removes
over 96% of the particulates can be demon-
strated, the EPA emissions standard would
be met.
The physical properties of the sus-
pended particulates in the flue gas may be
of interest. The size distribution and
morphology are:
<1 micron
1 to 3 microns
3 to 10 microns
>10 microns
Predominant shape:
Evident cleavage:
Structure:
Color:
7% wt.
39% wt.
21% wt.
33% wt.
Irregular
None
3 phase
White, black, red
Thus, 46 wt. percent of the suspended par-
ticulates collected by the source assess-
ment sampling system's cyclone (Refer-
ence 3) were less than 3 microns, and are
in the respirable range. The a radio-
activity averaged about 6 ± 3 pCi/g, and
the 0 radioactivity, 30 ± 10 pCi/g.
Table 12 indicates that the toxic and
volatile metal content in the particulates
is of real significance. Since most of
these particulates will be collected as
solid wastes by any effective particulate
collection system, the land MATE values
apply. Of these metals, all but Te and Hg
would exceed the land MATE even if the
particulates were removed from the flue gas
at a reasonable collection efficiency and
were considered for land disposal (see
Table 12). The second "Yes" in the MATE
comparison column indicates that the
particulates would still constitute a land
disposal problem even when diluted by a
factor of 100 by mixing with nontoxic mate-
rials (or some other technique) to reduce
the toxic metal concentration. These con-
clusions are valid for elemental metals,
but not necessarily true for compounds of
these metals.
Emissions in Bed Solids. Table 13
shows the inorganic analysis of the bed
reject solids. Several toxic metals exceed
MATE values, although their compound form
may be less toxic than implied by the com-
parison. The anion analysis indicates the
degree of conversion from carbonate to
sulfate within the bed. Table 14 shows the
trace metal content of this material.
Whether the chemical form in which the
metals are present is an unacceptable
hazard for simple land disposal still needs
TABLE 12. INORGANIC ANALYSIS OF PARTICULATE EMISSIONS
Is Land MATE Exceeded at
Substance
Size
1-3 v
Range
3-10 p
Land
MATE
Observed
Value
1/100 Observed
Value
Volatile and Toxic Elements, uq/q'a^
As
Be
Cd
Hg
Pb
Sb
Se
Te
Al
Fe
Si
K
Ca
C (total
carbon)
Cl~
F-
C03"
304"
so3-
s"
NO3~
N02~
45
15
2.1
<0.02
44
4.0
27
<0.5
200,000
60,000
200,000
3,000
30,000
12,000
0.011
0.031
<0.2
9.4
0.001
<0.03
<0.001
<0.001
36
11
1.3
<0.02
43
2.3
22
<0.5
Major
200,000
20,000
200,000
1,500
30,000
11,000
Anion s,
0.007
0.032
<0.2
8.7
0.004
0.03
<0.001
<0.001
0.1
0.03
Yes
Yes
0.004 Yes
0.02
0.1
0.4
0.05
3.0
Elements ,
2.0
0.5
300
1720
32.4
No
Yes
Yes
Yes
No
ug/g
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
Yes
No
Yes
No
Yes
Yes
Yes
NO
Yes
weight percent
(a) Atomic Absorption Spectroscopy Method used except for As
which was determined colorimetrically.
201
-------
to be determined. The concentration of
polycyclic organic materials was found to
be insignificant, as shown in Table 15.
The organic constituents in the bed reject
material pose no environmental problem for
land disposal. Table 15 also shows the
content of hydrocarbons with more than six
carbon atoms; these amounts appear fairly
low.
TABLE 13. INORGANIC CHEMICAL ANALYSIS
OF BED REJECT MATERIAL
Observation
Substance
Observed
Value
Volatile and Toxic
As
Be
Cd
Hg
Pb
Se
Sb
Te
21
2.7
.44
<0.02
8.0
0.8
0.5
<0.5
Land
MATE
Elements ,
0.1
0.03
0.004
0.02
0.1
0.05
0.4
3.0
Exceeds
MATE
yg/g(a)
Yes
Yes
Yes
No
Yes
Yes-
Yes
No
Major Elements, ug/g
Al
Fe
Si
K
Ca
C (noncar-
bonate)
15,000
10,000
15,000
10,000
200,000
1,600
2.0
0.5
300
1720
32.4
Yes
Yes
Yes
Yes
Yes
Anions, weight percent
Cl~
P~
C03=
S04=
SO3=
s-
N02~
N03-
.030
.003
15.1
27.7
.011
.005
•c.OOl
*.001
(a) Atomic Absorption Spectroscopy
Method used except for As which
was determined colorimetrically.
Leachates From Bed Reject Materials.
Leachates were generated in the laboratory
using the spent bed materials (SBM). The
720-hour shake method was employed. This
method involved shaking 33 grains of SBM
with 100 ml of triple-distilled water for
72 hours in a reciprocating shaker at 120
cycles/minute. After 72 hours, the liquid
was decanted and saved. A fresh 100 ml of
distilled water was again added to the same
SBM sample. This process was repeated 10
times to get 720 hours of shaking and
1000 ml of decanted leachates. By the
same procedure, leachates were generated
from the fly ash sample collected from the
second cyclone. Both leachates were
analyzed for inorganics, using Level 1
techniques (SSMS) for elements. Some
Level 2 techniques (wet chemistry or
atomic absorption spectroscopy) were em-
ployed to determine antimony, mercury and
arsenic. Leachates were also analyzed for
anions (SO4=, 803', etc.) by standard wet
chemistry. Organics were analyzed by
Level 1 techniques (liquid chromatography
separation and infrared analysis). Signif-
icant results of inorganic analysis are
presented in Table 16 in comparison with
water MATEs (based on ecological effects).
These results show that As, Ca, Ni, Pb, Li,
Se, SO4= and Al are present in concentra-
tions equal to or exceeding the MATEs.
Hence, these substances should be analyzed
by Level 2 techniques in future leachate
studies to accurately establish their con-
centrations and environmental effects.
Also, the compound forms in which they are
present should be investigated.
The results of organic analysis were
not as conclusive and further work is
needed to determine which specific organic
compounds in leachates are present in
harmful amounts.
Leachate Analysis and the RCRA. The
importance of the above results to FBC
waste disposal will be determined by com-
paring the leachate analyses with the re-
quirements proposed under the Resource
Conservation and Recovery Act (RCRA).
Under Section 3001 of RCRA, waste will be
defined as hazardous if it is inflammable,
corrosive, infectious, reactive, radio-
active, or toxic. Of these criteria, cor-
rosivity, reactivity, and toxicity are
likely to be pertinent to FBC residue.
Based on draft RCRA guidelines, FBC
waste will be considered corrosive if a
saturated solution of the residue in water
has a pH of less than 2 or greater than 12.
The results in Table 16 showed the leachate
from spent bed material to have a pH of
12.2 (corrosive). The pH of a saturated
solution could be expected to be somewhat
higher. However, the pH of the fly ash
leachate was only 9.0. Therefore, if dis-
posed together, the mixture of fly ash and
spent bed material may not be corrosive.
