SERA
United States
Environmental Protection
Agency
Industrial Environmental Research
Laboratory
Cincinnati OH 45268
EPA-600/2-79-211
December 1979
Raaearch and Development
Control of
Particulate
Emissions in the
Primary Nonferrous
Metals Industries
Symposium
Proceedings
-------
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 ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of pollution-sources to meet environmental quality standards.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-79-211
December 1979
CONTROL OF PARTICULATE EMISSIONS IN THE
PRIMARY NONFERROUS METALS INDUSTRIES
Symposium Proceedings
Del Monte Hyatt House
Monterey, California
March 18-21, 1979
Richard L. Meek, Editor
Southern Research Institute
2000 Ninth Avenue South
Birmingham, Alabama 35205
Grant No. 804955
Project Officer
John 0. Burckle
Industrial Pollution Control Division
Industrial Environmental Research Laboratory
Cincinnati, Ohio 45268
INDUSTRIAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
DISCLAIMER
This report has been reviewed by the Industrial Environ-
mental Research Laboratory-Cincinnati, U.S. Environmental
Protection Agency, and approved for publication. Approval does
not signify that the contents necessarily reflect the views
and policies of the U.S. Environmental Protection Agency, nor
does mention of trade names or commercial products constitute
endorsement or recommendation for use.
11
-------
FOREWORD
When energy and material resources are extracted, processed, converted,
and used, the related pollutional impacts on our environment and even on
our health often require that new and increasingly efficient pollution
control methods be used. The Industrial Environmental Research Laboratory -
Cincinnati (lERL-Ci) assists in developing and demonstrating new and improved
methodologies that will meet these needs both efficiently and economically.
To perform this duty effectively, the lERL-Ci coordinates and interacts
with industry through conferences, symposia, and cooperative projects. This
particular report is the proceedings of a speciality conference designed to
provide a forum for the exchange of knowledge and new ideas on particulate
control technology with emphasis on industrial applications of ^environmental
particulate control technology in the primary nonferrous industries. The
symposium, held at Monterey, California, March 18-21, 1979 was sponsored
by the U. S. Environmental Protection Agency, Office of Research and
Development, Metals and Inorganic Chemicals Branch, Industrial Environmental
Research Laboratory-Cincinnati.
David G. Stephan
Director
Industrial Environmental Research Laboratory
Cincinnati
ill
-------
ABSTRACT
The purpose of the symposium on "Control of Particulate
Emissions in the Primary Nonferrous Metals Industries" was to
provide a forum for the exchange of knowledge and new ideas on
particulate control technology with emphasis on industrial
applications of environmental particulate control technology in
the primary nonferrous industries. The symposium held at
Monterey, California, March 18-21, 1979 was sponsored by the
U.S. Environmental Protection Agency, Office of Research and
Development, Metals and Inorganic Chemicals Branch, Industrial
Environmental Research Laboratory-Cincinnati.
The symposium included presentations on foreign and
domestic technology applicable to copper, lead, zinc, and
aluminum as well as discussions of other advanced technology,
measurement techniques, and current EPA programs that are
pertinent to particulate control in the nonferrous metals indus-
try.
Speakers from England, Canada, Sweden, Japan, Australia,
and the United States discussed recent developments and tech-
nology for particulate control in nonferrous operations.
IV
-------
CONTENTS
Foreword iii
Abstract iv
Introduction - George S. Thompson, Jr., Chief
Metals and Inorganic Chemicals Branch
Industrial Environmental Research
Laboratory
Cincinnati, Ohio 1
EPA NONFERROUS RESEARCH
Session Chairman: Robert I. Jaffee
Electric Power Research
Institute
Palo Alto, California
"EPA's Nonferrous Metals Production Research, Develop-
ment, and Demonstration Program" - John 0. Burckle,
Environmental Protection Agency, Office of Research and
Development, Industrial Environmental Research
Laboratory, Cincinnati, Ohio 5
"Characterization of Emission Control Problems at a
Molybdenum Roaster" - Klaus Schwitzgebel, C. Dean
Wolbach, Radian Corporation, Austin, Texas 18
"Present and Future Control of Fugitive Emissions in
the Primary Nonferrous Metals Industry" - Alfred B.
Craig, Jr., IERL, U.S. Environmental Protection Agency,
Cincinnati, Ohio, and L. V. Yerino, M.D. Giordano, and
T.K. Corwin, PEDCo Environmental, Inc., Cincinnati,
Ohio 31
"Demonstration of the Bergsoe Agglomeration Furnace
and Best Management Practices at a Secondary Lead
Smelter" - Richard T. Coleman, Jr., Radian Corporation,
Austin, Texas; Alfred B. Craig, Jr., Environmental
Protection Agency, Cincinnati, Ohio; and Robert
Vandervort, Radian Corporation, Salt Lake City, Utah... 98
"Getting Ready for Inhalable Particles" - D. Bruce
Harris, Process Measurements Branch,Industrial Envir-
onmental Research Laboratory, Environmental Protection
Agency, Research Triangle Park, NC 123
v
-------
CONTENTS (continued)
LEAD AND ZINC INDUSTRY 134
Session Chairman: Jerome E. Cole
International Lead zinc
Research Organization, Inc.
New York, N.Y.
"An Overview of Controls in Primary Lead and Zinc"-
W. A. Lemmon and D. Haliburton, Air Pollution Control
Directorate, Environment Canada, Ottawa, Ontario 135
"Control of Particulate Emissions from the Imperial
Smelting Zinc-Lead Process" - Roger W. Lee, Imperial
Smelting Processes Limited, Avonmouth, Bristol BSll
9HP, England 154
"Particulate Control at the Port Pirie Lead Smelter"-
John D. Martin, The Broken Hill Associated Smelters
Pty. Ltd. , Port Pirie, South Australia 172
"Applications of Baghouses in Lead and Zinc Smelters"-
Knowlton J. Caplan, Industrial Health Engineering
Associates, Inc. , Minneapolis, Minnesota 188
"Review of Design and Operation of Electrostatic
Precipitators in Nonferrous Metals Applications" -
Heinz L. Engelbrecht, Air Pollution Control Division,
Wheelabrator-Frye Inc., Pittsburgh, Pennsylvania 204
ALUMINUM INDUSTRY 227
Session Chairman: Seymour Epstein
The Aluminum Association
Washington, D. C.
"An Introduction to the Primary Aluminum Industry and
Its Commonly Employed Emission Control Methods" - ,1
Lawrence C. Tropea, Jr., Reynolds Metals Company,
Richmond, Virginia 228
"Particulate Emission Control in an Automated Side-
work Pre-bake Aluminum Reduction Facility" - William
J. Janson, Eastalco Aluminum Company, Frederick,
Maryland 242
"Emission Control in Vertical Soderberg Furnaces" -
J. L. Byrne, Martin Marietta Aluminum Inc., The Dalles,
Oregon 257
VI
-------
CONTENTS (continued)
"Experience with Thirty-one Wet Electrostatic
Precipitators on HSS Aluminum Potlines" - Leland 0.
Slind, Reynolds Metals Company, Longview, Washington .. 258
"A Low Emission Process for the Melt Treatment of
Aluminum Alloys" - F. R. Mollard, J. E. Dore and
N. Davidson, Consolidated Aluminum Corporation,
Research and Development, St. Louis, Missouri 274
COPPER INDUSTRY 289
Session Chairman: Ivor E. Campbell
Clyde Williams and Company
Columbus, Ohio
"Inco's Oxygen Flash Smelting Process for Copper and
Nickel Concentrates Off-gas Handling and Impurity
Distributions" - M.Y. Solar, Inco Metals Company,
Toronto, Ontario, Canada, and A..D. Church and T. N.
Antonioni, Inco Metals Company, Copper Cliff, Ontario,
Canada 290
"Processing of Copper Smelting Gases at Naoshima
Smelter" - H. Uchida, Mitsubishi Metal Corporation,
New York, New York, and N. Kikumoto and M. Hayashi,
Mitsubishi Metal Corporation, Naoshima, Japan 314
"Particulate Emission Control at the Garfield Smelter"-
Frederick E. Templeton, Kennecott Copper Corporation,
Utah Copper Division, Salt Lake City, Utah 344
"The Recovery of Particulate Matter and Sulfur Dioxide
at the Hidalgo Smelter" - W. J. Chen, Phelps Dodge
Corporation, Playas, New Mexico 351
"On the Electrostatic Precipitation of Copper Smelter
Emissions" - W. R. Heifner, United McGill Corporation,
Columbus, Ohio 361
SPECIAL CONTROL SYSTEMS 381
Session Chairman: Grady B. Nichols
Southern Research Institute
Birmingham, Alabama
"Application of Cottrells in ASARCO's Nonferrous
Smelters" - E. S. Godsey, ASARCO, Inc., Salt Lake City,
Utah 382
vn
-------
CONTENTS (continued)
"Electrostatic Precipitators (ESPS) in Two Stages
Used for Arsenic Recovery at the Ronnskar Copper
Works" - Kjell Porle, AB Svenska Flaktfabriken,
Vaxjo, Sweden, and Bjorn Lindquist, Boliden Metall
AB, Skelleftehamn, Sweden 401
"The WSL Approach to Metallurgical Fume Emissions" -
P.R. Dawson, Warren Spring Laboratory, Stevenage,
Hertfordshire, England 414
"Advanced Techniques for Fine Particulate Control" -
James H. Abbott, EPA Particulate Technology Branch,
Industrial Environmental Research Laboratory, Research
Triangle Park, North Carolina 434
SYMPOSIUM ADDRESS 458
"And What of the Future?"
W. H. Dresher
Dean, College of Mines
The University of Arizona
Tucson, Arizona
CONTROL DEVICE SELECTION AND EVALUATION 468
(Panel Discussion and Open Forum)
Moderator: Richard L. Meek
Southern Research Institute
Birmingham, Alabama
Vill
-------
EPA WELCOME AND INTRODUCTION TO THE SYMPOSIUM
George S. Thompson, Jr.
Chief
Metals and Inorganic Chemicals Branch
Industrial Environmental Research Laboratory
Cincinnati, Ohio
Welcome and Acknowledgements
I would like to extend to you EPA's "welcome" to this
symposium on particulate control practices and problems in the
nonferrous metals production industries. I offer a special
thanks to Dick Meek and his colleagues of the Southern Research
Institute for their aid in developing and administering the
symposium, to the Session Chairmen, Bob Jaffee, Jerry Cole,
Ivor Campbell, and Seymour Epstein, to the authors of the pre-
sentations to follow, to our Monday Dinner Speaker, Dr. William
Dresher and the panel members for the discussion on Tuesday
evening, and to the audience, those who have come to learn and
to share their experiences and insights. We hope our efforts
are successful in providing you with an understanding and an
awareness of the objectives and status of EPA's R & D activities
impacting the nonferrous metals production industry.
We solicit your participation and comments after each
presentation and particularly at the Tuesday evening discussion
session. Members of my staff who are involved with the non-
ferrous metals production R & D program include John Burckle
and Fred Craig. They will be actively soliciting your comments
and reactions during this conference's discussion periods and
through informal contacts.
The Role of Research and Development
I represent EPA's Office of Research and Development,
specifically the Industrial Environmental Research Laboratory
in Cincinnati, Ohio. My Branch, the Metals and Inorganic
Chemicals Branch, is responsible for conducting R & D activities
on air, water, and solid waste pollution for a variety of indus-
tries, including the nonferrous metals industry, the inorganic
-------
chemicals industry, the metal finishing and fabrication industry,
and a number of miscellaneous industries such as glass, cement,
and asbestos. My Branch, as well as my Director's Division and
the Laboratory in Cincinnati, came into existence three years
ago as a result of a major reorganization within EPA's Office
of Research and Development. The major benefit from this
reorganization was the establishment of a new charter directing
one research group to address pollution control RD & D for all
three of the major environmental media—air, water, and land
for specific industries. Prior to the 1975 reorganization, the
programs for RD & D involving pollution in the different media
were conducted by different groups under different managment.
Often, they did not coordinate their work. In most cases, these
groups of Federal researchers were physically located in differ-
ent parts of the United States. This reorganization has been
instrumental in correcting these problems and bringing a cohesive
approach to our control technology programs.
Our broad charter is suited extremely well to conducting
valuable research programs; we are one of a small handful of
EPA activities that can address the "total pollution problem."
Also as a result of our broad charter, we have the capability to
interface with other EPA offices and Federal Agencies having
regulatory and enforcement responsibilities impacting the indus-
trial sector. Allow me to provide a specific example that
directly addresses your interests. My staff interfaces with the
regulatory offices, such as the Effluent Guidelines Division,
the Office of Solid Waste, the Office of Air Quality Planning
and Standards, and EPA's newly structured Office of Toxic Sub-
stances. We interface with EPA's air and water enforcement
offices as well as many of EPA's ten Regional Offices. What
does this interface provide? Awareness--a basic requirement
for conducting valuable research—and technical support—a "must"
for establishing firm technical foundations for regulatory and
enforcement actions. I must stress the following point: EPA's
Office of Research and Development is an independent function
within EPA; it does not report through line management to any
one EPA regulatory or enforcement activity. If you're asking
yourself "what does all of this mean?"—let me summarize. Our
interface with these programs along with our interface with you,
the industry, provides us with the awareness to structure our
research activities to be best "in tune" with the most important
needs. We in research can develop and implement programs that
provide answers to key technical and economical pollution prob-
lems impacting the industry.
The nonferrous metals research program is structured
around the following goals: (which will be described in
detail by John Burckle)
-------
• Advance the state-of-the-art in air, water, and solid
waste pollution control and treatment technology.
• Provide EPA with the best technological basis for the
setting and enforcing of regulations.
• Ensure both EPA and the industry that the abatement
of pollution from one media will not result in either
pollution to another media or excessive energy con-
sumption.
• Act as a focal point for information dissemination.
These goals are not easily achieved. Everyone must work
toward them. We in the Office of Research and Development must
have a clear understanding of the industrial processes and the
resulting air, water, and solid waste pollutants generated by
these processes; we must be perceptive to both the short and
long-term research needs within EPA and obviously outside of
EPA. We rely strongly on ideas, direction toward problems,
and expert advice from people like yourselves.
In conclusion, this symposium provides an excellent forum
for us to exchange ideas and viewpoints. Please comment on our
research program. While we are knowledgeable, we are not experts,
However, you are faced with the "real world" problems that pro-
vide you with insight and practical working knowledge. If you
share your insight and expertise to assist us in developing
better programs, we will all benefit.
-------
EPA NONFERROUS RESEARCH
Session Chairman: Robert I. Jaffee
Electric Power Research Institute
Palo Alto, California
-------
EPA's NONFERROUS METALS PRODUCTION
RESEARCH, DEVELOPMENT, AND DEMONSTRATION PROGRAM
John 0. Burckle
Environmental Protection Agency
Office of Research and Development
Industrial Environmental Research Laboratory
Cincinnati, Ohio
ABSTRACT
The U.S. Environmental Protection Agency is primarily a
regulatory agency whose mission it is to attain and maintain
environmental quality for the protection of human health and
welfare. As a complement to its regulatory activities, the
Agency supports research and development leading to improved
pollution control technology. The Office of Research and Devel-
opment is responsible for the development and management of a
National program of research, development, and demonstration
leading to an understanding of the origin and effects of pollu-
tion, techniques for the measurement and monitoring of environ-
mental quality, and technology for the control, management, and
reutilization of pollutants and wastes. The ORD's Nonferrous
Metals Program sponsors research to provide technology for their
solution in the nonferrous metals industry. This program is the
responsibility of the Industrial Environmental Research
Laboratory at Cincinnati. The Resource Extraction Branch is
responsible for the ore extraction and beneficiation operations,
and the Metals and Inorganic Chemicals Branch is responsible
for the production and fabrication operations. This paper
presents an overview of the research program dealing with the
production of nonferrous metals.
INTRODUCTION
The EPA's nonferrous metals program sponsors research,
development, and demonstration projects to define environmental
problems caused by the mining and winning of nonferrous metals
and to provide technology for their solution. This program is
located in the Industrial Environmental Research Laboratory
at Cincinnati. The Resource Extraction Branch is responsible
-------
for ore extraction and beneficiation aspects, and the Metals and
Inorganic Chemicals.Branch is responsible for the production
(smelting and refining) and fabrication (tube, sheet, wire, etc.)
operations. This paper presents an overview of the program
dealing with the production of nonferrous metals.
The program is structured into two major components. The
first component is problem identification and definition. Prob-
lem identification and definition is accomplished through a
program of engineering analysis of process inputs and operations
coupled with sampling and analyses of process inputs and waste
streams to identify and characterize waste streams having a
potentially serious effect upon the environment. A multimedia
approach is utilized and energy and economic factors are taken
into consideration in comparison of alternative process and
control technologies.
The second phase of the program, the control technology
RD & D phase, deals with the research, development and demon-
stration of feasible solutions. As a first step, the technical
and economic feasibility of selected conceptual approaches are
studied to identify the most promising candidates for support.
The transfer of technology developed and applied in other areas
or in foreign countries is the fundamental approach utilized.
PROBLEM IDENTIFICATION AND DEFINITION
Upon reassignment to Cincinnati in 1976, the program
undertook as a first step the assessment and definition of the
environmental problems arising from the production of nonferrous
metals. This step, has resulted in the development of a data
base containing a profile of the composition of the industry
process information with associated air, water and solid waste
emission data.
To date, the assessment and problem definition has been
completed for all primary and secondary nonferrous metals pro-
duction processes used in the United States. However, the
assessment is currently based on information existing in the
literature only. This task has resulted in a five volume report
in various stages of review.
1. Overview
2. Primary Copper, Lead, and Zinc
3. Primary Aluminum
4. Primary Minor Nonferrous Metals
5. Secondary Nonferrous Metals
-------
The first three volumes of this report have been distri-
buted within the Agency and the selected outside organizations
for review and comment. We are also preparing an update to sur-
vey newly developed information for future incorporation into
the report.
In brief, the assessment shows the following. The non-
ferrous metals production industry of the United States is
diverse and complex in its composition, product mix and environ-
mental problems. It is responsible for the production of 35
commercially significant primary metals and the recovery of some
18 nonferrous metals from waste materials.
The primary industry is composed of about 200 mines and
200 plants involved in the various phases of winning metals from
the ores. In order of commercial significance by tonnage pro-
duction, the industry produces aluminum, copper, lead, zinc,
and the minor metals. Sulfuric acid is a principal by-product.
Depending on the product and ore composition, a number of pro-
cessing steps are involved: mining, concentrating, metal win-
ning, and refining into a high purity form through a large
number of processes. These activities may be integrated into a
single site, or they may be widespread geographically. Most
domestic ores, excepting aluminum, are obtained from 75 large
mines producing copper, lead and zinc ores; some additional 13
minor metals are obtained as coproducts from these ores. A
number of these operations are located in highly urbanized
areas, while a significant number are situated in remote,
relatively unpopulated areas. Many of the plants are old, with
only a small portion having been built since 1960.
The secondary industry collects scrap materials, and.,,
through pyrometallurgical or chemical processes, recovers
specific nonferrous metals for recycling into commerce. It is
the most diffused segment being composed of 400 plants operated
by 300 firms. These plants are generally small with 85% having
less than 100 employees and 35% having less than 4 employees.
They tend to be concentrated in the urban, highly industrial
areas (East Coast, Great Lakes, and clusters on the West Coast).
Production of nonferrous metals in the United States
generates great quantities of pollutants that enter the air,
water, and soil. Many of the emissions having the potential for
adverse impacts are controlled, but others are not. Some could
be controlled by application of currently available technology;
others will require development of new control techniques and
systems.
The most important needs for improved environmental control
are as follows:
-------
1. Technology for more effective control of air
emissions containing sulfur oxides and significant
quantities of toxic metal compounds.
2. Methods for effective disposal of solid and sludge
residues containing leachable toxic components.
3. More effective methods for treatment of acidic
waters contaminated with toxic metal compounds.
4. More effective control methods for fugitive emissions
containing S02 and toxic metal compounds.
5. Better definition of specific emission streams sus-
pected to be hazardous, especially in plants pro-
ducing the minor nonferrous metals in the secondary
nonferrous metals industries.
6. Better definition of the potential impact of new
technologies that may be used in the future for
production of nonferrous metals.
Based upon what we learned from this first step, we have
incorporated a continuing assessment effort into our program.
This industry assessment is designed to provide information on
the production technologies, the process emissions, the control
technologies, and the environmental pollution problems of the
nonferrous metals production industry. The assessment activity
is composed of activities which lead to a highly specific
quantitative multimedia analysis and problem definition of an
industry segment. Information is identified, acquired, and
subjected to engineering analysis in order to identify specific
environmental pollution problems and potential technologies to
solve these problems. The results of these efforts provide the
specific program tactics for out-year efforts and redirection of
ongoing projects as appropriate.
The assessment program is now composed of a number of
projects which culminate in an assessment report to be updated
annually. First, current information in the worldwide literature
is screened on a bimonthly basis; the result of this screening is
a publication entitled "Technical Awareness Bulletin - Nonferrous
Metals." This Bulletin serves several useful functions: (1) it
is distributed to interested Agency personnel to furnish up-to-
date information on industry developments; (2) it is a source
tool for the update of the assessment activities;(3) it is dis-
tributed to industry to inform and aid in technology transfer
which will hopefully lead to improved environmental control.
-------
The Nonferrous Metals Production Resource Center is the
focal point of the assessment activities. The Center is being
set up to acquire information (spotlighted by the Technical
Awareness Bulletin, pertinent government reports, agency files,
and newly developed data from field test programs) and perform
an engineering analysis of selected situations to pinpoint
environmental problems and develop conceptual solutions. This
activity results in the annual update of the Environmental
Assessment reports which form the basis for the development of
specific program tactics for out-year planning. In addition,
the report will specifically identify key pollution problems.
Specific effort is made to acquire information from
foreign countries regarding advanced technologies and practices
for add-on control devices or improved process technologies
(.Analysis of Foreign Technology) . This project also furnishes
the capability to review and analyze the technical results of
our PL-480 programs. The scope of assessment activities is
currently being expanded to provide for capabilities for analy-
sis of
• existing technology, both add-on controls (retrofit
or new plant) and (new) demonstrated process tech-
nology such as the Noranda and Mitshubishi systems
• potential research and development initiatives
involving technology not yet applied to nonferrous
metals production
• field testing
CONTROL TECHNOLOGY R & D
The control technology RD & D program covers the develop-
ment of equipment and process modifications to control the
multimedia environmental impacts resulting from the production
of nonferrous metals to acceptable levels.
The control technology development program addresses
5 subareas: (1) weak S02 gas stream control, (2) particulate
control with emphasis on volatile and toxic trace elements,
(3) control of fugitive emissions, (4) solid waste management
techniques, and (5) treatment techniques for metal-bearing
acidic wastewaters. The program is structured to deal with these
areas identified by the environmental assessment as the most
important targets for improved control. Efforts programmed for
the support of regulatory and enforcement programs include a
demonstration of a flue gas desulfurization technology, prepara-
tory work for demonstrations of control technologies for fugi-
tives and other weak S02 gas stream controls, and research lead-
ing to the development of more effective air pollution control,
water pollution treatment and solid waste containment systems.
-------
• Air Pollution Problems
In recent years, the Agency has focused most of its
regulatory and enforcement activities for this industry toward
the abatement of SOa and total particulate emissions from pri-
mary smelters. These problems remain persistent and are still
to be resolved. Besides these criteria pollutants, this indus-
try has the potential to emit large tonnage quantities of air-
borne particulate composed of such metals as arsenic, cadmium,
lead, and copper. Control devices, which are conventionally
operated for recovery of metal values, a prevalent industry
practice, have thus far been found to be ineffective collectors
of the volatile form of these metals. Fugitive emissions con-
taining arsenic, lead and possibly cadmium are also a significant
source of toxic metals to the atmosphere. In the case of both
processes and fugitive emissions, the emissions contain appre-
ciable S02 and, therefore, the selection of control options must
consider how control of all pollutants can best be accomplished.
Add-on control requires the development of more cost
effective approaches to flue gas cleaning, especially where
retrofit is concerned. The objectives of this work area are:
(1) to identify the existence of more cost-effective add-on
control technology and, through applications research, transfer
existing technology from other fields into the nonferrous area;
(2) to develop and document the performance of existing and new
process technology having superior environmental and energy-
conserving attributes coupled with competitive economics.
Weak S02 Control
Studies evaluating the technical feasibility of applying
controls and process modifications to specific unit operations
and gas blending have been completed. These studies clearly
show that approaches such as oxygen enrichment, flue gas blending,
and various FGD methods can, from a theoretical standpoint, be
applied to metallurgical offgases to abate discharges of sulfur
dioxide. However, no one technology is applicable to all situa-
tions, with the possible exception of FGD. Control technologies
based upon process modifications, such as oxygen enrichment and
gas stream blending, are dependent upon specific process condi-
tions. The above studies were extended to investigate the tech-
nical and economic feasibility of maximizing total smelter SC-2
control to 95% or better of the smelter input sulfur. This study
revealed that certain "advanced" systems held promise in being
economically feasible for greenfield smelters.
We are going forward with our FY 78 initiative to develop
and demonstrate a suitable Flue Gas Cleaning (FGC) technology
for control of particulate and weak SOa from a molybdenum roaster.
The above studies will serve a vital role in defining systems
10
-------
to be piloted. The technical and economic conclusions drawn from
this demonstration along with the results of the above efforts
will provide the definitive data to settle the question of con-
trol for weak SOa gas streams for primary copper smelters and
support the Agency's NSO program. This demonstration is designed
to provide for effective control of volatile and other trace
metals contained in the weak S02 offgases; thus, an additional
control alternative would be available for toxic metals.
Particulate Control
The program for the control of particulate emissions is
being focused on the identification of technology which will
provide acceptable environmental levels of control for volatile
and toxic trace elements as well as gross particulate. The
basic strategy is to (1) locate and test full-scale systems
which are exemplary in control of trace, as well as gross, parti-
culates, including volatiles; (2) apply new technologies devel-
oped by IERL-RTP on smelter off-gas slip streams as Demonstra-
tions at pilot scale.
A baghouse controlling particulate emissions from a moly
roaster has been tested for emissions control performance and
as a clean-up device prior to SOa control. Fabric filtration
will also be experimentally investigated as a means for control
of particulate from a lead sinter machine and blast furnace in
Yugoslavia. An exemplary fabric filter/hot ESP system for
arsenic trioxide control with metals value recovery has been
tested.
The performance of electrostatic precipitators for
control of both gross particulate emissions and toxic, including
volatile, trace elements is being investigated (ESP Evaluation).
The first year's effort identified specific problems in the lack
of control of arsenic and other volatile trace elements by
electrostatic precipitation at elevated stack gas temperatures.
Plans are also being formulated to conduct (including volatile
and trace metal) studies on existing exemplary full-scale and
promising advanced technology pilot systems.
Wet scrubbers for particulate control are applied pri-
marily in the primary industry in the cleanup of acid plant feed
gases and in certain segments of the secondary industry. They
are also applied in the primary aluminum segment, but the pre-
ferred control today is "dry scrubbing." Our efforts in scrub-
ber development are limited to: (1) applications
testing of advanced technology-based systems as developed by
IERL-RTP which offer significant cost advantages; (2) a
scrubbing system which will permit simultaneous control of SOa
and serve to remove particulate from a gas stream before intro-
duction to a desulfurization scrubber. This latter aspect will
be investigated as a part of the FGC Demo project mentioned
previously.
11
-------
Fugitive Emissions Control
Fugitive emissions are major sources of uncontrolled
trace metal and SOz emissions to the workplace and to the gen-
eral environment. Several studies underway address the problems
of particulate (including volatile and (other) toxic trace
elements) and S02 fugitives control. This particular area is of
crucial interest; because of the high cost potential for con-
trolling fugitives to meet OSHA and EPA standards, the need for
control may pose significant constraints upon the use of exist-
ing reverb/converter technology. Suitable control techniques
or alternative process technologies are needed.
We began a comprehensive evaluation of existing and
proposed process control systems and process modifications for
copper smelters. This study included an examination of second-
ary hooding for Fierce-Smith Converters, the Hoboken Converter,
whole building evacuation, etc. It drew upon data generated by
other Agency components and foreign technology employed in
Poland, Belgium, and Japan.
The evaluation of the "second edition" of the Polish
version of the Hoboken converter is nearing completion. The
newest configuration ("third edition") will be investigated
to establish how design changes have improved the overall con-
trol of fugitive emissions from the converter operations.
In FY 78, we began a joint project with NIOSH to inves-
tigate fugitive emissions in the secondary metals industry.
This study is currently focusing on secondary lead smelters.
EPA and NIOSH are cost sharing the evaluation of fugitive and
process emissions at several exemplary lead plants including
several located abroad. The first phase of this study is
expected to characterize a system believed to be vastly superior
to any currently in use in the United States. If this expecta-
tion is realized, a cosponsored demonstration of the system with
NIOSH and a selected smelter will be considered. We have under
development a project with Trepca (Yugoslavia) which will address
the problem of fugitive emissions control at a primary lead/zinc
smelter under the Special Foreign Currency Program.
• Water Pollution Problems
The effluent limitations for the Best Available Technology
Economically Achievable, to be met in 1984, are currently being
developed for primary and some secondary metals industries for
the beneficiation, and the smelting and refining subcategories.
Many technical problems require resolution before the mandates of
the Consent Decree, the 1985 zero pollutant discharge goals, and
pending pretreatment guidelines can be met. Process wastewaters
generated by this industry contain all 13 of the metallic prior-
ity pollutants listed in the June 1976 NRDC/EPA Consent Decree
12
-------
as well as cyanides and various organic reagents used for ore
concentration. These create special problems where end-of-pipe
treatment must be upgraded to permit recycle and reuse of pro-
cess water. Improved metals removal will increase the problem
of sludge disposal. Current technology provides only for land-
filling or ponding, both of which can contribute to groundwater
pollution and contamination and which are coming under regulation
under the Resource Conservation and Recovery Act (i.e., solid
waste act).
The wastewater control program is addressing these prob-
lems on several fronts. The immediate goals of the pretreatment
regulations significantly affect the secondary metals industry,
which is located in urban areas and often discharges to POTW's.
For this industry, the treatment technology of choice may be
precipitation to control heavy metals. To assist in providing
a viable technology, the neutralization and precipitation
manual will be completed and published. The Neutralization/
Precipitation Manual of Practice will provide Regional personnel
involved in permit writing and evaluation and industry with
guidance on upgrading existing in-plant treatment processes as
well as a design aid for new facilities required to meet the
pretreatment and BAT regulations. The design parameters re-
quired to meet these less immediate 1984 BAT goals for primary
and secondary plants will be available through this manual.
Meeting the Agency's goal of zero discharge of pollutants
to surface waters by 1985 poses more difficult problems for
the nonferrous industry. Recycling treated process wastewater
back into the system, often to the ore concentrating stage, can
be utilized on a limited basis. However, when total recycle is
used with small amounts of make-up water, quantities of dis-
solved metals remain in the treated water at levels sufficient
to adversely affect ore beneficiation. To remedy this, the use
of sulfide precipitation has been investigated, on a limited
basis, for treatment of various acidic metal-bearing wastewaters
from nonferrous and metal finishing processes (charged membrane
ultrafiltration and sulfide precipitation). The pilot scale
studies show that a significant improvement in metals removal can
be affected by the use of sulfide precipitation. A full-scale
system has been installed at Boliden Metall's Ronskaar works in
Sweden. An evaluation is underway to-identify key design param-
eters and to determine effluent quality attainable by this treat-
ment mechanism. The results of the Boliden monitoring program
will document state-of-the-art performance for sulfide precipi-
tation of heavy metals. Field evaluations will be verified by
lab studies conducted at the pilot scale facilities. Sludge and
H2S generation will be studied to evaluate sludge recyclability
to roasters/reverbs as well as its stability in landfills.
A study of "Best Management Practices" has been under-
taken to identify .techniques which control water pollution from
13
-------
nonprocess related sources, i.e., runoff from storage piles
in-plant boundaries, tailings ponds, etc. This will incorporate
data collected on existing tailings ponds, their location, con-
dition and key operating parameters. Exemplary methods used by
industry to reduce surface and groundwater contamination from
spills and runoff of plant compounds (fugitive discharges) will
be identified and documented to assist in evaluating facets of
environmental protection not directly related to processes.
Treatability of the priority pollutants will receive major
emphasis but only as far as it promotes the transition into a
recycle and reuse program for water and solid waste residuals
resulting from wastewater treatment. Process water quality
criteria necessary for recycle or reuse of water back into flo-
tation circuits will be developed. These will be incident to
the treatability studies being conducted at Boliden, University
of Kentucky, and in-house at the Test and Evaluation facility.
The objective will be to optimize treatment technologies so that
treated process water meets industry intake water purity require-
ments. Technological deficiencies will be identified in this pre-
liminary recycle/reuse program for concentrated R & D funding.
Crossmedia impacts, in so far as they present an obstacle to wide-
scale recycle or reuse of water or conflict with RCRA require-
ments, will be considered in optimizing any system.
• Solid Waste Problems
The objective of the solid waste research program for the
nonferrous metals production industry is first to quantitatively
define the potential for environmental pollution resulting from
the disposal of process wastes and air and water pollution con-
trol residuals. This quantification will be followed by a
research, development and demonstration control technology
program leading to the establishment of environmentally accept-
able practices for smelter waste management for those wastes
found to have significantly harmful pollution potential.
Major smelter "solid" waste products are slag, flue dusts,
and wastewater pollution control residuals. In current smelter
practice, slags are sent to an open slag dump which is a specir..»
fied area for "storage" of slag materials. Flue dusts are
either recycled to the process or discarded to a segregated
"storage" area for some unspecified use in the future. Waste-
water, containing air pollution control residuals (including
acid plant blowdown), is either returned to the tailings pond
without treatment or is treated for discharge by sludge pro-
ducing hydroxide precipitation. All of these wastes contain
toxic trace elements in varying amounts. The environmental
acceptability of current tailings disposal practices is not
defined. However, it has been reported that the State of New
Mexico is experiencing contamination of groundwater by trace
metals, presumably from nonferrous mining and smelting operations.
14
-------
Waste disposal and the potential environmental damage that the
trace elements could cause, either through leaching and entry
into ground or surface waters or through entrainment into the
air as fugitive dust, are problems requiring attention. Solid
waste management technologies are required which will permit the
economic recovery of metal values from the waste coupled with
techniques which render the residual wastes environmentally
harmless.
The first phase of the program is to quantitatively define
the potential for environmental pollution from the disposal of
smelter process waste and air and water pollution control resid-
uals. The smelter environmental assessments, as a first step in
this problem definition exercise, has identified the points of
generation and general characteristics of many of the waste
materials produced. This study also identified arsenic-bearing
flue dust wastes as the first priority problem due to the
arsenic content and the potential for environmental contamina-
tion. The second step of this program is directed towards a
smelter-by-smelter survey of solid waste management problems and
practices. This study will lead to a sampling and analysis pro-
gram designed to test selected smelter wastes for leachibility
and environmental stability.
Arsenic Fixation
The second phase of this program is the research, develop-
ment, and demonstration of control technology suitable for abat-
ing pollution from solid waste management activities. In
response to the findings of the smelter environmental assessment
study regarding the hazard of arsenic bearing flue dust wastes,
we initiated a project to develop safe disposal/recovery
techniques (Investigation of New Techniques for Control of
Smelter Arsenic Bearing Wastes). This project is completing the
second year of its lab phase with efforts concentrated on the
development of waste fixation techniques and leach tests of the
fixed products. We are directing our efforts towards control of
arsenic-bearing flue dust wastes because (1) it is a large
source of soluble arsenic (2) the results are needed to support
the OAQPS planned regulation of arsenic emissions under Section
112 of the CAA.
Resource Recovery
Negotiations for a program with the Institute of Non-
ferrous Metals are underway (Poland) to study the feasibility
of copper smelter waste utilization under the Special Foreign
Currency Program.
Air Pollution Control Residuals
In conjunction with our flue gas cleaning projects, we
15
-------
will be investigating: (I) the characteristics of flue gas
cleaning wastes; (2) the possiblities for recovery and recycle
of metal and sulfur values from the wastes; (3) techniques for
safe ultimate disposal of residuals.
Water Pollution Control Residuals
The environmental aspects of water pollution control
residuals will be studied in conjunction with the treatability
techniques now under development. The program investigating
sulfide precipitation has produced results which indicate that
certain metals can be selectively removed from the precipitate
(sludge) for recovery. We are also exploring direct recycle of
sulfide sludge to permit recovery of metal values.
PROGRAM APPROACH
The goals of these R & D efforts are to develop and
evaluate pollution control methods that are broadly applicable
to the industry, that present cost effective options for meeting
regulations, and that minimize or eliminate intermedia transfer
of the pollution problem. Because of the federal manpower limi-
tations, we conduct all of our research activities via extra-
mural means, specificially: contracts, research or demonstra-
tion grants, and interagency agreements. Our research funding
is presented to us each fiscal year in the form of Congressional
appropriations. We in EPA's research office then allocate these
funds to projects based upon our awareness of technological
needs. Thus, our awareness to the most important research needs
is essential, since we are limited in our funding level. We feel
that industry input to this program is vital to ensure that
useful results - not just paper/reports - are obtained from the
expenditure of these funds. Industry participation is needed in
four areas: (1) in the activities leading to identification and
prioritization of research needs; (2) active participation in the
formulation of approaches to satisfy these needs; (3) active
participation in reducing selected control technologies to
practice under in-plant conditions; (4) review of research
findings and reports.
The nonferrous metals research program is available for
your use whether you are with EPA or with the industrial sector.
We have published reports and are currently addressing problem
definition and solution through active projects. We can provide
you with more detail on any one of these activities if you desire.
If you feel that you have a solution to a pollution
problem and would like to have our assistance in proving out
this solution, contact us for discussions of information perti-
nent to your project. We will be glad to work with you in pro-
viding background information which we have available and in
cosponsoring work for which we can make funds available.
16
-------
One last point regarding our working together - you may
have a potential solution to a problem that is not relevant to
the nonferrous industry, but is relevant to other segments of the
industrial sector. If this is the case, I can provide a list
of EPA researchers who have responsibility for specific indus-
trial areas.
17
-------
CHARACTERIZATION OF EMISSION CONTROL PROBLEMS
AT A MOLYBDENUM ROASTER
Klaus Schwitzgebel, C. Dean Wolbach
Radian Corporation
Austin, Texas
ABSTRACT
The roasting facility of the Molybdenum Company of America was
studied. The plant consists of two multi-level hearth furnaces, a high temp-
erature baghouse and a water quench scrubber. Molybdenum disulfide is
oxidized to molybdenum trioxide product and sulfur dioxide.
Removal efficiencies of the baghouse and quench-scrubber for heavy
metal compounds are 87% and 89% respectively. The low baghouse efficiency
during the testing period was due to broken bags. Sulfate contributions
to the grain loadings are 0.4% at the roaster exit (600°F), 17% at the
baghouse exit (450°F), and 91% at the quench scrubber exit (gas heated to
250°F according to EPA Method 5). Organic material does not contribute
significantly to the quench scrubber exit grain loading.
One roaster (1800 Ib/hr MoSz feed) produces 14000 scfm stack gas. The
main discharges to the atmosphere were:
• 1060 Ib/hr Sulfur dioxide
• 28 Ib/hr Particulates plus sulfates
• 0.1 Ib/hr Molybdenum
• 0.08 Ib/hr Selenium
• 0.015 Ib/hr Lead
• 0.007 Ib/hr Iron
» 0.015 Ib/hr Organic material
A bench scale wet electrostatic precipitator completely removed the mist
which presently forms the plume. A bench scale lime based S02 wet scrubber
was also operated and the scrubber products analyzed for impurities. The
impurities found were mainly introduced by the lime used and were not con-
stituents of the flue gas.
18
-------
INTRODUCTION
Molycorp's molybdenum sulfide roasting facility at Washington, Pennsylvania,
presently processes 10 million Ib/.vear molybdenum. Particulates and sulfur
dioxide escaping the multilevel hearth roasters need to be controlled to
bring the plant in compliance with local emissions laws.
Particulates are effectively controlled in a baghouse, quench cooler-
packed tower arrangement (quench scrubber). A perspective sketch of this
equipment arrangement is shown in Figure 1.
The most effective means for SO 2 control needs to be investigated. The
present study was made to supply information about the system performance
and to supply data for future changes. These goals were met by:
• Characterizing the roasting process with respect to the flow of
elements.
• Determining the effectiveness of the baghouse in removal of
particulates.
• Measuring the particulate removal efficiency of the quench scrubber.
• Analyzing the quality of the product of a lime based SOz scrubber.
« Determining all important baseline data of the present facility.
The main findings of the study are compiled in tables 1 through 5.
The results and their implications are described in the following chapters.
ROASTER OPERATION
During the sampling effort (June 23-78) only one roaster was operating.
The feed rate was 1800 Ib/hr MoSa consisting of a mixture of Questa ore
(1650 Ib/hr) and baghouse recycle dust (150 Ib/hr). The exit gas flow was
8500 scfra (70°F, 29.9" Hg) with an average grain loading of 4.68 grains/
scf and a SOa concentration of 13000 ppm (v/v) . A molybdenum trioxide pro-
duction rate of 1345 Ib/hr was calculated.
The analytical results for the roaster stream can be summarized as follows.
Semiquantitative survey analyses by Spark Source Mass Spectrometry (Table 5)
for 73 elements indicate that Twin Buttes and Questa concentrates show con-
centrations in the low ppm range for most heavy metals. Exceptions are lead,
zinc, copper, iron and manganese. Twin Buttes ore is high in copper (24,000
ppm versus 1000 ppm in Questa ore) , where as Questa ore shows a high lead con-
19
-------
centration (4300 ppm versus 930 ppm in Twin Buttes ore).
Volatilization of selenium and mercury is indicated during the roasting
process. The mercury and selenium flow rates in the feed stream are 2.3 x
10""* Ib/hr and 0.15 Ib/hr respectively. The corresponding figures in the
MoOs product are 0.3 x lO"* Ib/hr mercury and 0.003 Ib/hr selenium. Volatilization
at roaster temperatures gives an explanation for the depletion in the product.
Analytical data and flow determinations were checked by material balance
for 15 elements. The closure is good for lead, molybdenum, mercury, bismuth,
antimony, cadmium, silver, zinc, copper, manganese, and iron. The rhenium
concentration was below the detection limit. Marginal closures were found for
arsenic and selenium. The balance for nickel does not close. The reason is
not obvious.
BAGHOUSE BALANCE
Sampling to establish a baghouse balance was done simultaneously with
the roaster sampling. The following baseline conditions were measured:
Roaster Feed: 1650 Ib/hr Questa concentrate
150 Ib/hr recycle dust
Flow in: 8500 scfm
Flow out: 13100 scfm
Dust collection rate in baghouse: 272 Ib/hr
Grain loading in: 4.68 grains/scf
Grain loading out: 0.41 grains/scf
Collection efficiency 86% (corrected for cooling air dilution)
The chemical analysis of all ingoing and outgoing streams showed that the
element distribution of the dust collected in the baghouse is similar to that of
the ore. The main crystalline phase as determined by X-ray diffraction is
M6S2. Most of the mercury and a large amount of the selenium are not retained
in the baghouse.
Good material balance closure was found for: arsenic, lead, molybdenum,
mercury, bismuth, antimony, cadmium, silver, zinc, manganese, and iron.
Difficulties were encountered in the balance closure for selenium, copper
and nickel.
The removal efficiency of the baghouse based on element flow rates was
found to be 87%.
QUENCH SCRUBBER CHARACTERIZATION
The samples for the characterization of the Quench Scrubber were col-
lected on June 21-78. The following base line parameters were measured
20
-------
Roaster Feed: 1295 Ib/hr Questa Ore
159 Ib/hr Twin Buttes Ore
346 Ib/hr Baghouse Recycle Dust
Inlet gas Flow: 12200 scfm
Outlet gas flow: 14400 scfm
Inlet grain loading: 0.98 grains/scf
Outlet grain loading: 0.23 grains/scf
The water pickup was calculated from the water content at the inlet and outlet
gas streams to be 5 gallons/min. Part of the grain loading in both inlet and
outlet gas streams was due to sulfuric acid mist condensation. Sulfate forma-
tion contributed 0.17 grains/scf at the outlet.
The analytical findings indicate that 90% of the mercury escaping th;
baghouse is retained in the quench scrubber, and that 50% selenium removal is
achieved. The collection efficiency calculated from the elemental flow rates
at inlet and outlet is 89%. 20 Ibs per hour molybdenum are retained. This high
value was probably caused by broken bags lowering the efficiency of the baghouse.
Major discharges to the atmosphere are 28 Ibs/hr solids containing large amounts
of sulfates, and 1060 Ibs/hr S02. Heavy metal discharge with the flue gas amounts
to: 0.1 Ib/hr molybdenum, 0.08 Ib/hr selenium, 0.015 Ib/hr lead, 0.007 Ib/hr
iron, and 0.015 Ib/hr organic material.
LIME SCRUBBER - PRODUCT CHARACTERIZATION
The use of a lime scrubber for SOa control will produce about 2500 Ibs/hr
scrubber sludge from the off gases of one roaster. This sludge has to be disposed
of in an environmentally acceptable fashion.
Scrubber products with and without use of a wet electrostatic precipitator
(WEP) were generated in a bench scale experiment by scrubbing a slip stream of
the flue gas.
The wet electrostatic precipitator removed all of the acid mist. The
level of impurities in the product showed a strong correlation with the im-
purities in the lime. The three major components were iron, manganese, and
molybdenum.
Lime Product with WEP Product without WEP
Iron 1500 ppm 1550 ppm 1600 ppm
Manganese 72 ppm 95 ppm 70 ppm
Molybdenum 7 ppm 9 ppm 9 ppm
The concentration of the other elements was in the low ppm or ppb range.
The heavy elements in the scrubber liquor investigated were in all cases
below the drinking and irrigation water standards. These favorable results
are partly due to the short residence time of the liquid in the scrubber.
21
-------
Verification is needed in long term pilot plant studies.
ORGANIC COMPOUNDS IN PLANT STREAMS
The plume at Molycorp's facility was attributed by some earlier investi-
gators to the discharge of organic material introduced by flotation agents.
Extractions of selected solids, liquids and filters were, therefore, investigated
by gas chromatography using a flame ionization detector. The results of the
semiquantitative analyses show the following:
• Discharge rate with flue gas: 0.015 Ib/hr
• Quench water inlet: 0.045 Ib/hr
• Quench water blow down: 0.052 Ib/hr
• Ore feed: 0.1 Ib/hr
• Baghouse recycle dust feed 0.0001 Ib/hr
The discharge of 0.015 Ib/hr organic material can not be the cause of the plume
formation.
SULFATE CONTRIBUTION TO GRAIN LOADING MEASUREMENTS
The high oxygen, high SOz concentration of the roaster off-gases favors
the formation of SOs as the flue gas cools. The 80s reacts with solids to
form sulfates and with water vapor to form sulfuric acid mist. This increases
the grain loading results as the gas temperature decreases. Acid dew points
are 309°F at the roaster outlet, and 269°F at the baghouse outlet.
Grain loading determinations showed a particulate concentration at the
roaster exit (600 F) of 4.68 grains/scf. Analysis of the filter for sulfate
indicated a sulfate contribution of 0.4%. An average grain loading of .70
grains/scf was measured at the baghouse exit (450°F). Sulfate contribution
accounted for 17% of the particulate matter. Water droplets and acid mist
were present in the quench-scrubber exit gases. Water had to be evaporated
in order to determine a grain loading. The EPA method 5 was used. Contributions
of sulfuric acid mist accounted for 50%, 74% and 91% at sampling temperatures
of 380°F, 300°F and 250°F, respectively. These results point to sulfuric acid
mist as the cause of the plume formation. The mist was completely removed in
a bench scale wet electrostatic precipitator.
22
-------
BAGHOUSE EFFICIENCY AND PARTICLE SIZING STUDIES
Baghouse efficiency studies showed the baghouse to be operating below
expected levels. Efficiencies were found to be 85-90%. Particle sizing
showed a very high percentage of large (>2ym) material penetrating the
baghouse. This is indicative of unreplaced broken bags. Elements in the
vapor phase below 250°F were negligible.
CONCLUSIONS
1. Sulfate particulates and sulfuric acid mist, not organics, contribute
most to the grain loading observed in the stack gas.
2. The installation of a wet electrostatic precipitator will eliminate the
present plume problems.
3. The disposal of lime/limestone based scrubber byproducts needs to be
addressed. Areas of interest are:
• dumping sites
• sludge fixation
• leaching properties
4. The reason for the low baghouse efficiency during the testing period
appears to be unreplaced broken bags.
23
-------
to
cnmua 4 FT.
OF I>KUK A
NBIENI AID BLENDED [0 HAimMH IMHOUSE
EXIT ttHPEIWTURS At 400-450- F
BLOWDOffi LIQUID FED
TO K THICKEKER
(T)—SAW1.IK6 POINTS
FIGURE 1:
MIST COLLECTION SI5IEK AT HOLtC»»>
-------
TABU i. SDMURT OP SAXPLHIG croon JOHE-U;O
Dete
Roeeter Belance
6/23
Baghouee Balance
6/22
6/23
6/24
quench Scrubber
6/20
6/21
6/22
Site1
1
4
5
1
2
5
1
2
5
1
Efficiency
3
6
7
2
3
2
3
6
7
2
1
1
2
2
1
1
1
2
1
1
2
1
1
1
1
1
2
2
2
1
1
2
2
1
1
1
2
1
1
1
1
1
1
1
1
Sample
Grain Loading
SO, - SO,
vapor Phase Trace Elements
Total Clements
queata Ore Bin 6
Quests Ore Bin
Baghouae Recycle Duet
Product Molybdenum Oalde
Grain Loadings
Vapor Phase Trece Elements
Particle Size
Baghouse Recycle Dust
Grain Loadlnga
SO, - SOi
Vapor Phase Trace Element
Total Elements
Grain Loading
SOi - SO!
Total Elemente
Baghouse Dust
Particle Size
Grain Loadings
SO, SO,
Partlcls Size
Quench Scrubber
quench Scrubber
Greln Loading
SO, - SO,
Total Elements
Grain Loading
SO, - SOi
Total Element
Grain Loading
Vapor Phase Trace Elements
Total Elements
Particle size
Grain Loading
Vepor Pheee Trace Elements
quench Injection Hater
quench Blow Down Water
Procedure
EPA-3
EPA-8
Acld/Bsce Train
Radian HEP
24 Br Compoolte
24 ar Composlts
24 Ir Composite
24 tr Compoelte
EPA-5
EPA-8
In-steel Andersen
24 Br Coapoalta
ZPA-5
EPA-8
Acid/Base. Train
Radian VEP
EPA-5
EPA-8
Radian HEP
24 ar. Compoelte
In-etaek Brinks
EPA-S
EPA-a
Out-of-Staek Andersen
Injection Hater
Blow Down Voter
EPA-5
EPA-0
Radian HEP
EPA-5
EPA-<
Radian HEP
EPA-5
Add/Base Trsln
Radian HEP
In-Staek Andersen
KPA-5
AcU/Baee Train
Grab Sample 4 Plow Rate
Greb Sempls t Flow Race
Comments
Flltere st 3aO°F
Sempler et 300°F
Greb Semple e Flow Rate
Grab Sample > Flow Rate
la-Stack Filter
Filter et 300°F
In-Stack Filter
Filter At 2SO°P
Scrubber Product Cheracterizecion
6/20
6/23
3
3
1
1
Radian Scrubber Product
Radian Scrubber Product
without Radian HE?
With Radian HEP
Baghouea Efficiency Studiee
6/22 2
6/23 1
2
6/24
6/26
6/27
Grain Loading
Particle SU«
Grain Loadinga
SU; - SO,
Grain Loading
SO, - SOi
Particle Size
Grain Loadinga
SO, - SO,
Particle Siie
Grain Loadinga
SO, - SO,
Particle Size
Grain Loading
SO: - SOt
Particle Size
Grain Loading
SO, - SO,
Particle Size
EPA-5
In-Staek Anderaen
EPA-5
EPA-a
EPA-5
EPA-a
Iii-Stack Brink.
EPA-5
EPA-B
In-Stack Brink*
EPA-5
EPA-a
In-Stack Anderaen
EPA-5
EPA-8
In-Stack Brink.
EFA-5
EPA-a
In-Stack Anderaen
Atonapberlc Eniaaione
6/20
6/21
6/22
'Sanple Site
3 1
1
1
1
2
Greln Loading
SO."' aa B,SO.
Organic.
Grain Loading
SO,"' «• H,SO,
Organica
SO, - SO,
50, - SO,
Grain Loading
so.'2 as B9SOk
2 SO, - SO,
3 1 (Sreln Loading
JSO,"! «» H,SO.
(Organics
I
3 1
SO, - SO)
Grain Loading
SO.-'as 1,SO,
Organica
1 Between Roaster and Baghouae
2 Between Baghouae and quench Scrubber
EPA-S
EPA-5
EPA-8
EPA-a
EPA-5
BPA-8
EPA-5
EPA-a
EPA-5
Dual Filter! at 380°F
Vertical Traverse
Duel Flltere at 365°F
Horizontal Traverse
380°F
365»F
Dual In-Stack Flltere
Out of Stack Filter at
Dual Filters at 300*7
Dual Flltere at 250°F
Plus
300°P
After Quench Scrubber
Concentrate Peed at feeder
Baghouie Du*t Feed at Feeder
Recycle Punp By-Paw
Quench Scrubber Blow Down Drain
Produce Stream Fron Roaster
25
-------
TAKE 2. SOTIARY OF HASE LI»IE HF-AS'JRQKMT RESULTS
EXPERIMENTAL RESULTS
1.
2.
3.
4.
5.
6.
7.
8.
9.
Ore feed races
• Quest* 6 Ibs/hr
• Twin Buttes Ibs/hr
e Baghouae Dust Recycle
TOTAL FEED RATE
MoOi Production rate (Ib/hr)
Gas Flow
• Roaster Exit scfa
• Baghouse Exit scfla
• Stack Discharge scfm
Grain Loading
e Roaster Exit grains/scf
e Baghouse Exit grains /scf
• Stack grains/scf
Gas Composition
• Roaster Outlet - SO, ppm
- SO, ppn
HzO 7.
e Baghouse Outlet - SOi ppm
SO, ppB
H,0 I
e Stack Discharge - SO, ppm
SOi ppm
H,0 i
Baghouse Dust Catch Ib/hr
«
Intake Quench Water gal/mln
Quench Vater Bloudown gal/mln
Particle size distribution (X leas
e Brinks Roaster Exit
- Cyclone catch *<3.1 urn
- Plate 1 <1,2
- Plate 2 o 50
- Plate 3 <0.35
- Plate 4 <0 22
- Plate 5 <0.14
e Anderson Baghouse Exit
- Plate 1 <7.4 un
- Plate 2 <6,8
- Plate 3 <4.2
- Plate 4 <2~8
- Plate 5
37 30
25 20
12 10
1.1 3.2
70
58
51
42
33
24
17
11
June-27
ND
HD
ND
ND
9200
13000
ND
7.33
.71
ND
12700
114
4.4
5980
15
3.5
ND
HD
HD
HD
HD
ND
20
15
11
8 A
.8
6.6
69
53
hO
.1i
29
23
19
17
26
-------
J.
SUMrVW OF QUANTITATIVE
ANALYTICAL RESULTS
Sample
Description
Element" --x_^^
Arsenic
Lead
Molybdenum
Hercury
Selenium
Bismuth
Antimony
Cadmium
Silver
Zinc
Nickel
Copper
Manganese
Iron
Rhenium
ROASTER BALAHCK
Quests Twin Buttes MoOi
Ore Ore Product
ppm pp» ppm
12 9.9 19
4300 980 4800
470000 360000 540000
.IS .13 .02}
56 122 2.0
280 63 160
21 HD 19
18 14 20
38 62 49
1500 240 1700
96 20 59
950 24000 1SOO
730 840 890
11800 32500 17000
<10 <10 <10
HEP
Site 1
June-23-78
ug/scf*
IS
1780
149000
.22
1S7
113
7.0
7.6
14.2
485
8.4
401
197
3750
-------
TABLE 4. SUMMARY OF ELEMENTAL FUM
RATES IN ROASTER STREAMS
Ssnple
Element
Arsenic
Lead
Nalybdenua
Mercury
Selenlixi
Bismuth
AntiBKHiy
CadBluB
Silver
Zinc
Nickel
Capper
Mnnganeee
Iron
Rhenium
ROASTER STREAMS 6-23
quite Dust HoOi Boaster
*«•* Peed Product Dult
<1«0 Whr) (150 Ib/hr) (1345 lb/hr> (341 Ib/hr)
Ib/hr Ib/hr Ib/hr Ib/hr
•020 .007 .025 .017
7.1 .90 6.44 2.00
7BO 74 725 lit
2 * 10-% .3 > 10"% .3 > 10"- 2.5 x 10""
.092 .060 .003 .18
.46 .05 .21 .13
.035 .004 .025 .008
.030 .003 .027 039
.063 .008 .066 .016
2.48 .29 2.28 .55
.16 008 .079 .005
1.6 .56 2.0 .45
1.2 .12 1.19 .22
20 2.7 22.8 4.22
<.02 < 002 <.01 <..J
BACHOUSE STREAMS
6-23
Beghouse Beghouee
Dust Gea &, Dust
(272 Ib/hr) (46 Ib/hr)
Ib/hr Ib/hr
.013 .003
1.64 .24
134 20
.6 x 10" 2.0 « 10"'
.11 .16
.10 .008
.007 .001
.006 .001
.014 .002
.52 .08
.015 .001
1.0 .03
.21 .02
4.88 .05
•c.003 .«.OOS
HUHIDIFIEIt - SCRUBBER
STUUMS JUNE-21-78
Beghouie quench quench quench
Gas & Duet Off-Gases Hater Inlet Water Outlet
(103 Ib/hr) C16 Ib/hr) (31 gal/Bin) 26 gel/»ln
Ib/hr Ib/hr US/8 ugVg
.003 .0007 -020 .022
.21 .015 •" -134
19.8 .101 48 51
2.0 x 10"' .3 x 10"" 5 i 10— 8 > 10"'
.16 .08 -11 -1*
.008 4 i 10"% -02 -05
.001 .6 x 10"' -005 -005
.001 3 x 10" -008 .001
.002 7 x 10"* <.0008 <.0007
.08 5 x 10-> -6» .685
.001 <1 x 10" -03 .030
.033 1.5 x 10"' 2-5 *•*
.023 1.3 x 10-' -«2 -'1J
.054 7.3 x 10"' 13.2 13.3
<. 005 <.0003 <1.3 <.04
LIME 6 SCRUBBER PRODUCT
Gyp sum Cypsum
Lime Without Prescrubber With Prescrubber
1110 Ib/hr) (2590 Ib/hr) (2590 Ib/hr)
Ib/hr Ib/hr Ib/hr
1 > 10"' 3 x 10"' 3 x 10"'
-------
TABlf 5. SUfflftRY OF SURVEY WIALYSES
BY
SPARK SOURCE fttSS SPECTROHETW
10
vo
Twin Buttes Queata
Ore Ore
Element ppa ppm
uranium
Thorium
Bismuth
Lead
5 Thallium
Gold"17
Platinum
Irldlum
10 Cl,m1,m
Rhenium
Tungsten
Tantalum
Hafnium
^5 I.itr»f4i»i.
Ytterbium
Thulium
Erbium
Rolalum
20 uyflproa.! v*»
Terbium
Gadolinium
Europium
25 ^ZStSZL.
Praseodymium
Cerium
Lenthanun
Barium
30 Cesium
Iodine
Tellurium
Antiaony
Tin
35 Indium
8
3
90
430
NK
.2
140
26
< 4
.5
.6
*1
.2
.9
.5
2
6
*2
14
12
31
3
35
22
41
26
STD
15
22
40
>1000
.7
HR
11
5
<2
.9
1
2
.4
1
"2
6
3
12
43
44
1
38
72
19
STD
MoO,
Product
ppa
20
32
77
>1000
11
<"2
2
10
1
.2
.7
1
.7
3
.9
9
25
5
32
58
60
1
10
42
78
51
STD
June 23
HEP
Site 1
yg/scf
7
7
29
>360
3.5
HR
3."s
-
.1
.7
3.7
1.8
21
21
18
.3
3.5
25
14
1.8
STD
Baghouse
Dust
ppm
10
*17
300
>1000
.7
HR
13
9
<4
.5
.6
1
<.2
1
.5
2
13 .
6
30
30
56
2
<5
220
91
47
STD
June 23
HEP
Site 2
Mg/acf*
.3
3
IlK
5
-
-
.06
1
.7
.4
.6
STD
June 21
HEP
Site 2
ug/acf*
.3
.3
26
.02
HR
4
.4
<.07
.
.06
.1
'.4
.03
.2
.5
1
STD
June 21
HEP Quench
Site 3 Hater In
ug/ecf* ng/ml
.04
.2
HR
*3
-
-
.05
.07
.04
STD
.1
.2
.4
4
.01
nK
.2
.6
.005
.008
"..02
.008
.1
1
3
8
.02
<.001
.01
< 1
.2
STD
Quench
Hater Oat
Mg/ml
1
3
.1
.3
•c.04
2
3
<.08
.02
.1
.01
.1
.1
.05
.1
.1
.9
2
.2
.02
.05
.2
STD
Scrubber
Product
Without VEP
Line Prescrubber
ppm ppm
7
1
NK
-
-
1
.3
3
2
10
.7
STD
<2
1
HR
-
.
-
.3
6
.7
STD
Scrubber
Product
With- VEP
Prescrubber
ppm
<2
1 -
MR
-
.
•
.3
2
6
STD
*Mg/acf of gas sample
HR - Not reported
STO - Standard
Continued
-------
TABLE 5. (CONTINUED) S1WARY OF SURVEY ANALYSES
BY
SPARK SOURCE MASS SPECTRONETRY
U>
O
Element
Cadmium
Silver
Palladium
Rhodium
4H Rnt-h»n{im
Molybdenum
Niobium
Zirconium
Yttrium
Rubidium
Bromine
Selenium
Arsenic
SO nmunlun
Gallium
Zinc
Stele!
5S nrihilr
Iron
Manganeae
Chromium
Vanadium
^0 T4r«Hum
Scandium
Calcium
Pottaaalum
Chlorine
ej^ Sulfur
Phosphorus
Silicon
Aluminum
Magnealum
70 Sodium
Fluorine
Oxygen
Nitrogen
Carbon
Beryllium
Lithium
Hydrogen
Twin Buttea
Ore
ppm
170
120
>1000
3
19
17
47
25
10
290
49
7
10
>1000
>1000
210
>1000
>1000
6
22
INT
5
>1000
480
13
>innn
50
>1000
>120
>120
>300
>1000
NR
NR
NR
s
4
NR
Queata
Ore
ppm
67
220
>1000
11
27
14
8
.8
110
43
I
5
>1000
930
210
>1000
>1000
24
39
TNT
12
>1000
>1000
8
>innn
>1000
>1000
200
>1000
20
•v320
NR
NR
NR
7
NR
MoOf
Product
ppm
330
75
>1000
23
29
16
74
13
4
7
47
I
15
>1000
>1000
75
>1000
>1000
52
42
INT
3
>1000
>1000
42
>1000
>1000
>1000
>1000
1.160
OR
NR
NR
11
NR
June 23
WEP
Site 1
ug/acf
25
32
>360
3.5
1.4
2.5
7 1
3.2
.7
2.1
7.1
-1
2.5
>360
>360
14
7
>360
290
3.5
3.5
INT
.1
>360
>360
-1
>360
290
70
>360
>210
•(.140
NR
NR
NR
.7
1.4
3.5
NR
Bughouse
Dust
ppm
85
160
>1000
15
19
17
83
10
3
480
81
7
4
>1000
>1000
48
22
>1000
>1000
6
22
INT
>1000
>1000
13
>1000
>120
>1000
50
1.140
NR
NR
NR
.7
~4
NR
June 23
WEP
Site 2
Ug/acf*
.04
>95
.04
.6
.07
i
I
.05
.02
6
>95
50
2
1
2
INT
>.02
>95
10
.09
>?5
7
30
>20
40
>50
KR
KR
NR
.*
.07
NR
June 21
WEP
Site 2
Ug/acf*
1
>90
'.2
.6
.2
.04
9
2
.2
.1
27
90
2
1
>90
.8
.5
INT
>90
62
.5
>90
27
9
>90
NR
NR
NR
.04
NR
June 21
WEP
Site 3
ug/acf*
>;>
>90
.60
.08
.04
9
.9
>.04
.63
*2
.7
1
.04
27
'.3
.04
INT
>90
27
.1
>90
36
18
4
9
NR
HR
NR
.01
NR
Quench
Water In
ug/ml
.7
.08
.3
.1
.2
3
.02
2
1
. .02
.1
>10
>10
2
1
.5
INT
.UB
>10
>10
.3
>10
>10
>2
NR
NR
NR
.7
NR
Quench
Water Out
ug/ml
.8
.5
1
1
1
7
.3
.1
.5
.1
.4
7
>10
>10
1
INT
>10
>5
NR
NR
NR
6
'.a
NR
Lima
ppm
.7
10
.8
4
2
430
T4
1
3
1
9
3
8
.4
>1UUU
100
9
12
>1000
120
28
62
86
>1000
>200
>1000
^
5«T
NR
NR
NR
1
NR
Scrubber
Product
Without WEP
Rreacrubber
ppm
.9
5
.5
3
2
ISO
•C.I
'9
2
5
35
6
29
3
MOOO
22
1
*!
>1000
87
>100u
>1000
>210
>1000
1.190
HR
NR
NR
.5
.6
NR
Scrubber
Product
•With WEP
Preacrubber
ppm
.6
2
'"3
5
150
<1
9
3
3
35
9
>1000
36
2
7
•in
nooo
87
80
>ifl(K)
91
nooo
> 210
nooo
•^50
HR
NR
NR
.7
NR
*vg/scf of gaa sampled
NR - Not Reported
STO - Standard
-------
PRESENT AND FUTURE
CONTROL OF FUGITIVE EMISSIONS
IN THE PRIMARY NONFERROUS METALS INDUSTRY
Alfred B. Craig, Jr., IERL
U.S. Environmental Protection Agency
Cincinnati, Ohio
L.V. Yerino, M.D. Giordano, and T.K. Corwin
PEDCo Environmental, Inc.
Cincinnati, Ohio
ABSTRACT
The functions of the process equipment used in the primary
copper industry; sources and characteristics of fugitive emis-
sions; methods currently employed for control of fugitive emis-
sions and their collective efficiencies; future considerations
for controlling fugitives with existing technology and new con-
cepts in process and/or control equipment; and costs and energy
requirements for existing and future techniques for fugitive
emission controls are reviewed.
INTRODUCTION
This paper deals with fugitive emissions from the primary
copper smelter as it exists today and takes a look into the future
with respect to known technology and a combination of ideas from
other industries to minimize these fugitive emissions. The
thoughts and ideas presented are those of the PEDCo author. Their
presentation here does not represent an approval of the U.S.
Environmental Protection Agency.
Figure 1, which represents a conventional smelter layout,
shows the feed being conveyed to a multihearth roaster, then trans-
ported to a reverberatory furnace. The matte from the reverbera-
tory furnace is charged to a Peirce-Smith converter, and the slag
31
-------
is sent to a disposal area. The blister copper from the conver-
ter is ladled to an anode furnace, and the converter slag is
returned to the reverberatory furnace.
The fugitive emission figures shown on the following tables
were estimated from information supplied in reports and papers
presented at other meetings. These values can differ substantially
from plant to plant. Before making these estimates, we attempted
to secure published values that had been arrived at through field
measurements, but several inquiries proved unsuccessful. The fol-
lowing are the reports and papers from which the information was
extracted:
The Bureau of Mines report RI 7957, entitled "Cost of Produc-
ing Copper from Chalcopyrite Concentrate as Related to S02
Emission Abatement," by Masami Hayashi, H. Dolezal, and J.H.
Bilbrey, Jr. (1974).
The Radian Corporation report, entitled "Pollution Control
and Heat Recovery from Non-Ferrous Smelters," Volume II.
An Abstract, entitled "Oxygen Smelting in the Noranda Pro-
cess," by J.B.W. Bailey, R.R. Beck, G.D. Hallett, C. Washburn,
And A.J. Weddick. Presented at the AIME annual meeting in
New York City in February 1975.
An abstract, entitled "New Developments in Outokumpu Flash
Smelting Methods," by S.U. Harkki and J.T. Juusela.
Presented at the AIME annual meeting in Dallas in February
1974.
The Bureau of Mines report was the primary source of the
emission values, but we modified the numbers based on producing
equivalent tonnages of blister copper.
From a review of currently used methods for controlling fugi-
tive emissions, it appears such methods consist primarily of
hooding and building enclosures. The efficiency of these systems,
when retrofitted to existing reverberatory furnaces or converters,
is limited in all phases of the process operation. For example,
a Peirce-Smith converter equipped with fixed or secondary fugitive
hooding will capture at the most a small percentage of the fugi-
tive emissions created during the charging of matte or slagging
or blister copper pouring as shown in Figures 15, 16, and 17 (see
Appendix B-l). When slag from the Fierce-Smith converter is being
charged, the reverberatory furnace permits the fugitive emissions
to be discharged to the surrounding area and to the roof monitors.
32
-------
Table V
Estimated Fugitive Emissions
SO,
Cu
Fe
Others
Multihearth (Figure 2)
Reverberatory (Figure 3)
Peirce-Smith Converter (Figure 4)
Fluid-bed Roaster (Figure 5)
Reverberatory (Figure 6)
Peirce-Smith Converter (Figure 7)
Noranda (Figure 8)
Peirce-Smith Converter (Figure 9)
Dryer (Figure 10)
Outokumpu (Figure 11)
Peirce-Smith Converter (Figure 12)
Dryer (Figure 10)
Electric Furnace (Figure 13)
Peirce-Smith Converter (Figure 14)
These tables provide a comparison of the different process
systems. Percentage shown could vary substantially from plant to plant.
Percentage of sulfur charged to the process equipment, expressed as S02-
Percentage of the total charged to the process equipment.
4.2
6.5
Table 2a
4.2
6.5
Table 3a
2.9
1.5
Table 4a
2.9
2.1
Table 5a
3.0
6.5
0.16
0.16
0.16
0.16
0.11
0.03
0.15
0.03
0.15
0.19
0.10
0.21
0.09
0.21
0.07
0.33
0.15
0.38
0.07
0.21
0.10
0.11
0.12
0.13
0.04
0.12
0.07
0.08
0.04
0.06
33
-------
EMISSION CONTROL EQUIPMENT
co
CONVERTER
WASTE HEAJ
BOILERS
REVERBERATORY
FURNACE
Figuro 1. Conventional smelter layout.*
Source: "Background Information '7~~ Mew ^ourcp Performance Standards: Primary Copper, Zinc, Lead Smelters"
Volume I: Proposed Standards. EPA 450/2-74-002a. October 1S74.
-------
Feed
S02
C02
383.0
53.5
(54.8)
(7.6)
7.0% vol.
Cu
Fe
S
Si 02
CaO
Other
308.6
308.6
382.9
102.9
5.7
34.2
(44.1)
(44.1)
(54.7)
(14-7)
(0.8)
(4.9)
RABBLEK,
ARM
RABBLE
BLADE
Flux
CaO
Si 02
C02
Other
49.0
97.1
37.3
12.2
(7.0)
(13.9)
(5.3)
(1.7)
FUGITIVE a
EMISSIONS
02
C02
191.5
16.2
(27-4)
(2.3)
NATURAL
GAS
COOLING
AIR
Material balance in tons/day.
Numbers in parentheses indicate
capacity of each unit.
Other values are based on 363
tons/day of blister copper.
Values not available.
Figure 2. Multiple-hearth roasting furnace.*
^Source: Background Information; EPA 450/2-74-002a, October 1974-
35
Cu
Fe
S
Si02
Ca'0
Other
308.6
308.6
191.4
200.0
54.7
46.4
(44.1)
(44.1)
(27.3)
(28.6)
(7.8)
(6.6)
-------
Fuq Hives
S
Cu
CaO
02
Other
Cu
Fe
S
Si 02
CaO
Other
308.6
308.6
141.4
200.0
54.7
46.4
3.7
12.0
239.1
115 2 CONVERTER
5!0
69.2
6.9
AIR AND -*£
OXYGEN
BURNERS
MATTE
SLAG
FETTLING PIPES
SU2
Cu
Fe
S
E7T
0.5
0.5
0.2
FETTLING DRAG
CONVEYOR
OFF-GAS
S02|114.7
1 vol.%
SLAG Cu
Fe
S
K ST02
5 CaO
02
Other
4.6
308.7
6.8
315.2
56.0
63.7
53.1
MATTE
Cu
Fe
S
CaO
02
315.5
238.5
126.6
3.7
47.7
Material balance in tons/day.
All values are on basis of 303 tons/day
of blister copper (slightly below
capacity of unit).
Figure 3. Reverberatory smelting furnace*
*Source: Background Information; EPA 450/2-74-002a, October 1974.
-------
Matte
Cu
Fe
S
CaO
02
315.5
238.5
126.6
3.7
47.7
(129.3)
(97.7)
(51.9)
(1.5)
(19.5)
S00 237.0
TUYERE
PIPES
Slag
Cu
Fe
S
Si02
CaO
02
Other
12.0
239.1
3.7
115.2
5.0
69.2
6.9
(4.9)
(98.0)
0.5)
(47.2)
(2.0)
(28.4)
(2.8)
(97.2)
4% vol.
Fugitives
OFF-GAS
Blister Copper
Cu
Fe
Other
303.0
1.5
1.5
U24.2)
(0.6)
(0.6)
S02
Cu
Fe
S
8.2
0.5
0.5
0.2
(3.4)
(0.2)
(0.2)
(o.i)
SILICEOUS
FLUX
Fe
Si 02
CaO
02
Other
2.6
115.2
1.3
0.7
8.2
(1.1)
(47.2)
(0.5)
(0.3)
(3.4)
tf
AIR
02
143
.3
(58.
9)
Source:
Material balance in tons/day.
Values in parentheses indicate capacity
of unit; other values are on basis of
303 tons/day blister copper.
Figure 4. Copper converter.*
Background Information; EPA 450/2-74-002a, October 1974.
-------
Feed
Cu
Fe
S
Si 02
CaO
Other
308.6
308.6
382.9
102.9
5.7
34.2
oo
oo
FUGITIVE
EMISSIONS
Flux
CaO
Si 02
C02
Other
49.0
97.1
37.3
12.2
S02|255.8|13% vol
C02I 39.0
Cu
Fe
S
Si02
CaO
Other
308.6
308.6
255.0
200.0
54.7
46.4
'Source:
Material balance in tons/day.
Tonnages are for capacity of unit
a Values not available.
Figure 5. Fluid-bed roaster.
Background Information; EPA 450/2-74-002a, October 1974.
-------
Cu
Fe
S
SlOo
CaO
Other
308.6
308.6
255.0
200.0
54.7
46.4
-CALCINE
OJ
FETTUNG DRAG
CONVEYOR
S02I153.0|1% vol
SLAG Cu
Fe
S
Si 02
CaO
Other
Converter
Slag
Cu
Fe
S
Si 02
CaO
02
Other
12.0
240.0
5.0
115.2
5.0
91.2
7.8
MATTE
FETTLING PIPES
Material balance in tons/day.
Tonnages are for capacity of unit.
MATTE
Figure 6. Reverberatory smelting furnace*
*Source: Background Information; EPA 450/2-74-002a, October 1974.
-------
Matte
O
Cu
Fe
S
CaO
02
315.5
238.5
168.7
3.7
63.4
(119.8)
(90.6)
(64.1)
(1.4)
(24.1)
TUYERE
PIPES
SLAG
S
Cu
Fe
Si02
CaO
02
Other
5.0
12.0
239.1
115.2
5.0
91.2
6.9
0.9)
(4.6)
(90.8)
(43.8)
(1.9)
(34.6)
(2.6)
SOo 316.0
(120.0)
4% vol
OFF-GAS
Fugitives
S02
Cu
Fe
S
11.0
0.5
0.5
0.2
(4.
(0.
(o.
(0.
2)
2)
2)
1)
SILICEOUS
FLUX
/
ifffi-f
m
Fe
Si 02
CaO
02
Other
2.6
115.2
1.3
0.7
8.2
(1.0)
(43.8)
(0.5)
(0.3)
(3.2)
FLUX GUN
AIR
Cu
Fe
Other
303.0
1.5
1.5
(115.0)
(0.6)
(0.6)
AIR
°2
190.
8
(72.
3)
Material balance in tons/day.
Values in parentheses are
capacity of unit; all others
are on basis of 303 tons/day
blister copper.
'Source:
Figure 7. Copper converter.*
Background Information; EPA 450/2-74-002a, October 1974.
-------
Palletized
Concentrates
Flux
Cu
Fe
S
SiOz
357.7
395.7
421.9
67.6
(431.7)
(477.6)
(509.2)
(81.6)
Fe
S
Si02
11.7
2.4
181.6
(14.1)
(2.9)
(219.2)
FEEDER
so2
650.7
(785.
4)
6% vol.
S02
OFF-GAS
FUGITIVES
CONCENTRATE
•PELLETS AND FLUX
HOOD
S02
Cu
Fe
S
12.3
0.4
0.3
0.3
(14.8)
(0.5)
(0.4)
(0.4)
liiiiniiiiii iiiiiiminmim
COPPER
02 [331.5 [(400.1)
Material balance in tons/day.
Values in parentheses indicate capacity
of unit; all others are on basis of
303 tons/day blister copper.
me
Cu
Fe
S
n-braae watte
315.0
8.4
82.2
(380.2)
(10.1)
(99.2)
Figure 8. Noranda continuous smelting.
k
Source: Background Information; EPA 450/2-74-002a, October 1974.
-------
Mat1
Cu
Fe
S
315.0
8.4
82.2
te
(233.0)
(6.2)
(60.8)
TUYERE
PIPES
>£»
to
Slag
Cu
Fe
S
Si02
Other
11.9
7.5
2.0
25.5
0.8
(8.8)
(5.5)
(1.5)
(18.9)
(0.6)
S02|159.0|(117.5)
4% vol
Fugitives
SO?
Cu
fe
S
1.2
0.1
0.005
0.1
(0.9)
(0.1)
(0.004)
(0.1)
SILICEOUS
FLUX
.Jfci- =
Si02
Fe
Other
25.5
0.6
2.3
(18.9)
(0.4)
(1.7)
Cu
Fe
Other
303.0
1.5
1.5
(224.1)
(1.1)
(1.1)
°zl
80.1
[[59.2)
AIR
Material balance in tons/day.
Values in parentheses indicate capacity
of unit; all others are on basis of 303
tons/day blister copper.
Figure 9. Copper converter.'
'Source: Background Information; EPA 450/2-74-002a, October 1974.
-------
FEED CHUTE FRICTION SEAL
INLET HEAD
(COUNTER FLOW ONLY)
SPIRAL FLIGHTS
NO.l RIDING RING
TRUNNION AND
THRUST ROLL
ASSEMBLY
GIRT
GEAR
KNOCKER
BREECHING
SEAL
DRIVE
ASSEMBLY
LIFTING
FLIGHTS
BREECHING
NO.2
RIDING
RING
TRUNNION ROLL
ASSEMBLY
DISCHARGE
Figure 10. Pictorial of a countercurrent direct-heat rotary dryer .s
(Combustion chamber not shown.)
Source: Adapted from Surface Mining and Our Environment, U.S. Department
of the Interior. 1967.
-------
Dried
Concentrates
Fugitives
S
Fe
Cu
Si02
Ni
Other
499.6
467.6
330.6
64.1
1.6
93.1
FEED
1
Si 02
Fe
Ni
Other
205.7
8.1
2.3
177.6
, CONCENTRATE BURNER
PREHEATED cf>
AIR
MATTE
SLAG MATTE
SETTLER
Cu
Ni
Fe
S
Si 02
Other
15.1
0.4
449.5
16.1
269.8
262.3
Material balance in tons/day.
Values indicate capacity of unit.
Figure 11. Outokumpu flash smelting furnace.*
*Source: Background Information; EPA 450/2-74-002a, October 1974.
-------
Matte
Cu
Ni
Fe
S
Other
315.0
3.5
25.5
89.1
8.4
(214.0)
(2.4)
(17.3)
(60.5)
(5.7)
TUYERE
PIPES
Ul
Slag
S
Cu
Fe
Si02
Ni
Other
2\3
11.9
24.8
35.6
3.5
10.1
(1.6)
(8.1)
(16.7)
(24.2)
(2.4)
(6.9)
S0?
171.5
(116.3)
4% vol.
Fugitives
OFF-GAS
S02
Cu
Fe
S
1.9
0.1
0.01
0.1
(1.3)
(0.1)
(0.01)
(0.1)
SILICEOUS
FLUX
/
Fe
Si 02
Other
0.8
35.6
3.2
(0.5)
(24.2)
(2.2)
PNEUMATIC
PUNCHERS
BLISTER COPPER
Cu
Fe
Other
303.0
1.5
1.5
(205.8)
0.0)
(1-0)
AIR
"2!
86.
7
(58.
8)
Material balance in tons/day.
Values in parentheses indicate capacity
of unit; all other values are based
on 303 tons/day blister copper.
Figure 12. Copper converter.*
*Source: Background Information; EPA 450/2-74-002a, October 1974.
-------
Dried
Concentral
Cu
Fe
S
Si02
CaO
Other
308.6
308.6
379.1
102.9
5.7
38.4
,es
(404.9)
(404.9)
(497.4)
(135.0)
(7.5)
(50.4)
Fugitives
S02
Cu
Fe
S
11.6
0.5
0.5
0.3
(15.2)
(0.7)
(0.7)
(0.4)
Converter
Slag
CTi
Cu
Fe
S
Si02
CaO
Other
18.9
372.9
7.7
179.4
7.8
154.0
(24.8)
(489.2)
(10.1)
(235.4)
(10.2)
(202.0)
CONVERTER
SLAG
LAUNDER
n
OFF
i
V<
-G
fr
•:
•s
AS S02
238.0 (310.2)
FLUX
Si 02
CaO
C02
Other
32
48
37
9
5% vol.
.9
.0
.3
.0
(43.2)
(63.0)
(48.9)
(11.8)
Cu
Fe
S
CaO
Other
322.5
371.2
253.5
5.8
61.0
(432.1)
(487.0)
(332.6)
(7.6)
(80.0)
Air
°2
123.9
(162.
7)
Material balance in tons/day.
Values in parentheses indicate capacity
of unit; other values are on basis of
303 tons/day blister copper.
Figure 13. Electric smelting furnace*
*Source: Background Information; EPA 450/2-74-002a, October 1974.
Cu
Fe
S
Si 02
CaO
Other
4.5
309.8
9.1
315.1
55.7
140.4
(5.9)
(406.4)
(11-8)
(413.6)
(73.1)
(184.2)
-------
Matte
Cu
Fe
S
CaO
Other
322.5
371.2
253.5
5.8
61.0
(95.4)
(109.8)
(75.0)
(1.7)
(18.0)
TUYERE ™
PIPES
Cu
Fe
S
Si02
CaO
Other
18.9
372.9
7.7
179.4
7.8
154.0*
5.6)
(110.4)
(2.3)
53.1)
(2.3)
(45.6)
OFF-GAS
^S02[47'4.5
(140.3)
3.5% vol.
*Contains 0?. 80.4 (23.8)
Fugitives
§02
Cu
Fe
S
16. b
0.6
0.8
0.3
^^^^^^^^^^
(4.9)
(0.2)
(0.2)
(0.1)
Fe
Si Op
CaO
Other
4.0
179.4
2.0
14.1
(1.2)
(53.1)
(0.6)
(4.2)
FLUX GUN
AIR
PNEUMATIC
PUNCHERS
Blister Copper
Cu
Fe
Other
303.0
1.5
1.5
(89.6)
(0.4)
(0.4)
(96.4)
Material balance in tons/day.
Values in parentheses indicate capacity
of unit; other values are on basis of
303 tons/day blister copper.
*Source:
Figure 14. Copper converter?
Background Information; EPA 450/2-74-002a, October 1974.
-------
After looking at the existing process equipment, we suggest
an attempt be made to minimize fugitive emissions from these pro-
cesses by applying any of the following techniques where feasible,
As an illustration on most operations, increasing the available
draft for process gases reduces the fugitives but also dilutes
the gas stream's SOa content, increases the size of all following
equipment, etc. (We realize, of course, that each particular
smelter has its own unique problems.)
ROASTERS
Preventive maintenance to keep the unit tight and min-
imize leakage through the walls
A fugitive collection system to collect emissions during
the filling of a transfer or larry car from the roaster
area
A cylindrical envelope installed about the roaster.
(Sufficient space should be left between the envelope
and the roaster unit to allow maintenance; this space
would be connected to a fugitive emission collection
system.)
LARRY OR TRANSPORT CAR
A collection system for fugitive emissions at both
charging and discharging points
A cover over the car when it is in transit
Use of a pneumatic system for conveying the calcine
product to the furnace storage bins
REVERBERATORY FURNACE
Fugitive emission collection system, consisting of
hooding and envelopes, for the launders and ladles
Preventive maintenance on refractories, etc. to
minimize fugitive emissions
Converter slag return opening fitted with a pivoted
and method of slag return
48
-------
T.O. RAIL
SMOKE
HOOD
PLENUM
TO SECONDARY A
HOODING
MAIN DUCT
Figure 15. Secondary converter hood configuration,
49
-------
LADLE
Figure 16. Peirce-Smith converter—retracted
hooding, pictorial view.
50
-------
TO COMMON HEADER
HOIST DRUM
TO FAN y
SECONDARY DUCT
TO STACK
c-'DAMPER CONTROL
INSTALL LIMIT SWITCH TO
TRAVEL OF TROLLY FOR
CHARGING OR COLLAR PULLING
TO PROTECT HOODING: FOR
OTHER MAINT. WORK, ETC.
REQUIRING THE HOOK TO WORK
IN THIS AREA.THE LIMIT
SWITCH WILL ENERGIZE A GONG
& FLASHING LIGHT TO ALERT
THE CRANEMAN & CONVERTER
OPERATORS THAT THE LOAD ON
THE CRANE OR HOOK CAN
INTERFERE OR DAMAGE THE
HOOD UNLESS THE HOODING
IS RETRACTED.
BIN OR
TAKE AWAY
Figure 17. Peirce-Smith converter—side view,
blister pouring operation, hooding extended.
51
-------
ELECTRIC ARC FURNACE
0 Preventive maintenance
0 Feed transfer points to the furnace tied into fugitive emissions
system
0 A fugitive emission collector at the intersection of the electrode
to the arch roofline, to handle leakage or paste burning
0 Fugitive emission collection system of hooding and envelopes for
the launders and ladles
0 Pivoted cover fitted on slag return opening
DRYER
0 Fugitive emission envelopes at charging and discharging ends to
handle upsets
OUTOKUMPU
0 Fugitive emission collection system of hooding and envelopes for
the launders and ladles
0 Fugitive emission collector at the concentrate injection
0 Preventive maintenance on refractories, etc.
NORANDA
0 Preventive maintenance on refractories, etc.
0 Fugitive emission collector at the concentrate injection
0 Fugitive emission collection system of hooding and envelopes for
the launders and ladles
0 Secondary hooding on roll-out
PEIRCE-SMITH CONVERTERS
0 Rework the main hood for off-gases, with a minimum clearance to the
converter
0 Fugitive emission system hooding, including (where possible) swing-
away, movable, etc. (Figures 15-17)
0 Use of an air curtain control technique for fugitives (Appendix C)
52
-------
HOBOKEN CONVERTERS
0 The addition of a swingaway-hood type fugitive emission collection
system for slagging and blister copper operations would virtually
free this unit of fugitive emissions (Figure 18)
0 Maintenance of proper draft within the converter and good operating
procedures
ANODE FURNACES
0 Addition of a swingaway-type fugitive emissions collection system
MATTE, SLAG, AND BLISTER COPPER TRANSPORT
0 Use of a covered ladle with a stopper rod for bottom pouring to
minimize emissions during transit or the use of vermiculite or
equivalent as a cover for matte to minimize fugitives in lieu of a
fixed cover to the ladle in transport
0 Adaption of a telescopic stiff leg to the electric overhead travel-
ing crane (EOT) for evacuation of fugitive emissions during trans-
port of matte, slag, or blister copper or when charging or pouring
(Figure 19)
TOTAL BUILDING ENCLOSURE
0 Good control on air movements into and from the building
0 Fan capacity with sufficient acfm
0 Building design modification
The passing of time and the resulting need to update or replace facili-
ties to remain competitive make it necessary to plan ahead to minimize operat-
ing costs and meet possible mandated pollution controls in a specific industry
or area.
It is likely that primary copper smelters will be required to meet a
host of regulations limiting S02, arsenic, and trace metals emissions. What
methods are available to control these emissions and still remain cost com-
petitive? The future system for achieving, this goal may be one of the fol-
lowing methods or some combination of them.
CASCADING GRAVITY FLOW
This system (Figure 20a, b, c) would begin with the feed being conveyed
to storage bins and then to either multihearth or fluid-bed roasters or to a
dryer. The partially calcined product is collected and sent to storage bins
or silos and then conveyed as required to the smelter furnace (i.e., a rever-
beratory, electric arc, Noranda, or Outokumpu). Where necessary, the feed is
53
-------
Ul
CAPSTAN
LIFT
DEVICE
HOBOKEN CONVERTER ^^,-
' ~-
TO FUGITIVE
EMISSION
COLLECTOR
Yi
•i
i **
N M
\»,
Figure 18. Hoboken converter with swingaway hood.
-------
fA« Jltt- ID (AH
. AUK HOISI DHUH
llUUllllll
MCI
HU^OPU CAMUW
out i |s inn AN 10
Sim i>ai
HT CIMC)
U1
U1
DUCT
TO FAN
AM) UGHOUSC
(CM U NDMHO
OHHWAS
m HOf TWSSCS)
Figure 19. Ladle/EOT crane fugitive emission collection system.
-------
en
ANOW Will IE',
Figure 20a. Cascading gravity flow.
-------
CONVERTER
8LOG
Figure 20b. Schematic cascading building outline.
-------
Ul
00
SCHEHATIC
C«SC«OI»G
SVSTEd
«"OOE WHEELS
Figure 20c. Fugitive emission collection system for cascading gravity flow.
-------
RELATIVE COSTS, EMISSIONS, AND ENERGY EVALUATION
EXISTING VERSUS ALTERNATE SYSTEMS
Cascading System Versus Conventional Primary Copper Smelter
Costs
Building costs may be greater.
Operational costs may be higher.
Emissions
Minimal
Energy Requirements
Higher for concentrate feed delivery because of height
Minimal for transporting from furnace to furnace
Pros and Cons
Preventive maintenance must be practiced.
Proximity of furnaces to each other
Change in operating techniques
59
-------
prepared to the size required. After smelting, the matte flows by gravity,
via launders, to the holding furnaces (induction-type or supplemental fuel-
fired). From the holding furnaces, it flows by gravity, via launders, to
Hoboken converters and then on to the anode furnace, or to holding furnaces
and then to the anode casting wheels. The conventional transfer of the main
process off-gases would be controlled to electrostatic precipitators or bag-
houses, which exhaust to an S02 recovery system (acid plant, liquid S02,
etc.). The fugitive emissions (Figure 20b) would be ducted to one or more
main collection systems via a baghouse prior to being discharged to the
atmosphere. Collection would be made at all transfer points over launders
and at the discharge to slag pots or ladles or holding furnaces. The vessel
in which slag from the smelter is conveyed to a dump or refuse area would be
capped or covered, prior to leaving the collection station. The Hoboken
converter would have a swingaway hood to collect fugitive emissions during
slagging or pouring. The slag ladle would be a stopper rod-type, fitted with
a cover. During recharging to the smelter furnace, the ladle is positioned
and mated to the furnace so that when the stopper rod is lifted the slag
discharges into the furnace with minimal fugitive emissions.
CASCADING/INDUCTION/GRAVITY FLOW
This system (Figure 21a) would permit the use of conventional building
heights, whereas the cascading gravity flow would require taller buildings.
The roaster or dryers would be at ground level. The partially calcined feed
would be conveyed to storage bins or silos and then to the smelting furnace.
The matte would then be transported to the holding furnaces (Hobokens or
anode furnaces) by means of induction coils that move the molten metal from a
lower elevation to the higher level. The movement of the matte would be
through refractory-lined pipe. Such an arrangement would minimize the need
for fugitive emissions collection ducts over the launders to the holding
furnaces. It would require the use of tundishes or covered troughs to convey
the matte or blister copper to another furnace that is not in the direct line
of discharge from one point to the next. Again, collection of fugitive emis-
sions (Figure 21b) and handling of slag would be similar to that for the
cascading gravity system.
SMELTING/CONVERTER FURNACES OF THE NORANDA OR OUTOKUMPU
With some slight modifications in operating and process techniques, it
may be possible to bypass the holding and converter furnaces of the cascading
system discussed above. Extending the hearth area (Figure 22) of the Outo-
kumpu, using lances and supplemental feeder burners (i.e., pulverized feed
and oil or coal), and adding another off-gas duct (similar to the Hoboken
converter) for draft control might permit a continuous operation for discharg-
ing blister copper all in one vessel. Extending the Noranda center section
(Figure 23), using lances in the area of the matte collection, and slightly
increasing the off-gas hood size might enable delivery of blister copper in a
one-step unit.
60
-------
HOBOKEN
CQHVERURS
STORAGE
SIN
FROM
ROASTER-
Figure 21a. Cascading/induction/gravity flow.
-------
to
Figure 21b. Fugitive emission collection system for cascading/induction/gravity flow.
-------
RELATIVE COSTS, EMISSIONS, AND ENERGY EVALUATION
EXISTING VERSUS ALTERNATE SYSTEMS
Cascading/Induction/Gravity System Versus
Conventional Primary Copper Smelter
Cost
Building costs would be lower.
Operational costs would be higher.
Emissions
Minimal
Energy Requirements
Minimal for gravity transportation from furnace to furnace
High than crane transportation during induction pumping
Pros and Cons
Maintenance would be higher.
Proximity of furnaces to each other
Change in operating techniques
63
-------
a\
CONCENTRATE
OFF-
CONCENTRATE BURNER
UPTAKE •
TAPHOLE
SLAG
TAPHOLE
Figure 22. Modified Outokumpu.
-------
U1
CONCENTRATE
PELLETS AND FLUX
FEEDER
PARTICULATE
AND
SO,
FIXED
HOOD
AIR TUYERE
REDUCING GAS
TUYERE
BURNER
SLAG
COPPER
Figure 23. Modified Noranda.
-------
RELATIVE COSTS, EMISSIONS, AND ENERGY EVALUATION
EXISTING VERSUS ALTERNATE SYSTEMS
Modified Outokumpu or Noranda Versus
Smelting and Converters
Costs
Building costs would be less.
Operational costs would be less.
Emissions
Less
Energy Requirements
Less
Pros and Cons
New operating techniques
Refractory problems
Loss of one unit is loss of total production.
66
-------
RELATIVE COSTS, EMISSIONS, AND ENERGY EVALUATION
EXISTING VERSUS ALTERNATE SYSTEMS
Flash Smelters, Noranda, and Electric Arc Furnaces
Versus Reverberatory Furnaces
Costs
Building costs could be greater.
Operational costs slightly higher.
Emissions
Less
Energy Requirements
More efficient smelting
Pros and Cons
Change in operating techniques
Maintenance could be greater
Better control
67
-------
ENCLOSURES FOR INDIVIDUAL FURNACES
Each individual furnace (Figure 24) would be enclosed and self-contained
with its own captive EOT crane and a transfer car (for ladle additions,
etc.), tuyere puncher, and fugitive emissions system. The equipment would
all be remotely controlled except during maintenance, when it could be effec-
tively controlled manually. If needed, closed circuit televisions would
permit the operator to monitor all operations. As envisioned, the operator
would work from a control booth or room outside of the enclosure. For example,
the operation of a converter might be as follows: a covered ladle of matte
from the smelter furnace (reverberatory, electric arc, etc.) would be sitting
on a transfer car outside of the smelting furnace enclosure, having been
picked up by an EOT crane and positioned on the transfer car outside of the
enclosure of the converter. The operator would raise the sliding door and
move the transfer ladle car inside the converter enclosure and the fugitive
emission evacuation would be energized. The sliding door would then be
closed and the ladle cover removed with a jib boom-type hoist. The ladle
would then be engaged by the hooks on the crane, lifted, and poured into the
converter. Then the ladle would be returned to the transfer car and the
procedure reversed until a sufficient number of ladles of matte, reverts,
etc., are put into the converter. Blowing would then begin. The same operat-
ing methods that are now employed would then be applied. It is true that the
operator would now be working outside the enclosure and might feel he does
not have good control of the operation, but he would shortly become acclimated.
Control of fugitive emissions from a small confined area would be easier than
from the entire building, where open doors, etc., cause disruptions in the
flow of the fugitives to the overhead monitor collection system. The energy
expenditure for air movements also would be lower, whereas building costs
would be a little higher.
CRANE EVACUATION OF FUGITIVE EMISSIONS
As shown in Figure 19 this might be a relatively easier way of capturing
fugitive emissions than the use of an enclosed building evacuation. The
principle entails the use of a spreader beam with a telescopic emission cap-
ture duct built as portion of the beam. The telescopic duct would act in the
same manner as the stiff-leg telescopic soaking pit crane used in the steel
industry. The telescopic duct would be attached to a fixed duct (under the
trolley), which in turn is ducted to another transfer duct attached to one of
the crane girders. This duct would be rectangular, with a split rubber cover
that allows the nozzle from the proceeding duct to discharge into it with a
minimal opening. The transfer duct on the crane girder would in turn dis-
charge to another duct, either on the crane runway, walkway, or overhead to
the building trusses. This duct would discharge at a number of points to a
fixed collection system. The discharge fan could either be fixed on the
trolley, the crane bridge girder, or the building. Another possibility would
be to discharge the fugitives via a duct directly above the trolley and to
use a central monitor at the peak of the building trusses to collect the
fugitive emissions and discharge them to the main collection system.
The telescopic fugitive emission capture system would be in operation
during the picking up of a ladle filled with matte, blister copper, or con-
68
-------
vo
Figure 24. Individual furnance enclosure.
-------
RELATIVE COSTS, EMISSIONS, AND ENERGY EVALUATION
EXISTING VERSUS ALTERNATE SYSTEMS
Enclosing Each Furnace Versus Existing Open Layout
Costs
Building costs may be higher.
Operational costs may be higher.
Emissions
Minimal
Energy Requirements
Slightly greater
Pros and Cons
Preventive maintenance must be practiced.
Air movements or changes must be properly controlled.
Change in operating techniques.
Enclosing Each Furnace Versus Enclosed Building Evacuation
Costs
Building and evacuation equipment costs would be less.
Operational costs would be less.
Emissions
In each case minimal
Energy Requirements
Would be less
Pros and Cons
Control of emissions at source.
Preventive maintenance must be practiced.
Air movements or changes not as critical.
Change in operating techniques.
70
-------
RELATIVE COSTS, EMISSIONS, AND ENERGY EVALUATION
EXISTING VERSUS ALTERNATIVE SYSTEMS
Ladle Emission Control Using Covered
Bottom Pour Ladles and Crane Evacuation
Versus Secondary Hooding. Building Evacuation, etc.
Costs
Building costs should be less.
Operational costs should be less.
Emissions
Less than secondary hooding
Same as building evacuation
Energy Requirements
Less
Pros and Cons
Special crane and duct design
Special ladle design
Higher maintenance
Greater handling of ladles
Change in operational techniques
71
-------
verter slag; the transporting of ladles between process furnaces; the pouring
of the matte, blister copper, or converter slag into the respective vessels;
and also during the time the ladle is being filled. With a good capture
velocity, designed controlled leaks, and good design, this technique would
minimize fugitive emissions during ladle handling.
HOBOKEN CONVERTERS (FIGURE 18) WITH SWINGAWAY HOODS
As shown in Figure 18, the use of Hoboken converters with swingaway
hoods would minimize fugitive emissions over the Peirce-Smith converters
during slagging, pouring, blowing, etc. The Hoboken converter with a properly
designed draft control and gas take-off should be fugitive emission free dur-
ing blowing operations. During slagging and blister copper pouring, however,
the use of a swingaway fugitive emission hood would minimize secondary emis-
sions. Emission problems during matte charging could be handled by a tele-
scopic capture duct attached to the spreader beam of the EOT crane. A covered
ladle in conjunction with a bottom-pour stopper rod also could be used to
minimize fugitive emissions during transporting and discharging the molten
metal or slag to a respective furnace. The ladle cover would be placed on at
the station where the ladle is filled. The cover would be installed under
some sort of emission collection hood, either fixed or movable. The base of
the ladle would be adapted to allow the ladle to be discharged into the re-
ceiving vessel, and if the receiving vessels were operated at a slightly
negative draft, emissions would be minimal. The operation of the stopper rod
or the ladle would be handled by the crane operator via an "auto-pour unit,"
which is used for teeming operations in the steel industry (Figure 25).
ENCLOSED BUILDING EVACUATION TO A CENTRAL DISCHARGE SYSTEM
Enclosed building evacuation, with provisions to provide air changes as
required in the various areas of the building (for the operation and main-
tenance personnel) as well as air for the process combustion and in-leakage,
is another possible way to minimize fugitive emissions. (See Appendix Table
B-2 for a summary on enclosed building evacuation.)
OTHER POSSIBLE EMISSION CONTROL SYSTEMS
0 Floor operated charging machines (Figure 26a, 26b) or manipulators
installed under hooded enclosures in the vicinity of the process
furnaces
0 Q-BOP furnaces (Figure 27) (as used in the steel industry), using a
central injection lance for oxygen-enriched air and feed, installed
in a hooded enclosure with primary and secondary hoods
0 INCO flash smelting furnace (Figure 28) (would be essentially the
same category as the Noranda smelter)
Table 6 shows the author's estimates of energy requirements for the use
of secondary type hoods, individual furnace enclosure, or building evacua-
tion. The numbers indicate energy usage per unit; thus, if two reverberatory
72
-------
RELATIVE COSTS, EMISSIONS, AND ENERGY EVALUATION
EXISTING VERSUS ALTERNATE SYSTEMS
Hoboken Converters Versus Peirce-Smith
Converters (Equivalent Tonnage)
Costs
Building costs would be approximately the same.
Operational costs could be less.
Emissions
Much less
Energy Requirements
Less
Pros and Cons
Maintain refractory
Change in operating technique
Higher operating maintenance
73
-------
Figure 25. Hydraulic cylinder mounted on barrel of ladle rigging
raises and lowers stopper rod to control flow of molten steel from
ladle to ingot mold. Source: "Making, Shaping and Treating of
Steel" by U.S. Steel, Edition 1971. Copyright by U.S. Steel Corp.
1971. Reproduced with permission.
74
-------
Ul
MODIFIED
CHARGING
MACHINE
Figure 26a. Modified charging machine.
-------
Figure 26b . Typical charging machine in the steel industry
Making, Shaping, and Treating of Steel," by U.S. Steel,
Copyright by U.S. Steel Corp. 1971. Reproduced with
permission .
76
-------
RELATIVE COSTS, EMISSIONS, AND ENERGY EVALUATION
EXISTING VERSUS TOTAL BUILDING EVACUATION
Costs
Building costs would be greater.
Operational costs would be greater.
Emissions
Less
Energy Requirements
Greater
Pros and Cons
Air changes and associated problems
Maintenance would be greater.
Operational changes
77
-------
RELATIVE COSTS, EMISSIONS, AND ENERGY EVALUATION
EXISTING VERSUS ALTERNATE SYSTEM
Floor-Operated Chargers With Furnace
Emission Hooding Versus Existing System of
Ladle Handling Transporting, Charging, etc.
Costs
Building costs would be higher.
Operational costs would be higher.
Emissions
Less
Energy Requirements
Greater because of emission hooding
Pros and Cons
New operational techniques
Maintenance higher
78
-------
RELATIVE COSTS, EMISSIONS, AND ENERGY EVALUATION
EXISTING VERSUS ALTERNATE SYSTEMS
Q-BOP Versus Smelters and Converters
Costs
Building costs would be less.
Operational costs should be less.
Emissions
Less
Energy Requirements
Similar
Pros and Cons
New operating techniques
Higher maintenance
Higher tonnages
Good enclosures required
79
-------
SCRUBBER
STACK-
QUENCHER
Gas cleaning system
Figure 27. The 225-ton Q-BOP is enclosed in a doghouse
to prevent fugitive emissions. Source: Iron and Steel
Engineer, Noveinber 1978, reprinted with permission^
80
-------
CHALCOPYRITE
SAND CONCENTRATE
*
CONSTANT
•EIGHT FEEDERS
OXYGEN
OXYGEN
PYRRHOTITE, CHALCOPYRITE
CONCENTRATES, AND SAND
OXYGEN
SLAG MATTE
MATTE
Figure 28. INCO flash smelting furnace.
81
-------
TABLE 6. ENERGY REQUIREMENTS FOR CONTROL OF FUGITIVE (GASEOUS AND PARTICULATE)
EMISSIONS AT INTAKES AND DISCHARGE POINTS
Process equipment
Roaster
Dryer
Outokumpu
Noranda
Electric arc furnace
Reverberatory
Peirce-Smith
Hoboken
Anode
EOT crane
kWh/h (minimum /maximum )
Secondary hoods ;
capture hoods
192/200
156/162
30/149
60/149
89/176
53/149
168/216
17/108
55/108
7/73
Individual"'0
enclosure/unit
27
38
90
50
110
90
11
17
11
-
BuildingD»c
evacuation
271
38
411
172
444
355
365
305
365
-
00
to
a Operation only when emissions are being captured.
Continuous operations.
0 Air changes estimated at 12 per hour.
Note: Amounts shown are per unit and with a baghouse.
-------
furnaces were required, the secondary and individual energy would double,
whereas building evacuation would be 75 to 80 percent greater.
Appendix A shows the costs attributed to secondary hooding systems for
converters. In Appendix B, Table B-l shows the estimated hooding efficien-
cies, and Table B-2 presents a summary of current fugitive emission control
systems.
83
-------
APPENDIX A
COSTS OF SECONDARY HOODING SYSTEMS FOR CONVERTERS
This section includes a capsule discussion on costs of the following major
types of secondary hoods:
0 Fixed type -- made of structural steel with an elliptical cross-
section. It is attached to the primary or uptake hood.
0 Fixed and movable -- consist of a movable intermediate hood and a
hood fastened to the gate. Both hoods are made of structural steel
with elliptical cross-sections so that they telescope in the
retracted mode.
0 Swingaway type with a fixed overhead hood -- made of structural
steel, refractory-lined, and supported by pivot arms with a motor-
ized drive to permit positioning during blowing and pouring opera-
tions.
0 Combination of fixed, movable, and swingaway types.
Costs of Fugitive Emission Control Systems
Cost parameters - secondary hooding --
This cost section includes various items that must be installed or modi-
fied to achieve control of fugitive emissions from the Peirce-Smith converter
(this furnace being the worst source of secondary emissions). It does not
include costs of certain operating procedures that would minimize fugitive
emissions, e.g., allowing minimal clearance between the primary uptake hood
and the apron of the converter, or maintaining proper matte charges to pro-
vide for direct flow of gases from the mouth of the converter to the center-
line of the primary uptake hood.
The following items are considered in this cost section:
0 A fixed hood having an elliptical cross-section of 17 feet 6 inches
on the major axis and 7 feet on the minor axis, and 9 feet 6 inches
long. The plate is 3/8-inch carbon steel, with stiffeners of 7-inch
channels. The fixed hood is bolted to the primary hood and to the
smoke plenum of the secondary duct system*
0 A movable hood in the retracted position, parked above the fixed
hood. It has an independent track system with a five-speed, double-
grooved hoist unit and slack cable limit switch. The movable hood
84
-------
is 9 feet long and is elliptical, with a major axis of 18 feet 6
inches and minor axis of 7 feet 6 inches. These dimensions provide
a 3-1/3-inch clearance between the movable and fixed hoods. There
are mating plates on the lower end of the fixed hood and on the top
of the movable hood. The lower end of the movable hood is fitted
with a 12-inch asbestos-type curtain that follows the elliptical
perimeter to form a seal with the gate hood. The movable hood is
constructed of 3/8-inch carbon steel with stiffeners of 7-inch chan-
nels.
0 The gate hood, elliptical in cross-section, has a major axis of 16
feet 6 inches and a minor axis of 6 feet 6 inches and 9 feet long.
Clearance between the fixed hood and the gate is thus 3-1/2 inches.
The hood will be bolted to the gate. The plate is 3/8-inch carbon
steel reinforced with 7-inch channels.
The dimensions listed above would be modified for each converter layout
to provide the required clearances. Design considerations may dictate that
the fixed hood is the largest unit, with a movable hood under it and the gate
hood under the movable hood.
0 If crane runway rail height is a problem, the smoke plenum of the
secondary hooding duct system could be fitted as follows: the plenum
would span the primary uptake hood and;would have a secondary hood
dust bin affixed on each end. The dust bins would be equipped with
pneumatic dust valves and discharge pipes. No provision is made in
this study for removal of dust in the dust bins. The smoke plenum
for this cost study is 4 feet by 4 feet 8 inches by 21 feet. It is
constructed of 3/8-inch steel with 6-inch channel stiffeners.
0 The secondary hooding duct system would have an uptake from each
dust bin adjacent to the converter and then pass to its main ducting
header for fugitive emissions. The damper valve shown would be
adjacent to the main ducting header and would be closed when the
converter is out of service. Existing facilities would determine
the path of retrofit. The gases go to a dust bin ahead of the fans
and from there to the breeching into the main converter gas duct
and to the stack. For this study, the length of the main duct runs
is 600 feet.
0 The fans considered in this estimate are Buffalo Forge Type 1320 BL,
single-inlet, Arrangement 1, Class 3, with 145 bhp, 785 rpm, 80,000
cubic feet per minute at 200°F. There would be one fan for each
converter in the plant; as many fans as are required would be tied
into the system.
0 Support items for the system include piping, wiring, foundations,
supports for ducting every 20 feet, expansion joints, miscellaneous
platforms, and walkways. Valves, fans, dust bins, and similar items
are flanged for ease of maintenance.
85
-------
The retrofit factor was considered as being midway between a new install-
ation and an existing "difficult" installation.
Estimates of total installed costs are based on current (1978) costs for
major components of specified sizes, as provided by equipment suppliers. Es-
timates of fabrication co.sts and installation in the Southwest are based on
generally accepted engineering practices, i.e., Richardson' s, Mean1 s, the
Chemical Engineering Index, K.M. Guthrie, and data within the PEDCo organiza-
tion.
A 5.0 percent contingency factor to allow for emissions testing, equip-
ment changes, etc., is applied to the total of the direct and indirect costs.
An escalation factor of 7-1/2 percent per year is used to account for in-
creases in cost of equipment, labor, and services before and during construc-
tion. Direct capital costs include equipment, instrumentation, piping,
electrical, structural, foundations, site work, insulation, and painting.
Indirect capital costs include the following: engineering costs, contractor's
fee and expenses, interest (accrued during construction on borrowed capital --
estimated at 9 percent per year), freight, off-site expenditures, taxes
(sales tax of 4 percent of equipment cost), startup or shakedown, and spares.
Annualized costs include both operating costs and fixed capital charges
and consist of: utilities; labor and fringe benefits; maintenance; plant
overhead; and total fixed costs (amounting to 19,97 percent of total installed
costs and consisting of depreciation over 15 years at 6.67 percent unless
otherwise indicated, property insurance at 0.3 percent, property taxes at 4
percent, and interest on borrowed capital at 9 percent).
Capital costs --
Table A-l shows the direct costs, indirect costs including one year of
contingency and escalation, and total capital costs for plants containing one
to nine converters without a baghouse in the fugitive emission discharge sys-
tem. Table A-2 includes the costs in Table A-l plus those for a baghouse and
an appropriate increase in the fan pressure design.
Operating costs --
Operating costs include the following: operating labor at $8 per man-
hour, supervision at 15 percent of labor, maintenance for labor and supplies
at 2 percent of total capital costs, maintenance materials at 15 percent of
maintenance labor and supplies, electricity at 35 mills per kWh, plant over-
head at 50 percent of operations, and payroll at 20 percent of the operating
labor costs. The fixed costs include a straight-line 15 years' depreciation,
0.3 percent for insurance, 4 percent for taxes, and 9 percent for capital
costs. Table A-3 lists the operating costs for a multiconverter plant without
a baghouse in the fugitive emission discharge system. Table A-4 includes the
costs in Table A-3 plus those for a baghouse and appropriate increases in
energy and maintenance costs.
Additional handling of slag and blister ladles may cause delays in opera-
tion of the movable and swingaway converter hoods. It is estimated that a
86
-------
delay of 5'to 15 seconds may occur with each ladle movement, equivalent to a
total delay of 3 to 10 minutes per day or a 0.23 to 0.7 percent production
slowdown. This loss is calculated on an annual basis as part of the operating
cost since it is negligible in comparison with other delays that are encoun-
tered, such as delay because of lack of matte, because the anode furnace can-
not accept more blister copper, or because of maintenance of converters or
reverberatory or electric arc furnace, etc.
Each of these systems is connected to a main discharge duct and an ex-
hauster-type fan, which exhausts to an existing stack.
87
-------
TABLE A-l. ESTIMATED CAPITAL COSTS OF SECONDARY HOODING FOR
MULTICONVERTER PLANT WITHOUT BAGHOUSE
(dollars)
No. of
converters
1
2
3
4
5
6
7
8
9
Direct
costs
760,000
1,216,000
1,532,000
1,771,000
2,211,000
3,219,000
3,601,000
3,880,000
4,402,000
Indirect
costs
532,000
785,000
963,000
1,255,000
1,463,000
2,122,000
2,383,000
2,545,000
2,850,000
Total
costs
1,292,000
2,001,000
2,495,000
3,026,000
3,674,000
5,341,000
5,984,000
6,425,000
7,252,000
88
-------
TABLE A-2. ESTIMATED CAPITAL COSTS OF SECONDARY
HOODING FOR MULTICONVERTER PLANT WITH BAGHOUSE
(dollars)
No. of
Converters
1
2
3
4
5
6
7
8
9
Direct
Costs
1,122,000
1,616,000
2,127,000
2,587,000
3,103,000
4,850,000
5,523,000
6,093,000
6,877,000
Indirect
Costs
736,000
1,007,000
1,298,000
1,714,000
1,966,000
2,936,000
3,466,000
3,792,000
4,244,000
Total
Costs
1,858,000
2,623,000
3,425,000
4,301,000
5,069,000
7,786,000
8,989,000
9,885,000
11,121,000
89
-------
TABLE A-3. ESTIMATED ANNUAL OPERATING COSTS FOR SECONDARY HOODING
IN A MULTICONVERTER PLANT WITHOUT BAGHOUSE
(dollars)
No. of
converters
1
2
3
4
5
6
7
8
9
Labor and
supervision
10,000
21,000
31,000
41,000
51,000
62,000
72,000
82,000
93,000
Maintenance
labor, supplies,
and materials
30,000
46,000
57,000
70,000
85,000
124,000
138,000
148,000
167,000
Overhead
plant and
payroll
22,000
38,000
50,000
64,000
78,000
105,000
119,000
131,000
149,000
Utilities
1,000
3,000
4,000
5,000
6,000
8,000
9,000
10,000
11,000
Fixed
costs
258,000
400,000
498,000
604,000
734,000
1,077,000
1,195,000
1,283,000
1,448,000
Total
annual
costs
321,000
508,000
640,000
784,000
954,000
1,376,000
1,533,000
1,654,000
1,868,000
10
o
-------
TABLE A-4.
ESTIMATED ANNUAL OPERATING COSTS FOR SECONDARY HOODING
IN A MULTICONVERTER PLANT WITH BAGHOUSE
(dollars)
NO. Of
Converters
1
2
3
4
5
6
7
8
9
Labor and
Supervision
17,000
28,000
38,000
48,000
58,000
72,000
82,000
92,000
103,000
Maintenance
Labor , Supplies ,
and Materials
43,000
60,000
79,000
99,000
117,000
179,000
207,000
227,000
256,000
Overhead
Plant and
Payroll
33,000
50,000
66,000
83,000
99,000
140,000
161,000
178,000
252,000
Utilities
80,000
122,000
226,000
343,000
404,000
523,000
878,000
1,039,000
1,209,000
Fixed
Costs
371,000
524,000
684,000
859,000
1,012,000
1,555,000
1,795,000
1,974,000
2,221,000
Total
Annual
Costs
544,000
784,000
1,093,000
1,432,000
1,690,000
2,469,000
3,123,000
3,510,000
4,041,000
-------
APPENDIX B
TABLE B-l. PEDCo'S ESTIMATED HOODING EFFICIENCIES*
(values in percent)
Hood type
Fixed
Fixed and movable
Fixed and swing -away
Fixed, movable, and swing-
away
Enclosed building
Matte
or
hot metal
addition
30-50
30-50
30-50
30-50
90b-100
Blister
or
hot metal
pouring
30-50
40-70
80-90
80-90
95b-100
Skimming
or
rabbling
30-50
40-70
50-70C
60-80C
95b-100
Blowing
60-70
70-90
80-90
80-90
95b-100
Most system efficiencies would be higher if air motions (i.e., open
doors, man-cooling fans, monitors, etc.) could be eliminated.
Skimming is removal of slag from the converter by tilting of the
converter. Rabbling is removal of slag from the converter by tilting
of the converter and manual use of a rake to work the molten bath.
Reduced efficiency due to air motions; if doors are left open this
efficiency could drop to 50%.
c Efficiency during skimming would be similar to blister pouring.
92
-------
TABLE B-2. SUMMARY OF CURRENT FUGITIVE EMISSION CONTROL SYSTEMS
Type
Design and operation
Operational and maintenance problems
Efficiency
Monitor, Natural -
U.S.
Monitor, powered
vo
U)
Fixed hood with
secondary emission
ducting - U.S.
Enclosed converter
hood,swing-away type
with fixed hood
Enclosed building
Simple design. Relies on outside
air movement for removal of emis-
sions from the area.
Simple design. Large air movement
required at the fans. Removal
rate is constant.
Clearance problems for crane hood
and cables during collar pulling or
matte additions. Retrofit diffi-
culties for ducting, fans, breech-
ing, and dust bins. Operational at
all times that converters are on
line. Good face and capture
velocities required.
Clearance problems for crane hood
and cables during collar pulling or
matte additions (fixed hood).
Clearance problems for floor space
relationship to the fixed hood;
rugged drive mechanism needed for
swing-away converter hood.
Requires careful consideration of
all openings (personnel, truck,
rail, materials) to minimize air
motion. Roof monitor must handle
all ventilation air for workers,
emissions and in-leakages. In-
creases building costs because of
wind load design, tightness, and
close-fitting openings.
Haze in building during emissions; outside air movement
affects time required to clear the area. Crane oper-
ator and maintenance personnel working above the con-
verter may be required to wear face aspirators. Visi-
ble emissions in the monitor area. Maintenance in the
converter area, electric overhead traveling crane and
roof trusses for removal of the settled emissions other
than gases.
Blind pockets or short-circuited flows could cause haze
and emission buildup in the roof line area. Crane
operator may require the use of a face aspirator at
times. Maintenance of fans and drives. Visible emis-
sions at each powered monitor. Maintenance in the
converter area, electric overhead traveling crane and
roof trusses for removal of settled emissions other
than gases.
Operational damage to hood by swinging or uncontrolled
electric overhead traveling crane action during matte
addition or collar pulling. Maintenance is less in the
converter area, electric overhead traveling crane and
roof trusses due to particle buildup.
Space occupied in the aisle by the swing-away converter
hood when adding matte, rabbling, or skimming could
hamper crane movements. Crane must deposit the ladles
for pouring or skimming and at the completion of the
operation must await retraction of the hood before en-
gaging the ladle. Maintenance of the swing-away
mechanism and minimal maintenance for removal of par-
ticulate buildup that occurs during matte additions
and rabbling.
All openings must be maintained constantly against ex-
cessive air infiltration. Tight siding and roofing
required. Air circulation within the building for
the workers and process must be carefully controlled.
Intake and exhaust fans require preventive maintenance.
Cleanup maintenance for settled particulates in the
converter area is similar to that for a monitored
system, either natural or powered.
Dependent on outside air
currents and inside air
motion.
Dependent on number of
monitors, fan size, build-
ing design above the con-
verter proper; air motions.
Dependent on the distance
of the mouth of the fixed
hood from the emission
source; also on the capture
and face velocity created
by the fan at the mouth of
the fixed hood.
Dependent on operational
cycle. When pouring, blow-
ing or slagging the emis-
sions control woudl be good;
when adding matte, or rab-
bling, it would be similar
to the fixed hood. Air mo-
tions influence efficiency
in all operations.
Dependent on building
tightness, air motion con-
trol, monitor exhaust capa-
bilities.
-------
APPENDIX C
AIR CURTAIN CONTROL TECHNIQUE FOR FUGITIVES
A method somewhat similar to secondary hooding for controlling fugitive
emissions at a Peirce-Smith converter would be the use of an air curtain
technique as used by Mitsubishi in Japan. This technique could also be used
on other smelter operations, i.e., roaster, reverberatory furnace, etc.
This consists of steel side panels with a back and top panel, forming a
partial enclosure. The front or approach side by the crane with a ladle to
the converter is open. The top panel has an opening of 1.4 meters to permit
the overhead crane's cables to pass into the enclosure when charging matte or
removing converter slag or blister copper ladles.
Air is blown from one side of the top panel opening and collected on the
other side of the opening. The air flow rate across the opening is 1000
NM^/min. This flow'plus the rising fugitives are captured and discharged to
a fugitive collection system.
We developed costs using this air curtain fugitive control technique in
a manner somewhat similar to that shown in Appendix A for secondary hooding
systems.
Our design differs from the Mitsubishi in that the side panels were
extended to contain the location area of the ladle awaiting blister copper or
slag from the converter. Also partial covering was placed on the approach
side (Figure 29). 3
The air flow across the top was set at 1000 NM /min but the collection
side was designed with its own suction fan to handle 1,250 NM^/min to create
a flow pattern to capture the fugitives during charging, teeming, slagging,
etc.
Tables C-l and C-2 show the total installed capital and annual operating
costs.
94
-------
TABLE C-l. ESTIMATED CAPITAL INSTALLED COSTS FOR
AIR CURTAIN TYPE HOODING WITH BAGHOUSES
FOR MULTICONVERTER PLANT
No. of
converters
1
2
3
4
5
6
7
8
9
Direct
cost
$ 372,200
736,900
987,800
1 ,365,500
1 ,621 ,200
2,111 ,300
2,688,700
3,242,000
3,591 ,200
Indi rect
costs
$ 586,700
926,700
1 ,222,200
1 ,708,800
2,010,200
2,725,500
3,406,300
4,189,800
4,630,300
Total
costs
$1,082,000
1 ,878,000
2,495,000
3,470,000
4,099,000
5,460,000
6,880,000
8,389,000
9,280,000
a Includes a 5% contingency and 7-1/2% escalation for one
year.
95
-------
TABLE C-2. ESTIMATED ANNUAL OPERATING COSTS FOR AIR CURTAIN TYPE HOODING
WITH BAGHOUSES FOR MULTICONVERTER PLANT
No.
of
converters
1
2
3
4
5
6
7
8
9
Labor &
supervision
17,000
28,000
38,000
48,000
58,000
72,000
82,000
92,000
103,000
Maintenance,
labor,
supplies
& mat1 Is
25,000
43,000
57,000
80,000
94,000
126,000
158,000
193,000
213,000
Overhead
plant &
payrol 1
20,000
29,000
36,000
47,000
54,000
76,000
92,000
110,000
120,000
Utilities
59,000
67,000
80,000
145,000
168,000
225,000
240,000
297,000
349,000
Fixed
costs
220,000
382,000
506,000
704,000
832,000
1,109,000
1,398,000
1,704,000
1,886,000
Total
Annual
costs
341 ,000
549,000
717,000
1,024,000
1,206,000
1,608,000
1,970,000
2,396,000
2,671,000
-------
MAIN HOOD
AIR CURTAIN
CONVERTER
FRONT VIEW
PICKUP
AIR AND
AIR FLOW FUGITIVES OUT
PATTERN
Figure 29. Air-curtain fugitive control system.
97
-------
DEMONSTRATION OF THE BERGS0E AGGLOMERATION FURNACE AND
BEST MANAGEMENT PRACTICES AT A SECONDARY LEAD SMELTER
Richard T. Coleman Jr.
Senior Chemical Engineer
Radian Corporation
Austin, Texas
Alfred B. Craig Jr.
Physical Scientist
EPA, Cincinnati
Robert Vandervort
Senior Industrial Hygienist
Hadian Corporation
Salt Lake City, Utah
ABSTRACT
The United States Environmental Protection Agency has recently promul-
gated a new National Ambient Air Quality Standard for lead at 1.5 micrograms
of lead per cubic meter of air (quarterly average as measured at the fence
line). This standard, coupled with the OSHA occupational safety standard of
50 micrograms lead per cubic meter of air is anticipated to have a signifi-
cant impact on the primary and secondary lead industry in the United States.
This paper provides an overview of EPA's control technology study for
the secondary lead industry. Preliminary findings concerning the
exemplary practices utilized at a secondary lead smelter in Denmark are re-
viewed and the application of similar techniques to a U.S. smelter is
discussed. Indications are that major contributors to the violation of
this standard are fugitive sources such as piles of baghouse flue dust, raw
materials storage areas, and general inattention to plant housekeeping.
98
-------
INTRODUCTION
This paper describes a study being conducted by the Industrial Environ-
mental Research Laboratory (IERL-CI) of the U.S. Environmental Protection
Agency in Cincinnati. The study is a four part demonstration of fugitive
emission controls and what EPA terms "best management practices" (BMP's).
The objectives of the program are first, to determine what improvement can
be made in lead-in-air levels both inside and ouside a smelter by retrofitting
fugitive emission and instituting work practi.ce controls (BMP's) and second,
to determine the cost.
The first part of the study was conducted at the Paul Bergs^e and Son
secondary lead smelter in Glostrup, Denmark. As part of an inter-agency
effort, the Control Technology Assessment Branch of NIOSH and the Metals and
Inorganic Chemicals Branch of EPA performed this characterization of the new
secondary lead smelting technology being used at the Bergs^e smelter. The
processes characterized offer potential solutions to major occupational and
environmental problems associated with secondary crude lead production. Two
major processes were studied: special battery (SB) smelting and flash ag-
glomeration of the flue dust. The remainder of the program, which is also an
inter-agency effort, involves: first, measuring the present lead-in-air
levels at a U.S. smelter; second, retrofitting the Bergs^e dust agglomeration
furnace to an existing U.S. smelter, instituting BMP's, and cleaning the
smelter; and finally, measuring the reduction in lead-in-air levels both in-
side and outside the smelter.
This paper contains only a portion of the results obtained in the
characterization study of both the Bergs^e smelter and the initial charac-
terization of the U.S. smelter. The complete results of the Bergs«£e study
are available in a draft report entitled "Source Characterization of the SB
Smelting Furnace." Copies can be obtained from the Cincinnati offices of
EPA's Metals and Inorganic Chemicals Branch and NIOSH's Chemical Agents
Control Section.
PROJECT OVERVIEW
A special report prepared in 1976 for IERL-CI identified secondary lead
as an industrial segment of the secondary nonferrous metals industry posing
a serious environmental threat based on the data gathered for two environ-
mental assessment reports. As part of a larger study of the nonferrous
industry, IERL-CI commissioned an evaluation of new technologies and control
options for reducing process and fugitive emissions of lead. The primary
goal of this undertaking was to identify new approaches to metal production
which had been developed with environmental considerations given priority as
key design parameters. In addition, control options readily adaptable to
the industry were to be similarly sought. It was intended that, following
these preliminary fact-finding studies, more detailed investigation of each
important process could be undertaken.
99
-------
The study of the emerging and advanced processes for secondary lead
was to find solutions to the environmental and occupational problems of:
• battery breaking and consequent battery acid drainage and
spillage,
• stack emissions of lead, arsenical, and antimonial particulate
and fume,
• fugitive emissions such as unducted process gases and dust
sources such as air pollution control residues, and raw and
waste material storage piles, and
• employee exposure to lead in the workplace.
The most attractive potential solutions to the major environmental
problems associated with the secondary lead industry were analyzed to
identify the best available control technology. Two major processes were
identified which presented opportunities for significant environmental
improvement: whole battery smelting and flash agglomeration of flue dust.
These processes represented the most advanced technology for controlling
emissions from secondary crude lead production. Whole battery smelting
offered solutions to both environmental and occupational health problems. The
SB (special battery) smelting, system at the Bergs^e smelter in Denmark included
both environmental and occupational considerations as key design parameters.
Plant layout, raw material storage and handling, process and hygiene ventila-
tion, housekeeping, process control, and flue dust agglomeration were each
included in the fundamental design. By including these design parameters,
it appeared to be possible to produce crude lead without the usual environ-
mental and occupational health problems.
Flash agglomeration of flue dust appeared to be an effective method for
minimizing fugitive dust problems, improving baghouse performance, and in-
creasing smelting capacity. The major advantages that flue dust agglomera-
tion offered were:
• an 80 percent reduction in the specific volume of baghouse dust,
• a virtually dustless product in a form which is ideal for
charging to a blast furnace,
• reduced dust loads in the blast furnace off-gas thus improving
efficiencies and reducing operating costs of existing baghouses,
• increased blast furnace charge rates,
• elimination of fugitive emissions from piles ox stored flue
dust.
100
-------
Although these claims appeared to be reasonable, they needed to be
.substantiated, especially at a U.S. secondary lead smelter. Actual reductions
in stack and fugitive emissions which could be realized by installing an ag-
glomeration furnace have not been quantified. There are other factors such
as dust melting point and smelter capacity which could also limit the appli-
cability of the process. To answer these questions, both source characteri-
zation and technology demonstration programs were suggested for both new
and old smelters to quantify the reduction in fugitive emissions which could
be achieved by using flue dust agglomeration.
The Control Technology Branch of the National Institute of Occupational
Safety and Health commissioned a similar study to document and evaluate the
most effective control technology available to minimize worker exposure in
the secondary nonferrous smelting industry. This evaluation was accomplished
by review of existing information from government and other sources as well
as a series of preliminary and in-depth plant surveys.
The major result of this study was that successful control of employee
exposure was almost always dependent upon an integrated control program which
required the application of several different control elements, none of
which worked effectively without the others. In most of the smelters sur-
veyed, only a portion of a completely integrated control program was in use.
Thus, effectiveness of controls in use in one area of the smelter was
masked by contamination from other uncontrolled or poorly controlled sources.
All of the exposure control elements identified were applications of basic
industrial hygiene exposure control principles.
The EPA initiated an interagency agreement with NIOSH to measure fugi-
tive atmospheric emissions and workplace employee exposure levels from
secondary nonferrous smelting. These terms, synonymous expressions that the
two agencies use, describe airborne pollutants that are discharged at low
levels during raw material charging, slag tapping and metal pouring and are
essentially uncollected by process controls. The purpose of this program was
to jointly measure the levels of lead-in-air at the Bergs^e smelter and at a
U.S. smelter and then evaluate the effectiveness of a combination of controls
in reducing lead-in-air levels at the U.S. smelter. The controls of particu-
lar interest in this case were the agglomeration furnace, BMP's, and an im-
proved ventilation system. The joint program resulted in considerable cost
savings to the government in that a sampling program satisfying both R&D pro-
grams was fashioned at less cost than separate sampling by each agency.
RESULTS OF THE SB FURNACE CHARACTERIZATION
SB smelting is an important process because it eliminates the battery
decasing step when whole batteries comprise a portion of the furnace charge.
During the test period, whole batteries comprised 40 percent of the lead-
bearing charge. In addition, since flue dust agglomeration is an integral
part of the SB smelter, a major part of the smelter fugitive dust emissions
is eliminated. It is important to note, however, that SB smelting alone is
101
-------
applicable only for the production of crude lead. Additional refining
processes are required to produce soft lead, lead oxide, or lead alloys.
The SB smelter plant layout, raw material storage and handling facili-
ties, process and hygiene ventilation, and flue dust agglomeration system all
include both environmental emission control and occupational health considera-
tions in their design. Also, the SB furnace is controlled at conditions
producing minimum emissions. This approach has been implemented by Paul
Bergs«le and Son A/S of Glostrup, Denmark.
SB Smelting Furnace
The SB smelting shaft furnace has an oblong cross section unlike most
cylindrical secondary lead blast furnaces used in the United States. The
construction is similar to a primary lead blast furnace. Figure 1 is a
diagram of the furnace and associated gas treatment system. The furnace is
constructed so as to isolate the charging floor from the bottom of the
furnace. Thus, only the front-end loader operator works in a "dirty" area.
However, the front-end loader does have a filtered air supply. In addition,
the top of the blast furnace is provided with local exhaust ventilation to
minimize fugitive emissions.
At the bottom of the furnace, local exhaust ventilation is provided for
the four slag taps and the lead ladles. There are two rows of tuyeres, one
on either side of the furnace, designed to use air preheated to 500°C. The
tuyeres have special covers which minimize fugitive emissions during punch-
ing.
The local exhaust ventilation air and process flue gases are mixed and
all gases pass to four baghouses (at 100 to 125°C). The baghouses are a
Swedish design using felted polyester cloth. The dust is collected on the
outside of the bag and only a mild cleaning air stream is required to dislodge
the dust. This reportedly gives the bags an exceptionally long life. The
baghouses were designed to operate with three running and one spare. During
the test period, all four were running. The air-to-cloth ratio with four bags
operating is 34.5 m3/hr/m2 (1.88 ft3/min/ft2) . The effective c.loth area is
4720 m2 for four baghouses. Typical pressure drop across the baghouse is
150 mm (^6 inches) water gage.
The collected dust is conveyed in an enclosed screw conveyor system to
one of two small flash agglomeration furnaces. In this patented process, the
dust is melted and reduced in volume by about 80 percent. This reduction in
volume increases production per square foot of furnace cross-section by re-
ducing the dust recirculation. The agglomerated dust represents only 2 or
3 percent by weight of the furnace charge. Agglomerating the dust also reduces
the dust load circulating in the gas cleaning system. The flash agglomerator
furnace is oil fired and consumes approximately 7.7 liters of oil per hour.
The smelter area is paved and is wetted and swept periodically. This
practice minimizes fugitive emissions normally caused by the wind blowing dry
102
-------
O
U)
STIICAM
Ha.
LfAO kflL VCNT
to
ii
STREAM HAM
SLA6 TAP
S/W/rMT VfNT
sec. SIA&
TAf Y£WT
fttOCfSS GAJ
Off GAS
AOdtOMfXATfOH
mewces o*s
©+©+©
CQWUHCD
EXIT GAS
fUOUCR VENT
-LfAO W£U
Figure 1. Whole battery smelter furnace and flue gas treatment system.
-------
lead dust in the yard. This rinse water is collected and combined with the
acid drainage from cracked batteries stored in the yard and is finally
treated using a soda ash preqipitation process. The treated effluent is
discharged to the municipal wastewater treatment facility. The sludge is
withdrawn approximately once every two weeks and is charged back to the
furnace for additional recovery of metals.
This smelter is serviced by two additional sewer collection systems,
one for rainfall and one for sanitary sewage. The rainfall collected is
used as washdown water for the smelter yard. Additional makeup cooling
water is obtained from the municipal water supply. This water is softened
in an ion exchange unit before use. Sanitary sewage is discharged directly
to the municipal collection system.
Furnace Operation—
The operation of the SB furnace results in low stack and fugitive lead
emissions rates. The relatively low blast air rate (^3500 Nm3 air/hr) and
large furnace cross-sectional area (4.0 m2 at the tuyeres) results in a low
gas velocity. This combined with the low furnace top temperatures and the
absence of loose flue dust in the charge result in a low lead dust generation
rate. The large furnace cross-section and small production rate allows the
charge material to descend slowly through the furnace shaft. Thus the charge
heats slowly and is not hot enough at the top of the furnace to generate lead
fume.
This slow heating also prevents the rubber and polypropylene case
material from "burning thru." Burn-thru could occur when the charge material
is ignited throughout the furnace shaft rather than only in the smelting zone.
This occurs more readily in the SB furnace than in conventional blast furnace
because the rubber or polypropylene case material is present. It is there-
fore very important to control the furnace temperature both from an operating
and an environmental viewpoint.
Furnace Charging—
In order to help maintain constant temperature in the furnace, the feed
is carefully bedded on the chargeroom floor. A layer of coke is spread on the
floor first, followed by recycle slag, batteries, plates, scrap iron and other
feed materials. By doing this, each bucket of material charged to the furnace
contains roughly the right amount of coke. This practice maintains a homo-
geneous mixture of material in the furnace and helps avoid hot spots.
Each furnace charge during the test period contained roughly the same
ratio of the materials shown in Table 1. A front end loader is used to spread
the feed materials on the chargeroom floor. A large floor scale is used to
weigh the front end loader with a full bucket. The measurement is fed to the
mini-computer which monitors the smelter operations. The weight of the empty
front end loader is subtracted and the charge weight is recorded. A very
accurate measurement of the charge blend is possible using this technique,
104
-------
typically within 1 to 2 percent of the target on major components (batteries,
plates, etc.) and within 4 to 5 percent on the minor flows.
TABLE 1. TYPICAL FEED MAKEUP FOR SB SMELTING FURNACE CHARGE
Approximate Percent of Lead
Feed Material Weight Percent Bearing Charge
Whole Batteries
-polypropylene case
-hard rubber case
Battery Plates
Agglomerated Dust
(or battery mud
Drosses
Return Slag
Coke
Scrap Iron
FeO (mill scale)
CaC03
Total
12.6
12.6
31.5
3.2
3.2
22.1
5.7
1.9
6.3
0.9
100.0
20.0
20.0
50.0
5.0
5.0
100.0
Operating Paramters—
Scrap iron and mill scale (FeO) are added to the furnace so that a PbS-
FeS matte forms between the lead bullion and slag in the furnace. Most of
the sulfur input is trapped in this matte. At present, the matte is being
stored on-site because of its high lead content Cv>8%) . It must either be sold
to a primary lead smelter or disposed of in a protected landfill. Recovery
of the matte lead content at a secondary lead smelter is not possible because
of the sulfur dioxide emissions which would result. A primary lead smelter,
however, could combine this material with the normal lead sulfide materials
they process.
Oxygen enriched air is used in the SB furnace. Blast air is preheated
to 500°C then mixed with oxygen prior to entering the furnace. Preheating the
air reduces the amount of coke required in the furnace and allows smaller
blast air rates to be used. As mentioned earlier, this helps reduce dust
generation in the furnace. Blast air pressure at the tuyeres ranges from^OO
to 1200 mm H20. The blast air rate is typically between 3200 and 3700 Nm /hr
with between 60 and 115 Nm3/hr of oxygen added. These rates correspond to
a production rate of between 62 and 74 metric tons Pb per day.
105
-------
Flash Agglomeration Furnaces
The flue dust generated by automobile battery and battery manufacture
scrap smelting melts at approximately 400 to 900°C (750 to 1650°F). This is
the property which makes flash agglomeration of flue dust possible. Dusts
with higher melting points cannot be agglomerated using this technique with-
out causing the low melting point materials to volatize. A special furnace
was designed to take advantage of this property so that dust handling could
be completely avoided.
At most secondary lead smelters, it is common practice to return flue
dust directly to either the blast furnace or a reverberatory furnace. A
considerable amount of this dust is also entrained in the furnace flue gas
system. Agglomerating the flue dust prevents entrainment, thus reducing the
load on the baghouse and improving its performance.
The Bergs«5e smelter has two agglomeration furnaces serving the four SB
furnace baghouses. Figure 2 is a diagram of a flash agglomeration furnace.
The agglomeration furnace is fed directly from the baghouse dust hoppers via
screw conveyor. The dust drops onto the furnace hearth where it melts al-
most instantaneously upon contact with an impinging flame. The liquid runs
down the sloping hearth, through a permanently open taphole and into a cast
iron vessel where it solidifies.
The transfer of flue dust is usually a dusty operation and is a source
of fugitive emissions and high exposure to lead-in-air. The agglomeration
furnaces completely eliminate handling of the dust, the associated occupa-
tional hazard, and fugitive emissions from flue dust storage piles provided
that the agglomerated dust is stored indoors.
Tipping the contents of the cast iron vessels onto the floor is usually
sufficient to break the material into lumps suitable for recharging to the
blast furnace. It is simply mixed with coke and flux and loaded into the
top of the blast furnace along with other charge materials.
Because less flue dust is generated, and because the volume being re-
cycled is reduced by about 80 percent, additional material can be charged to
the furnace, thus increasing the smelting rate slightly. This is one eco-
nomic justification for the agglomeration furnace.
Plant Layout
A plot plan of the new Bergs^e smelter is shown in Figure 3. The old
lead smelter, the lead refinery, the small cast iron department, the copper
department, and the other small smelting operations are located between 100
and 400 meters southwest of the new smelter. The old smelter was built in
the 1930's. Construction on the new smelter began in 1973 and the initial
startup occurred in 1975.
106
-------
&UR.NER.
VENT
BAGHOUSC-
OUST
MOLTEN DUST
-COOLING /TRANSPQKTA TtON
CXUC/BLE
Figure 2. Flash agglomeration furnace.
0225281
1Q7
-------
M
O
00
XHX6-4
Figure 3. Plot plan.
-------
The SB furnace is located in the building labelled #1 on Figure 3. As
mentioned earlier, the furnace building isolates the top of the furnace where
the charging is done from the work area on the first floor. A large hood
over the charge area captures emissions escaping the furnace top. As shown
in Figure 3, the SB furnace and the two short rotary furnaces located in
Building #3 all discharge to a common stack. The rotary furnaces were not
operating during the SB furnace test period.
Raw materials are stored in concrete bins in Building #2 in Figure 3.
Building #2 also contains the charge bedding area. Building #2 is not large
enough to contain all of the raw materials because of the irregular receipt
of scrap material. As a result, several large piles of plates, unbroken
batteries, and clean lead scrap are stored to the west of Buildings #1 and #2.
The small sodium carbonate water treatment plant is located between the
old and new smelters to the south of Building #1. As can be seen in Figure 3,
the smelter fence is 200 meters east-southeast from the SB furnace build-
ing. This is the closest point from smelter to fenceline. In addition, the
prevailing winds are from either the west or west-northwest. However, no
ambient measurements were made at the fenceline because of the close proximity
of the rotary department (Building #3), the old smelter, and the motorway
which lies approximately 200 meters from the fenceline.
Summary of Results
The purpose of the testing performed for this program was to characterize
the operation of and emissions from the combined SB smelting and flue dust
agglomeration furnaces. The EPA characterization included: 1) an analysis
of the smelter feed materials, 2) a determination of the flows of lead, anti-
mony, arsenic, chlorine, and sulfur, 3) measurement of stack emissions,
4) calculation of an emission factor, 5) measurement of lead-in-air levels
in the smelter yard, and 6) a description of the furnace operation. The
NIOSH characterization included 1) workplace and personal monitoring,
2) evaluation of ventilation and other control systems, and 3) observation
and evaluation of work practices and personal protective equipment.
The major results are:
• The smelter feed consisted of the following materials:
Whole Batteries
-polypropylene case 12.6
-hard rubber case 12.6
Battery Plates 31.5
Agglomerated Dust 3.2
Drosses 3.2
Return Slag 22.1
Coke 5.7
Scrap Iron 1-9
Mill Scale (FeO) 6.3
CaCOa 0.9
Total 100.0
109
-------
The total input materials contained an average of 79 metric tons
lead, 1.5 metric tons antimony, 29 kilograms arsenic, 0.1 metric
tons chlorine, and 2.6 metric tons sulfur per day.
• During the test period, these five elements were distributed in
the smelter output materials as shown in Table 2.
Table 2. APPROXIMATE ELEMENTAL DISTRIBUTION IN SMELTER EXIT STREAMS
Percent of total incoming flow
Output
stream
Lead Bullion
Lead stone
Stack gas
Total
(Estimated
(matte)
accuracy of
Pb
87.7
0.90
0.0025
89
elemental
Sb
98.5
11.2
0.84
110
flow rates
As
10.4
75.7
0.07
86
is ±20%.)
Cl
12.8
36.8
75.8
125
S
0.14
98.0
7.2
105
The chlorine content of the agglomerated flue dust averaged
25 percent by weight. This indicates that chlorine accumu-
lates in the flue dust collection system.
The lead particulate emission rates based on the EPA Method 5
test results were:
-September 26, 1978 0.056 kg Pb/hr
-September 27, 1978 0.074 kg Pb/hr
The stack lead particulate emission factors based on the EPA
Method 5 test results were:
-September 26, 1978 0.019 kg Pb/metric ton Pb product
-September 27, 1978 0.016 kg Pb/metric ton Pb product
Total stack lead emissions, based on the Radian wet electrostatic
precipitator (WEP) experiments, ranged between 0.042 and 0.12 kg
Pb/hr.
Stack chlorine emissions ranged from 1.6 to 7.1 kg/hr.
Stack gas concentrations ranged from 39 to 54 ppm sulfur
during the test. This corresponds to sulfur emission rates
of 6.7 and 9.1 kg S/hr.
110
-------
Total stack antimony emissions ranged between 0.52 and 0 54
kg Sb/hr.
Total stack arsenic emissions ranged between 0.0005 and
0.0013 kg As/hr.
Only one high volume area sampler was run on each of the
two test days. The hi-vol was placed underneath the bag-
house nearest the east fenceline (see Figure 3). It was
approximately four meters from the east agglomeration furnace
and 80 meters from the east f enceline. The two twenty-four
hours average lead-in-air levels were:
-September 26, 1978 12 yg Pb/m3
-September 27, 1978 18 yg Pb/m3
Stack emissions based on the WEP tests are presented in
Table 3:
Table 3. STACK EMISSIONS DETERMINED USING WEP TRAIN
Element
Lead
Antimony
Arsenic
Chlorine
Sulfur
September
Concentra-
tion
(yg/Nm3)
1010
4390
4
13300
55700
26, 1978
Emission
rate
(kg/hr)
0.12
0.54
0.0005
1.6
6.7
September
Concentra-
tion
(yg/Nm3)
350
4370
11
59500
77100
27, 1978
Emission
rate
(kg/hr)
0.04
0.52
0.0013
7.1
9.1
The engineering and work practice controls of employee exposure
at this smelter are exemplary. The effectiveness of this system
of controls is evidenced by the control of employee exposures
to lead in all work activities associated with the SB furnace
to approximately 100 yg/m3 or less.
In general, the local exhaust ventilation systems provided for
the SB furnace are well designed and maintained. They provide
good enclosure of emission sources, vigorous hood face and
duct transport velocities, access openings and mobility to
allow efficient performance of routine work.
Yard sprinkling and washdown procedures appeared to greatly
minimize entrainment of dust into the air, tracking muddy
materials into other work areas, and splashing of mud on
employee clothing and plant equipment.
Ill
-------
• Washdown procedures employed in the SB furnace operating area
did maintain floors in dust free and clean conditions.
• Reported blood lead data ranged from less than 30 to 80 yg Fb
per 100 ml of whole blood.
• Respirators are not worn at the Bergs^e smelter except for unusually
hazardous jobs such as furnace maintenance.
• The SB furnace process control room was found to be contaminated
with lead (38 to 54 yg Pb/m3). Some improvement could be made in
this area by supplying fresh or filtered air to the control room.
The specific measurements and complete results are available from either the
EPA or NIOSH offices in Cincinnati as mentioned earlier.
RESULTS OF THE INITIAL DEMONSTRATION SITE CHARACTERIZATION
At present, only a portion of the results for the "before" or preliminary
characterization are available. The preliminary results indicate that lead-in-
air levels at the demonstration site are significantly higher than those mea-
sured at the Bergs«5e smelter. When the complete results are available, an
interim report will be issued with a complete list of recommended controls
and BMP's intended to reduce the lead-in-air to acceptable levels. Some
preliminary results are presented below.
Smelter Description
The secondary smelter demonstration site is typical of existing facili-
ties in the United States. Scrap batteries are broken using slow-moving
shears. The battery plates, mud, plant scrap from an adjacent battery
manufacturing facility, drosses, and a variety of other lead scrap are fed
to a vertical shaft blast furnace. This blast furnace is charged using a
skip hoist and traditional slag-tapping and metal-tapping equipment is
employed.
The smelter employs approximately 12 people per shift. Much of the work
is manual. Charging the furnace and handling raw and refined lead in the
smelting building involves simple materials handling equipment. A small
Bobcat loader is utilized to handle charge materials within the smelting
building.
The building which houses the blast furnace, refining kettles and other
smelting equipment is of open-air construction (refer to Figure 4). Several
of the walls are open to allow movement of fresh air in and out of the build-
ing. The floor of the building is paved but rough. It is heavily contaminated
with muds formed by water and paste from battery plates, as well as other mater-
ials tucked in from the yard surrounding the smelting building. Housekeeping
at this particular smelter is a difficult problem due to the rough surface of
112
-------
A '^^y
I -3, / / .
MJ Cr*AtLirt4
I i
O /^
Ssia
Figure 4. Work area sampling locations within smelter building.
113
-------
the floor and the lack of appropriate drains for flushing accumulated lead
particulate materials.
Local exhaust ventilation is provided for various points of emission
associated with the blast furnace and, to some extent, for the refining
kettles. The skip hoist shaft, the charging point at the top of the blast
furnace, the molten lead tapping port, the molten lead launder and molding
area, and the slag tapping port on the-blast furnace all receive some local
exhaust ventilation coverage at present. The two.refining kettles receive
minimal local exhaust ventilation. Only one of these kettles is provided
with back draft slot hoods to remove some heat and a portion of the emissions
from this source. The other kettle is not presently equipped with local
exhaust ventilation. The pigging machine is used to form ingots of finished
lead and is not local exhaust ventilated.
Employee Exposures
The general conditions described at this smelter are similar to many
other secondary smelters. The employee exposures which result from relatively
unsophisticated materials handling procedures and marginally controlled emis-
sion sources are high by comparison to the recommended limits for employee
exposure. The demonstration site smelter, at present, has a very severe
employee lead exposure problem. Air sampling data provided by the smelter
indicate that exposures have varied over a wide range and incorporated the
usual seasonal and process fluctuations. Exposures monitored during the
initial part of this survey confirm those sampling results and indicate that
the smelter does operate with workroom lead-in-air concentrations well above,
or many times the currently accepted exposure limit (refer to Tables 4 and 5).
Breathing zone and work area sampling for lead and antimony was performed
on January 23 and 24, 1979. January 23, 1979 was a relatively normal operating
day at the smelting complex. The weather was sunny and cold with light winds
blowing from several compass directions. Employees at the smelter performed
normal tasks and rotated jobs at the halfway point in the work shift as is
indicated by the change in employee initials associated with each of the jobs
or operations shown in Table 4. January 24, 1979 was a relatively unusual day
at the smelting complex. Heavy rains and strong winds were present throughout
the period monitored. The strong gusty winds caused vigorous drafts through
the smelting building which disturbed the capture characteristics of local
exhaust ventilation hoods and caused the introduction of dust into the atmos-
phere from the building surfaces. By the afternoon of January 24, large vol-
umes of runoff water were entering the smelting complex and leaving on the
down hill side. The smelting building itself was very wet with pools of water
in most of the work areas.
The sampling conducted on these two days indicates that exposures were
generally higher on January 24, 1979. This result is possible due to the
strong air currents which entrained contaminants from the capture zone of
the local exhaust ventilation hoods. As stated earlier, the smelting work
area was very wet. However, fume produced by slag tapping, metal tapping
114
-------
TABLE ,4. RESULTS OF BREATHING ZONE SAMPLING FOR LEAD AND ANTIMONY
1-23-79
Job/operation
Furnace :
Hoist:
Payloader :
Coke:
Lead Pot:
Foreman :
Industrial
Yardmen:
(LW)/(DH)
(NH)/(LW)
(DH)/(LW)
(RH)/(JM)
(NH)/(RH)
(LW)/(NH)
(RH)/(NH)
(DH)/-
(WE)/-
(WE)/-
(CS)/(CS)
(CS)/(DH)
Battery Breaker:
(JM)/(JM)
(JM)/-
(DS)/-
(DS)/-
(RM)/-
(RM)/-
Sampling
interval
0758-1232
1300-1547
0803-1205
1300-1547
0807-1300
1258-1547
0828-1233
1205-1547
0807-1206
1206-1425
0825-1207
1207-1547
0805-1300
1300-1615
0817-1235
1235-1615
0818-1236
1236-1615
Pb*
mg/m3
0.51
0.38
0.37
0.20
0.78
0.48
0.18
0.25
0.21
0.37
0.27
0.28
0.43
0.35
0.69
0.25
0.37
0.23
Sb**
mg/m3
0.051
<0.044
<0. 031
<0.044
<0.025
<0.044
<0.030
<0.033
<0.031
<0.064
<0.033
0.037
<0,025
<0.029
<0.029
<0.034
<0.029
<0.034
1-24-79
Sampling
interval
0803-1212
1234-1559
0801-1233
1238-1558
0807-1300
1343-1557
0818-1342
0821-1205
1214-1603
0815-1237
Pb*
mg/m3
0.62
0.87
0.87
7.8
1.2
0.83
0.66
—
0.66
1.2
0.73.
Sb**
mg/m3
<0.030
0.047
<0.027
<0.037
<0.025
<0.025
<0.023
—
<0.033
<0.032
<0.028
Battery Breaking:
(KA)/(KA)
(WD)/-
(KA)/-
- /(WD)
0833-1209 2.5 <0.034
0826-1211 3.5 <0.029
1210-1515 0.18 <0.040
0823-1200 0.26 <0.034
1203-1520 0.20 <0.037
*Time-weighted-average lead-in-air concentration for
**Time-weighted-average antimony-in-air concentration
the period sampled.
for the period sampled.
115
-------
and refining operations was still present. On both days, the amount of
antimony in the air was very low by comparison to the permissible exposure
limit of 0.5 mg/m3 (29 CFR 1910.1000).
Several employees were found to have extremely high exposures to lead
during portions of work shifts on each day sampled. Battery breakers on the
first day sampled showed high exposures to lead and the hoist operator
encountered very high lead exposure during the afternoon of the second day
sampledv In the case of battery breaking, a splash of lead laden solution
may have caused excessive contamination. In the case of the hoist operator,
the handling of flue dust as it is charged to the hoist could have caused the
deposition of a large lead particle onto the filter.
The work area sampling results shown in Table 5 indicated the same trend
from the first to second day of sampling. Sampling results of the second
day are generally higher than those of the first. Work area concentrations
tend to be much lower than breathing zone concentrations and reflect the fact
that employees are much closer to the source of emission within the smelting
building than were the stationary area samplers positioned at the locations
shown in Figure 4.
TABLE 5. RESULTS OF WORK AREA MONITORING
Sampling Pb* Sb**
Interval mg/m mg/m
Sampling Pb* Sb**
Interval mg/m mg/m3
Employee Breakroom/ 0838-1214 0.15 <0.034 0739-1145 0.11 <0.030
Lunchroom 1215-1559 0.17 <0.033
Location No. 1 (Refer
to Figure 4) 0852-1219 0.050 <0.034 0745-1144 0.37 <0.031
Location No. 2 (Refer
to Figure 4) 0854-1222 0.089 <0.036 0755-1142 0.33 <0.033
Location No. 3 (Refer
to Figure 4) 0910-1224 0.18 <0.038 0749-1144 0.25 <0.031
Location No. 4 (Refer
to Figure 4) 0901-1222 0.12 <0.037 0758-1158 0.30 <0.030
*Time-weighted-average lead-in-air concentration for the period sampled.
116
-------
PLANNED AND PROPOSED CONTROL MODIFICATIONS
Current Control Modifications
The demonstration site smelter is in the process of making major modifi-
cations to its smelting complex to better control ambient and workplace
emissions. An important element in this modernization program is the BergsjJe
Flash Agglomeration furnace. The furnace will work in concert with a new
baghouse facility. The new baghouse system is of much greater capacity than
the system which is currently in operation. The new baghouse exhaust venti-
lation system will provide the capability to exhaust large volumes of air
from the smelter building. This exhausted air volume will be gathered from
strategic exhaust ventilation hoods which will serve major sources of emis-
sion associated with the smelting facility. The following list of major
emissions sources will be served by the combined old and new ventilation
system:
• The charging location at the top of the blast furnace.
• The skip hoist charging elevator.
• The lead tapping and pouring station.
• The slag-tapping hood.
• The refining kettles.
The final hood designs for local exhaust ventilation of these emission
sources are in the development stage at present. The first step in implement-
ing the new exhaust ventilation control program will be to attach existing
exhaust ventilation hood structures to the new system. The next stage in
implementation of the total control program will be to complete the installa-
tion of the flash agglomeration furnace. Once the flash agglomeration fur-
nace is operating in concert with the dust collecting system, attention will
be given to the final design of local exhaust ventilation hoods for refining
kettles, etc. Completion of this control program will take several months.
Base-line exposures data were collected while the existing exhaust ventila-
tion system was in operation.
Optimum control of employee exposures and emissions to the ambient
environment will require more than simply increasing the amount of air
exhausted from the smelting building and improving local exhaust hood designs.
The smelter has plans to improve paving of outside yard surfaces to enclose
portions of the smelting building (cutting down cross drafts which would
interfere with contaminant capture), and to make materials handling modifi-
cations which should help to control escape of contaminants into the work
environment.
117
-------
Control Considerations
Through experience with many U.S. secondary lead smelters and in
particular the Bergs^e smelting complex in Glostrup, Denmark, it has been
learned that very rigorous attention to cleanliness and housekeeping must be
observed if a workplace lead-in-air concentration of 200 ug/m3 or less is to
be met. To approach 200 yg/m3, the demonstration smelter must place much
more attention on improved materials handling and housekeeping within the
smelter building.
There are a number of emission sources which contribute to workroom lead
in air contamination at this smelter. Among these emission sources are the
following:
• Pulverization of raw materials of lead particulate by materials
handling equipment (front end loaders, large and small)
• Manual shoveling and dry sweeping of lead bearing materials,
especially settled particulate, flue dust and plant scrap
• Charging of lead bearing materials to the skip hoist
• Emissions from the charging port at the top of the blast
furnace
• Emissions escaping capture during process upset conditions
such as blow holes
• Emissions from the slag tapping port during slag tapping
• Emissions from tuyere covers during tuyere punching
• Emissions from the lead well, launder, and molten lead
molding operation
• Emissions from cooling finished metal molds
• Emissions from floor surfaces disturbed by personnel and
machinery traffic
• Emissions from the melting of lead in refining kettles
• Emissions from dressing and skimming operations in refining
kettles
• Emissions from the pigging machine as ingots are formed
• Emissions from vibrating surfaces in the roof structure of the
building
118
-------
• Wind blown particulate from outside yard areas, piles of old
flue dust, and materials handling operations
• Emissions from particulate which settles on work clothing and
is introduced into the workers breathing zone
• Emissions from splashes of lead bearing solution which attach
to work surfaces, dry, and then are dislodged by movement of
air or rubbing off by equipment or personnel
• Emissions from tote bins used to collect large particles from
knockout boxes or at the base of dust collectors
Each of the above emission sources and possibly others will contribute
in varying degrees to employee exposures at this smelting facility. To
effectively control exposures attention must be given to each of the possible
emission sources listed.
Several general principles in secondary smelter design and operation
could be applied at the smelter to improve workplace lead-in-air concentrations.
These principles are basic to industrial hygiene and are the foundation of
what EPA terms "best management practices" (BMP's). These principles are:
• Engineering Control
- Isolation of process or employee
- Enclosure of process or employee
- Ventilation
• local exhaust (emission control at emission source)
• dilution of contaminant with fresh or less contaminated air
- Substitution of existing materials, processes or equipment with
less hazardous materials, processes or equipment
- Modification of the physical state of the source or contaminant,
e.g., wetting dry materials
- Denying emission energy to the source, e.g., not heating metals
to unnecessarily high temperatures
• Administrative Control
- Operations practices
- Work practices of employees
- Scheduling of work to minimize exposures during emission-
producing operations
- Rotation of employees to lower exposure areas
- Training of employees
- Housekeeping practices, policies, and procedures (although
housekeeping equipment such as a central vacuum cleaner would
be classed an engineering control)
- Maintenance practices and policies
119
-------
• Personnel Protective Equipment Control
- Respiratory protection
- Skin protection
Engineering Controls—
A major improvement could be made by separating the raw materials hand-
ling and furnace charging operation from the furnace tapping and kettle
refining area of the smelter. This could be done by constructing a partition
between the bucket elevator and the blast furnace. In this manner, contamina-
tion generated by the charging of scrap lead materials would not communicate
directly with the furnace operating and refining kettle portion of the smelter
building. This technique has been effectively demonstrated .at the Paul
Bergs«5e and Sons secondary smelter in Glostrup, Denmark.
Improvement in the furnace charging operation could be made by improved
materials handling techniques. Scrap battery materials and other furnace
charging materials could be conveyed directly to the skip hoist by belt
conveyor or other means. This could reduce the need for manual or payloader
handling of materials. The charge materials storage portion of the building
could be better enclosed to reduce cross drafts and entrainment of particulate.
A sprinkling system could be installed to wet down dry charge materials.
Should the bucket loader still be required, consideration should be given
to equipping this vehicle with a filtered-air, enclosed cab which would offer
some reduction in employee exposure.
The floors in the charge preparation and handling area and in the
smelting and refining area of the buildings should be modified to permit
effective cleaning. Modifications should include provision for wet washing
to sumps or drains which can be performed routinely and periodically during
the workday. These washdown procedures could be replaced by a central
vacuuming system. However, the labor and capital investment involved with
the vacuum system may be prohibitive. In either case, dry sweeping of lead
particulate materials and other waste should be prohibited.
The existing exhaust ventilation hoods serving the blast furnace and
refining kettles will be improved. It will be very important to the eventual
success of this control application to provide maximum capture of lead con-
tamination at each emission source. The slag-tapping hood, in conjunction
with the work practices utilized to tap slag, does appear to do an effective
job of controlling emissions. The hoods associated with molten lead tapping
can be improved to provide better capture of emission from the lead well,
launder and pouring stations. Improved designs for control of this emission
source will be developed in the weeks to come.
120
-------
The emissions associated with furnace charging will be much better
controlled by the new exhaust ventilation service. A much increased volume
of air will be withdrawn from the charging hood to insure that furnace upset
conditions will not create catastrophic release of lead into the work envi-
ronment. This charging hood will work in concert with improved ventilation
of the skip hoist operation.
Attention must be given to the emissions generated in the loading of the
skip hoist itself. Existing ventilation is not adequate to capture contami-
nation in this area. Attention should be placed to providing better cap-
ture through use of improved hood design along with the increased amount of
exhaust volume which will be afforded by the new system. Improved designs
will also be developed for refining kettle hoods which will provide control
of melting, dressing and skimming.
Administrative Control Programs—
While engineering controls are being designed and installed, it is
imperative that employees be afforded maximum protection from exposure through
use of respiratory protective devices and rigorous personal hygiene. The
present locker room, showering and lunchroom facility was found to be sig-
nificantly contaminated with lead. This condition offers the distinct pos-
sibility for employees to eat food and use smoking and chewing materials
which are contaminated by lead.
The contamination of the lunchroom area may result from lead from a
variety of sources. It is adjacent to an industrial battery department
which may give rise to infiltration of lead contaminated air. A more obvious
source of contamination is the traffic of employees dressed in work clothing
to and from the lunchroom. Lead particulate lodged on clothing, hardhats,
etc. may be dislodged while the materials are removed or during normal move-
ment within the lunchroom.
To solve lunchroom contamination problems will require the redesign of
the shower, locker room and lunchroom facility. It is understood that a new
building for industrial battery production is being constructed which will
make space available adjacent to the existing hygiene facility. This space
could be used for expansion and rearrangement of the facility.
Another means of reducing employee exposures during the installation of-
engineering controls would be to rotate workers to a lower exposure area
during a portion of their workshift. It is not known whether this form of
administrative control would be feasible since it would require the training
of many new operators to serve in the rotating work assignments.
121
-------
SUMMARY
The results of the sampling study at the Bergs^e smelter indicate that
the combination of control technology and BMP's used there for workplace/
fugitive emission control is exemplary. The lead-in-air levels measured
both inside and outside the smelter are as low or lower than any similar
measurements reported for U.S. smelters. Reported blood/lead levels for the
Bergs«Je smelter workers range from less • than 30 to 80 yg per 100 ml of whole
blood. These values are comparable to blood lead levels found at U.S.
smelters, however, respirators are not worn at the Bergs^e smelter except
for unusually hazardous jobs such as furnace maintenance.
The preliminary results from the U.S. smelter demonstration site indicate
that lead-in-air levels are extremely high, both inside and outside the smelter.
It is hoped that with the new ventilation system, the Bergs«5e agglomeration
furnace, the institution of BMP's and a general plant cleanup, lead-in-air
levels can be reduced to within legal limits. Of additional importance
will be the documentation of both the cost of these modifications and any
observed drop in blood lead levels in the smelter workmen.
122
-------
GETTING READY FOR INEALABLE PARTICLES
D. Bruce Harris
Process Measurements Branch
Industrial Environmental Research Laboratory
Environmental Protection Agency
Research Triangle Park, NC 27711
ABSTRACT
In support of a reassessment of the total suspended particu-
late standard now underway by EPA1s Office of Air Quality Plan-
ning and Standards (OAQPS), three laboratories in EPA's Office
of Research and Development (ORD—Health Effects Research
Laboratory (HERL-RTP), Environmental Sciences Research Laboratory
(ESRL), and Industrial Environmental Research Laboratory (IERL-
RTP)—are examining potential measurement requirements. HERL
has recommended a 15 ym upper cut size and a second division at
2.5 ym. Current particulate sampling techniques do not provide
such data for either ambient or stack. A recommended measure-
ment program has been developed with short-term modifications
for existing samples and a longer effort to fully investigate
the requirements for more information including stack conden-
sables.
INTRODUCTION
By December 31, 1980, amendments to the Clean Air Act
require EPA to update all the air quality criteria documents,
including that for particulate matter, and decide whether to
reaffirm existing standards or to issue revised standards. Three
of EPA's ORD laboratories—HERL-RTP, ESRL, and IERL-RTP, all at
Research Triangle Park, NC—are supporting the effort of the
Office of Air and Waste Management's (OAWM) OAQPS to reassess
the particulate standard.
HERL-RTP examined the data available relating deposition in
the human respiratory system to particle size. The summary of
their recommendation provides the guidance needed for measure-
ment requirements to support a revised standard:
"It is recommended that research to develop information for
a size-specific standard should focus on 'inhalable particu-
late1 (IP) matter defined as airborne particles <_ 15 ym
123
-------
aerodynamic equivalent diameter. This particle size range
relates to that fraction of particulate matter which can
primarily deposit in the conducting airways and the gas-
exchange areas of the human respiratory system during mouth
breathing. It is also recommended that a second particle
size cut-point of £ 2.5 ym diameter be incorporated in the
air sampling devices/ based upon considerations of the
chemical composition and the size distribution of airborne
particles, and on the predominant penetration of particles
£2.5 ym diameter into the gas-exchange region of the
respiratory tract. Data collected in this size range could
be used in conjunction with epidemiological health para-
meters to^refine an inhalable particulate standard in the
future."<2'
DEVELOPING A SOURCE MEASUREMENTS PROGRAM
Though the standards to be revised are ambient/ the modifi-
cations will directly impact upon source measurements. Both
source and ambient measurements must measure the same thing in
order to evaluate the impact of a source on the regional air
quality. A Task Force (headed by James Abbott, IERL-RTP) was
formed by ORD to support OAQP's need for emission data while
developing a standard. A subcommittee, consisting of John Nader
of ESRL and Bruce Harris of IERL-RTP, was appointed to define
the source measurement requirements.
EPA personnel involved in the inhalable particulate (IP)
program met to develop specifications and priorities to obtain
the IP data. The needs and priorities agreed to at that meeting
are presented in Table 1.(3) The measurements subcommittee felt
that a research program designed to respond to these needs would
be extensive and should be formulated with the best available
knowledge. A number of prominent aerosol scientists in the U.S.
were invited to participate in a workshop to lay the groundwork
for such a program.
IP SAMPLER CURVE
The workshop first tackled the development of performance
criteria for devices to be used to meet the 15 ym cut point
suggested by HERL-RTP. ESRL had- already developed a prototype
head for ambient samplers and had amassed some calibration data
(Fig. 1).(4) This curve approximates the "standard curve" for
IP samplers. The workshop attendees felt that these data need
to be confirmed and extended by independent researchers. Since
it may be some time before the standard curve is accurately
known, it was recommended that EPA adopt a standard mathematical
performance curve with specified tolerances. Then any newly
124
-------
TABLE 1. EPA NEEDS AND PRIORITIES FOR EMISSION DATA
FROM STATIONARY SOURCES
Priority Program Element
Description
la
3
4
5
Ib
or
2c
2a
2b
3
Total mass — 15 ym aerodynamic diameter
- cut to match ambient monitor
- DSQ = 15 ym ± 2 ym
- slope to match ambient monitor
Condensable fraction of total mass
Discrete size fractions — 15 ym
Size distribution - continuum — 80 ym
Size distribution - continuum ± 15 ym
Chemical composition of materials in
program elements 1 and 2
Bioassay of materials in program
elements 1 and 2
125
-------
I ' i
80
60
40
20
£ 10
ui
I 8
SIZE SELECTIVE HI-VOL
O 2 km/hr
• 8 km/hr
O 24 km/hr
A 2 km/hr
A 8 km/hr
D 2 km/hr BECKMAN DICHOTOMOUS INLET
SIZE SELECTIVE MEMBRANE
I
I i I
4 6 8 10
AERODYNAMIC PARTICLE DIAMETER,
20
40
Figure 1. Calibration of the Hi-vol and Dichotomous Sampler Inlets.
126
-------
developed devices (including the hi-vol and dichotomous sampler)
must be calibrated and shown to agree with the mathematical
curve within the specified tolerances.
Suggested performance criteria for IP samplers are shown in
Fig. 2. (5) All samples would be required to fall within the per-
formance indicated by the shaded area. The shaded area was
constructed by plotting two log-normal curves, with geometric
standard deviations (ag) of 1.3 and 1.7, through the 15 jam, 50
percent point and connecting the 16 percent and 84 percent pene-
tration points. After allowing 10 percent for wall losses of
small particles and 10 percent penetration of large particles,
the area is completely defined as shown. The tolerance in D50
is 15 ± 2 ym. After more thorough characterization of existing
and newly developed sampling systems, it may be possible to
place more stringent tolerances on a standard curve.
TOTAL INHALABLE MASS SAMPLING
Obtaining total inhalable mass will be markedly different
from current practice. Inertial precutters are necessary to
meet the proposed performance curve and must be operated at a
fixed flow rate to maintain constant performance. Therefore,
variation of the flow through the mass sampler, as now done with
Federal Reference Method 5 for total mass, cannot be tolerated
with existing technology. As a result, the precutters may limit
flow rates to a maximum sampling rate of 14 L/M (0.5 ft3/min)
due to the physical size constraints for instack operation.
HERL-RTP indicated the desirability of obtaining information
on particles less than 2.5 ym for possible future standards,
since particles in this size range are predominately associated
with the gas exchange regions of the lungs. Workshop partici-
pants felt that this could be met in situ by using a cyclone in
series with the 15 ym precutter. This would probably reduce
the probe losses reported in many field tests.
The original IP needs list called for the inclusion of
Condensable material with the total mass. This problem was dis-
cussed extensively (presented later in this paper). It was
suggested that no measurement system would be sufficiently de-
monstrated to support a 1980 standard. OAQPS has accepted this
but wishes to accelerate development of acceptable methods for
condensable material.
PARTICLE SIZING
Instack impactors and series cyclones have been used to
obtain extensive particle size data on a number of processes.
127
-------
99.99
MJ
99.1
95
90
70
50 5
40 £
1.0
2.0
4.0 *.0 t.0 10 20
PARTICLE DIAMETER, Jim
20
10
2
1
0.5
0.2
0.1
0.06
0.01
40 CO BO 100
Figure 2. Recommended IP Sampler Specification.
128
-------
Unfortunately, most current operating practices yield upper cut
points of 10 vim or less which do not meet the IP performance
needs. Most current inertial sizing devices can meet these
criteria by reducing sampling rates. Concern was expressed by
the aerosol scientists that the bounce and collection efficiency
roll-over observed under current practices may be aggravated by
the reduced flow rate. This potential problem and recalibration
under reduced flow are currently being investigated.
CONDENSABLE MATERIAL
Measuring accurately the nature of condensable material in
a process stream and determining its physical and chemical pro-
perties in the atmosphere presents perhaps the most important
and challenging task of the IP program. Many process streams
contain materials in the vapor state which condense to form
homogeneous and/or hetrogeneous aerosol particles. Some "conden-
sable" vapors will not actually appear as particles in the air
downwind of the source, since their equilibrium vapor pressure
may be high enough that they will not condense, or condensed
vapors may re-evaporate as dilution is increased. In order to
measure the properties of the aerosol emitted to the atmosphere,
the condensation in the sampling system must simulate plume
behavior. Lower dilution rates may maximize the supersaturation
and overestimate the contribution of condensed vapors to the IP.
The rate of dilution may also be important. In studies of
aerosols produced by diesel engines, very rapid dilution of
1000:1 has prevented condensation of organic vapors. 'Slower
dilution may lead to condensation which clearly occurs under
some circumstances in plumes near the stack outlet.
The contribution of condensed vapors to mass emissions can
be very significant under some circumstances. Consider a coal-
fired boiler burning 10 percent ash coal and with 99 percent
particle removal. The emitted ash particles will amount to 0.1
percent of the original coal mass. For a conservative estimate
of the contribution of sulfates, assume that the coal contains
1 percent S, and that 5 percent of this (or 0-05% of the coal
mass) is emitted from the stack as sulfates (primarily sulfuric
acid). If all of the sulfate were to condense, it would equal
one third the total amount of solid mass emitted.
In other processes, notably smelters, the fraction of mass
emitted as vapors and condensing at the stack/ambient interface
may far exceed the mass existing in solid form within the stack.
During sampling, it would be preferable to dilute and cool
a representative sample of the particulate and gas as it exists
in the stack. However, because of the extreme difficulty in
129
-------
transporting large particles through sampling lines and dilution
systems without unacceptably high losses, it will almost certain-
ly be necessary to collect and size particles greater than 2-3 ym
in the stack without dilution or cooling. The smaller particles
would then pass to the out-of-stack dilution system for further
analysis. Removing particles from the sample aerosol removes
surfaces where condensation likely occurs upon cooling and
dilution. Thus, to some extent, the experiment will be biased,
and the size distribution and mass concentration of the aerosol
in the diluter will be distorted. Experimental and theoretical
studies must be performed to determine the magnitude and nature
of the errors associated with the compromised procedure.
FUGITIVE EMISSIONS
Nonducted sources are among the most difficult to measure
accurately. Current practice is to use ambient sampling tech-
niques in upwind-downwind studies or stack methods for hooded
sources. The IP methods developed for the other areas can be
incorporated into the existing fugitive methods. Perhaps the
greatest need is to develop suitable sampling strategies. Some
special research fugitive samplers, such as lERL-RTP's FAST
(Fugitive Assessment Sampling Train) and ESRL's Mega-Vol, are
being modified to include a 15 ym first cut.
CALIBRATION
An area that cannot be neglected is the calibration of
instruments and systems with standard aerosols. The development
of calibration procedures is not straightforward and will require
considerable time, effort, and expertise. The minimum parameters
which must be considered are presented in Table 2, It is impor-
tant to be able to calibrate devices at more than one facility
so that results can be compared and verified.
A RECOMMENDED RESEARCH AND DEVELOPMENT PROGRAM
At the workshop, a program was developed to meet the IP
needs outlined at the beginning of this paper. This program is
presented in Table 3.(6) Some of the items have already been
discussed. Still others represent long-term development which
will be undertaken. Many are desirable but, due to funding
limitations and low priority, will not be supported by EPA.
130
-------
TABLE 2. CALIBRATION PARAMETERS
Particle Type - Wet, dry, monodisperse, polydisperse
Velocity of Gas
Particle Diameter - Aerodynamic, standard set for verifying size
Sampling Efficiency vs. Particle Size
Wall Loss vs. Particle Size
Aerosol Concentration - Can affect the operation of some devices
Temperature - Must simulate sampling conditions
Wind Direction - Will certainly have some effect
Loading - Reentrainment
Frequency of Calibration
Number of Devices Calibrated - Initial prototypes, or all?
131
-------
TABLE 3. TECHNIQUES RECOMMENDED FOR USE IN OBTAINING DATA TO SET IP
EMISSION FACTORS FOR INDUSTRIAL SOURCES
Time Scale
Program Element
Stack
Ambient
Fugitive
Immediate
la. Total Mass 5 15 ym
Ib. Condensable Mass
2c. Size Dist. S 15 ym,
Discrete
2a. Size Dist. 5 80 ym
2b. Size Dist. 1 15 ym
Cont.
Re-analyze data
to
Six Months
la.
Ib.
2c.
2a.
2b.
Cyclones w/impingers,
for la, Ib, 2c .
Also SASS train
Cyclones, impactors
Lo-Vol, Hi-Vol, Dichot.
(Similar inlets requiring more cali-
bration and performance verification)
Ib, N/A FAST train w/new inlet
Dust fall onto suitable substrate, with
automatic or manual image analysis
Longer Term
Nine months
la.
Ib.
2c.
2a
15 ym, 2.5 ym cyclones
followed by dilution
Run in parallel with
Method 17
15 ym, 2.5 ym cyclones
Dilution - Out of stack
sizing, (imp., opt., diff.,
elec.), body impactor for
2-80 ym
Rotating, body impactor
optical, diffusional, elec.,
centrifuge
Mobile labs
-------
SUMMARY
An IP standard is being actively considered by EPA. Health
considerations suggest an upper size cut at 15 ym, and all
sampling methods need to be modified to provide the cut.