According to draft RCRA guidelines,
FBC residue will be considered toxic if its
leachate (to be generated by a "standard"
method, not yet determined):
(a) has a concentration of any substance
greater than or equal to 10 times
the drinking water standard
(b) has a concentration of any substance
greater than or equal to 0.35 times
the lowest oral mammalian LDso (mg/
kg) for that substance, as listed in
the NIOSH Registry of Toxic Effects
of Chemical Substances
20£
-------
TABLE 14. TRACE METALS IN BED REJECT MATERIAL<
MATE Value for Land 1
Element
Li
Be
B
P
Na
Mg
Al
Si
P
S
Cl
K
Ca
Sc
Ti
V
Cr
Mn
Fe
CO
Ni
Cu
Zn
Ga
Ge
As
Se
Br
Rb
Sr
Y
Zr
Nb
Mo
Ru
Rh
Pd
Minimum
Value
vg/g
0.51
0.0300
324
75
90
36.5
2
21
0.001
—
0.0200
1720
32.4
1610
2
0.300
0.5
0.200
0.500
3
0.02
1
1
14.9
18
0.100
0.05
1
3640
92.2
30
15
650
14
, 0.300
—
Observation
Exceeds
MATE
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Observation
pg/g
200
0.3
30
0.3
300
20%
1.5%
1.5%
50
1.5%
40
1%
20%
3
200
20
30
100
1%
1.5
150
15
400
7
40
20
3
3
60
300
1.5
0.5
2
<1
<1
<0.5
MATE Value for Land 1
Element
Ag
Cd
In
Sn
Sb
Te
I
Cs
Ba
La
Ce
Pr
Nd
Sm
Eu
Gd
Tb
Dy
Ho
Er
Tm
Yb
Lu
Bf
Ta
W
Re
Os
Ir
Pt
Au
Hg
Tl
. Pb
Bi
Tn
U
Minimum
Value
ug/g
0.500
0.00400
3
3
0.400
3
30
2460
2.20
—
38.4
1540
1580
—
—
—
—
380
— —
—
—
—
—
1.50
150
30
—
—
—
0.0600
—
0.0200
0.5
0.100
23800
1.26
6
Observation
Exceeds
MATE
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Observation
ug/g
<0.1
<0.5
<0.6
2
1
<0.4
<0.07
4
30
7
7
1
1.5
0.5
3
1
0.3
<0.1
<0.3
<0.1
<0.3
<0.3
TO. 3
<0.3
<0.4
<0.4
<0.4
<0.2
<0.4
<0.1
<0.4
<0.2
2
<0.15
0.2
0.4
(a) Spark source mass spectrographic analysis.
203
-------
TABLE 15. ORGANIC CHEMICAL ANALYSIS OF BED
REJECT MATERIALS
Substances
Land
MATE
Observation
Polycyclic Organic Materials, nq/q
Anthracene/phenanthrene 14,500 <0.1
Others specifically sought No data <0.1
Total
Hydrocarbon Content Above C6, ug/g
Hydrocarbon according to
boiling point range
C7 No data 20
C8 Ditto 22
C9 • 12
CIO • 4
Cll " 5
C12 " 3
>C12 " 17
Reduced Sulfur and Other Compounds
Reduced sulfur, ug/g No data <0.2
Other compounds detected Ditto None
TABLE 16. INORGANICS IN LEACHATES
FROM SPENT BED MATERIAL
(SBM) AND FLY ASH (FA)
Substance
or
Parameter
Li
Na
Mg
Al
Ca
V
Fe
Ni
Se
Pb
As
CN~
S04*
PH
Concentration ,
nw/1
SBM
6
10.4
9.6
o.e
460
0.1
0.2
0.03
0.07
0.05
0.04
<0.03
1610
12.2
FA
20
52
6.4
1.4
1000
0.1
0.2
0.03
0.07
0.05
0.05
<0.03
1950
9.0
Hater
MATE
mg/1
0.38
800
87
1.0
16.2
0.15
0.25
0.01
0.025
0.05
0.05
0.015
250 (a)
5-9 l«)
MATE
Exceeded?
Yes
No
No
Yes
Yes
No
No
Yes
Yes
No
No
Undecided
Yes
Yes
(a)
Proposed US EPA drinking water standards;
no HATE value available.
TABLE 17. RESULTS OF THE BIOLOGICAL TESTING OF FBC WASTE STREAMS
Test
Ames
Cytotoxicity
HI-38
RAM
Rodent Acute
Toxicity
Aquatic
Freshwater
Algal
Daphnia
Fish
Saltwater
Algal
Shrimp
Fish
Terrestrial
Soil Microcosm
Stress Ethylene
Haste Streams
Test Flue Suspended Bed
Parameter Gas Particulates Solids
V- +
LD50
LD50 NA NA
U>50 NT
BC50
1*50
"=50
EC50
IC50
LC50
Ranked in 23
order of
toxicity
Percent of 0%
increase
over
control
Bed
Solids
Leachate
-
NA
NT
45%
40.9%
25.3%
NT
NT
NT
1
NA - Not available.
NT » No toxicity.
204
-------
(c) has a concentration of any substance
equal to 10 times the lowest 96-hour
LCso (mg/1) for that substance, as
listed in the NIOSH Registry.
In these studies, none of the primary
drinking water standards are exceeded by
a factor of 10 by either leachate. How-
ever, the results still must be measured
against criteria (b) and (c) above.
If the MATE values for calcium are
similar to the primary drinking water
standards, calcium in leachates may well
exceed the allowable standards by more
than a factor of 10, thereby causing the
FBC residues to be considered hazardous.
The importance of the designation
"hazardous" lies in the somewhat stricter
disposal requirements likely to be imposed,
and the additional permits, testing, and
record-keeping required by RCRA.
Bioassay Results. The biological
test results for four FBC waste streams
are given in Table 17. The health and
ecological tests on each stream were per-
formed according to the pilot Level 1 bio-
assay program. The results of the cyto-
toxicity tests are still being processed
at this time.
The flue gas stream was tested only
with the stress ethylene test. At present,
no other Level 1 tests are suitable for
testing gases. Results indicate the gas
was nontoxic. Some caution is associated
with this conclusion because the full
quantity of gas required for the test was
not available and the sample had to be
stored for some time before testing.
Several tests were performed on the
spent bed solids. This stream showed very
low or no toxicity in all tests, and would
not likely constitute a biological hazard
in its solid form. The leachate from the
spent bed solids showed some toxicity in
two of the tests. The results show that it
was nontoxic to mammaIB (health tests) and
marine organisms, but toxic to freshwater
and soil organisms. This stream would
likely require further study to determine
its potential biological hazard.
Only the suspended particulates stream
gave positive Ames test results. The par-
ticulates were mutagenic and thus may also
be carcinogenic. This stream was also
toxic to soil organisms.
Based on the biological tests alone,
the relative ranking of the four waste
streams, in order of decreasing toxicity,
is:
(a) Spent bed solids leachate
(2) Suspended particulates
(3) Spent bed solids
(4) Flue gas.
It must be emphasized that the test re-
sults provide only relative data, and the
actual hazard to humans and other orga-
nisms can be determined only through addi-
tional testing.
Summary and Conclusions on
Comprehensive Analysis Results
On Source Emission Data
The major conclusion is that compre-
hensive analysis of emissions from emerging
energy technologies yields useful results
for completing the environmental assess-
ment of the processes. Other conclusions
are:
(1) Pressurized coal-burning FBC units
can meet existing New Source Per-
formance Standards for SO2 and NOX
emissions from coal-fired steam
generators. Particulate emissions
control needs demonstration.
(2) MATES for SO2» N°x» CO and possibly
for other substances need
reevaluation.
(3) Polycyclic organic matter (POMs) in
flue gas or other effluents from FBC
units do not appear to be health/
ecological hazards. POMs are con-
centrated in fine particles (<3
microns) as opposed to coarse (>3
micron) particles in the suspended
particulates.
(4) Though biological assay data are
difficult to interpret at this stage,
spent bed material leachate and sus-
pended particulates do show a rela-
tive higher toxicity than flue gas
and bed solids. This trend corre-
sponds with the greater than MATEs
concentrations of many volatile and
toxic trace elements (As, Ni, Pb, Li,
etc.) in leachates from bed mate-
rials. These results, therefore,
indicate the need for further study.
The final interpretation of the bio-
assay results and their toxicity
ratings is being considered by the
US EPA.
(5) The results of this study do not
imply that fluidized-bed combustion
of coal generates solid wastes of
greater or lesser toxicity than other
methods of coal combustion, since the
solid wastes from other methods have
not been subjected to such compre-
hensive analysis. Also, the above
results need careful evaluation and
further validation.