Programs to provide the information needed to measure IP
emissions and to assist OAQPS in reassessing the particulate
standard have been developed with assistance from many U.S.
aerosol scientists. These programs have been prioritized and
many are now underway.
REFERENCES
1. "Amendments to the Clean Air Act of 1970", Public Law 9595,
August 1977.
2. "Size Considerations for Establishing a Standard for
Inhalable Particles", by F. J. Miller et al., accepted
by JAPCA.
3. Internal Memorandum, J. S. Nader (EPA/ESRL) and D. B. Harris
(EPA/IERL-RTP) to J. H. Abbott (EPA/IERL-RTP), November
1978.
4. Private communication, C. Rhodes (EPA/ESRL) to D. B. Harris
(EPA/IERL-RTP), November 1978.
5. "Sampling Techniques for Inhalable Particulate Matter",
ed. by W. B. Smith, Special Report for EPA Contract
68-02-2610, January 16, 1979, page 7.
6. Ibid, page 23.
133
-------
LEAD AND ZINC INDUSTRY
Session Chairman: Jerome E. Cole
International Lead Zinc Research
Organization, Inc.
New York, N.Y.
134
-------
AN OVERVIEW OF CONTROLS
IN PRIMARY LEAD AND ZINC
W. A. Lemmon
D. Haliburton
Air Pollution Control Directorate
Environment Canada
Ottawa, Ontario
ABSTRACT
The lead-zinc processes are briefly reviewed with emphasis
placed on the potential particulate emission points. Particulate
emissions from the industry are discussed with an emphasis on
specific problems such as lead and trace metals emissions. The
use of particulate control equipment is reviewed and a number of
outstanding control problems are highlighted. New or emerging
process technology is discussed especially in relation to its
impact on particulate emissions control. Existing North American
environmental legislation and pending legislation will be sum-
marized.
INTRODUCTION
Lead and zinc frequently occur together in the same ore.
Consequently the primary outputs of the two metals are often
interrelated and mines frequently produce significant tonnages
of both. The normal practice is to produce separate lead and
zinc concentrates, using froth flotation wherever possible, and
treat them in separate smelting and refining processes. In
order to provide a suitable grade of lead concentrate and zinc
concentrate, it is often necessary to extract a middlings
fraction as a lead-zinc concentrate; this can amount to between
10% and 15% of the total concentrate produced. Most of this is
treated in the Imperial Smelting Process. Several alternative
processes are available to recover the metals and producers
select and modify the processes to suit their particular condi-
tions. Important considerations include concentrate composition
and environmental and economic aspects.
In 1977 approximately 63 primary lead and 76 primary zinc
smelters operated throughout the non-communist world. Tables 1
135
-------
and 2 show the total rated output capacities according to the
principal processing methods.
TABLE 1. ESTIMATED LEAD OUTPUT CAPACITY OF PLANTS IN THE NON
COMMUNIST WORLD ACCORDING TO PRINCIPAL PROCESSING
METHODS (AMERICAN BUREAU OF METALS STATISTICS 1977)
Method No. of Plants Refined Lead Output STPY
Conventional 54 3,941,000
Imperial Smelting 8 290,000
Boliden Electric Furnace 1 55,000
Total 63 4,286,000
TABLE 2. ESTIMATED ZINC OUTPUT CAPACITY OF PLANTS IN THE
NON-COMMUNIST WpRLD ACCORDING TO PRINCIPAL
PROCESSING METHODS (ABMS, 1977)
Method No. of Plants Refined Zinc Output STPY
Electrolytic 44 4,182,000
Pyre-metallurgical 24 1,354,000
Imperial Smelting 8 672,000
Total 76 6,208,000
Table 3 summarizes the total world production of lead and
zinc, including the Communist Bloc. Note, Tables 1 and 2 refer
to the related capacities of the processes and not the actual
yearly outputs.
136
-------
TABLE 3. WORLD PRODUCTION OF REFINED LEAD AND SLAB ZINC FOR
1977 AND 1976 (ABMS, 1977)
Lead Ouptut Zinc Output
Region 1976 1977 1976 1977
N. America 1,042,472 995,066 1,249,792 1,167,875
S. America 171,800 183,000 157,722 164,580
Europe 993,614 1,045,290 1,752,765 1,847,000
Asia
Africa
Australia
250,969
114,700
235,300
257,521
123,700
233,100
878,383
204,300
275,100
934,100
201,800
282,400
Sub-total 2,808,855 2,837,677 4,518,062 4,607,180
Communist Bloc 1,212,000 1,211,800 1,873,300 1,884,900
Others 16,300 15,900
Total Production 4,037,155 4,065,377 6,391,362 6,492,080
Primary lead and zinc processing plays an important role
in an industrial society. The problems faced by lead and zinc
producers in controlling particulate emissions, and conforming
to occupational .hygiene standards are similar to those faced by
other sectors of the metallurgical industry2.
THE PRIMARY LEAD INDUSTRY
Lead Production Processes
Most of the world's primary lead is produced from sulphide
concentrates in reduction blast furnaces. Figure 1 shows a
generalized flow sheet. The conventional process consists of
feed preparation, sintering, blast furnace reduction, bullion
dressing and either electrolytic or fire refining. This process
accounts for approximately 92% of installed western world
capacity. A similar process, the Imperial Smelting process,
produces an additional 7% from zinc-lead sinter. Several new
processes have been developed that eliminate the need for
137
-------
FIGURE 1 GENERALIZED LEAD PRODUCTION FLOWSHEET
ESTABLISHED
ROUTE
FLUX
FUEL
COKE
AIK
I
MIXINC
MiXiNC.
SINTERING
J-J
BLAST
FURNACE
Slag to j
Unct'o ""^
Waste
LEAD BULLION
DE COPPER! XI N— Slag to Waste
138
-------
sintering and premit direct smelting of lead concentrates; these
offer advantages in environmental control and occupational
hygiene. Examples are the Kivcet, the Boliden TBRC and the
Boliden electric furnace.
In both the conventional and Imperial processes, concen-
trates are sintered to produce a hard, agglomerated oxide charge
for the blast furnace. The feed is first thoroughly blended in
bedding or blending plants and pelletized. A low sulphur content
is required, and necessitates the recycling of large amounts of
crushed sinter to dilute the fresh concentrate. Sintering
involves heating the charge to just below its melting point on
a moving grate. Up-draught or down-draught machines are used.
Temperatures of about 1000°C are attained at the combustion
front. Part of the volatile metals present are evolved, e.g.,
mercury, arsenic, lead, cadmium and zinc, and are contained in
the off gas together with 862 and entrained particulates. The
sintering machines are normally operated to produce gas with
4-7% SO2 which can be treated in conventional contact acid
plants. The gases from different wind boxes may be split into
strong and weak streams and the weak stream (less than 2% 862)
discharged following particulate recovery. Alternatively, the
weak gases may be recirculated and a single, strong off gas
stream produced. The bulk of the particulates are recovered in
electrostatic precipitators or fabric filters. The acid-plant
wet-gas cleaning train provides further control of particulate
emissions, by collecting volatile and particulate constituents
which escape primary collection, for disposal in the acid plant
sludge.
The sinter is crushed, screened and charged with coke and
flux to a blast furnace, where it is reduced and melted. Coke
reduces lead oxide and impure lead bullion separates from slag
in a settler. The furnace gas is drawn from the top and cooled
from about 750° to 140°C by the addition of cooler air. Sus-
pended particulates are recovered in fabric filters. The slag
is tapped and granulated with water, or delivered molten to a
slag dump. Depending on the zinc content, slag may be fumed by
the injection of coal in a reverberatory, or an electric furnace,
to recover zinc.
The Imperial Smelting Process is similar to the conven-
tional blast furnace and is considered primarily as a zinc
process. It is described in this section because of its simi-
larity to the lead blast furnace. The furnace uses preheated
sinter, hotter blast air and higher shaft operating temperatures.
Zinc is vapourized in the shaft and recovered by shock cooling
the gas in a lead splash condenser. A wet scrubber captures
residual vapours and particulates escaping the condenser.
The impure lead bullion is cooled in dressing furnaces.
Some impurities principally copper and nickel, and to a lesser
139
-------
extent antimony and arsenic, are precipitated according to their
solubility limits. Batch or continuous processes are used.
The drossed bullion is refined to produce lead and certain
by-products. Fire or electrolytic refining is used. Fire
refining involves the sequential removal of dissolved impurities
by reacting molten lead with reagents at controlled temperatures.
Various drosses are formed and removed mechanically. Electro-
lytic refining involves casting the bullion as anodes, and plat-
ing out lead on cathodes in electrolytic cells using hydrofluo-
silicic acid. The impurities are retained as a sponge on the
spent anodes, and are refined to recover by-products.
Several stages are involved in pyrometallurgical refining
and alternative methods can be used for each stage. The stages
are:
1. Softening
Molten lead is oxidized to form a dross which contains
the arsenic, antimony and tin. Oxidation may be
carried out using an air lance in a reverberatory
furnace, or by stirring lead with sodium nitrate and
caustic soda in the presence of air in a reaction
kettle at 700°C. The drosses are removed and smelted
to recover by-products. An alternative method passes
molten lead through an immiscible sodium salt melt.
The impurities dissolve in the salt melt and are sub-
sequently recovered hydrometallurgically.
2 De-silvering
Zinc is stirred into batches of softened bullion, or
bullion is poured through molten zinc in a continuous
desilvering column. The precious metals dissolve in
the zinc. Temperatures are controlled so that zinc,
containing the precious metals, precipitates from
solution to form a crust.
3. De-zincing
A vacuum is maintained over molten lead at 600°C and
zinc is distilled and condensed on water-cooled hoods.
Batch or continuous processes are used.
4. De-bismuthing
Calcium and magnesium alloys are reacted with the lead
in processes similar to desilvering. Bismuth dissolves
preferentially in the melt and forms a scum. Batch or
continuous processes are used.
140
-------
The final lead is treated with sodium hydroxide to
remove residual impurities and cast as ingots.
Gaseous and fume emissions are controlled by ducting
the gas to high efficiency control devices. The gases
have to be cooled prior to collection from operations
where arsenic is evolved.
Electrolytic refining isolates the impurities in the anode
slimes and minimizes the amount of material to be treated pyro-
metallurgically. The slimes are melted, with fluxes, in a small
reverberatory furnace. A tin-bearing slag and a metal are
formed. The slag is smelted to recover tin. The molten metal is
transferred to another furnace and lanced with air. Arsenic and
antimony are fumed off, and a lead-bismuth slag and a dore
metal formed. Gold and silver are recovered by cupelling and
electrolyzing the dore metal. The bismuth slag is smelted and
chloridized to recover bismuth. Gases containing arsenic and
antimony are generated from the pyrometallurgical process and
the gases have to be cooled to condense the vapours. Fine
particulates are formed and high-efficiency fabric filters have
to be used to control emissions. Arsenic trifluoride may be
formed by reaction with residual electrolyte; this has a rela-
tively high vapour pressure at the fabric filter operating
temperature, and secondary scrubbing may be required to reduce
arsenic and fluoride emissions to acceptable levels.
Alternative processes exist to smelt concentrates directly
to lead bullion. Boliden have developed a top blown rotary
converter (TBRC) to process high-grade reduction phases; lead
is smelted in a 60% oxygen stream and slag and bullion formed;
a high-grade S02 gas is produced which is treated by waste heat
boilers, electrostatic precipitators and sulphur recovery plants.
The gas volume per unit of concentrate treated is low by virtue
of the oxygen enriched air used. A large proportion of the lead
is contained in the slag and is reduced by the addition of coke.
The off gas is diverted to a baghouse during the reduction mode.
The process offers advantages in environmental control; the
unit is relatively small, SOa concentrations are high because
of the use of oxygen and a secondary hood can be used to capture
fugitive emissions.
Boliden also use a combination of an electric furnace and
converter to produce lead bullion. Dry concentrate and flux
are smelted in the electric furnace freeboard and additional
heat supplied by the electrodes. Complete desulphurization
is avoided to minimize slag losses; further desulphurization
takes place in a converter. Particulates are recovered from
the furnace and converter gases. The furnace gas is treated in
sulphur recovery plants. The magnitude of the emission control
problem is less than that of the conventional sinter plant-blast
furnace process.
141
-------
The Kivcet process has been developed in Russia to smelt
lead-zinc concentrates directly in a combination of an oxygen
flash and an electric furnace. Concentrates are melted in the
oxygen flash shaft furnace and flow under a partition wall into
an electric furnace. Coke is added to the slag and zinc fumed
off. Lead: bullion is tapped from the furnace. Two gas streams
are produced, a strong 862 flash stream and a weak SOa electric
furnace stream. The former will be treated in sulphur recovery
plants and the latter discharged following zinc recovery. The
process has not yet been applied in the Western world. It offers
advantages in pollution control but may be limited by capital and
operating costs.
Particulate Pollution Problems and Abatement
Lead is a widely dispersed element and is usually present
in trace amounts in food, air and water. It serves no useful
function in human metabolism and elevated levels cause adverse
effects. Lead is a cumulative poison and its concentration in
the environment has to be minimized. In the elemental and sul-
phide forms it is relatively safe from a health point of view
but as an oxide it is readily absorbed by the body, particularly
if present as fumes smaller than 10 microns, which are absorbed
by breathing. The control of particulate and fine oxide fumes
generated during smelting and pyrometallurgical refining opera-
tions is important. Emissions can be controlled using high-
efficiency particulate collection devices.
The emission of other metals may also occur, namely arsenic,
cadmium, zinc, antimony and mercury. Measures to abate lead
emissions are usually effective in abating emissions of these
other elements as well.
Emissions can occur from a number of sources in the plants,
including both the process off gas streams, and also the fugitive
escape of particulates and fumes to the general work area. The
latter include sources such as material storage and handling,
crushing and screening, leaks from process equipment, tapping
operations, molten metal transfer,etc. Ventilation systems are
installed to remove fugitive emissions from the working environ-
ment. Proper design and good housekeeping and maintenance l0?
practices can limit the severity of fugitive emissions. Two • '
types of gas stream have to be cleaned, process off gas and
ventilation-gas. The type of process used affects the magnitude
of the control problem.
The sinter plant-blast furnace processes represent the
major method of bullion production. Emissions from the sinter-
plant process off gas stream can be effectively abated by direct-
ing all the process gas to sulphur recovery plants. The bulk of
the particulates and condensed fumes are removed in electrostatic
precipitators and de-misting towers in the sulphur recovery
142
-------
plants. Contact sulphuric acid plants are the usual abatement
process and normally require S02 contents greater than 4%. This
requires specialized operation of the sinter strands with gas
recirculation. In some cases weaker gases from certain wind
boxes may be segregated and treated separately, usually being
diverted to the blast furnace baghouse. In other cases sulphur
is not recovered from the gas, and gas is discharged following
particulate recovery. Emissions are higher for these cases.
Some ores contain trace amounts of mercury which raises
concern because of its toxicity. Nearly all of the mercury is
driven off in sintering and most of it is collected in the wet-
gas cleaning section of the sulphur recovery plants, if sulphur
abatement is used. Otherwise, most of the mercury will be
emitted. The acid plant waste sludge has to be safely disposed
of to minimize potential environmental effects. The mercury
content of the acid may also have to be reduced to meet acid
product specifications. For instance, many acid consumers now
specify mercury concentrations below 0.5 ppm in their contracts.
Emissions from blast furnaces are controlled using high-
efficiency fabric filters. The gases are cooled prior to
collection using air drawn from plant ventilation systems or
outside. The former is preferred since it reduces the volume
of ventilation gas to be treated in other facilities. Emissions
from the Imperial Smelting furnace are controlled by wet
scrubbers downstream from the zinc condenser.
Fugitive emissions are usually a major problem for the
sinter plant-blast furnace process. Properly designed ventila-
tion systems have to be installed. The newer direct smelting
processes, such as the Boliden TBRC and the Kivcet offer advan-
tages in the elimination of sintering and the attendant handling
of large amounts of recycle material, coupled with more compact
plants, fewer unit-processes and lower off gas flow rates. The
volume of exhaust gas to be cleaned is lower and less effort and
expense are required to meet occupational hygiene standards.
Particulate, arsenic, and antimony emissions can occur
from dressing. Emissions are controlled by cooling the gas to
condense the volatile metal vapours and passing the gas through
high-efficiency fabric filters to capture the fine particulates.
Dressing removes much of the arsenic and reduces the magnitude
of potential arsenic emissions in down-stream refining opera-
tions. Continuous dressing processes eliminate the fugitive
emissions from hot metal transfer.
Lead can be refined electrolytically or by fire refining.
The electrolytic process offers advantages in industrial
hygiene and pollution control as pyrometallurgical processes are
limited to by-product recovery, and intermediate transfers of
molten material are absent. Electrolytic refining is used
143
-------
extensively in Japan where strict environmental standards apply.
Fire refining is favoured in many instances because of its lower
capital and operating costs/ and adequate controls can be
installed to meet environmental and occupational hygiene stand-
ards. The facilities have to be properly designed . The magni-
tude of particulate emissions and industrial hygiene problems
occuring from different stages in fire refining varies and
depends on the method used3. Emissions can be controlled using
properly designed flue and ventilation systems, duct process
exhaust and fugitive emissions'*. High-efficiency collection
devices are used. Emissions are most serious from the softening
and softening dross treatment processes, where arsenic and
antimony vapours are evolved. The gases have to be cooled to
condense the vapours and cleaned in high-efficiency fabric fil-
ters. The other operations (de-silvering, de-zincing, de-
bismuthing and by-product treatment) have to be properly hooded
and ventilated. Exhaust gases are cleaned in fabric filters.
Some fire refineries are designed so processes operate continu-
ously or semi-continuously. This minimizes fugitive emissions
from material transfer.
Slimes treatment, and anode cathode casting, are the only
sources of particulate emissions from electrolytic refining.
Slimes treatment is the major source of emissions. Large amounts
of arsenic and antimony are volatized in the process and the
gases have to be cooled, to condense the vapours and the fine
particulates captured in high-efficiency fabric filters. Arsenic
trifluoride and other fluorine compounds may be formed as a
result of reaction with residual electrolyte; significant con-
centration can be present as vapours even after cooling and
particulate recovery, because of the high vapour pressure.
Secondary cleaning in wet scrubbers may be used. The problems
may be minimized by washing the slimes to remove electrolyte
prior to treatment.
THE PRIMARY ZINC INDUSTRY
Zinc Production Processes
Zinc is produced from sulphide concentrates by processes
which include pyrometallurgical operations. Two processes are
used: electrolytic (hydrometallurgical) and pyrometallurgical
refining. Figure 2 shows a generalized flowsheet. The electro-
lytic process is dominant and accounts for approximately 67%
of the Western world capacity. The pyrometallurgical process
includes the Imperial Smelting Process described previously
under Lead Production.
Roasting is used in all the electrolytic and most of the
pyrometallurgical refineries; some pyrometallurgical refineries
sinter the concentrates directly without pre-roasting. Roasting
144
-------
FIGURE 2 GENERALIZED ZINC PRODUCTION FLOWSHEET
ELEC'I
RO
\
ZINC
CONCENTRATE
ROLYTIC . PYROMETALUIKCICAL
UTE ROUTJ
I
1
1
ROASTING •< AIR AIR *- ROASTING '
\
.MM«^^«M CTTT PTTTTPT C A r*T n f~" "~"
1«^ — o Ul*ru UK1 L AUJ.X) 1
>' t
LEACHING FUEL »- SINTERING
\
SOLUT
PURIFIC
1
' Waste ^ '
IQN FUEL >- FURNACE
ATION ' ' " "
' ^ S .1 a }; L o
W.-uiLc;
> '
ELECTROLYSIS Zinc may be refined
\
column , depending on
product required.
i
Zinc cathodes to
melting & casting.
145
-------
converts the zinc to an acid soluble form in electrolytic
refining, and a reducible, volatilizable form, in pyrometallurgi-
cal refining. Three types of roasters are used: fluid bed,
multi-hearth and suspension roasters. The fluid bed is the most
recently developed and has been installed in newer facilities and
replaced others in many existing plants.
Fluid beds provide fast intimate reactions between the
sulphides and air. High temperatures are generated by the auto-
genous combustion and heat is recovered from the off gas in
waste heat boilers. Up to 80% of the product exits in the off
gas and is recovered by waste heat boilers, cyclones and elec-
trostatic precipitators. The gas contains 8-13% SO2 and is
treated in sulphuric acid plants. These plants have wet scrub-
bers, wet electrostatic precipitators and demisting towers
ahead of the catalyst, to remove residual particulates to pro-
tect the catalyst beds. Particulate emissions are virtually
absent, since any material passing the electrostatic precipitator
is removed in the acid plant.
In multi-hearth roasters concentrate is roasted as it is
contacted by hot combustion gas while being raked down through
a series of hearths in a cylindrical vessel. Coal may be added
and fuel burners used to supply heat. The reaction is slow and
excess air has to be added to obtain the required desulphuriza-
tion. The S02 contents are lower than those of fluid beds and
SOz fixation is not always possible. Gases with approximately
5% SOz can be produced and treated in acid plants, if good seals
are maintained. The hearth temperatures vary between 450° and
700°C. Fewer solids are entrained in the off gas than in the
case of fluid beds. The multi-hearth roasters have greater
flexibility to eliminate volatile elements, such as arsenic
and lead, since the hearth temperatures and oxidation conditions
can be varied. There is a greater tendency to fix these as non-
volatile oxides in fluid bed roasters. The potential for parti-
culate emissions is greater when acid plants are not used.
Suspension roasters are a hybrid of fluid bed and multi-
hearth roasters. The concentrate is dried on the lower drying
hearths and then charged to the upper combustion chamber.
Approximately 60% of the product is entrained in the off gas and
recovered in waste heat boilers, cyclones and electrostatic pre-
cipitators. The balance comes from the lower roasting hearths.
Gas containing 10-13% S02 is produced and treated in acid
plants.
The roasted product is leached and zinc recovered elec-
trolytically, or sintered and treated in pyrometallurgical
volatilization processes. Electrolytic recovery results in only
minor particulate emissions; however, acid mist is generated
and good ventilation has to be provided. The calcine is leached
in sulphuric acid and zinc dissolved. The solution is purified
146
-------
in a number of stages and various by-products are produced.
Iron is precipitated by adding reagents, and copper, nickel,
cadmium and cobalt are precipitated by adding zinc and arsenic
trioxide dust. The purified solution is processed in electro-
lytic cells and zinc is plated on inert aluminum cathodes. Sul-
phuric acid is regenerated and recirculated to the leach process.
The cathodes are periodically stripped, and the zinc sheets
melted and cast as slabs or other shapes.
In the pyrometallurgical process the roaster product is
sintered to volatilize and recover lead and cadmium, and produce
a suitable fused, porous feed for the retorts. Down-draft
sinter machines similar to those used in lead practice are
employed. Some plants sinter mixes of fresh concentrate and
recycle sinter on the grates without the use of roasters.
Different quality feeds are required for the three retort pro-
cesses and the sintering practices differ.
Pyrometallurgical zinc volatilization processes utilize
the relatively low boiling point of zinc, 906°C, to volatilize
zinc from a reduced mixture of sinter and coke. The zinc
vapour is shock cooled to prevent reoxidation. Three types of
retort are used; horizontal, vertical and electrothermic. The
Imperial Smelting Process relies on the same principles. The
retort product may be marketed directly or refined further in
distillation columns.
*
Horizontal retorts are the oldest reduction system pres-
ently in use. Many of the plants have been phased out, particu-
larly in North America and Europe owing to the high cost of
labour and the inability to control emissions. The plants are
only viable in areas where labour costs are low. The furnace
consists of a number of cylindrical retorts packed in a slotted
shell. The tubes contain a mixture of sinter and coke. Hot
combustion gas is passed through the furnace around the tubes
and zinc volatilized and condensed in a clay outer wall con-
denser. The operation is a batch process requiring about 48
hours. The process is not very attractive because of excessive
operating costs, low productivity, high labour demands for
tedious tasks (e.g., packing and stacking tubes), difficulties
in controlling particulate emissions from the numerous retorts,
and low zinc recoveries.
The vertical retort was developed by New Jersey Zinc in the
1920's. The process is semi-continuous. Hard or briquetted
sinter is charged to the shaft in slugs and moves downwards. The
charge is heated indirectly through an outer shell and zinc vola-
tilized and shock cooled in a splash condenser. The condenser
gas is cleaned in a wet scrubber and injected into the fuel
burner. Particulate emissions can be minimal with good operating
practices.
147
-------
The electrothermic process was developed by St. Joe
Mineral Corporation in 1936 and is in operation at Monaca (U.S.A.)
and Toyama (Japan). Preheated, hard sinter and coke are charged
and heated by electric arcs between electrodes in the upper and
lower shaft. Zinc oxide is reduced and zinc volatilized. Gases
are drawn off midway from the furnace and passed through a con-
denser full of liquid zinc. The exhaust is cleaned in a wet
scrubber and the carbon monoxide containing gas used as auxiliary
fuel.
The zinc products from the different retorts may be mar-
keted directly or refined in vertical fractional distillation
columns to produce zinc to meet different specifications. Prin-
cipal elements removed are cadmium and lead.
Particulate Pollution Problems and Abatement Technology
Particulate and metal vapour emissions can occur from zinc
production. Emissions from electrolytic refining are more
readily controlled since only one pyrometallurgical process
(roasting) is involved. Potential emissions from the processes
include suspended particulates, and lead, zinc, cadmium, arsenic
mercury, etc., which may be emitted as vapours or fine condensed
fumes. High-efficiency control devices are required.
The roaster emissions are effectively abated when S02
recovery is used. This is usually achieved using contact sul-
phuric acid plants. The pollutants are captured in combinations
of waste heat boilers, cyclones, electrostatic precipitators, wet
scrubbers, wet electrostatic precipitators, and demisting towers.
The presence of elevated levels of mercury in some concentrates
causes difficulties as it is absorbed in the acid product,
leading to marketing difficulties. Several remedies are avail-
able5; the mercury vapour may be removed from the particulate-
free gas prior to acid manufacture by dry absorbtion in packed
towers containing activated carbon or selenium; by wet scrubbing
with concentrated sulphuric acid or mercuric chloride solution,
or by H2S injection and collection of a mercuric sulphide pre-
cipitate in coke filters. Alternatively, the contaminated acid
may be treated to precipitate dissolved mercury as mercuric
iodide, or some other compound, and filtered out. Combinations
of pre-removal and acid cleaning may be used. Emissions are
greater from multi-hearth roasters when sulphur recovery is not
used. Some operations produce dilute S02 gases (2%) and these
are discharged following particulate recovery. The gases con-
tain fine particulates and fumes which would be captured in the
acid plant wet gas cleaning train.
Acid mist is generated during electrolytic refining and
the tankhouses have to be ventilated.
148
-------
Pyrometallurgical refining includes roasting and additional
high temperature operations, i.e., sintering, retorting, and
possibly distillation refining. Sintering is used to eliminate
volatile components which would contaminate the zinc product.
Large amounts of lead, cadmium, arsenic and suspended particu-
lates are evolved and high-efficiency particulate control
devices, usually fabric filters, have to be used.
Emissions from vertical and electrothermic retorts can be
readily controlled using conventional equipment.
Emissions from horizontal retorts are a serious problem.
A large number of retorts are required to achieve a reasonable
capacity and each retort is a potential emitter. The trend has
been to phase out the retorts becuase of the inability to install
satisfactory pollution controls.
Fugitive emissions can occur from the processes. These
may be minimized through the use of well designed facilities
and good operating, maintenance and housekeeping practices.
Ventilation systems are installed. Fugitive emissions are more
serious for pyrometallurgical refining because of the large
number of high-temperature processes involved. Fugitive emis-
sions are very serious from horizontal retorts.
ENVIRONMENTAL LEGISLATION
Various regulatory agencies have issued standards limiting
the emission of particulates and certain elements from lead and
zinc production, as summarized in Table 4. Other authorities,
e.g., some states in the U.S.A., limit the amount of particulates
that may be emitted. The limits are normally based on the pro-
duction rate raised to a fractional power.
Occupational hygiene and ambient air standards also re-
quire emissions to be controlled. The U.S. Occupational Health
and Safety Administration limit the permissible exposure limits
to lead and arsenic in the working environment to 50 ug/m3 and
10 ug/m3 respectively over a time weighted eight-hour day6'7.
This requires particulate and fume emissions to be minimized
and/or ventilation provided for operations where lead and
arsenic are involved.
The U.S. EPA requires lead emissions to be controlled to
meet a limit of 1.5 yg/nm3 at the plant boundary8. This may
require many sources in the plant to be controlled to meet the
ground level standard.
Indirect control over particulate emissions is effected
through limits placed on S02 emissions. Sulphur recovery
processes have to be installed to comply with the limits and
these require the particulates to be removed for efficient
operation.
149
-------
OUTLOOK
In order to meet the standards summarized in Table 4,
substantial changes can be expected, in the next decade, in the
processing of lead and zinc concentrates. For instance, weak
S02 gas-cleaning systems are expensive and the emphasis in the
future will be on producing higher-strength gases, which can
be treated in, for example, a sulphuric acid plant, in order to
meet S02 limits such as those set by EPA.
This will necessitate the use of continuous processes and,
in the case of lead will result in the increased use of commer-
cial oxygen. The application of the TBRC by Boliden, and the
development of the Kivcet process by the U.S.S.R. are examples
of this trend in the lead industry.
The Finnish company, Outokumpu Oy, has for a number of
years used a flash smelting process for treating copper and
nickel sulphide concentrates. At their research station in Pori,
Finland, this company has tested, on a pilot plant basis, the
application of their flash smelting process to lead sulphide
concentrates. It has been found that lead concentrates can be
successfully flash smelted but the dust circulation can be high
because of the high volatility of lead compounds under these
smelting conditions. Nevertheless, a renewed interest in flash
and similar smelting processes, applied to lead concentrates,
can be expected.
Cominco has done pilot and commercial scale tests on an
oxygen flash smelting process producing a soft bullion and a
PbO slag (Canadian Patent No. 934,968, issued Oct. 9, 1973).
The Quenau-Schumann-Lurgi (QSL) process is presently being
piloted in Germany and could be considered an emerging process.
Pelletized concentrates are charged to a reactor. Oxygen and
fuel, as required, are blown in through tuyeres mounted in the
bottom of the reactor. A relatively high strength off gas is
produced which is cooled in a waste heat boiler prior to feeding
a sulphur fixation plant. Slag and lead bullion are removed on
a continuous basis. The process is still in the early develop-
ment stages and will require a number of years before it can be
considered for commercial operation.
Increasing concern in the industry about the emission of
lead to the working environment and ambient air have recently
created an interest in hydrometallurgical processes. A number
150
-------
TABLE 4 SUMMARY OF PARTICULATE EMISSION STANDARDS BASED ON
CONTAINMENT AT SOURCE
Agency
US- EPA9
US - EPA9
US- EPA9
US- EPA9
_ 10
Japan
_ 11
Germany
United Kingdom
British Columbia
British Columbia
12
British Columbia
Limit
50 mg/sm
<650 ppm SO2
2
50 mg/sm
<650 ppm SO-
30 mg/nm
20 mg/nm
20 mg/nm3
229 mg/sm3
286 mg/sm
343 mg/sm3
Comments
Lead blast fee, reverb dross fee, sinter mach
discharge gas
Lead sinter machine off gas, (particulate
control implied)
Zinc sinter machines
Zinc roaster off gas, (particulate control
implied)
Lead & Zinc Smelting furnaces (20 mg/nm
in more polluted areas)
For primary lead plants
For large lead works
Objectives for new facilities
Objective for existing facilities
Short-term objective for existing
facilities
151
-------
have been proposed and tested on a limited basis, but none have
reached the stage of development of the TBRC and Kivcet processes.
The control of intermediate compounds and plant effluents are a
few of the problems which must be overcome before the hydro
approach can be considered a viable alternative to pyro processes.
The zinc industry has seen more process innovation in
recent years than the lead industry, owing in part to the higher
value that zinc has commanded historically in the marketplace.
This has permitted producers to carry out process development and
plant improvements. The electrolytic zinc process, with its
various final residue alternatives, removes sulphur as a high
strength roaster gas which permits the sulphur to be fixed as
sulphuric acid. Hydro-metallurgical techniques are then applied
to extract the values and precipitate impurities. This approach
minimizes particulate emissions, particularly in the form of
fume, improves the working environment and provides the type of
operation that is amenable to automation.
Developments, therefore, in the zinc industry in the near
future will be less spectacular than can be expected in the
lead industry. They will be directed towards refinements to the
existing processes. The general trend will be away from pyro-
metallurgical refining processes, except for so-called dirty
concentrates, or specialized zinc oxide products. It is reason-
able to assume, therefore, that the ISP will continue to have a
role in the zinc industry, but other pyrometallurgical processes
will be phased out as labour, fuel, and emission constraints
increase.
ACKNOWLEDGEMENTS
The authors wish to thank Environment Canada for permission
to submit this paper and J. H. Reimers and Associates for
providing technical assistance.
152
-------
REFERENCES
1. American Bureau of Metal Statistics. Non-Ferrous Metal
Data 1977.
2. J.C. Taylor. Occupational Health and Safety in Metallurgi-
cal Plants. Canadian Mining and Metallurgical Bulletin,
June 1978.
3. J.F. Castle and J.H. Richards. Lead Refining: current
technology and a new continuous process.
4. H. Rozovsky. Lead and the Environment at the Smelting
Division of Brunswick Mining and Smelting Corporation.
Cadadian Institute of Mining and Metallurgical Bulletin,
April 1974.
5. Environment Canada, Internal Report. A Review of Process
Technology to Contain Mercury from Non-Ferrous Smelter
Metallurgical Gases. Oct. 1976.
6. U.S. Federal Register, Department of Labour. Occupational
Exposure to .Inorganic Arsenic. May 5, 1978.
7. U.S. Federal Register. May 14, 1978, Department of Labour.
Occupational Exposure to Lead.
^
8. U.S. Federal Register, Oct. 5, 1978. Environmental Pro-
tection Agency Ambient Lead Air Quality Standard.
9. U.S. Federal Register. Standards of Performance for New
Copper, Lead and Zinc Smelters. Jan. 15, 1976.
10. U.S. EPA Evaluation of Status of Pollution Control and
Process Technology - Japanese Non Ferrous Metals Industry.
Contract No. 68-02-1375.
11. E.C.E. Task Force for the Development of Guidelines for the
Control of Emissions from Non-Ferrous Smelters. Draft Lead
Chapter.
12. Province of British Columbia, Canada. Pollution Control
Objectives for Smelting Industries. 1976.
153
-------
CONTROL OF PARTICULATE EMISSIONS FROM
THE IMPERIAL SMELTING ZINC-LEAD PROCESS
Roger W. Lee
Imperial Smelting Processes Limited
Avonmouth, Bristol BS11 9HP, England
ABSTRACT
A brief description is given of the Imperial Smelting Process for
producing zinc and lead in a blast furnace. Typical limits on particulate
emissions are given for some of the 11 countries where the process is used.
The gas cleaning problems met with in various parts of the process are
quantified and the methods used to meet the required discharge limits are
described. These methods comprise: the use of an electrostatic precipit-
ator or a venturi scrubber to clean acidic sinter plant process gas; the use
of a disintegrator wet washing system to clean blast furnace process gas;
the use of venturi scrubbers and bag filters to clean sinter and blast furnace
plant ventilation gases.
Possible trends in gas cleaning technology 'applied to the process are
discussed.
INTRODUCTION
The Imperial Smelting process for producing zinc and lead in a blast
furnace which was developed at Avonmouth, UK, is now in operation at 12 sites
in 11 countries. During 1977 a total of 613,000 tonnes of zinc and 282,000
tonnes of lead was produced by the process.
As shown in Fig.1 the first part of the process consists of the updraught
sintering of zinc and lead concentrates, returns sinter and fluxes. The
concentrates are mainly sulphides and sintering removes the majority of the
sulphur as sulphur dioxide and this is converted into sulphuric acid. The
product sinter has a typical analysis of 42 - 44% Zn, 18 - 20% Pb, 0.5 - 1.0%
Cu and 0.7 - 1% S and is delivered to the furnace bunkers crushed and screened
to be 25 - 100 mm in size.
154
-------
I CONCENTRATE STORAGE BINS
2 UPDRAUGHT SINTER MACHINE
3 SINTER BIN
4 COKE PREHEATER
5 LEAD - ZINC BLAST FURNACE
6 LEAD SPLASH CONDENSER
7 COOLING LAUNDER
8 ZINC - LEAD SEPARATION
9 GAS WASHING SYSTEM
10 PRIME WESTERN GRADE ZINC
II LEAD BULLION
12 DECOPPERIZED LEAD BULLION
13 SLAG
14 COWPER STOVES
GAS TO ACID PLANT
TO STOVES
ZINC TO CASTING
OR REFINING
Figure I. FLOWSHEET OF THE IMPERIAL SMELTING PROCESS.
155
-------
The Imperial Smelting Furnace (ISF) is charged with coke preheated to 800C
and hot sinter. Inside the furnace the coke is burnt at tuyere level
with air preheated in Cowper stoves to 900°C and the product CO reduces the
zinc and lead oxides in the sinter. Zinc is produced as a vapour at the
furnace temperatures and travels with the furnace gases into a condenser where
it is absorbed in a spray of liquid lead. Liquid zinc is then separated from
the lead in an external cooling circuit and cast as Prime Western grade zinc
or thermally refined to give Special High Grade zinc.
Metallic lead produced in the furnace is tapped periodically from its
bottom together with the slag formed from the gangue constituents of the
charge. Separation of the lead bullion from the slag is carried out in a
forehearth after which the slag is granulated. The lead bullion is either
cast into blocks or may be pyrometallurgically or electrolytically refined.
The particulate control systems installed on the process are concerned
with cleaning the sinter plant and ISF process gases and also the ventilation
gases which are extracted to maintain a suitable working atmosphere. The
target emissions to be met vary in the different countries where the process
is operated. Some examples are given in Table 1.
Table 1. TYPICAL EMISSION LIMITS FOR PARTICIPATES
Plant
Commonwealth Smelting Limited*
(CSL) Avonmouth, UK
Hachinohe Smelting Limited
Hachinohe, Japan
Sulphide Corporation Pty.Ltd.
Cockle Creek, Australia
Pb
mg/Nm3
12
10
20**
Cd
mg/Nm3
-
0.8
-
* CSL is also subject to mass emission limits of
5.4 kg/h for Pb and 0.7 kg/h for Cd
** Total of heavy metals (Pb, As, Sb, Cd, Hg)
Since most of the dust and fumes dealt with contain at least 25% lead,
most of these limits are at least as stringent as the EPA limit of 50 mg/Nm3
for total dust from primary non-ferrous smelters.
SINTER PLANT PROCESS GAS CLEANING
A typical zinc/lead sinter plant produces about 60,000 Nm3/hr (dry) of
gas at a temperature of 300-350 C and containing about 6% S02- The fume
loading is usually 30-50 g/Nm3 (i.e. up to 70 tpd) with a typical size range
156
-------
and analysis as"given in Table 2.
Table 2. SINTER PROCESS GAS FUME SIZE AND CHEMICAL ANALYSIS
Fume size range
7=
Fume analysis
: %
<0.2 y
2.3
Pb
60-62
0.2-1 y
78.8
Cd
3.5-4.5
1-2 y
8.8
Zn
1-2
2-5 y
2.9
Total S
10-11
>5 y
7.3
This gas has to be cleaned and dried before conversion of its S02 into 803
and the subsequent production of sulphuric acid. The first zinc/lead sinter
plants cleaned the gas with venturi scrubbers followed by a cooling tower and
electrostatic mist precipitators. Subsequently the hot electrostatic
precipitator has become preferred to venturi scrubbers.
Venturi Scrubbers
Venturi scrubbers were successfully used on the prototype updraught
sintering machines at Avonmouth and subsequently were applied to large scale
machines at 4 sites. At a pressure drop of 600 mm, a gas cleanliness of less
than 1 g/Nm3 is obtainable and this was originally considered adequate in view
of the fact that the gas was subsequently cooled and cleaned in aa irrigated
packed tower and mist precipitators. It has been found in practice that this
quantity of particulates has adversely affected the operations of the remainder
of the gas cleaning and cooling circuit. A relatively large bleed off from
the cooling tower and more frequent washdown of the mist precipitators have
been necessary. Even so, some problems with system blockages have occurred.
To obtain more efficient gas cleaning a higher pressure drop venturi scrubber
would be required. This would not only result in increased power costs but
could also have an adverse affect on the life of the fan. The medium head
fans used at present only have a life of about 2 years even when constructed
from 25% Cr 5% Ni Mo W steel.
Electrostatic Precipitators
Typical resistivity curves for fume from zinc/lead sintering are shown in
Fig.2. These curves indicate that electrostatic precipitation should be
carried out at a low temperature of say 100°C with saturated gas or at temp-
eratures in excess of 200°C. In the early 1960's the experience which was
available on electrostatic precipitation of fume from lead sinter plants had
been gained only at low temperatures. This system, which required the
installation of a spray conditioning tower ahead of the precipitator was
accordingly adopted on several zinc/lead sinter plants. The precipitator
had to be lined with acid resistant bricks and further protected with carbon
bricks if any fluorine risk existed. Even so the precipitator had a very
limited life.
157
-------
10
E 5
o
12
52 10
co w
UJ
ac
10
10
18% H20
11-8% H20
16-8% HgO
21 8% H20
50 100 150 200 250
TEMPERATURE, 0°C
Figure 2. ELECTRICAL RESISTIVITY OF SINTER
PROCESS GAS FUME.
300
158
-------
As more experience was gained in the operation of large scale updraught
zinc/lead sinter machines, it became apparent that sinter gas temperatures of
over 300°C could be consistently attained. It was therefore proposed by
Lurgi that a hot gas electrostatic precipitator should be used which would
considerably reduce the capital costs since an unlined steel cased structure
could be employed. Lurgi installed a pilot precipitator on a sinter plant
at Avonmouth in 1964 and satisfactorily demonstrated the suitability of a hot
gas precipitator for dealing with the fume. The first full scale precipit-
ator was subsequently commissioned in 1966 following which it has become
accepted as the preferred method of cleaning sinter plant process gas. Whilst
the precipitator has a high capital cost compared with a venturi scrubber,this
is outweighed by the following operating advantages:
(1) Low power consumption
(2) High efficiency; this is typically 99.9% giving an outlet dust
loading of 20-50 mg/Nm3.
(3) Few operating problems; such problems as have occurred have mainly
been concerned with inadequate electrode rapping and bridging above
dust screw conveyors. The latter problem does not appear to arise
when drag link conveyors are used.
(4) The overall gas cleaning/acid plant water system is considerably
simplified when an electrostatic precipitator is used.
SINTER PLANT VENTILATION GASES
The total ventilation volume on a typical zinc/lead sinter plant is in
excess of 150,000 Nm3/hr. This volume is devoted mainly to ventilating the
handling and cooling of product and returns sinter. As shown in Fig.3 the
steps involved in this handling system are as follows:
(1) Sinter cake is discharged from the tip end of the machine and
crushed by a pronged breaker to <250 mm.
(2) The sinter is fed to a spiked roll crusher which reduces its top
size to <100 mm.
(3) The sinter is screened at 25 mm and part of the oversize (30-35% of
the total machine output) is fed hot to the ISF charge bunkers.
(4) The remainder of the oversize is crushed in a corrugated roll
crusher to <25 mm and is then joined by the screen undersize and
the fines recycled from the ISF plant. This material is then
crushed to <6 mm by a smooth roll crusher.
(5) The crushed sinter is cooled in a rotary drum by recycled slurries
and water and is then transported to the sinter returns bunkers in
the feed preparation section*.
* Note that on some plants the sinter is cooled in the rotary drum before
crushing in the smooth roll crusher.
159
-------
NEW FEED
MATERIALS
MIXING DRUM
CONDITIONING DRUM
TO ACID PLANT
tui»
YYYYYYYYYYYW
FRESH AIR FANS
IGNITION
WINDBOX FAN
PRONGED
BREAKER
RECIRCULATING
FAN
•>
,
^
DEWATEREC
SLURRIES ,
SINTER
SCREEN
Tt
i
i
, SINTER
Bl Ibll^CTBO
i ' i DuniNC.no
-------
Tip End Ventilation Gas
This gas, the volume of which is normally about 50,000 Nm3/hr, is
characterized by containing S02- The exact quantity of this and the proport-
ion of fine particulates in the gas depend greatly on the extent to which
process gas is sucked into the tip end enclosure. By using heat resisting
baffles between the machine hood and the enclosure and avoiding pressurizing
the machine hood and by ensuring the completion of sintering before the end
of the machine, the quantity of fine particulates can be minimized and the
S02 kept to less than 0.1%.
The following methods have been employed to deal with this gas:
(1) Recirculation to the sinter machine; this requires either that the
fan used is specially protected against erosion or,that the bulk of
the particulates are removed. Recirculation suffers from the
disadvantage that the fresh air volume required by the sinter machine
is usually less than that required to ventilate the tip end.
(2) Cleaning with a venturi scrubber; this method removes.,particulates
whilst obtaining some S02 removal at the same time. .However, it
has been shown that whilst an outlet particulates loading of less
than 50 mg/Nm3 can be achieved with a venturi pressure/drop of about
900 mm w.g., removal of significant quantities of S02 is difficult.
Unless scrubbing water is used on a once through basis, which is not
a practical proposition, then S02 can only be removed efficiently by
the use of lime slurry dosed recirculated water. This leads to
considerable problems with blockages of recycle liquor lines caused
by the combined effects of precipitated calcium sulphate/sulphite
and the solids scrubbed from the gas.
(3) Removal of particulates by bag filter followed by S02 removal by
venturi scrubbers; this method has been successfully practised by
Hachinohe Smelting in Japan. Data for their bag filter system
which handles 54,000 Nm3/hr is given in Table 3.
Table 3. BAG FILTER FOR TIP END GAS AT HACHINOHE
Temperature °C
Mean Dust Loading mg/Nm3
Gas Cooler Inlet
Outlet
Bag Filter Inlet
Outlet
140-170
85-100
81-95
76-85
7 000
1 000
1 000
0.2
Filter Type
Bag Material
Filtration Velocity
gross
net
Mechanically shaken and reverse air blown
Acrylic needlefelt
0.895 m/min
0.805 m/min
161
-------
The venturi scrubbers which follow the bag filter have a pressure drop of
810 mm w. g. and reduce the SC>2 content from 1500 to 200 ppm.
Ventilation Gases from Sinter Handling Systems
These gases generally contain a high particulates loading of over 35g/Nm3
but the particulates are relatively large as indicated in Table 4.
Table 4. TYPICAL SINTER VENTILATION DUST SIZE
Fume Size Range
7
/o
<2.5 y.
7.7
2.5-5y
6.1
5-10 y
9.7
10-20 y
13.7
20-30 y
13.6
30-43 y
23.1
>43y
26.1
Three methods of cleaning this type of dust have been employed:
(1) Impingement scrubbers; scrubbers of the Doyle type were favoured
initially for the local ventilation and gas cleaning of individual
dust emission points. However, apart from the problems of
maintaining a large number of widely separated scrubbers, it was
found that these scrubbers were not really suitable for handling
high dust loadings. Whilst relatively high scrubbing efficiency
could be obtained, daily cleaning of the scrubber was found to be
necessary.
(2) Venturi scrubbers; relatively low pressure drop venturi scrubbers
have been found to be very efficient in cleaning gases laden with
large quantities of coarse dust and are preferred to impingement
scrubbers. One example from CSL's plant at Avonmouth is the use
of a saturator tower and a 350 mm pressure drop venturi scrubber to
clean 35 000 Nm3/hr of ventilation gases from the product sinter
screen and the corrugated roll crusher. The results obtained are
given in Table 5.
Table 5. PERFORMANCE OF VENTURI SCRUBBER AT AVONMOUTH
Inlet gas
!
Outlet gas
Total Dust
mg/Nm3
63 200
Pb
mg/Nm3
16 190
2.1
Zn
mg/Nm3
22 570
1.1
Cd
mg/Nm3
430
0.16
Venturi scrubbers are also used to clean the gases from the cooling
drum. These gases are relatively saturated and the dust burden
can contain a substantial proportion of small particles. A 900 mm
pressure drop scrubber has been used successfully on this
application.
162
-------
(3) Bag filters; although low head venturi scrubbers are still an
economic gas cleaning method, bag filters are used at some sites and
are becoming preferred for dry gases since the only liquor circuit
possibly involved is that concerned with slurried filter dusts. The
majority of bag filters installed have been of the mechanically
shaken reverse air blown type. Rozovsky1 has suggested that pulse
jet filters are unsuitable for this type of application because of
the high dust loading. However, it is considered to be quite
feasible to use such a filter in conjunction with a cyclone.
ISF PROCESS GAS
An ISF with a shaft area of 17.2 m2 produces about 45,000 Nm3/h of gas
which leaves the lead splash condenser at a temperature of 440-450°C. The
gas usually contains 18-23% CO, 9-14% C02, 1% H2 with the balance being
nitrogen. The particulates loading is from 20-30 g/Nm3 with a typical size
range and analysis as given in Table 6:
Table 6. ISF GAS PARTICULATES
Size range
%
Analysis
%
< 3 y
24.3
Pb
30-45
3-5 y
14.4
Zn
30-45
5-10 y
20.4
Cd
0.5-2.0
10-20 y
18.4
S
1-3
> 20 y
22.5
Si02
1.5-2.5
The basic gas cleaning system adopted on all operating furnaces is shown
in Fig.4. It consists of a gas washing/cooling tower followed by a dis-
integrator and a cyclonic moisture separator. A wet gas cleaning system was
selected originally because of the explosion hazard with the gas and because
of the accreting properties of the particulates, which contain molten lead.
An unpacked gas washing tower is used and at most sites has a central
top entry. Two rings of sprays are mounted in the roof of the tower around
the gas entry and these are supplied with about 200 m3/h of recirculated
water. The use of this relatively inefficient co-current tower has been
dictated by the need to minimize the distance between the condenser offtake
and the tower inlet and to avoid any wet/dry transition zone. Tangential
entry towers are used by two operators and probably result in better gas
cleaning but suffer from accretion build up in the entry duct.
The gas emerging from the bottom of the tower is normally at 40-50 C and
contains 5-10 g/Nm3 of particulates. The gas is fed via irrigated ducts
into the inlet legs of the disintegrator. All operators have used disinte-
grators made by Theisen G.m.b.H of West Germany except the two Japanese plants
which have employed disintegrators made by a former Theisen licensee, Ebara
Manufacturing Co. of Japan.
163
-------
WATER SUPPLY
210 m 3/hr
WATER SUPPLY
80m3/hr
0 2700
MOISTURE
SEPARATOR
DISINTEGRATOR
/
0 4000
GAS WASHING
TOWER
INLET GAS
Nm3/hr
< AT 450° C
o
ox
10
Figure 4. TYPICAL ARRANGEMENT OF BASIC GAS
WASHING SYSTEM
164
-------
A disintegrator is virtually a complicated irrigated fan as can be
seen from Fig. 5 which shows the rotor. Water and gas are admitted to
the centre of the disintegrator and then pass radially outwards through a
series of rings of fixed and rotating bars. The rings of rotating
bars are mounted on a rotor which usually revolves at 740 rpm although one
plant uses a 960 rpm motor. The high relative velocity between the rotating
and static bars smashes the water into very small droplets.
High relative velocities also exist between these water droplets and the
particulates and result in good gas cleaning by inertial impaction. The
particulates loading of the gas after the cyclonic separator is typically
20-30 mg/Nm3. At three ISF plants the gas is given further cleaning by a
second disintegrator and at another plant a moving ball scrubber is used.
Secondary scrubbing results in a particulates loading of 5-15 mg/Nm3.
The dirty water from the gas washing tower, the disintegrator and the
separator is pumped to a thickener and about 75% of it is recycled to the gas
washing tower. The underflow from the thickener is filtered and the cake
recycled on the sinter plant.
The primary disintegrator has a power consumption of 7-8 kWh/1000 m3 of
gas and a water requirement of 1.2 m3/lOOO m3. This relatively high power
consumption has led to investigations into alternative gas cleaning methods.
However, since a wet system is considered necessary then the only alternative
would appear to be a venturi scrubber. Some pilot plant work with a venturi
scrubber was carried out at the Berzelius ISF in West Germany and gave the
results shown in Table 7:
Table 7. VENTURI PILOT SCRUBBER RESULTS
Venturi Pressure Drop
mm w. g.
800 - 1000
1100 - 1200
1500 - 1700
1600
1600
Inlet Dust Loading
mg/Nm3
2000 - 3000
2000 - 3000
2000 - 3000
2000 - 3000
5000 - 6000
Outlet Dust Loading
mg/Nm3
80 - 100
30 - 40
20 - 25
23
36
From these results it was concluded that to equal the performance of the
disintegrator a venturi pressure drop of over 2000 mm w.g. would be necessary.
Even when using a relatively efficient fan, the power required to produce such
a pressure drop would have been similar to that required by the disintegrator
and there was therefore no justification for changing from its use.
165
-------
Figure 5. ROTOR OF A THEISEN DISINTEGRATOR.
-------
ISF PLANT VENTILATION GASES
On a typical ISF plant a total volume of about 300,000 Nm3/h of vent-
ilation gases is extracted. Part of this volume is concerned with vent-
ilating the processes of feeding and screening sinter and coke in the charge
preparation section. These ventilation systems at Avonmouth have been
described in detail by Senn2. The particulates dealt with by the gas
cleaning equipment are basically similar to those dealt with in the sinter
plant sinter handling systems and the same remarks about gas cleaning systems
apply.
„ The remainder of the ISF ventilation systems are largely concerned with
molten metal and slag handling systems.
Zinc Separation Systems
In this system, lead at 550°C containing about 2.3% zinc is pumped from
the condenser into a refractory lined launder where it is cooled by immersible
water-cooled steel coolers. The temperature of the lead is progressively
reduced to about 440°C as it travels along the launder with the effect that
the solubility of zinc is reduced and molten zinc separates from the lead.
At the end of the launder the lead/zinc liquid mixture flows into a flux bath
and then into a large bath where the separation process continues in quiescent
conditions. Zinc is then recovered by overflowing into a holding bath whilst
lead returns to the condenser via an underflow weir.
Since the temperatures in this system are relatively low,'metallic fume
formation is not a major problem.-. The fume extraction system is primarily
applied to ventilate positions where zinc and lead drosses are removed and
also to remove products of combustion from the separation baths. However,
the removal of ammonium chloride fume from the flux bath is an important
ventilation requirement.
It has been found that a venturi scrubber with a pressure drop of 450 mm
w.g. has enabled satisfactory discharges to be maintained in this relatively
lightly loaded system. Corrosion of the ductwork after the scrubber caused
by ammonium chloride and 'its dissociation products is a problem. Several
ISF plants have installed bag filters instead of scrubbers to clean these gases.
Slag and Lead Bullion System
Slag and lead are tapped from the bottom of the furnace at intervals of
H ~ lj hours and flow for 20 - 30 minutes. The slag, at temperatures of up
to 1300°C, flows with the lead into a forehearth. Separation takes place by
gravity with the slag overflowing and being granulated by water whilst the
lead syphons out of the forehearth via an underflow weir and is then run into
a ladle. A typical arrangement of this area is shown in Fig.6.
The cleaning of the ventilation gases from this area has been described
in some detail by Lee and Coy3. It was shown at Avonmouth that the 600 mm
w.g. pressure drop venturi scrubbers originally installed could only obtain a
gas cleaning efficiency of 92-95% compared with the level of 99% which was
167
-------
FIGURE 6. SLAGGING FLOOR VENTILATION AT AVONMOUTH
168
-------
required to meet the discharge regulations. Size analysis indicated that
over 14% of the fume was sub-micron in size. It can be seen from Fig.7 that
the majority of the individual particles are sub-micron in size but may form
agglomerates.
It was concluded that to clean the gas satisfactorily either a venturi
scrubber with a pressure drop in the region of 2000 mm w.g. was necessary or
that a bag filter should be installed. After some trial work a pulse jet
filter was installed as illustrated by the .photograph in Fig.8. This unit is
equipped with envelope bags arranged in four tiers and made from 645 g/m2
terylene needlefelt. The filter, which is pulsed on line, has a specific
filtration velocity of 1.33 m/min . The fume collected by the filter is
slurried and recycled to the sinter machine via an existing dewatering system.
Results obtained with the filter are given in Table 8:
Table 8. SLAGGING FLOOR FILTER PERFORMANCE AT AVONMOUTH
Volume
handled Position
Nm3/h sampled
41,850 Inlet
Outlet
: 41,120 Inlet
Outlet
, j
Fume Concentration mg/Nm
Total
8,118
-
7,046
-
Pb
3,555
0.40
1,685
0.23
Zn
2,977
0.35
4,322
0.37
Cd
5.5
0.03
2.7
0.17
As
28.9
0.13
17 I
0.008
There has been a general trend by ISF operators towards the use of bag
filtration for cleaning slagging floor fumes. However the use of a venturi
scrubber is still necessary for dealing with the fairly saturated fume pro-
duced when the slag is granulated with water.
FUTURE DEVELOPMENTS IN GAS CLEANING TECHNOLOGY
In common with the other non-ferrous metals extraction processes the
Imperial Smelting process is being subjected to increasingly stringent
environmental regulations. Since these regulations cover the working
atmosphere, gaseous discharges and liquid effluents the following effects can
be anticipated:
(1) To meet tighter threshold limit values for various pollutants
increased ventilation volumes will be required despite efforts to
improve process systems and to use more effective hooding.
(2) Gases will require cleaning to higher levels of efficiency.
Removal of sub-micron sized fume particles will become increasingly
important.
169
-------
Figure 7. SLAGGING FLOOR FUME (Photo courtesy of
Skandiafelt A.B..Sweden)
Figure 8. BAG FILTER PLANT AT AVONMOUTH
170
-------
(3) The trend towards the use of the bag filter for gas cleaning will
continue:
(a) It is the most effective method on sub-micron sized fume.
(b) It presents no liquid effluent treatment problems.
(c) Maintenance problems have been reduced by improvements in
filter design and bag materials.
(d) There would appear to be further scope for improvements in
filter design and bag materials which would result in reduction
in the capital costs of filters.
(4) Venturi scrubbers will continue to be used extensively particularly
where:
(a) Moisture laden gases are to be cleaned.
(b) Removal of SC>2 is a requirement.
ACKNOWLEDGEMENTS
The author gratefully acknowledges the co-operation received from the
following companies during the preparation of this paper:
"Berzelius" Metallhutten Gesellschaft m.b.H Duisburg, West Germany.
Commonwealth Smelting Limited, Avonmouth, UK.
Hachinohe Smelting Co. Limited, Hachinohe City, Japan.
Theisen G.m.b.H Munchen, West Germany.
REFERENCES
Rozovsky, H. Lead and the Environment at the Smelting Division of Brunswick
Mining and Smelting Corporation, Bathurst, N.B. Canadian Institution of
Metallurgists Bulletin. April 1974, p.50-60.
Senn, H. Improvement of Ventilation and Scrubbing Systems in Metallurgical
Industry. In: Minerals and the Environment, Jones M.J. (ed). Institution
of Mining and Metallurgy Symposium, London 1974. p.477-496.
Lee, R.W. and Coy, C.M. Cleaning Fume from Zinc Lead Smelting, Filtration
and Separation. May/June 1978. p.197-203, 230.
171
-------
PARTICULATE CONTROL AT THE PORT PIRIE LEAD SMELTER
John D. Martin
The Broken Hill Associated Smelters Pty. Ltd.
Port Pirie, South Australia 5540
ABSTRACT
A lead smelter has operated at Port Pirie since 1889 to treat the rich
lead-silver-zinc ore mined at Broken Hill. The present production capacity
is in excess of 200,000 tonnes per annum, using a conventional updraught
sintering machine, blast furnace and lead refining units. Slag fuming fur-
naces and an electrolytic zinc plant were installed in 1967 to recover the
zinc from the lead blast furnace slag. Over the years, considerable devel-
opment has taken place in improving the control of particulate emissions
from the smelter, and installed collection equipment includes both fabric
filters and wet scrubbers B.H.A.S. experience with the various types of
collectors is discussed together with current views on equipment suitable
for a lead smelter.
INTRODUCTION
The development of the lead and zinc industry at Port Pirie, South
Australia, dates from the discovery of the rich lead-silver-zinc mining field
at Broken Hill in the state of New South Wales, in 1883. As the nearest sea
port to Broken Hill, Port Pirie provided a natural outlet for the products of
the field and following the erection of a lead smelting plant in 1889, has
continued to treat the major share of the lead concentrates produced. Port
Pirie is situated on the eastern shore near the head of Spencer Gulf in South
Australia, approximately 230 kilometres from the State Capital, Adelaide and
400 kilometres from the inland city of Broken Hill.
High grade lead sulphide concentrates containing about 75 percent lead
provide 85-90 percent of the input materials to the plant while the balance
is made up of purchased scrap battery plates and other minor lead bearing
materials. The lead concentrates contain approximately 3.6 percent zinc
which is collected in the lead blast furnace slag. By 1967 some 6 million
tonnes of slag containing around one million tonnes of zinc had accumulated
in the slag dump and a treatment plant was then built to recover this zinc.
The annual production capacities of the main sections of the plant are
as follows:-
Lead 230 000 tonnes
Silver 280 000 kilograms
Zinc 45 000 tonnes
Sulphuric Acid 90 000 tonnes
Other by-products include gold, cadmium, copper matte and antimonial lead
alloys.
Particulate emissions from pyrometallurgical processes are substantial
and the combination of small particle size and high carrier gas temperatures
172
-------
make collection difficult. In addition the toxicity of lead bearing materials
to man is well known and the ever decreasing emission limits being imposed by
environmental authorities makes it mandatory to utilise the most efficient
collection equipment available.
SOURCE AND TYPE OF EMISSIONS
The major particulate emissions in the lead smelter arise from the sinter-
ing, blast furnacing and refinery processes and are generated in sufficient
quantities to be economically recovered. While the gaseous constituents of
these emissions vary considerably between processes, the particulates have the
common property of being very finely divided with a significant proportion in
the fume category. Attempts have been made to conduct sizing analyses on the
different fumes but some difficulty has been experienced due to agglomerates
breaking up on impact. However, meaningful figures for blast furnace fume and
fume from dross kettles etc. have been obtained which showed that the fume was
almost entirely sub-micron size confirming data obtained elsewhere that high
temperature smelting processes generate mainly sub-micron fume with a lead
content in excess of 60 percent.
The sinter plant process gases are divided into two streams, the smaller
fraction containing up to 7 percent sulphur dioxide being delivered to the
acid plant while the balance containing up to 0.7 percent sulphur dioxide is
at present filtered in two baghouses. High pressure water sprays are used to
control the temperature at 200°C in the machine hood and the gas temperature
at the baghouse inlet is maintained at about 100°C. A water dew point of up
to 50°C has been measured and any condensate is acidic and severely corrosive.
A dry electrostatic dust precipitator followed by wet electrostatic mist
precipitators are used in the high sulphur dioxide stream going to the acid
plant. Dust loadings vary up to 16 grams per cubic metre of gas.
Blast furnace gases are induced into an overhead hood and together with
other process and hygiene gases from the copper removal stage in the refining
operation are filtered in a baghouse. The sulphur dioxide content of blast
furnace gases is about 0.25 percent and the carbon monoxide level can reach 1
percent. A lower dew point in the 15 to 20°C range means that less corrosion
problems_are experienced in this baghouse. However, the fume will at times
ignite at about 140°C and problems have been encountered with small smoulder-
ing pockets of fume in the collecting launders which have been caused by a
hot screw conveyor bearing. Dust loadings in blast furnace gases can be up to
6 grams per cubic metre but there is a drop out along the main smelter flue
which is a low velocity regime. Refining process gases are also handled in
the blast furnace baghouse, but they consist mainly of oxide fume with little
moisture or sulphur dioxide present.
(
A fume containing up to 50 percent arsenic, 10 percent lead and 20 percent
antimony is volatilised in a reduction process at the lead refinery and is
collected in a small baghouse. The volume of gas handled is only about 750
cubic metres per minute, but extra precautions are taken to protect the health
of operating personnel.
Hygiene gases draughted from belt conveyors, transfer points and crushing
rolls in the sinter plant have caused problems due to the wet sticky nature of
173
-------
the particulates. Lead dust is very difficult to keep in suspension if being
transported in a wet gas as it tends to settle out in the duct system, causing
flow restrictions. A wet scrubber is used to collect the particulates under
these conditions and the collection units must be installed close to the
emission source.
Particulate collection is also important in our zinc division, and a
baghouse is used to recover the zinc oxide fume liberated in the slag fuming
furnaces. The baghouse is similar to our lead division process baghouses
except that higher gas temperatures up to 230° are permitted by the use of
Nomex filter fabric. Dust loading is high at 65 grams per cubic metre.
Emissions from the zinc oxide roasting kilns have caused problems as
the high fluorine and chlorine content of the gases precludes the use of a
baghouse in this location. The particulate loading is about 25 grams per
cubic metre and it is a sticky fume causing rapid build up in the flue system.
Two short throat venturi scrubbers in series are used to collect the particu-
lates but emission levels are still unsatisfactory.
A kiln is also used to dry the residual material obtained after
leaching out the zinc from the zinc oxide fume and the emission control in
this unit is unsatisfactory. The gases are hot, moist and contain a high
proportion of sub-micron particulates. A low energy microdyne wet scrubber is
installed, but it is inadequate for the duty.
Electric furnaces are used to melt the cathode zinc produced in the
electrolytic plant and problems were encountered with the fume emissions
resulting from the use of ammonium chloride as a flux. Tests to control
emissions with a microdyne wet scrubber were unsuccessful and a reverse pulse
small baghouse was finally installed.
Small baghouses are installed to collect the product produced in the
zinc dust furnace and the zinc oxide dust recovered in the zinc dross treatment
plant but these are standard installation with no particular problems. Exist-
ing sources of particulate emissions together with the type of collecter
installed are summarised in Table 1.
PARTICULATE CONTROL AT B.H.A.S.
The regulatory authority responsible for controlling gaseous emissions
in the state of South Australia is the S.A. Health Commission, and control of
particulate emissions is achieved through the Clean Air Regulations, 1972
under the Health Act, 1935-71. These regulations took effect on 1st January.
1973 but included provision to obtain exemptions at the discretion of the
authority. B.H.A.S. meets the requirements in most cases but there are still
areas of concern where exemptions are being obtained on an annual basis.
Control standards are established at the emission source and standards
of concentration of air impurities are expressed as the amount of such impurity
in each cubic metre of effluent gas before admixture with air, smoke or other
gases. The regulations applying to the smelter operation at Port Pirie specify
the following emission concentration limits on a dry gas basis at 0°C and
760 millimetres pressure.
174
-------
Particulates - not to exceed 110 milligrams in each cubic metre of
effluent gas.
Heavy Metals - total mass of copper, lead, arsenic, antimony, cadmium
mercury or any of their compounds shall not exceed 20
milligrams in each cubic metre of effluent gas in existing
plant or 10 milligrams in each cubic metre of effluent gas
for all new plant additions.
General Philosophy
The first filtering device at the Port Pirie smelter was a baghouse
installed in 1907 and since this time there has been a progressive increase
in the use of bag filters for gas cleaning. We consider fabric filters to be
a most effective means of capturing lead fume from pyre-metallurgical activities
and providing a reasonable bag maintenance programme is maintained there is no
problem in meeting the Clean Air Regulation standards. The policy is to use
dry filtration equipment on the Works wherever possible.
Our experience with wet scrubbers has not been encouraging but the prob-
lems have been caused more by the difficult nature of the dust handled than by
the collection unit. In the lead smelter, wet scrubbers have been used mainly
to treat wet sticky gases in the sinter plant, and the sticky nature of the
collected particulates causes operating problems. The solids have accumulated
in the ducts and scrubbers to the point where draughting is reduced and the
plant has to be shut down for cleaning. However, up to the present time we
have been using low energy scrubbers not suited to the type of dust handled
and recently-installed venturi scrubbers are performing better. Emission
standards are being met and cleaning problems are reduced significantly. A
different problem exists at the zinc oxide roasting kiln of the zinc plant
where the installed high energy venturi scrubber is not meeting emission
standards. Efforts are being made to improve this performance and a longer
throat venturi will be tried in this position.
An electrostatic precipitator is installed to collect the majority of
particulates contained in the sinter plant gases used for sulphuric acid
production, but the collection efficiency of 93 percent is too low for general
use in other plants.
Baghouses
itOX
There are many factors to be considered when designing a baghouse for
handling large volumes of dirty process gases but the operating principles are
relatively simple. Many plants have developed their own design from practical
experience acquired over many years and this is certainly the case at Port
Pirie. We are at present constructing a centralised baghouse system to handle
most of the lead plant process gases, as part of an environmental project
costing about $23 million. The project revolves around the construction of a
205 metre tall stack designed to provide sufficient dispersal of the residual
sulphur dioxide in effluent gases to meet the Clean Air Regulation limit of a
maximum ground level concentration of 20 parts sulphur dioxide per hundred
million of air over a period of 3 minutes. A diagram showing the general gas
flow arrangement is shown in Figure 1 and the project is expected to be
operational in July, 1979.
175
-------
BLAST
FURNACE^
DROSSING
PLANT
F1
FROM
SILVER YARD
REFINERY
CLEAN GAS
REFINERY
Figure 1 — General Gas Flow Diagram
176
-------
Table 1. SOURCE OF EMISSIONS AND TYPE OF COLLECTOR INSTALLED
Source of Emission
Blast furnace and
refinery process gases
Sinter plant gases
(lower sulphur
dioxide)
Sinter Plant gases
(high sulphur
dioxide)
Sinter plant hygiene
gases
Lead refining
process gases
(high arsenic)
Slag fuming process
gases
Zinc Oxide roasting
kiln gases
.Leach residue drying
kiln gases
Zinc melting furnace
gases
Zinc dust manufacture
gases
Zinc dross treatment
gases
Vol.
Handled
nm-Vmin •
6770
4630
680
180
750
3070
240
130
350
110
450
Type of Collector
Bag filter
(mechanical shaking)
Bag filter
(mechanical shaking)
Dry electrostatic
precipitator
Venturi Scrubber
(medium energy)
Bag filter
(reverse air pulse)
Bag filter
(mechanical shaking)
Venturi Scrubber
(high energy)
Microdyne scrubber
(low energy)
Bag filter
(reverse air pulse)
Bag filter
(mechanical shaking)
Bag filter
(mechanical shaking)
Parti culate
Loading
g/m3
2.7
4.0
16.0
1.4
1.8
65.0
25.0
4.0
0 - 7.0
(Variable)
3.3
10.9
Collect.
%
Effic.
99.3
99.5
93.0
99.8
99.2
99.9
98.8
91.0
Not
Calculated
Not
Calculated
Not
Calculated
The design parameters applied to the new baghouse under construction are
discussed in succeeding paragraphs and they incorporate the factors we feel
are important in designing a baghouse system for filtering and recovering
particulates from a lead smelter process gas. A sketch of a typical baghouse
chamber layout incorporating a mechanical shaking arrangement is shown in
Figure 2. We also have several smaller packaged baghouses using mechanical
or reverse jet pulse shaking mechanisms and these will be discussed separately.
General Design - A standard compartment (Chamber) size of 17.5 metres x 4.5
metres x 11.5 metres high has been adopted for the lead smelter in which it is
possible to hang 530 filter bags, 229 millimetres in diameter by 8.84 metres
177
-------
OUTLET
FLUE
INLET
FLUE
»->->»--^Tir-irTrTr^f-ir-^-r
BACKDRAFT
FLUE
CLOTH
BAGS —
HOPPERS
4
A
T
CLOTH
BAGS
SCREW
CONVEYOR
OUTLET
-FLUE
Figure 2 — Typical Baghouse Chamber Layout
178
-------
long. Each bag has an effective filter area of 6.2 square metres, giving a
total filtering area of 3286 square metres per chamber.
There will be thirteen such chambers in the centralised baghouse system
in two separate buildings. One building (existing) comprises eight chambers
while a new building of five chambers is being built, with provision for
extension later if required. The new building will be cast in concrete as
opposed to mainly brick construction used in previous baghouses. A mastic
material called Stackfas has been thickly sprayed on to the inner walls to
reduce the possibility of corrosion by acidic condensate which may form when a
chamber is taken off line for checking. Corrosion does occur at times in the
present sinter plant baghouses if all openings are not sealed properly and cold
outside air is allowed to enter the chamber. The temperature of gases from the
sinter plant off-take will be raised to a maximum of 350°C and the moisture
content lowered so that the mixed sinter plant - blast furnace gases will enter
the baghouses at 110°C - 120°C and this should give less condensation problems.
Close attention has been given to reducing frictional losses and maintain-
ing streamline gas flow wherever possible in designing the new ducting and
damper system, as tests have shown that in the past much power has been wasted
in overcoming unnecessary resistance from these factors.
The fume dislodged from the filter bags falls into vee shaped hoppers at
the bottom of the chambers from where it is wormed out by screw conveyors set
in the hopper. It discharges into another collector screw conveyor outside
the chamber and is eventually mixed with cadmium plant filtrate in a sump and
returned to the sinter plant circuit.
A back draughting offtake is connected into the inlet duct of each chamber
on the baghouse side of the inlet damper. Its purpose is to draught away any
gas which may leak past a closed inlet damper while operators are working in a
chamber. The sulphur dioxide and carbon monoxide content of smelter gas are
sufficiently high to warrant this safety precaution.
Balanced Draughting System - A balanced draught will be maintained over the
baghouse in the new sytem. This means that two sets of fans will be required,
one delivering dirty gas to the baghouse and the other taking the clean gas
after filtration. The aim will be to maintain the null pressure point in the
bottom hoppers of the chambers where the dirty gas enters.
The decision to install a balanced draught system was not easy and was not
unanimous, but it followed a recommendation from the Fume Superintendent after
he had studied the practice of other overseas plants and compared it with our
own experience. Our present blast furnace baghouse works under balanced draught
conditions and apart from the inconvenience of having to clean the dirty side
fan blades about once each month, is regarded as satisfactory. On the other
hand, the existing sinter plant baghouses work under higher suction (induced
draught) conditions and serious corrosion problems have been encountered in
the chambers due to entry of cold false air through poorly sealed doors and
openings. The high dew point and sulphur dioxide content of sinter plant gases
contributes to the problem. False air leakage results in a greater volume of
gas to be. handled by the fan in an induced draught system and to some extent
179
-------
will counterbalance the extra power requirements for a balanced draught
arrangement.
Two fans in parallel will handle the dirty gases from the smelter flue
into the baghouses and they will be controlled automatically to maintain a set
suction level in the flue. They are centrifugal fans each designed to handle
200 cubic metres per second of air at 352 revolutions per minute shaft speed
and driven by a variable speed 500 kilowatt electric motor. The fans on the
clean side of the baghouse will each be axial flow type designed to handle 210
cubic metres per second of air at 735 revolutions per minute shaft speed and
driven by a variable speed 850 kilowatt electric motor. Suction on the bag-
house outlet will control these fans and there will be a feed forward control
system link between the fans on the inlet and outlet sides of the baghouses.
An emergency diesel generator will be installed on the axial flow fans
to provide minimum draughting in the event of a general power supply failure.
Air to Cloth Ratio - For baghouses using a mechanical bag shaking system, an
air to cloth ratio (filter velocity) of 5 millimetres per second has been used
at Port Pirie in the past, but testwork has shown that this can be increased to
9 millimetres per second without increasing particulate emission levels beyond
the regulation limit. The higher figure has been used in the centralised
baghouse system for an expected average pressure drop across the bags of about
80 millimetres water gauge.
A higher air to cloth ratio of 12 millimetres per second has been found
satisfactory in small prefabricated baghouses using a reverse pulse of com-
pressed air for cleaning the bags.
Filter Cloth - The original filter fabric used at Port Pirie was a plain weave
woollen cloth but over the intervening years tests have been conducted in
operating baghouses trying out different types of cloth, different weaves and
the effect of a raised surface on one side of the cloth. As a result of these
practical tests it was found that polyacrylic cloth performed best in the
sintering plant baghouses where the sulphur dioxide and moisture content of
the gas is highest. A bag life of approximately 3 years is obtained at present
in this location. The specifications of the cloth in use (Dralon, Orion,
Microtain) fall within the following limits.
Weave -2x2 twill
Weight - 187 to 250 grams per square metre
Permeability - 14 to 17 cubic metres per square metre per minute at
25 millimetres water gauge pressure.
Polyester cloth is preferred in the blast furnace baghouse due to its slightly
better abrasion resistance and lower cost, and will be used in the centralised
baghouses handling the mixed gases. These bags have a life of approximately
7 years in the present baghouse. The terylene cloth used has the following
specification -
Weave - Irregular twill
Weight - 224 grams per square metre
Permeability - 16 cubic metres per squ'are metre at 25 millimetres
water gauge pressure.
180
-------
Finish - Scour and heat set.
Nomex cloth is used in the slag fuming baghouse due to its ability to
withstand^emperatures up to 230°C. The normal operating gas temperatures are
about 200 C with odd peaks near the maximum. Apart from temperature
considerations, these bags are subjected to more frequent mechanical shaking
than the other baghouses and a bag life of just over 2 years is normal. This
shorter life combined with a cloth cost four times that of terylene raises
doubts on the economics of operating baghouses at higher temperatures instead
of increasing the gas cooling facilities. The specifications of the nomex
cloth used are as follows -
Weave -2x2 twill
Weight - 300 grams per square metre
Permeability - 16 cubic metres per square metre per minute at 25
millimetres water gauge pressure.
Shaking Arrangement - Dust build up on the cloth in our process gas baghouses
is removed by means of mechanical shaking applied to the top section of the
bags. Various types of shaking arrangements have been tried over the years,
but we have standardised on hanging the bags to a moveable top frame which is
given a sideways rocking movement through an electrically driven crank disc.
The length of stroke is fixed at 178 mm and the motor speed is 92 revolutions
per minute. Shaking time may be varied but it is kept to a minimum to reduce
bag wear and is presently controlled at 10 seconds shake every 4 hours. The
chamber being shaken is isolated for 15 seconds before the shake commences and
120 seconds after the shake cycle to allow dislodged dust to settle into the
screw conveyor troughs.
The new centralised baghouse will be on automatic control for shaking and
it will be possible to vary the cycles as required. Some of the chambers in
the existing blast furnace baghouse have been converted to automatic control
and a typical chart indicating gas temperature, gas volume and pressure drop
across the chamber is shown in Figure 3. The step trace in the pressure drop
and gas volume patterns reflects the changes as other chambers are shaken in
sequence. It is hoped that a much better control can be exercised over our
baghouse operations by having these recording facilities.
Small Prefabricated Baghouses - Several smaller type packaged baghouses are
in use on the Works and we are generally satisfied with their operation. Our
first baghouse using terylene needlefelt cloth and employing a reverse air
pulse system of bag cleaning was installed in 1975, to collect an arsenic-
bearing fume arising from our refining operations. It has performed to
expectations and confirmed that the higher filter velocity of 12 millimetres
per second is satisfactory for this particular fume. Attention was given to
designing a bag-changing arrangement whereby the operators would not be over-
exposed to the high arsenic fume collected. A vacuum cleaning system was
installed which cleans the outside of the bag as it is withdrawn from the top
of the chamber. The system has worked satisfactorily under close supervision.
Bags are changed as holed and the failures to date have been caused by burning
cinders getting into the baghouse in the gas stream and some rubbing of the
cloth on the support frames. It is estimated that a normal bag life of 4 years
should be obtained.
181
-------
Figure 3 — Typical Control Chart
182
-------
A similar baghouse using the reverse air pulse cleaning method was
installed to handle the dust and fume arising from the zinc melting furnace.
The fume originates from the use of ammonium chloride flux in the furnaces and
in combination with acidic electrolyte sticking to the cathode zinc sheet,
presented some filtering problems. Extra heat was bled into the system from
an adjacent combustion furnace to keep the gas temperature above the dew point.
The baghouse is performing satisfactorily and apart from some problems with
bags rubbing on the wire cages, it is expected that a bag life of about 3 years
will be achieved. Present cloth specifications are as follows -
Type - heavy duty terylene needlefelt
Weight - 400 grams per square metre
Permeability - 16 cubic metres per square metre per minute at 25
millimetres water gauge pressure.
General Operating Comments - Men working inside baghouses on bag maintenance
are highly exposed to particulates and with lead fume the health risk is
correspondingly greater. They have to be provided with a full cover of
protective equipment including suitable respirators and supervision should be
strong. With the new centralised baghouse it is hoped that all bag maintenance
will be done on daywork with one shiftworker controlling shift operations.
The total number of men employed in baghouse work will be reduced by at least
4 and perhaps 6 men.
The problem in continuously monitoring particulate material in the
filtered gases has not yet been resolved as preliminary trials with an optical
instrument were unsuccessful and tests are now being conducted with an 1KOR
monitor. This instrument measures the charge transfer which Occurs when
particles collide with a sensor in the gas stream. Normal emissions will be
in the order of 10 milligrams particulates in each cubic metre of effluent
gas and this is a very low level for detection. An instrument capable of
indicating changes in emission levels is required to enable bag maintenance
personnel to locate leaking baghouse chambers.
Wet Scrubbers
Wet scrubbing has been practised in the lead smelter for many years and
has mainly been used in the sinter plant area to handle the wet hygiene gases.
The particulates contained in these gases are of a sticky nature and caused
blockage problems both in the ducting and in the scrubber. While the units
were clean and working properly the emissions were reasonable, but the position
soon deteriorated and frequent plant stops were necessary to control the
problem. At this time a modified scrubber incorporating the jet impingement
and fluidised bed principles was adopted. The Scrubber design is shown in
Figure 4 and illustrates how the entering gas stream impinges on the water
surface before it rises and passes through two beds of marbles mounted on re-
taining grids. Water sprays are situated on the underside of the marbles and
also at the top of the scrubber to wash away the adhering particulates.
Unfortunately the marble beds tended to block up with mud and the cleaning
operation during plant stoppages was very messy. Venturi scrubbers have now
replaced the marble scrubbers and plugging is avoided by the relatively open
design of these units. Another type of scrubber tried unsuccessfully in the
sinter plant and subsequently installed in the leaching section of the zinc
183
-------
plant is the Joy Microdyne scrubber. A diagramatic sketch outlining its main
components is shown in Figure 5. This low energy scrubber is unable to achieve
the collection efficiency required in handling the hot fume laden gases from
the drying of leach residue material and will ultimately be replaced by a
venturi scrubber.
Venturi Scrubbers - High energy scrubbers were first installed in the zinc
plant to handle the gas from the zinc oxide roasting kilns. A unit consists
of two Venturis in series which are followed by a cyclone separator and the gas
is then discharged through a fan up the stack. The first venturi is designed
to cool and condition the gas while the major scrubbing action is achieved in
the second venturi. A volume of 240 cubic metres per minute is handled by the
fan and water additions are 70 litres per minutes in the first stage and 150
litres in the second stage. Pressure drops are approximately 1120 millimetres
water gauge across the Venturis and 150 millimetres water gauge across the
cyclone separator. The fan is driven by a 130 kilowatt electric motor.
Unfortunately, the collection efficiency of these units is slightly below 99
percent and the particulate emissions do not meet the Clean Air Regulation
limits. It is now intended to replace one of the existing units with a long
throat venturi to determine if better re'sults are possible with the longer
throat design.
A long throat venturi scrubber of the Stairmand - I.C.I, design has been
used successfully in the sinter plant hygiene circuit and was recommended by
our Consultant, Charles G. Martin . The design features are shown in Figure 6
and the unique characteristic is that the throat length should be at least
seven throat diameters. This is based on the theory that the turbulence
created in the longer throat section will ensure intimate contact of the water
and dust particles. Under these conditions the method of water injection is
not considered important and simple tangential entry is adequate.
Initial design parameters for the sinter plant units were established
from a test unit set up in the plant. The specifications for the final
arrangement was then submitted by our consultant and six similar units were
installed. They were designed to handle a nominal 250 cubic metres per
minute of air, throat diameter to be 240 millimetres and a pressure drop
across the installation of 660 millimetres. Water requirement was estimated
to be 154 litres per minute and a 50 kilowatt power requirement to drive the
fans was calculated.
In practice the volume handled is around 180 cubic metres per minute but
the other specifications are substantially correct. Collection efficiencies
of 99.8 percent by weight are common and blockages have been eliminated with
consequent better draughting. However, there are still problems to be solved
in the fan operation as there is a carryover of dirty water droplets from the
cyclone separator and the solids build up on the fan blades causes out of
balance problems. The fan blades also show the effects of corrosion and
erosion, and alternative materials of construction are being tried. It is
possible that a mist eliminator will be required on top of the cyclone
separator. We still have some way to go before mastering the techniques of
venturi scrubber operation.
184
-------
MARBLE
BED
MARBLE
BED
INLET
FLUE
WATER
LEVEL
WEIR
DRAIN
VALVE
Figure 4 — Marble Type Scrubber
MICRODYNE
FAN
Figure 5 — Joy Microdyne Scrubber
185
-------
FUME INLET
TANGENTIAL-1
WATER INLET
LONG THROAT
VENTURI
SEPARATOR
CLEAN GAS
OUTLET
FISHTAIL
UNDERFLOW
REGULATOR
MOISTURE
TRAP
FIBREGLASS
LAUNDER
Figure 6 — Long Throat Venturi Scrubber
186
-------
CONCLUSIONS
B.H.A.S. is committed to heavy expenditure on environmental projects at
the present time and there is still much to be done. The toxicity of lead is
very much under scrutiny and emission standards are being progressively
reduced, making it difficult for older type plants to comply. Serious
consideration must now be given to the new technology required in the future
to meet changed environmental considerations, otherwise control costs will
become prohibitive.
The Management of B.H.A.S. is very aware of its obligations and is
maintaining a policy designed to control particulate emissions to the limits
of the available equipment.
ACKNOWLEDGEMENTS
The author thanks the Management of The Broken Hill Associated Smelters
Pty. Ltd. for permission to present this paper.
REFERENCES
Stairmand C.J. - Removal of Grit, Dust and Fume from Exhaust Gases
from Chemical Engineering Processes. The Chemical Engineer, December 1965.
Martin C.G. - The Design of Industrial Air Pollution Control Devices.
Second International Clean Air Congress of the International Union of Air
Pollution Prevention Association 1970 Washington.
187
-------
APPLICATIONS OF BAGHOUSES IN LEAD AND ZINC SMELTERS
Knowlton J. Caplan
Industrial Health Engineering Associatesr Inc.
Minneapolis, Minnesota 55440
ABSTRACT
Fabric filters are widely and successfully used in lead-zinc smelters.
Special considerations need to be provided for hot gases, presence of SOX,
high dew points, and sometimes for fire prevention Efficiencies are adequate
for present emission standards, ranging 98-99 wt.%, and a clear stack is
usually obtained. Restoration of permeability is usually by shaking alone.
Air-to-cloth ratios are usually quite conservative because dust cake per-
meability is low. Dust is somewhat difficult to handle physically; problems
of employee exposure to toxic materials are difficult to control. Possible
means of improving efficiency and combating other problems are discussed. A
tabular summary of some operating data is presented.
GENERAL INDUSTRY DESCRIPTION
The ores which are the raw material for primary lead and zinc production
are usually found as combined lead-zinc sulfides. The major separation of
lead from zinc takes place in a flotation process at the mill, commonly
located at or near the mine site. Both lead and zinc ore concentrates are
usually processed in the same plant, or at least by the same company in a
complex of plants. There are some instances of buying and selling concen-
trates, intermediate products, or residues; these arrangements shift from
time to time depending on metals markets and other factors. Lead is
usually produced in 100 pound pigs or one ton blocks of metal, and zinc
is usually produced in the form of zinc oxide or metallic zinc. Metallic
lead is currently produced only by pyrometallurgical processes, whereas
metallic zinc is produced by electrowinning followed by remelting and
casting.
Lead and zinc smelters and refineries are large, capital intensive
establishments, most of which rank in size only below integrated steel
mills and copper smelters. The lead-zinc smelting industry in this country
is old, much of the installed capacity having been built around the turn
of the century and the most recent plants in the new Missouri lead belt
having been built in the middle sixties.
Raw material sources may be categorized into two groups, the Missouri
lead-belt ores and the Western ores. The Western ores are relatively high
in silver, copper, cadmium, and antimony, and are highly variable according
to specific mine source. The Missouri ores are much lower in these
other metals, with copper being the major contaminant.
Because the raw material sources are non-uniform with respect both to
time and location, and because different smelters may handle different
sources from time to time, the industry is not highly attracted to a
188
-------
sophisticated level of automation and mechanization. Most process steps
are either batch or continuous-batch so that time is available between
process steps for various necessary changes in process parameters or
reagent additions to permit the necessary metallurgical control. Only as
the final steps in the process are approached does one find an opportunity
for handling a reasonably standard, uniform product such as metallic lead,
metallic zinc, or zinc oxide. Uniformity of process streams practically
never occurs with the by-products, because these are the streams that
provide the "balloon" for necessary changes in producing the primary
products.