Careful attention to insure steady-
state operation of the process is necessary
before beginning sampling. Adequate steps
205
-------
to insure sufficient coal/sorbent supply
of uniform composition to last the duration
of sampling is important for obtaining use-
ful and reliable data.
Acknowledgments
The studies described in this paper
are part of the environmental assessment
of the fluidized-bed combustion process
which Battelle is conducting for the US
EPA, under the guidance of R. P.
HangebraucJc and D. Bruce Benschel of the
Agency's Industrial Environmental Research
Laboratory, Research Triangle Park, North
Carolina. The comprehensive analyses of
emissions were conducted in cooperation
with many individuals of EXXON. The
authors sincerely appreciate the assis-
tance and support of all these persons.
List of Acronyms
CA Comprehensive Analysis
FBC Fluidized-Bed Combustion
SASS Source Assessment Sampling System
SSMS Spark Source Mass Spectroscopy
AAS Atomic Absorption Spectrometry
AAS/GF Atomic Absorption Spectrometry/
Graphite Furnace Atomization
AAS/HG Atomic Absorption Spectrometry/
Hydride Generation
ASTM American Society for Testing and
Materials
FAAS Flameless Atomic Absorption
Spectrometry
POM Polycyclic Organic Matter
LRMS Low Resolution Mass Spectrometry
GC/MS Gas Chromatography
GC/FPO Gas Chromatography/Flame Photo-
metric Detection
GC/FID Gas Chromatography/Flame loniza-
tion Detection
FTIR Fourier Transform Infrared
Spectroscopy
SEM Scanning Electron Microscopy
NBS National Bureau of Standards
SRM Standard Reference Materials
LLL Lawrence Livermore Laboratory
ORNL Oak Ridge National Laboratory
MATE Minimal Acute Toxicity Effluent
RCRA Resource Conservation and Recovery
Act
NIOSH National Institute for Occupa-
tional Safety and Health
References
(1) D. B. Henschel, "The EPA Fluidized-Bed
Combustion Program: An Update", Fifth
International Conference on Fluidized-
Bed Combustion, Washington, D.C.,
December 12-14, 1977.
(2) J. G. Cleland and G. L. Kingsbury,
"Multimedia Environmental Goals for
Environmental Assessment", Vols. 1
and 2, Final Report by Research Tri-
angle Institute of U.S. Environmental
Protection Agency/ Contract No. 68-
02-2612, EPA-600/7-77-136a and b
(November 1977).
(3) J. W. Hamersma, et al., IERL-RTP Pro-
cedures Manual; Level 1 Environmental
Assessment, EPA-600/2-76-160a, June
1976, U.S. Environmental Protection
Agency, Washington, D.C., pp 29-46.
206
-------
ENVIRONMENTAL ASSESSMENT PROGRAM FOR THE HYGAS PROCESS
by
L. J. Anastasia, W. G. Bair, and D. P. Olson
Institute of Gas Technology
3424 South State Street
Chicago, Illinois 60616
ABSTRACT
The main objective of the HYGAS^ Environmental
Assessment Program is to systematically obtain
and interpret experimental data from the HYGAS
pilot plant to estimate pollutant production
for demonstration and commercial-scale HYGAS
coal gasification plants. Priorities have
been established for operating systems for
sampling, analyses, and data evaluation to
define the fate of potential pollutants
generated during pilot plant operation. The
assessment methodology for environmental data
acquisition and interpretation features these
sequential objectives:
• identification of potential pollutants
in plant effluent streams,
• development of sampling, preservation,
and analytical techniques.
process unit, strea
and
and species selection,
• quantitative descriptions of significant
pollutants.
Introduction
The Institute of Gas Technology (IGT) is
performing an environmental assessment of the
HYGAS Process under contract to the U. S.
Department of Energy (DOE). HYGAS is a
second-generation coal gasification process
which uses hydrogen in a fluid!zed bed at
high pressures and temperatures to maximize
production of high-Btu substitute natural gas
(mostly methane) from all types of coal.
About two-thirds of the total methane produced
in the process is generated by direct conversion
in the gasifier.
The process is currently being developed
at the HYGAS pilot plant under funding from
DOE and the American Gas Association. The
pilot plant has a design capacity of 3 tons/hr
of coal to produce a nominal 1.5 x 106
SCF/day of pipeline quality gas at 1000
Btu/SCF. The main objective of the HYGAS
Environmental Assessment Program is to
systematically obtain and interpret experimental
data from the pilot plant to estimate pollutant
production from demonstration- and commercial-
scale HYGAS coal gasification plants. Priorities
have been established for operating systems
for sampling, analysis, and data evaluation
to define the fate of potential pollutants
generated during pilot $lant operation.
The HYGAS Process
Figure 1 shows the current processing
steps and the major plant streams for the
HYGAS pilot plant. After the coal is crushed
and dried, the pretreatment step is used only
for agglomerating coals, as pretreatment is a mild
surface oxidation used to destroy agglomerating
tendencies. Coal is introduced into the
high-pressure reactor in a light oil slurry
containing up to 45 weight percent coal. The
coal passes through four fluidized beds
during gasification and spent char is removed
from the bottom of the gasifier. Crude
product gas leaves the gasifier and passes
through a cyclone separator to a quench
system for the separation, recovery, and
recycling of both the light oil and water.
Acid gases (H2S and CC^) are removed from the
cooled product gas with a diglycolamine-water
solution which is regenerated and recycled.
Subsequent washing steps further prepare the
gas for upgrading to essentially pure methane
(SNG) by methanation of residual CO and H2 in
a packed bed of nickel catalyst.
The Environmental Assessment Program
Pilot Plant Sections of Interest
The environmental assessment is being
focused upon portions of the pilot plant
considered "scaleable" to larger plant designs
(Figure 1). A scaleable unit is defined as
an existing pilot plant unit which will have
a duplicate counterpart in a demonstration or
commercial HYGAS plant (such as the pretreater
and the hydrogasifier). Other parts of the
pilot plant (such as the wastewater treatment
section) are designed specifically for pilot
plant operation and are not scaleable.
Because a wide variety of process designs
and flow sheets can be expected for the
demonstration and commercial plants, many
HYGAS process variations are being considered
and evaluated. A methodical chemical engineering
analysis of existing facilities is necessary
to characterize probable pollutant distributions
in streams designed into the large-scale
plants.
Definition of Steady-State Operation
Data acquisition from the pilot plant
requires defining the concept of "steady-
state operation". Normal operation of develop-
mental pilot plants, such as HYGAS, is intermittent
with considerable periods of unsteady operation.
Problems hampering continuous operation
include process upsets and mechanical failures.
The criteria for "steady-state" operation has
been defined by HYGAS plant operators according
to the following sequence:
• steady feed rates (of coal or pretreated
char) to the gasifier,
• constant density of feed slurry to the
gasifier,
207
-------
• self-sustained operation of the gasifier
(that is, all heating requirements for
the gasification reactions are supplied
by the steam-oxygen section with no
external heating),
• steady fluidired-bed levels in the four
beds of the reactor,
• essentially constant off-gas composition
and solids flow rate from the gasifier,
and
• steady operation of downstream process
units, such as the quench and light oil
stripping sections, which influence
operation of the gasifier.
All pilot plant recording instruments
transmit process data to a central data
processing system which files the inputs for
future retrieval and data reduction by computer
calculations. Steady-state periods are
defined during a subsequent review of the
operating data and are normally identified
after completion of each HYGAS test.
Sampling Techniques
Once steady-state, self-sustained operation
is achieved in a specific HYGAS test, routine
grab sampling of the process water, solids,
and oil streams is started. In addition to
specialized sampling for specific purposes
(such as time series studies for specific
pollutants), on-line analytical instrumentation
is also being developed to study sulfur
species in process gas streams plus total
organic carbon (TOC) and total oxygen demand (TOD)
in process water streams. Techniques which
are currently applicable to the analyses of
solid, liquid, and gaseous phases are listed
in Table 1.
Sample Preservation
Analyzing HYGAS wastewater samples
presents significant problems because known
preservation and analytical techniques are
not effective for all species. HYGAS wastewaters
are not unique in this respect; similar
results have been observed in other coal
gasification plants. State-of-the-art procedures
developed in earlier HYGAS work were the
starting point for analyzing the water samples.
Appropriate refinements to the wet-chemical
water analyses have been made due to continuing
laboratory studies of preservation and analytical
techniques at IGT and Carnegie Mellon University
(CMU). Efforts to improve sample preservation
and analysis at each of the coal gasification
pilot plants are being coordinated by CMU.
The cooperative efforts between IGT and
CMU have established that current preservation
methods for the water samples prevent significant
degradation with storage time and that HYGAS
samples can be analyzed off-site, preferably
within 24 to 72 hours after collection. The
comprehensive data base includes sets of
daily mean concentration values for each
pollutant in each water stream as well as
simultaneous process conditions such as
pressures, temperatures, and flow rates.
These data are later evaluated to normalize
the results to obtain pollutant production
rates in terms of coal feed on a moisture and
ash free basis. A set of interim results for
a recent HYGAS test with Illinois bituminous
coal is given in Table 2. These data are not
yet final because the process data are still
being refined and applied to the normalization
procedure.
Sulfur-by-Species Analyses
A sulfur-by^species analysis has been
conducted on the process solids streams as
part of the environmental assessment. These
solids have included the coal feed, the
pretreated char, the spent char, and char
from the first and second gasification stages.
The gasification of sulfur by species
for one HYGAS test, from fresh coal feed to
the pretreater through the gasified spent
char, is reported in Table 3. Overall, the
sulfur gasified in this HYGAS tost was 27%
during pretreatment, 69% and 79% for the
first and second gasification stages, respectively,
and 96% for the spent char. Pretreatment was
more severe than expected under optimized
conditions because gasifier operability was
the main objective for this HYGAS test.
Minor and Trace Elements
The fate of minor and trace elements in
the solids is also being monitored during
the pretreatment and gasification steps.
These trace element recoveries in solids
obtained from gasification of subbiluminous
coal (with no pretreatment) are shown in
Table 4. As shown in the table, these
elements can be grouped into several categories
according to their recovery in the spent
char. In these tests, the more volatile
elements were chlorine, mercury, selenium,
and (in one test) cadmium. The trace elements
are expected to exhibit individual reactivities
in the various gasification steps, depending
on the chemical state of the elemental species
and the processing conditions (such as large
temperature differences and oxidizing or
reducing atmospheres). The HYGAS process is
not unique in this respect, but is comparable
to other industrial processes (power plants,
steel mills, glass plants and other coal
gasification processes) where large temperature
gradients exist.
Light-Oil Composition
The light oil used to slurry the coal
feed for introduction into the high-pressure
gasifier is being sampled and analyzed by
GCMS techniques for organic composition.
This oil is recovered and recycled during
each pilot plant test. The HYGAS process
will produce a net make of light oil (benzene,
toluene, and xylene) and toluene is used in
the pilot plant to represent the oil product.
The oil used in the coal feed slurry represents
a concentrated source of polynuclear aromatics,
potential carcinogens, and potentially toxic
208
-------
organic compounds. As each HYGAS test progresses.
this oil changes in composition as a function
of operating time and processing conditions.
Changes in composition as a function of
operating time and coal type are shown by the
data in Table 5. The differences in oil
composition near the beginning of a test
occur because the plant inventory of approximately
6000 gallons remaining after a pilot plant
test is reused to slurry the coal feed for
the subsequent test.
Conclusion
A brief overview of the environmental
assessment program for the HYGAS process has
been presented. Data from this program are
now providing a strong baseline characterization
of pollutant species in major process streams.
The program is now entering an advanced phase
which requires a more thorough material
balance characterization of specific pollutants
around the scaleable HYGAS pilot plant units.
CLEANED (MS
TOMONEUTOft
Figure 1. SCALEABLE PROCESS UNITS AND MAJOR STREAMS IN THE HYGAS PILOT PLANT
(Major Streams: Reference Numbers 1-13)
Table 1. SUMMARY OF APPLIED ANALYTICAL Table 2. WATER-SOLUABIE POLLUTANT PRODUCTION:
TECHNIQUES FOR HYGAS PROCESS STREAM CHARACTERIZATION HYGAS TEST 64, ILLINOIS NO 6 COAL FROM PEABOOY
»»ei«« idttru. MINE 10 (INTERIM RESULTS)
Nui «B»rtioicopy'
Gu OnOMtOfnpajr
Suite
Atomic
Hydrocarbon
ttftiacabtm*
Nui Spoctracapr
Atoaic AbtoptioB
GOB'
IOC
TOO Analyser*
act-lab Aulrtlt
Hus Spoctnscopy
Atonic Absorption
X-tay Diffraction
•it-lab Analysis
Sim Analysis
Tnco Elaaonts
Trace Orioles (TNA)
TOC
TOD. (COD. MO)
Of. HJ,. «-,. F-.
".S™ COO,510D..
als. TOC, ft^s*
Solid IttMSt
Sp«ci«»
Tot»l Dissolved Solids
Phenols
Cyxnide
Totml Orfinic Carbon
Iliiocyuute
Sulfid*
Oilorida
Pretr*«ter
Ib/Ton Co»l (MAT)
67
S.2
0.0002
12
1.7
0.0004
0.4
S.4
Guifier
Ib/Ton Oi«r (MAT)
11
10.2
0.0006
20
1.1
3.3
24.9
1.2
Table 3. GASIFICATION OF SULFUR BY SPECIES FOR HYGAS
TEST 64, KITH ILLINOIS BITUMINOUS COAL
TO, lUr
i Alkalinity.
(Oil
Dissolved Oxgna. TSS.
pH, a". TOO. S-Total.
K-total. SCH . SOj . S04
Coil
Moor od Trmc* U«MU
NUer nd trie* U«r««T««
S-total
N-total
PirticaUtM
S.ftul
N-IMal
Sulfide
Sulfate
Organic
Pyritic
Total
Ib Sulfur/100 Ib Coal
0.04
0.21
2.57
1.71
Spent
Char
0.012
0.007
0.1S9
0.022
0.200
% Art. » VoUtllM
% NtUtan
Hut Vala*
1. Bvirieal nalrais of all lUtW
2. Oa-liao coBtlBunu aal]n*r.
S. to» CkroHtotnpk-Nu
209
-------
Table 4. TRACE ELEMJNT* RECOVERY IN
SPENT CHAR
Table 5. LIGHT OIL COMPOSITION IN HYGAS
COAL-FEED SLURRY
HYGAS TESTS SS AND S<: ROSEBUD SUBBITtJNINOUS COAL
90 TO 100%
. Zn, U. Cr. Pb. Ba. U, M^
SO TO 90*
B. Qi. F. Fe. No. Tl. Ba, M^. Pb
LESS THAN 50%
Cl. Uf. Se. CA
m Elements belov detection limit* in both feed md char
vere Sb. Co. Te, md SB.
t> Underlined elewmts could be placed in more than one froup.
Coal Type
Days of Operation 1_
Umite
I li
SubbltiMinou*
Aliphatlei
Aroaatics~
One- Ring
TWO- Ring
Three- Ring
Four- Ring
Five- Ring
Hlfcellaneaus
Unknowns
0.5 1.1
97.5 96.6
0.7 1.8
0.06 0.1
0.013 0.027
— —
— —
1.2 0.3
6.8
85.7
5.0
0.9
0.21
0.01
0.7
0.6
7.3
84.6
5.5
0.8
0.13
0.004
0.7
0.8
210
-------
ANALYSIS OF SYNTHANE/SYNTHOIL PRODUCTS
A. G. Sharkey, Jr. and J. L. Shultz
U. S. Department of Energy
Pittsburgh Energy Research Center
4800 Forbes Avenue, Pittsburgh. PA 15213
Abstract
An overview of the compositional data for the main
streams from coal gasification and coal liquefac-
tion processes, important for environmental as
well as process development considerations, is
given. A summary of the information available
concerning the chemical constitution of coal,
derived by some of the latest spectral techniques,
is presented. Challenges for the analyst in terms
of developing methods to obtain data required for
product characterization for some of the newer
fossil energy processes are highlighted.