The nonferrous smelting industry is sometimes accused of being "back-
ward". However, in view of the circumstances—high capital investment,
large and uncontrollable variations in raw materials, and immediate and
direct impact of world metal markets, a tendency to go slow is under-
standable. It is my impression that there is actually a high degree of
technology involved in primary nonferrous smelting, but it is a different
kind of technology than one finds in an automobile assembly plant or a
plant producing synthetic chemicals.
There are only six primary lead smelter plants in the United States,
and although I do not have an exact count, a correspondingly small number
of primary zinc producing plants. Although the basic types of processing
are usually quite similar, the number of plants is not sufficiently large
to permit arriving at a "normal" or "standard" description of processes
and related air pollution control facilities.
GENERAL FABRIC FILTER APPLICATION
In subsequent sections, tabular data is presented showing the typical
range of application of fabric filters to smelter operations. However,
some aspects of fabric filter application are rather common throughout and
can be described in a single discussion. It will be obvious when the
tables are studied that much data is missing; much of the information was
collected incidental to other work, and much of the missing data probably
has never been determined.
One of the most notable omissions is particle size information.
Although this is regrettable from a scientific point of view, for practical
applications it is not too important. Virtually all the processes generate
emissions which are mostly fine metallurgical fume which would be expected
to, and does, have a mass median aerodynamic diameter of one micron or
less. If the fume is zinc oxide, it agglomerates readily so that if
sufficient time is allowed, the end product is more like a mechanical dust
than a fume. Of greater practical importance is the fact that the fume
is "difficult"; that is, the permeability of the accumulated filter cake
is low. The collected dust is, in general, somewhat difficult to handle,
somewhat sticky, etc., as distinguished from a free-flowing granular
material.
189
-------
It will be noted that all the filter applications involve very con-
servative air to cloth ratios in order to achieve differential pressures
across the fabric in the normally expected range, not exceeding 5 inches
w.g. Differential pressures across the fabric are usually in that expected
range, although accurate data are not available due to the large variations
in operating conditions, which are commonly not measured at the same time
that the pressure drop is monitored.
Efficiencies are usually in the range of 99 wt.% or better and seldom
as low as 98%. The stack is usually visually clear and the outlet grain
loading comfortably below present emission standards, whether official or
unofficial, of about 0.02 grains per cubic foot. The efficiency is usually
somewhat less than the 99.9+% that can usually be expected for a fabric
filter under ideal conditions, for several reasons. One reason is that
the baghouses are usually quite large, and in view of the high temperatures
and noxious conditions, finding and fixing small mechanical leaks through
the fabric is difficult and is not actively pursued as long as the
emissions are satisfactory. Another factor is that, due to the very
large square footage of fabric filter required, optimizing efforts in the
past have usually been directed toward bag life, cost, and pressure drop
with the fine points of increased efficiency taking a somewhat secondary
role.
The hot gases from the furnace may be cooled to suitable baghouse
temperatures by any of the classical methods. If acid gases such as SC>2
are present, the most favorable design would be to use radiant coolers to
avoid the introduction of additional water vapor. When water sprays are
used, they are usually used only for partial cooling, the remainder being
by dilution in order to avoid too high a dew point.
If acid gases are present in significant quantity or if water spray
cooling is used, the fabric filter must be kept above the dew point.
Keeping the hopper above the dew point requires provisions over and above
that required for the main housing. The collected bed of dust in the
hopper provides an insulating effect so that the metal temperature may
be far below the gas temperature. However, the vapors will still migrate
through the bed of dust, condensing on the cold metal walls and forming
a cake or mud.
Smelter baghouse testing has been conducted by EPA contractors.
Such testing has focused almost entirely on quantifying the emissions
from the baghouse and provides little or no design or operating data con-
cerning the baghouses themselves or the inlet loading conditions. This
limitation on such projects may or may not be appropriate, depending on
one's point of view. However, one aspect of such testing is of prime
importance.
In at least one such test with which we are familiar (1), the stack
sampling was stopped during and for a short period following the shaking
of a section in the baghouse, even though it was noted that a visible plume
190
-------
persisted for a significant fraction of a minute after the shake cycle.
Obviously taking the data in this fashion resulted in a lower measured
emission than would be obtained if the shake cycle were to be included.
While this may, at first thought, be gratifying to the smelter manager,
it in fact represents a strong potential for inequity in regulations.
Regulations seeking "best available control technology" or its several
variations would naturally be based on the data furnished to EPA. Thus
the potential exists for the regulation to be based on a technology which
is, in fact, somewhat less effective than the test data represent. One
possible compensation for this bias would be to formulate emission
standards based on the same testing methodology, but there are two
objections to this. One would be the obvious, that such testing doesn't
best represent the actual emissions; and the other would be that, in many
installations, programming automated stack monitoring equipment in such
a fashion would be difficult or impossible.
Fires in Metallurgical Baghouses
Fires are not uncommon in metallurgical baghouses. They usually but
not always occur in the hopper of the baghouse. However, the circumstances
under which they occur are highly variable, and immediate investigations
following a fire frequently lack the application of scientific expertise.
As far as we know, no concerted research has been done on the subject.
Fires occurring in the hopper are almost always caused by the presence
of finely divided unoxidized metallics of a pyrophoric nature. Metallic
zinc is probably the most common source of such fires, and metallic lead
is probably the second most common source.
Prevention of fires in baghouse hoppers has been most successfully
accomplished by the addition of an inexpensive diluent material. The most
common diluent is agricultural lime. It obviously serves as a diluent to
the combustible materials and also, if a fire does start, heat will be
absorbed in dehydrating the lime thus furnishing an additional cooling
effect. Some installations have used various carbonates such as soda ash
or ground limestone, the theory being that the heat of the fire will re-
lease carbon dioxide which will tend to smpther the fire. This latter
action probably would not be effective for a fire in the bags since the
filtering air flow would rather quickly flush the blanket of carbon
dioxide away from the combustion area.
There is an empirical test called a "burning test" which is applied
to controlling fires in metallurgical baghouses (2). In general the test
consists of taking a shallow pan full of the collected dust and attempting
to light it at one end by ordinary means such as a match or a burning
paper towel. The time of burning from one end to the other of the pan is
then measured, and if it is less than some specified value, more diluent is
added to the duct entering the baghouse. The test is strictly empirical and
the parameters need to be developed for each individual plant or baghouse
situation.
191
-------
One plant used a pan one-half inch deep, 2 inches wide, and 6 inches
long. In this case, the burning was not actually timed; if the dust
burned at all, more diluent was added. The diluent was hydrated lime
(agricultural lime) and was added in amounts varying from 5% to 20% of
the total dust load.
In another plant the test pan was 1 inch deep, 3 inches wide and 9
inches long. If the length of time required to burn from one end to the
other was greater than 2 minutes, it was judged acceptable; if less than
2 minutes, agricultural lime was added to the flue.
Fires starting on the bags rather than in the hopper are usually
blamed on the carryover of sparks although there are some indications that
these fires also are sometimes started by metallics. In one interesting
case (not a primary smelter), the bags of a reverse pulse filter with
bottom inlet caught fire at the top of the bags and there was no fire in the
hopper. Because at times sparks had been observed in the flue, the fire
was blamed on sparks. It is difficult to understand how such sparks would
preferentially migrate to the top of the bags. Another possible explanation
would be that the reverse pulse air, having more energy and physically
agitating the dust cake more at the top of the bags than further down, may
have started the fire in the metallics present. In any case, the provision
of a spark trap such as a cyclone or a spray chamber would prevent fires
from sparks if they cannot otherwise be prevented. If both sparks and
unoxidized zinc or lead are present, both types of precautions may be
required.
Dust Handling
The handling of recovered dust in most secondary smelters is a
difficult problem because such plants usually do not have a multiplicity
of options for disposal of such material. The material may also have a
significant monetary value. Its disposal by means other than by return
to the process would be a serious problem in solid waste disposal.
The dust is usually somewhat difficult to handle. It does not flow
readily from hoppers, etc., unless special provisions are made. Pneumatic
conveying should be an applicable technique, but some operators have
experienced difficulty with this operation due, in our opinion, to
inadequately designed systems. Screw conveyors work, but they must be
enclosed and dust control applied. Belt conveyors or other dusting type
of material handling equipment is to be avoided.
The chemical composition of the dust is usually significantly different
from the parent material and metallurgical considerations may govern the
point in the process to which it is returned, or even if it is to be
returned. Leaving aside the metallurgical considerations, dust returned
in that form to a blast furnace, for example, would be mostly re-entrained
and would immediately reappear in the baghouse, constituting a circulating
192
-------
dust load. The circulating dust load may be an operating advantage if
such recirculation accumulates unwanted impurities so that a bleed-off
or programmed disposal of the dust can be made. Otherwise, the dust will
be recovered more efficiently by pugging, pelletizing, or briquetting of
dust, or by incorporating it in sinter feed materials.
Baghouse Design
Considerable variation is found in baghouse design features, involving
the basic type of structure, whether the baghouse is enclosed or insulated,
whether there is a one-fan or a two-fan system, and whether section dampers
are single or double. Almost all the baghouses applied to primary nonferrous
smelting to date are of the shaking bag type with minimal application, to
date, of pulse jet or reverse flow types.
The basic structure of the baghouse may be literally a building, con-
structed of brick or monolithic concrete, within which the cell plate, bags
and shaking mechanism are located. The other extreme is the structural
steel baghouse with sheet metal panels, insulated or uninsulated, and with
or without auxiliary heating of hoppers. Many older baghouses are of the
building type construction, but such construction is relatively expensive
as compared to the other types. The relative merits of reliability,
durability, problems of preventing condensation, damage by fires if such
occur, are reasonably obvious.
Another design variable, usually associated with the era and type of
basic construction, is whether or not the dust is handled in a cellar or
by hoppers and conveying equipment. The cellar type of construction has
the advantage that the collected bed of dust may be deliberately ignited so
as to reduce any fire hazard in the following handling operations, and to
densify the dust. As far as we know, such deliberate ignition has not
been used for hopper and conveyor type systems. Problems with the cellar
type of construction revolve mostly around the exposure of the front-end
loader operators to toxic dust, and the creation of fugitives by the
handling of the collected dust in this manner. These disadvantages could
be overcome by providing a greater height of the cellar so that front-end
loaders with pressurized clean air cabs could be used in them, and dust
control can be visualized for subsequent handling of the material from
the front-end loader into hoppers, trucks, or cars. All such provisions
are fairly expensive, high labor cost solutions, so that all factors
considered, it is doubtful that any more cellar-type baghouses will be
constructed in the industry.
Without making a specific cost evaluation, it would appear that the
scheme of constructing an ordinary baghouse inside a secondary structural
steel and siding building offers the best overall solution in climates
where condensation is a problem. Little or no supplemental heating is
required for the surrounding building because most of the heat is
193
-------
furnished from the baghouse itself. Maintenance on conveying equipment,
etc., would be conducted inside, minimizing fugitives and providing a
better opportunity for dust control if necessary.
Some metallurgical baghouses have a two-fan system, one on the dirty
side and one on the clean side of the baghouse. This permits the baghouse
enclosure to operate at virtually zero pressure differential to atmosphere,
which may be important in some cases. In general, the fan operating on
the dirty side of the baghouse will suffer from increased erosion, corrosion,
or caking and will be a high maintenance item. Further, in general, one
blower can provide adequate static pressure for the system. Therefore,
unless an unusually high static suction is required for the system or
unless the structural problems are unusual, use of one blower would seem
to be more cost-effective. If conditions are not unduly onerous for
blower operation, it may be on the dirty side of the baghouse, otherwise
it should be on the clean side. Locating the blower on the clean side
of the baghouses means that it is under suction, resulting in fewer
problems when it is necessary to enter a section for maintenance, but
higher structural cost.
Some smelter baghouses are provided with section dampers on both the
inlet and outlet of each section. These are usually found where the outlet
gases from the baghouse go to a tall stack and the system fan is on the
dirty side of the baghouse. The outlet damper is then necessary to stop
the significant draft that would occur due to the stack height when the
section is isolated for maintenance.
APPLICATION TO SPECIFIC SOURCES
Lead Smelters
The major sources controlled by baghouses in lead smelters are the
sintering machine and the blast furnace. In many cases a dross reverbera-
tory furnace is associated with the blast furnace and served by the same
baghouse.
The sintering machine is similar in principle to the well-known sinter
machines in the steel industry. Most lead smelter sinter machines are up-
draft units. First an ignition layer is placed on the grates, ignited,
and then covered by a much thicker main layer. The sinter feed is basically
ore concentrate, return sinter, recycled lead-bearing materials and perhaps
additives to promote slagging later in the blast furnace. The sinter
machine off-gas typically is used as a feed gas to an acid recovery plant,
and is desired to contain about 5% or more SO2. The circuit of gases
through the sinter machine may be somewhat complex, with weak gases pro-
duced toward the end of the machine, where S02 evolution is low, being
recirculated through an upstream section of the machine to enrich the SO2
content before going to the acid plant.
194
-------
The sinter machine off-gases contain dust and fume and must be cleaned
before being processed in the acid plant. This is typically accomplished
either by fabric filter or by electrostatic precipitator. For fabric
filters, the available data is presented in Table 1.
Blast Furnace
The product of the sinter plant is lump sinter, fused chunks of lead
oxide, lead, slag and other impurities. The sinter is charged into the
blast furnace with coke as a reductant and with more fluxes. The blast
furnace off-gas contains a lead oxide fume, dust, trace amounts of SO2 and
significant amounts of carbon monoxide. Most smelters control the emissions
from the blast furnace by use of a baghouse. The same baghouse frequently
serves the dross reverberatory furnace, where the dross from lead kettles
is remelted to recover the contained metallic lead and to produce matte
and/or speiss in a handleable form. The blast furnace gases typically go
through a settling chamber or a long low-velocity flue to provide cooling
and to settle out any sparks. Sometimes water cooling sprays are installed
in the flue.
Table 2 shows the parameters involving the baghouses and their
operations on blast furnace/reverb furnace duty.
The zinc fuming furnace is, in some instances, physically located in
the lead smelter because it can be charged with molten slag from the blast
furnace to recover the contained zinc as zinc oxide. However, since the
product is zinc, and since the zinc fuming operation is not necessarily
conducted in the lead plant, it is included under zinc plants.
ZINC PLANT
Zinc sulfide ore concentrates are oxidized in one of several possible
styles of roaster. Since the desulfurized product is to be dissolved in
acid and used as feed for electrowinning, there is no need to prepare a
sinter product. The roasters may be a hearth type roaster or a combination
flash and hearth type roaster. Off-gases from the roaster are rich in SO2
and are usually the feed for an acid recovery plant. Either electrostatic
precipitators or fabric filters are used to clean the off-gas for acid
plant processing. Characteristics of fabric filters used for this purpose
are shown in Table 3.
Zinc Fuming Furnace
In the zinc fuming furnace, metallic zinc vapor is evolved from the
blast furnace slag and oxidized to zinc oxide fume. Radiant hairpin coolers
or long low-velocity flues, or a combination of both, are used to cool the
gas. The long transit time allows agglomeration so that the product, cap-
tured in a fabric filter, is more like a dust than a fume in its appearance
and handling characteristics. Fabric filter applications in this duty are
shown in Table 3.
195
-------
TABLE 1
LEAD SMELTERS
SINTER MACHINE
FABRIC FILTER ITEM NO.
CHARACTERISTIC
F.F. Type
Fabric
Temp °F
A/C Ratio
acfm/ft2
AP in.w.g.
Inlet gr/scf
Outlet gr/scf
Cooling
SO2 %
Bag Life Mos.
Dust Removal
7 8
Shaker Shaker
Acrylic
200 215
1.1 1.1
3.4
2.0 1.4
0.001
Water
Spray
5.5
36+
Screw Cellars
Conv-
9
Shaker
-
260
1.8
4.0
2.1
0.003
Water
Spray
-
23+
Cellars
10
Shaker
Polyester
240
1.2
-
-
-
-
5.0
-
Screw
Conv.
196
-------
TABLE 2
LEAD SMELTERS
BLAST FURNACE
FABRIC FILTER ITEM NO.
CHARACTERISTIC
F.F. Type
Fabric
Temp °F
A/C Ratio
acfm/ft2
AP in.w.g.
Inlet gr/scf
Outlet gr/scf
Cooling
S02 %
Bag Life mos.
Dust Removal
1
Shaker
-
200
1.8
2.0
2.0
0.05
—
0.1
30
Screw
Conv-
234
Shaker Shaker Shaker
Wool Acrylic
200 160 160
0.5 1.6 1.6
2.5 3.5 3.5
1.1 2.6
0.008 0.006
Infilt. - Water
sprays
_
53 40
Cellars Cellars Cellars
5
Shaker
Acrylic
150
0.55
2.0
0.9
-
Infilt.,
water
sprays
-
56
Cellars
6
Shaker
Acrylic
-
1.3
3.5
1.24
0.01
-
-
Screw
Conv.
197
-------
TABLE 3
CHARACTERISTIC
F.F. Type
Fabric
Temp °F
A/C Ratio
acfm/ft2
AP in. w.g.
Inlet gr/scf
Outlet gr/scf
Cooling
S02 %
Bag Life mos.
Dust Removal
Application
ZINC
11
Shaker
Polyester
300
2.1
5.0
-
0.028
Radiant
-
-
Screw
Conv.
Zinc
Fuming
Furnace
PLANT
FABRIC
12
Shaker
Nomex
300
1.8
3.2
6.7
0.003
Radiant
-
25
Screw
Conv-
Zinc
Fuming
Furnace
FILTER ITEM
13
Shaker
Acrylic
240
1.06
5.0
15.0
0.0055
Radiant
-
30+
Screw
Conv.
Zinc
Fuming
Furnace
NO.
14
Reverse
Glass
500
1.96
8.0
5-8
0.05
WHB
6.5
24
Screw
Conv.
Roaster
198
-------
MISCELLANEOUS OPERATIONS
A few miscellaneous operations are shown in Table 4.
Fabric filters are not widely used in the primary smelting industry on
miscellaneous dust sources but it is predictable that their use will become
more widespread. In the industry, dust from material handling operations
is distinguished from furnace ventilation by calling the former "mechanical
dust" or "hygiene ventilation". For such applications, low to medium
pressure drop wet scrubbers have been applied more frequently than have
fabric filters. One reason for using wet scrubbers is that, in the sinter
preparation and sinter return circuits, the materials being handled are
often hot and moist, ruling against the use of fabric filters without
extensive gas pretreatment. However, from both operational and hygiene
points of view, it is coming to be regarded as a net advantage to handle
such materials dry. Even though much more dust is generated when handling
dry materials, the associated problems with duct work, housekeeping, etc.,
are much greater than if dry materials were handled. Well-designed dust
control hooding can avoid excessive dust pickup while still preventing
dispersion of the dust into the workplace air, and all other problems are
simplified. For such applications, the use of fabric filters of course
would give a higher efficiency or a lower emission than would the use of
low or medium pressure drop wet scrubbers, and fabric filters are usually
more cost-effective than high pressure drop scrubbers. A further
consideration is the fact that many nonferrous smelters are located in
areas where water supply is restricted.
MODIFICATION FOR IMPROVEMENT
As previously described, the rather difficult duty for baghouses on
smelter furnace emissions, combined with the onerous working conditions
inside, results in a somewhat lower degree of maintenance and correction
of minor fabric leakages than is possible. There are several design
changes that could be made to result in better efficiency performance.
It is not uncommon to see even the best of metallurgical baghouses
with an inch or more of powdery dust on the cell plate on the clean side
of the fabric. An improvement-in collection efficiency could be achieved
by use of more efficient media, and a philosophy of cleaning up the clean
side of the baghouse and keeping it clean.
Reported experience with development projects for fabric filters in
power plants offers a clue. Such installations have stack monitors for
total particulate. It has been observed that, by correlating stack monitor
readings with the cleaning cycle for different filter compartments, an early
detection can be obtained of minor leakages. If the offending compartment
is then shut down and inspected, it is easier to find the leakage since the
"clean side" of the bags really is clean, and leakages can be found and
fixed before they become serious.
199
-------
TABLE 4
MISCELLANEOUS
FABRIC FILTER ITEM NO.
CHARACTERISTIC
F.F. Type
Fabric
Temp °F
A/C Ratio
15
Reverse
& Shake
250
1.92
16
Shaker
Acrylic
200
2.1
17
Shaker
Acrylic
200
1.67
acfm/ft^
AP in. w.g.
Inlet gr/scf
Outlet gr/scf
Cooling
SO2 %
Bag Life mos.
Dust Removal
Application
Screw
Conv.
Dross Arc
Furnace
4.0
2.6
1.56
30
Screw
Conv.
4.0
2.6
1.56
30
Screw
Conv.
* Weak end of sinter machine plus misc. sources
200
-------
Along with the early warning system described above, it would be
necessary to have a well-designed vacuum cleaning system so that the cell
plate and housing can be cleaned following every maintenance operation.
Bag testing experiments could be made in which the emphasis is shifted
more strongly to considerations of efficiency. Unfortunately, efficiency
experiments cannot be done with one or a few bags in each compartment like
experiments on bag life. Following preliminary small-scale experiments,
a real operating experiment would require resulting an entire compartment.
The design of fabric filters commonly provides enough sections so that
the net filtering area is adequate when one section is off-line for cleaning.
Some baghouses provide a second extra section so that the net capacity is
adequate when an additional section is off-line for maintenance. This
practice should be expanded, in order to properly compensate for the
operational changes suggested above.
OSHA VS. EPA
For various reasons, government regulation of emissions and ambient
air quality is compartmentalized in the Environmental Protection Agency,
and government regulations concerning contaminant concentrations in the
workplace are lodged with the Occupational Safety and Health Administration
of the Department of Labor. This is unfortunate in several respects. Of
major importance is the fact that contaminant concentrations inside the
plant at the workplace, and contaminant concentrations in the stack or
ambient air, are the two opposite ends of the same, engineering problem.
A second historical fact is that EPA got started first. Industry frequently
compartmentalized itself to match the government agency structure. The two
functional problems, similar as they are in technical content, thus meet
only at the engineering function where equipment design and installation is
involved. Since compliance with most EPA requirements is regarded as a
constraint to be met with minimum cost, and since EPA requirements preceded
OSHA requirements, there has been a trend in recent years to place
engineering emphasis on EPA requirements; attention to the OSHA-type
requirements was minimized or forgotten.
Now we have both requirements with us. There is a degree of tunnel
vision on the part of both agencies regarding single-minded pursuit of their
objectives without consideration of other government requirements, due largely
if not entirely to the legalities and bureaucratic structure involved. As
a result, there are many severe problems looming on the horizon. Some of
the requirements are complementary to each other, as in the example given
above for keeping the clean side of the baghouse clean. The steps
described lead to both improved efficiency performance of baghouses and
to reduction of exposure to the workers doing the baghouse maintenance.
That fortunate compatibility does not exist in another and very important
area.
201
-------
EPA, as you all know, is beginning to emphasize fugitive emissions
from the smelter, as well as stack emissions. With a stringent fenceline
lead-in-air standard looming, it is predictable that there will be major
emphasis on controlling fugitive emissions. All of you who have had an
opportunity to contemplate this problem realize its magnitude. Fugitive
sources include the discharge of ventilation air from the smelter, the
emission of contaminated air through roof monitors, etc., storage piles,
roadways, etc.
One technique that has been attempted, in focusing on EPA requirements,
is what I call building containment. The entire building is enclosed and
the roof monitor exhausted through a fabric filter and up a tall stack if
necessary. If one can live with the results of such an installation, it
does solve the problem of fugitive emissions from the smelter building.
Such installations are very expensive, requiring perhaps as much as one
million cfm capacity, in an attempt to emulate mother nature's ventilation.
The baghouse is so big that it usually must be located remotely, so that
long, large and expensive flues are required, the power requirement is
significant, etc., etc. For these reasons, if EPA compliance is all that's
in the designer's mind, the capacity of such an installation is cut to
the bone and working conditions inside become significantly worse.
Two such building containment installations have been made in the
copper smelting industry, both of which are serious failures in terms of
internal working conditions. Both installations were designed on an
"air changes per hour" basis which, aside from the fundamental question of
whether such an installation is appropriate in the first place, is the
poorest design basis that could be selected. Designing general ventilation
on the basis of air change rates results in a capacity related to the size
of the building rather than to the size of the problem. While that approach
may be appropriate for some relatively standard situations like school
classrooms, where the students can be packed only so close and the ceiling
height ranges only between 8 and 12 feet, it does not apply to most
industrial operations since there is such a wide variation from one plant
to another. This variation is even more extreme in the nonferrous smelting
industry than in most others.
Proper design of a building containment system must be based, as a
minimum, on an attempt to remove hot contaminated air at least as effectively
as occurred with natural ventilation before the containment. This means
estimating the velocity and volume of thermal plumes, which is exceedingly
difficult in view of the size of the operations and their irregular nature.
Because workplace conditions seldom meet OSHA standards now, general building
containment ventilation based on this concept would still be inadequate, but
at least the conditions inside would not be made worse than they are now.
Many of the sources in the smelter can be controlled locally, although
in view of the nature of the operations, extraordinary effort may be re-
quired, and more than one try may be required, to achieve success. Assuming
that success is eventually achieved by local control, the payout is signifi-
cant.
202
-------
The potential payout assumes one of two forms. It is possible, but
not highly probable, that major advances in local control of fume and dust
would make the roof monitor emissions of the smelter building so low as to
escape the need for further control. That would indeed be a happy state
of affairs. A much more likely result is that the local control would
reduce the contaminant burden of the containment air sufficiently so that
air cleaning equipment much less expensive than heavy duty fabric filters
would be adequate. The particulate collected in the containment systems
mentioned amounts to about 100 Ibs/hr. With better local control of
sources, this rate could be reduced to the point where much less expensive
filtering equipment, or perhaps none at all, would suffice for the con-
tainment system. Much higher general ventilation rates would then be
affordable. The combination of local control and copious general ventila-
tion would improve, rather than degrade, working conditions inside the
building.
Assuming that good faith efforts on the part of the industry can
be established and recognized, it seems only reasonable to ask both EPA and
OSHA to be cognizant of each other's impact on the industry. The agencies
should refrain from rule-making incompatible with the other agency's
objectives and rules. The agencies should recognize the increased cost and
time required for the industry to cope with both sets of requirements, in
what is essentially two artifically defined parts of the same engineering
problem.
ACKNOWLEDGEMENT
The kind assistance of the many individuals in the industry, who
furnished much of the tabulated data, is sincerely appreciated.
REFERENCES
(1) Tussey, Robert C. Jr., "Emissions from a Primary Lead Smelter Blast
Furnace", Midwest Research Institute, EPA Contract No. 68-02-0228,
ETB Test No. 72-MM-14, Kansas City, Mo., January 1973.
(2) Godsey, E. S., "Application of Fabric Filters to Lead and Zinc Smelters"
Proc. APCA Conference "The User and Fabric Filter Equipment Specialty
Conference" Buffalo, NY, October 1973.
203
-------
REVIEW OF DESIGN AND OPERATION
OF ELECTROSTATIC PRECIPITATORS IN
NONFERROUS METALS APPLICATIONS
Heinz L. Engelbrecht
Air Pollution Control Division
Wheelabrator-Frye Inc.
Pittsburgh, Pennsylvania
ABSTRACT
Electrostatic precipitators for applications in the nonferrous metals
industry belong in a group representing one of the earliest applications for
electrostatic precipitators. For example, White1 lists several "firsts";
Lurgi, one of the major suppliers of electrostatic precipitators world-
wide, started with an electrostatic precipitator for a Herreshoff furnace
in 1913. Despite this long history, the application is not considered easy
and free of risks. The diversity of processes and equipment used in this
industry has made every installation a unique system and a challenge for its
designers and operators.
This paper is written with the intent to provide a review of some prac-
tical design considerations, applications, and operating experiences with
electrostatic precipitators in the nonferrous metals industry, specifically
for applications in primary smelters for copper, zinc, and lead.
This paper is based on information provided by Lurgi based on 65 years of
service to this industry.
Review of Processes and Electrostatic Precipitator Applications
The selection of methods to separate dusts from waste gas streams and to
recover entrained metals is largely dependent on the process, i.e., type of
smelter and the. particular equipment, such as roaster, converter, dryer,
etc., and on the raw materials used in the process. A further complication
is caused by the sulfur oxides generated at various stages, which not only
require a high degree of dust collection prior to their own removal from the
waste gas stream, but which also to a large extent determine and complicate
the dust collection system. Costs of gas cleaning systems depend basically
204
-------
on the gas volume to be treated and increase progressively with the required
collecting efficiency.
A review of various gas producers in nonferrous metal smelters was made
by Wanner2. Table 1 shows typical gas producers and recommended dust collect-
ors in copper, zinc, and lead smelters. Dust collectors for waste gases in
nonferrous metal smelters are:
- electrostatic precipitators
- fabric filters-
- scrubbers
- cyclones
Operating conditions may limit collectors to specific applications.
In general, dust collectors have to be designed for continuous heavy-duty
operation, a condition which again reflects itself in the cost of such equip-
ment.
So-called hot-gas electrostatic precipitators are used to collect dusts
and to a certain degree fumes from waste gas having an appreciable SOX con-
tent. Because of the sulfuric acid dew point around 150 to 22Q°C., operating
temperatures of 300 to 400°C. are normally selected. This applies to pre-
cipitators for reverberatory furnaces, electro-furnaces, flash furnaces,
dryers, converters, and roasters.
Fabric filters with glass fabrics with silicone finish can be used for
gas temperatures up to 25QQC. Their application is limited because of the
high sulfuric acid dew points and H2F2.compounds in some waste gases. With
sticky dust, cleaning of the fabric filters proves to be difficult, but
fabric filters offer a definite advantage where extreme low residual dust
contents, for example, less than 10 mg/m^ NTP are required.
Scrubbers are often used for the cleaning of waste gases in this indus-
try; for example, venturi scrubbers in various designs. Energy requirements
increase with the pressure drop through the scrubber, but for fine particles,
high pressure drops are required to achieve an acceptable collecting effici-
ency. For example, the pressure drop of a Venturi scrubber for the separation
of zinc oxide particles of 0.5 micrometer diameter is approximately 2,000 mm
W.G. for a collecting efficiency of 96 percent. The main advantage of scrub-
bers are their low investment cost; the main disadvantage are the need for
sludge processing equipment and corrosion-resistant material.
Cyclones are used in cases where particles are larger than 5 micrometers
and collecting efficiencies are low. Particles larger than 30 micrometer
can be collected at almost a 100 percent level. In cases of extremely high
dust loads, cyclones are often used as pre-collectors. Advantages of cyclones
are simplicity of design and capability of handling gases up to 1000°C., be-
sides low investment costs and small space requirements. Their main disadvan-
tage is the rather limited collecting efficiency for fine particles.
The traditional design of wet-process precipitators is the tube or
plate-type lead precipitator, but lead as a material of construction for the
precipitator; i.e., shell and internal parts, has disadvantages, and wet-
process precipitators made out of plastics have been developed. Lurgi has
205
-------
Table 1. GAS CLEANING IN NONFERROUS SMELTERS
SOURCE OF GAS
COPPER SMELTER
Converter
Reverberatory Furnace
Flash Smelting Furnace
Concentrate Dryer
Fluid Bed Roaster
(Partial)
Fluid Bed Roaster
(Dead)
Electro Furnace
Slag Furnace
ZINC SMELTER
Fluid Bed Roaster
Sintering Machine
ISP Furnace
Waelz Kiln
LEAD SMELTER
Sintering Machine
Shaft Furnace
GAS TEMPERATURE
300-400
250-350
350
100-120
400
300
350
500-1,000
350
300-400
350
350
300-400
150-250
U
S02 CONTENT
5-9
1-1.5
8-11
.2
13-14
8
1-6
.5
9-11
5-7
-
-
5-7
~*
DUST CONTENT
g/Nm3
2-9
5-10
30-80
100-200
1,000
300
30-50
5-10
100
20-30
-
50-100
20-30
10
RECOMMENDED
DUST COLLECTOR
Electrostatic
Precipitator (ESP)
ESP
ESP
ESP
Cyclone, ESP
Cyclone, ESP
ESP
Scrubber
Cyclone, ESP
ESP
Scrubber
ESP, Fabric Filter
ESP
ESP, Fabric Filter
-------
supplied more than 120 electrostatic precipitators made mainly out of plas-
tics or with plastics used for major components. Compared with lead precipi-
tators, plastic precipitators offer considerable advantages such as:
- design suitable for high negative pressures
- low maintenance and repair costs
- modular design
- low erection costs
- maximum continuous operating temperature of 80°C.
Design of Electrostatic Precipitators
The basic design of hot-gas precipitators is similar to the standard de-
sign of dry-process precipitators used in other industries, except that the
design used in nonferrous metal applications is suitable for gas temperatures
up to 450°C. If the gas does not contain any components causing corrosion,
normal carbon steel can be used in this temperature range since its
scaling rate is minimal as long as the gas contains less than 5 percent 02-
Electrode systems consisting of discharge wires and collecting surface
plates are grouped in single or multiple electrical fields in series or
parallel, installed in single or double-chamber casings. Precipitators are
built with horizontal or vertical gas flow, the former restricted to plates
as collecting surfaces and the latter normally restricted to a single field
design, but available as plate-or tube-type precipitator.
Collecting surface plates oppose the discharge wires in the electrical
field; electrically they are connected to the positive outlet of the power
supply and grounded with the precipitator casing. Their support structure
can be part of the casing. Collecting surface plates form individual gas
passages (Figure 1) with discharge wires centered between rows of plates,
resulting in a generally non-uniform electrical field.
Collecting surface plates are made out of cold-rolled steel sheets nor-
mally 18 to 24 in. wide and up to 45 ft. high. Assembled they form curtains
up to 18 ft. long. Roll-formed profiles offer rigidity, pockets with little
or no gas flow, and good vibration characteristics. Wet-process precipitators
use a simplified collecting plate design in heavier gauge material.
Tube-type collecting surfaces offer a uniform electrical field but only
the inside can be used as collecting surface; hexagonal, square, or segmented
collecting surfaces allow the use of both sides, but lose some of the uniform-
ity of the electrical field of round tubes. Wet-process precipitators use
collecting surfaces made out of lead or plastic, in tube-or plate-type design.
The discharge system, including its support structure, is connected to
the negative outlet of the high-voltage power supply. Discharge wires are
either suspended from a horizontal support system or individually mounted in
vertical frames. Dry-process precipitators use rapping systems to keep the
discharge wires clean; spray systems in wet process precipitators clean both
collecting surfaces and discharge wires. Discharge wires with round, square,
or star-shaped cross-sections (Figure 2) are used. Spiked discharge wires
207
-------
Figure 1. Collecting Surfaces
208
-------
u
Figure 2. Discharge Wires
HAMMER
HEX SHAFT
ANVIL
Figure 3. Collecting Surface Rapper
209
-------
offer higher currents at somewhat lower voltages. Wet-process precipitators
for sulfuric acid mists use star-shaped lead discharge wires with a steel core
for strength. Special wires with spikes have been developed for precipitators
with square tubes as collecting surfaces. Removal of precipitator dust is
done with rapping systems in dry-process precipitators, and with continuous
or intermittent spray systems in wet-process precipitators. Various mechan-
isms enter into the removal of the precipitated dust from collecting surface
plates by rapping, the most important being the acceleration of the plate
normal to its surface, commonly measured in g (Ig = 9.81 m/sec^). Experience
dictates a minimum of 200g at any given point of the collecting surface plate.
Periodical rapping of the collecting surface plates is achieved by rotating
hammers impacting on the anvils of the guide bars at the bottom of the col-
lecting surface plates (Figure 3).
Wet-process removal of the accumulated dust by flushing requires a good
distribution of the water jets over all of the collecting surfaces, including
the discharge system. Continuously operating spray systems may cause exces-
sive arcing on top and bottom of the electrical field. Discharge wires
mounted in vertical frames are kept clean by an effective rapping system us-
ing rotating hammers (Figure 4); rapping accelerations have been reported by
Engelbrecht^. Frames are made out of square or tubular pipes with suitable
cross-sections for structural strength.
The discharge system is supported and electrically isolated from the pre-
cipitator casing by four support insulators for each electrical field. Insu-
lator materials are porcelain, alumina, or quartz. The insulators also serve
as a seal between the inside of the precipitator and the ambient air. Heaters
are used to prevent condensation on the surfaces of the insulator during start-
up and shut-down periods. Purging systems with heated air or clean gas are
used in special cases to provide additional protection against condensation.
Electrode systems and their supports are reinforced to account for the
possibility of dust build-ups; also reinforced are the rapping mechanisms and
the dust handling systems due to the possible stickiness and abrasive charac-
teristics of the dust and clearances outside of the electrical field are in-
creased to allow for dust build-ups without causing short circuits.
Normally, electrode systems, i.e., discharge wires and collecting sur-
face plates are enclosed in the "smallest possible" housing; in some cases,
additional access and space for manual cleaning has to be provided. Addition-
al rapping systems can be used for the hopper walls. Monitors for the con-
tinuous operation of the dust discharge system and the dust temperature are
recommended. Dusts can be pyrophoric; continuous operation of the electrode
rapping systems should be monitored.
The precipitator casing encloses and supports the electrical field(s).
Casings are designed as self-supporting structures using steel plates with
outside stiffeners as supporting elements, they are welded gas-tight and
thoroughly insulated against heat losses. Inlet and outlet gas ducts are
connected to the precipitator casing with transition pieces, incorporating
gas distribution devices, such as vanes, baffles, or perforated plates.
Various designs are used with good success (Figure 5). Minimum requirements
210
-------
DISCHARGE
FRAME
SUPPORT
BRACKET
SUSPENSION
FRAME
ANVIL
HAMMER
SHAFT
CRANK
RAPPER
(GRAVITY HAMMER)
Figure 4. Discharge Wire Rapper
Figure 5. Gas Distribution Plate
211
-------
for gas velocity distribution have been set by the Industrial Gas Cleaning
Institute^; flow model test results have been reported by EngelbrechtS.
Rapping systems for gas distribution plates are .required for all dry-
process precipitators in this industry. Drag conveyors to evacuate the dust
through hoppers are used; air inleakages through hopper air locks, inspec-
tion doors, drive shaft seals, etc., have to be kept to an absolute minimum
to prevent severe corrosion problems.
High-voltage power supplies are not different from those used for other
precipitator applications. Automatic voltage control systems are designed to
keep the precipitator voltage below the fluctuating arc-over voltage, to re-
duce the voltage immediately after an arc-over inside of the precipitator and
restore the operating voltage immediately after arcing. The last two require-
ments minimize the periods of reduced collecting efficiency and prevent dam-
age to the precipitator by electro-erosion. Automatic voltage controls have
been developed to satisfy these requirements (Crynack^).
Wet-process precipitators require that the gases entering the precipitator
be saturated; a wet-scrubber can be used for this purpose. Dust particles and
mist droplets are precipitated together on the immersed collecting surfaces
forming a thin slurry. To maintain a free flow of this slurry, it is recommen-
ded to keep the dust burden at the inlet of the precipitator below Ig/m^. The
collecting surfaces and the discharge system are flushed periodically with the
precipitator de-energized in this section. Due to the lack of particle re-
entrainment, wet-process precipitators can achieve high collecting efficien-
cies. Lead has been a time-proven material of construction for sulfuric so-
lutions in an atmosphere containing oxygen. The shell of the precipitator is
made of lead sheets supported by a steel structure. The collecting surfaces
(plates or tubes) are made out of lead; all other components of the electrode
system are made out of lead-clad steel parts. Lead is not always sufficiently
corrosion proof, especially if substantial amounts of halogenides are present.
Neither are the mechanical characteristics of lead satisfactory for long time
use.
Therefore, it became necessary to develop an alternative design using
fiberglass reinforced plastic as material of construction (Figure 6). Approx-
imately 120 plastic precipitators have been built by Lurgi during the last
20 years. The collecting surfaces, located between the upper and lower hous-
ing, are made out of non-conducting plastic. The electrical conductivity
of these collecting surfaces is provided by a uniform liquid film (Figure 7).
Special consideration has to be given to the grounding of the collecting sur-
faces. The discharge electrodes are either lead-clad steel rods or made en-
tirely out of high-alloy steels. Spiked electrodes are used extensively.
Precipitator casings are also made out of plastics using, for example,
a glass-fiber reinforced polyester resin with a PVC liner.
Since plastic is rather expensive as a material of construction, it is
important to use all advantages this material offers to the fullest extent.
212
-------
Figure 6
Plastic Precipitator
(Schematic)
COLLfCTINS
SURFACE
NON-
GONDUCTiVE
>IOBOHM-C
k//
I
:M.
^
1
CORONA
CURRENT.
LlOUIOFlLM
<10*OHM-CM
1
DISCHARGE
ELECTRODE
Ix*
'
r "*
Figure 7. Collecting Surface (Schematic)
213
-------
Operation of Electrostatic Precipitators
The continuous operation of an electrostatic precipitator at a high col-
lecting efficiency level depends normally on:
- adequate size
- suitable design
- proper operation
- good maintenance
The first two conditions have to be met by the supplier. Proper opera-
tion and maintenance procedure can be established by the plant operator using
information and manuals provided by the supplier, past experience or informa-
tion provided in the literature, (Englebrecht7).
The dust collection process in an electrostatic precipitator starts with
the charging of particles, their transportation to, and precipitation on the
collecting surfaces. In this process, particles will be deposited on top of
other particles causing the formation of a substantial dust layer on the col-
lecting surfaces. This causes no serious problems as long as the dust parti-
cles are of a relatively low specific electrical dust resistivity which allows
for the electrical charge to bleed off to the grounded collecting surface.
However, in the case of a high resistivity, for example, in excess of 2x10^0
ohm-cm, these charges cannot be reduced quickly enough and a layer of charged
dust particles forms on the collecting surfaces (Figure 8). This will even-
tually lead to a reduction of the voltage gradient between the discharge elec-
trode and the outmost dust layer; and consequently, to a reduction of the pre-
cipitator collecting efficiency (Wanner^, Figure 9).
It is important for the supplier of the electrostatic precipitator to be
aware of the dust resistivity level, which typically depends on the gas tem-
perature (Figure 10). The upper curve exceeds the critical level of 2x10^0
ohm-cm within a certain temperature range whereas the lower curve representing
a carrier gas with an 863 content of 3 mg/m^ is completely below the critical
level. The dust resistivity can be reduced by conditioning, i.e., by adding
certain condensable substances such as 803, NH3, HCL, Na2C03. Conditioning
"by water injection into the process is also feasible (Eisert^, Figure 11).
An 803 level of approximately 20 ppm is normally sufficient for condi-
tioning. For dusts with a high basicity, for example Pb-Zn fumes, 863 levels
up to 200 ppm may be required. Conditioning with 803 is often achieved by
generating 803 in the system itself, provided the gas has sufficiently high
802 an<* °2 levels and dust oxides are available as catalysts. The process
requires that a conversion temperature of 420 to 800°C. is maintained for a
sufficient period of time.
Problems with Electrostatic Precipitators
As much as 803 conditioning seems to be beneficial for the operation of
the electrostatic precipitator, a high 803 level can cause a series of prob-
lems, such as:
214
-------
DUST LAYER.
DISCHARGE WIRE.
COLLECTING PLATE
A REDUCED VOLTAGE IN THE ELECTRICAL FIELD
b BREAKDOWN OF VOLTAGE IN THE DUST LAYER.
Figure 8. Effect of dust resistivity
DUST RESISTIVITY
OHM-CM.
ARC-OVER VOLTASe >
KV.
IOO ZOO 3OO 4OO
6ASTEMPERATURE °C
Figure 9. Dust Resistivity and Arc-Over Voltage
215
-------
SPECIFIC DUST
RESISTIVITY.
10"
OHM-CM |
10"
K)10
to9-
a
b
\
\\
100 ZOO 300 400 *C.
6AS TEMPERATURE.
SO3 . ZERO
SOs = 3mg/m3-
Figure 10. Dust Resistivity as a function of
gas temperature and 863 content
DUST RESISTIVITY OHM-CM.
10"
10"
10"
I09
10"
7
oO
4O 6O 80 100 200 300 400
<5AS TEMPERATURE °C
Figure 11. Dust Resistivity as a function of
gas temperature and dewpoint
216
-------
- increase in corrosion rate by increase of the sulfuric acid dew point.
- S03 can sulphate the dust, making it sticky and difficult to handle.
_ the size of the mist precipitator in the acid plant has to be
increased.
Thus, the .rate of cooling,.i.e., the residence time in the temperature
range between 420 and 800°C. has to be carefully considered.
An example of an electrostatic precipitator for waste gases from a zinc
roaster is used to explain a major problem of precipitator applications in
this industry, namely, the dust build-up on collecting surface plates and
discharge wires.
The reaction in a zinc roaster generates SC>2 in the roaster, leading to
503 in the waste-heat boiler at temperatures below 820°C. The reaction de-
pends on the presence of catalysts and drops to zero at gas temperatures be-
low 420°C., resulting in S03 and zinc in the form of zinc sulphate. Normally,
this does not necessarily lead to build-ups in the precipitator. But in a
number of cases, a heavy build-up is occurring on both the discharge wires
and the collecting surfaces; the mechanisms leading to these build-ups can be
analyzed.
Dust build-ups consist of dust particles sticking on other dust parti-
cles sticking on other dust particles; thus, the conditions on the surface of
the dust particles have to be investigated. Two different kinds of build-ups
can be observed:
- in vicinity of the waste heat boiler, the cyclones, the
roaster precipitators, the fans, and in the ductwork
- on the discharge wires of the roaster precipitator
The latter consisting of nodules and rings (donuts).
Dust particles in the effluent after the roaster are zinc oxide. The
803 generated in this process cannot exist with the zinc oxide, but forms a
zinc sulphate. The rate of this reaction depends upon the penetration of the
803 through the zinc sulphate shell towards the zinc oxide core. Thus, the
surface of the dust particles consists of zinc sulphate and some "free" 863
exists in the gas.
Roaster gas contains a noticeable H20 component and the "stickiness" of
the dust can be plotted as a function of the H20 component with 803 as a pa-
rameter. Obviously, "stickiness" is not a physical or technical quantity,
thus, no dimensions can be given. The graph (Eisert9, Figure 12) shows that
the stickiness of the dust increases with increasing H20 content, accelerated
by an increasing 803 content.
The H20 content of the roaster gas is caused by the moisture content of
the raw material (approximately 8 percent), of the combustion air, and the
water sprayed into the roaster for temperature control.
217
-------
DUST STICKINESS
a
b
S03
SOj
K> 15 20 25 3O
WATER VAPOR. % BY VOLUME
3g/m3 N.TR
O.Zg/m* NT. P.
Figure 12: "Dust Stickiness" as a function of water
vapor and 803 content
218
-------
Modern roasters produce waste gases with moisture contents of only 8
to 10 percent by volume, and the stickiness of the dust is normally low. A
moisture content of 13 percent or more will cause considerable build-ups; and
in excess of 20 percent by volume will limit operating periods to not longer
than one week at a time. The stickiness of the dust causes undesirable dust
build-ups in the waste heat boiler limiting the heat recovery and increasing
the gas outlet temperature. Pressure drop in the system increases. Of dire
consequences are the dust build-ups in the precipitator causing arcing, and,
eventually, short-circuits. The hypothesis of the increase of the stickiness
of the dust analog to the increase in moisture content of the gas is based
on observations. An exact chemical reaction causing this increase in sticki-
ness has not yet been established, but an analogy to the specific dust re-
sistivity, which is also dependent on the conditions at the gas/dust inter-
face should be mentioned.
The build-ups on the discharge wires (modules) are of a different nature.
They are limited to the electrostatic precipitator and can occur with no oth-
er build-ups present. They are also zinc sulphates, formed by a different
mechanism. Based on a number of observations, it can be concluded that they
are caused by inleakage of cold air upstream to the electrostatic precipitator.
The mechanism can be explained as follows:
Cold air causes a local temperature drop into the range of the sulfuric
acid dew point, forming a sulfuric acid mist. Since the rate of evaporation
in this temperature range is rather slow, the sulfuric acid mist reaches the
electrical field of the precipitator. Thus, the precipitator collects dust
and sulfuric acid mist. The predominant collection of the dust particles is
on the collecting surface plates, whereas the sulfuric acid mist particles
are predominately precipitated on the discharge wires. Thus, the ratio dust/
mist is in favor of dust on the plates and in favor of mist on the discharge
wires. The dust can be removed from the plates by normal rapping, but the
wires show a layered build-up, which in one case reached about 200mm with
an air in-leakage of less than two percent. The build-ups on the wires are
relatively sticky and spongy at operating temperature, and, thus, cannot be
easily removed by rapping.
In general, plate build-ups can be reduced by reduction of the moisture
content of the waste gas, and wire build-ups can be reduced and often elimina-
ted by preventing in-leakage of cold air.
Review of Typical Installations
A typical gas cleaning installation for a roaster may include —
- cyclones
- hot-gas precipitator
- scrubber
- gas cooler
~ wet process precipitator
219
-------
HOT-SAS VENTURI
CYCLONE PRECIPITATOR SCRUBBER
WET PROCESS
PRECIPITATOR..
DUST
DUST
5LUD6E ACID ACID
OONDENSATE CDNDENSATE
Figure 13: Gas Cleaning System (Schematic)
220
-------
Figure 14. Gas cleaning system for turbulent
layer roaster
221
-------
i
Figure 15. Hot gas precipitator for lead
sinter roaster
Figure 16. Electrostatic precipitator with
gas conditioning tower
222
-------
Figure 17. Wet-process lead precipitator
with gas cooler
223
-------
A hot gas precipitatbr for gas from a sinter machine for Pb/Zn - concen*
trates for a gas volume of 100,000 Nm3/hr. at 300 to 400°C. is shown in Fig-
ure 15.
Using a conditioning tower allows operation of the precipitator in a more
favorable temperature range. An installation to collect lead-oxides is shown
in Figure 16. A wet-process precipitator with two electrical fields in ser-
ies, separated by a gas cooler, is shown on Figure 17. This precipitator
and its electrode system are made out of lead. This design is now replaced
by precipitators using plastic as the basic material of construction as shown
on Figure 18 for an installation with a gas volume of 50,000 Nm3/hr.
Summary
Electrostatic precipitators are available in a wide range of designs
suitable for applications in the nonferrous metals industry. Highly special-
ized gas cleaning systems have been developed for smelters, including mechan-
ical collectors, dry-and wet-process precipitators, scrubbers, and fabric
filters. Although the increasing demands on air pollution control systems
will require more use of two-stage systems, i.e., a dry-process collector
followed by a wet-process collector, it is expected that electrostatic pre-
cipitators will almost always be the main device in each system.
But it has to be recognized that, in some instances, the capital invest-
ment necessary to provide adequate performance, i.e., collecting efficiency
will be very high and that in other cases, the operation of the gas cleaning
system may not be without problems. In those cases, a search for better so-
lutions has to center on the process itself. A typical example is changing
the process to operate above the acid dew point to eliminate operational and
corrosion problems in the gas cleaning system.
Of equal importance is to avoid dust build-ups in the dry-process pre-
cipitator. It is not practical to provide mechanical means of cleaning by
rapping for all precipitator internal parts. The magnitude of accelerations
to keep, for example, support systems clean of build-ups would be such that
these components would require expensive high-strength materials of construc-
tion. Thus, it becomes important to reduce or eliminate conditions which
could result in dust build-ups. Surface characteristics of particles and
their reaction with the surrounding gas have to be influenced to prevent par-
ticles from becoming "sticky" by chemical reactions and surface condensation.
To do this, changes in the metallurgical process may again be required.
The extent of these changes is still a matter of experience, since reliable
data and information on the actual reactions and phases of all materials
present in the process are not fully available.
In the past, gas cleaning systems were sometimes regarded as a nuisance
for the plant and were often neglected in design, operation, and maintenance.
But they have to be considered essential to the operation of the plant and
treated accordingly.
224
-------
Figure 18. Wet-process plastic precipitator
Acknowledgement:
All photographs courtesy of Lurgi Umwelt und Apparate-Technik Gmbh,
Frankfurt, Germany
225
-------
During the planning of gas cleaning systems for the nonferrous metals
industry, especially when new processes are involved, it has to be considered
that the characteristic of the gas/dust system are decisive for the efficien-
cy and operation of the gas cleaning system and that these characteristics
result from the metallurgical process and are influenced also by changes up-
stream of and in the dust collector. An optimum solution of the gas clean-
ing problem, thus, requires close co-operation between the operator of the
plant, his engineer, and the designer of the air pollution control equipment.
ACKNOWLEDGEMENT
The assistance of Dipl. Ing. Werner Eisert of Lurgi Umwelt und Chemotechnik
GmbH, Frankfurt, Germany, in the preparation and review of this paper is
gratefully acknowledged.
REFERENCES
1. White, H. J., "Industrial Electrostatic Precipitation", Addison Wesley
Publishing Co., Inc., Reading, Mass. 1963
2. Wanner, D., "Gas Cleaning and Gas Handling in Metallurgical Plants",
Lurgi Publication, Frankfurt, Germany, 1976
3. Engelbrecht, H. L., "Rigid Frame Precipitators", Specialty Conference on
Operation and Maintenance of ESP, APCA Michigan Chapter, Detroit,
Michigan, April 1978
4. Industrial Gas Cleaning Institute (IGCI), "Gas Flow Model Studies",
Publication No. EP-7
5. Engelbrecht, H. L., "Air Flow Model Studies for Electrostatic Precipita-
tors", U.S. EPA Symposium on "Transfer and Utilization of Particulate
Control Technology", Denver, Colorado, July 1978
6. Crynack, R. R., "A Review of the Electrical Energization Equipment for
Electrostatic Precipitators, 71st Annual Meeting of the Air Pollution
Control Association (APCA) Houston, Texas, June 1978
7. Engelbrecht, H. L., "Plant Engineer's Guide to Electrostatic Precipitator
Inspection and Maintenance", Plant Engineering, April 1976
8. Eisert, W., "Electrostatic Precipitators for the Collection of Dust from
Gases in Non-ferrous Metallurgical Smelters". Paper presented at the
8th Metallurgical Seminar of the Society of German Metallurgical and
Mining Engineers - Arolsen, Germany, March 1978
9- Eisert, W., "Aspects of Gas Cleaning Systems for Roaster Gases from
Turbulent Layer Roasters for Zinc Blende". Paper presented at
Symposium of the Society of German Metallurgical and Mining Engineers -
Zinc Committee - Frankfurt, Germany, April 1977
226
-------
ALUMINUM INDUSTRY
Session Chairman: Seymour Epstein
The Aluminum Association
Washington, D.C.
227
-------
AN INTRODUCTION TO THE PRIMARY ALUMINUM
INDUSTRY AND ITS COMMONLY EMPLOYED
EMISSION CONTROL METHODS
Lawrence C. Tropea, Jr.
Reynolds Metals Company
Richmond, Virginia 23261
ABSTRACT
Aluminum bearing ores make up about eight percent of the earth's crust.
Aluminum occurs in nature in a variety of forms but the most important natu-
rally occuring aluminum bearing ore is bauxite.
Alumina is extracted from bauxite by the Bayer process. Alumina is then
converted to aluminum in primary aluminum plants.
Alumina is electro-chemically reduced in primary aluminum plants to
aluminum by the Hall-Heroult process. This process basically involves two
steps. First, alumina is dissolved into molten cryolite. Second, an electric
current is passed through this mixture, thus decomposing the alumina into
metallic aluminum.
The electro-chemical reaction which takes place to produce aluminum is
carried out in electrolytic cells or "pots". Aluminum reduction cells are of
two general types: Soderberg cells and prebake cells.
In many primary aluminum plants, cell hoods are employed to- capture the
particulate matter and gases evolved from the reduction cells. These captured
gases are, at some prebake plants, directed to an air pollution control sys-
tem consisting of dry scrubbers. At many Soderberg plants, captured cell
gases are directed to wet electrostatic precipitators. Both dry systems and
wet electrostatic precipitators have demonstrated a high degree of efficiency
in removing particulate matter.
INTRODUCTION
Aluminum is the "magic metal".r The following excerpt from a recent
article in National Geographic accurately captures the essence of aluminum
and provides insight into the ever increasing role which aluminum is playing
in the world.
228
-------
"It is 1942: A Nazi U-boat surfaces in foggy darkness off a
lonely Long Island beach and four invaders land stealthily.
Armed with sophisticated explosives, the saboteurs seek to
cripple America's burgeoning air armada. Among the targets -
aluminum smelters in New York and Tennessee. Their plan -
destroy the cables that carry electricity for processing
the aluminum. Long before power can be restored, the molten
metal will solidify in the furnaces (reduction cells) so that
only blasting can remove it, knocking out the smelters for
months.
Before the enemy agents can strike, the FBI scoops them up.
But the lesson is clear. The light, shiny metal that began
its working life as kitchen pots and pans has emerged as a
sinew of industrial society." *
The Aluminum Association has recently launched a program to make the
public more aware of the essential role which aluminum plays in each American's
life. The following excerpt points out that aluminum plays an important role
in each citizens day-to-day activities and is vitally important to our na-
tion's efforts to recycle and reuse limited resources and to conserve energy.
"America needs aluminum for cutting down fuel consumption in
transportation. For packaging. For building. For recycling.
For transmitting electricity. For protection from the heat
and cold. America needs aluminum because it's light weight
and strong and versatile. In fact, Americans need aluminum
for a thousand different reasons every day of the year."2
The remainder of this paper will provide an overview of the primary
aluminum industry and the emission control techniques which are commonly
employed. By necessity, the paper will be directed towards providing the
reader with a general overview of the subject matter.
PRIMARY ALUMINUM CAPACITY
The primary aluminum capacity in the United States, at the end of 1977,
was 5,193,350 (short) tons. The worldwide primary aluminum capacity was
approximately 15,000,000 (short) tons at the end of 1977.,3 As such, for the
year 1977, the primary aluminum industry in the United States accounted for
approximately thirty-five percent of the worldwide primary aluminum capacity.
The companies which produce primary aluminum in the United States, and
their respective rated capacities, are shown in Figure 1. "The percentage
distribution of U.S. primary capacity by company at year-end 1977 was: Alcoa
32.3; Reynolds 18.8; Kaiser 13.9; Consolidated 6.8; Anaconda 5.8; Alumax 4.2;
Howmet 4.2; Martin Marietta Aluminum 4.0; Revere 3.9; Noranda 2.7; National
Aluminum 1.7; and Southwire 1.7."
229
-------
Figure 1. PRIMARY ALUMINUM PRODUCTION CAPACITY 3
AT THE END OF 1977
(Short Tons)
Company
Aluminum Company of America
Alumax
Anaconda Aluminum Company
Consolidated Aluminum Corporation *
Martin Marietta Aluminum
Howmet Corporation
Kaiser Aluminum & Chemicals Corporation
National-Aluminum
Noranda Aluminum Company
Revere Copper & Brass, Inc.
Reynolds Metals Company
Southwire
Annual Capacity
1,675,000
218,700
300,000
351,950
210,000
218,700
724,000
90,000
140,000
200,000
975,000
90,000
5,193,350
* Includes the aluminum capacity acquired from Olin Corporation in 1974,
230
-------
GEOGRAPHICAL DISTRIBUTION OF PLANTS
Figure 2 shows the graphical distribution of primary aluminum plants
located in the United States. ** As depicted in Figure 2, there is a concentra-
tion of primary aluminum plants in the Pacific Northwest, the Tennessee Valley
region, the Ohio Valley region and the Gulf Coast area. The locations of the
existing plants were selected after careful consideration of the numerous
factors which must be considered in the site selection process; however, the
availability of an adequate supply of electrical energy has traditionally
been an important factor in site selection decisions.
GENERAL DESCRIPTION OF THE PRIMARY ALUMINUM REDUCTION PROCESS
Basic Cell Construction and Operation
In the Hall-Heroult process, aluminum is produced by the electrolysis
of alumina dissolved in a fused salt bath. The reduction of alumina to
aluminum is carried out in reduction cells, or, as they are commonly called,
pots. A reduction cell is essentially a rectangular steel shell which is
lined with heat insulating materials and with carbon blocks. The carbon liner
of the cell serves as the cathode for the electrolytic cell. The carbon liner
contains collector bars which are connected to a collector bus, which is, in
turn, connected to the main potline service bus.
The anode, or anodes, of a reduction cell are also composed of carbon.
The anode is suspended over the cell and is immersed in the molten electrolyte
bath.
Individual reduction cells are electrically connected in series and are
typically housed in buildings, commonly referred to as potlines. A single
plant typically has two to as many as seven potlines with up to 1000 cells.
The basic reduction process begins when the cell is fed a supply of
cryolite and alumina. Direct electrical current is then passed through the
alumina/cryolite solution causing the solution to decompose into metallic
aluminum and oxygen. The aluminum migrates to the bottom of the cell where
it can later be siphoned off. The oxygen released in the decomposition re-
action combines with the available carbon and is released from the cell.
The two materials consumed in the reduction process are alumina, which is
periodically fed to the cell, and carbon, which is consumed from the anode.
A small amount of cryolite is also consumed in the electrolytic production
of aluminum. A general material flow diagram for a primary aluminum plant
is shown in Figure 3.
The reduction process is carried out in two main types of reduction cells,
prebake and Soderberg. The characteristics of each type of cell are des-
cribed in the sections which follow.
231
-------
Figure 2. GEOGRAPHICAL DISTRIBUTION OF
PRIMARY ALUMINUM PLANTS'*
to
u>
tO
i- Signifies Primary
Aluminum Plant
Location
-------
Figure 3. MATERIAL FLOW DIAGRAM FOR A
PRIMARY ALUMINUM PLANT 6
CO
Carbon Plant
Potroom
Refining
and Casting
n
Shipment
Raw Material
Cryolite
-------
Prebake Cells
Two basic variations of prebake (PB) cells are employed in the primary
aluminum industry. One type is called a center-work prebake cell while the
other is called a side-work prebake cell. These two types of prebake cells
are illustrated in Figures 4 and 5.5
Prebake cells, as their name implies, employ prebaked carbon blocks to
serve as the cell anodes. The prebaked anodes are pressed, or jolted, from
a mixture of petroleum coke and pitch binder and baked prior to use in the
cell. The prebake anodes are progressively lowered into the electrolyte as
they are consumed and, of course, eventually replaced by new anodes.
The two variations of the prebake cell which are employed in the primary
aluminum industry are, as previously mentioned, center-work and side-work
cells. A center-work prebake cell, which is shown in Figure 4, derives its
name from the manner in which the cell is "worked". In the case of a center-
work prebake cell, alumina is fed to the cell along its centerline and the
crust is also broken along the centerline of the cell.
In the case of a side-work prebake cell, the cell is "worked" from the
side. A side-work prebake cell is shown in Figure 5.
Soderberg Cells
Soderberg cells are also commonly found throughout the primary aluminum
industry. The two types of Soderberg cells used in the industry are the
horizontal stud Soderberg (HSS) cell and the vertical stud Soderberg (VSS)
cell. These cells are illustrated in Figures 6 and 7.5
In both types of Soderberg cells, the carbon anode is contained in a
rectangular steel shell suspended over the cell. The anode can be lowered
and, as carbon is consumed, carbon paste is added periodically to the top
surface of the anode. The heat of the cell progressively bakes the new ma-
terial as the anode is lowered.
Figure 6 shows the general arrangement of a vertical stud Soderberg
(VSS) cell. The steel studs carrying electrical current to the cell are in-
serted into the anode on a staggered schedule and become firmly baked into
position as they approach the cell working surface.
In this type of cell the anode casing typically carries a small hood
or skirt which collects the gases evolved around the periphery of the anode.
These gases are normally burned in a special type of burner located close
to the "gas skirt", as shown in Figure 6.
The horizontal stud Soderberg (HSS) cell is shown in Figure 7. It
is similar to the VSS cell, except that the steel studs are inserted through
the long sides of the anode casing. Crust breaking operations, and the
addition of new alumina, are carried out along the sides of the cell, as is
the case with VSS cells.
234
-------
Figure 4.
CENTERWORK
PREBAKE CELL
Alumina Hopper
Gas Offtake
Hood
Carbon Anode
Crust
Molten Bath
Molten Aluminum
Carbon Cathode
Figure 5.
SIDEWORK
PREBAKE CELL
235
-------
Figure 6.
VERTICAL STUD
SODERBERG CELL
Carbon Anode
Anode Studs
Fzgure 7.
HORIZONTAL STUD
SODERBERG CELL
236
-------
EMISSION CONTROL IN THE PRIMARY ALUMINUM INDUSTRY
Primary Fume Collection
In the primary aluminum process, certain gaseous and solid constitu-
ents are evolved from the reduction cells. The amount and type of evolved
constituents is a function of a number of operational conditions such as:
bath ratio, bath temperature, bath chemistry, and the frequency of anode
effects.
The major constituents evolved from the reduction cells are particu-
late matter along with gaseous and particulate fluoride. The exact amounts
of each of the constituents evolved will, as previously stated, vary with a
number of operational conditions. In fact, the evolution rate would be
expected to vary, for a variety of reasons, even within an individual plant
employing uniform operational conditions.
Efficient cell hoods can be employed, in cases where they are neces-
sary, to capture a high percentage of the constituents evolved from the
individual reduction cells.* The small portion of the evolved constituents
which escape capture by the primary collection system are released through the
potroom roof ventilators. The control of these secondary emissions is not
routinely practiced in view of the .low concentration of constituents found in
the large volume of air exhausted from the potroom roof ventilators. Further,
control of secondary emissions is not routinely practiced since it is not
cost or energy effective. Control of the dilute, large volume secondary air
stream requires a large energy and financial investment to achieve a relative-
ly small amount of additional emission reduction.*
The basic purpose of the primary collection hoods is, of course, to
capture the constituents evolved from the cells and direct them to the pri-
mary emission control system. A variety of successful hood designs have been
employed within the primary aluminum industry. Typically, hood design efforts
require a considerable amount of engineering and research time in order to
tailor the hood design to each individual plant's physical facilities and
It should be noted that cell hoods are not necessary or warranted in some
individual situations. Fume collection for VSS cells is, in many ways,
dissimilar to fume collection for other types of cells. For this reason,
fume collection for VSS cells will not be addressed in detail in this
paper.
The conclusion as to the general infeasibility of secondary emission con-
trol is supported by EPA's proposed position on emission guidelines for
existing primary aluminum plants. In the proposed guidelines, EPA concludes
that secondary control for HSS and CWPB "does not appear justified" and
that secondary control at VSS and SWPB plants is only justified depending
"on (the) severity of the fluoride problem," ("Draft Guideline Document:
Control of Fluoride Emissions from Existing Primary Aluminum Plants,"
U.S. Environmental Protection Agency, June 1978, Table 1-4).
237
-------
Primary Emission Control Equipment
Over the years, a variety of air pollution control equipment has been
successfully employed in the primary aluminum industry- In recent years,
the most common control techniques employed in the primary aluminum industry
have involved the installation of spray towers, wet electrostatic precipita-
tors (WESP's) and dry scrubbing systems. It should be noted, however, that
these types of control systems may not be appropriate or necessary at all
plants.*
Spray Towers -
Spray towers have long been a traditional method of control in the
primary aluminum industry. A number of plants currently successfully employ
spray towers to reduce atmospheric emissions. "The term spray tower is
applied to gas scrubbing devices in which the gas passes through an enclosure
at relatively low velocity and is contacted by water, alkaline liquor or limed
water liquor, sprayed from headers usually in counterflow with the gas . . .
A typical spray tower . . . uses water or limed water and consists of an open
top redwood tower . . . with cyclonic inlet breeching and a mist eliminator
at the top. Liquor may be sprayed down from the top or at several elevations
in the tower."*
Spray towers have been found to achieve excellent reductions, in the
range of 93 to 99 percent, in the gaseous fluoride level of the captured
gases. Singmaster and Breyer have reported that particulate removals of 64
to 80 percent can be achieved by spray towers at primary aluminum plants."*"
Wet Electrostatic Precipitators -
The gases captured by the cell hoods can be directed to wet electrostatic
precipitators (WESP's). "The electrostatic precipitator is a relatively large
chamber through which cell gas streams pass at low velocity, usually 3 to 5
feet per second (ft/sec). In its usual form, high negative voltage corona
06
In some specific plant locations the high level of removal provided by
WESP's, dry scrubbers, or other equipment is not warranted or necessary.
A number of factors must be considered in developing a control strategy for
an individual plant. Among the factors which must be considered are local
conditions, economics, dispersion characteristics, plant operational prac-
tices, courtyard space limitations, equipment maintenance considerations,
etc.
Singmaster and Breyer caution that care should be exercised in applying
the reported removal efficiencies for spray towers since the data base
for their conclusions was- sparse and because of limitations in the col-
lected data.
238
-------
discharge wires are suspended across the air stream and grounded collector
plates form parallel passageways for the air. The ionizing field surrounding
the discharge wires ionizes part of the gas stream and imparts electric
charge to most particles, some positive but most negative. Positively charged
particles migrate toward the discharge wires and negatively charged particles
migrate to the grounded collection plates." 6 Water sprays are then typically
employed to wash the collected materials from the plates.
Wet electrostatic precipitators typically have a scrubbing section in the
inlet section of the unit. The water sprays,,among other things, help to con-
dition the gas stream. Gas conditioning may be very important since some
experts believe that improved removal is achieved, for certain types of par-
ticulate, by raising the moisture content of the gas stream.6 Gaseous fluo-
ride removal is, of course, also enhanced by the water sprays.
Singmaster and Breyer state that wet electrostatic precipitators "...
operating on potline primary effluent streams report design and operating
removal efficiencies ranging from 60 percent to 99 percent on particulate.
Unlike many types of control equipment, electrostatic precipitators may be
designed for almost any selected efficiency. By using conservative design
dimensions, by controlling (the)humidity of the incoming gas, and by operating
at high voltage. . . (wet electrostatic precipitators) can achieve 98 to 99
percent removal of potline effluent particulates."'1*
In its deliberations on establishing Section llld guidelines for existing
primary aluminum plants, EPA has proposed that "... spray tower-electrostatic
precipitator combinations can readily achieve 98.5 percent total fluoride
removal." 6 Further, it should be noted that EPA is presently considering
establishing 95 percent as the target removal efficiency for wet electrostatic
precipitators alone. *
Dry Scrubbing Systems **
Dry scrubbing systems have also been successfully applied at a number of
primary aluminum plants. Dry scrubbing systems typically employ alumina to
absorb gaseous fluoride along with mechanical separation to remove particu-
late matter. The three types of dry scrubbing systems employed in the indus-
try are: coated filter dry scrubbing system, fluid bed dry scrubbing system,
and injected alumina dry scrubbing system. The names of the individual types
of dry scrubbing systems are derived from their basic method of operation.
At some existing plants which have recently installed new WESP installa-
tions, space has not been available to leave existing spray towers. EPA
has, however, stated that the level of control corresponding to an in-
stallation of spray towers-electrostatic precipitators is not mandatory
except in special situations where it may be warranted because of local
conditions. This policy was announced by EPA technical staff at the
22 August 1978 meeting of the National Air Pollution Control Techniques
Advisory Committee,
239
-------
In all cases, the alumina used in the dry scrubbing system is eventually re-
turned to the reduction cells. This recycling of the reacted alumina allows
the recovery of the major portion of the fluoride value of the captured gas
stream without, in most cases, the need for elaborate further processing.
In the coated filter dry scrubbing system, alumina is injected into the
gas stream to form an alumina layer on the fabric filter bags. The alumina
layer is then available to absorb gaseous fluoride as the gas stream migrates
through the alumina layer. The alumina layer also traps a large portion of
the particulate matter in the gas stream.
The fluidized bed dry scrubbing systems, as their name implies, employ
a fluidized bed of alumina. In this type of system alumina is continuously
directed to a fluidized bed reactor and the reacted bed material is allowed
to overflow the system. The fluidized bed traps a large amount of the par-
ticulate matter in the gas stream and, as previously noted, the alumina
absorbs the gaseous fluoride. A baghouse is located above the reactor bed to
remove the alumina and trapped particulate matter from the gas stream.
Injected alumina dry scrubbing systems have also been successfully uti-
lized in the primary aluminum industry. In this type of system, alumina is
injected into the gas stream and provided with an adequate opportunity for
contact with the gas stream. The alumina laden gas stream is then directed
to bag filters.
Singmaster and Breyer report that dry scrubbing systems are capable of
achieving from 97 to 99 percent removal of particulate matter.4 Gaseous
fluoride removal efficiencies of up to 99 percent have been reported for dry
scrubbing systems.6
In its deliberations on establishing Section llld guidelines for existing
primary aluminum plants, EPA has proposed that dry scrubbing systems, can
achieve 98.5 percent total fluoride removal. Further, it should be noted
that EPA is presently reconsidering the issue and will likely establish the
control efficiency range for dry scrubbing systems at 95 to 98.5 percent.*
FUTURE PERSPECTIVES ON ALUMINUM
The future for aluminum is bright and aluminum justly deserves being
called the "magic metal". "Indeed, just as earlier ages of human development
have taken their names from the distinctive material that nurtured them -
Stone, Bronze, Iron - there are those who believe our era may be called the
Aluminum Age." Aluminum "... offers a shining hope for energy conserva-
tion. Once made, the metal can be recycled over and over for only a fraction
of the energy used in making it originally. Also, by putting our overweight
autos on a strict aluminum diet, we can drastically reduce their weight and
thus their thirst for gasoline "
This policy was announced by EPA technical staff at the 22 August 1978
meeting of the National Air Pollution Control Techniques Advisory Committee.
240
-------
An interesting perspective on the future course of society and the role
which aluminum will play was outlined in the recent National Geographic, article.
Glenn T. Seaborg, former chairman of the Atomic Energy Commission and a Nobel
Prize-winning chemist is quoted to say that he ". . . envisions a society
in which 'the present materials situation is literally reversed; all waste
and scrap - what are now called secondary materials - become our major re-
sources , and our natural, untapped resources become our back-up supplies.'
. . .When and if this day comes, versatile aluminum will be playing a role -
or maybe two or three." *
FUTURE PERSPECTIVES ON ENVIRONMENTAL CONTROL
In the environmental area, the primary aluminum industry can be proud of
its recent history. The industry has, where necessary, developed innovative
approaches to controlling emissions. The development of functional hooding
systems and the successful application of wet electrostatic precipitators
and dry scrubbing systems are significant achievements for which the industry
can be justifiably proud. The challenges facing the primary aluminum industry
in the future will likely center around continuing to optimize the operation
of environmental control facilities and their relationship to the process.
Further, attention will likely continue to be directed toward the challenges
of additional recycling.
ACKNOWLEDGEMENTS
The able assistance of Ms. Janet Ashman in the preparation of this manu-
script is greatly appreciated.
REFERENCES
1. Canby, Thomas Y., "Aluminum, the Magic Metal," National Geographic,
August 1978, pages 186-211.
2. Aluminum Association, "American Needs Aluminum".
3. Aluminum Association, "Aluminum Statistical Review 1977".
4. Singmaster and Breyer, "Air Pollution Control in the Primary Aluminum
Industry," PB-224 282, Report No. EPA-450/3-73-004a, Volume I, July 1973.
5. International Primary. Aluminum Institute, Environmental Committee Report,
"Fluoride Emissions Control: Costs for New Aluminum Reduction Plants,"
April 1975.
6. U.S. Environmental Protection Agency, Draft Guideline Document: Control
of Fluoride Emissions from Existing Primary Aluminum Plants, June 1978.
241
-------
PARTICULATE EMISSION CONTROL IN AN AUTOMATED SIDE-WORK
PRE-BAKE ALUMINUM REDUCTION FACILITY
William J. Janson
Eastalco Aluminum Company
Frederick, Maryland 21701
ABSTRACT
Automated and advanced technology process equipment in an aluminum
reduction potline can require added maintenance and supervisory control to
continuously assure maximum particulate collection. At the Eastalco Alum-
inum Company reduction plant at Frederick, Maryland, alumina feeding and
crust breaking is accomplished with an internally programmed computer
operated pot working machine. The computer opens the pot hoods when working
the pot is required and closes the pot hoods on completion of the work.
Slight mechanical adjustment changes during the day of either the pot working
machine or hoods can result in damage to the hoods by the pot working ma-
chines. Such damage, if not corrected, could result in loss of negative
draft under the hoods and loss of particulate to the less efficient secon-
dary scrubbers. Manned pot tending machines utilized to change anodes, tap
metal and add or remove electrolyte, also require close operator and super-
visory attention due to close tolerances around the pot hoods. Eastalco
employs a maintenance crew to respond to required adjustment changes and to
repair the pot hoods due to equipment malfunction or operator error.
INTRODUCTION
Eastalco Aluminum Company operates a two potline primary aluminum reduc-
tion facility south of Frederick, Maryland, with an annual capacity of
176,000 tons. The plant began operation as a one potline facility in 1970.
The second potline was constructed over the 1974-1975 period and became fully
operational in 1976. The Hall-Heroult reduction process is utilized to pro-
duce primary aluminum in 480 cells or pots.
The original one potline configuration utilized Pechiney designed side-
work pre-bake pots.1 Individual pots were not hooded. Control of emissions
utilized the potroom as a containment device with pot emissions being collected
242
-------
in sixty-four (64) spray screen wet scrubbers located in the roof of the pot-
room buildings. Effluent gas flow was maintained by exhaust fans in each
roof scrubber producing a negative pressure in the potrooms.2
Beginning in 1973, the original potline was retrofitted with hoods over
the individual pots. The hooded effluent was ducted to injected alumina
(A1203) dry scrubbers to provide a primary control system. Twelve (12) of
the roof wet scrubbers were removed to provide the required room for primary
system ducts. The remaining wet roof scrubbers were retained as a secondary
control system.3 The second potline was constructed utilizing side-work
pre-bake pots and the identical primary and secondary control systems retro-
fitted to the original potline.
The retrofit reduced particulate emissions from the original potline by
sixty three (63)% f and the entire two potline complex had twenty seven (27)%
less particulate emissions than the original one potline complex.
Figure 1 shows an aerial view of the present reduction facility. The
four 1500 feet long buildings contain the two potlines. The four injected
alumina dry scrubbers can be seen standing between the potline buildings.
The stacks of the many secondary control system wet scrubbers can be noted
along the roof of each of the potline buildings. Figure 2 illustrates the
dual Eastalco emission control system.
Particulate emissions from pre-bake primary aluminum potlines are well
documented4f5 and consist primarily of alumina (A12O3). Other components
include carbon (C) , cryolite (Na.AlF,.) , aluminum fluoride (A1F.,) , calcium
fluoride (CaF2), chiolite (Na5Al3F14J and iron oxide (Fe2O3).
Installation of best achievable control technology at Eastalco obviously
had a marked effect on reduction of particulate emissions. As is usually the
case, further reductions are achievable only in small increments and require
additional demands on both personnel and equipment. Emission control, main-
tenance and potline personnel at Eastalco are showing success at achieving
those increments of progress.
SIDE-WORK PRE-BAKE REDUCTION CELL HOODS
As the development of primary aluminum reduction cells or pots progressed
since the Hall-Heroult discovery, a divergence in the pre-bake technology oc-
curred between the United States and Europe. In the United States, alumina
feeding and metal tapping was accomplished down the center of the pot and the
anodes were widely separated to provide for these activities. Only anode
changing required access to the side of the pot. In Europe, the pre-bake pot
was designed so that all activities occurred on the sides of the pot and the
anodes were placed closer together on the superstructure. Automated pot work-
ing machines and manually operated pot tending machines were developed to
accomplish the many pot activities along the sides of the pot. The United
States pre-bake pot was termed the "center work pre-bake" pot and the European
pre-bake pot became known as the "side-work pre-bake" pot.6
243
-------
' '
J.
FIGURE 1. Aerial view of Eastalco primary aluminum reduction facility.
-------
to
£*•
cn
EMISSION CONTROL SYSTEM
FIGURE 2. Eastalco Primary and Secondary Emission Control System.
-------
The center work pre-bake pot was very amenable to hooding which provided
a relatively low volume high pollutant concentration primary collection stream
allowing treatment in classical control systems. Crust breaking and alumina
feeding mechanisms utilizable during hood closure, metal tapping through a
small end door and removal of only a portion of the segmented hood during
anode change were added features in the center work pre-bake pot.
Side-work pre-bake pots, on the other hand, were difficult to hood due
to the physical nature of the superstructure and frequent access requirements
to the length of both sides of the pot by the pot working and tending ma-
chines. The earlier approaches to pollutant control were to use a higher
than normal ventilation volume and to treat the ventilation air in roof-
mounted wet scrubbers. The entire potroom thus became a hood as such for
the pots contained therein.
Beginning in 1970, an extensive effort began to develop a hood for the
side-work pre-bake pot. Approximately a dozen different concepts were tested
before an acceptable hood was achieved. This allowed hooding of the Eastalco
side-work pre-bake pots and installation of a primary control system. This
primary control system together with the retained secondary control system
then provided a total control system equivalent to the more efficiently
hooded center-work pre-bake pot with a primary system control alone.
The features of the hood include a permanent top pan mounted to the
superstructure beneath the anode bus. The pan is slotted to allow anode stem
to bus contact. Stem seals are mounted on the anode stems to seal off the
slotted areas of the pan and, at the same time, to allow stem movement during
jacking operations. Movable side doors operated by electrical motors and
associated gear boxes provide access to the pot when required by pot working
and tending machines. Permanent end panels and wing panels on each end of
the pot complete the hood closure. Effluent fume removal is via ducts
mounted to the top pan. Figure 3 shows a closed hooded pot. Figure 4 de-
picts an Eastalco potroom with one side door open on the pot during metal
tapping with a pot tending machine. The ducts carrying the pot effluent to
the dry scrubbers can be noted between the pot and the wall on the right.
POT HOODS AND AUTOMATED POT WORKING MACHINES
Automated pot working machines termed "semi-gantries" conduct the func-
tion of alumina feeding and crust breaking at the Eastalco plant. The semi-
gantry conducts its functions on both sides of the pot at the same time and
is monitored and controlled by the potline process computer system. The
computer has the capability of conducting the semi-gantry activities in
various ways, depending upon how the computer is programmed and also sends
the semi-gantry to centrally located stations to re-fill its self-contained
ore bins when required.
Installation of hoods on the Eastalco pots required modifications to both
the pot superstructure and the semi-gantries. The entire superstructure was
raised to reduce the potential of top pan damage by the crust breaker during
its upward stroke. Since the semi-gantry breaks and feeds both sides of the
246
-------
•
I
FIGURE 3. Side-work pre-bake hooded pot.
-------
1- ;
4.
00
FIGURE 4. Eastalco potroom with one hood door open for tapping operation.
-------
pot at the same time, both hood side doors are required to be in a semi-open
position. The bridge on which the semi-gantry travels was raised to provide
room between the bridge and the superstructure to accommodate both hood doors
in this semi-open position. Hood door opening and closure for semi-gantry
operation is also controlled by the potline process computer. Unless both
hood doors are in their correct open position, the computer does not allow
semi-gantry operation along the sides of the pot. Figure 5 shows the semi-
gantry breaking the crust with the hood doors in the semi-open position.
With the hood retrofit, an area with very close tolerances exists be-
tween the moving semi-gantry and the tubular arms on the hood side doors as
they are raised by the computer to facilitate semi-gantry operation. Figure
6 depicts this very close tolerance. If the semi-gantry trolley hits the
tubular arm nut, the tubular arm can be bent which causes binding of the door
motion. This causes further strain or warping of the pivot points and gear
box mechanisms and therefore improper door closure. The nut can be completely
sheared off and cause the door to fall down on that end of the pot. Limit
switches on the semi-gantry do not allow motion of the semi-gantry if the
trolley moves forward toward the pot. Setting of these limit switches is
sometimes difficult due to wear of the trolley rails, which can cause the
trolley to drift forward. The weight of the cables carrying electrical ser-
vice to the semi-gantry along the bridge also has enough weight to pull the
trolley forward if they are not properly supported.
Other factors influencing hood damage during semi-gantry operation in-
clude damage to the top pan from the crust breaker if the stroke of the crust
breaker is out of proper adjustment. Likewise, if the clutches slip on the
hood door mechanism, the doors can fall on the semi-gantry while it is working
the pot. This results in damage, including loss of alignment of the hood door.
Hood damage, if allowed to occur and to routinely continue to occur, can
result in a loss of negative draft under the hoods and therefore reduced pick-
up of the particulate effluent by the highly efficient primary control system,
resulting in collection of a portion of the particulate by the less efficient
secondary control system. This problem is magnified if repairs to the hood by
the hood repair crew require portions of the hood to be removed from the pot.
As the above noted difficulties have manifested themselves since the
hood retrofit, additional supervisory and maintenance efforts have been
initiated to inspect and make the required repairs and adjustments on a daily
basis. These repairs and adjustments can include replacement of worn trolley
rails, proper setting of limit switches, and adjustment of cable supports,
crust breaker stroke and clutches on the hood door mechanisms. These efforts
have resulted in a significant reduction in hood damage and they are playing
an important role in the achievement of reduced particulate emissions from
the Eastalco potlines since the hood installation.
POT HOODS AND MANNED POT TENDING MACHINES
Pot operations such as anode changing, aluminum metal tapping and trans-
fer of liquid electrolyte at Eastalco utilize manned pot tending machines.
These manned pot tending machines can also perform the semi-gantry functions.
249
-------
I
I '
FIGURE 5. Semi-gantry breaking crust with hood doors in semi-open position,
-------
Semi-gantry trolley
Tubular arm nut
FIGURE 6. Close tolerance between semi-gantry trolley and hood tubular
arm nut.
251
-------
Due to close tolerances around the hoods, close operator attention is
continuously required to prevent damage to hood components. A responsibility
of the potroom supervisor is to assure that the operator is adequately trained
to perform the required functions correctly and understands the consequences
of hood damage.
During anode change, the fixed top pan can be struck by the anode assembly
if excess crust is not swept out of the way prior to moving the assembly into
position. Moving the anode stem into the assigned slot in the pan also re-
quires operator diligence to assure that the pan is not bent by the stem.
Figure 7 shows an anode assembly being moved into place. Similar problems
occur while tapping metal or transferring liquid electrolyte. The tap hole
must be located correctly and be large enough so the spout does not hang up
during insertion or removal. The tapping spout can strike the pan and bend
it or break the pan's angle supports and the pan will then sag. Figure 8
shows the proper tapping procedure.
The wing panels on the ends of the pot can also be damaged during chang-
ing of end anodes. Operator care and attention are required to assure that
such anodes are set in without striking these wing panels and requiring repair
or replacement by the hood repair crew.
During anode change, metal tapping and liquid electrolyte transfer, the
hood side door,is opened on only one side of the pot. However, to accomplish
these required activities, the hood side door is moved to its full open
position. This full open position is best illustrated in Figure 4. During
anode changes, it is important that the crane also be moved inward with ex-
treme care. If the crane strikes the bottom of the hood side door during
its inward movement, the door is pushed across the top of the superstructure.
This can result in considerable damage to both the door and the arm supports.
Considerable accomplishment has been achieved in supervisor training to
assure operator care when using pot tending machines around the retrofitted
pot hoods. Operator experience over time is also a factor in reducing hood
damage. The problems discussed herein are becoming the exception rather than
the rule. As was the case with the automated pot working machines, these
improvements are reducing particulate emissions from the Eastalco potlines.
SUMMARY
The installation of hoods and ducting of the hoods to injected alumina
dry scrubbers to provide a primary control system and the retention of the
existing wet roof scrubbers as a secondary control system reduced particulate
emissions from Eastalco's original side-work pre-bake potline by 63 %. Con-
struction of the second potline identical to the environmentally retrofitted
first potline resulted in 27 % less particulate emissions from the two potline
complex than was the case with the one potline complex.
Further reductions have been achieved in small increments. Figure 9
illustrates the reduction in grain loading from the secondary wet scrubbers
252
-------
tv)
Ul
LO
FIGURE 7. Anode assembly being moved into place on the pot.
-------
! •
'••
'
FIGURE 8. Aluminum metal tapping.
-------
CO
\
o
O
CD
(0
I
w
§
H
CO
W
H
S
H
3.0
2.5
2.0
1.5
1.0
0.5
I
I
1976
1977
1978
YEAR
FIGURE 9. Secondary control system particulate emissions,
255
-------
over the past three years. Such an achievement was accomplished by increasing
the hood collection efficiency of generated particulate effluent. This in-
volved increased supervision, maintenance and operator education efforts to
assure rapid repairs, frequent equipment adjustments and operator care and
attention to prevent hood damge during pot operations using automated pot
working machines and manned pot tending machines.
REFERENCES
1. Peria, Paul. "New Pot Fume Control and Cleaning Methods",
Revue de L1Aluminium No. 428, April 1974, p.8.
2. Oehler, R. H. "Emission of Air Contaminants in Aluminum Electrolysis",
paper A-70-11 presented at the TMS-AIME Annual Meeting, Denver,
February 16-19, 1970.