Introduction
Increased emphasis on developing processes for
substitute gaseous and liquid fuels from coal has
highlighted the need for additional compositional
data for coal as well as for many of the process
streams. The purpose of this contribution is to
summarize some of the latest information concern-
ing the so-called chemical structure of coal as
determined by newer spectral techniques and to
give an overview of the composition of the streams
common to many gasification and liquefaction
processes under development.
Data from the SYNTHANE POT gasifier(1) and 1/2
ton per day liquefaction unit at the Pitts-
burgh Energy Research Center are used to describe
the composition of the various process streams.
Table 1 shows the process streams from the SYN-
THANE gasifler Including, in addition to the feed
coal, the gas, char, condensate water, and tar.
Streams from the liquefaction PDU are shown in
Table 2 and include the feed coal, the feed paste,
which is a mixture of feed coal and recycle oil
from the centrifuged liquid product, the raw
product, centrifuged residue consisting of
unreacted coal and mineral matter and some of the
scrubber effluents.
Discussion
Structure of Coal
The structure of coal was a major topic for re-
search both here and abroad in the 1950's and
early 1960's. During the 1960's and extending
into the 1970'a, the study of basic coal chemistry
decreased; the remaining effort was concentrated
in a few laboratories, including the U. S. Bureau
of Mines at Bruceton, Pennsylvania.
The following is « summary of information obtained
from a number of sources including the Chemistry
of Coal Utilization and the Supplementary Volume
by.Lowry,* * books by Francis and van Krevelen,
^ several publications by Given and others,
and a recent excellent review by J. Gibson de-
livered, as the Bobens Coal Science Lecture in
1977. ' Also Included are results of several
studies at the Pittsburgh Energy Research Center.
Reflectance and x-ray diffraction are two of the
techniques used extensively in early studies of
the structure of coal. Reflectance has the ad-
vantage that individual particles can be Investi-
gated as against bulk property determinations by
most other techniques. Reflectance is related to
the refractive index which in turn Is related to
the molar refractivity, an additive property of
atoms and bonds. Coking ability and similar
utilization information can be Inferred from
reflectance measurements.
X-ray data show that coal is not crystalline but
contains much amorphous material depending on the
rank of the coal. The types of Information that
can be provided by x-ray diffraction are the sizes
of the layers and the number of layers In parallel
stacks. The consensus.of information derived by
Hirsch and coworkers indicates that in the 78
to 94 percent carbon range the largest clusters of
atoms contain approximately 32 atoms. The amor-
phous content decreases with increasing rank. Low
rank coals have small layers that are randomly or-
iented. Medium rank coals show layers of moderate
orientation, while high rank coals into the anthra-
cite range contain the largest layers and have a
high degree of orientation.
Aromaticlty (f • aromatic carbon/total carbon)
can be derived oy a number of techniques Including
magnetic resonance, density determinations, sound
velocity, heats of combustion and ultraviolet
spectrophotometry. The range in f is from ap-
proximately .7 for low rank coal to 1 for the
higher rank coals. These data were derived pri-
marily fromgX-ray determinations by Hirsch and
coworkers. J. K. Brown' obtained hydrogen
aromaticity values from infrared spectra. In low
rank coals about 80 percent of the hydrogens are
on aliphatic carbons while anthracitic coals show
essentially zero association of hydrogens with
aliphatic carbons.
Data for the heteroatoms, oxygen, nitrogen and sul-
fur, are meager and less authentic than for carbon
and hydrogen. Total oxygen is determined by dif-
ference. Unweathered coals do not indicate car-
bonyl oxygen. At present there Is no direct way
to establish how nitrogen and sulfur occur in
coals. The information obtained thus far has been
derived from pyrolysls products such as tars or
from extracts of coals. Pyridine nucleus com-
pounds predominate and benzothiophene is many
times used as a model for the sulfur in coal.
Coal liquefaction products contain both basic and
acidic nitrogen in the fora.of pyridine nucleus
compounds and carbazoles.
Information concerning possible hydroaroaatic rings
in coals has been obtained by catalytic dehydro-
genation and other methods. The indications are
that 40 to 50 percent of the total hydrogen.In
bituminous coal Is in hydroaromatic rings.
211
-------
Vapor phase osmometry and other techniques have
been used to determine the molecular weight of
solubilized material from coals. Intermolecular
interactions can lead to erroneous results but
values have ranged up to several thousand.
Several attempts have been made to reconcile the
somewhat divergent data concerning the aromatlcity
of coal. As previously mentioned, Hirsch and co-
workers have evidence from x-ray data that the
carbon atoms in coal are primarily aromatic while
the infrared data of Brown concerning the distri-
bution of hydrogen atoms indicates that the hydro-
gen la.primarily.associated with aliphatic carbons.
Given,1 ' Gibson* ' and H111UZ' have proposed
models that incorporate much of the structural
Information derived by a variety of techniques.
Perhaps the most popular model is that proposed
by Given in 1960 In which he used an 82Z carbon
coal and the associated data to derive his model.
The carbon aromaticity of the model was .69, con-
sistent with much of the data for coal with this
carbon content. The mean structural unit used in
Given's model contains two condensed rings.
Composition of Coal Gasification Products
Mass spectrometry has played a major role. In the
analysis of coal gasification product streams. Gas
analysis methods, including techniques for the
analysis of trace and minor components, were stand-
ardized by the petroleum laboratories la the
1940's. Analyses of several streams from the
SYNTHANE process are used to illustrate the appli-
cation of mass spectrometry to coal gasification
products. In addition to the major gaseous compo-
nents (i.e., hydrogen, methane, carbon monoxide,
ethane, and carbon dioxide), Table 3 lists the
concentrations of traca.components found In •
typical product gas. Hydrogen sulflde, at
6500 ppm, is of the greatest concern from the
environmental point of view. An analysis of the
methylene chloride extract of a condensate water
is shown in Table 4. ' The extract is In the
range of 0.6 to 2.4 weight percent of the water;
60 to 80 percent of the extract is phenolic In
character. A relatively small amount of tar,
highly aromatic in nature, la also formed as a
by-product of coal gasification; a low ionising
voltage analysis of this tar Is shown in Table 5.
An appreciable quantity of heterocyclic
material containing oxygen, nitrogen, and sulfur
has been identified.
Composition of Coal Liquefaction Products
The analysis of coal liquefaction products re-
quires s combination of solvent fractionalIon,
chromatographic separation, and chemical deriva-
tization prior to application of spactrometrlc
methods, including Infrared and mass spectrometry.
The centrlfuged liquid product (CLP) from the 1/2
TPD liquefaction unit at the Pittsburgh Energy
Research Center Is used to Illustrate the analy-
tical techniques involved. The first solvent
separation, with benzene and pentane (hexane and
cyclohexane have also been used) yields a heavy
oil and an aaphaltene fraction. The heavy oil,
-701 of the CLEj,is further separated, using the
SARA technique or other chromatographic
methods, into saturate, aromatic, basic nitrogen,
neutral nitrogen, and acid (phenolic) fractions.
The saturate and aromatic fractions are amenable
to mass spectrometrlc type-analysis methods
developed.for petroleum and coal liquefaction prod-
ucts. '*' Table 6 shows a typical analysis of
s heavy oil, combining the mass spectral data for
the saturate and aromatic fractions with the chro-
matographic data for the other fractions. The
nitrogen heterocyclics have been analyzed by low
ionizing voltage; the oxygen-containing species
have been classified as alkylated phenols, indanols/
tetralinols, phenylphenola, and cyclohexylphenols.
* ' New analytical methods will be required to
further identify the specific compounds present in
these fractions. Methods are not available to de-
termine the degree of hydrogenation of the various
ring systems, particularly in the basic nitrogen
fraction. Thin-layer chromatography and capillary
gas chromatography are possibly the best solutions
to this problem but are hindered by the lack of
representative pure compounds with hydrogenated
ring systems.
Screening for Hazardous Compounds
High-resolution mass spectral data (HBMS) are rou-
tinely computer processed to obtain the precise
masses and elemental compositions of all ions in a
mass spectrum. To screen for toxic or hazardous
compounds, an array of the precise masses of ap-
proximately 300 chemicals whose toxic limit values
are known was added to the mass spectral data pro-
cessing program. Screening is effected by computer
matching the experimental masses against the toxic
compound masses contained in the reference list.
While the elemental composition can be determined,
the particular isomerlc form cannot be identified
by HUMS data alone. This preliminary screening,
however, limits the number and class of possible
hazardous compounds that must be considered, thus
defining the analytical determinations required
for specific Identification. Table 7 lists a por-
tion of the screening data obtained for coal
liquefaction process streams.
Analytical problems concerned with the conversion
of coal to gaseous and liquid fuels present many
challenges to the analyst. These challenges are
summarized In Table 8. Methods are needed for
the direct determination of the physical and
chemical properties of coal. Data for several of
the prime species, such as the heteroatoms sulfur,
oxygen and nitrogen, are obtained by indirect
methods. Speciation is Important particularly
from the viewpoint of possible pollutants from
the use of fossil fuels. The characterization of
asphaltenes in coal-derived liquids is important
as asphaltenes are thought by many to be inter-
mediates in the conversion of coal to liquids.
The extent of conversion is generally based on
the yields of oil and asphaltenes. Currently,
many methods are in use but standard methods are
not available that could provide reliable com-
parisons of conversions from various processes.
Only limited data for a few elements are avail-
able showing the fate of potentially hazardous
elements in coal gasification and liquefaction
processes. Elemental balances require determi-
nations in various matrices and new methods must
be developed. The concentration and nature of
•various heteroatoms, the source of many environ-
mental pollutants, must be determined. Charac-
terization of the heavy-ends is difficult, as
many high molecular weight components with low
212
-------
volatility are included in the residues. Compo-
sitional data for the residues are Inportant from
the viewpoint of waste disposal. Many of the
current coal conversion processes are catalytic
and, particularly in the case of liquefaction,
commercially available catalysts used in the
petroleum industry are not adequate as daactiva-
tion occurs quickly. In summary, many of the
standard methods available from the petroleum
industry cannot be applied directly to coal-
derived fuels and new methods must be developed.
References
1. Forney, A. J., S. J. Gaslor, W. P. Haynes,
and S. Katell, BuMines TPR-24, Department of
the Interior, Washington, 1970.
2. Yavorsky, P. M., S. Akhtar, and S. Friedman,
Chen. Eng. Prog., 69, 1973, p. 51.
3. H. H. Lowry, ed., "Chemistry of Coal Utili-
zation," Vols. 1 and 2, Wiley, New York,
1944; Supplementary Volume, ed. I. G. C.
Dryden, Wiley, New York, 1962.
4. Francis, W., "Coal," 2nd edn., Arnold, Lon-
don, 1961.
5. Van Krevelen, D. W., "Coal Science, Elsevler,
Amsterdam, 1961.
6. Given, P. H., Fuel. v. 39, 1960, p. 147.
7. Gibson, J., "The Constitution of Coal and Its
Relevance to Coal Conversion Processes," The
Robens Coal Science Lecture, London, 3rd
October, 1977 under the auspices of the
British Coal Utilization Research Associa-
tion Ltd.
8. Hlrach, P. B., Phil. Trans. Roy. Soc., 252A.
1960, p. 68.
9. Brown, J. K., J. Chen. Soc., 1955, p. 744.
10. Shultz, J. L., C. M. White, F. K. Schweig-
hardt, and A. G. Sharkey, Jr., PERC/RI-77/7.
11. Reggel, L., I. Wender, and R. Raymond, Fuel,
v. 47, 1968, p. 373.
12. Hill, G. R. and L. B. Lyon, Ind. 6 Eng. Chem.,
v. 54, 16, 1962.
13. Sharkey, A. G., Jr., J. L. Shultz, C. E.
Schmidt, and R. A. Frledel. PERC/RI-75/5.
14. Schmidt, C. E., A. G. Sharkey, Jr., and R. A.
Frledel, PERC/TPR-86, 1974.
15. Jewell, D. M., J. H. Weber, J. W. Bunger, H.
Plancher, and D. R. Latham, Anal. Che*., v.
44, 1972, p. 1391.
16. 1972 Annual Book of ASTM Standards.
17. Swansiger, J. T., F. E. Dickson, and B. T.
Beat, Anal. Chem., v. 46, 1974, p. 730.
18. Schweighardt, F. K., C. M. White. S. Friedman,
and J. L. Shultz, Preprints of ACS Div. of
Fuel Chemistry, v. 22, #5, August 29-September
2, 1977. Chicago. Illinois, p. 124.
19. Lett, R. G., C. E. Schmidt, R. R. DeSancls,
and A. G. Sharkey, Jr., PERC/RI-77/12.
Table 1. Process Streams From a SYNTHANE Gaslfler
Feed Coal
Gas
Char
Condensate Water
Tar
Table 2. Process Streams From a SYNTHOIL PDU
Feed Coal
Feed Paste
Recycle Oil (CLP)
Raw Product
Centrlfuged Residue
Scrubber
Table 3. Trace Components In Gas From Coal
Gasification
Compound
H,S
COS
Benzene
Toluene
C_ Aromatic
Tniophene
C.-thiophene
Cj-thlophene
Kethylmercaptan
Concentration (ppm)
6,500
107
480
66
34
43
24
5
80
Table 4. Condensate From Gasification
of Illinois No. 6 Coal
Compound
Phenol
Cresols
C.-Phenols
C.-phenols
Dlhydrlc phenols
Indanola ~j
AcetophenonesJ
Bydroxybenz aldehyde")
Benzole acids J
Naphthola
Indenols
Blphenols
Benzothlophenols
Weight, ppm
3400
2840
1090
110
250
150
60
160
90
40
110
213
-------
Table 5. Mass Spectroaetrlc Analyses of the
Benzene Soluble Tar, SYNTHANE
Gaslfler
Structural type
(including alkyl
derivatives)a/
Ho.
Illinois . ,
6 coal.^'
vol. Z
Benzenes 2.1
Indenes 8.6—
Indans 1.9
Naphthalenes 11.6
Fluorenes 9.6
Acenaphthenea 13.5
3-ring aromatics 13.8
Phenylnaphthalenes 9.8
4-ring peri-condensed 7.2
4-ring cata-condensed 4.0
Phenols 2.8
Naphthols £/
Indanols 0.9
Acenaphthenola
Phenanthrols 2.7
Dlbenzo furans 6.3
Dlbenzothlophenea 3.5
Benzonaphthothigphenes 1.7
N-heterocycllcs=7 (10.8)
— Average molecular weight, 212.
— Spectra indicate traces of 5-ring
aromatics.
— Includes any naphthol present (not
resolved in these spectra).
—Data on N-free basis sine* isotope
corrections were estimated.
Table 6. Typical Analysis of Bexane Soluble Material
From a Coal Liquefaction Product
Compound Class
Paraffins
Non-condensed Naphthenes
2-rlng Naphthenes
3-rlag Naphthenea
4-ring Naphthenes
5-ring Naphthenes
6-rlog Naphthenes
Benzenes
Tetrallna
Naphthalenes
Tetrahydroacenaphthenes
Tetrahydrophenanthrene
Nitrogen Bases
Acids-Phenollcs
Type Analysis^
1.7
1.0
1.6
1.3
1.1
0.5
0.2
0.6
7.1
1.6
2.0
11.9
10.