3. Mantle, E. c. "Pollution Problems of the Primary Aluminum Industry",
In: Prevention of Air Pollution in the Non-Ferrous Metals Industries,
BNF Metals Technology Centre, August 1974, p. 153-157.
4. Grjotheim, K., H. Kvande, K. Motzfeldt and B. J. Welch, "The Formation
and Composition of the Fluoride Emissions from Aluminum Cells",
Canadian Metallurgical Quarterly, Volume 11, Number 4, 1972, p. 585-598.
5. Singmaster and Breyer, "Air Pollution Control in the Primary Aluminum
Industry", Volume I, Sections 1 through 10, Environmental Protection
Agency, July 1973.
6. International Primary Aluminum Institute. "Fluoride Emission Control:
Costs for New Aluminum Reduction Plants", April 1975, p. 24-25.
256
-------
EMISSION CONTROL IN VERTICAL
SODERBERG FURNACES
J. L. Byrne
Martin Marietta Aluminum Incorporated
The Dalles, Oregon
ABSTRACT
A practical vertical stud Soderberg anode for aluminum
reduction cells was developed in the late 1930*s. The rigid
anode casing allows the fixture of a collection skirt that
captures the cell off-gases in a small volume very near the
working face of the anode. Potential capture rate is somewhat
less than that inherent in hooded prebake and side pin Soderberg
cells. Sources of particulate and size distribution are not
significantly different from those of the other types of reduc-
tion cells. Overall particulate control may require primary
(pot gas) and secondary (room air) control devices. Several
control schemes are discussed.
(We regret that the manuscript for this paper
was not received for publication in this
proceedings for reasons beyond our control.)
257
-------
EXPERIENCE WITH THIRTY-ONE WET ELECTROSTATIC PRECIPITATORS
ON HSS ALUMINUM POTLINES
Leland 0. Slind
Reynolds Metals Company
Longview, Washington 98632
ABSTRACT
Thirty-one MikroPul wet electrostatic precipitators are presently
collecting emissions from the six horizontal stud Soderberg potlines at
the Reynolds Metals Company's Reduction Plant in Longview, Washington.
Gaseous fluoride and particulate collection efficiencies of the precipi-
tators are discussed. Also described is a major operational problem with
the precipitators and its effect on collection efficiencies.
INTRODUCTION
The first precipitator, hereinafter called the pilot model, began
operation in October, 1971,. It fulfilled its purpose as a pilot model
by being tested, modified and many times remodified. The information
gained from it was used in the design of the subsequent 30 precipitators.
The first of these began operation in May, 1973, with the last starting
up in July, 1975.
All precipitators are preceded by cyclonic spray towers which scrub
and cool the pots' emissions from about 240° F. to about 110° F. The flc
rate of the alkaline liquor through the spray towers is about 180 gpm.
Six of the 31 precipitators have a design gas flow of 50,000 cfm at
70° F. and the remaining 25 are designed for 100,000 cfm at 70° F.
Inlet and outlet particle and fluoride concentrations listed in this
paper were measured by methods which are modified versions of EPA Methods
5 and 13, respectively. These methods have been used concurrently with the
State of Washington's EPA approved compliance test methods and test results
were comparable. For comparative purposes and additional data the reader
is referred to the particle size and concentration measurements conducted
by Southern Research Institute on a Longview Plant precipitator. 1
258
-------
Figure 1 shows a cross-sectional view of a typical precipitator.
The channel-shaped baffles, parallel collecting plates and rigid-type
discharge electrodes are sprayed continuously with an alkaline liquor,
PRESSURIZED
HIGH VOLTAGE INSULATOR
COMPARTMENT-
_HIGH VOLTAGE
INSULATOR
ACCESS
DOORS TV
/r
/ • ^
r- / , * 4—
SPRAY
NOZZLES
INLET
BAFFLES COLLECTING
PLATE
SCRUBaNG &
GAS DSTRBUTION
BAFFLES
SCRUBaNG LIQUOR
EXIT TROUGH
tTACK
MIST ELIMINATING
BAFFLES
-\'.._
yi RIGID TYPE \
DISCHARGE ,\
ELECTRODE*-"
Figure 1. Cross-sectional view of typical precipitator
The scrubbing liquor circulates through a treatment system which
thickens the solids to about 30 percent for pumping to the Cryolite Re-
covery Plant. The clarified liquor is recirculated back to the precxpi-
tators.
259
-------
Table 1 summarizes the specifications for the precipitators at the
Longview Plant. For a more complete description of MikroPul's wet electro-
static precipitators and the theory of their operation, the reader is re-
ferred to the paper published by E. Bakke.2
Table 1. SUMMARY OF PRECIPITATORS' SPECIFICATIONS
Gas Flow
Gas Velocity
Treatment Time
Maximum Pressure Drop
Minimum Liquor Flow
@ 60 psi
100,000 cfm
-------
EFFECT OF GAS FLOW ON COLLECTION EFFICIENCY
Soon after startup of the pilot model precipitator a series of tests
were conducted to determine its collection efficiency under conditions of
increased gas flow. Gas flow was increased by diverting the emissions
from additional pots through the pilot model. The results are listed in
Table 2.
Table 2, INCREASED GAS FLOW THROUGH PILOT MODEL
Gas Flow^ cfm Inlet Particulate, Outlet Particulate,
Number of at 70 F0 gr/Dscf gr/Dscf
Tests Run Average Average Std. Dev. Average Std. Dev.
3 47,500 0.0793 0.0052 0.00273 0.00018
4 52,500 0.0912 0.0104 0.00297 0.00033
5 54,600 0.0982 0.0079 0.00367 0.00052
4 61,200 0.0875 0.0110 0.00460 '0.00054
4 62,300 0.1090 0.0246 0.00463 0.00037
Reynolds ' specifications stated that the total particulate discharg-
ing from the pilot model's stack could not exceed 0.003 gr/Dscf with a gas
flow of 48,500 cfm at 70° F0 It was believed that a particulate loading
greater than 0.003 gr/Dscf would exceed the 20 percent opacity standard
imposed by the State of Washington on point source emissions within an
aluminum reduction plant.
Table 2 clearly shows the pilot model met this specification. Gas
flow was increased nine percent, to about 53,000 cfm before the 0.003
gr/Dscf was exceeded. The opacity of the pilot model's emissions did in-
crease as gas flow increased, but even at the highest average gas flow,
62,300 cfm, opacity was no more than about 25 percent for an average out-
let total particulate loading of 0.00463 gr/Dscf.
The data from these tests were factored into the design of the second
generation precipitators of which 21 were designed for a gas flow of
100,000 cfm at 70° F. Results from increased gas flow tests on one of
these precipitators are listed in Table 3 and plotted in Figure 2.
Table 3. INCREASED GAS FLOW THROUGH 100,000 CFM PRECIPITATOR
Gas Flow, cfm Inlet particulate, Outlet Particulate,
Number of at 70 F. gr/Dscf gr/Dscf
Tests Run Average Average Std. Dev. Average Std. Dev.
3 104 600 0.09534 0.01689 0.00155 0.00004
3 126,000 0.08924 0.02124 0.00314 0.00036
3 139,800 0.06814 0.01330 0.00471 0.00046
261
-------
OUTLET FLUORIDE,
ppm
1.0
0.0
0.005
TOTAL
OUTLET
PARTICULATE,
gr/Dscf
0,003
0.001
I
100,000 120,000 140,000
AVERAGE GAS FLOW, CFM AT 70° F.
Figure 2. Increased gas flow through 100,000 cfm precipitator
Figure 2 shows that the 100,000 cfm precipitator has a greater capa-
city for collecting particulate from increased gas flows than does the pilot
model. Its gas flow can be increased about 25 percent, versus the pilot
model's nine percent, before the 0.003 gr/Dscf is exceeded. Opacity at the
highest average gas flow listed in Table 3, 139,800 cfm, was about 30 per-
cent for an average outlet total particulate loading of 0.00471 gr/Dscf.
As a matter of interest, outlet fluorides are also plotted in Figure 2.
Inlet emissions to the pilot model and 100,000 cfm precipitator, during
the increased gas flow tests, were not prescrubbed. The collector electrode
area inside both precipitators is about 295 ft.^/1000 cfm.
There are primarily two reasons why the 100,000 cfm precipitator has a
greater collecting capability than the pilot model . Its electrode alignment
and gas distribution are far better.
262
-------
Electrode alignment and gas distribution inside the pilot model were
found to be very inadequate. Modifications were made soon after startup
to improve both, but results were generally unsatisfactory. These prob-
lems were specifically addressed in the design and erection of succeeding
precipitators and both electrode alignment and gas distribution were sig-
nificantly improved.
EFFECT OF FIELD REDUCTION OK COLLECTION EFFICIENCY
A series of tests were conducted on the first 100,000 cfm precipitator
to determine if three fields were required to meet the specified 0.003
gr/Dscf. The results are listed in Table 4.
Table 4. REDUCED FIELDS INSIDE 100,000 CFM PRECIPITATOR
Number of Outlet Particulate
Number of Fields gr/Dscf Outlet Fluoride, ppm
Tests Run Operating Average Std. Dev. Average Std. Dev.
3 " 3 0.00117 0.00009 0.18 0.03
4 2 0.00205 0.00030 0.55 0.03
4 1 0.00682 0.00049 1.22 0.15
Test results show that with one field de-energized the average total
outlet particulate was only 0.00205 gr/Dscf which, of course, easily com-
plies with the 0.003 gr/Dscf specification. Opacity of the emission was
about ten percent. However, the specification was significantly exceeded
with two fields de-energized when the average total outlet particulate was
0.00682 gr/Dscf and opacity was about 35 percent.
After reviewing the results from these tests, it was tempting to fab-
ricate future precipitators with only two fields. However, it was decided
that this would be imprudent because of probable deterioration in the pre-
cipitators ' collection efficiency. This has occurred and is discussed in
a later section of this paper.
Also, electrical supply or other problems periodically occur which re-
quire the shutdown of a field. This does not cause an environmental prob-
lem because the precipitator is still able to comply with air regulations.
As it turns out, then, the third field is a good insurance policy.
All operating conditions such as gas flow, inlet particulate loading,
etc., were similar for these tests.
EFFECT OF SCRUBBING LIQUOR FLOW RATE ON COLLECTION EFFICIENCY
The scrubbing liquor flow rate into the fields of the Longview Plant's
precipitators is adjusted to a minimum--just enough to satisfactorily keep
263
-------
the electrodes clean. The reason for this is that as liquor pressure in-
creases, the number of droplets spraying from the nozzles rapidly in-
creases and their mean diameter decreases. This increases the space
charge density which, in turn, decreases the corona current for a given
voltage. This decrease in corona power decreases collection efficiency.
This phenomenon is clearly shown in Table 5 where the liquor flow
rate was varied to the inlet scrubbing chamber and to the electrostatic
fields of a non-prescrubbed precipitator. The flow rate was varied by
simply adjusting the liquor pressure.
By decreasing the liquor pressure over the fields from 80 psi (test
series A) to 50 psi (test series B), the average corona power increased
from 7.4 to 10.2 kva. As a result, total outlet particulate decreased
from 0.00458 to 0.00323 gr/Dscf.
Contrary to earlier predictions, this reduction in liquor flow did not
adversely affect fluoride collection. Outlet fluorides decreased from 0.97
to 0.71 ppm. This indicates that any loss of gaseous fluoride collection
efficiency because of reduced liquor flow over the fields is more than off-
set by the increased collection efficiency of particulate fluorides.
Test series C goes a step further in liquor reduction. Liquor pres-
sure to the inlet scrubbing chamber was also reduced from 100 psi to 50
psi. As a result, total outlet particulate increased from 0.00323 to
0.00353 gr/Dscf, but the major change was in the outlet fluorides which
increased from 0.71 to 0.99 ppm. The inlet scrubbing chamber is the major
area for contact between the alkaline liquor and gaseous fluorides and any
reduction in its liquor flow will significantly affect fluoride collection.
The gas flows listed in Table 5 are high, about 34-38 percent greater
than the 100,000 cfm for which the precipitator was designed. This pre-
cipitator and nine others were initially installed without prescrubbers.
The fans for these precipitators were intentionally oversized so that if
prescrubbers were installed, there would still be a minimum gas flow of
100,000 cfm.
The ratios of condensable hydrocarbons to solid particulate in both
the inlet and outlet grain loadings listed in Table 5 are typical. The
magnitude of the ratios will vary from test to test, but the pattern of
greater solid particulate in the inlet, and greater condensable hydro-
carbons in the outlet is typical.
This is well illustrated in Figures 3 and 4. The data in these fig-
ures is from tests on non-prescrubbed precipitators.
As would be expected, the inlet condensable hydrocarbon fraction is
relatively constant as shown in Figure 3. There are no changes in the
anode baking process that would cause a significant change in hydrocarbon
emission rate.
The same cannot be said for solid particulate which are by far the
264
-------
TABLE 5. EFFECT OF LIQUOR FLOW RATE ON COLLECTION EFFICIENCY
Test Series A
Inlet Scrubbing Chamber's
Liquor Pressure, psi 90
Electrostatic Fields'
Liquor Pressure, psi 80
Total Calculated Liquor
Flow Rate, gpm 835
Average Gas Flow,
cfm at 70° F. 134,800
Number of Tests Run 3
Average Total Corona
Power, kva 7.4
100
50
775
138,400
3
10.2
50
50
640
136,100
3
8.6
Inlet Particulate,
gr/Dscf
Average Std. Dev. Average Std. Dev. Average Std. Dev
Condensable
Hydrocarbons 0.01309 0.00029 0.01279
Solid Particulate 0.04218 0.00198 0.03793
0.00321 0.01198 0.00072
0.00935 0.04104 0.00608
Total Particulate 0.05527 0.00178 0.05072 0.01232
Inlet Fluorides,ppm 37.60 1.71 49.57 18.46
0.05302 0.00549
39.53 2.47
Outlet Particulate,
gr/Dscf
Condensable
Hydrocarbons 0
Solid Particulate 0
Total Particulate 0
Outlet Fluorides, ppm
Average Efficiency, %
Condensable
Hydrocarbons
Solid Particulate
Total Particulate
Fluorides
.00261 0.00010
.00197 0.00016
.00458 0.00021
0.97 0.03
80.1
95.3
91.7
97.4
0.00177 0.00027
0.00146 0.00017
0.00323 0.00044
0.71 0.09
86.2
96.2
93.6
98.6
0.00184 0.
0.00169 0.
0.00353 0.
0.99
84.6
95.9
93.3
97.5
00006
00006
00011
0.01
265
-------
0.15
0.10
COMPONENT
CONCENTRATION,
gr/Dscf
0.05
• SOLID PARTICULATE
O CONDENSABLE HYDROCARBONS
o0oo
O 00
OO O
O o
«o o
0.05
0.10
0.15
0.20
TOTAL INLET PARTICULATE, gr/Dscf
Figure 3. Composition of inlet particulate
0.006 I—
0.004
COMPONENT
CONCENTRATION,
gr/Dscf
0.002
• SOLID PARTICULATE
O CONDENSABLE HYDROCARBONS
o°o°
o
0 ••
0.002
0.004
0.006
0.008
TOTAL OUTLET PARTICULATE, gr/Dscf
Figure 4. Composition of outlet particulate
266
-------
major contributor to an increase in total inlet participate. Routine op-
erations on the pots such as "crustbreaking", anode "flex raises" and "pin
and channel pulls" will have a very significant effect on the solid par-
ticulate emission rate.
The reverse is true for outlet emissions from the precipitator. With
few exceptions, Figure 4 shows that the condensable hydrocarbon fraction is
the major contributor to the total outlet particulate. This is to be ex-
pected because the hydrocarbon droplet has a lower dielectric constant than
the solid particle. As a result the saturation charge on the hydrocarbon
droplet will be less, making the droplet more difficult to collect.
The uniformity of the ratio of condensable hydrocarbons to solids with
increasing total outlet particulate is interesting to note in Figure 4.
EFFECT OF SCALE ON COLLECTION EFFICIENCY
There were concerns about emissions from the pots accumulating on the
electrodes because of the reduction in liquor over the fields. However, this
problem never occurred.
What has occurred is a buildup of scale on the electrodes and inlet gas
distribution and scrubbing baffles. This scale is a densely packed, tightly
adhering and gray-brown layer that covers all surfaces which are wetted by
the scrubbing liquor.
It is not to be confused with the buildup of constituents in the emis-
sions from the pots. This buildup is a blackish-colored deposit, high in
hydrocarbons, that can be washed off by the spray discharging from the pre-
cipitators' spray nozzles. In fact, it occurs because of plugged nozzles
or nozzles not discharging enough liquor in an area.
The formation of scale was no surprise because the recirculating scrubb-
ing liquor that passes through each precipitator is alkaline (pH 8.8 - 9.4),
with a large content of dissolved solids (40,000 - 50,000 ppm). It was the
scale's rapid rate of formation that was not anticipated.
The magnitude of the scale problem can be divided into separate cate-
gories--prescrubbed and non-prescrubbed precipitators. Until recently, ten
100,000 cfm precipitators were not preceded by prescrubbers as the other 21
precipitators were. The rate of scale buildup inside these non-prescrubbed
precipitators was far greater than the rate inside the prescrubbed precipi-
tators. This is well illustrated in Table 6.
The reason for the greater buildup of scale inside the non-prescrubbed
precipitators versus the prescrubbed precipitators and the explanation for
the scale thickness gradient from inlet to outlet, that is seen in Table 6,
are discussed in a previous paper by this author.3 The paper also describes
the scale's composition and the chemical reactions involved in its formation.
Scale thicknesses of the magnitude shown in Table 6 for non-prescrubbed
267
-------
precipitators have a very significant effect on precipitator operation. Gas
flows decreased up to 22 percent because the gaps between the inlet baffles
were blinded with scale. The resulting gas pressure against the baffles
caused many of them to bow. Corona currents dropped more than 50 percent
because of the thick scale on the discharge electrodes.
Table 6. SCALE THICKNESS AFTER TWO YEARS OF OPERATION
Scale Thickness, Inches
Non-Prescrubbed Prescrubbed
Component Precipitators Precipitators
Inlet Baffles 1 1/2 3/8
Inlet Field Electrodes 3/4 3/8
Middle Field Electrodes 1/2 1/4
Outlet Field Electrodes 1/4 1/8
As a result of these effects, cleaning the non-prescrubbed precipi-
tators was necessary. The most successful method to remove scale has been
to vibrate the discharge electrodes and baffles with pneumatic chipping
hammers. The scale is brittle when it dries and will readily shake loose.
Unfortunately, the collecting plates are too stiff to vibrate the scale
loose. Even so, Table 7 clearly shows the dramatic change produced by
cleaning the scale from the inlet gas distribution and scrubbing baffles
and the discharge electrodes inside-a precipitator. Corona power increased
by 77 percent.
As a result of this increase and improved gas distribution, total out-
let particulate decreased 63 percent and fluorides decreased 54 percent—
even though there was a 12 percent increase in gas flow.
DECREASE OF PRECIPITATORS' COLLECTION EFFICIENCY SINCE STARTUP
As would be expected, the collection efficiency of the Longview Plant's
precipitators has decreased since their startup. This is graphically il-
lustrated in Figure 5 where outlet fluorides and total outlet particulate
are plotted versus time since startup. The data is from tests conducted
on the first prescrubbed 100,000 cfm precipitator to begin operation May 9,
1973.
From this date to November, 1978, five and one-half years later, Figure
5 shows the total outlet particulate increased from about 0.00080 gr/Dscf
to 0.00189 gr/Dscf. However, the current average, 0.00189 gr/Dscf, is still
well below the 0.003 gr/Dscf that was originally specified by Reynolds for
satisfactory performance. Also, the present plume opacity is less than ten
percent.
The major reason for the increased total outlet particulate is the in-
creased scale buildup which has decreased corona power by about 73 percent
since startup--from about 72.2 kva to about 19.7 kva.
268
-------
Table 7. EFFECT OF SCALE ON COLLECTION EFFICIENCY
Average Gas Flow,
cfm at 70° F.
Average Total Corona
Power, kva
Before Cleaning
141,300
6.6
After Cleaning
158,000
11.7
Inlet Particulate,
gr/Dscf
Average Std. Dev.
Condensable
Hydrocarbons 0.02517 0.00019
Solid Particulate 0.09591 0.00694
Total Particulate 0.12108 0.00713
% Decrease
Average Std. Dev.
0.02544 0.00105
0.12977 0.03416
0.15521 0.03521
Inlet Fluorides,
ppm
48.61
1.53
59.80
14.25
Outlet Particulate,
gr/Dscf
Condensable
Hydrocarbons 0.00549
Solid Particulate 0.00512
Total Particulate 0.01061
0.00030
0.00000
0.00030
66.5
59.2
63.0
0.00184
0.00209
0.00393
0.00002
0.00000
0.00002
Outlet Fluorides,
ppm 3.26
0.05
54.3
1.49
0.03
Average Efficiency,
Condensable
Hydrocarbons 78.2
Solid Particulate 94.7
Total Particulate 91.2
Fluorides 93.3
92.8
98.4
97.5
97.5
269
-------
OUTLET FLUORIDES,
ppm
0.002
TOTAL OUTLET
PARTICULATE,
gr/Dscf
0.001
0246
YEARS SINCE STARTUP
Figure 5. Increase in outlet emissions since startup
PRECIPITATORS' ELECTRICAL CHARACTERISTICS
Typical V-I curves for the three fields inside a 100,000 cfm precipi-
tator are plotted in Figure 6. The data for these curves were obtained by
manually increasing the applied voltage and recording corona current read-
ings at the selected voltage levels. For a given voltage such as 40 kv,
the outlet field has the highest current, 300 ma, followed by the middle
field, 250 ma, with the inlet field having the lowest current, 200 ma.
Such a uniform difference between V-I curves is certainly not typical
with all of the precipitators. The V-I curves vary considerably. How-
ever, the general pattern of increasing corona current toward the outlet
is typical and is explained by the higher space charge density that re-
sults from the higher particulate loadings in the inlet field.
This same effect on current because of space charge density shows up
270
-------
400
CORONA CURRENT, ma 200
INLET FIELD
MIDDLE FIELD
OUTLET FIELD
0
SPARKING
20 40
APPLIED VOLTAGE, kv
60
Figure 6. V-I curves for three fields inside a
100,000 cfm precipitator
600i—
400
CORONA CURRENT, ma
200
SCRUBBING LIQUOR PLUS EMISSIONS
SCRUBBING LIQUOR ONLY
STATIC AIR ONLY »— MAX. POWER
SPARKING
20 40
APPLIED VOLTAGE, kv
60
Figure 7. V-I curves for an inlet field during varying
conditions of particulate loading inside a
100,000 cfm precipitator
271
-------
in Figure 7 which plots V-I curves for a precipitator while energized under
three different conditions; (1) static ambient air only, (2) scrubbing
liquor only, and (3) scrubbing liquor plus potroom emissions. For a given
voltage, such as 40 kv, the highest current, 355 ma, occurs with the static
ambient air condition where, of course, the concentration of particulate is
much less than in the other two conditions.
The average and range of voltage and current readings from meters on
the field control panels are listed in Table 8. These readings are from
100,000 cfm and 50,000 cfm prescrubbed precipitators soon after their start-
up.
Table 8. PRIMARY AND SECONDARY ELECTRICAL READINGS FOR 50,000 AND
100,000 CFM PRECIPITATORS SOON AFTER STARTUP
Field
Inlet
Meter
Middle
Outlet
Prim, volts
Prim, amps
Sec. kv
Sec. ma
Prim, volts
Prim, amps
Sec. kv
Sec. ma
Prim, volts
Prim, amps
Sec. kv
Sec. ma
Average
Range
Average
Range
Average
Range
Average
Range
Average
Range
Average
Range
Average
Range
Average
Range
Average
Range
Average
Range
Average
Range
Average
Range
Readings For
50,000 cfm
Precipitators
276
230-305
43
32-60
49
46-53
149
120-200
287
200-360
61
34-90
49
44-54
236
140-310
269
230-300
61
48-71
44
42-46
243
175-320
Readings For
100,000 cfm
Precipitators
295
270-320
88
65-100
51
46-56
281
220-420
312
280-390
106
100-110
47
42-53
376
320-520
323
290-350
108
95-130
49
45-54
420
350-580
The magnitude of the readings is less now because of the previously
discussed scale buildup.
The total average power requirements of a 50,000 and 100,000 cfm pre-
cipitator are listed in Table 9.
272
-------
Table 9. PRECIPITATORS ' POWER REQUIREMENTS
Gas Flow, cfm Item Power Requirement, kw
100,000 Three Fields 36
Twelve Insulator Heaters is
One Insulator Positive Pressure
Blower, 3 hp 1.3
Two Potroom Emission Fans,
125 hp each 168
Total 223.3 kw
50,000 Three Fields 13
Twelve Insulator Heaters 18
One Insulator Positive Pressure
Blower, 3 hp 1.3
One Potroom Emission Fan,
125 hp 84
Total 116.3 kw
REFERENCES
1. Gooch, John P. and Joseph D. McCain. Particulate Collection Efficiency
Measurements On A Wet Electrostatic Precipitator. Southern Research
Institute, Birmingham, Alabama. Publication No. EPA-650/2-75-033.
2. Bakke, E. On The Application Of Wet Electrostatic Precipitators For
Control Of Emissions From Soderberg Aluminum Reduction Cells. AIME
Annual Meeting. Dallas, Texas, 1974.
3. Slind, L. Operational Problems With Thirty-One Wet Electrostatic
Precipitators. AIME Annual Meeting. Denver, Colorado, 1978.
273
-------
A LOW EMISSION PROCESS FOR THE MELT TREATMENT OF ALUMINUM ALLOYS
F. R. Mollard, J. E. Dore and N. Davidson
Consolidated Aluminum Corporation
Research and Development
St. Louis, Missouri
ABSTRACT
A low emission process has been developed for degassing, fluxing and
refining molten aluminum alloys. This process utilizes a mixture of an inert
gas, i.et nitrogen or argon, with 3 to 5% dichlorodifluoromethane (CC12F2).
CC12F2 is a non-toxic compound, commercially available under the Freon 12
trademark, among others.
The gas mixture is bubbled through the melt in coated steel pipes or
graphite tubes with or without a protective molten salt cover; the exact
nature of the salt depends upon the composition, especially magnesium level,
of the alloy. Technically, a mixture of inert gas with 3 to 5% Freon 12 is
equivalent to pure chlorine in terms of rate of removal of hydrogen and non-
metallic particulates. At the same time, and in contrast with chlorine, the
corresponding emissions levels are well below prevailing environmental
protection standards. Because of its lack of toxicity, accidental release of
Freon 12 creates no occupational hazard, as opposed to chlorine. Also, since
Freon 12 completely breaks down at molten aluminum temperature, none is
released into the atmosphere under normal operating conditions.
274
-------
Introduction
Untreated aluminum alloy melts commonly contain various impurities which
can be quite detrimental to the quality of the ingots and products fabricated
therefrom. Most troublesome are: dissolved hydrogen, non-metallic inclusions,
principally aluminum oxide, and alkali metals such as sodium.
To control these impurities, the most common approach consists in fluxing
the melt with a gas stream in order to remove the hydrogen and float up a
portion of the non-metallic inclusions. A molten metal filter, either combined
with or following the fluxing operation, serves to eliminate most of the re-
maining inclusions. The alkali metal level is controlled by using an active
gas for the fluxing operation.
This melt treatment is carried out either as a batch process in the
melting or holding furnace or as an in-line process in the transfer trough
between the furnace and the casting station. Any melt treatment must obviously
be technically effective while meeting environmental and occupational safety
standards at minimum energy consumption levels.
Several in-line processes have been developed that appear satisfactory
from both a melt cleanliness and emissions standpoint. Examples include
Alcoa's 469 (1) and flux washing (2) processes, Linde's spinning nozzle inert
flotation system (3-5) and British Aluminium's fumeless in-line degassing and
cleaning method (6). These processes fit well in new cast shop facilities
specifically designed to accommodate them. However, because of large space
requirements, relatively high capital cost and energy consumption, they are
not universally applicable. Therefore, batch melt treatment in the furnace
using simple coated steel pipes or graphite tubes to introduce a fluxing gas
into the melt are still in favor. Chlorine has long been considered the most
efficient fluxing gas in this application. However, because of its toxicity
as well as the large amount of HC1 and Al^Og emissions it creates, alternate
gases have had to be considered, initially diluting chlorine with nitrogen,
and later developing the use of N2-CO-C12 (7) gas mixtures. Although these
gas mixtures reduce stack emissions, they are not as effective as Cl2 in
removing hydrogen and non-metallicparticulates from aluminum and aluminum
alloy melts.
These considerations, coupled with indications obtained elsewhere (8) of
the degassing effectiveness of Freonl^have led to the work reported here, -
which culminated in the development of an industrial, low pollution, non-toxic
inert gas-Freon 12 degassing process (9-10) for aluminum alloy melts.
275
-------
Description of the Inert Gas-Freon 12 Melt Treatment Process
The inert gas-Freon 12 treatment consists of fluxing aluminum alloy melts
with a small amount (3 to 5%) of Freon 12 (CCl2F2) diluted in an inert gas such
as nitrogen or argon. A thin layer of molten halide salts designed to enhance
degassing efficiency is used when nitrogen is employed as the carrier gas. A
NaCl-KCl salt with a small amount of fluoride salt is preferred because of
cost considerations for aluminum alloys containing less than 1% Mg. A NaCl-
free salt flux consisting primarily of MgCl2 and KC1 is preferred for alloys
containing more than 1% Mg to enhance sodium removal.
In practice, the inert gas-Freon process is used for conventional batch
fluxing, in either melting or holding furnaces; a simple gas mixing device is
all the equipment that is required, in addition to the piping and degassing
tubes conventionally available for gas fluxing.
Briefly stated, this process is technically equivalent to pure chlorine
fluxing in removing hydrogen, non-metallic particulates and sodium from
aluminum alloys. At the same time, it reduces stack emissions by a substan-
tial factor over chlorine, allowing the cast shop operator to meet existing
air pollution standards. Freon 12 is non-toxic so that even its accidental
release in the plant will be of no consequence to the workers from an occupa-
tional safety standpoint. In the melt, Freon 12 thermally decomposes com-
pletely into carbon, fluorine and chlorine; therefore, no Freon 12 is released
into the atmosphere. Chlorine and fluorine subsequently react with aluminum
and other elements such as magnesium to form various compounds.
and/or MgF£ are solids at molten aluminum temperature and remain on
the surface of the melt . The MgCl2 formed is a liquid and also remains on
the surface of the melt. A1C13, which has a low sublimation temperature is
in the gaseous phase, and reacts with moisture in the furnace atmosphere or
stack to form hydrated A^Og and HC1. Also, a very small amount of AlFg and
hydrolyze to form minute amounts of HF.
Basic Characteristics of Freon 12
The following presents some of the basic characteristics of Freon 12 as
they relate to its environmental suitability as a molten aluminum treatment
gas. Dichlorodifluoromethane (CC12F2) is a fluorocarbon widely used as a
refrigerant. It is available commercially under various trade names, one of
the most common being Freon 12 , a trademark of the original manufacturer, E.
I. duPont de Nemours. Freon 12 has very low toxicity, rated in Group 6 by the
Underwriter's Laboratories Report MH-2375,(11) with a threshold limit value
of 1000 ppm (12) which is the allowable exposure to time weighted concentra-
tions for a 7 or 8 hour work day and 40 hour work week. By comparison, the
threshold limit value for chlorine is 1.0 ppm (13). Accidental release of
Freon 12 in the work environment would therefore be of no consequence to the
health of the workers.
276
-------
However, since concern has been expressed over the potential harm that
fluorocarbons may cause to the ozone layer, care must still be taken to avoid
accidental release of Freon 12. On the other hand, because Freon 12 completely
breaks down at molten aluminum temperature, typically above 650C,'(1200F) there
is no risk of introducing unreacted Freon 12 into the atmosphere as a result
of the fluxing process. This is supported by data (12) that indicate that
Freon 12 in quartz starts to decompose at 540C (1000F). Since it is also
known (14) that aluminum lowers the stability of Freon 12 , it is very unlikely
that any Freon 12 can withstand the temperatures encountered in aluminum alloy
melts.
The possible existence of C, C±2 and N2 in presence with moisture (HoO)
in the air also raises the concern of forming cyanide CN and/or
phosgene COC12, both highly toxic compounds. However, measurements pre-
sented later will show conclusively that this is not the case.
Emission Sampling Procedure
Throughout the experimental development and initial plant implementation
of the nitrogen-Freon 12 melt treatment process, continuous sampling was per-
formed for particulate and soluble halide gases. Sampling took place in the
stack, close to the roof line of the cast house, with a gas train* and pro-
cedures (15) recommended by the Federal Environmental Protection Agency. The
sampling was isokinetic in that the average velocity of the sample stream was
the same as the average gas velocity within the stack. Depending on the stack
cross-section, sampling was conducted at either one central location in the
smaller stacks or 4 to 8 locations in the larger ones.
The solid particulates were weighted and fired at 1000C (1830F) for 1 hour.
The aluminum was analyzed and reported as A1203, the remainder consisting of
oxides which were not analyzed qualitatively and were reported with the
ignition losses. The caustic and water solutions used to trap soluble halide
gases were also analyzed for Al+3, Cl~ and F~. Since negligible quantities
of Alf^ were found, results were reported on the basis of emission rates of
HC1 and HF. Carbon content was analyzed when Freon 12 was used. In this
case, a Dragon technique was also used to measure phosgene and carbon monoxide
levels; cyanide was checked for with a colorimetric technique using a cyanide
test outfit.**
Experimental Development of Inert Gas-Freon Melt Treatment
Initial experimental work was carried out on 450 Ibs of induction melted
pure aluminum (99.85% Al) at a total gas flow rate of 15 1/min (0.53 scfm).
* RAG Model 2343 Train Staksampler, Research Appliance Co., Allison
Park, Pennsylvania.
** Model SCF, LaMotte Chemical Products Company, Chestertown, Maryland
277
-------
This brought out two major conclusions:
1. Nitrogen-10% Freon 12 and argon-10% Freon 12 gas mixtures were
more efficient than either N2-10% CO-10% C12 or N2-10% C12 in
terms of degassing rate.
2. Nitrogen-10% Freon 12 and argon-10% Freon 12 were more
efficient than pure chlorine.
3. A molten salt cover flux was effective in suppressing fume
emission and increasing degassing rates.
Subsequent pilot plant tests with a gas fired furnace and 4500 Ibs melts,
again of pure aluminum (99.85% Al) showed that Freon 12 levels in the 5-10%
range, lower than previously thought, degassed as effectively as pure chlorine.
The total flow rate was then 90 1/min (3.2 scfm) and no flux cover was used.
The emission rates shown in Table 1 and illustrated in Figure 1 indicate a
significant reduction in emission rate, for an equivalent degassing efficiency,
as a result of using N2-10% Freon 12.
Table 1 - Emission rates during batch fluxing of 4500 Ibs melt of 99.85% Al.
Comparison between chlorine and nitrogen-10% Freon 12 .
Fjnission Rate (Ib/hr)
Fluxing Gas
ci2
N2-10% Freon 12
* Flow rate required to degas to 0.07 cc H2/100g Al in 30 min.
** Ignition losses and unanalyzed oxides.
Plant Tests of Inert Gas-Freon Melt Treatment
Routine production tests were carried out on 45,000 Ib melts, in a gas
fired open hearth furnace equipped with a flux bay with 10 graphite tubes
(1/2 in.diameter I.D.) lowered through the roof, allowing for continuous
fluxing during casting.
The test program consisted in a comparison between fluxing with pure
chlorine and nitrogen-5% Freon 12 . The alloys included 1145, a Mg-free alloy
and 3004, a Mg-containing alloy, and the emission rates are shown in Table 2.
Flow*
Rate
scfm
3.2
3.2
21.
3.
1
6
2
AlCli
0.4
0.1
A1F3
0.01
0.03
Solids
**
Others
8.8
0.9
Gases
Total
30.8
4.2
HC1
42.3
4.7
HF
0.01
0.03
278
-------
w
o
tH
CO
en
80
70
60
50
40
30
20
10
HC1
Other
Particulate
A1203
N2-10% FREON 12 CHLORINE
FLUXING GAS
FIGURE 1 - Emission rates during batch fluxing of 4500 Ibs melt
of 99.85% aluminum. Comparison between chlorine and
nitrogen-10% Freon 12 at 90 1/min (3.2 scfm) flow
rate.
279
-------
Table 2 - Emission rates during continuous fluxing of <45,000 Ibs melts of
alloys 1145 (Mg-free) and 3004 (Mg-bearing). Comparison between
chlorine and nitrogen-5% Freon 12 .
3004
Fluxing
Gas
C12
Nitrogen-
5% Freon 12
C12
Nitrogen-
5% Freon 12
5.5
3.2
Fjnission rate (Ib/hr)
Flow*
Rate
(scfm)
4.5
3.2
Solids
Cast
No.
1
2
1
2
Al
16
16
2
2
22l
.50
.60
.26
.00
A1C1-3
0.05
0.58
0.01
0.01
A1F3
0.00
0.00
0.02
0.01
**
Others
3.
0.
1.
0.
45
92
31
48
Total
20
18
3
2
.00
.10
.60
.50
Gases
HC1
42.30
41.90
6.00
5.40
HF
0.10
0.04
0.26
0.28
1
2
1.37 0.14 0.00 0.31
1.82 54.50 0.07
0.07
0.02
0.02 0.01
0.01 0.01
0.44
0.20
0.54
0.24
3.20 0.14
2.00 0.40
*Flow rates were.adjusted to give equivalent degassing for both fluxing
gases, down to 0.06-0.09 ccH2/100g Al.
**Ignition losses and unanalyzed oxides.
At the time of the test, the proposed state regulation called for a maxi-
mum solid particulate emission rate of 20 Ib/hr for the entire plant. As
shown in Table 2, this cannot be met with C12 for Mg-free alloys if more than
one casting unit is operating at one time.
The Mg-containing alloys could be fluxed with C12 even if all 8 furnaces
were operating at once, on the basis of solid particulate emission. However,
the HC1 emission rate for both Mg-free and Mg-bearing alloys are so high with
Cl? as to rule out its use.
Conversely, the nitrogen-5% Freon 12 clearly appears capable of meeting
regulations as well as efficiency requirements for hydrogen removal.
Additionally, the stack gas sampling indicated that the phosgene (COC12)
content was below 0.1 ppm and the CO content was below 10 ppm, detection
limits of the analytical method respectively, and OSHA standards as well.
Because nitrogen-Freon 12 mixtures release N, C, F and Cl, there was a
theoretical possibility that C and N might react to form cyanide (CN) which
would be likely to be in the dross since this is where the carbon is present.
This was not expected to be a real problem since the earlier 4500 Ib melt
fluxing had resulted in less than 1% carbon in the particulate collected
during fluxing with a nitrogen-30% Freon 12 mixture. This was confirmed by
the actual level of CN found in the dross which was only in the 0.5-4.5 ppm
range, the range in the reagent blank considered free of cyanide.
280
-------
Visible Emission;
An idea of the limitation placed on visible emission can be obtained from
the following quote from the Visible Emission regulations of the State of
Tennessee: "No discharge of a visible emission ... with a density greater
than number one (1) on the Ringelmann chart or an opacity in excess of
twenty (20) percent for an aggregate of more than five (5) minutes in any one
(1) hour or more than twenty (20) minutes in any twenty-four (24) hour period."
Visible emissions tending to produce a black plume are evaluated in terms
of the Ringelmann scale, while visible emissions tending to produce a non-
black plume are evaluated in terms of equivalent opacity and expressed as per-
cent opacity. The major emission problems are encountered with Mg-free alloys;
therefore, the potential of inert gas-freon treatment in reducing emission
levels was initially tested with such alloys.
Table 3 - Comparison between visible emissions during batch fluxing of
25,000 Ibs melts of Mg-free aluminum alloys.
Flow
Rate Visible emission**
Fluxing Gas (scfm) Flux* (Ringelmann No.)
C12 6 None 5
4 None 4 to 5
C12 4 A 5
N2-5% Freonl2 6 A 1/4 - 1/2
N2-3% Freonl2 6 B 1/4
4043 C12 6 None 5
C12 4 None 4 to 5
N2-3% Freonl2 6 B 1/4
*Flux A (demagging flux): 50% AlFo, 24% NaCl, 24% KC1, 2% K3A1F6
Flux B: 48% NaCl, 48% KC1, 4% K3A1F6
**Maximum allowed is Ringelmann No. 1
As summarized in Table 3, it is apparent that fluxing Mg-free alloys with
nitrogen-3 to 5% Freon 12 results in visible emissions well below the required
Ringelmann No. 1, whereas Ringelmann numbers in the 4-5 range are observed
when using pure chlorine. Figure 2 dramatically illustrates this point.
Another type of emission problem results during dross treatment of alloys
with a salt flux. For example, dross treatment of a Mg-free^ alloy (3003)
containing 25% painted scrap and using a commercial flux (55% NaCl, 15% KC1,
15% Na9SiFfi, 7% Na2S04, 7% NaN03) created visible emissions above the
specified 20% opacity level during the first 5 minutes of the treatment, still
281
-------
a)
FIGURE Typical visible emission during fluxing of a Mg-free aluminum
alloy melt.
a) Chlorine fluxing - Ringelmann 5
b) Nitrogen-5% Freon 12 - Ringelmann 1/4-1/2
282
-------
staying above 10% for the next 15 minutes. When switching to N7-3% Freon 12
at 4 scfm, visible emissions never exceeded 5% opacity. To get equally low
metal losses from the dross, it was however found necessary to increase the
Freon 12 content from 3 to 10%. Even though no measurement was made to con-
firm this point, it was expected that the resulting opacity of the plume would
not exceed 15%.
As indicated earlier, chlorine fluxing of Mg-bearing alloys poses no real
visible emission problems. However, some tests were run to confirm the
expectation that inert gas-Freon 12 melt treatment would do as well, if not
better, than chlorine. For example, chlorine fluxing at 8 scfm for a clean
melt of Mg-bearing alloy (5056 alloy) was found acceptable in terms of visible
emission, with a Ringelmann number of 1/2. However, nitrogen-3% Freon 12 with
a 6 scfm flow rate resulted in still slightly lower Ringelmann numbers,
around 1/4. In the case of 6063 alloy, stack emissions were also about 10%
opacity or 1/2 on the Ringelmann scale, when fluxing with N2-3% Freon 12 at a
total flow rate of 6 scfm with a 48% NaCl, 48% KC1 and 4% K3A1F6 flux.
Technical Performance of Inert Gas-Freon Process
The success of the inert gas-Freon process in meeting environmental re-
quirements has been amply demonstrated above. It is important to emphasize
that this has been achieved at no sacrifice in the quality of the aluminum
melt. A few examples are reported below.
Hydrogen removal:
A low hydrogen level in the melt is required to give ingots free of
porosity. A series of degassing tests were run under essentially identical
conditions to compare chlorine and N2-10% CO-10% C12 to nitrogen-5% Freon 12 .
Hydrogen levels were measured in the transfer trough between melting and
holding furnace, using the initial bubble method (FMA hydrogen tests).
Table 4 - Typical hydrogen levels in plant fluxing tests -
45,000 Ibs melts.
Hydrogen level (cc/lOOg Al)
Alloy
Fluxing Gas 1100/1145 3004
N2-5% Freon 12 0.08 0.07
C12 0.08 0.09
N2-10% CO-10% C12 0.11
The typical results in Table 4 illustrate for several alloys (1100, 1145,
3004, 5052) the fact that nitrogen-5% Freon 12 is as effective as pure chlorine,
and somewhat more effective than N2-10% CO-10% C12 in terms of hydrogen
removal.
283
-------
A continuous production evaluation over a 3 month period demonstrated the
consistency of these results as'shown in Table 5,
Table 5 - Distribution of hydrogen levels during 3 month plant production.
45,000 Ibs melts. 1XXX, 3XXX and 5XXX series alloys.
Number of casts per hydrogen level
Fluxing gas Total <0.10cc/100gAl 0.11-0.15cc/100gAl >0.15cc/100gAl
N2-3% Freonl2 50 34 (68%) 15 (30%) 1 (2%)
C12 83 59 (71%) 22 (27%) 2 (2%)
Melt cleanliness;
Melt cleanliness, that is the amount of non-metallic particulate in the
melt, can be assessed directly by means of a pressure filter test. Approxi-
mately 20 Ibs of melt is ladled into a preheated crucible and then forced
under air pressure through a fine porous filter disc which catches all the
non-metallics. After sectioning and metallographic examination of the disc
and the adjacent residual metals, melt cleanliness is rated by the thickness
of the band of non-metallics.
Table 6 - Melt cleanliness as a function of fluxing gas and alloys, as
measured by pressure filter test of continuously fluxed metal -
45,000 Ibs melts.
Average
Non-metallic
Alloy Fluxing Gas Band Thickness (mm)
1100, 1145 N2-5% Freonl2 0.13
(Mg-free) CI2 0.35
N2-10% CO-10% C12 0.13
3004, 5052 N2-5% Freon 12 0.11
(Mg-bearing) C12 0.11
N2-10% CO-10% C12 0.50
The test results shown in Table 6 demonstrate that whether for Mg-free
or Mg-bearing alloys, N2~5% Freon 12 is equivalent to the best gas mixture,
which is N2-10% CO-10% C12 for Mg-free alloys and pure C12 for Mg-bearing
alloys.
284
-------
Later production control tests for 5252 alloy over a 3 month period
related the cleanliness of the final sheet material to the nature of the
fluxing gas mixture. In this case, the number of linear defects in a
lOin x lOin sample was measured and reported in Table 7.
Table 7 - Metal cleanliness as a function of fluxing gas - 3 month plant
production - 5252 alloy
Average
Number of Total Number Defect Per
Fluxing Gas Samples of Defects* Sample
N2-3% Freonl2 73 129 1.8
Cl2 51 163 3.2
* Type II linear defects in lOin x lOin sample.
The use of N£-3% Freon 12 resulted in an average of 1.8 defect/sample,
whereas with Cl2, this number was 3.2, or 80% higher.
Denatrification;
Sodium levels below about 5ppm are desirable in order to improve cast-
ability and hot workability of high Mg, 5XXX series alloys. In tests with
30,000 Ibs melts of 5056 alloy, it was demonstrated that ^-3% Freon 12 is as
effective as Cl2 in removing Na provided a Na-free cover flux (55% KC1,
42% MgCl2 and 3% CaF2> is used. Detailed results are presented in Table 8
for one cast with C12 and two casts with N2~3% Freon 12 .
Table 8 - Denatrification of 30,000 Ib melt of 5056 alloy.
Fluxing Gas
N2-3% Freon 12
C12 1 0.0014 <0.0007* 160 70
* Below detection limit.
Cast
No.
1
2
Na level (%)
Initial Final
0.0010 <0.0007*
0.0020 <0.0007*
Fluxing Sas
Usage (scf)
120
120
Time
(min)
65
150
285
-------
Melt losses;
For purely economic reasons and to minimize the amount of material which
must eventually be disposed of after reclamation of the metal content, it is
important that melt losses through dioss generation be minimized.
Based on plant scale tests, it was determined that for Mg-free alloys,
melt losses with N2~3% Freon 12 were equivalent to N2~10% CO-10% Cl£ and
slightly higher than Cl2- For Mg-bearing alloys, however, N£-3% Freon 12
resulted in the lowest melt losses, Cl2 and N2~10% CO-10% C12 giving
respectively higher losses.
Therefore, depending on the exact product mix between Mg-free and Mg-
bearing alloys, the net melt losses with N2~3% Freon 12 are expected to be
equivalent to C12 and better than N2~10% CO-10% C12-
Conclusion
The inert gas-Freon 12 melt treatment process has been demonstrated, up
to large scale industrial production condition, to be technically equivalent
to pure chlorine treatment in all important aspects including hydrogen removal,
non-metallic inclusions control and denatrification. At the same time, it
solves all the environmental and occupational safety problems caused by
chlorine. The major favorable aspects are the non-toxicity of the gas mix-
ture itself and the substantial reduction, well below existing standards, of
both air pollution and visible emissions levels.
This process is now used exclusively for melt treatment in the
Consolidated Aluminum Corporation rolling mill at Hannibal, Ohio.
ACKNOWLEDGEMENTS
The authors wish to acknowledge the extensive contributions made by
P. E. Sevier and J. C. Yarwood of Olin Metals Research Laboratories to the
development and plant implementation of the inert gas-Freon 12 melt treatment
process. The cooperation of the personnel of the various Consolidated
Aluminum Corporation cast shops during plant trials was also invaluable.
286
-------
REFERENCES
1. L. C. Blayden, K. J. Brondyke and R. E. Spear, "Process for Treating
Molten Aluminum," U.S. Patent No. 3,737,304, June 5, 1973. Assigned to
Aluminum Company of America.
2. M. J. Bruno, N. Jarrett and B. L. Slaugenhaupt, "Treatment of Molten
Aluminum with an Impeller," U.S. Patent No. 3,849,119, November 19, 1974
Assigned to Aluminum Company of America.
3. A. G. Szekely, "Apparatus for Refining Molten Aluminum," U.S. Patent
3,743,263, July 8, 1973, Assigned to Union Carbide Corporation.
4. A. G. Szekely, "An Alternative to Chlorine Fluxing of Aluminum: The
SNIP Process," Proceedings of the 2nd International Aluminum Extrusion
Technology Seminar, Nov. 15-17, 1977, v. I, pp 35-41.
5. W. C. Rotsell and C. E. Cawthorn, "An Alternative to Chlorine Fluxing of
Aluminum: The SNIP Process - Operating Experience and Results,"
Proceedings of the 2nd International Aluminum Extrusion Technology
Seminar, Nov. 15-17, 1977, v. I, pp 25-34.
6. M. V. Brant, D. C. Bone, and E. F. Emley, "Fumeless In-Line Degassing and
Cleaning of Liquid Aluminum," Journal of Metals, March, 1971, Vol. 23
No. 3, pp. 48-53.
7. H. 0. Titze, "Tri-Gas Fluxing, Metallurgical, Technical and Economic
Aspects," Light Metals, AIME, v. 2, 1973, pp 451-457.
8. H. Yamada, T. Kitomura and 0. Iwao, "Degassing Media for Molten Aluminum:
Degassing Capacity of Freon 12 (CC12F2) and a Nitrogen-Freon Mixture,
AFS Cast Metals Research Journal, v. 6, No. 1, March 1970, pp 11-14.
9. J. E. Dore, P. E. Sevier and J. C. Yarwood, "Purification of Molten
Aluminum and Alloys," U.S. Patent 3,854,934, Sept. 17, 1975. Assigned
to Swiss Aluminum Ltd.
10. J. E. Dore, J. C. Yarwood and J. A. Ford, "A New Low Emission Process for
Degassing and Treating Aluminum Alloy Melts",Light Metals, AIME, v. 2,
1976, p. 567-583.
11. Thermodynamic Properties of Freon 12 , Bulletin T-12, E. I. duPont de
Nemours, .Wilmington, Delaware, August 1975.
12. Freon Product Information Bulletin B-2, E. I. duPont de Nemours,
Wilmington, Delaware, June 1976.
13. Handbook of Industrial Toxicology, Edited by E. R. Plunkett
14. H. M. Barmelee, "Sealed-tube Stability Tests on Refrigeration Materials,"
Bulletin RT-42, E. I. duPont de Nemours, Wilmington, Delaware, April 1976.
287
-------
15. W. S. Smith, et al, "Stack Gas Sampling Improved and Simplified with
New Equipment," Air Pollution Control Association, Paper No. 67-119,
Annual Meeting, June 11-16, 1967, Cleveland, Ohio
288
-------
COPPER INDUSTRY
Session Chairman: Ivor E. Campbell
Clyde Williarns and company
Columbus, Ohio
289
-------
INCO'S OXYGEN FLASH SMELTING PROCESS
FOR COPPER AND NICKEL CONCENTRATES
OFF-GAS HANDLING AND IMPURITY DISTRIBUTIONS
M.Y. SOLAR1, A.D. CHURCH2 AND T.N. ANTONIONI3
Inco Metals Company
Toronto, Ontario, Canada
ABSTRACT
The Copper Cliff Smelter of the Inco Metals Company has been
autogenously smelting concentrates in an oxygen flash furnace of
Inco design since 1952. Over the past twenty-seven years, some
seven million tonnes of copper concentrates have been treated using
this technique. During a recent demonstration run, some 11,000
tonnes of nickel concentrates were also smelted using the same
process. The nominal capacity of the present furnace is 1400 tonnes
of concentrate per day, but throughputs as low as 600 - and as high
as 1650 tonnes per day have been achieved while remaining in an
autogenous mode of operation.
Two main gas streams are produced in this process. The two
top-discharged fluid bed reactors, in which the furnace feed is
dried, average 650 m3/min (24,300 scfm) of a gas-solid mixture at
105°C. The dried feed is separated in product baghouses which
operate at collection efficiencies of 99.9%.
The off-gases from the furnace proper amount to only 115 m3/
min (4300 scfm) at the current throughput of ^1100 tonnes per day.
They contain less than 3% of the feed as dust and between 25 to
70% of the bismuth, 20 to 50% of the lead and arsenic, and 5 to 15%
of the selenium in the concentrates smelted, depending on type of
concentrate and oxygen potential. About 85% of these gases are
scrubbed and sent,at 70 to 80% SO2 by volume, to a liquefaction
plant averaging ^270 tonnes of liquid S02 per operating day. The
dust and impurity collection efficiency of the furnace gas handling
system is 99.9+%.
The impact that incorporation of the Inco oxygen flash smelting
process would have on dust, impurity and S02 containment in an
integrated smelter is also discussed.
1. Area Process Engineer, Canada, Process Development
2. Superintendent, Process Technology, Copper Cliff Smelter
3. Superintendent, Furnace Department, Copper Cliff Smelter
290
-------
THE COPPER CLIFF SMELTER
In a typical year, the Inco Metals Company mines, and processes,
at its Copper Cliff complex, some 15 million tonnes of ore from
which the main products are some 160,000 tonnes of nickel, 140,000
tonnes of copper, 600,000 tonnes of iron ore, 80,000 tonnes of
liquid sulfur dioxide and 800,000 tonnes of sulfuric acid.
Significant amounts of cobalt, precious metals, selenium and
tellurium are also recovered.
At the heart of the pyrometallurgical activities are the
nickel and copper circuits of the Copper Cliff Smelter through
which some 1.6 million tonnes of "nickel" concentrates and 0.4
million tonnes of chalcopyrite concentrate are treated each year.
Copper reporting in the "nickel" concentrates (about 25% of copper
input) and nickel reporting in copper concentrate (only 3% of nickel
input) cause substantial intercircuit metal transfers, as the current
smelter flowsheet shown in Figure 1 illustrates. The reasons for
this interdependence merit further description, particularly since
it influences directly the mode of operation of the oxygen flash
furnace.
The "nickel" concentrates, averaging 10% Ni and 2% Cu, are
partially calcined in multi-hearth roasters, smelted in reverber-
atory furnaces and blown in Peirce-Smith converters to produce a
sulfur deficient nickel-copper-cobalt sulfide phase called Bessemer
matte. This Bessemer matte is cast into 22 tonne molds, cooled at
a controlled rate and then separated in the Matte Processing Plant
to produce heazlewoodite, metallic and chalcocite concentrates.
The heazlewoodite concentrate is treated in fluid bed roasters to
produce various grades of nickel oxide products while the metallic
fraction becomes the feed to the Copper Cliff Nickel Refinery.
The chalcocite concentrate, contaminated with ^5-6% Ni as heazle-
woodite (Ni3S2) is blown to blister in the copper circuit converters.
The chalcopyrite concentrate, analyzing about 30% Cu and 1%
Ni, is dried in fluid bed reactors, smelted in the oxygen flash
furnace and blown to blister copper in Peirce-Smith converters.
The blister copper composition is controlled at ^98% Cu and <0.7%
Ni (the rest being dissolved oxygen) and most of the nickel entering
the copper circuit is thus concentrated in the slag raised during
the copper converters' slag blows and the dry mush raised during
the finishing blows. Consequently, the only way of controlling the
nickel levels in the copper circuit is to return the copper converter
slag and mush to the nickel circuit.
The fact that no copper converter slag is returned to the
copper flash furnace is thus a response to the nickel bleed
requirements and not an intrinsic short-coming of the furnace
itself. As a matter of fact, recent tests (Dec. 1977-Jan/Feb.
1978) have indicated that recycle of copper converter slag to
291
-------
FIGURE I COPPER CLIFF SMELTER FLOWSHEET
Scrap
Nickel concentrate
Sand
Miscellaneous
Roasters
Calcine
ESP dust
lQjNickel
converters
Copper concentrate
Sand
Chalcocite concentrate
from Matte Separation
P""""1 1
T~
Mag to
i
lump
^ ^
\
.n—
' ^
P
^1
Nickel
reverb
furnaces
Matte
r' Misc
r— Flux
^^fmm & •• _ • A
s
Slag
Cooling
converters
Matte
casting
Bessemer matte
to Matte Separation
Fluid bed
****
S02 to
liquefaction
A
Slag to
dump
Reverts
Slag
Oxygen
Copper
flash
furnace
Matte
— Miscellaneous
— Flux
— Chalcocite concentrate
from Matte Separation
Copper
converters
Blister copper
to Copper Refinery
292
-------
the flash furnace is possible with very little disruption of normal
operation. A permanent slag return launder is consequently being
installed on the furnace.
OXYGEN FLASH SMELTING PROCESS
The oxygen flash smelting process was developed in the 1940's
by the then International Nickel Company of Canada to treat either
copper or nickel concentrates, as the earliest references to this
work indicate1i2. The first commercial unit was built in 1952 to
handle 450 tonnes per day (tpd) of concentrates. The current
furnace dates from 1968 and has a nominal capacity of ^1400 tpd.
Over the past twenty-seven years, some seven million tonnes
of copper concentrates have been treated by the oxygen flash
smelting process. The process was never commercially used on
nickel concentrates for two main reasons: a) the Copper Cliff
Smelter was extensively enlarged and modernized in the ten years
prior to 194O3 and b) further improvements to the existing nickel
circuit1*'5, particularly extensive use of tonnage oxygen6 , con-
siderably improved its efficiency. In 1976 however, a nickel
concentrate campaign was carried out in the commercial unit, during
which some 11,000 tonnes of various nickel concentrates were smelted
over ten days of operation. The process was found to handle nickel
feed with the same ease as it handles copper concentrate7.
The oxygen flash smelting process, its evolution and current
practice have been extensively discussed in the literature (8~ltf) .
Figure 2 illustrates in some detail the arrangement of the furnace
and its ancillary equipment. Typical operating conditions are
quoted in Tables 1 and 2, the reader being referred to the litera-
ture for description of the parts of the process beyond the scope
of the present paper.
Fluid Bed Dryers and Baghouses
Prior to 1971, the feed to the furnace was dried in small
rotary kilns, 1.2 m in diameter and 5.5 m long, that were oriqi-
nally fired with pulverized coal8 and later with natural gas10.
Five such kilns were used, four on concentrates and one on sand
flux. The kilns were "air-swept" to pick-up, transport and elevate
the dried products to the furnace feed bins^. The dried concentrate
and sand were separated from the gas stream in a series of dry
cyclones.
The total drying capacity of the concentrate kilns was only
950 tpd, well below the capacity of the flash furnace itself.
Moreover fugitive emissions from the kilns and associated multi-
clones were difficult to control and periodically caused unaccept-
able conditions in the workplace.
293
-------
FIGURE 2 FLASH FURNACE GENERAL ARRANGEMENT
N)
Fluid bed dryer
ri n n n n r
11111111111
11111111111
11111111111
| W W i> 14 \/
•Exhaust
• Exhaust
Baghouse
Oxygen
I
Baghouse
From second FBD
Dry feed
storage
Oxygen
Slag pot
\ / »Dump
Copper converters < r—7 Matte
-------
Table 1. INCO'S OXYGEN FLASH FURNACE
1977 OPERATING DATA FOR SMELTING OF COPPER CONCENTRATE
Dry Composition, Wt. %
Inputs Tonnes/Day Cu Ni+Co Fe_ Si02 S_
Concentrate 1,101 30.1 1.2 31.4 1.7 33.5
Cu Converter Dust 28 41.8 3.4 8.0 2.5 25.0
Flux 96 - - 2.2 77.8
02 219 _____
TOTAL 1,444
Outputs
Matte 794 42.1 1.7 29.2 - 24.6
Slag 289 0.6 0.1 40.9 32.4 1.4
Scrubber Sludge* 17 42.0 1.8 20.0 2.8 30.6
S02 343 - - - - 50.0
TOTAL 1,443
* Before neutralization
295
-------
Table 2. INCO'S OXYGEN FLASH FURNACE
1976 TEST OPERATING DATA FOR SMELTING OF NICKEL CONCENTRATE
Dry Composition, Wt. %
Inputs Tonnes/pay Cu Ni_ Co Fe_ SiQ2 S
Concentrate + Flux 1,232 2.6 8.0 0.29 34.1 20.2 25.0
Recycles* 36 24.3 5.8 0.25 20.0 10.0 15.0
02 294 ----- -
Residual Cu Matte** J^ 45.1 3.2 0.10 25.3 - 24.5
TOTAL 1,571
Outputs
Matte 435 9.7 23.2 0.58 35.2 - 27.1
Slag 659 0.18 0.34 0.14 36.6 38.0 1.0
Scrubber Sludge*** 37 2.9 4.4 0.24 26.8 2.4 19.5
SO2 405 - - 50.0
TOTAL 1,536
* Contain dust from copper converter operations.
** Cu matte washed out of furnace during nickel campaign.
*** Before neutralization, including sulfates in solution.
296
-------
Consequently these driers were replaced in 1971 by two fluid-
ized bed drying lines each capable of handling 900 tpd of a pro-
portioned concentrate-sand mixture. Each drying line (Figure 2)
consists of a 210,000 kJ/min combustion chamber fired with natural
gas or light industrial oil, a 335 kW blower capable of delivering
670 m3/min of dilution air at 19 kilopascals relative pressure
(kPa), a refractory lined fluid bed vessel and a collection bag-
house .
The wet concentrate-flux mixture (at ^8-10% H20) is introduced
to the dryer through a rotary sealing feeder of Inco design. The
solid particles are suspended in the upward stream of hot combustion
gas (the gas bed velocity is about 3.5 m/s at 10-14 kPa) and are
thus dried to 0.1-0.3% free H20- The moist gases and dry solids
are drawn out through the dryer roof for separation in the product
baghouse. Normal operating conditions are:
- a wind box temperature of 315°C
- a bed temperature of 120°C
and - an off-gas volume of 650 m3/min at ^10% moisture and 105°C.
The product baghouses, are 7.6 m high (including hoppers) by
9.8 m long and 3.7 m wide. Each contains 672 bags, 11.5 cm in
diameter by 2.4 m in length, which are suspended vertically and
supported internally by wire-frame retainers. A tube sheet supports
the bags which are clamped onto Venturis. Each baghouse has two
sides, each side containing five banks. Purging by reverse pulse
air is sequential and accomplished automatically. Various filtering
media were tested, the current bags being Nomex felt fabric. The
exhaust fans are driven by 75 kW, 650 rpm motors and are rated at
•WOO m3/min. Normal operating conditions are:
- an inlet loading of ^900 g/m3
- an inlet temperature of 105°C
- an outlet loading of ^0.7 g/m3 (0.2 g/m3 with new bags)
- an outlet temperature of '^80 to 100°C
and - and outlet pressure of -2 to -2.7 kPa (baghouse differential
varies between 0.5 and 2.0 kPa, i.e. 2" and 9" H20).
The product baghouses thus clean some 650 m3/min of gas at
better than 99.9% collection efficiency.
The baghouse off-gases are piped to the copper converter flue
and are further treated in a dry electrostatic precipitator (ESP)
that handles approximately 27,000 m3/min of gas at 150°C. The
metal collection efficiency of this ESP is currently ^92%, the dust
recovered being recycled to the flash furnace. The fluid bed dryer
off-gases are therefore de-dusted with an overall efficiency of
>99.99%, the dust loadings being brought down to <0.06 g/m3 (<0.026
grain/scf) before the gases are ducted to the 380 metre chimney.
297
-------
The overall availability of each drying line is better than
98%. Because of limiting factors beyond the furnace itself,
operating time is only 60 to 80%.
Flash Furnace and Off-Gas Handling System
The oxygen flash furnace is fired from both ends by four
burners, two at each end. The solids are withdrawn from individual
burner feed bins at a controlled rate and transferred to the burners
by screw conveyors and gravity, the screw conveyors being vented to
the copper converter flue. Recycled dust and touch-up sand, if
required,- are added to the screw conveyors, also at controlled rates.
At the burners, the feed is suspended in a horizontal flow of
oxygen (>96% purity, 'v-lOS kPa) and injected into the furnace. The
oxygen readily reacts with part of the sulfur and iron in the feed
to form S02 and iron oxides. Silica, contained in the flux and
concentrate, combines with the iron oxides to form a fayalite slag.
Copper, nickel and cobalt and the remaining iron and sulfur collect
in the matte.
The main oxidation reactions are:
Copper Concentrate
2CuFeS2 + 2.575 02 = Cu2S + FeS +0.55 FeO + 0.15 FesO^ + 2 S02
Nickel Concentrate
2FeSlelif + 1.855 02 = FeS + 0.55 FeO + 0.15 Fe3O4 + 1.28 S02
Hence, about 1.55 and 1.38 tonnes, respectively, of S02 are
produced per tonne of 02 used while flash smelting copper or nickel
concentrates. These figures translate to 1.50 and 1.33 tonnes,
respectively, of S02 per tonne of commercial oxygen used at 96%
purity. Oxygen utilization is 100%.
The flame gases analyze 94.5% S02 , 1% N2, 1.7% C02 (from
combustion of flotation reagents and wood debris in the concentrates)
and 2.8% Ar (argon from the tonnage oxygen production process) before
dilution by inleakage to the furnace. In the furnace proper, this
inleakage occurs mostly around the burners and is estimated at about
16 m3/min. The oxygen in the inleakage is known to react with the
feed; the furnace off -gases therefore assay 83% S02 , 13% N2 , 1.5%
C02 , 2.5% Ar and <0.02% 02 when they leave the furnace. They also
carry 2-3% of the feed as dust during copper concentrate smelting
(^3-4% of nickel concentrates which are finer) and significant
fractions of the impurities contained in the feed.
Notes : 10 m^/min at 0"C = 373.2 scfm at 60°F
1 kPa relative pressure =4" H20 or 0.15 psig
298
-------
At a typical production rate of 1,100 tpd of concentrate, the
off-gases have a volume of only 104 m3/min and exit the furnace at
1260°C. These gases are thoroughly cleaned prior to delivery to
an S0? liquefaction plant operated by Canadian Industries Ltd.
(CIL)15.
The furnace gas handling system, shown in Figure 3, was
originally built in 1952-53lg. This system has been renovated
from time to time and somewhat improved, but the current design
is basically the same as the original. The main units are:
- a 9.4 m high (including hoppers) by 4.9 m wide by 18.3 m
long settling and radiation chamber originally constructed
of mild steel, but currently made of 316 stainless. About
one third to one half of the entrained dust is collected
in this chamber (and pneumatically returned to the furnace)
while the gases are cooled to 700°C at the same time.
- a splash tower which is an unpacked chamber 1.2 m in
diameter and 5.5m high. The chamber and its two spray
nozzles are constructed of 316 stainless. The gases rise
in the chamber countercurrently to the water sprays and
are cooled to ^60°C in the process; scrubbing efficiency
is ^93%.
- Three venturi scrubbers in series, mounted on a common
tank. Again construction is all of 316 stainless. The
common tank is divided into three sections by baffles in
such a way that gases cannot bypass any unit, the spray
water accumulating to form a water seal at the bottom of
the tank. Flow in these Venturis is cocurrent, the gases
being cooled to 'v32°C; scrubbing efficiency is ^95%.
and - a wet electrostatic precipitator constructed of antimonial
lead and acid resistant brick. This wet ESP is protected
from any sudden increase in draft, which could collapse
the lead tubes, by a water seal connected to the unit near
its inlet. Gases leave the ESP at 25-35°C. Collecting
efficiency in this ESP is '^97%.
Draft through the gas handling system is provided by one of
two centrifugal fans rated at 120 m3/min. These fans have double
inlets, the impellers being constructed of a cast spider with
welded blades. Again construction is all 316 stainless. The
gases leaving the fan, compressed to 25 kPa and heated to 65°C
in the process, are carried to the CIL liquefaction plant in a
0.5 m 0 pipeline constructed of fiberglass-reinforced-polyester
wrapped polyvinyl-chloride (FRP/PVC). The line is covered with
5 cm of fiberglass insulation to prevent heat loss and condensation
within the pipe.
299
-------
FIGURE 3 FLASH FURNACE GAS CLEANING SYSTEM
Furnace
gases
Lime
Neutralizing
tank
By-pass
exhaust
I
Settling
chamber
Splash
tower
Air
i
Settling [
cone
v /
Stand
pipe
Venturi
scrubbers
Aerator
Stand
pipe
Heat
exchangers
Fresh
water-
Wet
precipitator
SOg to liquefaction •
Neutral pulp to thickener
Acid
300
-------
The scrubbing system operates under draft which varies from
-0.12 kPa (0.5" H20) in the settling chamber to -3 kPa in the wet
ESP. Because of this vacuum, and in spite of all efforts to com-
pletely seal the settling chamber, inleakage in this unit occurs:
an estimated 11 m3/min of air enters the chamber diluting the
furnace off-gases from 83% S02 to about 75% S02 (mixed with 20% N2,
2% Ar, 1% CO? and <1% 02) and increasing their volume from 104 m3/
min to 115 mVmin. No further dilution occurs in the wet scrubbing
system itself.
Scrubbing water flows are shown in Figure 3. The underflow
from the splash tower and the first venturi scrubber, about 1200
liters per minute, is taken to a 316 stainless steel cone thickener
for settling. The settling cone overflow, with make-up fresh water,
is cooled from 55°C to 20°C through a series of shell and tube heat
exchangers, also constructed of 316 stainless. The cooled water is
recycled to the third venturi scrubber.
The underflow from the settling cone, ^60 £/min at 20% solids,
is stripped of S02 by air purging at a rate of 3 m3/min, neutralized
with lime and reverted to the concentrator nickel circuit.
The underflow from the second and third venturi scrubbers,
^1500 £/min, is recycled to the splash tower and the first two
scrubbers.
Overall scrubbing efficiency is ^99.99% with less than 2 kg/
day of dust being collected on the liquefaction plant filters (see
Table 3). For several reasons, one being the need to increase draft
on the furnace when changing burners, another the limited capacity
of the scrubbing system and the liquid S02 plant, only about 85%
of the flash furnace gases are handled through the system described.
The remaining 15% are exhausted from the settling chamber through
a bell damper from which they are vented to the converter flue which
leads to the dry ESP mentioned in the previous section. Since the
efficiency of CIL's liquefaction plant is 95%, liquid S02 production
averages 270 tpd out of the 340 tonnes of S02 produced while the
furnace is in normal copper service.
Impurity Elimination
The Copper Cliff concentrates are, by most standards, rela-
tively free of impurities. The copper concentrates generally
average less than 0.01% Bi + As + Se and less than 0.10% Pb. The
nickel concentrates assay less than 0.01% Bi+Se and less than 0.04%
Pb+As, these four impurities being the only ones that have tradi-
tionally been followed.
For this reason and others, such as smelting and converting
oxygen potentials or amount and point of re-entry of various
301
-------
Table 3. FLASH FURNACE GAS HANDLING SYSTEM
COLLECTION EFFICIENCIES
Collected after
Cumulative, %
U)
o
to
passage through
Settling chamber
Splash tower
Venturi scrubbers
Weight
42
96
99.8
Cu
37
95
99.8
Ni+Co
49
96
99.8
Fe
52
97
99.9
As
38
96
99.8
Zn
46
96
99.9
Se
25
94
99.8
Pb
18
93
99.7
Bi
19
94
99.
7
Wet Cottrell
99.995 99.999 99.985 99.999 99.995 99.998 99.991 99.989 99.976
-------
recycle streams, the impurity distributions obtained in our
circuits, particularly in the oxygen flash furnace, are repre-
sentative only of our own operation.
Distribution of the four impurities mentioned between matte,
slag and scrubber slurry are given in Figures 4 to 7. Data are
quoted for nickel and copper flash smelting without slag return
as well as for conventional nickel roaster+reverb smelting and
copper flash smelting with slag return. The data are plotted
versus the oxygen potential sustained by the slags under the
various operating conditions (calculated from the Fe3+/Fe2+ ratio
in the slags).
As shown, about 20 to 30% of Bi, Pb and As and only 3% of Se
are recovered in the scrubber slurry under normal copper flash
furnace operation (i.e. without slag return). In normal nickel
service, these fractions are greater amounting to 50-70% for Bi,
30-50% for Pb, ^40% for As and 7-15% for Se depending on the
oxygen potential.
As shown in Table 3, these impurities are collected in the
gas handling system in somewhat different manners depending on
their physical state. As and Zn behave in much the same fashion
as Cu, Ni, Co and Fe: 40 to 50% of the total amount of these
elements carried by the gas stream are recovered in the settling
chamber. This behaviour indicates that As and Zn are already in
a solid state when they enter the settling chamber or that, at
the very least, they condense early enough in the settling chamber
to behave as Cu, Ni, Co and Fe (settling chamber temperatures drop
from 1300°C to 700°C). On the other hand, only 20-25% of the Se,
Pb and Bi carried by the gas stream are recovered in the settling
chamber. A fraction of these impurities could thus still be in a
volatile state at the settling chamber outlet temperature.
In any case, about 99.99% of the various elements are scrubbed
out in the remainder of the gas handling system. The neutralized
settling cone slurry is then recycled to the reverberatory furnace
circuit where the impurities are distributed mostly between slag
and matte (see Figures 4 to 7).
EMISSION CONTROL USING OXYGEN FLASH SMELTING PROCESS
The nickel and copper circuit interdependence, and the mode
of operation of the oxygen flash furnace described in the previous
sections are unique to Inco Metals and have grown out of the
difficulties encountered in the treatment of the Sudbury ores.
This, and the previous reluctance of the company to market its
technology, probably account for the fact that only two oxygen
flash smelting furnaces are in operation at present (Inco's and
an identical unit located in Almalyk, USSR) in spite of the fact
303
-------
FIGURE 4 IMPURITY DISTRIBUTION • ARSENIC
100
80
60
40
S.
o
| 20
1
* 10
20
10
8
6
o Ni roasters and reverb
o—o Ni flash
* Cu flash
A Cu flash with partial
slag return
o»
a
«
o
o 4
100
88°
| 60
* 40
20
10
I HIM I I I
1864 2 186
10-7 IQ-8
Slag oxygen potential, atm
I 8
10-9
304
-------
FIGURE S IMPURITY DISTRIBUTION • SELENIUM
100
80
I 60
o
5 40
S.
20
10
8
6
o Ni roasters and reverb
<^> Ni flash
* Cu flash
A Cu flash with partial
slag return
r *
i
* .
20
10
S 8
i, 6
8 2
£
1864 2 186
10-7 10-8
Slag oxygen potential, atm
I 8 6
10-9
305
-------
FIGURE 6 IMPURITY DISTRIBUTION • LEAD
100
80
60
40
o
$ 20
8
* 10
40
20
•§ 10
$ 8
I '
^ 4
60
40
20
10
g 8
i 4
&
8 2
o Ni roasters and reverb
«-^> Ni flash
* Cu flash
A Cu flash with partial
slag return
I
II
1864 2 186
10-7 IQ-8
Slag oxygen potential, atm
I 8 6
10-9
306
-------
FIGURE 7 IMPURITY DISTRIBUTION • BISMUTH
100
80
60
40
«
o
*
20
10
10
8
6
f
in
£ 2
I
100
80
60
40
20
I
f ,0
4
o Ni roasters and reverb —
°-^> Ni flash
» Cu flash ~
A Cu flash with partial —
slag return
i
HIM I I I HIM I I I I I I I
1864 2
10-7
Slag oxygen potential, atm
86
I 8
10-9
307
-------
that the process is generally regarded as one of the least (if
not the least) expensive process in terms of energy consumption,
capital expenditures and operating costs (17~19).
In the field of S02, particulate and impurity emission control,
the oxygen flash smelting process offers further advantages which
are best emphasized by focusing on the impact that its incorpo-
ration would have on a "typical", independent copper smelter.
Since sulfur dioxide and arsenic compounds are some of the
species under closest scrutiny at this time in North America,
sulfur and arsenic distributions were calculated, together with
gas volumes and S02 concentrations, for the two flowsheets shown
in Figures 8 and 9. Both plants were assumed to handle 1,100
tonnes per day of concentrate at 28% Cu, 29% Fe, 32.6% S and "low"
arsenic, typical of many feeds being treated in the U.S. and Canada.
To maintain a constant basis of comparison, the furnace matte grades
were fixed at 42% in both cases.
The conventional copper smelter flowsheet involves:
- four multiple hearth roasters,
- one reverberatory furnace,
- three Peirce-Smith converters, one on stand-by and two
operating at 50% blowing time,
- one 2160 m3/min (81,000 scfm) acid plant handling converter
gases at 4.4% SOa with 95% efficiency. Acid production
would then be ^570 tonnes/day.
By contrast, the oxygen flash smelter consists of:
- one 350 tpd oxygen plant,
- two fluid bed dryers,
- one oxygen flash furnace, without scrubbing system,
- the same converter operation, including slag return to
the furnace,
- one 2275 m3/min (85,000 scfm) acid plant handling the 7.4%
SC>2 mixture obtained by enriching the converter gases with
the flash furnace gases, 100 m3/min at 75% SC>2. Acid
production would then be 1010 tonnes/day. The acid plant
sludge would either be discarded (at a loss of <0.5 percent
of the copper in concentrate) or treated in a separate
plant since it cannot be recycled to the smelter in this
case because of impurity build-up considerations.
Sulfur and arsenic distributions in the various units were
taken as:
308
-------
FIGURE 8 SULFUR AND ARSENIC DISTRIBUTIONS
CONVENTIONAL REVERBERATORY SMELTING WITH S02 FIXATION ON CONVERTER GASES
Throughput' = 1100 tonn«s/Vay at28% Cu, 29% Fe, 32.6% 5> "low"arsenic
Concentrate
and flux
Emitted to
atmosphere
Condensed phases
Gaseous phases
Inleakage
•S distribution
• As distribution
-% S02
• Volume, m3/min
1000 m3/min = 37,324 scfm
309
-------
FIGURE 9 SULFUR AND ARSENIC DISTRIBUTIONS
OXYGEN FLASH SMELTING WITH S02 FIXATION ON SMELTING AND CONVERTING GASES
Throughput =IIOO tonnes/day at 28% Cu, 29% Fe, 32.6% S, "low" arsenic
Concentrate
and flux
Emitted to
atmosphere
Fluid bed
dryers
ESP dust
iiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiijiiiiiiiiiiiiiiniiiiiiiiiiiiiiiiiiiiiii h=
Dump!
J
IIIIIIIIIIIIIIIIIIIIIIIIU:
42
69
75
98
Sludge to
seperate
treatment
or discard
I
Condensed phases
Gaseous phases
Inleokoge
• S distribution
•As distribution
%S02
• Volume, m5/min
lOOOmVmin = 37,324 scfm
310
-------
Distribution, % of input to unit process
S AS
To Gas To Slag To Gas To Slag
Multihearth roasters 27 27 -
Reverberator y furnace 20 3 62 20
Fluid bed dryers 0 - 0 -
Flash furnace 42 3 60 15
Converters 98 2 92 5
The diagrams clearly illustrate the potential impact that
incorporation of the oxygen flash furnace would have on this
"typical" copper operation. The two acid plants are virtually
of the same size (^2,200 m3/min for the conventional smelter,
2,300 m3/min for the flash smelter) . However, because of the
extremely small volume and high strength of the flash furnace
gases blended with the converter gases, the flash smelter would
have emissions of only ^5% of the sulfur and virtually 0% of the
arsenic. The conventional smelter, on the other hand, would emit
of its sulfur input and 79% of its arsenic input.
Reductions of the same magnitude would also be obtained on
dust, lead, bismuth and other volatile impurities, since the low
volume of the flash furnace gases allow their mixing with converter
gases, normally the first stream to be controlled, without increas-
ing this volume significantly. Incorporation of the oxygen flash
furnace process in a conventional copper smelter with SC>2 fixation
on converter gases could thus lead to virtually zero emissions
without increasing scrubbing and acid plant facilities, the only
other expense being disposal of the acid plant sludge.
311
-------
REFERENCES
1. "Flash Smelting of Inco Ores Creates New Industry", Northern
Miner, July 8, 1948.
2. "Autogenous Smelting of Sulfides", J.R. Gordon, G.H.C. Norman,
P.E. Queneau, W.K. Sproule and C.E. Young, U.S. Patent
2,668,107, Filed June 25, 1949, Issued February 2, 1954.
3. "Metallurgical Improvements in the Treatment of Copper-Nickel
Ores", The Staff, International Nickel Company of Canada
(Presented by P. Queneau), CIM Transactions, Vol. 51, pp. 187-
98, 1948.
4. "Smelting: I. Copper Cliff Smelter", Canadian Mining Journal,
pp 431-45, May 1946.
5. "Treatment of Nickel-Copper Matte", K. Sproule, G.A. Harcourt
and L.S. Renzoni, Journal of Metals, Vol. 12, No. 3, pp. 214-
19, March 1960.
6. "Tonnage Oxygen for Nickel and Copper Smelting at Copper Cliff",
R. Saddington, W. Curlook and P. Queneau, Journal of Metals,
Vol. 18, No. 4, pp. 440-52, April 1966.
7. "Smelting Nickel Concentrates in Inco's Oxygen Flash Furnace"
M.Y. Solar, R.J. Neal, T.N. Antonioni and M.C. Bell, Journal
of Metals, Vol. 31, No. 1, pp. , January'1979.
8. "The Oxygen Flash Smelting Process of the International Nickel
Company", The Staff, CIM Transactions, Vol. 58, pp. 158-66,
1955.
9. "Oxygen Flash Smelting Swings Into Commercial Operation", The
Staff, International Nickel Company of Canada, Journal of
Metals, Vol. 7, pp. 742-50, June 1955.
10. "Recent Developments in the Inco Oxygen Flash Smelting Process",
S. Merla, C.E. Young and J.W. Matousek, 101st AIME Annual
Meeting, San Francisco, 1972.
11.. "Copper Smelting by the International Nickel Company of Canada"
The Staff, Copper Cliff Smelter, International Symposium on
Copper Extraction & Refining, Las Vegas, 1976, in Extractive
Metallurgy of Copper, Vol. 1, ed. by J.C. Yannopoulos & J.C.
Agarwal, TMS-AIME, pp. 218-33, 1976.
12. "Flash Smelting", J.W. Matousek, Seminar on Extractive Metal-
lurgy of Copper, McGill University, Montreal, Canada, January
1977.
312
-------
13. "The Inco Oxygen Flash Smelting Process", C.M. Diaz, Seminar
on Copper Metallurgy, Bogota, Columbia, February 1978.
14. "Inco's Oxygen Flash Smelting Process for Copper and Nickel
Concentrates", C.M. Diaz, H. C. Garven and M.Y. Solar, Non-
Ferrous Metals Mission to the People's Republic of China,
Peking, China, January 1979, available through Inco Tech.
15. "Sulphuric Acid and Liquid Sulphur Dioxide Manufactured from
Smelter Gases at Copper Cliff, Ontario", R.W. Allgood, CIM
Transactions, Vol. 55, pp. 123-5, 1952.
16. "Handling Acid Slurries at the Flash Smelting Plant of the
International Nickel Company of Canada", J.N. Lilley, R.W.
Chambers and C.E. Young, CIM Annual Meeting, Toronto,
April 1960.
17. "Energy Use in Sulfide Smelting of Copper", H.H. Kellogg
and J.M. Henderson, International Symposium on Copper
Extraction and Refining, Las Vegas, February 1976, in
Extractive Metallurgy of Copper, Vol. 1, ed. by J.C.
Yannopoulos and J.C. Agarwal, TMS-AIME, pp. 373-415, 1976.
18. "Pollution Control and Energy Consumption at U.S. Copper
Smelters", P.A. Schultz, Journal of Metals, Vol. 30, No. 1,
pp. 14-20, January 1978, with further discussion by H.H.
Kellogg and D.A. Schultz, Vol. 30, No. 6, p. 39, May 1976.
19. Private communication to Inco Tech, 1978.
20. "Standards Support and Environmental Impact Statement, Vol.
I: Proposed National Emission Standards for Arsenic Emissions
from Primary Copper Smelters (Draft)", U.S. EPA, Office of
Air and Waste Management, June 1978.
313
-------
PROCESSING OF COPPER SMELTING GASES AT NAOSHIMA SMELTER
H. Uchida, N. Kikumoto and M. Hayashi
Mitsubishi Metal Corporation
Naoshima, Japan
ABSTRACT
The emission levels of sulfur oxides and particulate matter are strictly
regulated in Japan. To comply with these regulations, all the gases produced
from various furnaces, such as fluo-solid roasters, reverberatory furnaces,
converters, and the new continuous smelting furnaces, are now treated through
acid plants at the Naoshima Smelter of Mitsubishi Metal Corporation. At
present., the Naoshima Smelter is the only copper smelting plant in the world
which produces sulfuric acid from the calcine-fed reverberatory furnace gas
low in sulfur dioxide concentration. This paper describes mainly the techni-
cal aspects of the total gas handling system at Naoshima with other features
of the closed gas scrubbing solution treatment, gypsum production and newly
installed desulfurizing plant for the acid plant tail gases.
INTRODUCTION
The Naoshima Smelter of Mitsubishi Metal Corporation is a custom smelter
of copper located on the Naoshima Island in the Seto Inland Sea. The present
total production capacity is 168,000 metric tons of electrolytic copper per
year as shown below.
Production Capacity
(mt cathode/year)
Continuous smelter line from concentrate
48,000
Reverberatory smelter line from concentrate
96,000
from scrap & blister
24,000
Total 168,000
The history of the Naoshima Smelter dates back to 1917 when the production
started with a capacity of 3,600 metric ton blister/year by a green-charge
reverberatory furnace line. The smelting capacity was gradually expanded there-
after together with various improvements and modifications. Then in 1969, an
additional calcine-charge reverberatory furnace line was constructed to expand
the over-all production capacity close to the present level.
On the other hand, it was becoming difficult to emit the reverberatory
314
-------
furnace off-gas to atmosphere under growing pressures from pollution-control
requirements and the old No. 1 reverberatory smelter facilities were becoming
obsolete. At the same time, the Mitsubishi Continuous Smelting and Converting
Process was then under successful semi-commercial operation. Under such cir-
cumstances, the No. 1 reverberatory smelter was scrapped down in 1974 to be
replaced by a new continuous smelter with a similar production capacity.
Concurrently with the replacement of the old smelter line, a series of
pollution-control facilities were constructed, including an acid plant to treat
low SC>2 gas from the reverberatory furnace, a plant to neutralize weak sulfuric
acid bled off from the new acid plant with calcium carbonate for production of
gypsum, and an effluent treatment plant to treat the washing water for the gas
purification of the acid plants. For compliance with further stringent pollu-
tion-control regulations put into effect thereafter, the construction started
for a desulfurization plant to treat the tail gas of the single-contact acid
plant for the reverberatory furnace gas, which was completed and put into
operation in March, 1978. The smelter operation has been stable with all the
new facilities and the over-all sulfur fixation rate has now reached to 99%.
This paper outlines the present operation and describes the pollution-
control measures incorporated at Naoshima.
OUTLINE OF SMELTER OPERATION
Raw Materials
During the initial period of operation, the Naoshima Smelter treated a
considerable tonnage of domestic copper concentrate. At present, however, the
domestic concentrate accounts only for 3% of the total concentrate treated due
to a rapid increase of the production capacity and depletion of minable domestic
ores. This percentage is expected to decrease further in future.
The Smelter is thus heavily dependent on raw materials imported from
several foreign countries, such as Philippines, Australia, Canada, etc., which
naturally differ from each other in tonnages and compositions. The Smelter is
provided with a 50,000 mt capacity storage house to meet fluctuation in the
tonnages of concentrates received, and with a bedding house to meet fluctuation
in the composition and the contents of impurities. Secondary materials are
also treated, mainly with converters, the tonnage of which tends to vary con-
siderably depending on the prices of copper.
Smelting Step
The over-all flowsheet of the Smelter is shown in Figure 1. The major
specifications of the reverberatory smelter facilities are given in Table 1.
In the reverberatory line, the bedded concentrate mixed with fluxes is
first fed to a fluo-solid roaster for partial roasting, in which approximately
40% of the sulfur content is removed. The main portion of calcine (80 - 90%)
is carried over with the gas from the roaster. The flux materials with under
315
-------
Figure 1. Over-all flowsheet of Naoshima Smelter
to
CRUDE TIN
ImJ
OXYGEN PLANT
ANODE
CONVERTER DUST
BYPRODUCT
PLANT
• DI±
rh
TIN TANK HOUSE
1 i
NAOSHIMA PIER
STORAGE BUILDING
^^BEDDING ^^BEDDING
• YARD • YARD
FLUO-SOLID
ROASTER
f of
DRYER
REVERBERATORY •
FURNACE T
"J
S-F'CE
f*—Tf * r*jr SH-~F CE
CONVERTER I I MITSUBISHI
T PROCESS
ANODE FURNACE ANODE FURNACE
CONVERTER SLAG
4
MILL
•I"
I
5?
t
ANODE
TANK HOUSE
GRANULATED ^^
ELECTROLYTIC Pb-Sn SLAG COPPER ANODE SLIME CRUDE NICKEL SULFATE
TIN ALLOY ZTNC IRON CONCENTRATE ELECTROLYTIC
SULFATE CONCENTRATE COPPER
MITSUBISHI COMINCO
LEAD SMELTER
CONC ACID GYPSUM
FIRE REFINED LEAD
-------
Table 1. MAJOR SPECIFICATIONS OF SMELTER FACILITIES
Facility Number
Fluo-solid roaster 1
Reverberatory furnace 1
Converter 3
Anode furnace 2
f
Dimensions
Charge
Bed temp
Boiler
Dimensions
Capacity
Boiler
Dimensions
Capacity
Boiler
Dimensions
Capacity
Specifications
5.3m6 x S.OmL
55 wet ton/hr
630°C
7 . 5 ton steam/hr
9mW x 33mL
1,200 ton charge/day
33 ton steam/hr
4m x 9mL
150 ton blister/cycle
15 ton steam/hr
4m(j> x lOrnL
350 ton anode/batch
3 mm size, which act as a constituent to form a stable fluidized bed, are dis-
charged through the roaster bottom. The off-gas is sent through cyclones for
collection of calcine, cooled through a waste-heat boiler to approximately
350°C, passed through an electrostatic precipitator for removal of the dust
content to 0.3 g/Nm and then sent to an acid plant. Superheated steam re-
covered by the boiler is sent with the steam recovered by the reverberatory
furnace boiler to a power generator. The calcine collected through the cyclones
and the precipitator is transferred with the under flow materials by chain con-
veyor to the reverberatory furnace.
The calcine is fed to the reverberatory furnace by Wagstaff guns to be
spread over the molten bath for quick smelting. Six low-pressure burners fire
5,500 £/hr of heavy oil with the unit oil consumption of 110 £/mt solid charge.
Matte is controlled at 43% copper grade and discharged through four tapholes
(two holes provided on each side of the furnace). Slag is skimmed through an
overflow hole located near the furnace uptake, granulated by sea water and
marketed. The off-gas is cooled through a waste-heat boiler from 1,300°C to
around 350°C, passed through an electrostatic precipitator for dust removal to
0.1 - 0.3 g/Nm3 and sent to an acid plant.
Three Fierce-Smith converters are installed, two in hot state and one on
repair or cold standby. Each converter is equipped with 52 tuyeres, all punched
by a Gaspe-type mechanical puncher. The off-gas is cooled through a boiler,
passed through an electrostatic precipitator for dust removal to 0.3 g/Nm3 and
sent to an acid plant. Steam recovered by the converter boilers is used as
317
-------
process steam for the smelter and refinery operation, while the precipitator
dust is sent to a dust treatment plant for recovery of valuable elements. The
converter slag is slow-cooled and sent to a flotation plant to recover copper-
rich concentrate and iron-rich concentrate, the former to be recycled to the
fluo-solid roaster and the latter to be sold mainly to cement mills. By flo-
tation of the converter slag, magnetite build-up in the reverberatory furnace
was reduced and the over-all copper recovery was improved .
Two 350 mt capacity anode furnaces are provided, using ammonia as a reducing
gas. Each anode weighs around 380 kg. The anodes are pressed by a straighten-
ing machine and sent to the tankhouse.
The Mitsubishi Process continuous smelter entered operation in March, 1974.
The features of the Mitsubishi Process and the operation of the commercial unit
have been described in the previous papers2' 3> "*» 5' 6. The off-gas from the
smelting furnace and the converting furnace is cooled through the respective
boiler, passed through an electrostatic precipitator and sent to an acid plant.
A lead smelter of Mitsubishi-Cominco Smelting Co., Ltd., a joint venture
company between Mitsubishi Metal and Cominco Ltd., is located adjacent to the
copper smelter. The lead smelter treats concentrate from the Pine Point mine
with a production capacity of 36,000 mt pyrometallurgical lead per year. The
off-gas from the sintering furnace is treated at an acid plant of the copper
smelter.
Table 2 shows the production figures in May, 1978 at Naoshima.
POLLUTION CONTROL REGULATIONS
Along with rapid economical growth in Japan, there has been an increasing
social pressure on control of pollutants generated at industrial plants since
around the latter part of 1960 's and the pollution-control regulations by the
Central Government gradually became stringent. The emission level of SOX is
regulated by the so-called "k-value limitation" shown below.
q = k x He2 x 10~3
where q : Emission of SOz (Nm3/hr)
He: Effective height of stack (m)
k: Constant
The emission level of SOz through each source is regulated by the above
equation with the k value determined for the respective area. Since a higher
stack would increase the allowable emission level, the industry constructed high
stacks to comply with the regulations. Table 3 shows the change of the k value
in the Noashima area together with the maximum ground level concentrations of
calculated in accordance with diffusion formula of Sutton and Bosanquet.
Since a few years ago, however, the local residents started to increase
their voice for pollution control, particularly in the heavily-concentrated
industrial areas, resulting in more stringent regulations contracted between
318
-------
Table 2. TYPICAL PRODUCTION FIGURES
(May, 1978)
(metric tons)
Continuous Smelter Line
Concentrate smelted
Anode produced
Granulated slag
Reverberatory Smelter Line
Concentrate smelted
Anode produced
(from concentrate)
(from blister & scrap)
(from anode scrap)
Granulated slag
16,100
4,500
11,600
28,200
12,800
(8,500)
(1,500)
(2,800)
10,600
Converter Slag Flotation
Copper cone produced
Iron cone produced
Tankhouse
Cathode produced
Crude Ni sulfate produced
Acid Plant
Cone sulfuric acid
Neutralized gypsum
Desulfurdized gypsum
1,100
8,000
13,700
20
40,200
4,000
1,800
Table 3. CHANGE OF K VALUES IN NAOSHIMA AREA
Effective Date
K Value
Maximum Ground Level
S02 Concentration (ppm)
June 1971
April 1972
June 1974
December 1976
26.3
22.3
11.7
11.5
0.0447
0.0379
0.0199
0.0197
319
-------
the industries and the respective local government authorities. In most areas,
these new regulations are based on the "total quantity emission limitation"
and requires the control of the emission level of S02 regardless of the stack
height.
The Naoshima Smelter entered into an environmental protection agreement
with the local government with respect to both emissions and effluents, which
came into effect in April, 1978. Table 4 shows the regulations by the local
government for emission of S02, the S02 level in the environment, emission of
particulate matters, and disposal of liquid effluents, as compared with those
by the Central Government. For S02, these new regulations require that the
total emission through the stacks of the Smelter should not exceed 295 Nm /hr
and the average emission per day should not exceed 185 Nm /hr. For the ambient
S02 concentration, the prefectural government can now order the Smelter to
reduce the operation if the concentration of over 0.2 ppm continues for three
hours, over 0.3 ppm for two hours, or over 0.5 ppm for one hour. All the emis-
sion levels are monitored by the telemetering system. At the Naoshima Smelter,
the volume of, and the S02 strength in, the off-gas through the three major
stacks, and the ambient S02 in the selected points on the Naoshima Island are
constantly telemetered to the prefectural government. Under such regulations,
even a minor misoperation would not be permissible.
While the emission levels of particulate matters and the disposal levels
of liquid effluents are not yet telemetered, the regulations are far more
stringent than those by the Central Government.
MEASURES FOR STRINGENT REGULATIONS
Gas Train System before 1974
Figure 2 shows the gas flow at Naoshima before 1974. The off-gas from the
two reverberatory furnaces were, after dust removal, emitted to atmosphere.
The roaster off-gas and part of the converter off-gas of the No. 1 reverberatory
smelter was treated at the No. 1 Lurgi-type single-contact acid plant (called
"L-l" at Naoshima) for production of concentrated sulfuric acid. The remaining
part of the converter gas and the sintering furnace gas from the lead smelter
were treated at the then-existing Petersen-type acid plant for production of
70% sulfuric acid. The tail gas from this acid plant was washed, together
with the tail gas from the L-l, with sea water and emitted to atmosphere. The
roaster gas and the converter gas of the No. 2 reverberatory smelter were
treated at the No. 2 Lurgi-type double-contact acid plant (called "L-2" at
Naoshima) for production of concentrated sulfuric acid. The tail gas from the
L-2, with 200 ppm S02, was emitted directly to atmosphere.
With the gas train system described above, the over-all sulfur fixation
was 85%. It was then expected that direct emission of the reverberatory furnace
gas to atmosphere would be restricted in future and there was concern over
considerable leak gases at the old No. 1 reverberatory smelter.
Sulfuric Acid Production from Reverberatory Furnace Gas
320
-------
Table 4. POLLUTION CONTROL REGULATIONS
Central Government
Kagawa Prefecture
S02 Emission
Below k = 11.5
for each emission source
Maximum: below 295 Nm /hr
Average per day:
below 185 NmVhr
Ambient SOa
Average per day:
below 0.04 ppm
averaging the hourly
values
Maximum hourly value:
below 0.10 ppm
20% operation reduction:
over 0.2 ppm for 3 hrs
over 0.3 ppm for 2 hrs
over 0.5 ppm for 1 hrs
50% operation reduction:
over 0.5 ppm for 2 hrs
80% operation reduction:
over 0.5 ppm for 3 hrs
over 0.7 ppm for 2 hrs
Particulate Total:
Matters
Pb :
Cd :
Effluents pH :
& • O • •
COD :
Zn :
Cd :
Pb :
As :
Cu :
below 200 mg/Nm3
Q
below 10 mg/Nm
below 1.0 mg/Nm3
5-9
150 ppm
120 ppm
5 ppm
0 . 1 ppm
1.0 ppm
0 . 5 ppm
1.0 ppm
Total:
Pb :
Cd :
Cu :
PH :
S.S.:
COD :
Zn :
Cd :
Pb :
As :
Cu :
below 100 mg/Nm
below 5 mg/Nm
below 0.5 mg/Nm
below 5 mg/Nm
5-9
35 ppm
10 ppm
-
0.05 ppm
0 . 5 ppm
0.3 ppm
1.0 ppm
321
-------
Figure 2. Gas treatment flowsheet (before 1974)
NO.2 REVERB LINE
NO.l REVERB LINE
MITSUBISHI
COMINCO
HIGH STACK
HIGH STACK
ROASTER REVERB
SINTERING
F'CE
NEUTRALIZA-
TION TOWER
STACK
STACK
322
-------
Due to stringent pollution control requirements and economical reasons,
it was decided in 1972 to replace the No. 1 reverberatory smelter with the new
continuous smelter. The problem remaining to be solved was then treatment of
the 1-2% S02 gas from the No. 2 reverberatory furnace.
Two alternative methods were studied for treatment of the reverberatory
gas.
(1) Direct desulfurization of the gas to produce stable sulfur products, such
as gypsum, or to produce concentrated S02 gas by an absorption-desorption
cycle.
(2) Mixing the gas with the high SOz gas from the continuous smelter for
production of sulfuric acid.
Various desulfurization methods were under development or in actual prac-
tice at that time, but all the processes had some disadvantages and no single
process was available to meet all the requirements. The main problems of
desulfurization were as follows.
(1) While less energy-consuming than acid production, desulfurization requires
considerable costs for process supplies, such as a neutralizing reagent.
(2) There are few experiences with treatment of 1 - 2% S02 gas.
(3) If a solid material, such as lime, is to be used for neutralization, there
is concern over various operational difficulties for handling of slurries
throughout the plant.
On the other hand, the acid plant operation was fully established and
reliable. Acid production had an additional advantage in that industrial-
grade sulfuric acid was already produced in large tonnages and, among the
sulfur compounds, most stable in prices and marketable tonnages. Reliability
was given the highest priority in order not to disturb the smelter operation
and it was finally decided to produce sulfuric acid from the reverberatory
furnace gas.
While it would be possible to produce sulfuric acid only from the rever-
beratory gas, this-process would consume considerable energy for water and
heat balancing. The Onahama Smelter of Onahama Smelting & Refining Co., Ltd.
experienced the production of sulfuric acid only from the 2.4% SOa gas from
the green-charge reverberatory furnace7. This method was featured by the
following.
(1) Use of a refrigerator in the gas purification step to reduce the gas tem-
perature to below 6°C, thus maintaining the moisture content entering the
converting step at a sufficiently low level.
(2) Constant firing of a preheater to provide supplemental heat.
The plant was, therefore, a large energy consumer and the operation was
discontinued one year after the start-up at Onahama. The gas is now treated
323
-------
p
by a gypsum plant .
The off-gas from the calcine-charge reverberatory furnace at Naoshima is
further lower in the SOa strength than at Onahama. If mixed with the gas from
the continuous smelter, the SOa strength will increase to 4 - 5%, which is the
minimum strength to self-sustain the heat balance in acid production without
external heat supply.
At Naoshima, sea water is used as a coolant in the acid plants, the
temperature of which rises close to 30°C in the summer season. With a carbon-
cooler, the economical limitation of cooling is considered to be 40°C and as
far as this system is employed, it would be impossible to maintain the water
balance in summer.
Figure 3 shows water vapor pressures over sulfuric acid and the corres-
ponding SOa concentrations required to maintain the water balance. As shown
in the Figure, an increase of the acid concentration would decrease the water
vapor pressure and the corresponding SOa strength. For example, if 50% sulfuric
acid is to be circulated, the water balance can be maintained with the SOa
strength of 2.8% even at the cooling temperature of 40°C. By application of
this relationship, the new No. 3 acid plant (called "L-3" at Noashima) incor-
porated a drying tower using dilute sulfuric acid. Dilute sulfuric acid bled
off from this drying tower is neutralized with limestone for production of
gypsum, thus maintaining the water balance without a refrigerator. This method
may be called "indirect desulfurization of the reverberatory furnace aas via
sulfuric acid".
Other Measures
Concurrently with the construction of the L-3 and the neutralization
plant, the following additional measures were taken.
(1) The Petersen-type acid plant would be shut down due to obsolescence of
the facilities and the high NO content in the tail gas.
(2) The tail gases from the three acid plants would be blended and passed
through a mist precipitator to be newly installed and emitted to atmos-
phere through the stack 270 m high from the sea level.
(3) Scrubbing water in the gas purification step would be treated in the
newly-instailed liquid effluent treatment plant to cope with the expected
regulations on the effluent control. While the waste scrubbing water
contained volatile elements carried from the smelting operation, such
as As, F, Pb, Zn and Se, it was only neutralized with NaOH before 1974.
(4) The spill gases around the furnaces, which were causing some problems to
the ambient SOa concentration and the working conditions, would be collec-
ted as much as possible, passed through a bag filter for dust removal and
then emitted to atmosphere through the high stack.
324
-------
Figure 3. Vapor pressure and S02 strength over sulfuric acid
lo-
g-
s'
S 7"
H 6_
I 5"
§ 4-
co
CM
o
CO
3-
2
1
0
g. 50H
40-
i 30~
PM
g 20H
10-
o-
S02 STRENGTH LIMIT
FOR WATER BALANCE
70%
WATER VAPOR PARTIAL
PRESSURE OVER H2S04 BATH
70%
n 1 1 1 1 1 1 I r~
10 20 30 40 50 60 70 80 90
TEMPERATURE (°C)
325
-------
Measures for Pollution Control Agreement with Kagawa Prefecture
The following additional measures were taken for stringent pollution con-
trol requirements imposed by the prefectural government, which were all completed
by March, 1978.
(1) The tail gas from the single-contact L-3 normally contained 2,000 ppm of
SOa, which corresponded to the S02 emission of 250 Nm3/hr and already
exceeded the limitation level. It was then decided to install a desul-
furization plant. As to the process, a basic aluminum sulfate method
was chosen for low investment cost, easy operation and low running cost.
(2) For the off-gases from heavy oil burners, which had been emitted directly
to atmosphere, it was decided to use low-sulfur oil (less than 1.2% sulfur)
instead of Bunker C oil containing 2.7% sulfur.
The current gas flow at Naoshima, incorporating all the measures described
so far, is shown in Figure 4 together with the gas balance.
DESIGN AND CONSTRUCTION
L-3 and Neutralization Plant
Construction of the L-3 and the neutralization plant was started in 1973
and completed in February, 1974.
Figure 5 shows the flowsheet of the L-3 and Table 5 gives the major speci-
fications as compared with the L-l and L-2. The design features of the L-3
are as follows.
(1) The No. 1 drying tower using 45 - 60% sulfuric acid was installed to
maintain the water balance.
(2) Heat-insulation of the converting section was strengthened to maintain
the heat balance with the 4-5% SOa gas. In order to maintain as con-
stant as possible the heat load in the converting section even with
fluctuation of the gas volume and S02 strength, a fully automatic pre-
heater was installed so as to make possible the operation only with the
reverberatory furnace gas.
(3) The converters are designed so as to effectively absorb thermal expansions
and shrinkages to prevent gas leakage.
(4) For cooling of concentrated sulfuric acid, an irrigation cooler was em-
ployed in the previous installations. However, the irrigation cooler
had certain disadvantages, such as dissolution of Fe into acid, generation
of steam mist and requirement of a large area for installation, Through
the experiences at Onahama, it was also known that a considerable volume
of NOX would be contained in the acid produced from the reverberatory
furnace gas, which would accelerate the corrosion of the cooler. Consider-
ing these factors, a Teflon-coil cooler was chosen for the L-3. There
326
-------
Figure 4. Gas treatment flowsheet (after 1974)
REVERB LINE
CONTINUOUS SMELTER MITSUBISHI COMINCO
0 3 REVERB NO. 2 NO.l
TGH STACK HIGH STACK HIGH STACK HIGH STACK
j
t *
BAG BAG
FILTER FILTER
1 SINTERING
ROASTER REVERB CONVERTER S F CE C F'CE | F'CE
1 |
1
1 ! 1
CYCLONE BOILER BOILER BOILER BOILER FILTER
BOILER COTTRELL J COTTRELL
COTTRELL |
COTTRELL
1 V V V '' 'F "
L-2 -T
(DOUBLE j
CONTACT) j
1
L-3 1 L-l ,
(SINGLE (SINGLE
CONTACT) " CONTACT)
h
I iin
MIST !
COTTRELL '
DESULFUR NEUTRALIZA-
PLANT TION TOWER
t „ ,
L _ _L
WEAK ACID
50% H2S04
f A *7 T TMT? """" ""* "™" "™ ™ '
* *
EMERGENCY GAS LINE NEUTRALIZA- EFFLUENT
TION GYPSUM TREATMENT
T TfYTTTTV T TXTC1
PLANT PLANT
GAS BALANCE
(Dry Nm3/min)
L-l
L-2
L-3
L-3 DESULFUR
TOTAL
ROASTER
600
600
REVERB
1,200
200
1,400
CONVERTER
700
200
900
CONTINUOUS
SMELTER
200
550
750
MITSUBISHI
COMINCO
200
200
TOTAL
400
1,300
1,950
200+a
327
-------
Figure 5. Flowsheet of L-3
INLET GAS
EFFLUENT
TREATMENT
SEA WATER
VENTURI SCRUBBER
GAS COOLER
1
1
1
MIST COTTRELL
NO.l DRYER
9 6%
NO. 2 DRYER
CONVERTERS
ACID TANK
ABSORBING TOWER
ACID COOLER ">
NEUTRALIZATION
50%
TEFLON COOLER
TEFLON COOLER^
CONC ACID
DESULFUR PLANT
GAS LINE
LIQUID LINE
328
-------
Table 5. MAJOR SPECIFICATIONS OF ACID PLANT
L-l
L-2
L-3
Type
Capacity (t/d, 100% H2SOO
Start of operation
Main blower
Number
Gas vol (m3/min)
Pressure (mm Aq)
Power (KW)
Catalyst (m3)
S02 in inlet gas (%)
Converting efficiency (%)
SOa in tail gas (ppm)
Lurgi single
contact
240
1957
2
650
2,500
380
65
av 6.0
97
2,000
Lurgi double
contact
700
1969
2
1,500
3,800
2,200
165
av 8.5
99.8
200
Lurgi single
contact
500
1974
2
2,100
3,600
2,500
175
aV4.5
96
2,000
were, however, no actual experiences in Japan with the Teflon cooler in
acid plants at that time and there was also some concern over plugging
or damage of tubes by cooling sea water. After survey of the actual
practices in foreign countries and extensive pilot plant tests at Naoshima,
the Teflon cooler was finally employed.
(5) A tube-type mist precipitator was installed, thus improving the space
velocity and reducing the installation area as compared with a conventional
plate-type precipitator.
(6) FRP was used for the gas purification section. For the venturi scrubber,
organic resin was used instead of glass fiber considering resistnace
against fluorine.
Figure 6 shows the flowsheet and basic design figures of the neutralization
plant. The plant has a capacity to produce 10,000 mt gypsum per month. The
plant may be one of the largest plants in the world to produce gypsum directly
from commercial-grade sulfuric acid. Prior to the plant design, a series of
semi-commercial tests were-conducted to determine the important parameters,
such as pH, temperatures, residence time and recycled quantity of seed crystals.
The maximum tonnage of dilute sulfuric acid which must be held off from the
L-3 to maintain the water balance in the sunaftet season corresponds to the gypsum
329
-------
Figure 6. Flowsheet of neutralization gypsum plant
LIMESTONE
(6,OOOT/M)
BALL MILL
SLURRY TANK
NEUTRALIZATION TANK
CRYSTALLIZATION TANK
CENTRIFUGAL SEPARATOR
GYPSUM (CaS04-2H20)
(10,OOOT/M)
r
50% ACID (FROM L-3)
(5,300T/M
AS 100% H2S04)
ACID TANK
EFFLUENT
TREATMENT
PLANT
AIR COOLING TOWER
OF
THICKENER
UF
330
-------
production of 8,000 rat/month, and the tonnage is close to zero in winter.
The plant was designed flexible to produce either sulfuric acid or gypsum
within the allowable range. Gypsum produced is sent to an adjacent gypsum
board plant by belt conveyor for manufacture of gypsum boards.
Liquid Effluent Treatment Plant
Figure 7 shows the flowsheet and normal operation figures of the liquid
effluent treatment plant, which entered operation in January, 1974. The
effluent from the acid plant purification section contains 2-3% sulfuric
acid and considerable metal elements. Two-stage neutralization is, therefore,
employed. The effluent is first mixed with calcium carbonate in the gypsum
crystallization tank controlled in the pH range of 3 - 4, where the sulfate
ion is separated as gypsum. The overflow from the No. 1 thickener is further
neutralized with lime in the second neutralization tank controlled in the pH
range of 10 - 11, where most metal contents are removed. For removal of As,
FeCls is added for coprecipitation with Fe. The sludge precipitated in the
second neutralization tank is sent through the No. 2 thickener to be then
filter-pressed. The overflow from the No. 2 thickener is neutralized with
sulfuric acid to the pH of 7 and disposed. The tonnages of gypsum and sludge
produced varies depending on the fluctuation of the tonnage of effluent treated.
The average tonnages are 600 mt/month for gypsum and 150 dry mt/month for
sludge.
Spill Gas Treatment
For handling of spill gases, maximum possible enclosure with ventilation
would be most effective and also desirable from the energy-saving standpoint.
In conventional reverberatory smelters, a large volume of spill gas is generated
for a short period of time during matte tapping, slag skimming, crane trans-
portation of melts, and converter charging and discharging. Converters are
particularly difficult to seal since the rope handling a ladle would come
above the furnace. Control of the converter spill gases presents difficult
problems in all the smelters in Japan. At Naoshima, the secondary hood was
modified for higher efficiency and an air-curtain system was employed above
the converter. The total volume of spill gases ventilated is now 4,500 Nm3/min
for both the reverberatory furnace and the converters, which is passed through
a bag filter for dust removal and emitted through the No. 2 stack 270 m high
from the sea level.
The continuous smelter employs stationary furnaces interconnected by
launders. The volume of spill gases is, therefore, much smaller and, since
the process is continuous, handling of the gases is much easier. Approximately
1,000 NrnVmin of spill gases is ventilated, mixed with the off-gas from the
concentrate dryer, passed through a bag filter for dust removal and then emitted
through the No. 1 stack 170 m high from the sea level.
Acid Plant Tail Gas Desulfurization
331
-------
Figure 7. Flowsheet of effluent treatment plant
LIMESTONE
INLET LIQUID
300T/M
15,000-20,OOOT/M
(H2S04 2-3%)
LIQUID TANK
GYPSUM CRYSTALLIZATION TANK
LIME
50T/M
FLOCCULANT
200kg/M
H2S04
DISPOSAL
(PH 3~4)
NO.l THICKENER
U.F
O.F
NEUTRALIZATION TANK
(FILTRATE)
CENTRIFUGAL
SEPARATOR
(PH 10~11)
NO.2 THICKENER
FILTER PRESS
(FILTRATE)
SLUDGE
150DT/M
GYPSUM
600T/M
332
-------
The acid plant tail gas desulfurization is based on the basic aluminum
sulfate method, for which Dowa Mining Co., Ltd. owns the basic patent. The
flowsheet is shown in Figure 8 and the main specifications in Table 6. The
Table 6. MAJOR SPECIFICATIONS OF DESULFURIZING PLANT
Type Basic aluminum sulfate method
Inlet gas L-3 tail gas
2,100Nm3/min, 3,000ppm S02
Reverb off-gas
200Nm3/min, 15,000ppm
Total
2,300Nm3/min, 4,000ppm S02
Absorption efficiency Over 95%
Gypsum produced 3,000 t/month (moisture 10%)
Blower 2,500Nm3/min x 500mm H20
700KW
SOa in off-gas Below 200ppm
plant is designed to enable treatment of a part of the reverberatory furnace
gas, in addition to the L-3 tail gas, so as to alleviate the operation load
of the L-3.
For the acid plant tail gas desulfurization, a conventional gypsum method
was initially considered. However, the basic aluminum sulfate process was
found to have the following advantages.
(1) Liquid is handled through the absorption and oxidation section and there
is no concern over plugging or erosion of the facilities associated with
the slurry-handling system.
(2) Calcium carbonate is used as an absorbent instead of lime. In Japan, the
price for one ton of calcium is 3 - 5 times more expensive with lime than
with calcium carbonate.
(3) The construction cost is lower and the over-all running cost, including
energy consumption, is lower.
(4) The process is simpler to operate.
Reliability of the desulfurization plant is an extremely important factor
in stabilization of the over-all smelter operation. For this reason, the
basic aluminum sulfate method was finally chosen for the plant.
The outline of the process is as follows.
333
-------
Figure 8. Flowsheet of desulfurization plant
COTTRELL
L-3
OFF-GAS"
U)
OJ
[CENTRIFUGAL
SEPARATOR
If
DISPOSAL
ALUMINUM SULFATE TANK
-------
(1) Absorption step
In the absorption tower, the gas comesx in contact with basic aluminum
sulfate solution and desulfurized as follows,
A1203 -AlaCSOOa + 3S02 -* A12 (S03)3 -A12 (SOO 3 .................. (1)
(2) Oxidation step
The solution is then oxidized with pressurized air in the liquid
phase.
A12 (S03) 3 ' A12 (SOO 3 + 3/2 02 -»• 2A12 (SO^ ) 3 ....................... (2)
The liquid is handled up to this step without any slurry.
(3) Neutralization and filtration step
The oxidized solution is partly discharged to be neutralized with
limestone for formation of gypsum.
2A12 (SOOs + 3CaC03 + 6H20 ->• A1203 -A12 (SOO 3 + 2CaS(V2H20 + 3C02
.................. (3)
The slurry after neutralization is thickened and fed to a centrifugal
separator. The filtrate is recycled to the absorption tower. To prevent
lowering of the aluminum concentration in the solution by water carried
into the system, part of the filtrate is bled off to be sent to the alumi-
num recovery step and, after recovery of aluminum, the water is disposed.
(4) Aluminum recovery step
An excess quantity of calcium carbonate is added to precipitate
aluminum hydroxide and dispose overflow water.
A1203-A12(S01()3 (Aq.) + 3CaC03 + 8H20 -»• 3CaS(V2H20 + 4A1(OH)3 + 3C02
.................. (4)
The precipitated aluminum hydroxide is recycled.
The by-product gypsum is transported by truck to the adjacent gypsum
board plant.
ACTUAL OPERATION
After some start-up difficulties, the new gas treatment system is now
continuing smooth operation.
Figure 9 shows the yearly averages of the S02 concentration at the location
approximately 2 km from the Smelter. As shown in the Figure, the ambient S02
level has lowered year by year along with incorporation of various pollution
335
-------
Figure 9. Change of ambient S02 concentration
0.06
0.05
0.04
3 0.03
C-J
o
w
0.02
0.01
1973 1974 1975 1976 1977 1978
YEAR
control measures at the Smelter. The ambient S02 level is now nearly equal to
that in the atmospheric background.
Table 7 shows the emission levels of SOa; and particulate matters. The
present emission level of SO^ from the Smelter is considerably lower than the
average limitation level of 185 Nm3/hr imposed by the regulations. Table 8
gives a typical sulfur balance at Naoshima, which shows that the present sul-
fur fixation rate is 99%. As to the particulate matters, the present operation
satisfies all the requirements. Generally, it is impossible to reduce the
dust content in the furnace off-gas to less than 0.1 g/Nm3 only with electro-
static precipitation. Since all the furnace gases are treated through acid
plants at Naoshima, the emission of particulate matters through the No 3
high stack is negligible as shown by the actual emission data. On the other
hand, the gas emitted through the No. 1 high stack consists of the ventilated
gas and the concentrate dryer off-gas, and the dust load is considerably high.
The bag filters for dust removal are, therefore, under close vigilance and if
the off-gas from the stack is colored even slightly, the bag filters are
adjusted. The off-gas through the No. 2 high stack is only the ventilated
336
-------
Table 7. EMISSION DATA
Gas
862 Emission Actual Emission Data
of Particulate Matters
Regulated Levels Actual Emission Data (mg/Nm3)
(Nm3/hr)
„» M i Gas Vol. SOS SOa
Maximum Normal (Nm3/min) (ppm) (NtnVhr)
n,, r-^ &<*
Cu Cd As
No. 1 High Stack Continuous smelter
ventilation
44 25
1,500 200 18 80 4.2 0.3 2.5
No. 2 High Stack Reverb smelter
ventilation
90 70 4,300 194 50 30 2.5 0.1 0.5
No. 3 High Stack Acid plants
off-gas
Others
140 80
21 10
3,900 180 42
5 Tr Tr Tr
Total
295 185
115
-------
Table 8. SULFUR BALANCE
(May, 1978)
(ton) (%)
Concentrate
Reverberatory line 9,181 62.0
Continuous smelter line 4,993 33.7
Oil
Reverberatory line 95 0.6
Continuous smelter line 32 0.2
Others 2
Mitsubishi-Cominco gas 514 3.5
Total 14,817
Output
Sulfuric acid*
Neutralized gypsum
Desulfurized gypsum
Effluent gypsum
Granulated slag (reverb)
Iron concentrate
Granulated slag (continuous)
Converter dust
Effluent disposed
Atmosphere
13,369
674
305
100
80
9
70
20
60
130
90.2
4.5
2.1
0.7
0.5
0.1
0.5
0.1
0.4
0.9
Total 14,817
(Note) * The tonnage includes fuming sulfuric acid and liquid S02.
gas and presents no problems.
Smelter Operation and Gas Treatment
At Noashima, the gases from five sources are now treated by three acid
plants. The gas flow interconnects the furnaces, acid plants and desulfuri-
zation plant by the respective booster fans and is considerably complicated.
For this reason, the operation of each plant affects the over-all operation
of the smelter. Reliability of each operation is, therefore, indispensable.
If the operation of any one of the L-2, L-3 or desulfurization plant is sus-
338
-------
pended, the operation of the entire reverberatory furnace smelter must be also
suspended. To cope with such possibilities, the blowers and other important
devices are provided with stand-by units and the respective gas treatment plant
is constantly under close vigilance. Various emergency measures are also in-
corporated. For example, inter-lock circuits are provided for the respective
gas flow connections to prevent gas leakage from the furnace in case any blower
of the gas treatment side fails.
Behavior of Impurities
The Naoshima Smelter treats more than ten different kinds of concentrates.
Aside from precious metal elements, it is desirable that the concentrates are
low in other impurities, such as As, Pb, Zn, Sb, F, etc., from the standpoints
of pollution control, cathode quality and sulfuric acid quality. In the smelt-
ing step, such impurities are removed either with the discard slag or flue
dusts.
At Naoshima, the converter precipitator dusts of the reverberatory smelter
are all bled off and treated separately. The details of the dust treatment
plant have been already reported9. At this plant, the respective impurities
are fixed in marketable forms as follows.
Zn : zinc sulfate crystals
Cd : sludge
Pb and Sn: Pb-Sn alloy
Bi : Pb-Sn anode slime
In the continuous smelter line, the generation of mechanical dusts is
smaller due to its unique feeding method. The dusts are in principle recycled
to the furnaces, but if the impurities become concentrated in the gases, the
precipitator dust is bled off10
Arsenic and fluorine are both highly volatile elements and also have high
condensation temperatures. These elements are, therefore, difficult to collect
in the precipitator. These impurities enter the acid plant, are fixed by
scrubbing water and sent to the effluent treatment plant. Since arsenic is
coprecipitated with iron in the effluent treatment plant, the arsenic content
in the scrubbing water must be checked carefully. While fluorine is easily
precipitated by neutralization, its corrosive properties cause problems in the
gas purification section of the acid plant since some fluorine dissolves as
HF and corrodes the refractories or FRP materials. Although the L-3 is provided
with fluorine-resisting materials, corrosion of the refractories sometimes
causes operational troubles in the L-l and L-2.
The arsenic balance and the fluorine balance in case that the by-products
of the effluent treatment plant are not recycled to the smelter are shown in
Figures 10 and 11 respetively.
Previously, part of the gypsum was recycled to the smelter as a calcium
source for slag making, while the sludge, with virtually no precious metal
339
-------
Figure 10. As balance
CONTINUOUS
SMELTER
INPUT
(31.8)
REVERB
INPUT
(52.7)
CONTINUOUS
REVERB II SMELTER
SLAG I \. SLAG
(20.5) \ (18.2)
CONVERTER
DUST
(23.2)
EFFLUENT
(20.0)
EFFLUENT
TREATMENT
SLUDGE
(19.5)
SLIME
(0.9)
IRON CONC
(1-8)
GYPSUM
(0.5)
(0.03)
340
-------
Figure 11. Fluorine balance
CONTINUOUS
SMELTER
INPUT
(32)
REVERB INPUT
(68)
EFFLUENT
TREATMENT
CONVERTER ACID &
SLAG OTHERS
(3.1) (0.4)
GYPSUM
(49.1)
DISPOSAL
(3.7)
341
-------
contents and a high moisture content, was stored in the stockyard. Under pollu-
tion control regulations, however, simple storing of the sludge became difficult.
There was also limitation on the space available for storage and it then became
necessary to recycle the sludge to the smelter. However, if both gypsum and
sludge are recycled, fluorine was expected to exceed the maximum allowance
limitation. It was finally decided to market all the gypsum and recycle only
the sludge. This recycling method started in March, 1978 and has been satis-
factory to the present.
CONCLUSION
All the furnace gases are now treated through acid plants at Naoshima and
the tail gas from the single-contact acid plant is desulfurized. The rever-
beratory furnace gas low in SOz is treated by indirect desulfurization combin-
ing an acid plant and a gypsum plant. The washing water for the acid plants
is treated by the effluent treatment plant. The gas treatment plants are all
in satisfactory operation, resulting in the over-all sulfur fixation of 99%.
ACKNOWLEDGEMENT
The authors wish to thank Dr. T. Nagano, Senior Managing Director, for his
permission and encouragement to publish this paper. They also wish to thank
all the staffs and employees at the Naoshima Smelter for their support and
cooperation in construction and operation of the gas treatment facilities.
REFERENCES
(1) Itakura, K., Nagano, T. and Sasakura, J.; "Converter Slag Flotation - Its
Effect on Copper Reverberatory Smelting Process", The 99th AIME Annual
Meeting, February, 1969.
(2) Suzuki, T. and Nagano, T...; "Development of New Continuous Copper Smelting
Process", Joint Meeting MMIJ-AIME, May 24 - 27, 1972, Tokyo, Japan.
(3) Suzuki, T.; "The Mitsubishi Process - Operation of Semicommercr'.al Plant",
The Latin American Congress of Mining and Extractive Metallurgy, August
25 - September 3, 1973, Santiago, Chile.
(4) Suzuki, T., Ohyama, I. and Shibasaki, T; "Computer Control in Mitsubishi
Continuous Copper Smelting and Converting Process", The 103rd AIME Annual
Meeting, February, 1974, Dallas, Texas.
(5) Nagano, T. and Suzuki, T.; "Commercial Operation of Mitsubishi Continuous
Copper Smelting and Converting Process', The 105th AIME Annual Meeting,
February, 1976, Las Vegas, Nevada.
(6) Nagano, T. and Suzuki, T.; "Process Development for Continuous Copper
Smelting and Converting", The 15th Annual Conference of Metallurgists,
CIM, Ottawa, Canada, August, 1976.
342
-------
(7) Niimura, M., Konada, T. and Kojima, R.; "Sulfur Recovery from Green-Charged
Reverberatory Off-Gas at Onahama Copper Smelter", The 102nd AIME Annual
Meeting, February, 1973.
(8) Itakura, K., Ikeda, H. and Goto, M.; "Double Expansion of Onahama Smelter
and Refinery", The 103rd AIME Annual Meeting, February, 1974, Dallas,
Texas.
(9) Suzuki, T., Uchida, H. and Mochida, H.; "Converter Dust Treatment at
Naoshima Smelter", The 106th AIME Annual Meeting, March 6-10, 1977,
Atlanta, Georgia.
(10) Suzuki, T. and Shibasaki, T.; "Behavior of Impurities in Mitsubishi Con-
tinuous Copper Smelting and Converting Process", The 104th AIME Annual
Meeting, February, 1975, New York, New York.
343
-------
PARTICULATE EMISSION CONTROL
AT THE GARFIELD SMELTER
Frederick E. Templeton
Kennecott Copper Corporation
Utah Copper Division
Salt Lake City, Utah 84147
ABSTRACT
The smelting system employed by Kennecott at the Utah Copper Division
smelter at Garfield, Utah meets the particulate emission limitations which
have been selected by EPA to achieve ambient air quality standards for
particulate matter. This paper describes how this is accomplished.
Furthermore, in achieving this clean air goal, we have identified an
opportunity to reduce the cost of pollution cleanup by changing the manner
in which emission limits are specified.
THE PROCESS
In the pyrometallurgical smelting of copper concentrates, particulate
emissions and sulfur dioxide emissions share common sources of generation,
namely the smelting vessels. The Utah Copper Division smelting system is
composed of a two-stage process which is illustrated in Figure 1. The
first stage is composed of three "reactor" smelting vessels which transform
concentrate into an intermediate material called "matte" which contains
approximately 70% copper. The production schedule utilizes two reactors in
an operating mode with one down for maintenance. Matte is transferred from
the reactors to the second stage of the smelting process via ladles carried
by overhead cranes. The second stage of smelting is composed of four
converter smelting vessels which transform the matte into blister copper.
PARTICULATE SOURCES
Particulates are generated in the smelting vessels in two different
ways. First, condensed volatiles are particulates that form from metal
fumes emitted from the molten furnace bath. As the gases from the vessels
cool, the fumes condense into particulates. Second, the violent action of
the process within the smelting vessels generates mechanically entrained
dust which is carried away with the process gases. Another source of
particulate is fugitive dust that is produced in the material handling
systems and dust kicked up by the motion of heavy equipment and wind action
in the plant area.
344
-------
Figure i. FACILITIES FOR CONTROL OF SMELTER EMISSIONS
PICTORIAL FlOW SHEET (SIMPLIFIED)
KENNECOTT COPPER CORPORATION
UTAH COPPER DIVISION
U)
tfe
Ul
owioimbn GAS CLEANING PLANTS
nuiMtm (TYPICAL)
Exisime Mean?AIODS
-------
Two basic systems are used to control particulate emissions from the
smelter. The primary gas handling system treats gases from both the reactors
and converters. The secondary gas handling system collects the gases that
escape from the primary sources. Gases from the primary gas handling
system are cleaned in pollution control systems and then emitted from the
1200-foot stack. Gases from the secondary gas handling system are routed
to the 1200-foot stack.
PARTICULATE CONTROL SYSTEM
By far the greatest amount of particulate is generated in the operation
of the reactors and converters, so naturally the emphasis of the particulate
control system is on the primary gas stream. Gases from the reactors and
converters join at the new electrostatic precipitators (ESP's) and from
there are treated in identical cleaning units. Before arriving at the new
ESP's, however, the gases from the reactors and converters are treated
somewhat differently. The gases from the reactors pass through water-
cooled hoods to waste heat boilers where they are slowed down and the
larger particulate drops out. The heat extracted from the gases is used to
generate steam for plant use. From the waste heat boilers, the gases enter
hot cyclones which serve to further remove particulate matter. Gases
emerging from the hot cyclones pass through fans and are directed to the
new ESP's.
The off-gases from the converters flow through water-cooled hoods to a
dropout chamber where the larger particulate is removed. From there, they
go to shot coolers where more particulate is removed before entering the
new ESP's where they mix with the gases from the reactors.
The new ESP's serve to remove approximately 80% of the particulate.
From the new ESP's the gases go to four separate acid plants or a visible
train. Each of the acid plants is composed of a gas cleaning plant section
and an acid production section. The gas cleaning section serves to remove
the condensed volatile particulate before the gases enter the acid production
section of the acid plant. The gas cleaning section is composed of three
process steps. First, the gases pass through a humidifying tower where
they are cooled and some particulate is removed. The second step is to a
packed scrubbing tower where most of the condensed volatiles are removed.
Finally, the gas is directed to a wet electrostatic precipitator where SO,
aerosol mist is removed from the gas prior to entering the drying tower
which is the first step of the S0? removal process. The visible train is
simply the gas cleaning portion of an old acid plant. This was restored to
serve as a backup whenever there is insufficient acid plant capacity to
handle all of the gas in the S0? removal section of the acid plants. By
doing this, we can remove the particulate from the process gas stream even
when an acid plant is malfunctioning.
Another feature of the particulate control system is the secondary gas
collection system which serves to control gases that escape from the primary
sources. This is composed of secondary hooding around the reactors and
converters and at the tapping ports on the reactors. These hoods remove
gases that escape from around the primary hoods on the converters and
346
-------
reactors and also clean out the gases that escape during the slag and matte
tapping operations at the reactor. The gases collected in the secondary
gas collection system are directed to mixing chambers where they are blended
with the outlet gases from the acid plants and then released from a 1200-
foot stack.
SYSTEM PERFORMANCE
Most of the particulate in the stack gases comes from the secondary
gas collection system. Consequently, the emission rate varies dramatically
over the course of time. This is so because the operations such as tapping
which produce much of the secondary particulate emissions occur inter-
mittently. Another source of particulate variation, however, is the aerosol
mist that is occasionally produced when an acid plant is malfunctioning or
is experiencing a temperature unbalance. Taking these variations into
account, the stack particulate emission rate can be expected to follow a
statistical distribution such as the one shown in Figure 2. The average is
13.4 percent of the maximum, and the standard deviation is 17.6 percent of
the maximum.
The foregoing result demonstrates an important fact that has been
overlooked by most regulatory agencies in setting emission limitations for
stationary sources. This fact is that constant emission control technology
rarely, if ever, results in a constant rate of emission.
AN IMPROVED APPROACH TO EMISSION LIMITATION SPECIFICATIONS
Little, if any, attention has been given to how emission limitations
should be specified and what effect such specifications have on the cost
and effectiveness of environmental improvement efforts. This section
analyzes the current practices for developing emission limit specifications
and sets forth an alternative approach to specifying emission limits that
in many cases can reduce control costs and eliminate litigation and contro-
versy between regulatory agencies and sources.
Emission limitations have been based on two types of criteria. One
approach has been to select emission limits directly on the basis of what
is needed to meet ambient air quality standards. Another approach has been
to base emission limits on the degree of control that can be achieved by
specific technology—thus, the genesis of a series of four-letter symbols
such as BACT, RACT and BCPT, etc., which denote various levels of control
technology. In any case, the emission limitations that have been applied
by regulatory agencies to most stationary sources do not reflect what is
achievable or practical. This has resulted or can result in unnecessary
expenditures in an attempt to achieve a degree of control that cannot be
justified by any environmental benefits.
The current practice for specifying emission limits on the basis of
ambient standards is to use the "worst case" approach. This involves a
three-step procedure. First, a maximum ambient concentration is estimated
by use of a diffusion model or it is obtained from actual field measure-
ments. Next, an emission which produced the maximum concentration is
347
-------
0.400,
u>
00
FIGURE 2: DISTRIBUTION OF PARTICULATE EMISSIONS
0.002
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100
PARTICULATE EMISSION - PERCENT OF MAXIMUM
-------
estimated or obtained from measurements, and finally, a proportional or
rollback calculation is performed to determine the maximum emission that
would be permitted without exceeding the worst case concentration. The
result of this procedure is the specification of a single fixed maximum
emission rate for the source. This specification is often based on two
uncertain parameters. First, the maximum concentration can rarely be
determined with a degree of precision that is anywhere near what is needed
for the development of a fixed result, and second, the emission rate
corresponding to the maximum concentration is rarely known if for no other
reason than the transport delay between the source and receptor precludes a
precise correlation between concentration and emission rate. Consequently,
this procedure utilizes variables that are subject to a great deal of
uncertainty and pops up with a resultant number that must be achieved with
absolute certainty. Now this is inconsistent with the way in which ambient
standards are specified. The ambient standards are set on a statistical
basis, that is, a given concentration cannot be exceeded more than a certain
percent of time. Most stationary sources do not emit at a constant rate;
therefore, in order to determine whether the emissions from the sources
will exceed the ambient standards, one must evaluate the probability that
the emission rate will be high enough to exceed the standard at the same
time that worst case dispersion conditions exist. If the probability that
the emission rate will exceed a certain value at the same time that the
poorest atmospheric dispersion conditions exist is sufficiently small, then
ambient standards will not be exceeded. This suggests the notion of a
statistical specification for emission limitations instead of the current
practice of specifying a single maximum emission limit.
When emission limitations are based on control technology, a similar
problem arises. The present practice is to estimate the efficiency of
various pieces of control equipment operating under ideal conditions and
then calculate an emission limit based on a fixed or constant input to the
control equipment with a constant resultant output. Most processes—and
this is particularly true of smelting processes—do not produce constant
outputs of either product or gas streams. Furthermore, most control equipment
does not operate at a constant efficiency. Combining these two variable
factors results in an emission output that is variable, not one that is
constant. Again, the variation in the output depends on the variations in
the process and in the control equipment. These variations in output from
the control equipment result in an emission that has statistical properties.
Generally, regardless of the nature of the process or control device, the
resulting emission limitation would be expressed or specified in terms of a
statistical distribution. By doing this, the emission limitation can be
made to reflect a realistic assessment of what technology can actually
achieve.
If emission limitations were specified as a statistical distribution,
consistent with the performance of the applicable control technology,
substantial cost savings could be achieved. First, attempting to add
controls such that the maximum emission is equivalent to the average emission
achievable by a given technology can be extremely costly. This is because
such controls require backup equipment or production curtailments, either
of which is costly. Second, even though some regulations have permitted
349
-------
excursions over a fixed limitation under certain malfunction conditions,
the cost of analyzing and filing the reports to explain the excursion can
be substantial for both the source and the regulatory agency. A statistical
specification of emission could, in many cases, eliminate these costs and
make the achievement of environmental goals more economical.
350
-------
THE RECOVERY OF PARTICULATE MATTER AND
SULFUR DIOXIDE AT THE HIDALGO SMELTER
W. J. Chen
Assistant Smelter Superintendent
Hidalgo Smelter
Phelps Dodge Corporation
Playas, New Mexico
Abstract
The Hidalgo Smelter of Phelps Dodge Corporation in New Mexico utilizes
the unique flash smelting process developed in Finland to achieve primary
smelting of copper. The flash smelting process generates a high-strength
sulfur dioxide gas stream that may be treated by various processes to allow
the removal of the sulfur dioxide gas. At the Hidalgo Smelter the continuous
gas stream of high-strength sulfur dioxide from the flash furnace is mixed
with the intermittent, relatively low-strength sulfur dioxide gas stream from
the converters prior to treatment at the sulfuric acid facilities. The con-
tinuous feed stock of relatively high-strength sulfur dioxide gas enables the
sulfuric acid plants to maintain a stable and efficient operation.
High efficiency electrostatic precipitators are installed at the Hidalgo
Smelter to remove particulate matter from the flash furnace off-gases as well
as from the rotary dryer off-gases. The efficient operations of the electro-
static precipitators and the sulfuric acid plants have enabled the smelter to
comply with environmental regulations governing particulate emissions and
sulfur capture.
General Description of the Hidalgo Smelter
The Hidalgo Smelter consists of a rotary dryer, a flash smelting furnace,
an electric furnace, three converters, two anode furnaces with a casting wheel,
two sulfuric acid plants and an elemental sulfur plant. The elemental sulfur
plant is not presently in operation. Supporting facilities include a power
plant, a crushing plant, a bedding plant, an analytical laboratory and vari-
ous shops and warehouses.
351
-------
The copper concentrates and silica flux are received by rail. The con-
centrates are blended in the bedding plant and reclaimed by a bucket-wheel
reclaimer for transport to a(rotary dryer. The concentrates and silica flux,
which contain approximately six to eight percent moisture, are dried in the
dryer to approximately 0.2 percent moisture prior to transportation to the
furnace feed system by a pneumatic lift. Exhaust gases from the dryer pass
through a high efficiency electrostatic precipitator to remove the particu-
late matter prior to discharge to the atmosphere through the 600-foot stack.
The dried mixture is fed by drag conveyors to the flash furnace through
four concentrate burners where the concentrates react with the oxygen in the
preheated process air. Preheated air and fuel oil are controlled by computer
depending on the chemical composition of the concentrates and the desired
matte grade. The furnace gases containing about ten percent S02 travel
through the settler portion and the uptake shaft of the flash furnace prior
to entering the waste heat boiler. The furnace gases are cooled from about
2,200° F to 650° F in the waste heat boiler to produce saturated steam
which is superheated for use in power generation. The cooled gases enter
three electrostatic precipitators in parallel where the majority of the par-
ticulate matter is removed. Three induced draft fans provide the driving
force to move the cleaned gases to either the sulfuric acid plants or the ele-
mental sulfur plant (when in operation) for removal of the sulfur dioxide.
The flash furnace matte assaying about 60 percent copper is transported
to the converters for further processing into blister copper. The anode
furnaces ultimately refine the blister copper into anode copper which is cast
into 750-pound anodes. Off-gases from each converter are cooled in a waste
heat boiler to produce steam before entering two electrostatic precipitators
in parallel for dust removal. The cleaned converter gases are mixed with the
flash furnace gases in a mixing chamber prior to treatment in the sulfuric
acid plants.
Flash furnace slag containing approximately 1.5 percent copper is skimmed
continuously to an eleven megawatt electric furnace for cleaning of its copper
content. Converter slag of about five percent copper is returned to the
electric furnace for further processing. Final slag of about 0.7 percent cop-
per is discarded in a designated area. Figure 1 shows the flowchart of the
Hidalgo Smelter. The smelter was made operational June 1976 and the current
daily production at the Hidalgo Smelter is 500 tons of anode copper and 2,000
tons of sulfuric acid.
Modern instrumentation including a computer is utilized at the Hidalgo
Smelter for process control. An operator's console is provided at the central
control room to permit the technicians to implement the computer control
techniques.
352
-------
CO
(Jl
U)
FIGURE 1
HIDALGO SMELTER FLOW CHART
-------
Electrostatic Precipltators and Control of Particulate Emission^
Electrostatic precipitators are installed at the Hidalgo Smelter to con-
trol particulate emissions. Electrostatic precipitator installations include:
- Three precipitators for the flash furnace off-gases
- One precipitator for the rotary dryer gases
- Two precipitators for the converter off-gases
The flash furnace precipitators and the converter precipitators were con-
structed in parallel to increase operational flexibility and to permit main-
tenance of one of the precipitators while the others are in operation.
Each flash furnace precipitator consists of a large metal casing with two
continuous V-shaped hoppers at the bottom. The casing and hoppers are of all
welded, gastight mild steel. Vertical baffles are installed in the hoppers to
prevent any gas flow underneath the electrode systems. The hopper bottoms are
welded to fit drag conveyors which remove the dust. Rappers are installed on
the side of the hoppers to prevent dust accumulation.
Each precipitator for the flash furnace gases has four fields. One
transformer-rectifier set is provided for each field. Proper gas distribution
is accomplished by the installation of two gas distribution baffles at the in-
let of the precipitator. The baffles are perforated stainless steel plates.
A rapping mechanism is provided for the distribution baffles.
The internals of the precipitator consist of collecting plates and emit-
ting plates which are made of stainless steel. The emitting plate has a peak
type electrode where corona discharge takes place. The emitting plates are
hung between collecting plates. A rapping mechanism is provided for the emit-
ting system and the collecting system.
Actual operating experience of the flash furnace precipitators has indi-
cated that the dust removal efficiency of this type of precipitator is in
excess of 99.0 percent with inlet gases containing approximately 20 grains of
dust per dry standard cubic foot and outlet dust grain loading of less than
0.03 grain per dry standard cubic foot of gas. The high efficiency is attrib-
uted to the following factors:
1. Proper gas distribution.
2, Plate-to-plate configuration on the collecting and emitting systems
with appropriate guide bars resulting in proper alignment of the
emitting electrodes with respect to the collecting plates.
3. Adequate rapping mechanism on both the collecting and the emitting
systems.
4. Low resistivity of the flash furnace dusts.
5. Proper gas temperature as a result of continuous operation.
6. Proper maintenance of the precipitators.
The gases leaving the flash furnace precipitators are treated at the
sulfuric acid plants.
354
-------
The rotary dryer precipitator is similar in construction to the flash
furnace precipitator with the following exceptions:
1. A "plate to wire" configuration is employed.
2. The plates are made of mild steel.
3. The precipitator casing has a flat bottom with two scrapers to
remove the collected dusts.
4. The precipitator has three fields with related transformer-rectifier
sets.
Operation statistics for the dryer precipitator have demonstrated that
its dust removal efficiency is in excess of 99.0 percent. The precipitator
is subject to an inlet dust grain loading of about 100 grains per dry stan-
dard cubic foot. Dusts collected are returned to the pneumatic lift for
transportation to the flash furnace for smelting. Gases leaving the pre-
cipitator contain less than 0.03 grains of dust per dry standard cubic foot.
These gases are discharged into the atmosphere through a 600-foot stack.
The converter precipitators utilized at the Hidalgo Smelter are of
plate to wire configuration for the collecting and emitting systems. The
precipitator is made of a mild steel casing and has a single mild steel hop-
per with screw conveyors at the bottom. The precipitator has three fields
with related transformer-rectifier sets.
The efficiency of the converter precipitators ranges from 85 percent to
90 percent. Low efficiency is attributed to:
1. High resistivity of converter dust causing degradation in precipi-
tator efficiency.
2. Intermittent operation of converters causing wide fluctuation in
operating temperature which ultimately results in excessive cor-
rosion.
3. High adhesiveness of dust causing buildups on collecting electrodes
and short circuiting.
The gases leaving the precipitators are treated in the sulfuric acid
plants.
Table I indicates some typical operating parameters of the various
precipitators.
Maintenance on precipitators has high priority at the Hidalgo Smelter.
A maintenance team of six persons is assigned for the routine cleaning and
maintenance of the precipitators in order to increase the equipment avail-
ability and to maintain efficiency.
355
-------
Table I; Comparison of Precipitator Performances
Precipitator
Application
Flash
Furnace
Rotary
Dryer
Converters
Primary
Amperage
AC-Amperes
100
70
50-75
Load Amperage
DC-Amperes
0.5
0.3
0.25
Voltage
DC-Kilovolts
40
40
40
Efficiency
%
99.0
99.0
85.0
Availability
%
85
95
70
Baghouses are also used at the smelter to control dust emissions from the
following smelter activities:
-Crushing plant operation.
-Transportation of concentrates. Baghouses are installed at transfer
towers at various locations.
-Transportation of fine reverts to the electric furnace.
The maintenance team is responsible for changing bag filters and other
routine maintenance on the baghouses. There are thirteen baghouses in use
at the Hidalgo Smelter.
Operation of the Sulfuric Acid Plants
The Hidalgo Smelter was designed originally for the flash furnace gases
to be treated in an elemental sulfur plant and the converter gases to be
treated in a sulfuric acid plant with capacity of 1,180 tons per day. However,
energy costs have made it uneconomical to operate the sulfur plant. Therefore,
a second sulfuric acid plant of 2,200 tons capacity per day was constructed.
Both plants treat the mixed flash furnace gases and converter gases. Gases
from the flash furnace and the converter are mixed in a mixing chamber. The
combined gases of approximately eight percent S02 can be routed to either or
both acid plants for production of sulfuric acid. With the flash furnace op-
erating at its capacity of 2,000 tons of concentrate per day, gas volume is
approximately 190,000 SCFM with 8.0 percent S02- The original sulfuric acid
plant is designed for 90,000 SCFM at 7.3 percent S02 while the second sulfuric
acid plant can handle a maximum of 129,000 SCFM containing 8.3 percent S02.
356
-------
Figure 2 shows the sources of the smelter gases and their treatment in the
sulfuric acid plants.
Sulfuric Acid Plant
Both plants are double absorption sulfuric acid plants. Each acid plant
contains a gas cleaning section and a double absorption section. The smaller
acid plant employs a gas scrubbing/cooling tower, four lead mist precipitators
as equipment in the gas cleaning section while the larger sulfuric acid plant
utilizes two venturi scrubbers, five lead "star coolers" and six plastic/lead
mist precipitators as equipment for the gas cleaning section. The scrubbers
and coolers adiabatically cool the incoming gas from about 500° F to about
90° F. The majority of the dusts and sulfur trioxide in the incoming gases
are removed from the gas stream in the scrubber to form weak sulfuric acid and
sludge. The remainder of the dusts and sulfuric acid mists are removed by the
mist precipitators. The gas leaving the gas cleaning section is essentially
clean but saturated with moisture.
The S02 blowers draw the gases through a drying tower countercurrent to
the circulating 93 percent sulfuric acid. Moisture in the gas is absorbed by
the sulfuric acid. The dried gases pass through a mesh pad filter to remove
droplets of acid prior to entering the S02 blowers that deliver the gases
to a catalytic converter through a series of shell and tube heat exchangers.
The 93 percent sulfuric acid is diluted by the absorption of moisture. How-
ever, its concentration is restored by cross-flow of 98 percent acid from the
absorbing towers.
In the catalytic converter, the sulfur dioxide is converted to sulfur
trioxide in the presence of vanadium pentoxide in four catalyst passes.
357
-------
CO
O1
CO
10
ra
vo
CONVERTER
ELECTROSTATIC
PRECIPITATORS
FIGURE 2
TREATMENT OF SMELTER GASES IN THE
SULFURIC ACID PLANTS
o
STACK
STACK
-------
Oxygen to sulfur dioxide ratio is controlled by a process computer to allow
proper conversion of sulfur dioxide to sulfur trioxide. Normal conversion
efficiency is in excess of 99 percent. The sulfur trioxide from the second
pass is absorbed in the primary absorbing tower by 98 percent acid, and the
sulfur trioxide from the fourth pass is absorbed by 98 percent acid in the
final absorbing tower. Acid strength in the 98 percent acid circuit increases
constantly due to absorption of sulfur trioxide. However, the acid strength
is maintained by cross-flow of 93 percent acid from the drying circuit or by
addition of water. Heat generated by dilution in the 93 percent acid circuit
and by absorption in the 98 percent acid circuit is removed with cooling water
by means of shell and tube heat exchangers.
The tail gases pass through high efficiency "candle filters" installed at
the upper portion of the final absorbing tower. Gases leaving the acid plants
contain less than 650 ppm S02 • The two sulfuric acid plants enable the
Hidalgo Smelter to meet the 90 percent sulfur capture requirement imposed by
New Mexico's air pollution control regulations applicable to new copper
smelters.
The control room for the sulfuric acid plants is equipped with modern
analog controllers for automatic control. The computer control loops enable
the sulfuric acid plants to achieve a high degree of productivity and maxi-
mum sulfur recovery.
Sulfuric Acid Plants Control Room
359
-------
The operation of the sulfuric acid plants is very energy intensive.
Large amounts of electrical energy are required to operate two 5000-HP motors
in the larger acid plant and two 3000-HP motors in the smaller acid plant.
In addition, numerous pumps of various sizes are required to circulate the
sulfuric acid through the towers and to circulate cooling water through the
acid coolers. Fuel oil is consumed to operate the preheaters in order to
maintain operating temperature at the catalytic converter during periods of
temporary inactivity and prior to startup following a shutdown. Experience
at Hidalgo indicates that the average energy consumption is 2.7 MMBTU per
ton of sulfuric acid produced. Approximately 40 percent of the total energy
consumed by the entire smelting complex is attributable to pollution abate-
ment requirements.
Conclusion
With the installation of the high efficiency electrostatic precipitators
and the double absorption sulfuric acid plants, the Hidalgo Smelter is able
to comply with environmental regulations regarding particulate emission,
sulfur dioxide emission and overall sulfur capture. At the Hidalgo Smelter,
particulate emission is less than 0.02 grain per dry standard cubic foot of
gas as compared with the applicable State regulation of 0.03 grain per dry
standard cubic foot of gas. Sulfur dioxide emission at the sulfuric acid
plants is less than 650 ppm while the overall sulfur recovery is above the
90 percent requirement.
360
-------
ON THE ELECTROSTATIC PRECIPITATION OF
COPPER SMELTER EMISSIONS
W. R. Heifner
United McGill Corporation
Columbus, OH 43216
ABSTRACT
The electrostatic precipitation of high concentration, low
resistivity dust is primarily limited by two mechanisms: corona
quenching and its effect on the electrohydrodynamic secondary
flow, and the restitution and subsequent migration of charged
particles from the EP electrodes. These mechanisms are,
discussed as they apply to the conventional Cottrell geometry
and to a modified two-stage EP design. Pilot test results from
a copper smelter application are presented to illustrate
electrostatic precipitation in the latter electrode arrangement.
INTRODUCTION
Particulate emissions from copper smelting can be collected
by the electrical precipitation process and good tools for
predicting EP performance from physico-chemical data are
beginning to emerge^. Although considerable effort has been
expended towards an understanding of the particulate collection
problem, particulate collection does not appear to be the
primary concern of suppliers and users of pollution control
equipment in the metallurgical industry. The central problem
lies in the control of the hostile components of the stack
gas. These condensibles plague particulate control systems
through the deposition of strongly adhering dusts, corrosion,
and the secondary disposal of toxic substances. Problems
361
-------
associated with these aspects of pollution control are not as
predictable as particulate collection, and thus pilot testing
often becomes necessary.
The goal of the present paper is to affirm the necessity of
pilot testing for each pollution abatement or process material
recovery application. This theme is echoed from the aisles of
a recent USEPA conference on Particulate Collection Problems in
the Use of Electrostatic Precipitators in the Metallurgical
Industry (Denver, 1977). The paper begins with a discussion of
the electrostatic precipitation process in general and then
proceeds to a review of some recent work done at United McGill
on electrohydrodynamic effects in electrostatic precipitators.
Finally, some recent exploratory investigations at a copper
smelter facility will be reviewed. Although the subject
application constituted a pilot scale investigation, it is
worth sharing as an amplified example of some traditional "EP"
problems experienced in the copper smelting industry.
FUNDAMENTALS
Electrostatic precipitation employs the electrical force,
between charged bodies, to separate particulate matter from a
gas stream. The electrical force, which is responsible for the
precipitation, arises from opposing charge accumulations on
separated EP electrodes. This charge separation is
accomplished by a high potential dc power supply (transformer
and rectifier and possibly auxiliary pulsing equipment) with
its terminals connected to the EP electrodes.
The electrical precipitation process is simple:
Particulate matter is charged, then transported to a collection
surface by an electric field, and then removed to an external
receptacle. For the efficient operation of an EP, the order of
these events cannot be changed.
Application of the electrical precipitation process to
physical situations is not simple. There are several charging
mechanisms, migration and electrohydrodynamic effects, and
adhesive mechanisms which interact to determine the performance
of an electrode geometry in a given application.
The most commonly discussed charging mechanism in
electrostatic precipitation is the dc corona2. Other gaseous
charging schemes, under laboratory or Pilot scale
investigation, include pulsed corona^f^'5, ^on beam
systems6, and radioactive chargers7. Each of these devices
362
-------
serves the same purpose — gaseous ion charging. Within a few
tenths of a second after entering the charging region, a
particle gains a charge near the Pauthenier limit.
«• J+f
where e is the dielectric constant for the particle, e is the
permittivity of vacuum, Eo is the static field at the °
particle, and a is the radius of the particle.
When a particle lies on an electrode surface, it gains a
charge by induction. The charge gained by a particle, exposed
to this second charging mechanism, is near that which the
particle would have gained by the gaseous charging mechanism
(1). This charge is called the Maxwell or Felici charge8.
q = i_ (4lT£; E a2)
6 o o (2)
The charging time, corresponding to this mechanism, is on the
order of the permittivity-resistivity product — microseconds
for many metallurgical dusts.
Inductively charged particles (almost all electrostatically
deposited particles) tend to detachment unless resisted by
strong adhesive forces. The repulsion force on the particles
is on the order Of
F = 1.37 (47reoE^a2)
(3)
As a particle becomes embedded in a surface, its charge and
detachment tendency decrease to zero.
A third charging mechanism is a solid state effect, which
occurs when particle surfaces of differing composition come
into contact'. These contact effects result in a small
charge transfer which depends on the interfacial capacitance
between the materials and on the difference between their
surface state potentials. Although less significant as a
particle charging mechanism, the charge transfer due to
non-uniform composition results in very strong adhesive forces
in electrostatically deposited dust layers10. The natural
charge of dispersed dust obtained its charge from this
mechanism.
Another charging mechanism is the microdischarge process,
or corona discharge, of a particle as it approaches an
electrode surface11. This mechanism leads to charge reversal
and difficult collection conditions for low resistivity
particles.
363
-------
Finally, in real situations, the corona is bi-ionized.
That is, it contains ions of both polarities. Streamers and
back ionization, both of which contribute ions of the opposite
polarity from those of a simple dc corona, are exhibited in
Fig. 1. At present, a number of investigators are looking at
electrostatic precipitation in bi-ionized systems. 12, 13, 14
In addition to its charging function, the corona induces
secondary gas flows on the order of 1-2 m/s in the vicinity of
the grounded electrodes and on the order of 15-20 m/s in the
vicinity of the corona electrodes15. These flows transport
particles, irrespective of their charge, to the boundary layer
near the collection surfaces where they must collect
electrostatically. Particles which entrain and/or are of
opposite polarity continue with the return flow or migrate to
the corona electrode. Recently, T. Yamamoto succeeded in
numerically solving the Navier-Stokes equations for the
electrohydrodynamic field of a positive polarity EP of the
Cottrell design16. Although limited to the particle free
conditions of no cross flow, the results confirm the flow
patterns he observed using Schlieren photography. His model is
presently being extended to include the cross flow components.
The choice of EP electrode system requires a consideration
of all these mechanisms as well as a concern for dust adhesion
and removal. The many designs currently employed in the
metallurgical industry can be divided into single and two-stage
geometries. The Cottrell or wire in duct geometry employs a
series of wires (or rods) to establish a corona and to collect
the particles. To serve this dual role, a high-potential
difference is required (50KV). Low resistivity particles,
which easily reverse their charge (often before they reach the
surface) and require some tenths of a second to recharge in a
corona, are thus difficult to collect in such a geometry.
Fortunately, most metallurgical dust adheres to the electrode
surfaces on one of the first few impacts with a surface and
thus collection is achieved.
In two-stage EP's, a parallel plate collection field is
employed to collect particles of either polarity. The
potential difference between the electrodes in two-stage units
is lower than that for single stage units because it only needs
to serve one role, to establish a collection field. One loses
the local electrohydrodynamic flows, but gains more collection
area per unit EP volume and a lower electrode power
consumption. The role of the corona in two-stage EP's is to
provide the initial particle transport to the collection
electrodes.
364
-------
Fig. 1. Positive dc corona from a linear array of needles. (Note pre-breakdown
streamers and reverse ionization.)
365
-------
An industrial scale, modified two-stage EP design is
exhibited in Fig. 2. This design was employed in the pilot
testing program for two reasons: 1) The modified two-stage
design was found to perform well on two other high
concentration dusts (Iron oxide 25 gm/Nm3, Alumina 9
gm/Nm3) with similar resistivity in the EP operating range,
and, 2) There was an interest in determining the feasibility of
copper separation from fluid-bed roaster emissions. From a
research standpoint, successful collection of high
concentration dust is an indication of the lifetime for charge
on particles suspended in a bi-polar charge system.
About a year ago, C. Noll and T. Yamamoto observed
particulate collection in this geometry using mist
particles14. In Fig. 3, a mineral oil/CC>2 mist was
injected as a tracer with the gas stream. The electrodes were
operated near sparkover potential (20kV). The gas speed was
near 0.65 m/s. Note that the particles were propelled to the
collection electrode surfaces. At a much lower potential (13.5
kV) , yet still above corona onset, a similar particle transport
was observed (See Fig. 4). In this case, the mist was
oil-soaked cigar tobacco fumes. The gas speed was 1 m/s.
There are two tracer streams in the photo. A further
experiment employing Schlieren photography, revealed vortex
formation near the grounded electrode surfaces (See Fig. 5).
Similar investigations were carried out for the Cottrell
geometry. Gas velocity/pressure studies, using hot wire
anemometry and micromanometer techniques were also performed to
support the numerical studies mentioned above.
The interest in this work was motivated by the casual and
almost universal observation of dry particulate collection on
both polarity electrodes in EP's (See Fig. 6). It was
concluded that both polarity particles can, and do, normally
exist in the electrode region of an EP and that the role of the
corona in two-stage designs is to effect initial particle
transport to the grounded collection surfaces. This was
confirmed by field experience which indicated a rapid rise in
efficiency with corona onset to some steady state level
determined by entrainment — See Figs. 7 and 8. To obtain a
greater efficiency, a larger EP system would be needed.
One of the traditional problems of the implementation of
two-stage precipitators to industrial processes lies in corona
quenching1'. The current drawn by an EP electrode system is
composed of two parts: ionic and particulate.
J « cpK E + c±K1E (4)
366
-------
Fig. 2. Modified two-stage EP design. (Needle-plate in duct geometry (NPD).)
367
-------
Figure 3. Mist collection in NPD geometry (Potential at sparkover 20 kV).
Figure 4. Mist collection in NPD geometry (Potential below sparkover 13.5 kV)
368
-------
Figure 5. Electrohydrodynamic secondary flow in needle to plane geometry.
Schlieren photograph recorded with needle shank parallel to plane.
369
-------
Figure 6. Dry particulate collection on both polarity electrodes of NPD
geometry.
370
-------
15
O 10
g
Q.
UJ
5
<
Ul
5 5
o
1.0
• 1 I
I I I I-
2.0
3.0
4.0
AVERAGE FIELD -V./H, (KV/CM)
Figure 7. Corona current vs. mean field data for pilot unit.
371
-------
100
99
98
7" 97
55
5 96
95
O»
fc
2 94
o
C
93
92
91
• i i r
i i i I i i i B
6800
O 6800
0 5700
O 6300
5800
i i i i i i i
07600
o °6800
O 4600
7600
'6100
NUMBERS NEAR CIRCLES
REPRESENT EP INLET FLOW
RATE (Nm3/hr).
lilt
I i i i i I > ii i I * i i i t i • i i
1.5
2.0
2.5 3.0 3.5 4.0
AVERAGE FIELD,V0/ H - (kV/cm)
Figure 8. Efficiency vs. mean field corresponding to current data in
Figure 5.
372
-------
where c, K, E represent the concentration mobility, and static
field for the particle and ion systems (subscripts),
respectively. When the concentration of particles is high or
the dust is very fine, attachment of the ion concentration to
the particulate system leads to corona quenching. The
particles have a far lower mobility than the ions and thus the
field near the corona electrode is suppressed or bucked.
Consequently, the secondary flow of the corona is suppressed.
Although significant corona quenching was observed in the first
field, on the EOF application (Iron oxides), it did not
penetrate the second or later fields and impair the systems
collection. It was thus proposed to test this modified
two-stage design on a pilot scale smelter application. This
pilot work would extend the useful concentration range of this
geometry to at least 75 g/Nm3.
PILOT INVESTIGATION
a. Description of Pilot facility
At the copper smelter, the primary objective was to
separate copper compounds and elemental copper from the
emissions of a natural gas fired, fluid-bed roaster. It was
the contention that if the EP inlet temperature was maintained
at a sufficiently high level, volatilized arsenic and other
condensibles wo.uld pass through the precipitator, while solid
copper emissions would be collected^*.
A temporary 0.51 m diameter duct system, with dampers, was
erected, allowing a portion of the flue gas from the roaster to
enter the pilot precipitator (See Fig. 9). The total length of
the duct was approximately 40 m and was insulated. A cyclone
precleaner was installed 9 m upstream of the precipitator.
The EP exhaust fan was then used to control the volume of
flue gas from the existing emission system, with the EP's
horizontal evaporative cooler used to control the flue gas
temperature when necessary. The complete specifications of the
pilot precipitator installation is summarized in a
reference^.
The test schedule centered around running the precipitator
at the highest attainable inlet temperature while altering the
precipitator gas flow-through velocity.
b. Description of Sampling Procedure
Emission testing was accomplished by conventional EPA
Method 5, as outlined in the Federal Register20, with several
373
-------
CJ
If
o
H»
P.
I
DAMPER
ROASTER FLUE
TEST
STATION
I
CR3T
EXISTING FLUE
TEST
STATION
3
-------
exceptions. All testing was performed using an in-stack
Alundum filter ahead of the standard sample train hot box
filter. The filter, filterholder, and stainless steel nozzle
were preheated to stack temperature, with the filter gasket
tightened before each test. The in-stack filter was contained
entirely inside the test duct and prevented the traversing of
the stack as outlined in EPA Method 5. As a result, testing
was performed at the duct center line, with a center line
coefficient applied to the results.
Simultaneous measurements were secured at the inlet and
outlet of the precipitator and at the inlet to the cyclone
pre-cleaner. Also, transmissometers, mounted on the
precipitator inlet and outlet, were employed to keep a 24-hour
record of the inlet and outlet opacity.
The probe unit was attached to a pyrex filter holder,
containing one Gelman "A-E" glass fiber filter. The filter
holder was maintained at 395K (250°F) during tests and was
connected to the first of four impingers.
All four impingers were of a modified Greenburg-Smith
design, reducing pressure loss across the sample train. The
first three impingers contained 250 ml, 100 ml and 100 ml
respectively, of 30% sodium hydroxide. The fourth contained
500 g of dry silica gel. All four impingers were placed in an
ice bath to prevent the sample from rising above 295K (70°F)
in the last impinger.
The liquid in Impinger 1, the combined liquid in Impingers
2 and 3, and the impinger washings were measured and placed in
three separate polyethylene bottles. The silica gel in
Impinger 4 was weighed immediately to determine its weight
gain. The two filters (in-stack and out-of-stack) were placed
in watch glasses, over dried at 105°C, desiccated and
weighed. The in-stack and out-of-stack portions of the
sampling probe were washed with water into two separate
polyethylene bottles.
The filter particulate was dissolved with nitric acid and
analyzed along with the other solutions. The in-stack and
out-of-stack washes were evaporated at 378K (220°F),
desiccated and weighed. Weights and solution volumes were
reported for each sample, along with the assays for copper,
arsenic, zinc, bismuth, selenium, tellurium, lead and
antimony. The filter catches, plus the evaporated probe
washes, represented the dry particulate collected. All
chemical analyses were performed in the sponsor's laboratories.
c. Discussion of Results
Tests on the gas-fired fluid-bed roaster were characterized
by both high particulate loading and very high moisture
375
-------
content This combination caused problems in the existing
system by frequently plugging the cyclone pre-cleaners, and was
a particular problem if the system was allowed to cool after a
period of down time. Condensation during start-up also
resulted in cyclone plugging.
The pilot's EP's horizontal evaporative cooler was used to
obtain test data in the EP inlet temperature range from 590K
(600°F) to 645K (700°F). No tests at temperature below
590K were performed, as the EP outlet temperatures were deemed
too low for useful arsenic separation information to be
generated^.
Particulate collection efficiency across the EP was
maintained at high levels, with copper recovery rates as high
as 99% on an inlet to outlet kg/s basis (See Table I). Total
particulate collection efficiencies were also at this high
level. As a result, the EP flow through velocity was controled
at average gas speeds in the range from 0.75 to 1.1 m/s. These
tests, plus tests with only one and two EP fields energized
provided additional design data as well as arsenic separation
information. A total of 27 EPA Method 5 tests were secured
from various points along the roaster duct system.
Method 8 test results indicated extremely high sulfuric
acid and sulfur dioxide concentrations in the gas stream: 0.5
x 10^ mg/Nm^ S02- In addition, tests across the cyclone
indicated that its efficiency was low at the inlet velocity of
9.0 m/s.
Heavy flue gas dust loading and high moisture content
produced several problems. During EP startup, high moisture
condensation was observed. The initial hopper and cyclone
discharge dust was of a thin muddy consistency, and during this
period, the cyclone would easily plug. To circumvent this
problem at the EP, the EP hopper rotary valves were kept in
continuous operation.
Heat loss through the system was also a problem. Due to
height limitations, while transporting the Mobile Precipitator,
the high potential insulators were located inside the main EP
chamber. A flow of clean purge air was used to prevent the low
resistivity dust from building up on the insulators and causing
shorting or low potential arcing.
Initially, this flow of air was kept very low and several
tests were completed. However, the insulators later became
coated with dust, which was not acceptable, and the volume of
purge air had to be increased. After this adjustment, the
precipitator ran well and continued to effectively collect
376
-------
particulate emissions. Unfortunately, as a result of the
increase of insulator purge air, the EP outlet temperature
showed a significant decrease.
Dust fallout in the duct was evident with gas flow
velocities of 8.0 - 12.7 m/s. Twice during the testing
program, the inlet duct filled halfway with dust, requiring
that the precipitator be shut down and the duct cleaned.
TABLE I* Typical Mobile EP Performance Data
(Gas-fired fluid-bed roaster)
Average EP velocity
Gas Volume
Moisture
Inlet loading
Efficiency (kg/s at outlet
vs. inlet basis)
EP Field potentials
EP Field currents
Total Energy (Power) supplied
to the EP fields
EP inlet temperature
EP outlet temperature
Cu recovery
As recovery
1.13 m/s
2.49 m3/s
43%
1.96 kg/s
98.9%
26/18.5/23 kV*
2/5.7/4.2 mA
554 joules/Nm3
«
620K
500K
99.0%
68.6%
(3.70 ft/s)
(5279 ACFM)
(259 Ib/hr)
(243 watts/
1000 SCFM)
(653°F)
(445°F)
* For the high dust loading case, the plate spacing in the
first field was increased from 8 cm to 10 cm.
377
-------
CONCLUSIONS
The following conclusions were drawn from the Mobile
Precipitator pilot testing program and the fundamental studies.
1. The particulate emissions from the gas-fired, fluid-bed
roaster at this copper smelting site were easily
collected, as evidenced by a consistent 97-99% recovery
rate for the three field mobile pilot unit. This
conclusion was further based on similar experiences
with the same mobile unit on over 25 other
applications-^.
2. The separation of arsenic and copper was observed
during the Mobile EP testing program, but the results
were not definitive because of the temperature drop
across the EP. The potential for success with an EP
operating with an inlet temperature near 700K (800°F)
and an outlet temperature near 645K (700°F) remains
only speculative. Cold walls and unintentional ambient
air leakage lowered the Mobile Precipitator temperature
below that which was desired. During the testing
program, the ambient temperature was typically below
273K (0°P). Additional testing with a heated shell
precipitator is recommended for obtaining more
representative data.
3. Corona quenching is indicated in the data of Table I by
a suppression of the first field current. However, in
a test where only the first field was energized, 64% of
the inlet loading was shown to be collected in that
field. These results demonstrate that the modified
two-stage geometry would successfully collect dust
under conditions of corona quenching.
ACKNOWLEDGEMENT
The Author wishes to acknowledge Charles G. Noll, Ph.D.
for his technical assistance during the Pilot Testing Program
and during the preparation of this paper.
378
-------
REFERENCES
1 "Proceedings: Particulate Collection Problems Using
ESP's in the Metallurgical Industry", USEPA
Environmental Protection Technology Series:
EPA-600/2-77-208, October 1977, 259 pages.
2 S. Oglesby, et al. "A Manual of Electrostatic
Precipitator Technology, Part I — Fundamentals", NTIS,
Springfield, Va.: PB-196380, August 1970, 340 pages.
3 S. Masuda and J. Harai, "A Pulse Voltage Source for
Electrostatic Precipitators", Conference Record of the
IEEE-IAS 13th Annual Meeting, Toronto: Cat. No.
78CH1346-6IA, October 1978, p. 23.
4 H. I. Milde and P. L. Feldman. "Pulse Energization of
Electrostatic Precipitators", Conference Record of the
IEEE-IAS 13th Annual Meeting, Toronto: Cat. No.
78CH1346-6IA, October 1978, p. 66.
5 Private Communication with G. W. Penney regarding a
pulsed biased supply for a tri-electrode EP.
6 R. H. Davis and W. C. Finney. "Design Concepts for a
Laboratory Electrostatic Precipitator Using Electron
Beam lonization". Conference Record of the IEEE-IAS
13th Annual Meeting, Toronto: Cat. No. 78CH1346-6IA,
October 1978, p. 59.
7 R. J. Harrison, et al. "Radiation Charging: A Novel
Way to Charge Fine Particles Electrically", JAPCA 2j>
(2), 179 (1975).
8 A. Y. H. Cho. "Contact Charging of Micron-Sized
Particles in Intense Electric Fields", J. Appl. Phys.
3_5, 2561 (1964).
9 L. B. Loeb. "Static Electrification, Parts I-II" in
Progress in Dielectrics 5, j>: Plenum Press, NY,
1963-1964.
10 G. W. Penney and E. A. Klingler. "Contact Potentials
and the Adhesion of Dusts", Trans. AIEE 81, 122 (1975).
379
-------
11 S. K. Berger "Reduction of Breakdown Voltage in Uniform
and Coaxial Fields in Atmospheric Air Through Moving
Conducting Spheres", Gas Discharges IEE Conference
Publication 118, p. 380 (1974).
12 K. P. B. Kinkelin, "Elektrofilter Theorie mit
Ruckspruher: Die "Scheinbare Kombinierte
lonenmobilitat1", Staub. Reinhauting der Luft 36 (12),
469 (1976) and 3_7 (1) , 6 (1977).
13 J. S. Chang, K. Kodera and T. Ogawa. "A Computer
Experiment of Electrostatic Charging of Aerosol
Particles by Bi-polar Ions", Conference Record of the
IEEE-IAS 13th Annual Meeting, Toronto: Cat. No.
78CH1346-6IA, October 1978, p. 38.
14 C. G. Noll and T. Yamamoto. "Electrostatic
Precipitation by Contact Charging in an External
Field", Conference Record of the IEEE-IAS 13th Annual
Meeting, Toronto: Cat. No. 78CH1346-61A, October 1978,
p. 53.
15 s. Masuda. "Recent Progress in Electrostatic
Precipitation", Inst. Phys. Conf. Ser. No. 27 (Japan
1975), Chapter 3, p. 154.
16 T. Yamamtoto, Private Communication
17 M. B. Awad and G. S. P. Castle. "Efficiency of
Electrostatic Precipitators under Conditions of Corona
Quenching", JAPCA 25 (2), 172 (1975).
18 J. S. Roberts, "Controlling Roaster Off Gases at '
Campbell Red Lake Mines", Can. Mining J., p. 54 (July
1976).
19 C. G. Noll and T. Yamamoto. "A Mobile Electrostatic
Precipitator for EP Sizing and Research", Conference
Record of the IEEE-IAS 12th Annual Meeting, Los
Angeles: Cat. No. 77CH1246-8-IA, October 1977, p. 892
20 USEPA, "Standards of Performance for New Stationary
Sources", Federal Register 42 (160), Part II (1977).
380
-------
SPECIAL CONTROL SYSTEMS
Session Chairman: Grady B. Nichols
Southern Research Institute
Birmingham, Alabama
381
-------
APPLICATION OF COTTRELLS IN ASARCO'S NONFERROUS SMELTERS
E. S. Godsey
Chief Fume Recovery Engineer
Central Engineering Department
ASARCO
Salt Lake City, Utah
Abstract
For the past 66 years ASARCO has designed and built
cottrells for particulate recovery in our various copper, lead,
and zinc smelters. The ASARCO design has advantages and dis-
advantages when compared to commercially available units for dry
application. Data on inlet and outlet grain loading and par-
ticulate size distribution have been accumulated for both
roaster-reverberatory and converter cottrells. ASARCO recog-
nizes the effectiveness of gas conditioning with respect to the
control of particulate resistivity. Conditioning agents used
are: water, sulfuric acid, S03, and hydrated lime. Some ASARCO
plants use only one agent while others use a combination of two
or more. Measurements have been made that show a marked improve-
ment in particulate and 303 recovery from a gas stream when water
conditioning was used.
Introduction
The involvement of ASARCO Incorporated with the application
of cottrells at nonferrous smelters began nearly 71 years ago.
For the past 66 years the experience has included the use of
ASARCO's own designed cottrells. The company exposure to opera-
ting cottrells began in July 1907 when Dr. F. G. Cottrell
Installed the first commercially successful cottrell in the
United States, at ASARCO's Selby, California lead smelter.(1,3)
This unit, as described by F. G. Cottrell, was made up of ^
several rows of four inch wide by four feet long lead plates —
installed vertically, on four inch, centers, within a four foot by
four foot lead flue. Lead covered iron rod discharge electrodes
with mica, which gave a point discharge for the current, were
placed between each pair of plates. Much time and effort were
expended in experimentation to develop the ideal discharge
382
-------
electrode for the Selby unit. The design that was accepted as
the ultimate was of lead impregnated with mica.(2)
Generally speaking, cottrells we have built in ASARCO are
categorized on the basis of the design of the collecting elec-
trode. ASARCO's experience in dry cottrell operation over the
past 66 years has been directed to four basic designs of collect-
ing electrodes, namely; pipe, corrugated plate, wire screen, and
wire cage.
ASARCO started to design and build cottrells for our own
smelters in 1912 when the first large multiple-pipe precipitator
was built at our Garfield smelter near Salt Lake City, Utah.(3,4)
ASARCO installed only pipe, screen, and wire cage type cottrells
in the period between 1912 and 1924. After this date, all our
dry cottrell installations have been of the corrugated plate
design.
Many of the earlier design cottrells have been abandoned and
Table I lists those currently installed in ASARCO's lead, zinc,
and copper smelters. Two cottrells listed, for the El Paso lead
sinter and the Tacoma anode furnace operations, are scheduled to
start this year. All the cottrells in Table I are of the plate
design except the Tacoma reverberatory installation which con-
sists of a pipe and plate cottrell in series. Eleven of those
tabulated are ASARCO design and two, the Columbus and the Hayden
converter cottrells, are commercially available.
Table I
Cottrell Operations in ASARCO
Location
Columbus
Corpus Christi
East Helena
El Paso
El Paso
El Paso
El Paso
Hayden
Hayden
Tacoma
Tacoma
Tacoma
Tacoma
Operation
Conditioning
Zn. Roaster
Zn. Roaster
Pb. Sintering
Pb. Sintering
Cu. Converter
Cu. Roaster
Cu. Reverberatory
Cu. Converting
Cu. Roaster-Reverberatory
Cu. Reverberatory
Cu. Converter
Cu. Converter
Cu. Anode Furnace
None
None
Water
Water
Water-Acid
Water-Acid
Water-Acid
Water
Water
Water-Acid
Water-Lime
Water-Lime
Watera
Gas
Tempera-
ture °C
380
300
225
230a
138
116
116
230
125
100
138
138
93a
a Anticipated
383
-------
Design Features
The cottrells we are building today are essentially the same
design used in the mid 1920's. Over the years modifications have
been made that result in improved performance. Insofar as I have
been able to determine, we have always used corrugated plate com-
posed of iron alloyed with 0.2% copper. Rows of corrugated
plates are spaced at 6" centers. Round wires have been used as
discharge electrodes since 1912. (4)
The number of gas passages per field and the active lengths
of field vary since some of the early cottrells were installed
in existing flues and were designed to fit local conditions. The
plate height has varied from 6' to 12' with 12' plates made from
16 gauge sheets being the current standard.