Compound Class
Hexahydrophenanthrene
Octahydrophenanthrene
Dihydropyrene
Pyrene/Pluoranthene
Tetrahydrofluoranthene
Bexahydropyrene
Decahydropyrene
Chrysenea
5-ring peri-condensed
5-ring cata-condensed
6-ring peri-condensed
Coronenes
Neutral Nitrogens
Losses
Type Analysis
0.3
2.3
3.6
4.1
2.7
9.2
3.4
0.5
0.6
— Type analysis data derived from saturate and aromatic fractions. Methods are ASTM
D-2786, C20 matrix and a tentative ASTM method for aromatic hydrocarbons.
—'Data from SABA separation procedure.
214
-------
Table 7. Screening of Coal Liquefaction Process Streams for Possible Hazardous Compounds
Precise
Mass
202.0780
213.1153
217.0891
228.0936
242.1095
243.1048
251.1674
252.0936
Formula
C14H15HO
C16H11N
C18H12
C19H14
C18H13B
C18H21H
C20H12
o
01 O
Possible Hazardous . *• • u*rH
or Toxic Compound (s)^ rS " So
Pyrene X X
4-Amlno-4 -Hydroxy-azobenzene X
1,2-Benzcarbazole X X
Benz (a) anthracene
Chrysene
7-Methyl Benz (a) anthracene X X
Amino-l,2-Benzanthracene^' (2) X
Diethylamlno Stilbene X
8,9-Ace-l,2 Benzanthracene
4,10-Ace-l,2 Benzanthracene x X
«l « • M
y » o ta
9 IH a b w k> u
•DO -H T3 O 09 £^11 3 J3 «
iH -H S h -H 3 00 -O « 9 O J5 3
« 3 TJ U9-O g-H • 3 -H 0) 3 -H
4J O1 O CCTO B •> 00 M U-. 3t
-------
CHARACTERIZATION OF OIL SHALE PROCESSES
J. E. Cotter,
TRW Environmental Engineering Division
Redondo Beach, California 90278
Abstract
Experimental Characterization of an oil shale
process 1s reported, based on an environmental sam-
pling and analysis program conducted at the Paraho
retorting operations. Multi-media sampling ap-
proaches are described. Analysis techniques are
summarized with particular reference to shale oil
process stream characteristics. Implications for
future work are outlined.
Current Status
The development of oil shale deposits 1n
Colorado 1s moving ahead on a firm basis, with two
federal lease tracts (C-a and C-b) beginning mine
excavation work this year, and demonstration re-
torts already operating at Anvil Points (Paraho)
and Logan Wash (Occidental 011 Shale). First
production runs are expected to take place at the
federal lease tracts in the early 1980*s. With
this activity in motion, the U.S. EPA began an
environmental assessment program of shale oil
recovery processes over two years ago.
The characterization of oil shale processes
has been difficult and speculative without having
a full-scale Industry 1n being. Nevertheless, the
need 1s urgent for an understanding of the kind of
pollution control needs that will be faced by the
new Industry.
The Paraho Process
Arrangements were made 1n 1976 to conduct
environmental sampling and analysis at the Paraho
demonstration site at Anvil Points, CO. The test
plan was quickly put Into action, since the demon-
stration plant was about to conclude a 30-month
program.
The demonstration plant operations, Indicated
schematically 1n Figure 1, consisted of mining,
raw shale hauling, crushing and screening, retort-
Ing, and retorted shale disposal. Crude shale oil
was stored 1n tanks for subsequent shipment to an
off-site refinery. The heart of the demonstration
plant 1s the Paraho retort (Figure 2), which can
process about 400 metric tons per day.
Provision has been made for operating the
retort 1n either the direct mode or Indirect mode.
In the direct mode the carbon on the retorted shale
1s burned 1n the combustion zone to provide the
principal fuel for the process. Low calorie retort
gases are recycled to both the combustion zone and
the gas preheating zone. In the Indirect mode heat
for retorting 1s supplied by recycling off-gases
through an external furnace, thus eliminating
combustion In the retort and producing a high heat-
Ing value, 8000 kcal/std cu meter off-gas.
In either mode of operation, raw shale 1s fed
Into the top of a Paraho retort and passed downward
by gravity successively through a mist formation and
preheating zone, a retorting zone, either a combus-
tion zone (direct mode) or heating zone (Indirect
mode), and finally, a residue cooling and gas pre-
heating zone. It 1s discharged through a hydraul-
Ically-operated grate, which controls the throughput
rate and maintains even flow across the retort. The
retorted shale 1s discharged from the retort at about
150°C (300°F), and sent to the shale disposal area.
The shale vapors produced 1n the retorting zone
are cooled to a stable mist by the Incoming raw shale
(which 1s thereby preheated), and leave the retort.
this mist 1s sent to a condenser, and finally a wet
electrostatic predpltator, for oil separation. The
resulting shale oil 1s transported to storage.
The demonstration plant differs considerably
from a comnerclal facility design, so that 1t cannot
be considered a scale model of a full-size operation.
The product gas at the demonstration plant was com-
busted 1n a thermal oxldlzer prior to atmospheric
discharge; 1n a commercial facility, this gas would
be cleaned and used as a fuel 1n process heaters and
boilers. Material handling 1n a commercial plant
would most likely rely on conveyors, rather than
trucks, and the disposal of retorted shale would be
a major portion of the operation.
Sampling Methodology
Sampling methods, equipment, and stations were
specific to the nature of the specie monitored Multi-
media sampling approaches Included
• Proportional and grab sampling of the recycle
process gas stream and thermal oxldlzer dis-
charge for analysis by Instrumental and wet
chemical methods.
• Grab sampling of the product crude oil, re-
cycle process gas condensate water, and
retorted shale for organic and trace element
analysis.
• Grab sampling of recycle process gas stream
condensate water for gross parameter analy-
sis—for example, biochemical oxygen demand.
216
-------
chemical oxygen demand, total organic car-
bon and pH.
• High-volume sampling of particulate emis-
sions from raw shale mining, crushing, and
retorted shale transfer.
• Cascade Impactor sampling for particulate
sizing of airborne emissions from raw and
retorted shale processing.
Standard EPA sampling procedures for ambient
or stack monitoring were modified as necessary to
be usable In a process environment. This required,
for example, sample train 1mp1nger solutions to be
at much higher concentrations for recycle gas
sampling, in order to avoid Incomplete capture of
high-concentration constituents 1n the gas stream.
Collection of fugitive dust 1n close proximity to
shale-handling equipment was done with Mgh-vol
sampler durations of less than an hour. Bulk
retorted shale samples, taken fresh from the retort
discharge conveyor, were Immediately capped, since
the volatile emissions from retorted shale were as
much of Interest as the solid material. Preserva-
tion methods for shale oil process condensates must
be carefully selected to prevent modification of
the constituents. Freezing was eventually deter-
mined to be the best method.
Sample Processing and Analysis
The Initial sampling and analysis program was
conducted at the Paraho facility In 1976, shortly
before the plant was scheduled to shut down. There-
fore the analysis effort went considerably beyond
the Level 1 approach (1), since there was no oppor-
tunity to conduct a follow-up effort at the time.
(A resumption of operations 1n 1977 afforded the
chance for a second test effort).
Standard analytical methods were used wherever
possible. Specific techniques were developed to
handle some of the various process samples. Figures
3, 4, and 5 show plans for processing and analysis
of the Paraho gaseous, liquid, and solid samples.
Analytical methods Included
• Inorganic and trace elements analysis
—wet chemistry
—atomic absorption spectrophotometry
—gas chromatography
—spark source mass spectrometry (SSMS)
• Organic analysis, separation and Identifi-
cation with
—gas chromatography (GC)/mass spectrometry
(MS)
—thin layer chromatography (TLC)
—high-pressure liquid chromatography (HPLC)
and Level 1 LC methods
—spectrophotofluoremetry (SPF)
• Polynuclear aromatic hydrocarbons analysis
—two-dimensional elutlon with TLC.