We have experimented with aluminum plates and wires in at
least two installations and found little difference in the life
of aluminum compared to the steel alloy. Economics ruled out the
use of aluminum since the cost of aluminum plates was several
times the cost of iron plates.
A few of the parameters for a single electrical field are
given in Table II.
Table II
Parameters of a Basic Electrical Field
Item English Metric
Collecting electrode area 2604 ft. 242 m2
Number of gas passages 15 15
Collection electrode spacing 6 in, 0.15 m
Gas passage length (active) 7.5 ft. 2.3m
Discharge electrode spacing (along row) 3 in, 0.08 m
Discharge electrode diameter 0,08 in. 2.03 mm
Electrical volume (active) 651 ft.3 18.44 m^
Field length (active) 7.5 ft. 2.3 m
Field width (active) 7.5 ft. 2.3 m
Field height (active) 12 ft. 3.6 m
The active electrical volume in Table II is less than the
total indicated by the field dimensions because there are four
voids in each field. The lower frame of the discharge electrode
assembly is supported by beams attached to the upper support and
these beams are inside the field. Spacing between the grounded
electrodes and the emitting electrode bottom support beams is
not considered part of the active electrical field.
Our cottrells have been built with as few as two fields in
series and with as many as six. However, the most common design
384
-------
is four fields in series. The number of parallel rows of fields
is determined by the expected gas volume and our experience in
operating cottrells on a similar gas stream, in some of our
smelters we have either replaced a cottrell or expanded the
existing cottrell. In our recently designed cottrells, we
generally have had the measured efficiency for a similar type of
operation and have used these data to size the new cottrell.
We normally design a cottrell on the basis of retention time
in the electrical field. Cottrells built since 1974 have ranged
from 7.9 seconds to 16 seconds retention time. These retention
times result in a gas velocity of 3.8 (1.16 m/sec) and 1.9 feet
per second (0.58 m/sec) respectively in a four field series. The
cottrell designed for a 16 second retention time was installed on
a copper converter gas stream. Our experience has indicated that
particulate from a converter is the most difficult to collect and
the combined particulate from copper roasters and reverberatory
furnaces is one of the easiest to collect.
A basic electrical field contains fifteen rows of 26 dis-
charge electrodes suspended at three inch intervals from a three
inch diameter pipe. Each electrode is tensioned by a twelve
pound "U" shaped iron weight attached to the free end. Beneath
each row of wires, a 2-1/2" x 5/8" bar is so installed that the
tensioning weights are suspended over it. The bar arrangement is
such that when an electrode breaks the "U" shaped weight is
retained on the bar thus preventing it from falling into the hop-
per and conveyors which are underneath.
The electrode suspension frame is supported by "I" beams
placed at either end. The "I" beams are, in turn, supported by
four stacks of insulating brick. Nominally, each stack is made
up by alternating 7 and 9 inch square bricks, each 1-3/4" thick,
for a total of eight bricks. Each stack of bricks is housed in a
metal box located in the gas atmosphere below floor level. Since
the supported "I" beam passes through an opening in this box the
supporting bricks are exposed to the gas and particulate. Because
of this there is a propensity for current to track or short across
the stack of bricks. Cleaning the brick at weekly intervals
reduces the frequency of shorting.
Each electrical field has a multiple number of cover plates
to allow easy access for replacing broken electrodes and for
major repair work. Our latest designed covers have an elliptical
shaped gasket fastened to the bottom edges of each cover to
exclude air infiltration. The older design covers do not have
gaskets and the floor is generally covered with sheets of either
plastic or rubber to reduce air infiltration.
Dust recovered in our cottrells is removed by either screw
conveyors or drag conveyors, with screw conveyors the usual
method for cold cottrells and drag conveyors for hot cottrells.
385
-------
One plant has a combination screw conveyor system and air con- v
veying system to return dust to the smelting circuit. In three
of our cottrells dust is returned directly to the smelting cir-
cuit in closed conveying systems. Dust collected in other
cottrells is put through a dedusting device prior to returning it
to the charge preparation system.
Electrical
Our early cottrell electrical power supply included either a
15 or 25 KVA transformer for each field, a mechanical rectifier,
and a hand-controlled power selector operating through a series
of resistors for voltage control.
Over the years, starting in the 1950fs and during the 1960's,
the mechanical rectifiers were gradually replaced, first with
selenium conversion kits, and then eventually with silicon con-
version kits connected to the original transformers.
Starting in 1963, we experimented with an automatic voltage
control device which adjusted the primary voltage through the use
of a multiple number of mercury switches connected to existing
resistor banks. As the state of the art advanced, we replaced
the original automatic voltage regulators with saturable core
reactors. As new cottrells were built, we installed saturable
core reactor controls. Subsequently, thyristor controls have
replaced saturable core controls on two of our converter
cottrells. Our most recently commissioned cottrells have all
been equipped with thyristor controls. The power supply for
these cottrells consist of 45 KV-1000 MA transformers with built-
in rectifiers and thyristor automatic voltage regulators.
We found by experimenting on a copper converter cottrell
that thyristor controls gave a substantially higher power input
than was possible with saturable core reactor controls on this
variable gas stream. Some of our plants which have a constant
gas condition, and in which saturable core controls were origi-
nally installed, continue to operate with this type of control.
During the transition from mechanical rectifiers to the pre-
sent automatic controls, the number of electrical fields connected
to each transformer was changed. We currently have 3, 4, or 5
fields per transformer on our own designed cottrells depending on
the cottrell configuration. We have one cottrell, of a commer-
cial design, which has a power supply for each field.
In the ASARCO plan, the overhead power distribution system
is arranged so that any transformer can be connected to any
electrical field in the cottrell. The normal arrangement is four
fields connected to each transformer with no two fields in series
on the same transformer. Under these conditions, a short in one
386
-------
field still leaves three fields operating in each of our adjacent
sections. When one transformer becomes inoperative, the four
connected fields are individually connected to four other trans-
formers. These four transformers are then temporarily operated
with five connected fields.
Our reason for connecting a minimum of three fields per
transformer is dictated by the method used in cleaning the wires
and plates in our cottrell. We clean by interrupting gas flow
with dampers and disconnecting power to each of the fields in
series with pneumatically operated switches. Disconnecting all
units on the transformer rectifier (TR) sets could result in
damage to the electrical equipment during a cleaning cycle if
only two units were connected to a TR set and one of these units
was previously disconnected for maintenance.
Vibrators
The early cottrells installed in our company employed
specially-designed Chicago Pneumatic or Sullivan Company air
vibrators to clean the plates and wires. These original vibra-
tors were mounted above the floor with normally four vibrators
being mounted on the plate frame and one on the wire frame. The
wire frame vibrator incorporated a spring in its design to re-
tract the pounding bar to prevent shorting after the cleaning
cycle was complete.
The cost of the originally designed vibrators became too
expensive since they were no longer a standard production item.
Consequently, in 1968 we built our first cottrell with standard
vibrators of a different design mounted under the floor and
bolted directly to the support frames. In this instance, two
vibrators were installed on the grounded frame and one on the
wire frame. Air to the vibrators was piped through the floor
with flexible tubing. This first cottrell with under the floor
vibrators operate at 250-300°F (121-149°C) and the vibrators
function quite well.
Our next cottrell was installed on a 600°F (315°C) gas
stream. Due to the elevated temperature the vibrators had to be
mounted above the floor. The discharge electrode frame vibrators
were mounted on tapered ceramic shafts to provide electrical
insulation. A small box was placed above the floor around each
ceramic shaft and a thin plastic membrane was installed across
the bottom of the box and around the shaft to provide a gas free
environment for the upper portion of the shaft. The need for
providing a gas free environment was dictated by the high acid
dew point of the gas stream which could result in acid condensing
on the ceramic shaft, thus causing current tracking to ground
through the conductive film of condensed acid.
387
-------
Maintenance
Wire electrodes are the major source of daily routine
maintenance in many of our cottrells. Two of the cottrells
which use both acid and water for conditioning experience 2 to 4
broken wires per day, as an average. In other cottrells, 30 to
50 broken wires per year are a common occurrence. Many of the
broken wires are caused by arcing, particularly in those
cottrells which use either no conditioning or water only for
conditioning. At plants that have acid conditioning, corrosion
is a major cause for wire failures.
A major overhaul of an electrical field is quite often per-
formed by plant personnel with the crew consisting of an
experienced cottrell operator, plus two laborers. One electrical
field can be completely rebuilt in ten operating shifts with this
three man crew. A complete overhaul consists of removing all the
wire electrodes, the electrode support frame, and all the plates.
Individual corrugated plates are either 24" or 31" wide and 12'
long and are light enough for one man to handle although two men
generally are needed in lining up the plates during installation.
The only shop personnel needed is a welder to tack weld the
sheets to the top support.
The frequency of a major overhaul in our smelter cottrells
is quite variable ranging from four years to over twenty years,
depending primarily on acid condition and operating temperature
of each cottrell.
The cottrells which discharge to a stack can undergo major
repairs with no upset in the operation. During an overhaul of
one electrical field, the practice is to close both inlet and
outlet dampers in that section and remove all covers over the
field to give an unobstructed access to plates and wires.
Repair work is generally conducted on day shift only. In some
instances, temporary covers are installed on the unit under
repair so the other three fields can be placed back in operation
during afternoon and night shift. Those cottrells which are on
chemical plant gas streams cannot be completely overhauled unless
the chemical plant is down since air leakage past the damper
dilutes the SO2 content of the gas stream. In the context of
this paper, the term chemical plants refers to our liquid S02
plant or to our seven sulfuric acid plants,
Labor
Labor required to operate our cottrells range from two men
per shift, plus a dust puller on day shift, to some cottrells on
chemical plants where the chemical plant operators are assigned
the cottrells as a part-time job along with their regular job.
388
-------
There are eight chemical plants in the company, six chemical
plant cottrells have men who monitor the cottrell as part of
their other duties, one plant has two men on day shift only, and
one plant has operators on all three shifts. Two cottrells on
reverberatory or roaster-reverberatory gas have two men on each
shift because both acid and water are used for conditioning,
while a third reverberatory-roaster cottrell operates with two
full-time operators on day shift since water only is used for
conditioning and water conditioning is automated in all our
cottrells.
The operator's job varies to some degree among plants; but
generally, they include removing shorts, cleaning insulator boxes
and applying hydrated lime on the support insulators, taking
power readings, cleaning water sprays since acid plant scrubber
water is used for cooling gas ahead of the cottrells, adjusting
the acid addition, adding hydrated lime to the gas stream for
conditioning, etc. Two of our cottrells do not have an auto-
matic cleaning system and for these two units, the operators
manually close dampers, disconnect power, and activate air
hammers.
Operation
All of our cottrells are monitored by either cottrell
operators, chemical plant operators, or by supervisory personnel
inspecting the cottrell at frequent intervals.
A broken wire electrode is indicated by erratic fluctuations
in the voltmeter and ammeter connected to the transformer which
supplies power to the field which has a short. These meter
anomalies can be in the form of wide swings in the readings, as
is the case of an oscillating wire, or very low but steady volt-
meter reading if there is a dead short. Most of our older
cottrells have, in addition to meters, a light connected to the
primary circuit, which glows bright during high power input, dims
with a voltage decrease,and flickers when there is a fluctuating
short which usually indicates a broken wire. This light arrange-
ment gives the operator a visual indication of applied power and
is used in locating which field has a short. When a broken wire
is indicated in one of the fields served by a specific trans-
former, the operator disconnects power on each field one at a
time until the faulty field, as indicated by a return to normal
of the light or meter, is located.
In those cottrells discharging directly to stack, the inlet
and outlet dampers are then closed, one cover removed near the
area of arcing, the safety bar lifted, power reapplied, and the
specific point of arcing noted. Broken wires can be located
very rapidly with this method and can be replaced with a new wire
in a matter of some fifteen minutes. In those cottrells which
389
-------
are part of a chemical plant gas cleaning system, broken wires
are located and removed as described above but generally are not
replaced at that time. Air leakage through the dampers dilute
the gas stream so the time covers are open is kept to a minimum.
A low voltage and higher than normal steady milliampere
reading generally means that current is tracking to ground
through one of the four stacks of insulator bricks supporting the
discharge electrode frame. When this occurs the operators
replace the stack of insulating tile which is tracking to ground.
As regards to power-on cleaning versus power-off cleaning,
we have experimented with keeping power on the last field during
the cleaning cycles to recover dust drifting from the other
fields during a cleaning cycle. The last field was cleaned only
once each shift during these tests. EPA Method 5 sampling did
not indicate any significant difference in grain loading with
and without power on the last field during the cleaning cycle.
Dust in our precipitator is normally allowed to accumulate
over night in the hoppers and is removed on day shift. We seldom
have trouble with accumulated dust catching on fire. However,
this has happened once at our East Helena lead smelter when dust
inadvertently accumulated and spontaneous combustion ignited the
dust causing it to fuse into clinkers. Spontaneous combustion
in reverberatory and roaster cottrells is a very rare occurrence
and has never happened to the extent of damaging any of the con-
veying systems or the cottrells.
Several of our cottrells are designed with the inlet distri-
bution plenum at right angles to the cottrell sections. This
arrangement is generally dictated by the plant configuration,
and as a result of these poor inlet conditions, we balance gas
flow through the cottrell by adjusting either the inlet or the
outlet dampers. This system is monitored by the use of a mano-
meter, connected to the first and last field in each section,
which measures the pressure drop across each section. If the
manometer deflection on all sections is not uniform, dampers
are biased toward the closed position on those sections exhibi-
ting the higher manometer deflection.
All but two of our own designed cottrells have an automatic
cleaning cycle which is activated at either one hour or two-hour
intervals, depending on the type of operation. The sequence of
actions during a cleaning cycle are: close inlet and outlet
dampers, disconnect power on each field with a pneumatically
operated switch, activate air hammers for 30 to 60 seconds,
reconnect power, and open the dampers. Once the cleaning cycle
starts, all sections are cleaned one at a time.
All three of our cottrells on reverberatory or roaster-
reverberatory gas streams which discharge direct to stack are
390
-------
cooled to 225-300°F fl07-149°C) with water sprays. These
cottrells are operated at lower temperatures to condense volatile
metal oxides and to increase recovery of these compounds. Our
four cottrells on copper converter gas are part of chemical
plant gas cleaning systems, three of which cool the gas to
280-300°F (138-149°C) while the fourth cools the gas to 400°F
(204°C). Cottrells on our fluid bed zinc roasters operate at
600-700°F (315-370°C). The only cottrell operating on an up-
draft lead sinter machine at the time this paper is being
written operates at 400-450°F (204-232°C). Dry cottrells
installed on the zinc and lead sintering operations are part of
chemical plant gas cleaning systems. All our chemical plants
have a second stage gas cleaning system following the dry
cottrell and volatile metals escaping the dry cottrell are
recovered in this system. The second stage systems consist of
scrubbers and wet cottrells.
Particulate Parameters
Particulate loading into our cottrells is quite variable
depending on the particular operation. The measured inlet grain
loading for one cottrell, cleaning a combination reverberatory
and roaster gas, is 2.5 gr./SCF (5.7 gm/m^). Our Hayden copper
converter cottrell has a measured grain loading of 2.0 gr./SCF
(4.6 grn/m^). The calculated grain loading, based on tons of
dust collected and normal gas flow rate, for the lead sinter
plant cottrell is 1.5 gr./SCF (3.4 gm/rn^). For one of our zinc
plant cottrells the calculated grain loading is 0.6 gr./SCF
(1.4 gm/m^).
A series of tests for particle size distribution for a
roaster-reverberatory cottrell and for a copper converter
cottrell were conducted using a Mark III impact sampler with a
preimpactor and using Reeve Angel substrates which were not
conditioned prior to the sampling. Results of these tests are
shown on Figure I.
A series of tests were conducted on the outlet of the
Hayden roaster-reverberatory cottrell for particle size distri-
bution and the results are shown on Figure II. These particle
size samples were collected at a single point approximately 30'
downstream from the cottrell outlet.
Data collected from traverse sampling of the Hayden roaster-
reverberatory cottrell for grain loading, and the particle size
distribution data were used to calculate the cottrell efficiency
for various size fractions and these data are shown on Figure III.
391
-------
0)
-p
CD
e
<0
•H
Q
O
•H
CD
H
(0
>
•H
3
O1
H
fl
0)
rH
(d
>
-rH
3
O1
W
20
10
8
6
4
1
.8
.6
II II I I I I I I I I.I Mill
Q Combined Roaster
Reverberatory
Particulate
^ Converter
Particulate
I 1 I I lilt
J I I !
1 5 20 40 60 80 95 99
Accumulated Weight % Less Than Indicated Size
Fig. I - Particle size distribution at the inlet of
two cottrells.
20
-------
in
CD
>
O
O
0)
100
98
96
94
92
90
II
I
I
I
I
1 23 5 7 10 15
Equivalent Aerodynamic Diameter, jam
Figure III - Roaster-reverberatory cottrell
efficiency with reverberatory
gas only conditioned with water,
The particulate characteristics play a major role in deter-
mining the current flow in a cottrell. The figures below
indicate the two extremes in our operations for resistivity of
particulate collected and characteristics of the gas stream to
conduct current. Particulate collected in our roaster-
reverberatory cottrells contain copper, generally in the range of
15 to 24% of the total dust content, with 2 to 5% each of lead
and zinc. The particle size measured at the inlet of our Hayden
roaster-reverberatory cottrell is advantageous in efficient
recovery with only 12% by weight being under one micron in size.
Particulate collected in our copper converter cottrells contain
lead, generally in the range of 30% of the total dust content,
with 3 to 5% copper and 7 to 12% zinc. The particle size enter-
ing the Hayden copper converter cottrell has 60% by weight under
one micron in size.
Data for a copper converter cottrell are given in Figure IV
where current flow per square meter of plate surface is plotted
for the three electrical fields. This cottrell has four parallel
sections each of which is three electrical fields deep. Each
field is connected to its own individual TR set with the inlet
field in each section being designed as the first field of the
cottrell. Gas to the cottrell is conditioned by the addition
of water.
393
-------
CM
e
•H
ra
c
CD
Q
-P
fi
1
-P
•H
CO
0)
Q
4J
G
(I)
H
H
3
O
0.6
0.4
0.2
\\
1st
2nd
3rd
4th
Fig. V - Current flow in a roaster
reverberatory cottrell.
394
-------
Gas Conditioning
ASARCO first discovered the importance of conditioning
shortly after starting a pipe cottrell on converter gas at our
Garfield smelter in 1912. (4r5) it was discovered in the pre-
liminary experiments that dust collected on the pipes were
conductive during the finish blow due to the higher acid content
in the gas stream. During the slag blow a non-conductive dust
layer was formed and further precipitation ceased. It was found,
however, that by injecting a small amount of moisture the dust
layer was rendered conductive and the process worked satisfacto-
rily at about 194°F (90°C).
Some of our early smelters had no need for conditioning,
except for copper converters, due to the high sulfur content of
ores being smelted at this time and with natural formation of
SOs for conditioning. Later with a change in ore supply, with
lower sulfur content, the natural 803 formation decreased and
conditioning assumed an important role in our cottrell system
design. Water conditioning to humidify the gas was one of the
first steps in conditioning followed later by the use of sul-
fur ic acid.
One of the first efforts in acid conditioning occurred in
the 1920's at our Murray, Utah lead smelter where sulfuric acid
was boiled in iron pots and the acid vapor plus combustion gases
were inspirated into the flue system by flue draft.(6) Our
Tacoma smelter was also using this method for acid conditioning
during this same period. Boiling sulfuric acid limited the
quantity of acid which could be evaporated and the second step
in acid conditioning was through the use of a stoker-fired coal
box with sulfuric acid sprayed into the products of combustion.
This system allowed the use of up to 2-1/2 tons of acid per day.
This method was later refined by the use of a small brick-lined
furnace in' which sulfuric acid was sprayed into the flame of a
gas burner and then was inspirated into the flue. This proce-
dure proved to be wasteful because a portion of the acid was
broken down to S02. This situation, coupled with the fact that
the furnace was a high-maintenance item, led to the discontinu-
ance of the gas fired furnace.
Our next step in developing a system for acid conditioning,
and one which we are currently using, was to atomize acid
through sprays with high-pressure air. The acid is sprayed into
a flue system ahead of any water conditioning and where the gas
temperature is 300°F (149°C) or higher.
One of the primary deficiencies in the use of acid condi-
tioning is the lack of a method for optimizing acid content of
dust collected in the cottrell. We have experimentally tried
several systems to control acid addition but have not been
successful with any method tried and, as a result, all our
395
-------
cottrells using acid conditioning have two men around the clock
to control acid addition along with other duties.
An experiment was conducted at one of our copper converter
cottrells to determine the effectiveness of hand controlled acid
addition using the milliamp meter and spark rate in the cottrell
as a guide for acid conditioning control. Acid content of
collected converter dust samples taken each hour were used as the
criteria for judging optimum range. The empirical method used
for acid determinations involved leaching a sample of dust in
water, filtering the sample, and titrating to a methyl orange
end point with sodium hydoxide. Acid content less the 0.5% is
too low, between 0.5 to 1.5% is considered optimum, between 1.5
and 5.0% is acceptable, and over 5.0% is too high. An example
of the difficulty in acid addition control is indicated by the
results of this test. Acid control was good for 15.7% of the
time, acceptable for 39.3% of the time, and either not enough or
too much 45% of the time.
S03 conditioning was installed at one of the Tacoma cottrells
when changes in the operating conditions resulted in gas too cold
for effective evaporation of acid. The source of 803 was the
copper converter acid plant where a small side stream of gas was
taken from the acid plant converter and introduced into the flue
system. This resulted in an improvement in cottrell effective-
ness over that possible with acid addition. However, we were
faced with problems similar to those faced when using acid, no
method for automatic regulation of 803 addition.
Lime conditioning is used at two of our converter cottrells
during the copper blow when the natural formation of 803 is high-
est. The segment begins at the first part of the copper blow
and continues through the rest of the cycle. Formation of 803,
with little or no production of particulate, results in wet
plates and wires in the cottrell when operating in the 250-300°F
(121-149°C) temperature range. This condition results in a drop
in voltage and an increase in milliamps. Such an electrical
change is an indication to the operators that acid neutralization
is required. Achievement of control is realized by blowing
hydrated lime into the gas stream at a point upstream from the
cottrell. The conditioning continues until the acid is neutra-
lized to a point where the voltage and milliamp readings return
to normal range.
The cottrell installed on our Hayden copper converter gas
stream ahead of the acid plant is not ASARCO's design and was
installed with no provisions for conditioning the gas stream.
The intent was to operate at 750°F (400°C) to control resistivity
of the dust in an acceptable range. Temperature control was
seldom uniform enough for good resistivity control due to varia-
tions in converter activity and the designed cottrell efficiency
396
-------
was seldom, if ever, realized. The efficiency was not acceptable
and a cooling chamber was installed ahead of the cottrell where
both water and acid could be used to condition the gas.
Saturable core reactor controls which were installed as
original equipment on this cottrell were found to react too slowly
to rapid changes in gas conditions as converters were turned in
and out of the hoods. A thyristor control was installed on one
inlet field to compare the performance of these two different
types of controls. The thyristor controlled field and other
inlet fields with saturable reactors had essentially the same
power input. The two fields, which were still saturable core
controlled, following the thyristor control indicated 49 to 68%
increase in applied power over similar fields with the original
control on the inlet field. This indicated to us that the
thyristor control resulted in a lower particulate loading than
the saturable core reactor could achieve. As a result of this
test, thyristor controls were installed on all twelve fields.
Several emission tests were conducted on the outlet of one
of the four sections using a ceramic in-stack filter. These
tests cover the period before and after thyristor controls were
installed and with and without water conditioning. Results of
these tests are given in Table III.
Table III
Mass Emission Data - Copper Converter Cottrell
Conditioned
Item
Gas Temperature °C
Water Content % by
Volume
Current Density
Milliamps/m2
Current Density
Milliamps/1000 ft. 2
S03 Concentration PPM
Particulate Concentra-
tion Gr./DSCF
Particulate Concentra-
tion mg/DSCM
Unconditioned
Saturable Saturable
Reactor Thyristor Reactor Thyristor
183
6.1
0.33
30.6
0.0546
125
180
3.5
0.45
41.6
209
0.0226
52
217
1.6
0.16
14.5
0.271
621
227
1.3
0.32
29,6
600
0.0545
125
Due to its high resistivity particulate from copper con-
verters is the most difficult dust to recover of any we have in
our smelters. This condition is indicated by lower current
density than we have on other operations shown in Table IV.
397
-------
Table IV
Typical Cottrell Operating Parameters
Reverber-
atory Lead Zinc
Item Converter Roaster Sintering Roaster
Gas Velocity FPS 1.5 11 3.2 2
Time in Electrical Field
Seconds 20 2.7 9.4 11.2
Gas Temperature °C 140 121-149 200 300
Water Content % by Volume 8.0 9.0 7.5 N.A.
Current Density Milliamps/
m2 0.28 0.52-0.84 0.7 0.85
Current Density Milliamps/
1000 ft.2 26 48-78 64 79
Applied Voltage KV 24 28-31 25 27
Copper converters operate as a batch process in which during
the first part of a cycle, the slag blow, most of the lead and
zinc is fumed out of the matte. During this part of the cycle
some of our smelters add acid to condition the dust even though
the gas stream contains 803. A series of samples were taken at
one of our copper plants to determine the S03 concentration
during the slag blow and during the finish blow. During the slag
blow, the SOs concentration ranged from 59 to 315 ppm with 156 ppm
as an average for the nine samples taken. Power to a cottrell is
lower during this phase, due to the dust burden, than it is
during the second phase of the cycle called a finish blow.
During a finish blow there is relatively little particulate
formed and the 803 content of the gas stream average some 1240
ppm. The range of 803 concentration for sixteen samples
collected during a finish blow when no water was used to quench
the gas temperature was from 600 to 2654 ppm. The ideal situa-
tion would be to blow one converter on slag, another on finish,
and combine the two gas streams but it is not always possible to
keep this combination of one on finish and one or more converters
on a slag blow.
There are two principle reasons for lower 803 during a slag
blow than during a copper blow in our operations. Temperature
plays a major role in the conversion of 802 to 803. Tempera-
tures measured during these tests ranged from 800°F to 950°F
(427-510°C) during a slag blow and from 800°F to 1200°F
(427-649°C) during a copper blow. The higher temperature during
a copper blow increases the SO2 to 803 conversion rate. The
second reason for greater 803 formation during a copper blow is
the higher concentration of SO2 during this phase of the opera-
tion.
398
-------
When acid plants were first installed on converter gas and
it became necessary to reduce air infiltration into the con-
verter hoods, the increased hood gas temperature and higher
grade SO2 in the gas stream promoted the formation of SOa. This
resulted in a wet, sticky dust in the cottrell. One of our
first attempts to overcome the problem of high acid dust collected
was to blow hydrated lime into the converter hood; up to five
tons per day at one plant. The cost of lime, and maintenance of
the lime feeder, were factors which made this method of control-
ling S03 concentration undesirable. The problem was alleviated,
but not completely solved, by spray cooling the converter gas to
800°F (427°C) at the hood.
We ran a series of tests at one of our copper plants to
determine the effectiveness of water cooling to control 303
formation during a slag blow with and without water cooling due
to the normal lower gas temperatures during this phase. During
a copper -blow the pounds of 803 were reduced by 50% with water
cooling the gas to 800°F (427°C) . Figure VI is a plot of con-
verter gas temperature versus Kg. per minute of 803 measured when
the converter was on a finish blow.
10
fi
•H
n
O
CO
0
_L
_L
_L
300 400 500 600 700
Temperature - °C
Fig. VI - 503 formation in a copper converter
gas stream during a finish blow.
Gas volume 40,500 scfm.
399
-------
References
1. Cameron, Frank: "Cottrell: Samaritan of Science", Review
by Dorothy S. Keesling, The Micron, September 1952.
2. Cottrell, F. G.: "The Electrical Precipitation of Suspended
Particles", J. Industrial and Eng. Chem., August 1911, 542.
3. Cottrell, F. G.: "Recent Progress in Electrical Smoke Pre-
cipitation", Engineering & Mining Journal, February 26, 1916.
4. Howard, W. H.: "Electrical Fume Precipitation at Garfield",
Salt Lake Meeting, August 1914, Transactions AIME, Vol. 49,
540-560.
5. White, H. J.: "Fifty Years of Electrostatic Precipitation",
Golden Jubilee Meeting of APCA, June 2-6, 1957.
6. Labbe, A. L.: "Acid Conditioning of Metallurgical Smoke for
Cottrell Precipitation", J. of Metals, April 1959, Trans-
actions AIME, Vol. 188.
400
-------
ELECTROSTATIC PRECIPITATORS (ESPS) IN TWO STAGES
USED FOR ARSENIC RECOVERY AT THE RONNSK&R COPPER WORKS
Kjell Porle
AB SVENSKA FLA'KTFABRIKEN
Industrial Division
S-351 87 VAXJO, SWEDEN
Bjorn Lindquist
BOLIDEN METALL AB
Research and Development
S932 00 SKELLEFTEHAMN, SWEDEN
ABSTRACT
A modern dedusting plant was placed in operation in 1975 at the
Ronnskar Copper Works in northern Sweden for the treatment of gas from
the roasters. A selective recovery of dust has been achieved through the
use of electrostatic precipitators (ESPs) in two stages. In the first
stage copper, zinc and lead are collected at temperatures exceeding
300 C. The solid arsenic is collected in the second stage after the gas
has been cooled to 120 C. The remaining impurities including gaseous
components are washed in a wet stage before the gas is transported to
the sulphuric acid plant.
This paper reviews the background - a prestudy phase, including
pilot-testing - design of the full-scale plant, and operating experiences.
BACKGROUND
The Process
The Ronnskar Copper Works treats sulphidic concentrates with complex
composition. Half of the concentrates come from Boliden's own mines in
Sweden. The other part is purchased. The concentrates contain approximately
20 - 25% copper and 1% arsenic. Lead, zinc, nickel, arsenic, selenium,
sulphur and precious metals are recovered besides copper. Roasting, the
first step in the copper process is done in four multi-hearth furnaces.
401
-------
The Old Plant
Previously, the gas from these furnaces was cooled to a temperature
3<
tion.
of about 150°C and then treated in four old ESPs for particulate collec-
The dust collected in the coolers and the ESPs contained about 9%
copper, 4% zinc, 5% lead, 20% arsenic and 0.6% mercury. Because of its
high content of arsenic and mercury this dust could not be completely
recycled to the process. Consequently about 1500 tonnes of collected
dust had to be disposed of every year, resulting in a large wastage of
useful metals.
Due to extensive corrosion in the old installation, high volumes of
air in-leakage diluted the process gases. This increasing treatment
costs in the sulphuric acid and liquid sulphur dioxide plants.
The New Concept
In 1972 Boliden and AB Svenska Flaktfabriken undertook a joint
research and development project aimed at solving these problems. The
objective was to separate the dust into two fractions, one containing
copper, lead and zinc, the other mainly arsenic and mercury. The first
fraction would be recycled to the smelter, the second would be treated
for recovery of arsenic and mercury.
Most arsenic and mercury compounds gasify at a temperature of 300 C
while the copper, zinc and lead compounds remain in a solid phase.
Filtration of the gases at or above this temperature in the first
filtering stage, would keep the recovered dust relatively free from
arsenic and mercury and allow for recycling. The gases free from copper,
zinc and lead could then be cooled to solidify the arsenic and mercury
which can then be removed in the second filtering stage. Fabric filters
were judged unsuitable for this service because the gas is rich in S09
(approx. 5%) and there is a risk of condensation. ESPs were suggested
for both high- and low- temperature service in the two stages of filtra-
tion.
A prestudy, including pilot plant tests, was undertaken in order to
investigate the operating conditions for the two ESPs.
PRESTUDY
Prevention of Glass Arsenic Formation
One point of concern when cooling gases containing As is the
formation of so called glass arsenic. This is a very hard sintered
compound which could be formed during condensation of vaporized arsenic
in the temperature range of 175 - 250°C. Glass arsenic cannot be easily
dislodged by conventional rappers and vibrators and frequent manual
clean-up is required. Furthermore dust transportation systems can be
damaged or blocked by hard pieces of glass arsenic. Temperatures within
402
-------
the above range must therefore be carefully avoided for all materials
that come into contact with the gas. This critical gas temperature range
should also be passed as quickly as possible which could be achieved by
using an evaporative cooler, where atomized water is sprayed into the
gas stream.
Possible Resistivity Problems with AsO
At the time the electrostatic precipitation properties of As 0
were largely unknown. The operating temperature of an ESP used for this
should not exceed 130 C to avoid loss of arsenic in vapour form. Accor-
ding to the International Critical Tables the.,amount of As?0 in vapour
form at 130°C can be in the order of 60 mg/Nm .
The resistivity of As-O- as measured in the laboratory at 130°C and
a water dew point of 40 C was found to be as high as 10 ohm cm. ESP
operation. with such a resistivity would certainly create severe back
corona on the collecting plates, resulting in arcing and/or reentrainment
of the dust. The above effects would reduce the efficiency of the ESPs.
The upper limit of resistivity of dust for..ef f icient precipitation
is normally in the order of magnitude 10 - 10 ohm cm, depending on
the application, as well as the electrode geometry. The extent of back
corona would be minimized by using discharge electrodes which give an
even current distribution over the entire collecting surfaces. This is
discussed in detail in (1) .
The dust resistivity can be reduced, for example, by increasing the
water content of the gas (2) . The necessary humidif ication would be
provided by the evaporative cooler. Due to the high SO., content of the
gas - resulting in a high acid dew point - a reduction of the resistivity
in the actual gas was also to be expected.
Pilot P lant Tes t ing
Two major items were to be investigated:
1. Extent of As^O- separated in a hot-side ESP.
2. Conditions under which As-O can be precipitated.
A two field pilot ESP was used for the testing. Each field had a
height of 2 m, depth of 2.4 m and a width of 0.25 m. The temperature
drop across the ESP was minimized by using external heating of the
walls. The gas to be tested was sampled at high temperature from the gas
duct after one of the roasters. The desired operating gas temperature
was obtained by cooling the gas upstream the ESP.
Different dust and gas conditions were tested by varying the type
of concentrate in the roasters, table 1.
403
-------
Table 1. RESULTS OF TESTS WITH THE PILOT ESP
Test
se-
ries
A
B
C
As in
cone.
%
0.4
5.5
12
T
oC
300
300
130
A/Q
m2/m3/s
40
70
70
Effici-
ciency
particu-
late %
85
95
98
S. g/Nm3
total
4.5
10
24*
AS2°3
0.4
1.0
23*
AS2°3 total
g/Nm
0.8
20
50*
T
A/Q
S.
in
The As?0 and dust concentrations were measured at around 180 C.
= Mean temperature in the pilot ESP
= Specific collecting area
= Inlet dust concentration (particulates) to the ESP
As 0 tot = Total As content (as As 0.) in flue gas before the ESP
Dust concentrations were measured using an in-the-duct-type sampler
with glass fibre thimbles. The sampled gas was then passed through
impinger bottles at ambient temperature where condensables were collected,
A very high As content in the concentrate was used during test
series C to achieve conditions in the pilot plant ESP to resemble those
in the cold-side ESP in the full scale plant. No hot-side collector was
used during this test series. The series C resulted in small amounts of
metal "impurities" in the collected dust, table 2.
Table 2. ANALYSIS OF DUST COLLECTED IN PILOT ESP, IN %
Test series
(As in cone . )
As
Cu
Pb
Fe
Zn
Hg
A
(0.4)
6-12
10 - 12
17 - 21
4-7
5-6
0.062
B
(5.5)
14 - 20
15 - 20
6-11
6-12
3-4
0.001- 0.03
C
(12)
66 - 71
0.5 - 1.7
0.4 - 1.1
0.5 - 2.4
0.2 - 0.5
0.02 - 0.04
404
-------
The amount of As collected in the hot-side pilot ESP compared to
total As in flue gas as a function of temperature is shown in figure 1,
o
Q_
ffi
LU
U)
in
80
60
40
20
0
111*111 i r
H>
I I I I ( 1. I I
240
260
280 300 320
TEMPERATURE,°C
Figure 1. Influence of gas temperature on percentage
of As collected in the pilot ESP
A minimum gas temperature of 300 C should be maintained in the hot-
side ESP to reduce the As_0- content of the collected dust recycled to
the smelter. On the other hand, as the dust might be of a pyrophoric
nature, the temperature in the hot-side ESP should not be higher than
necessary. The operating temperature must therefore be very carefully
controlled.
FULL-SCALE PLANT
Layout in Principle
Figure 2 gives an overall view of the ESP stages. Cyclones placed
close to the outlet of each furnace collect the coarse dust in the off
gases. This dust is returned to the furnaces. The gases from the four
furnaces are then mixed together in a mixing chamber before entering the
ESPs in order to equalize variations in the gas temperature and gas composi-
tion. The outlet gases from the mixing chamber are fed to two parallel
gas circuits each comprising one hot-side ESP, one cooling tower and one
cold-side ESP in series. These are followed by one single wet stage,
where the gas is washed in a tower and cooled indirectly in a cooler
with cold sea water before the gas can be utilized for production of
sulphuric acid and liquid S02, figure 3. The used water containing As as
arsenite is circulated back to the cooling tower between the ESPs.
405
-------
Figure 2. Overall view of the selective dust cleaning at Ronnskar Copper Works
-------
Surplus water is cleaned by using sodium sulphide in the central water
cleaning plant of the Ronnskar Works.
EVAPORATIVE COOLERS COLD-SIDE ESPs SATURATION COOLING TOWERS
TOWER
=0
pH-CONTROL
TANK FOR
COOLING WATER
CIRCULATION BUFFER WATER TO
TANK TANK CLEANING
Fig 3. Sketch of the wet stage and the water flows
Design of the ESPs
The ESPs are of heavy-duty type for industrial applications. The
gas flow is horizontal. The collecting plates hang from roof beams,
forming gas passages in which the high voltage frames are located. Each
frame is supported by four insulators. The discharge electrodes are of
the spiralized wire type and are hooked on to the frames. The spiral
shape provides a number of evenly distributed points from which corona
takes place, and the design is well-proven in high-resistivity applications.
Owing to the properties of a wire the micro- and the macro- current
distributions are excellent. Each bus section is energized by a thyristor
controlled transformer rectifier set equipped with automatic spark rate
controllers.
407
-------
Rapping is achieved by tumbling hammers mounted on rotating shafts.
An alarm device signals if a shaft stops rotating.
Tne ESPs are equipped with trough hoppers; the hopper walls are
electrically heated to a controlled temperature in order to avoid conden-
sation and prevent formation of glass arsenic. Dust storage time in the
hot-side ESPs is minimized due to the pyrophoric character of the dust.
Redler conveyors operating continuously quickly remove the collected
dust which is further transported by a drag-chain convyeyor to the
electric smelter.
The collected dust in the cold-side ESPs is transported to the arsenic
refinery in an enclosed system comprising redler conveyors and a vacuum
operated pipe line. The dust is then processed to pure arsenic oxide.
A flushing system protects the ESP insulators from condensation and
dust build-up. Air is heated to 300°C and 150°C respectively in the two
stages and is fed through the insulators into the ESPs.
The hot-side ESPs are equipped with a preheating system heating all
internals to above 300°C before any gas is allowed to enter. A 300 mm
mineral wool insulation ensures that the walls exposed to the gas during
normal operating are maintained at high temperatures.
Due to the risk of occasional acid dew point peaks the internals as
well as the casings of the cold-side ESPs are fabricated from AISI 316
stainless steel. Mild steel is used for the hot-side ESPs.
The Cooling Tower
The cooling tower is basically a standard Flakt design with downward
gas flow. Water atomizing nozzles spraying in the gas flow direction are
located just above the cylindrical part of the tower, figure 2. The
inlet is designed to provide an even and stable gas distribution all the
way through the tower. This ensures that the water is well mixed with
the gas in order to obtain complete evaporation. The evaporation time
depends mainly upon the water droplet sizes - average diameter as well
as size distribution - and gas temperatures. Each tower is designed for
a dry bottom - a complete evaporation - and for a gas flow of 20,000
Nm /h. Fine droplets are achieved by using air pressurized nozzles. Wear
of the nozzles is minimized by using low operating pressures for the
water and the air (< 6 x 10 Pascal).
To meet the special process demands the cylindrical part of the
tower consists of two concentric shells. Heated air in the annular space
between the shells maintains the inner wall at fixed temperatures to
avoid glass arsenic build-up. The upper part of the tower is kept at
300 C and the lower part at 150°C. The inner shell is fabricated of
stainless steel, mainly AISI 316. The duct between the tower and'the
cold-side ESP is provided with equipment for steam conditioning to be
used in case back corona effects are observed. The injection is automa-
tically controlled from the electrical readings of the cold-side ESP.
408
-------
Design Data
Design data for the dust collecting plant are given in table 3
Normal operation of the plant is based on equal gas flows in each circuit
However, the entire gas flow can be handled in one circuit with reduced
collecting efficiency.
Table 3. DESIGN DATA FOR THE FULL-SCALE PLANT
Number of parallel ESPs
Number of fields /ESP
Total gas flow, Nm /h
o
Gas temperature range, C
Nominal gas velocity in ESPs,
two lines in operation, m/s
3
Inlet dust concentration, g/Nm
Guaranteed outlet dust concentration,
g/Nm
two circuits in operation
one circuit in operation
Stage 1
Hot-side ESPs
2
2
40,000
300 - 350
0.60
10
0.25
0.70
Stage 2
Cold-Side ESPs
2
3
46,000
120 - 130
0.53
25
0.25
0.90
Operating Experiences
Selective collection of two separate dust fractions has been success-
ful. However, two major operating problems have been experienced since
the start-up of the plant. The first involved sticky dust at low temperatures
- the second back corona phenomena. These problems were minimized in
various ways.
Due to the sticky dust one circuit often had to be taken off-line
for cleaning. During such periods the entire gas volume had to be
cooled in just one cooling tower. This resulted in a wet bottom and dust
build-up due to overloading. The problem was remedied with the installation
of new air-pressurized nozzles. The bottom has since been dry.
Sticky dust and some corrosion in the cold-side ESP was caused by
high concentrations of sulphuric acid generated in the roasting process.
The excess of generated acid was reduced by increasing arsenic content
of the concentrates and by lowering the temperature in the roasters.
These steps eliminated problems of corrosion and sticky dust. At times,
however, dust still builds up in the cold-side ESPs, requiring that a
circuit be shut down after 3-4 months of operation. The build-up of
As 0 is removed manually using water. The collected As20_ is of good
quality and can be handled easily in the transporting system.
409
-------
In figure 4, severe back corona conditions are contrasted with
normal operating conditions in the cold-side ESPs by the current-voltage
characteristics for different fields. The resistivity of the collected
As»0_ dust has been measured at varying H-SO, contents of the dust,
figure 5. The dust samples contain about 95% As-O plus H SO, and 5%
other compounds. Efficient ESP operation requires a certain range of
resistivity levels. This means that only a limited variation of the
H?SO, contents - resulting in resistivities within the required range -
is acceptable.
60
LU
£50
o
30
20
10
0
1 I I I Tl 1
NORMAL
CONDITIONS
BACK-CORONA
CONDITIONS
A-FIELD
FFELD
J I I I I I I
0
100 200 300 400 500600700800
CURRENT DENSITY,jj,A/m2
Fig. 4 Current-voltage curves for the cold-side ESPs
41Q
-------
E
o
c{
10
10
16
>
i—
>
15
10*
1013
10
10
^ in 12
IU
10
10
10
10
10
9
8
H2S04
CONTENT
I- 5.1%
H-6.7%
EL-7.8%
EZ"-9.4%
31-13.0%
I
1
50
100
200 300 £00
TEMPERATURE °C
Fig. 5 Resistivity of As^O™ as a function of
temperature and H SO^-content of the dust
It is preferable to operate with a moderate H«SO, content in the
dust to avoid sticky dust in the event of increased SO. generation
resulting from disturbances in the roasting process. Thus back corona
conditions will still occur occasionally.
Analyses of the collected dusts are shown in table 4. These values,
as well as those given in tables 5-8, are mean values for one year and
are based on operation with the entire gas volume through one circuit.
Dust from the cold-side ESP often contains approx. 65% As, corresponding
to 85% As20_.
The relative distribution of different elements in the two stages
is shown in table 5. The total recovery of copper, zinc, lead and
arsenic has been excellent, tables 6 - 8. The recovery of arsenic
trioxide in the ESPs under normal operating conditions approximates 97 A
(for comparison, see table 7). The corresponding concentration of As^
(solids and gaseous) after the ESPs is less than 250 mg/Nm .
411
-------
Table 4. ANALYSIS OF COLLECTED DUST IN HOT-SIDE
AND COLD-SIDE ESPS
Hot-side ESP
Cold-side ESP
S
%
11.2
6.3
Fe
%
19.6
0.21
Cu
%
15.7
0.16
Pb
%
7.0
0.18
Zn
%
5.6
0.09
As
%
5.6
55.1
Hg
g/ton
75
3200
Table 5. RELATIVE DISTRIBUTION OF ELEMENTS BETWEEN
HOT-SIDE AND COLD-SIDE ESPS
Hot-side ESP,%
Cold-side ESP,
%
S
75.1
24.9
Fe
99.4
0.6
Cu
99.4
0.6
Pb
98.5
1.5
Zn
99.1
0.9
As
14.7
85.3
Hg
3.8
96.2
Table 6. RECOVERY OF ELEMENTS IN THE INSTALLATION
INCLUDING THE WET STEP
Total recovery, %
Cu
99.96
Pb
99.91
Zn
99.73
As
97.81
Hg
84.02
Table 7. RECOVERY OF ELEMENTS IN THE HOT-SIDE
AND COLD-SIDE ESPS
Recovery, %
Cu
99.91
Pb
99.63
Zn
98.71
As
88.69
"Table 8. CONCENTRATIONS IN THE GAS AFTER
THE COLD-SIDE ESP
mg/Nm
Hg
as dust
4.5
Hg
as vapour
4
As
as dust
and vapour
470
Hg
69.04
412
-------
Installation cost of the full-scale plant approximated 20 million
Swedish Crowns. Operating cost is about 4 million Swedish Crowns a year,
including labour, electricity, steam, water treatment and maintenance.
In arriving at the above figures, no consideration was given to the
value of collected metals. In view of the recovery value of the metals,
the incremental cost of the hot-side ESPs is justified.
Future Operation
Plans call for the multi-hearth furnaces to be replaced by a fluosolid
furnace within two years. The dust collection plant will remain unchanged.
SO., generation from the new furnace should be easier to control thereby
assuring stable conditions in the cold-side ESP.
References
1. Matts, S. Staubwiderstand und Wahl der Spiihelektroden. VDIBerichte
Nr.294, 1978.
2. Maartmann, S. The Effect of Gas Temperature and Dew Point on Dust
Resisitivity and thus the Collecting Efficiency of Electrostatic
Precipitators. Academic Press, New York, 1971. Proceedings of the
Second International Clean Air Congress.
413
-------
THE WSL APPROACH TO METALLURGICAL FUME EMISSIONS
P.R. Dawson
Warren Spring Laboratory
Stevenage
Hertfordshire, England
ABSTRACT
Research is in progress at Warren Spring Laboratory (WSL) to reduce
the quantity of metal lost as fume and to minimise the energy required for
its collection in pyrometallurgical operations. The paper is in three
parts:
(1) a brief discussion of the mechanisms of fume formation and growth;
(2) a description of the research into fume suppression using flux covers,
particularly for brass melting;
(3) a review of the available agglomeration techniques and WSL's work in
this field.
INTRODUCTION
Metallurgical fume emissions make a significant contribution to
pyrometallurgical processing costs for two reasons. Firstly, valuable
metal is lost and often has to be replaced by high cost virgin metal which,
in the UK, is almost certainly imported either as concentrate for
processing or as metal. Metal losses reduce production rates and hence
profitability. Secondly, fume emissions have to be controlled to protect
the environment and/or to recover valuable metals such as tin, silver and
gold. Consequently, costly gas cleaning plants have to be installed and
these are often expensive to run both in terms of energy requirements and
maintenance. For these reasons a research programme was started at WSL to
investigate methods of reducing the cost of fume emissions in
pyrometallurgical processes, particularly by suppression and agglomeration.
414
-------
MECHANISMS OF FUME FORMATION
*
An understanding of the mechanisms of fume formation is vital to the
WSL programme, particularly with respect to particle growth. A
comprehensive report covering formation mechanisms has been produced by
Ollerenshaw l and other publications by Fuchs2, Ramabhadran et al3, Hidy and
Brock and Buckle are cited here. The fuming process begins with
vaporisation of the more volatile metals such as zinc, lead and cadmium in
metal melting and refining operations. The various stages of this process
are outlined briefly in Figure 1.
MELT
I
VAPOUR (Zn.Pb.Cd. etc)
Non-reactive atmosphere
Reactive atmosphere
Superset tu ration
Supersaturation
NUCLEATION NUCLEATION GAS PHASE
(homogeneous or
heterogeneous)
I
CONDENSATION CONDENSATION COAGULATION
REACTION
1 1
COAGULATION REACTION
I
FUME COAGULATION
I
FUME
Fig. 1 Mechanisms of fume formation
In an inert atmosphere, nucleation occurs as supersaturation of the
vapour phase is reached. This can be either homogeneous or heterogeneous,
the latter occurring when foreign particles such as dust etc are present.
The nuclei so formed act as vapour 'sinks' and grow by condensation of
vapour onto their surfaces. Further growth by coagulation results from
collisions between particles/droplets. With a gradual decrease in
temperature, substantial growth can be achieved chiefly by condensation as,
for example, in metal powder production. Similar mechanisms apply to
volatile metal oxides, chlorides etc such as those of lead and antimony.
415"
-------
In reactive atmospheres two formation mechanisms are possible depending
on whether reaction occurs before or after nucleation. This in turn depends
mainly on the temperature gradient and gas composition near the melt
surface. When the vapour nucleates as metal, growth by condensation can
proceed and hence some control of particle size is possible. However, if
nucleation as oxide is the predominant mechanism growth by condensation
usually cannot occur because many oxides, particularly the refractory ones
such as zinc oxide, vaporise only at very high temperatures (>2000°C) by
disproportionation. Consequently the only significant growth mechanism is
coagulation as a result of Brownian diffusion and this effect becomes
insignificant as the partcle size reaches about 1 ym. Thus there is no
control of particle size for this mechanism.
In practice most metallurgical fumes are produced in reactive
atmospheres, the most commonly arising being zinc oxide. Almost every
sample of zinc oxide fume characterised in the laboratory has been extremely
fine with a mean particle size below 1 Mm, which suggests formation by
nucleation as oxide. However, one sample taken from a rotary kiln operation
to remove zinc and lead from steelmaking fume was found to be much coarser
as shown in Figure 2. As reducing conditions existed whilst the metal was
in the vapour phase, it is possible that nucleation occurred before
oxidation and hence growth by condensation was responsible for the observed
increased particle size.
ex brass melting
ex Zn recovery from
Steelmaking fume
1OO 6O 4O 2O 1O 6 4 2 1 O-6O-4 O-2
EQUIVALENT SPHERICAL DIAMETER,
Fig- 2 Particle size distributions for ZnO fumes
416
-------
In most practical situations control of temperature gradients and
atmosphere composition is not possible and particle formation occurs in a
narrow boundary layer close to the melt surface. Hence there is little
likelihood of particle size control at the source of generation. For this
reason attempts have been made to find alternative solutions to the fume
problem and work on suppression and agglomeration have featured prominently
in WSL's programme.
FUME SUPPRESSION
Fume emission rates are dependent on the vaporisation rate from the
melt surface and can be reduced by using a flux or slag cover. Secondary
brass production using an induction furnace was selected for laboratory
studies in this field and comprehensive reports have been published
elsewhere 6'7. Flux covers are often used in this industry to suppress zinc
oxide fume emissions.
Procedure
A 50 kW, medium frequency, coreless induction furnace was used to melt
approximately 27 kg of 60/40 brass scrap consisting of swarf, stampings,
strip and ingot. The fume generated was collected in a small bag filter
unit as shown in Figure 3.
TO IMPACTOR
SAMPLE NOZZLE
— BAGFILTER
FLUX AND DROSS
INDUCTION COIL
FUME
— COLLECTION
VESSEL
TO
AIR
Fig. 3 Schematic of Apparatus
417
-------
The melting procedure simulated typical plant practice. Scrap was
charged to the furnace in the order swarf, zinc, stampings (or strip), ingot
and flux, and melted into a molten heel of metal. A typical charge is given
in Table 1.
Table 1. TYPICAL CHARGE COMPOSITION
Heel
Swarf
Cuttings
Ingots
Zinc
Flux
11.17
4.54
5.39
5.58
0.49
0.20
40.8
16.6
19.7
20.4
1.8
0.7
27.37 100.0
The metal was heated to 1080°C and then the dross was removed before pouring
into three ingot moulds. A selection of commercial and non-proprietary
fluxes was used, the latter consisting of borax, borosilicates, pyrex and
cullet glass, and sodium carbonate.
All products were analysed chemically for copper, zinc and lead.
Drosses were rod-milled and screened at 850 ym (18 mesh), the oversize being
recycled as metal whilst the undersize was chemically analysed. Fume
particle size distributions were determined using a cascade impactor and
also a Sedigraph 5000 D analyser.
Results
The results have been described in detail elsewhere 6'7 and are
summarised briefly in Table 2. The following observations can be made:-
(1) Fume emissions were reduced by up to 40% by some fluxes but their size
distributions were very similar.
(2) Dross losses varied considerably from 1.8 to 3.5% of the charge weight,
but their metal contents were similar except when liquid fluxes (borax
based) were used, when metal entrainment was greatly reduced.
(3) The flux had no effect on the chemical composition of the metal.
418
-------
Table 2. THE EFFECT OF FLUXES IN A BRASS MELTING OPERATION
Flux
None
A
B
C
D
E
F
G
H
I
J
Borax
Borosilicate
Pyrex
Gullet
Sod. Garb.
Fume Loss,
% Charge
0.17
0.22
0.10
0.10
0.33
0.12
0.14
0.12
0.18
0.27
0.14
0.11
0.11
0.22
0.20
0.06
Dross Loss,
% Charge
5.05
2.01
2.66
2.94
2.33
2.33
3.01
2.64
2.65
2.16
2.73
1.86
2.10
3.43
3.71
2.86
Cu Content
of Dross,
%
23.0
22.2
18.5
18.9
14.8
12.9
22.1
25.8
17.2
14.7
23.1
4.5
4.7
27.7
28.9
14.2
Metal Entrained,
% Charge
1.93
0.75
0.82
0.92
0.57
0.50
1.10
1.13
0.77
0.53
1.05
0.13
0.16
1,58
1.78
0.70
Flux Classification
The fluxes used in this work are classified in Table 3.
categorised into three types as follows:
They can be
(1) Powdery, reducing fluxes containing mostly charcoal - these fluxes
formed a porous layer through which vapour escaped readily and high fume
emissions were recorded. Metal entrainment was also high.
(2) Hard crust - this group consisted of the remaining commercial fluxes.
They melted at low temperatures but re-solidified before the end of the
melt cycle due either to loss of water of crystallisation or to uptake
of zinc oxide as reported by Deev et al8. The crust so formed could not
be penetrated by zinc vapour from the melt except at weak points where
the flux met the crucible wall. Here vapour burnt and gradually built
up tubes sometimes growing as long as 5-10 cm, as shown in Figure 4.
Most of the metal loss was swarf and other unmelted scrap entrained in
the dross layer.
(3) Liquid layers - fluxes which melted and remained liquid until the end of
the melt cycle were most effective. The low viscosity compounds covered
the entire melt surface forming a most effective barrier to vapour
escape. As the pressure built up, bubbles formed and broke through the
layer oxidising rapidly as they burst. This phenomena has also been
419
-------
reported by Goloveshko . Metal entrainment was greatly reduced because
heavy metal particles sank through the liquid layer. Viscous fluxes
were less effective because they did not cover the melt but formed a
sticky mass, which entrained a high proportion of metal. Consequently
fume loadings and metal losses were very high for these compounds.
4 Development of ZnO spirals from melt surface
420
-------
Table 3. FLUX CLASSIFICATION
Category
Fume
Emission
Dross
Loss
Fluxes in
Category
1. Powdery, sometimes
reducing
2. Flux melts and
resolidifies giving
hard crust.
(a) Non-volatile
flux
(b) Volatile flux
3. Liquid layer
(a) Low viscosity
(b) High viscosity
H
M
H
L
H
H
M
L
L
H
A,D
B,C,E,F,G,H,J,K
I
Borax, BS2
Pyrex, Gullet
Approximate ranges (in
Fume
Dross
of charge weight)
L M
<2.2
0.1-0.2
2.2-3.0
H
>0.2
>3.0
The main problem with molten fluxes is removal prior to metal pouring.
Conventionally the drosses are solid and are raked off but this is impossible
with liquid drosses. They are not easily poured off either. They can be
thickened or 'killed' with sand but this creates large quantities of dross
for subsequent handling, treatment and recycle. The ideal flux should be
liquid through the melt cycle, solidifying only at the end.
Research at WSL has shown that mixtures of sodium carbonate and borax
make suitable fluxes. As Figure 5 shows, the optimum composition is roughly
50% each of borax and sodium carbonate and this is the minimum level of the
latter if a solid dross is required.
AGGLOMERATION
The use of fluxes is restricted generally to remelting processes.
Clearly they are undesirable in refining processes where impurities such as
zinc, cadmium etc are deliberately fumed off and so techniques to facilitate
fume collection have been studied, notably by agglomeration. Extensive
reviews on this subject have been presented by Dawson ' , Hegarty and
Shannon J 2 and Ananth and Shannon 13"1 .
421
-------
10
O
UJ o
2 o
? o
01
O
1OO
I iquid dross
produced
solid dross —J
produced
25
75
50
5O
75 1OO°/o Na2CO3
25 O°/o borax
0
a:
o
z «- O-8
a
tr o o-4
r*
5*0.2
_l
2 C
1(
-
-
-
n , nn
5 25 5O
DO 75 5O
i
75 IOC
25 C
1OO°/o Na2CO3
O°/o borax
ig. 5 Fume and metal losses from a brass melting
operation using borax / NO2CO3 fluxes
422
-------
Energy Requirements in Gas Cleaning
Gas cleaning is a major contributor to the capital and operating costs
of many pyrometallurgical processes. In particular the energy requirements
can be very high and can be related to the fume particle size as shown by
Teller16. As the power requirement rises exponentially with decreasing
particle size, submicron fumes present the greatest problems in terms of
collection efficiency and cost. Only high efficiency collection plants are
effective for control of such fumes and high capital and operating costs are
incurred. However, by agglomerating 0.1 ym particles to 1 pm particles, for
example, power requirements could be reduced by a factor of 6 or 7 according
to Teller a 6.
Agglomeration can also be used to uprate existing gas cleaning plants.
As environmental legislation has become more stringent, the demands on gas
cleaning plant efficiencies have increased dramatically. For wet scrubbers,
in particular, increased efficiencies can be achieved by utilising more
powerful fans or by changing venturi geometries etc but this increases the
power requirements. Moreover, electricity consumption17 increases
exponentially for efficiencies greater than 95%. Consequently a critical
economic situation can be reached when it may be cheaper in the long run to
build new plant rather than incur the increased energy costs. Agglomeration
offers a third choice.
Theoretical Considerations
Agglomeration of small particles suspended in a gas stream occurs when
they collide and coalesce. The basic coagulation equation is given by Hidy
and Brock 4 ,
dn, r- i=k-l
G, = -x\ I b., n, n, 1-n, I Z b... n. I (1)
r- i=lc-l -. i- » -i
4 I b. . n, n. - n. I b., n.
2 L j=1 iJ i J J k L i=1 lk i J
dt k
>
1-k-J
where the first term on the right hand side represents the gain to species k
by collisions between i and j particles, while the second term denotes the
loss by collisions of k species and all other particles.
Using the definition of the continuous spectrum function n(v,t), eqn (1)
can be rewritten as an integro-differential equation as follows:
v *f°
~ = -J- f b(m,n) n(m,t) n(v-m,t)dm - n(v,t) j b(m,v) n(m,t)dm (2)
3t 2 J £
The collision frequency term b(m,n) is of major significance in
agglomeration studies. In some cases coagulation is a spontaneous process
due to Brownian diffusion. It can be shown that provided particle diameter d
is much greater than X, the mean free path of surrounding gas atoms,
423
-------
b(m,n) - 4 » (Dj + DZ) (^ + r^ ..... (3)
where r^ is the radius of particle of volume ta, and r£ that of particle
of volume n. The diffusion coefficient of particle i in the gas is given by
(Kennard 18):
At N.T.P., the mean free path of a gas atom is about 0.1 vm which could be
greater than some particle diameters. In this case Williams19 gives the
following equation for collision frequency:
w ,
'
Agglomeration due to Brownian motion is inversely dependant on particle size
and consequently is less effective for larger particles. In practice ;Lt is
insignificant for particles larger than 1 ym and only of minor importance for
those larger than 0.5 ym.
Agglomeration rates can be increased by increasing the collision
frequency b(m,n). This can be done by applying external fields such as sonic
or electric/magnetic, by creating turbulence, or by increasing the number of
collision sites as in steam condensation and a fluidized bed.
Sonic Agglomeration
Collisions can occur between particles of different size or density
moving at different velocities as a result of vibrations induced by sound
waves. The smaller particles take part in the sound wave vibrations with
larger amplitude due to their small mass and they collide with larger
particles which hardly vibrate at all. The ratio between particle and gas
amplitude is given by Williams 19 as:
A
/= (6)
* J K2 + »2
Q_
where K =
For water droplets in air at NTP the ratios were 0.93 and 0.39 for 1 ym and
2 Jim particles respectively at the same frequency. This shows that quite
small differences in mass led to very different amplitudes and strong phase
differences. Hence acoustic excitation of a fine particle suspension could
increase the agglomeration rate.
An expression for coagulation rate was derived by Mednikov20 :
424
-------
•jr = - F n f-j\
at s v''
If the particle count at time t = 0 is n0, then the solution to eqn (7)
s
-Ft
n = n e
o .....
Hegarty et allz derived the following relationship for F:
F = 2irf5nLer£aA ..... (9)
A can be related to sound intensity J according to:
Hence agglomeration rate is dependent on frequency, particle number density
and size distribution, sound intensity and exposure time.
To achieve significant growth, high intensities have to be used.
Hegarty and Shannon12 related residence time and intensity to a growth ratio
t rf
t J ^ 3 In— (11)
r
o
For 5-20 fold growth, intensities of 150-165 dB would be needed with
corresponding residence times of 6-200 seconds. Hegarty and Shannon
also calculate that the energy requirement would be at least 53.6 kW per
100 m^ min~l for a 4 m chamber operating at 160 dB. Mednikov 20
estimated energy requirements between 10.7 and 26.8 kW per 100 m-* min~l
air flow. These energy levels are very high, greater than many gas cleaning
plants and severely restrict the application of sonic agglomeration.
Agglomeration in Laminar and Turbulent Flow Fields
Agglomeration can be promoted by flow fields due to the relative
velocity between particles. Smoluchowski21 has shown that in laminar flow,
the collision rate is given by:
b(m,n) = y G (rl + r^ (12)
dU
where G is the constant shear rate = -7—.
425
-------
Levich 22derived the following equation for agglomeration rate, Y, in
turbulent flow:
32 » r F n = 25 - r n
(13)
A similar equation was derived by Beal23 who included Brownian motion in
Levich 's theory:
= Q X 2/3 ^ r1/3 n ..... (14)
y o n o
where XQ is the microscale of turbulence =
The agglomeration rate is strongly dependent on particle size and
decreases rapidly for submicron particles. It can be increased by increasing
the gas velocity but this raises the energy requirements. Ananth and Shannon
calculated that 107.2 kW per 100 m3 min~l air flow might be necessary
to agglomerate 0.5 Pm particles to 2-3 urn. Even if this energy level could
be tolerated, the time required for growth would be about a minute which is
unacceptable.
Agglomeration due to Applied Electric/Magnetic Fields
Externally applied electric/magnetic fields induce dipole moments on
^^ t)
particles resulting in interactions and possible agglomeration. Fuchs
showed that the mean value for the ratio of agglomeration constants for
charged and uncharged particles, z is given by:
—2 3
- 1 E
Z =12 -
The agglomeration rate is thus strongly dependent on field intensity and
particle size. Magnetic fields are by far the stronger of the two and can be
applied to much smaller particles. However as the rate is proportional to
r3, this technique is of little value in agglomeration of submicron
particles. Ananth and Shannon 1 5 also point out that magnetic fields can only
be applied to ferromagnetic particles which limits their use considerably.
Steam Condensation
Collision frequency can be increased by injecting foreign particles or
droplets into the fume laden gas. This can be achieved by injecting steam
into the particle suspension and allowing it to condense. Schauer
collected 99% of 0.3 ym diameter dioctyl phthalate particles in a wet
scrubber after injecting steam where previously collection had been very
poor* Prakash and Murray 25claimed a tenfold increase in particle size by
using up to 10 kg steam per 100 m3 air. Steam production is expensive and
426
-------
such injection levels would be costly if waste heat is not available.
Steam injection is particularly attractive as an agglomeration technique
prior to a wet scrubber. Calvert and Jhaveri 26 have reported promising
results for a flux force/condensation (FFC) scrubber. They show that the
scrubber collection efficiency increases with increasing vapour condensation.
The FFC method becomes attractive economically when high collection
efficiencies of submicron particles are sought. These are the most difficult
and the most energy consuming operating conditions for the conventional wet
scrubber. Calvert and Jhaveri26 also suggest that this technique could be
suitable for uprating existing scrubbers, but in all cases, the cost of the
steam is a major consideration.
Fluidized Bed Agglomeration
The use of fluidized beds for fume agglomeration has not previously been
described although Pilney and Erickson27 reported that flyash particles in
the effluent air stream after fluidized bed filtration were in fact
agglomerates; Conventionally fluidized beds have been used to remove dust
and aerosol particles for some time, the earliest reported work being that of
Meissner and Mickley28 in 1949. As a collection device, it has found
applications in primary aluminium production as the Alcoa 398 process29,
and in asphalt recycling30.
The most serious problem in fluidized bed filtration is retention of the
fume once it has impacted. Usually to avoid excessive re-entrainment the bed
is changed continually and the used collection media can be either dumped,
recycled (as in Alcoa 398) or cleaned for recycle. Particles within a
fluidized bed collide or rub against their neighbours (or the walls) and
consequently it is probable that surface layers of collected fume will break
off. These new species will be agglomerates having a significantly larger
particle size than the inlet fume. Hence if the bed is not changed and
layers of fume are allowed to build up on the collector particles, it could
operate as an agglomerator and this is the basis of WSL'S fluidized bed
agglomerator.
According to Jackson 31, the principal collection mechanisms are Brownian
diffusion for particles smaller than 0.5 ym, inertial impaction for particles
larger than 2 ym, and interception in the intermediate region. Knettig and
Beeckmans 32and Jackson31 concluded that most collection (or collisions)
occurred near the distributor. Doganoglu et al33 explained the effect of gas
velocity on collection efficiency. They used a multi-orifice distributor
with low free area (1.5-6.7%) and related collection efficiency with an
impaction parameter or Stokes number, St, given by:
C p d2 U
Ma a or
St = —5 -\
9 n dp
Where Uor is the gas velocity through a distributor orifice.
427
-------
Most collection occurs in the grid jet region. With small grid jets,
orifice velocities are high and fume particles impact on the bed particles
there. High collection efficiencies result as found by Meissner and
Mickley28, Knettig and Beeckmans 32, Doganoglu et al 3 3, and Chaturvedi and
Reed 3 **. On the other hand when an open perforated plate is used as a
distributor, the orifice velocity Uor is too low to gain advantage of
inertial collection in the grid jets.
Using small diameter grid jets, small bubbles form and they coalesce
rapidly giving rapid transfer between the bubble and particulate phases. In
contrast using the more open distributors, larger bubbles are formed and
interphase transfer is slower. Bypassing is common and consequently the
collection efficiency decreases as gas velocity increases.
It is possible for impaction to dominate collection in the grid jet
region whilst diffusion dominates in the bed itself* If this is the case
then collection efficiency should increase with deeper beds. This has been
confirmed experimentally by Doganoglu et al and by Chaturvedi and Reed • .
As has been pointed out, retention of the fume on collector particles
may be the critical factor, for it is by no means certain that two such
particles will stick together after collision. If agglomeration is the aim,
however, retention is less important because re-entrainment is part of the
process. The physics of retention have not been studied closely. Claes et
al35 and Dumont et al 36have examined the retention of soot particles on 150
and 220 urn alumina particles. They found that the bed could accommodate
about 0.05% of its mass of fume before efficiency declined.
Retention can be improved considerably by incorporating a liquid binder
to render the collector 'sticky1 to dust or fume particles. Doganoglu et
al 33 coated their bed particles with a non-volatile liquid (dioctyl
phthalate) and this greatly improved retention. Pilney and Erickson27
showed that fly ash particles could be filtered in a fluidized bed when a
small amount of water was introduced in the bed, but there are practical
difficulties in the use of such collection aids.
Anderson and Silverman 37, and Jugel et al38 examined the effect of
electrostatics in fluidized beds and retention therein. Bed particles become
charged due to friction with their neighbours and the walls of the bed
(triboelectrification). Anderson and Silverman37, and Jugel et al3'8 have
shown that this improves retention but few details are given. Ciborowski and
Zakowski 39 showed that the effect decreased as air humidity increased.
Zahedi and Melcher '*0''11 have reported high collection efficiencies using
their electrofluidized bed.
Theoretically, fluidized beds incorporating low free-area distributor
plates are capable of the highest collection because the high orifice
velocities promote inertial impaction - the dominant collection mechanism. A
free area of about 3% appears reasonable from WSL's work. However, as the
collision frequency is particle size dependent, higher orifice velocities are
needed for smaller particles and the pressure drop across the distributor
plate is thus higher. For submicron particles a pressure drop of about 5 ins
428
-------
w.g. is indicated, which is roughly the same as that of a bag filter or
electrostatic precipitator but much less than that of a high efficiency
venturi scrubber*
It is difficult to convert a pressure drop to energy terms because
factors such as fan efficiency, rotor speed, type of impeller etc must be
considered. However, the increased fan speed is related to absorbed horse
power according to:
for a particular fan assuming constant air density. The proportionality is
based on a constant fan efficiency but this can also vary in practice and
might be improved using a different impeller. Hence, if used to uprate
existing plants, the fluidized bed agglomerator would be most attractive for
those incurring high operating costs, e.g. high efficiency venturi scrubbers.
A different situation exists when installing new plant. The capital
costs of bag filters and electrostatic precipitators are high, much higher
than for a fluidized bed, for example. Hence a combination of agglomerator
and low capital cost plant, such as low efficiency venturi, cyclone etc,
might also be considered on economic grounds, assuming that comparable
efficiency can be attained.
WSL has commenced work on fluidized bed agglomeration of zinc oxide
fume. So far the bed has been operated in the collection mode only using
250 pm and 500 wm ballotini particles and roughly 80% collection efficiencies
have been obtained. The next step will be continuous operation, monitoring
fume particle size and loading before and after the bed to assess retention
and agglomeration potential.
CONCLUSION
WSL's fume programme has been in existence for three years during which
time fume formation mechanisms have been reviewed extensively and two key
areas of research have emerged - suppression and agglomeration. Fluxes for
use in the secondary brass industry have been classified into three groups
and a suitable flux has been prepared to minimise metal losses in fume and
dross.
Fume agglomeration techniques have been reviewed and their potential for
reducing gas cleaning costs has been assessed. Agglomeration rate is
dependent on collision frequency and can be increased by increasing the
number concentration of particles as in steam condensation, and using a
fluidized bed. Techniques using applied fields have been discounted because
of the high energy requirements and/or long reaction times.
429
-------
NOTATION
A amplitude of displacement
a entrainment factor
b,bji collision parameters
b(m,n) collision frequency between drops or particles of volumes m and n
C Cunningham slip factor
D diffusion coefficient
d diameter
E electric field intensity
e energy dissipated per gram second
F coagulation constant
f frequency
J sound intensity
k Boltzmann's constant
ni,nj,nk number density of particles of species i,j,k
n(m,t) particle volume distribution at time t
Q constant
r radius
T absolute temperature
t time
U gas velocity
V volumetric flow rate
Greek Letters
e radius of assumed cylindrical coagulation zone
X mean free path of gas atoms
n gas viscosity
u angular frequency of sound wave = 2irf
p density
Subscripts
a aerosol or fume particle
f final
g gasesous phase
i,j,k index for particles of species, i,j, or k
L large particle
o initial
p collector particle
s small particle
REFERENCES
1. Ollerenshaw, R.J. Fundamental Processes Involved in the Formation of
Metallurgical Fume. Warren Spring Laboratory, Stevenage. Report
LR 286 (ME), 1978.
2. Fuchs, N.A. The Mechanics of Aerosols. N. York, Pergamon Press, 1964.
430
-------
3. Ramabhadran, I.E., T.W. Peterson, and J.H. Seinfeld. Dynamics of
Aerosol Coagulation and Condensation. Am. Inst. Chem. Eneng. J.
22:840-51, September 1976.
4. Hidy, G.M., and J.R. Brock. The Dynamics of Aerocolloidal Systems.
Oxford, Pergamon Press, 1970.
5. Buckle, E.R. Nucleation in the Condensation of Aerosols. Chem. and Ind.
475-8, June 1974.
6. Wilkinson, A., and P.R. Dawson. The Use of Fluxes in Reducing Metal
Losses as Fume and/or Dross in Secondary Brass Production. Warren
Spring Laboratory, Stevenage. Report LR 256 (ME), 1977, 20 p.
7. Dawson, P.R., and A. Wilkinson. The Effectiveness of Fluxes in
Secondary Brass Re-melting. Warren Spring Laboratory, Stevenage.
Report LR 259 (ME), 1977, 8 p.
8. Deev, V.I. , V.E. Sovkov, V.I. Smirnov, A.A. Ganichev, and
Yu. 0. Rahnmanov- Effect of ZnO Concentration on the Physicochemical
Properties of Fluxes. Tsvet. Met. 14:44-48, 1971.
9. Goloveshko, V.F. Wettability of Metals and Alloys by Acid Fluxes.
Sov- J. Non-ferrous Metals. 10:93-94, September 1969.
10. Dawson, P.R. The Evaluation of Agglomeration Systems as an Aid in Fine
Particle Control. Warren Spring Laboratory, Stevenage. Report
LR 249 (ME), 1977, 24 p.
11. Dawson, P.R. Agglomeration of Fume Particles - a Review. Filtration
and Separation. 535-8, November 1978.
12. Hegarty, R., and L.J. Shannon. Evaluation of Sonics for Fine Particle
Control. Midwest Res. Inst., Kansas City. Report EPA/600/2-76-001,
1976, 45 p.
13. Ananth, K.P., and L.J. Shannon. Evaluation of Thermal Agglomeration for
Fine Particle Control. Midwest Res. Inst., Kansas City- Report
EPA/600/2-76-067, 1976, 14 p.
14. Ananth, K.P., and L.J. Shannon. Evaluation of Turbulent Agglomeration
for Fine Particle Control. Midwest Res. Inst,, Kansas City. Report
EPA/600/2-76-006, 1976, 10 p.
15. Ananth, K.P., and L.J. Shannon. Evaluation of Magnetics for Fine
Particle Control. Midwest Res. Inst., Kansas City. Report
EPA/600/2-76-173, 1976, 27 p.
16. Teller, A.J. Air Pollution Control. Chem. Eng. N,Y., Deskbook Issue.
79:93-98, May 1972.
431
-------
17. Cooper, H.B.H., and W.J. Green. Energy Consumption Requirements for Air
Pollution Control Equipment at an Iron Foundry. J. Air. Pollut. Control
Assoc. 28:545-548, May 1978.
18. Kennard, E.H. Kinetic Theory of Gases. New York, McGraw Hill, 1938.
19. Williams, M.M.R. An Introduction to the Coagulation of Aerosols. J.
Inst. Nuc. Engrs. 17:83-8, 1976.
20. Mednikov, E.P. Acoustic Coagulation and Precipitation of Aerosols.
New York, Consultants Bureau, 1965.
21. Smoluchowski, M. Z. Physik. 17:557, 1916.
22. Levich, V.G. Physicochemical Hydrodynamics. Englewood Cliffs, New
Jersey, Prentice Hall, 1962.
23. Beal, S.K. Turbulent Agglomeration of Suspensions. J. Aerosol Sci.
3:113-25, March 1972.
24. Schauer, P.J. Removal of Submicron Aerosol Particles from a Moving Gas
Stream. Ind. Eng. Chem. 43:1532-1538, 1951.
25. Prakash, C.B., and F.E. Murray. Particle Conditioning by Steam
Condensation. A.I.Ch.E., Symp. Series. 71:81-8, 1975.
26. Calvert, S., and N.C. Jhaveri. Flux Force Condensation Scrubbing.
J. Air Pollut. Control Assoc., 24:946-51, October 1974.
27. Pilney, J.P., and E.E. Erickson. Fluidized bed fly ash filter.
J. Air Pollut. Control Assoc., 18:684-5, Oct 1968.
2'8. Meissner, H.P. , and H.S. Mickley. Removal of Mists and Dusts from Air
by Beds of Fluidized Solids. Ind. Eng. Chem. 41:1238-42, 1949.
29. Cook, C.C., G.R. Swany, and J.W. Colpitts. Operating Experience with
the Alcoa 398 Process for Fluoride Recovery. J. Air Pollut. Control
Assoc. 21:479-83, August 1971.
30. Zieve, P.B., K. Zahedi, J.R. Melcher, and J.R. Denton. Electro-
fluidized Bed in Filtration of Smoke Emissions from Asphaltic Pavement
Recycling Plants. Environ. Sci. Technol. 12:96-99, January 1978.
31. Jackson, M.L. Fluidized Beds for Submicron Particle Collection.
Am. Inst. Chem. Eng., Symp. Series. 70:82-87, 1974.
32. Knettig, P., and J.M. Beeckmans. Capture of Monodisperse Aerosol
Particles in a Fixed and in a Fluidized Bed. Can. J. Chem. Eng;
52:703-706, December 1974.
432
-------
33. Doganoglu, Y. , V. Jog, K.V. Thambimuthu, and R. Clift. Removal of
Fine Particulates from Gases in Fluidized Beds. Trans. Inst. Chem.
Engrs. 56:239-48, 1978.
34. Chaturvedi, R.K., and I.C. Reed. Aerosol Removal Efficiency of
Fluidized Beds. Proc. 2nd Pacific Chem. Eng. Congress (Pachec 77),
held in Denver, Colorado, 28-31 August 1977.
35. Claes, J., G. Dumont, F. Decamps, and W. Goosens. Application of a
Fluidized Bed and a Fixed Bed for the Removal of Submicron Particles
from-a Gas Stream. 1st World Filtration Congress, Paris, 1974.
36. Dumont, G., W. Goosens, A. Taeymans, and W. Balleux, Het.
Ingenieursblad. 10:3, 1973.
37. Anderson, D.M., and L. Silverman. Development of a Triboelectrified
Fluidized Bed for Aerosol Filtration. Proc. 5th Air Cleaning Conf. of
AEC.
38. Jugel, W., E.D. Reher, E. Grobler, and A. Tittman. Removal of Dust
Particles from a Contaminated Gas Stream Using a Fluidized Bed Process.
Chem. Tech. 22:403, 1970.
39. Ciborowski, J. , and L. Zakowski. Dust Removal in a Fluidized Bed.
II Electrostatic Effects in a Fluidized Bed and Their Effect on Dust
Removal Process Efficiency. Int. Chem. Eng. 17:538-48, July 1977.
40. Zahedi, K., and J.R. Melcher. Electrofluidized Beds in the Filtration
of Submicron Aerosols. J. Air Pollut. Control Assoc. 26:345-52,
April 1976.
41. Zahedi, K., and J.R. Melcher. Collection of Submicron Particles in
Bubbling Electrofluidized Beds. Ind. Eng. Chem., Fundamentals.
16:248-54, May 1977.
433
-------
ADVANCED TECHNIQUES FOR FINE PARTICULATE CONTROL
James H. Abbott
Particulate Technology Branch
U. S. Environmental Protection Agency
Industrial Environmental Research Laboratory
Research Triangle Park, North Carolina 27711
ABSTRACT
The Particulate Technology Branch of the U. S. Environmental Protection
Agency's Industrial Environmental Research Laboratory, located at Research
Triangle Park, North Carolina, has for the past 5 years had the major
responsibility for carrying out a research and development effort aimed
at evaluating existing novel devices for fine particulate control, and
discovering and bringing to a commercial feasibility stage devices based
on new collection principles or concepts or new combinations of existing
concepts. This paper presents test results of the concepts and devices
which have been developed and/or tested by EPA. Emphasis is placed on
those concepts or devices which may find application in the metallurgical
industry.
INTRODUCTION
Particulate emissions are usually collected from an industrial
smoke stack by one or more types of three "conventional" devices—fabric
filters, wet scrubbers, and electrostatic precipitators. The design of
each of these three devices has until recently remained relatively
unchanged since the turn of the century. In recent years because of the
need to collect smaller and smaller particles, meaning higher and higher
collection efficiences, many improvements have been developed for the
"conventional" devices, and a number of novel devices and/or concepts
have been proposed, developed, and/or marketed.
As the requirement to collect finer and finer particulate has been
imposed, the cost of conventional particulate control has risen. Since
many important collection mechanisms become much less effective on fine
particles (particles less than 1-3 micrometers in diameter), conventional
devices (with the exception of fabric filters) have become larger or
have required more energy to operate and thus have become more expensive.
In order to minimize the impact of these increased costs on our national
434
-------
clean air policy, the EPA has supported a research and development
effort aimed at evaluating existing novel devices and discovering and
bringing to a commercial feasibility stage devices based on new collection
principles or concepts or new combinations of existing concepts.
The Particulate Technology Branch of EPA's Industrial Environmental
Research Laboratory, located at Research Triangle Park, North Carolina,
for the past 5 years has had the major responsibility for carrying out
the R&D effort mentioned above. This R&D program has been divided into
the following program areas:
1. Characterization and Improvement of Conventional Control
Devices.
2. Development of Technology for Control of Particulate
Emissions from the Combustion of Low Sulfur Coal.
3. New Idea Identification and Evaluation (Novel Device
Evaluation).
4. New Particulate Control Technology Development (Novel
Concepts).
5. High Temperature and Pressure Particulate Control.
6. Accelerated Pilot Demonstrations.
This paper will discuss only program areas 3 and 4, above, associated
with novel device and novel concepts research and development.
NOVEL DEVICE PROGRAM
The objective of lERL-RTP's novel device program is to identify and
evaluate new technology or new combinations of well studied mechanisms
in order to achieve cost effective control of fine particulate. A novel
particulate collection device is a device or a dust collection system
based on new collection principles or on radical redesign of conventional
collectors which is available for testing as a pilot scale or full scale
unit.
In the fall of 1973 the novel device evaluation program was initiated
to identify, evaluate, and develop, where necessary, those devices or
systems which showed the most promise for high efficiency collection of
fine particulate. Contracts with four research organizations were
funded to test and evaluate novel devices.
More than 40 novel particulate collectors have been identified.
About half of the devices identified have been of sufficient interest to
justify technical evaluations. To date 14 devices have been either
field or laboratory tested:
435
-------
Braxton - Sonic Agglomerator
Lone Star Steel - Steam-Hydro Scrubber
R. P. Industries - Dynactor Scrubber
Aronetics - Two-Phase Wet Scrubber
Purity Corporation - Pentapure Impinger
Entoleter - Centrifield Scrubber
Andersen 2000 - CHEAF
Rexnord - Granular Filter Bed
Air Pollution Systems - Electrostatic Scrubber (Scrub-E)
Air Pollution Systems - Electro-Tube
Century Industrial Products - FRP-100 Low Energy Wet Scrubber
American Precision Industries - Apitron
Particulate Control Systems - Electrified Bed
Ceilcote - Ionizing Wet Scrubber
Future plans include testing four additional devices if satisfactory
test sites can be located:
Combustion Power - Dry Scrubber
United McGill - NAFCO ESP
Dart Industries - Hydro-Precipitrol Wetted Wall ESP
DuPont Company - DuPont Scrubber
The devices which have been tested by lERL-RTP's Particulate Technology
Branch have not been selected for any particular industry. Although all
were tested on sources with a high fraction of fine particulate, several
were tested on metallurgical sources.
Novel Device Tests
This section discusses some of the more successful novel device
tests with emphasis on those devices which may find application in the
non-ferrous metals industry. Figure 1 presents a comparison of the
fractional particulate penetration for most of the novel devices tested
by EPA.
Of all the devices tested, the Lone Star Steel Steam-Hydro Scrubber
gave the highest efficiency on fine particulate—99.9 percent efficiency
on a fine fume from an open hearth steel furnace. This scrubber uses a
high pressure steam nozzle surrounded by a number of water nozzles which
inject water into the steam, generating very fine high velocity water
droplets which are discharged into a venturi throat for capture of dust
particles. A schematic of the system tested is given in Figure 2.
The Steam-Hydro system provides 99 percent collection efficiency
over a wide range of particle sizes. However, intrinsic power consumption
is quite high, an order of magnitude greater than an ejector venturi
scrubber. If fuel must be purchased to make the steam, the Steam-Hydro
system is not an economic alternative to such systems as fabric filtration
or electrostatic precipitation. If sufficient waste heat is available,
the Steam-Hydro system becomes economically feasible. It should be
noted that this system creates its own draft, thus obviating the need
436
-------
_o
'+3
I
«.
z
o
cc
H
LU
2
UJ
a.
0.001
\ * % °>
\ \ \ ^
1 \ \ \
* * \
>
. JL. \
I I I I LX^ I I I I I I I I I
0.01
0.1 1.0 10
AERODYNAMIC DIAMETER, jumA
Figure 1. Penetration vs. particle size for novel devices tested by EPA.
437
-------
OUTLET SAMPLING
LOCATIONS
MIXING TUBE.
INJECTION WATER
STEAM NOZZLE
INLET
INLET DUCT
FLUE GAS FROM WASTE
HEAT BOILER. FED BY
OPEN HEARTH FURNACE
PARTICLE
ACCELERATOR.
ATOMIZER CHAMBER
ATOMIZER WATER-
INLET SAMPLING
LOCATIONS
ATOMIZER SLURRY
CYCLONES
CYCLONE
SLURRY
Figure 2. The Lone Star Steel steam-hydro air cleaning system.
438
-------
for fans, blowers, etc. The Steam-Hydro system has been installed in
open hearth furnaces, cupolas, iron ore kilns, paper mills, sinter
plants, ammonium nitrate fertilizer plants, and copper smelters. Since
the Steam-Hydro system was tested by EPA, Lone Star Steel has reported
that they have made significant reductions in power requirements without
loss of collection efficiency.
The Aronetics Scrubber is similar to the Steam-Hydro Scrubber but
instead of steam it discharges high-pressure/high-temperature (200°C)
water through a nozzle to provide small high velocity water droplets.
The Aronetics unit, tested on a ferroalloy plant, was not quite as
efficient as the Steam-Hydro unit. Collection efficiency was found to
be greater than 50 percent for particles larger than 0.5 ym aerodynamic
diameter, and greater than 99 percent for particles larger, than about 1 ym
aerodynamic diameter. A schematic of the system tested is given in
Figure 3.
The Steam-Hydro and Aronetics Scrubbers are very high energy users,
but in both EPA tests waste energy was used to generate the steam or hot
water.
In a field test on a diatomaceous earth calcining and drying process,
the Cleanable (media) High Efficiency Air Filter (CHEAP) had an overall
mass efficiency of 95 percent and maintained the efficiency down to
about 0.3 ym. A laboratory test of this device confirmed that relatively
high collection efficiencies can be maintained on submicron particulate.
A schematic of the CHEAP System tested in the field is given in Figure
4. '*
The manufacturer claims that the CHEAP system is well suited to the
collection of water-soluble submicron particulates and aerosols, including
ammonium nitrate and urea prill tower emissions, soda and borosilicate
glass furnace emissions, phosphoric acid mists, phosphorus pentoxide
fumes, emissions from inorganic chemical calciners and dryers, food
product spray dryer emissions, galvanizing fumes, and sulfuric acid
mists.
A small pilot scale laboratory test of the Air Pollution Systems
(APS) electrostatic scrubber (Scrub-E) showed this system to be equal to
a conventional yenturi scrubber with a power requirement 1-1/2 to 2-1/2
times as great. The Scrub-E is basically an electrostatic charger (or
ionizer) followed by a venturi scrubber. Figure 5 is a schematic diagram
of the pilot system. An electrode, upstream of the venturi, charges the
inlet particles, which then enter the venturi throat. The gas stream
atomizes the central water spray in the venturi throat. The charged
particles, according to APS, are then attracted and collected by the
highly polarized water molecules. The charged particles are also collected
on the walls of the ionizer section prior to the throat of the venturi.