• Particulate size analysis by gravimetric
determinations of Impactor fractions, and
scanning electron microscopy of low-volume
sampler filters.
Although a detailed description of the analysis
methods are available in a separate publication (2),
some particular characteristics of shale oil process
stream analysis can be highlighted. Recycle gas
condensates and aqueous samples from crude oil/water
separation must have the oily and aqueous phases
separated under standard conditions to obtain repro-
ducible results. Organic components were typically
0.2 - 0.3 wt. percent, including acids, bases, and
neutrals (primarily aromatics). Inorganic consti-
tuents were chiefly ammonium carbonate and bicarbon-
ate, although a wide variety of trace elements were
found 1n small quantities. (Trace elements contain-
ed in oil shale appear to be mostly retained in
retorted shale, with some carryover to the crude
shale oil). The presence of a high concentration of
ammonia Interferred with standard titrimetric deter-
minations for carbonate and bicarbonate alkalinity,
so that Inference from total Inorganic carbon was
used.
Ammonia was also found 1n the recycle gas stream
at about 1 volume percent levels, along with hydro-
gen sulfide (in the 0.1 volume percent range). These
constituents would be removed in a gas-cleaning unit
under full-scale operations. Emissions from burning
the treated gas would then be similar to natural gas
combustion.
Considerable extraction and separation of in-
organic and organic components from solid samples
was done, using both bulk retorted shale and air-
borne particulate samples. Classification was done
in the case of particulates for organic components
separated by Level 1 liquid chromatography. Retort-
ed shale was singled out for thin layer chromato-
graphy analysis, 1n order to determine the ratio of
polynuclear aromatics to polar compounds (about
50/50 by wt. 1n retorted shale, 15/85 in raw shale
parti cul ate matter).
Implications for Future Environmental Work
Modular and full-scale shale oil plants will
have to be monitored for a number of atmospheric and
aqueous waste constituents. Automatic measurement
techniques will have to be developed, to avoid the
cost of manual sampling and laboratory analyses.
Sampling and analysis programs such as this one serve
to characterize various constituents found in pro-
cess and waste streams, but do not answer the ques-
tion of monitoring priorities. Health effect stu-
dies, comparison with natural environment character-
istics, and studies of residual behavior are all
needed to determine priorities of concern. For ex-
ample, the Interaction of retorted shale with water
from natural sources or process wastewater needs to
be understood before groundwater monitoring programs
can be well designed.
The same comments apply to the development of
appropriate pollution control technologies, since
only a few of the constituents found In shale oil
by-product streams are covered by existing federal
or state regulations. The development and Improve-
ment of oil shale retorting processes needs to be
done in parallel with pollution control evaluation,
since process methodologies will affect the charac-
teristics and quantities of waste streams.
217
-------
Finally, characterizations of shale oil pro-
duct uses will be needed. Combustion of crude
shale oil has been monitored 1n one case (3), and
future work should examine the Impacts of refined
product consumption.
References
(1) IERL-RTP Procedures Manual: Level 1
Environmental Assessment, EPA-600/2-76-160a,
June 1976
(2) Sampling and Analysis Research Program at the
Paraho Shale Oil Demonstration Plant, IERL-
C1, to be published (TRW Contract 68-02-1881,
May 1977)
(3) 0. G. Jones, M.N. Mansour, Low NOV Combustion
of Paraho Shale 011 1n a 45 MN Utility Boiler,
ASME 77-UA/Fu-l, Atlanta, December 1977
218
-------
Figure 1. Anvil Points operations.
HAULING
RETORTED SHALE
TRANSFER
CRUSHING
Figure 2
PARAHO RETORTING, INDIRECT MODE
Rotating spreader,
Collecting tubes
Mist formation
and preheating
Distributors
Retorting zone m
Distributors
Heating
Residue cooling and
gas preheating
Moving grates
Feed shale
Product
Recycle gas
blower
• Retorted shale to disposal beds
219
-------
Figure 3
SEPARATION AND ANALYSIS SCHEME. GASEOUS SAMPLES
Gas
bottles
Tenax
absorption
Mine Safety
Appliances
charcoal
absorption
Bendix direct
reading
Gastec
tubes
ro
f\>
Impmger
solutions
Wet
chemistry
analysis
Add ben*
and neutral
organic
extraction
Aqueous
phase
HPLC
Fluorescence
Ultraviolet
Organic
carbon
Total carbon
anaryito
Not used because of the presence of aerosol.
-------
Figure 4
SEPARATION AND ANALYSIS SCHEME, WATER SAMPLE:
Phase
separation
(from product
Bondapak
absorption
oforganics
Solvent
extraction
Preseparatory"
(silica)
TLC
Acidification
pH=2
Organic
extraction
Headspace
(nitrogen
stripping)
Tenax
absorption
Aqueous
phase
Flasher
SSMS
Acid, base,
and neutral
organic
extraction
TLC analytical scheme.
221
-------
Figure 5
SEPARATION AND ANALYSIS SCHEME,
RETORTED SHALE AND HIGH-VOLUME SAMPLER AIR PARTICULATES
Retorted shate
and high-volume
air paiticulates
C, H, O, N, S
ash and sieve
analysis
Water
extraction,
(leachate)'
Flasher
(volatites)
Solvent
extraction
HPLC
Fluorescence
Ultraviolet
Carbon, hydrogen, oxygen, nitrogen, sulfur.
Not performed on air particulates
TLC analytical scheme.
222
-------
TECHNICAL REPORT DATA
(Please read Inunction* on the reverse before completing)
1. REPORT NO.
EPA-600/7-78-168
2.
4. TITLE AND SUBTITLE gympQ8^um Proceedings : Process
Measurements for Environmental Assessment
(Atlanta, February 1978)
7. AUTHOR(S)
Eugene A. Burns , Compiler
9. PERFORMING ORGANIZATION NAME Ah
TRW Systems Group
One Space Park
Redondo Beach, California
4D ADDRESS
90278
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
3. RECIPIENT'S ACCESSION NO.
6. REPORT DATE
August 1978
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO.
10. PROGRAM ELEMENT NO.
EHE624
11. CONTRACT/GRANT NO.
68-02-2165, Task 24
13. TYPE OF REPORT AND PERIOD COVERED
Task Final; 4/77-2/78
14. SPONSORING AGENCY CODE
EPA/600/13
is. SUPPLEMENTARY NOTES yjERL-RTP project officer is James A. Dorsey, Mail Drop 62, 919/
541-2557.
16 ABSTRACTThe report documents the 26 presentations made at the Process Measure-
ments for Environmental Assessment Symposium, held February 13-15, 1978, in
Atlanta, Georgia. The symposium was sponsored by the Process Measurements
Branch of EPA's Industrial Environmental Laboratory, Research Triangle Park,
North Carolina. The objective of the symposium was to bring together people who
were responsible for planning and implementing sampling and analysis programs for
multimedia environmental assessment. The program consisted of sessions defining
the uses of environmental assessment data, the techniques for acquiring information,
and recent user field experience with environmental assessment measurement pro-
grams.
17.
a. DESCRIPTORS
Pollution
Measurement
Industrial Processes
Assessments
Sampling
Analyzing
18. DISTRIBUTION STATEMENT
Unlimited
KEY WORDS AND DOCUMENT ANALYSIS
b. IDENTIFIERS/OPEN ENDED TERMS
Pollution Control
Stationary Sources
Environmental Assess-
ment
19. SECURITY CLASS (This Report!
Unclassified
20. SECURITY CLASS (This page 1
Unclassified
c. COSATi Field/Group
13B
14B
13H
234
22. PRICE
EPA Form 2220-1 (9-73)
------- |