A thin film of water is run down the inclined surfaces to keep the walls
clear and prevent high voltage arcing. The particle laden water droplets
are then collected by a cyclonic separator and sent into a settling tank
(clarifier). Although the water can be recycled to the scrubber system,
439
-------
TWO-PHASE
JET NOZZLE
HEAT EXCHANGER
MIXING
SECTION
MAKE-UP-
WATER
r.
PUMP
rln
| | |STACK
»V
SEPARATOR
WASTE WATER
TREATMENT
HOT GAS
Figure 3. Aronetics two-phase wet scrubber system.
-------
INLET SAMPLE PORTS
STRAIGHTENING
VANES
OUTLET SAMPLE
PORTS
SAMPLE PORT FOR
PARTICLE GROWTH
TESTS
TO ROTOCLONE
STACK
ROTATING
FILTER
DRUM
CYCLONIC
PRECLEANER
DUAL BLOWER
UNIT
DIATOMACEOUS
EARTH CALCINING
AND DRYING PROCESS
CHEAP UNIT
Figure 4. Andersen 2000 CHEAP system.
-------
IONIZER
SECTION
CLEAN
GAS
OUT
TO INDUCED
DRAFT FAN
ELECTRODE
WATER
TO WASH
IONIZER
WALL
VENTURI
SPRAY
CYCLONE
ENTRAPMENT >\
SEPARATOR
HIGH
VOLTAGE
POWER
SUPPLY
RECYCLE
PUMP
SLUDGE-
Figure 5. Air Pollution Systems Scrub-E.
442
-------
fresh water was used during the test program.
The APS Electro-Tube, which is similar to a wet wall electrostatic
precipitator, gave some very high efficiencies on fine particulates—as
high as 98.9 percent on 0.5 ym particles. This performance is similar
to that which can be achieved in small wet electrostatic precipitators
with the same ratio of plate area to volumetric flow rate.
The pilot scale APS Electro-Tube is basically a tube electrostatic
precipitator with a central rod electrode and wetted wall collector.
Figure 6 is a schematic diagram of the pilot system. The inlet particles
are charged in a high energy field (12 kV/cm) by a high intensity ionizer
at the base of the electrode. The charged particles then migrate to the
wetted wall in the body of the device in a field of 5-10 kV/cm. APS
indicates that an initial saturation charge on the particles, higher
than the usual 4-5 kV/cm for a conventional ESP, facilitates increased
migration in the collecting electric field.
The R.P. Industries' Dynactor Scrubber uses a proprietary nozzle
design to produce a water spray which serves as an air mover and as a
scrubber, thus cleaning and propelling the gas simultaneously. In an
EPA sponsored laboratory test, the Dynactor Scrubber performed as a well
designed scrubber but was found to have poor efficiency on fine particulate.
A schematic of the system tested is given in Figure 7.
In FY-77 a competitive procurement was issued to demonstrate at
pilot or small full scale the technical and economic feasibility for the
most promising existing novel particulate collection system for control
of fine particulate emissions from industrial sources. This competitive
procurement was won by Air Pollution Systems, Inc. (APS) for demonstration
of the Scrub-E which was tested in the laboratory under the novel device
evaluation program. A contract was funded in September 1977 with APS to
demonstrate a 280 to 570 m /min (10,000 to 20,000 cfm) Scrub-E on a fine
particulate source. A magnesium recovery furnace has been selected as
the site for this demonstration. Startup is scheduled for April 1979.
NOVEL CONCEPTS PROGRAM
The major difference between the Novel Device and Novel Concepts
program areas lies in the fact that the Device Program deals with existing
equipment which either is offered commercially or is available for pilot
scale testing, and the Concepts Program deals with the development of
equipment from untried ideas or from the results of proven research
studies which show that the development of an idea is feasible.
Particulate collection mechanisms utilized by scrubbers and fabric
filters are impaction, interception, and diffusion and by ESP's are
field and diffusion charging. This combination of mechanisms gives rise
to a minimum in efficiency at the 0.2 to 0.5 ym range for conventional
devices. Under optimum conditions, this minimum may be greater than
443
-------
AEROSOL
INLET
u
HIGH
VOLTAGE
SOURCE
ELECTRODE
OPTIONAL
SECONDARY
AND TERTIARY1
IONIZATION
ZONES
AEROSOL OUTLET
TO BLOWER
on
S
WATER
INLET
BODY
HIGH INTENSITY
-./"IONIZER SECTION
TANK
OUTLET
WATER DRAIN
Figure 6. Air Pollution Systems Electro-Tube.
444
-------
AIR INPUT, LOW VELOCITY,
AMBIENT PRESSURE
-HIGH VELOCITY,
SUB-AMBIENT PRESSURE
REACTION COLUMN-* . •.
I
TURBULENT
MIXED
FLUID
FILMS AND PARTICLES
GAS OUTPUT
~^A^^~
—
LIQUID
RESERVOIR/SEPARATOR
LIQUID LEVEL DETER-
MINING TRAP
LIQUID OUTPUT
Figure 7. Dynactor single-stage scrubber.
445
-------
90% for ESPs, and greater than 99 percent for fabric filters. However,
under other conditions (e.g., high temperature, high ash resistivity,
sticky particulate, and corrosive or explosive flue gases), new concepts
specific to a problem will have an advantage.
Most work to date has been directed toward combining electrostatic
removal mechanisms with scrubbing or filtration mechanisms. The first
area to be developed was charged droplet scrubbing, with a feasibility
studyftat M.I.T. and a pilot demonstration at TRW on a Kaiser coke
oven. Electrostatics and filtration have been studied at both Battelle
Northwest and Carnegie-Mellon: the former with bed filters; the latter
with baghouses. At least two new concepts, a ceramic membrane filter
and a magnetic fiber bed, are oriented toward cleanup of high temperature
gases (1000-2000°F/500-1100°C). Other new concepts which have been
studied are foam scrubbing and pleated cartridge filters of a novel
material.
Of the advanced electrostatic collection concepts studied, those
employing water droplets or filters have demonstrated enhanced performance
and should be considered for future applications. Electrostatic collection
with water drops shows high removal efficiencies for 0.5 ym particles
which are difficult to capture. Electrostatic collection with filters
shows the potential for operation at either lower pressure drops or
higher filtration rates. The conventional approach in using electrostatics
for abatement of airborne particle emissions is to charge the particles
and then collect them in a high voltage DC field. The particles can be
both charged and collected in a single device. Alternatives to this
approach involve precharging the particles in one device, and collecting
them in another. The collection device may use electrostatic forces or
may introduce a collecting medium such as water droplets or a filter.
Some electrostatic collection concepts are given in Table 1.
Details for these concepts are given below. In general, the entries in
Table 1 reflect the three major categories of collection mechanisms:
electric field effects, scrubbing, and filtration. For the last two,
combining electrostatic effects and conventional mechanisms enhances
performance over conventional devices of the same type. The possibility
of enhanced performance stimulated EPA's involvement in developing
advanced electrostatic collection concepts.
Electrostatic Collection with Droplets
The use of water droplets in an electrostatic collection device
will differ according to whether the drops are charged or not and whether
there is an ambient electric field. The usual configuration for a
charged droplet scrubber is "an ambient field with the drops charged
oppositely to the particles. The scrubber consists of three chambers:
(1) a corona discharge section for particle charging; (2) a spray chamber
which introduces oppositely charged droplets; and (3) a mist eliminator.
The University of Washington electrostatic scrubber is a charged droplet
scrubber tested by EPA. The scrubber, shown schematically in Figure 8,
was tested on a coal fired boiler side stream at 1655 to 1802 am /hr
446
-------
Table 1. ELECTROSTATIC COLLECTION CONCEPTS
Name
Collection by Means of
EPA Activity
Electrostatic Scrubbing
Electric field and water droplets Contract 68-02-0250, MIT
Contract 68-02-1345, TRW
Charged Droplet Scrubbing Water droplets
Grant 803278, U. of Washington
Grant 804393, U. of Washington
Electrostatic Fiber Beds
Filter fibers
Grant 801581, Battelle Northwest
Electrostatic Effects in
Fabric Filters
Filter fabric
Grant 803020, Carnegie-Mellon U.
-------
OO
GAS INLET:;:./:-.-'.-.
\<-:-:&&&-?*?',
)•.••'••••»'.ft
CORONA
(PARTICLE CHARGING)
MIST ELIMINATOR
SCRUBBER \
CHARGED
WATER
SPRAYS
(COLLECTION OF CHARGED PARTICLES
BY OPPOSITELY CHARGED WATER DROPLETS)
Figure 8. University of Washington electrostatic scrubber.
-------
and on an electric arc steel furnace side stream at 1754 to 3031 am3/hr.
In both cases the overall efficiency was greater than 96 percent;
performance of the scrubber was greatly influenced by the water to gas
ratio. At the coal fired boiler site, doubling the water to gas ratio
from 0.32 to 0.76 liters/am halved the overall penetration from 3.9 to
1.9 percent. The power consumption was approximately 13 W/(m /min) and
the residence time was 8 sec.
A second case, which does not use a precharging section as other
devices in this group do, has an ambient field and the charge on the
drops is imposed by the electric field. This device, referred to as a
charged droplet precipitator, has only one chamber: water is introduced
into its middle from nozzles at high voltage. The drops leave the
nozzles with a charge and are accelerated to the walls by the electric
field. The TRW charged droplet scrubber, a charged droplet precipitator
tested by EPA, was demonstrated on a coking oven battery exhaust. The
test unit, shown schematically in Figure 9, has a capacity of 31,000
m /hr at a gas flow rate of 1.83 m/s with a residence time of 3 sec.
The scrubber operated with a low water rate of less than 0.14 liters/am
(1 gal./lOOO acf). The total power consumption was 28-42 W/(m /min).
The demonstration testing showed that the TRW unit could be operated
with very infrequent wall wash and that the overall efficiency (approxi-
mately 94 percent) was highest under this condition.
A comparison of advanced electrostatic concepts using water droplet
collectors is given in Table 2. These results are from development
programs, as noted in the introduction, or from field tests and may have
been tested under the Novel Device Program. All concepts provide high
overall collection efficiencies and high efficiencies in removing the
0.5 ym diameter particle. It should be noted that penetrations range^
from 30 to 50 percent for 0.5 ym particles in conventional scrubbers.
Since penetrations for the 0.5 ym particle through the advanced electro-
static collection devices range from 1 to 10 percent, these concepts
have demonstrated significantly enhanced performance. It is possible to
increase the pressure drop in a venturi scrubber to decrease the penetration
of 0.5 ym particles to 10 percent. The pressure drop required can^e
estimated by Calvert's equation 5.3.6-12 in the Scrubber Handbook. The
energy requirement for a venturi scrubber to collect 90 percent at a
0.5 ym particle size is 325 W/(m /min). This is an order of magnitude
higher than the 13-80 W(m /min) reported for the same performance with
electrostatic droplet concepts.
Electrostatic Collection with Filters
The use of filters with electrostatics varies according to the
collector (i.e., fibers or a fabric). For a fiber collector, precharging
the particles enhances collection efficiency. For a fabric collector,
precharging the particles changes the nature of the deposited cake,
correspondingly lowers the pressure drop, and maybe improves collection
efficiency.
449
-------
HIGH
VOLTAGE
POWER
SUPPLY
GAS EXIT
COLLECTOR PLATES
-* XT
GAS
DISTRIBUTION
LOWER
SECTION
CHARGED
WATER HEADERS
WATER SUPPLY
SLURRY
DISCHARGE
Figure 9. TRW charged droplet scrubber.
450
-------
Table 2. COMPARISON OF ELECTROSTATIC DROPLET CONCEPTS/DEVICES
Power
Overall Requirement Water to Gas Ratio Percent Efficiency at
Unit Tested Efficiency, % w/(m /min) (I/Am) 0.5 ym 1.0 ym
U. of Washington3 98.1-99.5 13 0.76-0.77
Electrostatic Scrubber
U. of Washington13 96.4-98.6 13 2.23-2.29
Electrostatic Scrubber
APS Electrostatic Scrubber 83-97 80 1.44
TRW Charged Droplet Scrubber 93.5 28-42 0.11-013
99 97.5
95 97
96 90
90 85-95
At a coal fired boiler.
At an electric arc furnace.
-------
The first case, referred to as an electrostatic fiber bed, was
studied by Battelle Northwest. The three aerosols studied (NH.C1,
Na00, and MgO) had mass median diameters of less than 1 ym. The freshly
generated particles were drawn first through a corona charging section
and then, through the fiber bed. The fiber bed had a void fraction of
0.96, was 15 or 30 cm thick, and was composed of stainless steel,
polypropylene, or Teflon. With a 30 cm bed, collection efficiency was
greater than 95 percent for bed velocities up to 1 m/s with a pressure
drop through the clean bed of less than 1 cm WC.
The second case, precharging particles before a fabric filter or
electrostatically augmented filter, was studied by Carnegie-Mellon University.
The filter was a 10 cm ID by 30 cm long bag made of woven material such
as polypropylene without antistatic treatment. The aerosol was a silica
dust charged either by impingement against a tungsten carbide surface or
by corona discharge. The bags were pulse-jet cleaned. At an air-to-
cloth ratio of 6, the electrostatically augmented filter had a pressure
drop of 6.4 cm WC; a comparable bag without precharged particles had a
pressure drop of 16.5 cm WC. Similar results have been reported by
American Precision Industries, Inc. Testing their electrostatically
augmented filter, call the APITRON, American Precision Industries reports
a reduction in pressure drop from 10 cm WC for a conventional filter to
approximately 1 cm WC for the APITRON, maintaining the same filtration
rate on a steel furnace fume. If the pressure drop is the same for both
filter types, the APITRON may be operated at a filtration rate 4 times
greater. Since electrostatically augmented filters have demonstrated
significant reductions in pressure drop at the same filtration rate (or
significant improvements in filtration rate at the same pressure drop in
conventional filters), this concept also has proven enhanced performance
with electrostatics.
High Gradient Magnetic Collection
Another collection concept which has been developed by EPA/IERL-
RTP's Particulate Control Program uses a magnetic force as the prime
collection mechanism. A device based on this new technology, magnetic
filtration,has been developed. We call it high gradient magnetic separation
(HGMS).
Recentjg HGMS has been used successfully to purify various types of
wastewater. Conventional magnetic separation devices are generally
useful for separating relatively large particles of strongly magnetic
materials such as iron and magnetite. HGMS, by contrast, makes possible
the efficient separation of micron-size particles of weakly paramagnetic
materials from liquids at high process rates. Generalized theory indicates
that HGMS can also be used to remove particles from gas streams. Magnetic
forces can be cost effective if the control device is properly designed.
In its simplest practical form, the HGMS is a canister packed with
ferromagnetic fibers (such as a stainless steel wool) that are magnetized
by a strong external magnetic field (Figure 10). The strong magnetic
forces produced near the edges of the fibers are capable of trapping
452
-------
MAGNET
COIL
cn
u>
PARTICLE-LADEN
GAS IN
CLEAN GAS
OUT
STAINLESS
STEEL WOOL
MATRIX
Figure 10. Schematic representation of a high gradient magnetic separator.
-------
fine, weakly magnetic particles with high efficiency.
The packing density of the separation zone can be adjusted to
obtain a proper balance between conventional filtration phenomena and
the HGMS effect. Reducing the packing density lowers the pressure drop
through the separator but requires a higher magnetic field for efficient
particle collection. The overall particle collection efficiency is
theoretically a function of the applied magnetic field, of filter mesh
parameters (fiber diameter and magnetization, packing density, and
length of mesh in the direction of flow), of particle parameters (magnetic
susceptibility and diameter), and of fluid parameters (superficial
velocity and viscosity).
This concept has been proven to be feasible for the collection of
fine dusts which are weakly magnetic in nature. Laboratory pilot scale
experiments have been completed using BOF dust as a model (see Figure
11), and the construction of a portable pilot unit is now underway.
This unit is scheduled for its first field test on a gas stream from an
iron and steel sinter plant in 1979.
CONCLUSIONS
The feasibility of enhancing the collection of fine particulate by
conventional devices utilizing electrostatic techniques has been demonstrated.
In addition, devices utilizing new collection mechanisms or new combinations
of old mechanisms have shown promise and are currently being developed.
The economics of these new approaches have not yet been established.
454
-------
TO
BAGHOUSE
WIND TUNNEL
FROM
"BAGHOUSE
SAMPLE
HGMS
SAMPLE
DUST
AIR
WASTE
Figure 11. Diagram of bench-scale apparatus used in preliminary HGMS
experiments.
455
-------
REFERENCES
1. McCain, J. D. and Smith, W. B., "Lone Star Steel Steam-Hydro
Air Cleaning System Evaluation," EPA-650/2-74-028 (NTIS No. PB 232-
436/AS), April 1974.
2. McCain, J. D., "Evaluation of Aronetics Two-Phase Jet Scrubber,"
EPA-650/2-74-129 (NTIS No. PB 239-422/AS), December 1974.
3. Calvert, S., Rowan, J., and Lake, C., "Johns-Manville CHEAF
Evaluation," EPA-650/2-75-058a (NTIS No. PB 256-311/AS), July 1975.
4. Rei, M. T. and Cooper, D. W., "Laboratory Evaluation of the
Cleanable High Efficiency Air Filter (CHEAF)," EPA-600/2-76-202 (NTIS
No. PB 256-689/AS), July 1976.
5. Calvert, S., Rowan, J., Yung, S., Lake, C., and Barbarika, H.,
"A.P.S. Electrostatic Scrubber Evaluation," EPA-600/2-76-154a (NTIS No.
PB 256-335/AS), June 1976.
6. Calvert, S. , Christensen, C., and Luke, C., "A.P.S. Electro-
Tube Evaluation," EPA-600/2-76-154b (NTIS No. PB 258-824/AS), July 1976.
7. Cooper, D. W. and Anderson, D. P., "Dyuactor Scrubber Evaluation,"
EPA-650/2-74-083a (NTIS No. PB 243-365/AS), June 1975.
8. Melcher, J. R. and Sachar, K. S., "Charged Droplet Scrubbing
of Submicron Particulate," EPA-650/2-74-075 (NTIS No. PB 241-262/AS),
August 1974.
9. Pilat, M. J. et al., "Fine Particle Control with UW Electro-
static Scrubber," in Second EPA Fine Particle Scrubber Symposium, (May
1977, New Orleans, LA), EPA-600/2-77-193 (NTIS No. PB 273-828/AS), pp.
303-318, September 1977.
10. Krieve, W. F. and Bell, J. M., "Charged Droplet Scrubber for
Fine Particle Control: Pilot Demonstration," Report EPA-600/2-76-249b
(NTIS No. PB 260-474/AS), September 1976.
11. Drehmel, D. C., "Fine Particle Control Technology: Conventional
and Novel Devices," J. Air Poll. Control Assoc. 27:138 (1977).
12. Calvert, S., et al., "Wet Scrubber System Study, Volume I.
Scrubber Handbook", EPA-R2-72-118a (NTIS No. PB-213-016), August 1972.
13. Reid, D. L. and Browne, L. M., "Electrostatic Capture of Fine
Particles in Fiber Beds," EPA-600/2-76-132 (NTIS No. PB 260-590/AS),
May 1976.
14. Penney, G. W., "Electrostatic Effects in Fabric Filtration:
Vol. I. Fields, Fabrics and Particles (Annotated Data)," EPA-600/7-78-
142a (NTIS No. PB 288-576/AS), September 1978.
456
-------
15. Frederick, E. R., "Electrostatic Effects in Fabric Filtration:
Vol. II. Triboelectric Measurements and Bag Performance (Annotated
Data)," EPA-600/7-78-142b (NTIS No. PB 287-207/AS), July 1978.
16. Helfritch, D. J. and Ariman, T., "Electrostatic Filtration and
the APITRON - Design and Field Performance," in Proceedings: Symposium
on New Concepts for Fine Particle Control, EPA-600/7-78-170, pp 286-304,
August 1978.
TM
17. Water Reclamation Using SALA-HGMS Magnetic Separation Equipment
and Filtration Processes, Sala Magnetics, Inc., Cambridge, Mass. (1975)
18. Drehmel, D., and Gooding, C., "High Gradient Magnetic Particulate
Collection," AIChE Symposium Series, Volume 74, 1978.
457
-------
SYMPOSIUM ADDRESS
"And What of the Future?"
W. H. Dresher
Dean/ College of Mines
The University of Arizona
Tucson, Arizona
458
-------
AND WHAT OF THE FUTURE?
W. H. Dresher
Dean, College of Mines
The University of Arizona
Tucson, Arizona
It is a distinct honor and a privilege for me to have been
invited here to address your conference tonight. When Dr.
Oglesby invited me, I accepted with great haste thinking that
this would be one more opportunity for me to practice my favorite
pastime of complaining about what the Federal Government is doing
wrong today, when he interrupted me to say..."but the Environ-
mental Protection Agency is sponsoring the conference". My
bubble of euphoria burst immeidately when I had to agree with
what he was implying, that it would be extremely rude for an
invited guest to insult a host by throwing bricks at him. I am
here tonight, so obviously I accepted the invitation and prom-
ised to be on my best behavior. But more than that, I accepted
a challenge to cause myself to think and to bring to you tonight
a message which, in turn, hopefully will cause you to think.
Specifically, where is the road we are taking in American society
today leading us? Will the actions which we are taking truly
enable us to arrive at where we think we are going? All too
often today our actions are taken in the name of the future and
the well-being of future generations of Americans. I have to
ask myself, and tonight I will ask you to ask yourselves, "and
what of the future?" In saying this, I am reminded of a bumper
sticker I once saw which said, "Due to uncertainties today,
tomorrow is cancelled." Will our tomorrow be cancelled due to
the uncertainties of the actions we take today? Oh yes, tomor-
row will come all right, but will it hold for us the opportuni-
ties and the qualities which we expect?
I have become intrigued recently to try to satisfy myself
with an understanding of just what we are all about today. We
have come a long way as a country and as a culture and we of the
present generation have seen more human progress on many fronts
than any other generation of people. Some of us have seen the
whole range of technological development from the horseless
carriage to space travel. We have seen a virtual explosion of
communications and information systems. We have seen once fatal
and disabling diseases all but eliminated and we have seen a
459
-------
Green Revolution provide a bountiful harvest where none was
possible before. But, on the other hand, we have seen some very
disturbing things, too. We have seen the race riots of the
1950's and the 1960's. We have seen some of our most capable
leaders (including a president) gunned down in cold blood. We
have seen urban problems come to a head that no amount of public
money has been able to solve. More in keeping with the theme of
this conference, we have seen an environmental concern grow to a
magnitude that would have been totally unbelievable twenty years
ago if a modern day Nostradamus had forecasted the future. To
top it all off, we have seen the development of an energy crisis
which promises to complicate any goals which we might set for the
future.
There is no doubt that we are in a period of change, but it
is difficult to understand the nature of the change which is upon
us. In a generation which should be, and probably is, better off
than any previous generation in history, we are beset with prob-
lems and faults. As Pogo has often been quoted as saying,
"...we have met the enemy and they are us." Perhaps it is
because of our prosperity that we have the time and the luxury
to find fault with ourselves. I would like to think that this
is so, for it gives me the ability to rationalize, and perhaps
to begin to understand, what is happening to us. If this is
possible, I believe we should be able to plan our route into the
future. If it is not possible, then we are like Dorothy in the
Wizard of Oz....we may find the yellow brick road and the Wizard
at the end of our traumatic travels, but then again we may land
in a pile of fragments and splinters in the middle of a Kansas
wheat field, our society never to rise again to its former
stature in the world.
About a decade ago, after the Korean War and the race
riots of the fifties, and the Vietnam action and the youth
rebellion of the sixties, we saw a number of people attempt to
deduce and to articulate what was happening to our society.
Alvin Toffler's "Future Shock" and Charles Reich's "Greening of
America" are two notable examples of these attempts. As Reich
put it,
"There is a revolution coming. It will not be like revo-
lutions of the past. It will originate with the individ-
ual and with culture, and it will change the political
structure only as its final act. It will not require
violence to succeed, and it cannot be successfully
resisted by violence. This is the revolution of the
new generation."
"Revolution" seems to be the key word in most people's
interpretation of what we are about. John D. Rockefeller III,
in his book, "The Second American Revolution", describes the
steps which our society is taking to evolve from "materialism"
460
-------
to "humanism". He points to the Declaration of Independence and
the Constitution of the United States as the basic credo of the
revolutionary change which he sees in progress. According to
Mr. Rockefeller,
"The Second American Revolution—the humanistic revolution
—is .one of fulfillment, designed to bring finally to
fruition in modern times not only the letter, but the full
spirit, the intent of these great documents. To grasp
this, it seems to me (Mr. Rockefeller), is essential to
understand what is going on today."
I like to think that I understand what these people are
telling me. I like to think that I might even believe in the
cause which is being espoused. However, it scares me to think
of how these concepts can be misused: how overly-dedicated
zealots can demand Utopia now, at a rate of change that we
cannot accommodate; how the unscrupulous might use this period
of amorphous behavior of our society to impose their will of yet
another society upon us. Perhaps this society will be the one
which former Secretary of the Treasury, William Simon, describes
in his book, "A Time for Truth"—one in which individual freedom
is lost once our industrial civilization collapses and "humanism"
gives way to panic, chaos, riots, starvation, and death. I share
this fear for the future with Secretary Simon.
For want of a better description, many futurists have
termed the era which we are approaching as the "Post-industrial
Society". Unfortunately, this terminology has two meanings and
unless our status is precisely described and understood, the
term, post-industrial society, can lead us to trouble. This,
I firmly believe is part of our problem today. The first mean-
ing, and in my opinion the wrong meaning, stems from the concept
that industry is bad, it is the source of all of our problems
today, it must be eliminated, and hence our society will cer-
tainly be a post-industrial society. I firmly believe that many
of the radicals and zealots which we see using the environmental
movement as their rallying point espouse this interpretation of
the post-industrial society and use their stand for environmental
protection as the vehicle to destroy the industrial structure of
our country. These are the people who are likely to lead us to
the society which Bill Simon fears. The other interpretation of
the "Post-industrial Society" can be likened to the period in our
history when we evolved from an agrarian society to an industrial
society. In this process we did not destroy our farms for we
still required food for sustenance. The evolution from one
status to the other merely meant that we no longer had to focus
a majority of our efforts on growing food. With the development
of machinery, less and less of our populace was required to
provide our food needs. In today's language, agriculture was
becoming less and less a percentage of our gross national product
as manufacturing took up a larger and larger scale. Similarly
461
-------
today, our transition is marked by the manufacturing component
declining in our Gross National Product, with the service com-
ponent expanding beyond 50 percent of our national efforts.
Machinery now allows us more leisure time and more freedom from
working, merely to survive. Machinery, if you will, gives us
the time and the resources to seek humanistic goals.
Probably the clearest explanation of the phenomena which
we are experiencing today is that proposed by Abraham Maslow
of Brandeis University in 1954 (which was far too early to have
predicted our status today). Maslow?s "Hierarchy of Human Need"
of "Theory of Self-Actualization", in my opinion, holds the key
to what we are observing today and, consequently, holds the
answer to what we must do to preserve and accomplish our goals
for the future. According to Maslow, mankind aspires to a
successive series of needs and/or desires in his quest for total
fulfillment. First come physiological needs, such as food and
shelter. Once these needs are satisfied, safety needs become
the primary focus followed by social needs, ego needs, and
finally self-fulfillment needs. Dr. Maslow proposes that a
satisfied need is no longer a motivator and therefore tends to
be forgotten; however, he also points out that a once-fulfilled
need, if removed, causes a reversion to the lower level of con-
cern. Thus, based on Dr. Maslow's theory, a post-industrial
society cannot truly exist, by its creation, the industrial
society upon which it is based, is destroyed. If it is true
that our society is moving forward to a new higher level of
human existence, then a new question arises. Will society be
wise enough to make this transition in an uninterrupted manner
or will it be caused to fall back to a lower level due to the
failure of its economic system? Many believe that an adequacy
of raw material and energy supply are at the heart of the answer
to this question. I tend to be of this persuasion. Let's look
now at some of the details of my concerns and relate these
concerns to the subject of this conference, pollution control in
the nonferrous metals industry.
In the last fifteen or so years, many Americans seem to
have concluded that industry, particularly the basic industries,
cannot or will not undertake change for social benefit without
the imposition of government authority. The net result has been
the promulgation and the proliferation of a literal niagara
of federal regulations. None of these mandates individually
has been cognizant of the relative costs and benefits to society,
or (more importantly) of the aggregate influence of the total
regulatory body. One recent study has determined that the
copper industry, and indeed most nonferrous metals industries,
is currently subject to more than 130 major regulatory mandates
incorporating several thousand individual regulations. I have
no doubt that each of these pieces of legislation, when
originally derived, was designed to correct a real or perceived
malfunction in industry's relationship to our society.
462
-------
Unfortunately, as we are now learning, the cost of the resulting
body of regulation (which, incidentally must be borne by society
sooner or later) was ill-defined at the time of passage of the
legislation. Only today are we becoming aware of the high cost
of these regulations as our inflation spirals upward. The
President's establishment of the Regulatory Council, with EPA's
own Douglas Costle as its chairman, is the first step in
attempting to take a positive step in the direction of reducing
the regulatory cost impact on our society. It would seem that
we have reached that unenviable state of requiring regulators to
regulate regulators. The NRDC vs. Schultze case has underlined
this absurdity by implying that the President, through his
Council of Economic Advisors, has no business interfering in the
regulatory activities of the agencies under his control. In
spite of this evidence of concern on the part of the President,
I cannot help but wonder if the full impact of the regulatory
morass will be brought to light by the Council. Let's look at
what I mean by this comment.
As I postulated earlier, the key to the success of our
society in reaching the goals it appears to have established
for itself has to lie in the preservation of the industrial
base which supports it. And yet, as I overview today's
happenings, I see one impediment after another being erected to
undermine the stability of our industrial base. The EPA and the
CEQ, in their recently released report, "Microeconomic Impact of
Federal Pollution Control Programs: 1978 Assessment", gives
cold, lifeless statistics. Wholesale prices will increase 0.4
percent per year, consumer prices 0.3 percent per year and
productivity will decline 0.3 percent per year—all as a result
of federal pollution control regulations. Others, such as
Chase Manhattan Bank, tell us that the cost of government regu-
lation to United States society in 1977 was $100 billion—almost
$500 for every man, woman, and child that year. What these
numbers do not tell us is where we will obtain the raw materials
and energy to operate our society if we destroy our basic
industries. Because of federal environmental, health, and
safety regulations, two things are happening: 1) we are rapidly
destroying our ability to obtain the vital fuel and nonfuel
minerals from domestic resources which are required to feed our
industry, and 2) we are rapidly destroying our ability to pro-
cess these minerals into important raw materials. In both
instances we are rapidly increasing our dependence on the
vagaries of foreign governments for the fulfillment of our
basic raw materials and energy needs. The energy scenario is
rapidly unfolding and its consequences are already becoming
evident. The nonferrous metals scenario is just beginning to
develop and its direction is already just as evident...we are
literally exporting our basic industries to foreign shores where
foreign influences can control our future!
Between 1969 and 1975 eight United States zinc smelters
463
-------
closed down in Oklahoma, Montana, West Virginia, Illinois, and
Texas. During this same decade, imports of zinc ores and con-
centrates dropped 77 percent while imports of zinc metal
increased 89 percent. The net result has been that United States
manufacturers are placing a far greater reliance on imported pro-
cessed zinc than in the past. Zinc, I will point out, is vital
to make automobiles and appliances more resistant to corrosion
in order that they may last longer—one of the major goals of our
new "conservation ethic".
For lead, the effect of regulation is not yet so obvious
as for zinc. We are less reliant on foreign sources for lead
than we were a decade ago. However, this statistic does not
reveal the problem for lead. Recently adopted OSHA standards of
50 micrograms of lead per cubic meter of air based on an 8-hour
time weighted average at the work place promise to force
secondary lead smelters out of business. OSHA has recently esti-
mated that the newly-imposed EPA and OSHA standards for the
allowed concentration of lead in the general environment and the
work place will require the expenditure of $1 billion and an
operating cost of $300 million per year. By comparison, the
annual product of the entire domestic lead industry is worth
only $800 million! As a result, NL Industries, the largest
producer of primary and secondary lead, is rapidly divesting
itself of its lead operations, leaving the business of lead
recycling to the smaller operators who can ill afford the costs
of conforming to the regulations. Lead will be required in even
greater quantities if electricity-powered vehicles are to help
to relieve our petroleum deficiency or if solar and wind energy
are to supplement fossil fuels. Lead is the only metal for which
today's needs are satisfied by more than 50 percent recycled
material. An average of 80 percent of the lead used in storage
batteries is recycled lead. Again, recycling is another of our
"conservation ethic" goals.
In copper, the capital outlay already invested in pollution
abatement facilities has moved the industry to its limit of
capital indebtedness. The long-term debt of the domestic copper
industry today is 30 to 40 percent of its net worth. An origi-
nal Federal government estimate in 1969 of a capital cost of
$33.8 million to $81.0 million to remove 98.8 percent of the
sulfur dioxide effluent from the entire nonferrous metals indus-
try has already cost the copper industry alone in excess of
$1.1 billion and 90 percent abatement has not yet been reached.
Department of Commerce estimates are that the industry will spend
$4.5 billion by 1987 in capital and operating costs and will be
forced to reduce production by 36 percent relative to 1974
production in order to meet present EPA and OSHA requirements.
According to this estimate, by 1987 our dependence on foreign
copper will increase to almost 46 percent of our needs compared
with an average of less than 10 percent over the last 20 years.
Our present smelter capacity is just below our domestic
464
-------
consumption. Newly-imposed OSHA regulations on arsenic in the
work place environment promise to close down at least one, and
possibly two, domestic copper smelters. United States reliance
on foreign sources has already increased from 6 to 8 percent of
consumption in the early 1970's to almost 20 percent in 1978.
Copper is a vital material in.the generation, transmission, and
conversion of electricity and will be vital to the substitution
of electricity for natural gas in household use. This should be
one of the major goals of the country if we are to reduce our
import dependence for fuels.
The aluminum industry has recently revealed that the
uncertainties of the nation's energy policy and the environ-
mental obstacles to expansion have caused it to look outside the
United States (especially to Brazil and Australia) to place the
major share of its future investment. The immediate projection
is for a shortage of aluminum and a drastic increase in prices.
Ironically, the industry predicts that it may not be able to meet
the needs of the automobile industry as it attempts to meet 1985
mileage standards by increasing the amount of aluminum in an
automobile. Again, a conflict with a major goal of the United
States in reducing its import dependence for petroleum.
Jordan Baruch, Assistant Secretary of Commerce for Science
and Technology, stated the problem well in a speech recently
when he said, "The whole body of regulation is primarily an ethic
to stop industry from doing things perceived as against the
public interest. Regulation is a tool for stopping industry.
Regulation is negative in concept. It's like brakes on a car—
essential but negative." It seems to^me that it is most impera-
tive that we find ways of preventing these brakes from locking.
Regulations cannot be permitted to "stop" any industry so vital
to the very existence of our society as our nonferrous metals
industry. It is clear to me from the examples I have cited that
in our haste to achieve our goals, ^we are risking the danger of
undermining the fundamental institutions on which our society
is based—our basic industry—and we are removing our ability to
meet established goals in other areas! There is a strong like-
lihood, in my opinion, that the very steps which we are taking
to preserve our quality of life could very well place in jeopardy
the very quality we are attempting to preserve. Frankly, I see
no winners if our present course of action is not changed.
Obviously, we cannot afford to cure one ill if it means creating
another.. .particularly if it carries with it the warning of
former Secretary Simon of a loss in personal liberties. We
simply cannot afford to see our industrial civilization collapse
...the consequences are far too great!
What we see happening is remarkably like the condition
described by John Kenneth Galbraith in his book, "The New Indus-
trial State". Mr. Galbraith perceives that modern industry, in
the United States and elsewhere, is a collection of inflexible,
465
-------
self-satisfied colossi which must be brought "under control" by
such strong medicine as regulation, anti-trust control, price
control and, in the extreme, nationalization. I tend not to
agree with Mr. Galbraith's views about modern industry and sub-
mit that his description is more that of industry in years gone
by than that of "new" industry. His recommendations for the
problem which he perceives are diametrically opposite that of
what I would consider to be the correct approach. Mr.
Galbraith's solution has been applied in many countries of the
world with the result that these are now second rate countries.
To this observer, the question "what can be done about what seems
to be an inevitable downfall of all things we hold dear to our
American way of life?" (That is, how can we reassure those
unalienable rights spelled out in the Declaration of Independence
...Life, Liberty, and the pursuit of Happiness...in view of the
many restraints being applied to the industrial base of our
country?)
My solution to these problems may be naively simple for
they are derived of a practical, not an idealistic, mind.
First, in my opinion, we must do everything in our power to
return the image of American industry in the eyes of our people
to the stature which it once held. A popular American concept
(espoused particularly by our news media and our politicians)
is that industry is nothing short of a "communist-like con-
spiracy" to undermine the country and to rip off the consumer.
Industry is alleged to be the true cause of the "energy crisis",
and the core of most of what is wrong with the country today.
This is not "a collection of inflexible, self-satisfied colossi"
we are speaking of, but the very backbone of our society—a
vital component from which we gain our sustenance. Without it we
cannot walk erect to meet the challenges of the future.
Second, we must modify our regulatory regime to accomodate
to the cruel facts of reality. A physician is a poor practi-
tioner indeed if in his attempt to cure the ills of his patient
he causes the death of the patient. Regulations are a vital
tool to establish expected and standard performance. However,
if misused, they can be debilitating and even lethal. The basic
industry of our country is too frail a patient to be tampered
with in this manner. Standards must be demonstrably essential—
not based on emotion and scant scientific evidence.
Third, whereever possible, regulations should be approached
from the incentive point of view rather than from the punitive
point of view. Competition and the freedom to innovately meet
challenges is what made American industry great. There is no
reason why these initiatives cannot also be used to fulfill
society's expectations of industry. For example, why don't we
apply the same medicine to the nonferrous metals industry which
is being used on the automobile industry. Appeal to the
466.
-------
•competitive nature of the industry by taxing emissions by the
pound just as "gas guzzlers" are to be taxed by their gasoline
consumption.
Fourth/ don't stifle innovation by pre-establishing not
only the standards but the methods by which the standards are
met. To do so is wasting America's most valuable resource—
Yankee ingenuity, the ability to get tough jobs done with a min-
imum amount of red tape. The goals which have been established
are tough ones to be sure but I, for one, would like to see how
clever we can be in meeting these goals.
Fifth and last, we must restore the attitude of government
to its rightful place of protecting the security and welfare of
the people while helping to maintain an environment in which all
of the instruments of society, industry included, can be free to
fulfill their rightful function with a minimum amount of inter-
ference.
"And what of the future?" you may ask...I am hopeful,
optimistic, and looking forward to the challenges it promises.
I hope that you are, too.
467
-------
CONTROL DEVICE SELECTION AND EVALUATION
(Panel Discussion and Open Forum)
Moderator: Richard L. Meek
Southern Research Institute
Birmingham, Alabama
J. H. Abbott
Environmental Protection Agency
Industrial Environmental
Research Laboraotry
Research Triangle Park, N.C.
S. Calvert
A.P.T. Incorporated
San Diego, California
E. L. Coe, Jr.
Western Precipitation Division
Joy Manufacturing Company
Los Angeles, California
F. R. Culhane
Wheelabrator-Frye Incorporated
Pittsburgh, Pennsylvania
E. S. Godsey
ASARCO Incorporated
Salt Lake City, Utah
D. B. Harris
Environmental Protection Agency
Industrial Environmental
Research Laboratory
Research Triangle Park, N.C.
S. R. Orem
Industrial Gas Cleaning Institute
Alexandria, Virginia
F. E. Templeton
Kennecott Copper Company
Salt Lake City, Utah
L. C. Tropea, Jr.
Reynolds Metals Company
Richmond, Virginia
468
-------
CONTROL DEVICE SELECTION AND EVALUATION
(Panel Discussion and Open Forum)
1. R. L. MEEK—Welcome to the panel discussion and open forum
on Control Device Selection and Evaluation. For the past
day and a half, we've heard quite a bit about control device
operation and some of the difficulties encountered. Perhaps
in some of the cases, they wished they had had a better
method of selection and in some cases a little less severe
method of evaluation. In any event, tonight is your oppor-
tunity to exchange views with what I consider to be a very
distinguished panel. We have, sitting to my left, a panel
with well over 100 years of experience in air pollution
control technology. Our panel tonight is representative of
industry, engineering concerns, and experts from the Envir-
onmental Protection Agency. Our ground rules for tonight
are for you to direct your questions to the panel members,
preferably by name, but if you wish, to the panel in gen-
eral. I will not entertain any discourses or presentations
from the floor. However, if you have any experiences that
you would like to share with the panel and with the audience,
these will be perfectly in order.
Our panelists this evening, seated from my left: James
H. Abbott. Jim is Chief of the Particulate Technology
Branch of EPA's Industrial Environmental Research Laboratory,
Research Triangle Park, North Carolina. Dr. Seymour Calvert,
President of Air Pollution Technology Incorporated, San
Diego, California. Mr. Frank R. Culhane is Vice President,
Sales and International Operations, for the Air Pollution
Control Division of Wheelabrator-Frye Incorporated,
Pittsburgh, Pennsylvania. Mr. Culhane is also President of
IGCI, the Industrial Gas Cleaning Institute. Mr. E. L. Coe,
Jr. is Manager of Advanced Technology for Western Precipita-
tion Division, Joy Manufacturing Company, Los Angeles, Cali-
fornia. Mr. Bruce Harris is a Sanitary Engineer in the
Process Measurements Branch of EPA's Industrial Environmental
Research Laboratory, Research Triangle Park, N.C. Mr. E. S.
Godsey is Chief Fume Recovery Engineer in the Central Engi-
neering Department of ASARCO, Inc., Salt Lake City, Utah. Mr.
Sidney R. Orem is Technical Director of IGCI, the Industrial
Gas Cleaning Institute, Inc., Alexandria, Virginia. Next to
last, Dr. F. E. Templeton is Division Environmental Engineer
for Kennecott Copper Corporation with the Utah Copper
469
-------
Division at Salt Lake City, Utah. And, Mr. Lawrence C.
Tropea, Jr., who is Director of Environmental Control for
Reynolds Metals Company, Richmond, Virginia. Larry is also
Chairman of the Environmental Committee of the Aluminum
Association. So we have representatives from EPA and the
three major nonferrous metals areas and for the major control
technologies.
Jim, will you start the opening remarks.
2. J. H. ABBOTT—Dick asked us to tell you a little bit about
what we do and what our background is to start things off.
I'm Chief of the Particulate Technology Branch of the Indus-
trial Environmental Research Laboratory located at Research
Triangle Park, North Carolina. There are two of these
Industrial Environmental Research Laboratories, one is
located in Cincinnati, Ohio and one in Research Triangle
Park, North Carolina. The one in Cincinnati is the labora-
tory that's sponsoring this conference. Our branch, the
Particulate Technology Branch, is responsible for the devel-
opment of particulate control technology not related to any
specific industry. It's kind of generic work oh various
devices. We look at new technologies which we call novel
concepts and novel devices, and we also do improvement and
development work on conventional devices (fabric filters,
electrostatic precipitators and scrubbers). Our R&D program
is a rather large program; our budget runs something like
6 or 7 million dollars a year.
3. SEYMOUR CALVERT—Air Pollution Technology, or A.P.T. is an
engineering company and our major competence is in fine
particle control equipment design. Wet scrubbers are our
principal specialty. We've done many EPA funded R&D pro-
grams, such as the wet scrubber system study which led
to our writing the "Scrubber Handbook".
We did several projects on flux force/condensation
scrubbing (F/C scrubbing) which means scrubbing involving
the condensation of water vapor. And we have studied the
effect of that condensing water vapor to enhance scrubber
efficiency. We have adapted existing theory, developed
mathematical models, done laboratory work, two pilot plants,
and two demonstration plants; one on a nonferrous metals
recovery furnace and the other on an iron melting cupola.
As a matter of fact, the second one is just about to start
up; the construction is being completed on that now.
We have done several projects on entrainment separators
(or mist eliminators). We have looked at the theory and
done pilot plant studies to determine the collection effi-
ciency for the drops, the reentrainment limits on the
separators, solids depositions, and methods of washing the
470
-------
mist eliminators. A.P.T. has done many field tests of large
scrubbers on a variety of sources. The primary reason for
doing these tests was because we, and other people, had
developed methematical models for predicting scrubber effi-
ciency, but we didn't have very much data to work with. In
order to validate these mathematical relationships, we had
to go out in the field and get data, and there were other
people who have done the same thing.
A.P.T. has been working on high-temperature and high-
pressure gas cleaning (by high-temperature and high-pressure,
we mean up to 2100°F and pressures up to 250 psi). These
studies are related generally to advanced energy products
or processes and bear on the need to clean up the gas before
it can be passed into a gas turbine. We worked on a basic
study of particle collection under these extreme conditions
and evaluated inertial impaction, diffusion-, and electro-
static deposition. We have done a study on granular bed
filters with and without electrostatic augmentation, and
another on what we call the PxP, or particles by particles,
dry scrubber; also with and without electrostatic augmenta-
tion. We're now working on fugitive dust control by means
of electrically charged sprays in the spray charging and
trapping (SCAT) system.
For commercial clients, we've done some projects on
such things as sulfuric acid nucleation and condensation,
especially in oil-fired power plants. We've done a study on
power plant plume opacity as related to particle size and
other particle properties. We have worked on flue gas
desulfurization scrubber performance improvement and other
projects of that type. Recently we have established APT
Equipment Inc., a subsidiary, for the sale of air pollution
control equipment.
We have made some proprietary developments on several
designs, including what we call a Triple Scrubber which is
a three-stage high efficiency scrubber which is especially
well suited to situations where there is heavy dust loading
and where the scaling tendency is severe. We've developed
a spray scrubber which is especially good when moderate
efficiency is adequate and when power requirement is
extremely important. We have developed a high efficiency
scrubber which in some ways surpasses Venturi scrubber per-
formance. We have also developed some new mist eliminator
designs for both vertical and horizontal gas flow orientation.
E. L COE, JR.—I am with the Western Precipitator Division
of Joy, a successor company to the Western Precipitation
Company, one of the old line manufacturers of dust control
equipment. I find it somewhat less comfortable than it used
to be to speak to this industry inasmuch as the type of
471
-------
equipment that we supply is now being forced down the throats
of the industry. The previous arrangement where it was sold
for an economic purpose is considerably.,more comfortable, at
least in my view. I've been with Western Precipitation
since 1963, all of that time in the R&D Department concerned
with the development of mechanical collectors, scrubbers,
fabric filters, and both wet and dry electrostatic precipi-
tators. I would say that if there's been a change or is an
ongoing change in the precipitator business, it's in the
greater rationalization of selection and application of pre-
cipitators, principally as a result of the greater use of
EDP equipment.
5. F.R. CULHANE—I am associated with Wheelabrator-Frye, a
leading manufacturer of air pollution control equipment. We
think of ourselves as one of the old guards in the business,
having manufactured our first fabric filter for the foundry
industry in 1913, and we became fully established in the non-
ferrous metals industry right after the second world war,
specializing in the filtration of gases at elevated tempera-
tures. Since 1913, after our beginning in fabric filtration,
our product line has included shaking baghousees, reverse air
baghouses, pulse fabric filters and Lurgi high-efficiency
precipitators. In addition to our control work of particu-
late emissions, we have also done significant work in control
of gaseous emissions, having pioneered the early work in dry
scrubbing using limestone to scrub gaseous fluorides. This
early work was performed by continuously seeding a bag filter
with limestone on a ferrous open hearth facility in the early
fifties. This was followed with a dry scrubbing pilot unit
at a western aluminum smelter. This work was the fore-runner
of dry scrubbing florides from aluminum reduction plants.
We are licensees of the Alcoa A398 process. In addition, we
pioneered the first gaseous control of SOa using spray drying
as the reactor in the power industry. We have also done
activated charcoal solvent recovery work. In addition, we
have done control work on what could be called gunk, not
gaseous or solid particulate emissions, but liquid aerosols
which is somewhat of a no-man's land and something that is
not talked about too much. I should also add, that in addi-
tion to manufacturing this equipment, we also manufacture'
filter media. We have three plants in this country, one in
France, one in England and a couple in Canada. Internation-
ally, we manufacture equipment either through equity holdings
or license agreements in Canada, Mexico, Brazil, England,
France, South America, Australia, India and Japan. In order
to serve the nonferrous industry, we have become involved as
engineers in the entire process.
6. E. S. GODSEY—I started working for ASARCO, known at that
time as American Smelting and Refining Company, in 1949 at
the smelter we have in El Paso, Texas, as a Metallurgist.
472
-------
The El Paso smelter is a primary lead and copper smelter and
also recovers zinc from,blast furnace slag in a flaming fur-
nace. Part of my duties as a metallurgist included the day
to day supervision of the baghouses and Cottrells in the
plant. Eventually, I was transferred to the Salt Lake
engineering office and to the Fume and Dust Recovery Section,
and in 1973 became head of the department within the Central
Engineering Department. Part of my duties include selecting
the type of device which will be used in a particulate recov-
ery system, whether it be a baghouse, Cottrell, scrubber, or
whatever, writing specifications for the equipment we buy
from vendors, evaluating the bids, selecting the equipment
and ultimately running stack type tests to determine if the
efficiency meets the guarantee or if it meets the regulations.
7. D. B. HARRIS—I've been with EPA for ten years. The last
eight I've been involved in particle measurement development
in support of the other branches in the Industrial Environ-
mental Research Laboratory, as well as the other sections of
EPA throughout the country. Special interest has been
particle size measurements as we relate to control device
evaluation, in direct support of some of Jim Abbott's work
in developing performance models and testing novel devices.
We have evaluated the performance of the measurement devices
themselves, including impactors, cyclones, and some automatic
instrumentation. For the last two years, I've been involved
as project officer for the methods developments for fugi-
tive measurements. We're working on a Fugitive Assessment
Sampling Train (FAST) for these,, that involves cyclone and
probably elutriator sizing with backup filter, as well as a
small organic module to collect vapors from fugitive sources.
8. S. R. OREM—The Industrial Gas Cleaning Institute is the
national association of manufacturers of air pollution con-
trol equipment. It was founded in 1960 with 16 charter mem-
bers, and it now has 42 members. As our chairman mentioned
earlier, Frank Culhane is in the middle of his term as
President of the association, and has previously served as
the Fiscal Officer and Vice President. In the last year,
last summer in fact, we relocated our headquarters from
Stamford, Connecticut, to Alexandria, Virginia, for what I
think must be obvious reasons. The association is divided
in several ways, perhaps one that is of more interest to
this group is by product categories. We have six divisions:
an electrostatic precipitator division, a fabric filter
division, a wet scrubber division, a flue gas desulfurization
division, a mechanical collector division, and finally, a
gaseous emission control division which embraces such pro-
ducts as flame afterburners and catalytic oxidation equip-
ment. These divisions meet twice a year, usually having an
outside speaker on some timely topic in the morning and a
business session working on publications, test procedures,
473
-------
or something of that sort in the afternoon. We also organize
and sponsor seminars on specific topics. Two recent topics
were a seminar on explosion protection in air pollution con-
trol devices. This brought a turnout of about 200. An
earlier seminar on construction problems, erection problems,
again brought a very promising turnout. The big affair of
our year is our annual meeting beginning the end of next
week in Phoenix. We have the usual standing committees to
deal with such things as preparing a budget and policing it,
government relations, industry relations, publications, etc.
We are proud of some work we have done for the EPA,
totaling, I believe, 6 contracts since 1971 which essentially
were study contracts to (1) identify a preferred control
technology for a particular problem, and (2) develop costs,
installed costs, for that preferred control equipment. One'
of the important functions that we serve our members, and for
that matter, the public, is to measure the market. We compile
quarterly reports of new orders booked in the market.
Of particular interest to our topic tonight are some
of the publications that have been prepared by task forces
over a period of some years. For example, a bulletin on
information for the preparation of bidding specifications on
precipitators. A bulletin on a bid evaluation form for
precipitators, bulletins on types of equipment and terminol-
ogy commonly used for fabric filters, for wet scrubbers, and
for mechanical collectors. Those are some instances and
although some of you, I'm sure, have used some of these
bulletins, I believe they have not been as widely exposed
in the nonferrous metals industries as they have in the
electric power, fly ash, side of the marketplace.
9. F. E. TEMPLETON—I manage a department in Kennecott's Utah
Copper Division that's responsible for all aspects of the
environmental control program of the division, including air,
water, solid wastes. We do such things as identifying prob-
lem areas, developing solutions, and monitoring emissions
for air, water, and solid waste to make sure the division is
in compliance with regulations. Kennecott just recently
completed a $280,000,000 program at its smelter to control
air emissions. This includes both particulate and sulfur
dioxide. We are proud of the program and feel it is a
successful one. We are meeting the standards, both the
particulate standards and the SO2 standards and I guess I'm
a little chagrined to read in this EPA bulletin, "Nonferrous
Technical Awareness", that we're still having problems. I'm
not sure if the problems are all technical ones; I think
there are some regulatory ones, but we think the job we have
done in selecting the course of action to take and the out-
come of it is the correct one, and we hope as time goes by,
we'll be able to persuade some of the regulatory elements
474.
-------
of EPA and the state that we're correct.
10- L. C. TROPEA, JR.--I manage the environmental control program
at Reynolds which encompasses the air, water, and solid
waste areas. In addition, I have responsibilities for the
corporate part of the fire protection program and the noise
control engineering program. Activity-wise, I am chairman
of the Environmental Committee of the Aluminum Association,
and also participate in the United Nations Environmental
Program consultive group on the aluminum industry.
*
I'm not heavily involved in the selection of equipment,
but I do get some involvement in that area. Most of my time
is spent in the environmental policy area. The major part of
my activities involves interfacing with EPA. Also, recently
I have been involved with Clean Air Act Amendments of 1977.
That really pretty well summarizes my background.
11. R. L. MEEK—Okay, we'll throw the floor open to questions,
comments, observations. Please share your experiences with
us. I ask that you not speak from your seat but use one of
the two mikes in the center aisle. Please go to one of those
two microphones. Identify yourself and your company and ask
any questions or comments.
12. ROGER W. LEE—Imperial Smelting Processes-I'd like to direct
two questions to Mr. Culhane as a supplier of equipment, and
subsequently, to Mr. Godsey as a user of equipment. Both of
these questions relate to bag filters. I'd like, if possi-
ble, an assessment of the relative roles, bear in mind
today's technology, of the shaker type, reverse air, and
pulse jet filters. That's the first question. And the
second one relates to cooling of gases before bag filters.
Have you done any recent work to evaluate, respectively,
zig-zag coolers, forced-air coolers or water-cooled ducts for
cooling gases before putting them into bag filters. Also,
have you done any work on use of high temperature filters,
for example, stainless steel or ceramic-fiber filter bags.
13. F. R. CULHANE—That was a planted question, because last
night that question was asked of one of the speakers and
after the session was over, I suggested to Mr. Lee that per-
haps he should ask the panel that question. The question of
why not a shaker-type or why not a pulse-type is a question
that I've heard many times. We have product people within
Wheelabrator-Frye that are working on different product lines,
be It the pulse jet fabric filter, the shaker baghouse, the
reverse air baghouse or the electrostatic precipitator. The
question of which collector is the best for a given emission
problem comes up frequently, and we have found that the best
answer to the question is economics. We all have personal
preferences; and with our preferences, we have prejudices,
475
-------
particularly when we are dealing with devices that we are
comfortable with, devices that we know well, and ones in which
we have a high degree of comfort, and our fears direct us
into dealing with known successes. The first step to the
answer is the proper and accurate identification of the
emission problem. You could almost make a lifetime study
when you get into the critical areas that affect the size
of equipment, such as particle size distribution, particle
chemistry, gas volume, turn down ratio, temperature, gas
chemistry, and resistivity of the dust. As soon as the
problem is well identified, then the equipment can be proper-
ly sized and the capital costs established. In addition to
that, capital cost is necessary to run an analysis on the
operating and maintenance costs, utilizing accurate cost
values and recognizing this variation of these costs. For
instance, the cost of energy is different in the northwest
than what it is in the northeast. This sort of cost analysis
will give you direction in the selection of the best kind of
equipment for a given problem. We've recently done this on
a major project just a month ago, pulse versus shaker. The
pulse collector first costs were much more attractive, but
the operating costs in time ate up the capital savings. The
particular funding of the project was such that it was
attractive not to realize the initial savings in first costs
and to minimize the operating and maintenance costs. The
shaker baghouse, in this case, was accordingly selected. I
would like to make another key comment on this question. You
have a countryman, Mr. Lee, by the name of George Bernard
Shaw, who said, "Generalizations are not worth a damn, includ-
ing this one." That is the crux of the problem and Mr.
Englebrecht said it well yesterday, that he would have a hard
time giving a specific answer to a general question. But our
best answer to your question, Mr. Lee, is that we would make
that determination on economics. On the other phase of the
question regarding gas cooling, perhaps Mr. Godsey would like
to respond.
14. E. S. GODSEY—You mentioned the different types of pulse shaker
or reverse air cleaning. ASARCO has traditionally used the
shaker type baghouse. The oldest one that I know of in the
company started in 1902 and is still operating. We have
traditionally used the shaker type and in this application we
hold the air-to-cloth ratio as low as 0.5 to 1. This is a
large baghouse built for an expanded operation which is now
reduced in size but with the same bag surface. Our normal
design ratio would be about 1.5 to 1.7 to 1 on the shaker
type. We are now getting more into the pulse type cleaning,
and we started using them primarily because of space restric-
tion. We couldn't get enough bag surface within the alloted
area with the shaker type so we went to the pulse type clean-
ing. I, personally, am completely happy with all the pulse
type applications that we have. Some are on sub-micron
476
-------
specif y^'more- ani ™ "e eXtremely Careful t0
On coolina ?r-^£- ?T °~1 air-to-cloth for the pulse type.
to 500 psi'wat^ nallY WS have started out at 300 and gone
success using air atomizing sprays. The grain loading on our
baghouses compared with what I hear in other papers is quite
low, generally about 2 grains per standard cubic foot.
Although in some of our zinc-fuming furnaces, they can go as
high as 15 grains per standard cubic foot. The higher grain
loading is on a shaker type baghouse and we always build at
least one extra compartment, as someone mentioned this morn-
ing, for the cleaning cycle and again for maintenance so that
we can take off the stream and still have adequate capacity
to ventilate the furnace that it's connected to.
15. F. R. CULHANE — I would like to briefly add the question of
gas cooling. Typically, we use several means of cooling the
gas for fabric filter applications. One of the problems in
selecting a system for a given problem is thermal control.
We have used air dilution, bleeding in outside air to cool
the gases. We have used spray towers utilizing high pressure
water to atomize the water in order to completely flash
evaporate the water to water vapor resulting in operating
evaporative cooling towers without a run-off. We have used
waste-heat boilers, air-to-air heat exchangers, radiant
coolers, such as the trombone coolers which are so character-
istic of the nonferrous industry.
With regard to your question of our work on high temper-
ature filters, commercially, the highest operating fabric
filters today employ fiberglass at 550°F. We have looked in
the laboratory at various materials that would withstand
temperatures in excess of that, such as sintered and woven
stainless material as well as ceramics. Qne of the major
disadvantages when we get into operating at high temperature
is that under expanded conditions we have more gas to filter,
so we're losing ground when we get into temperatures above
500-550 °F. In addition, we get into special housing mater-
ials and hardware because of higher temperatures of the gas.
16. CARL H. BILLlNGS—Arizona Bureau of Air Quality Control-For
the past two years Arizona has had no particulate standards.
We've stayed our particulate standards for copper smelters.
We've been carrying out a program to try and determine what
the standards should be. The data are just beginning to come
in during the last three or four months and we've been
attempting to analyze our data and it's a little difficult.
We have removals from 74.5 up to 99.9%. About half of the
installations do not meet the standards we previously haa
477
-------
which were the two EPA curves sometimes known as the San
Francisco curve and the Colorado curve which are process
weight curves. My question tonight is: Given this wide
variation of measurements, how is one supposed to evaluate
what is proper control technology? I don't care whether it's
best available technology or reasonably available technology,
but what is proper control technology for electrostatic pre-
cipitators on copper smelters? And I direct it to the whole
board. I'd like to hear a discussion of it.
17. E. L. COE—I think that's a question that has as many answers
as there are people in this room. The performance of equip-
ment on a copper smelter or on any other application depends
on not only the initial sizing of the equipment but on the
way that it's operated and maintained, and on the way it's
tested. I think that you've cited some very low figures
which may either relate to very old equipment or may very
possibly relate to equipment which was designed to be tested
by the ASME method. Very possibly, different results were
obtained when tested by other methods such as the EPA
Method 5.
I think as far as electrostatic precipitators are con-
cerned, they've been applied in the smelting industry for
years, and their performance and faults are well known. I
think that I am beating around the bush as far as your ques-
tion is concerned and I think that I have to do so because
there is not a single definite answer. As Mr. Englebrecht
pointed out fairly well, you can design a precipitator as
large as you need and it'll catch as much as you wish. Bag
filters have well-known high performance capabilities for
collection of particulate, but they also have limitations.
So does any other piece of equipment that you can think of.
If you can guide me through this maze that I am constructing
here, as to what kind of path you want followed toward your
answer, perhaps I can be of greater assistance.
18. F. R. CULHANE—You know it's really an interesting question,
because you can carry things to an extreme from the stand-
point of efficiency, 8 fields of precipitator, high SCA's,
800, 900, 1000 square feet of collecting surface per 1000 cfm.
As engineers, if somebody gives some requirements as to how
clean we must clean the gas...we can provide the means to
achieve the required results. In the case of beryllium, we
have installed baghouses in series. The first baghouse
removes 99.9% and the second one we pre-coat with a filter
aid and get another 99% out of the 0.1% that we lose in the
first unit. Is that BACT? That would scare me. I think we
really should get into BAPCT—Best Available Practical Con-
trol Technology. Perhaps that's a question best directed to
the EPA.
478
-------
range COImnents on that broad
'
20 ' vr t.ink When it: C0ines to evaluations of specific
T? H !S/ Jim Abb°tt would be able to comment moire on
comment questions on test methods, I'd be glad to
21. J. H. ABBOTT— Well, I ' n say that you haven't really told us
very much except that you've been out testing equipment and
have a broad range of results. I'm not sure what you were
looking for when you tested it, but certainly you can go in
the field and find as broad a range of efficiencies as one
wishes to find depending on how poor or how good the equip-
ment that you test happens to be.
22. C. H. BILLINGS — To begin with, you are probably all familiar
with the development of process-weight type curves and the
way that the so-called San Francisco curve was developed
whereby they did go out and test a number of industries, and
then they created an envelope curve that included all those
industries. All the industries tested therefore passed. We
attempted to do the same thing with copper smelters and of
course we haven't succeeded. We even developed our own test
method to obtain consistent results, and in fact, it did
give more reproducible results than Method 5. When we began
this study we found out that Method 5 was very inconsistent
due to the generation of acid mist in some of the ducts and
one thing and another in these copper smelters. Therefore,
we've developed a test method in which the hot box was kept
at a higher temperature and all the condensables were sup-
posedly carried through the filter and collected in an
impinger. So that all we were left with was mineral particu-
late. So, we feel that our methods of testing, and Mr.
Godsey was on the panel that supervised some of these tests,
were well done and that they were done with about as good
technology as you probably could muster. The only other
remark I could make is that I am virtually certain that 74.5
is too small an efficiency for an electrostatic precipitator
for only one reason, that if I've got to go that low with an
electrostatic precipitator, I'd rather replace it with a
multiple cyclone and save the electricity.
23. E. S. GODSEY — There is no such thing as a typical smelter.
But within our operations, we cool the gas to about 250-275 F,
particularly if we are going to stack. If we are going to an
acid plant, we have some that operate in the range of 600 F.
As you say, the volatiles go through the Cottrell. They have
no chance of being recovered unless they are as a solid
particulate. Now, if the company that has volatiles going
through is forced to cool down to improve the efficiency,
they are going to condense the acid mist; it's going to be
479
-------
recovered in the Cottrell, and you can almost stand there and
watch it melt in front of your eyes. I've done that. I've
watched it. Not day by day, but go back in two or three
months and see holes in it that weren't there before. So the
acid content is very important so far as Cottrell efficiency
is concerned, and too much, you're spending an awful lot of
time with the unit down to consider the gas stream in addi-
tion to the efficiency of the Cottrell. As I say, there is
no such thing as a typical one. Each one has enough differ-
ence in it that it's a special problem in itself. In ours,
we have some that we couldn't lower the temperature to 250
or 300°F. It would kill us. Others we have do operate in
the low temperature range, and generally speaking, on these
we get better efficiencies than we do with the ones operating
at the 600° range. We started out with a Cottrell purchased
from another company, not one of our own design, and operat-
ing it at the temperature recommended which is about 400°C.
With a copper converter operation, you can't maintain 400°C.
You can't maintain any specified temperature. The intent
was to operate it at 400°C, reduce the resisitivity of the
dust, and get what they indicated as 95% efficiency. It
turned out to be more like 85%, primarily because we couldn't
maintain the temperature. So, in our case, we changed the
flue configuration, we added water cooling, and we improved
it up into the 90%. But, again, this is one going to an acid
plant, and efficiency is not as critical as if we were going
to stack with it.
24. P. E. TEMPLETON—I'd just like to support Ed's point. It's
wrong to generalize all copper smelters, and I think the
notion of the so-called curves and the process weight limita-
tions applied across an industry is really the wrong approach
to take. Each situation is unique, and the results that
we've gotten indicate that you get a pretty broad range; with
any given control system you're going to get a broad range of
emissions. You'll get a cluster of values around an average
sort of area, but you'll get some higher values and some
lower values. If you do enough tests, you'll end up with a
distribution curve. So, if you ask what the efficiency of
the control system is, it really depends on when you measure
it. There will be a maximum and a minimum and, as I men-
tioned, an average. To try to say that all plants should be
able to achieve x percent control using a certain kind of
precipitator unit just won't work. It just isn't appropriate
to put a fixed emission limitation on all smelters of the
same size or proportionate size and try to develop a regula-
tion that way. I think what you've got to do is to have
a regulation that allows for the variations in the process
and in the units that control the process gases.
25. J. H. ABBOTT—I'm still not sure that I totally understand
the question, but I do have a comment. Regardless of how
480
-------
good your test method may be, when you go to a precipitator
at a given plant in a given situation to make test measure-
ments, before you factor the data into any kind of standard
or regulatory plan or whatever it is you're going to use it
for, I think you have to apply some good engineering judg-
ment to the data. First of all, you have to examine the
precipitator, what it was designed to do, what it is doing,
what are the inlet conditions, what are the outlet conditions,
what are the characteristics of that dust, the resistivity,
the particle size distribution. There are mathematical
modeling techniques that one can look at to try to explain
an anomaly that you might have seen, and I think in some
cases you might want to reject some of these points as not
being typical. You might want to utilize what one would con-
sider to be the typical points in whatever stragety you might
be utilizing or attempting to develop with this data. I
don't think you can just go out and blindly take a series
of points and then factor them all into something, whatever
it is, and come out with anything useful. There just has to
be something more to it than that.
26. F. E. TEMPLETON—We've had EPA come out and make measurements
on our plant and there have been times when they've turned
off the measuring equipment because they didn't feel enough
units were running, or it wasn't typical, and I guess my
reaction is that you'll never have a typical condition. It's
what it is and I think it's a little bit misleading sometimes
to try to achieve a single number that expresses the entire
control scheme. I think Arizona right now is doing some work
with its multi-point emissions limitations concept. I don't
see any reason why that same concept, for example, couldn't
be applied to particulate emissions because you're going to
get the same kind of results whether the technology is for
SOz control or particulate control. You're going to have a
distribution around an average and if you specify the number
as the average, no matter when you go out and measure, you'll
probably never measure it right at that particular time.
You'll either be above it or below it. So, rejecting the
measurements you make, when you go out and do tests in order
to try to find some typical condition, I think is just a
fruitless search. I don't think you'll ever find it.
27. E. S. GODSEY--I think we've found that in our operations, too.
Absolutely, anything that happens is typical. The best way
to get the reverb arch to fall in is to put a pitot tube in
the flue system. I know, I've done it. And, breakdowns are
common. They're part of the job, part of the operation.
Maintaining maximum smelting rate for the four-hour period
that it takes to collect a sample, at least in our flue
system, it's extremely difficult to do. You have not only
mechanical breakdown, but you fill up the hoppers, you fill
up the reverb, the converters are down. All of this is
481
-------
typical operation but it's not maximum smelting rate. Makes
it very difficult to sample at maximum smelting rate over a
four- or six-hour period. I agree, I have to take the ups
and downs and try to estimate or guess what it would be at
full smelting rate and that is extremely difficult. We don't
smelt at full smelting rate long enough to call it a typical
operation.
28. J. H. ABBOTT—I would like to say that I think you'll have to
agree that a precipitator or any other control device, and
there are those operating this way, that perhaps would have
half the amount of plate area that it really needed for a
given situation. Maybe one that was designed in early 1900
and still operating today certainly couldn't be considered
typical. I don't think we at EPA could accept that as being
typical. There are good units operating well, and there are
those that aren't and that's what I had reference to in my
comment.
29. JOHN 0. BURCKLE—EPA-We've run some particulate emission con-
trol tests on some southwest copper smelter ESP's, and we
found a great deal of difficulty getting really good sampling
conditions at the inlet of the precipitator. I would suggest
that some of your problem would lie in that area, getting a
good inlet sample, both in flow measurement and the problem
of getting a good isokinetic traverse across those cross-
sectional areas from which you are. trying to obtain your
samples.
30. KONRAD SEMRAU—SRI International-It occurs to me that at
least part of the question here revolves around definitions
of particulate matter. In some instances, particulate matter
is assumed to include materials condensable down to a certain
point. If you have a device that's operating at elevated
temperature, there's no point in calculating its efficiency
on a basis that would require a temperature much lower than
any part of the system is operating. - Because, quite plainly,
this device, whether it's a bag filter, precipitator or any-
thing else, is not going to collect material that is not
particulate at the temperature of operation. If you have a
condensable present, and you want to specify a control effi-
ciency with reference to that component, it's going to have
to be defined in a way that can be interpreted in terms of
the performance of the unit or the temperature of operation
of the unit.
31. C. H. BILLINGS—In the beginning of my remarks, I think I
stated, but maybe not, that we had developed a method that,
in a sense, rules out condensables. Our method of testing,
theoretically, collects all the condensables in impingers.
Then if you wish to add them back in as a regulatory measure,
why that's fine, but you're really only collecting on the
482
-------
Cal1 mineral particulates. So this would rule
do with t- i * that Were n0t taken On condensables. What we
Th<* rn-vwat -I ' as far as regulation, is something else.
Was °n the basis of the difficulty of getting
We did have Difficulty in getting inlet
we have cases, for instance, where one of the
precipitators that we're dealing with was designed to meet
the more stringent of the process weight curves, and a great
deal of money was spent to change that precipitator from
meeting the old process weight curve to the new process
weight curve when it came along. And this precipitator on
the outlet has never even come anywhere near meeting that
curve. So not all of our problems were on the basis of inlet
and outlet.
32. CHARLES D. HENDERSON—St. Joe Lead-I would address the ques-
tion to Mr. Abbott and Mr. Harris, but I'm not sure that
they're the appropriate ones. In our industry, as you know,
we've been forced by recent regulations to look intently at
the exit of collection devices. We believe in order to cap-
ture fugitive emissions, whether you use an electrostatic
precipitator, a bag collector or a scrubber, you're likely
to get exit emissions in the neighborhood of 0.01 grains per
cubic foot. We calculate this to be about 23,000 micrograms
per cubic meter. Then, if the dust is typical, it contains
50% lead. My question is this: how far away, in terms of
horizontal or vertical distance or a combination of both,
would the exit of such a collector have to be from your
measuring point in order to meet the EPA standard of 1.5
micrograms?
33. D. B. HARRIS—Well, I'm not a dispersion modeler, and I think
you're probably going to have to go to them to get your
answer for that. Sounds like a long ways.
34. J. H. ABBOTT—Also sounds like an awfully low collection
efficiency for a baghouse or for even a good scrubber. The
figure of 0.01 is a fairly high emission.
35. C. D. HENDERSON—In the survey that Mr. Caplan gave us during
yesterday's session, 0.01 was about in the middle. Actually
there were some that were 0.05 and higher on the exit of the
collectors in his industry survey. Some of them were listed
at 0.001 and I would personally question that quite seriously
as being a practical, good operating collector.
36. D. B. HARRIS—I don't see why you question that. We've seen
some efficiencies that low. I can't say specifically for
lead operations—it's been five years since I've tested any-
thina like that. But for other operations, some times when
we're testing we can calibrate our equipment better using the
exit exhaust of the collector than we can the ambient air.
483
-------
The collectors are working that well-
37. J. H. ABBOTT—They're operating at .OOX on fairly fine fumes.
Now again, I don't know that much about the lead problem that
you speak of. It seems that's an awfully high emissions rate.
38. C. D. HENDERSON—Maybe Mr. Culhane and Mr. Coe can give us
some guidance on that. What kind of exit dust conditions
would you be willing to certify for your customers?
39. F. R. CULHANE—Well, there is a difference in fabric filter
performance in the laboratory and in the field... and a dif-
ference between a fabric filter that is tuned up and one
that is not. There are three major size baghouses at Hercu-
lanium and a fourth smaller one with a concrete casing. As
the bags fail, they have to be replaced, and it's possible
you may have some bags that need replacing and under those
kinds of conditions the emission values could get up perhaps
to the 0.01 value. I think the maintenance at Herculanium is
outstanding, and I think you've got an outstanding crew.
Theoretically, we should see two zeros day in and day out. I
would say your outlet should be in the 0.007-0.008 range.
40. C. D. HENDERSON—Well, all right. The difference between
0.007, 0.008, and 0.01 isn't all that great. I was just
seeking a number to be typical of a well-operating collection
device that's been out in the field for three or four years.
And, further, of course there are many applications where a
fabric filter is not appropriate because you are handling a
wet material and you have to go to scrubbers, and I believe
that the 0.01 is not a bad description of a scrubber and
might even be low for the typical scrubber. The question
still remains, whatever the collection device, you're talking
at the exit of the collection device, dust loadings in the
thousands, and you have to go to 1.5 micrograms at some
distance away where the measurement is going to be at the
fence line.
41. F. R. CULHANE—This is outside my field, but it would seem to
me that the height of the stack and the distance from the
plant are the critical values that would establish the rela-
tionship between stack outlet emissions and ambient background
emissions, along with plume buoyancy, wind velocity and topo-
graphical conditions.
42. C. D. HENDERSON—I realize that the true solution is a model-
ing problem but I thought amongst the experts on the panel
there would be some locus of distances and elevations in
typical environmental and meterological situations.
43. R. L. MEEK—I don't think we have any modeling experts on the
panel, Mr. Henderson, unless I'm mistaken. Do any of you
484
-------
gentlemen have that kind of expertise?
h IIm n0t an exPert' but I *> have some
insights. We've done a lot of modeling studies, and been the
recipients of a lot more. I don't think there's any fixed
answer to the question. It depends on the circumstances, on
tne terrain,, the frequency of occurrence of inversion condi-
tions, and a lot of meterological factors that are totally
different from one area to another. I can tell you this:
depending on the terrain you're in, a 1200-ft stack will sure
do a lot of good. I don't think there's any given answer.
If you have really high emissions, and they are near the
ground and at locations relatively close, within a mile of
the plant, you're going to have high concentrations.
45. D. B. HARRIS — My suspicion would be that your low level
emissions are the ones that you're going to have to worry
about to meet your fence line fugitive emissions level, not
your stack emissions. You're going to have to worry too about
your slag dumps and storage piles and such things to meet
1.5. The stack is not going to be the one that's going to
impact that standard.
46. F. E. TEMPLETON — If your fence line is close to the plant,
fugitives are going to be the major problem you are going to
have.
47. C. D. HENDERSON — Just in response there, to control the fugi-
tives, where you have particularly moist dust situations,
a typical thing is a scrubber with normally an outlet stack
on the order of 15 meters or something of that sort which is
fairly low level compared to a 1200-ft stack. And with the
scrubber discharges, you have the additional problem of the
power, reheating, or whatever is necessary to get it to move
up the stack and to keep it from tumbling right down to earth
again as soon as it gets out of the stack.
48. R. L. MEEK — Anyone in the audience have any experience in
dispersion modeling that might contribute to this? I don't
think our panel members are dispersion modeling people.
49. J. H. ABBOTT — Dick, I'd like to ask the gentleman a question,
if I might. What is the inlet loading to the devices that
you're speaking of here? What is the concentration at the
inlet?
50. C. D. HENDERSON— On the order of 8 to 12 grains per cubic
foot.
51 S CALVERT--I don't really know what your specific question
is about dispersion, but, in general, it isn't just a matter
of a dilution factor, but you have to know the mass rate of
485
-------
emission. If you know the kilograms per hour being emitted,
the stack height, the wind velocity, the turbulence factors
for the atmosphere, and any other special conditions, you
can predict ground level concentrations at different points.
I don't have a chart here to do it or I'd do it before your
very eyes, but I'm not sure I understand what the question
is.
52. C. D. HENDERSON—I guess my question is really to point out
what I feel is a dilemma in our industry in that we have a
typical sinter plant, and in an area of 20 meters square you
may have 3 or 4 or 5 or 6 fugitive transfer points and loca-
tions and collecting these into scrubbers or some other
device to collect the dust, each of which has an exit at the
loadings we've been talking about. And with this concentra-
tion of necessary devices in order to keep the internal part
of the plant clean, we are moving to the outside ambient air.
And then this poses a problem, even with well run operating
clean-up devices, we are emitting into the ambient air fairly
high concentrations of particulates and lead. This then puts
a quite practical limit on how far away from a fence line
these devices can be located. Quite often, depending upon
the geography of the plant location and the territory around
it, you soon go through your modeling and your examination
of the numbers, and you find that you don't have enough
property in order to possibly meet this. We're into techni-
cal problems that appear to me to be quite severe for our
industry.
53. L. C. TROPEA—I'd tell you that maybe your answer lies in
litigation rather than in dispersion. Second, there has been
a great deal of discussion about ambient standards being
applicable at plant fence lines. I might point out that
there have been some documents in circulation that have been
pretty quiet for a while which propose to change the tra-
ditional definition of ambient air that would have rather
interesting ramifications if you have to meet that standard
inside your fence rather than just at your fence line. So,
land acquisition may not be the answer.
54. R. W. LEE—I would like to support what Mr. Henderson has
said. I think his figures are very typical, in fact. He's
talked in terms of 101 grains per cubic foot, which is 23
milligrams per cubic meter and as he says, 50% for lead con-
tent is quite typical. This ties up with, for example, the
British Standards for lead which are 12 mg/m3, Japanese
standards are either 10 or 5. So the figures he's talking
about are really quite typical of standards elsewhere and
what is achievable on lead practice. Normally, I find, which
I think he is referring to also, that as far as I know, the
EPA standards for primary smelters is 50 mg. So we are talk-
ing about much lower levels than the EPA levels, anyway.
486
-------
Then you try to relate this to this fence-line standard which
is 15 micrograms per cubic meter and you've got quitea
a^aro^d ?** ?*Peri?nc! ^ that' ^ fact, the maximu^ impact
at ground level is about 10 times the stack-height distance
away from the stack. So, if you're talking about a 60 meter
stack, under our sort of conditions, anyway, which may not be
typical, you're talking about maximum impact on the ground
about 660 meters away from the stack which may or may not be
within your fence line. What determines what happens to your
fence line is more likely to be these fugitive emissions
which are windblown dusts of one kind or another which repre-
sent a serious problem if you are ever going to get down to
the levels which are being talked about. But, like many
people, I would dispute the basis on which the EPA has
selected this level of 1.5 micrograms per cubic meter. It
seems to bear little reference to actuality. There seem to
be a number of imponderables added together, the end of which/.
1.5 is the allocation given to the smelter. But I think it
includes such things as the allowance in the air from other
sources and so on. It's really quite a complicated addition
of one thing and another. Also, it doesn't seem to take into
account either the particulate sizing or fugitive emissions
at the fence line or the chemical composition. You know,
generally, in the industry it has been regarded that sulfide
lead is not particularly harmful; it doesn't get ingested,
but as far as I know, the EPA standard doesn't take this into
account.
55. FRED (A.B.) CRAIG—EPA-In reference to Mr. Henderson's (of
St. Joe) question, he referred to fugitives being a problem
versus the process emissions. Mr. Calvert, I believe, when
he gave his opening remarks, made some mention to the use of
charged fogs, or something of that nature for dust control.
The discussion we had following the paper I gave on Monday
morning centered around dust disposal. What I'd like is for
Mr. Calvert to make some comments as to the use of charged
sprays with regard to the type of problem Mr. Henderson
mentioned in order to find out if there is a possibility
that maybe there may be an option or an alternative.
56. S. CALVERT—The project on charged sprays on fugitive dust
emissions is a new one, and the equipment is just about
built now. So I can't really comment on our work. There
has been work done at the University of Arizona with charged
sprays directed at the fugitive dust source, and they have
had reasonably good results with the control of several types
of fugitive emissions.
57. J M HENDERSON—ASARCO-1 don't have a question but I do
think It would be appropriate in connection with your ques-
tion, sir, to remember that the new source performance
standard for lead smelters, blast furnaces, tip end of the
487
-------
sinter machine, etc., is 0.02 grain per standard cubic foot.
58. K. SEMRAU—I would like to address this question to Mr.
Culhane, but with comments by anyone else who wants to offer
them. What has been your experience in the pulse jet type
filters with the use of either woven or felted type fabrics?
A good deal has been made in the past of the importance of
the dust layer as a filter medium in itself and with the
woven type fabrics, of course, the reverse jet tends to dis-
rupt that. Can you give us any information or leads on
your experience with the two types of fabric?
59. F. R. CULHANE—Well, the principal with the woven fabric is
that the dust does the filtering. The interstices between the
threads or the fibers of the cloth will allow initial seepage
until the cloth heals and you establish a dust cake. As you
know, we collect particulate below submicron in size with a
fabric filter. The finest particles that I have witnessed
collected with a fabric filter are particles that have been
measured in Angstroms, 10,000 Angstroms to a micron, and
we've collected particles as fine as 320 Angstroms. The only
reason that we can collect this fine particulate matter is
because these particles, fortunately, agglomerate, and in
agglomerating they form an excellent filter cake achieving
high filtration efficiencies in the 99.9% range. When we get
into a pulse-type fabric filter, our design goal is to
increase throughput at comparable filtration efficiencies,
and a tight dust cake is a limiting factor in achieving our
goal. Accordingly, felt is substituted for the combined
woven fabric and the dust cake. The felt requires more
energy to clean and the deacceleration forces effectively
clean the felt on a frequent sequence. Substitution of
woven fabric in a pulse jet will result in over-cleaning and
loss of the filter cake and reduced efficiencies.
60. MARK HOOPER—EPA-One of the things I would like to comment
on is that I think the lead standard was promulgated as a
result of a court order handed down to EPA within the frame-
work of Section 307 of the Clean Air Act. The lead industry
does have a forum to bring suit, and I think they may have
done that in the Circuit Court of appeals for the District
of Washington, D.C. I would imagine and hope that the lead
industry can develop a good case at least to bring forth
their concern. I'm not qualified to say whether that stand-
ard is right or wrong. But the forum is there for the lead
industry to speak its piece and hopefully, at least, if they
have a valid case they could convince the court. I believe
that's where it is now.
61. GEORGE THOMPSON—EPA-1 think in the wastewater area, we're
coming up with new techniques that are very good in advancing
the state-of-the-art. We have some good ideas in the" solid
488
-------
waste area. Considering the expertise on the panel the
Pa28iv, iS9 5ePrfsented very well by industry/manufacturers,
and the EPA itself 1 ask, where do we go in the future'
nology? ^ tO *A™C* &il Poll«tioi control tJch-
62. J. H. ABBOTT—That's the topic of the paper I'm scheduled
to give the last thing tomorrow. So I'll defer my answer
until then.
63. F. R. CULHANE—That' s a question that I hear often. We are
still using the same basic principles of fabric filtration
that Mr. Godsey tells me were first employed in 1850 in a
zinc operation along with the work in the early 1900s. Also,
the early work Dr. Cottrell did in the first decade of the
century employed the same electrostatic principles that we
employ today. It's interesting that a lot of our gas clean-
ing experience today evolved out of the nonferrous mining and
metallurgical industry. With 6-8% of your material going up
the stack, your industry has some economic motivation to do
something about it. In those early days of stack gas clean-
ing, basic laws of filtration and precipitation were applied
to gas cleaning and they're still in force today. We have
not had a major breakthrough in basic principles since the
first Cottrell in 1912 or the first baghouse in 1900. We
bump into the same drag loss in a fabric filter for a given
filtering velocity. We are fighting the same migration
losses that were fought in the early days. The progress that
I see in the gas cleaning industry is the application of the
control equipment. In this area, our progress has been out-
standing. Certainly, there has been work done in the refine-
ment of the equipment and progress has been achieved in
materials, better electrical controls for precipitators, and
after the Second World War, the synthetic fibers opened up
new applications of fabric filters for us as we were able to
operate above the dew point at 300°, 400° and 550°F. But
our principal progress has been in the application of this
equipment.
What's around the corner? something that can cut the size
of the precipitator in half? Or cut a fabric filter in half?
Or cut the energy requirements of a scrubber in half? I
don't see that sort of a thing in the picture if we're look-
ing for that as progress. You're in R&D, Mr. Coe. What
are your thoughts?
64. E. L. COE--I concur, at least in part, with what Frank
Culhane says in that most of the possibilities for improving
particulate collection appear to lie in the enhancement of
the basic physical principals that we've been working with.
There is work going on to decrease the size of the precipi-
tators through the use of a better charging mechanism for the
489
-------
precipitator. This work is sponsored by the EPRI and is
taking place at the Arapaho Station of the Public Service
Company of Colorado. It has promise of reducing the size of
precipitator required for a given performance. There is a
good deal of work yet to do on that project before it's
determined whether it is in fact a God-send or an also-ran.
Similarly, enhancement of fabric filter collections through
electrostatics in combination with scrubbing has been invest-
igated, and I believe that Dr. Calvert is one of the prime
investigators of such phenomena. Unless there is a genius
who has invented a new physical principal that we've not dis-
covered, we're going to collect dust in the same way, but I
also believe that we're going to improve its reliability.
65. S. R. OREM—Having watched this business since 1937, I like
to feel some remarkable progress has been made. I think
particularly of the number of power plants that you can look
at today and later be told that all the generating units are
on the line and you can't see anything. Well, there weren't
many like that fifteen years ago. And there have been a
combination of motivations that have achieved this. It's not
just the hardware suppliers that have done it. It's been a
motivation drive on the part of the plant owner for a
standard of excellence, his willingness under various pres-
sures, his willingness to invest more money on a bigger box.
As I say, these motivating forces have combined to produce
some excellent achievements in improved hardware and improved
operating techniques.
66. F. E. TEMPLETON—I'd just like to make one comment. I'm a
little perplexed, Mr. Thompson, at your question. We just
got through spending hundreds of millions of dollars on what
we've got now and you're suggesting what's the next step and
what are we going to replace all this stuff with? I think we
ought to concentrate on what we have now that works rather
reliably. The other thing is, if the problem's solved, that
is, if pollution is abated to the degree that our society
is agreed that it ought to be, then I really have to question
how much more effort in research to improve efficiency should
be put forth. Perhaps we should focus research on making the
existing systems more reliable and on processes that will use
less energy.
67. G. THOMPSON—Mr. Templeton, I didn't intend to challenge the
industry's record in application of existing technology to
control existing plants. My concern is the future those
cases that involve the installation of complete new plants,
expansion of existing ones, or replacement of worn out equip-
ment. We should be concerned with a number of factors: con-
trol efficiency, equipment reliability, maintainability, pre-
vention of performance degradation, energy consumption, and
the like. What efforts are being made to make air pollution
490
-------
control more cost effective, more affordable? Are there
technological improvements on our horizon which may come to
have a significant impact in reducing the costs of air pollu-
tion control? ^
68. F. E. TEMPLETON--I think your point is pretty clear, but I
do think that you have to recognize it both from a regulatory
standpoint and what kind of rules should we be looking at in
the future as well as what kind of technology should we have.
There seems to be a tendency to want to apply any new tech-
nology to everything that's already in as soon as it comes
out, and I'm not just accusing EPA of that becuase I think
there're some pretty sensible things done there, but I do
think a lot of environmental groups, for example, as soon
as word gets out, the, "Here's a new technique or a new
technology, ^there's pressure to apply it everywhere, even
with existing plants, and often before it has been thoroughly
demonstrated and proven in actual practice. Researchers
sometimes forget about the need for sustained reliability.
So while it may be that future plants, as they are designed,
can be built with better technology from an energy standpoint,
there is sure a tendency to run with the ball as soon as it
pops up even with the plants that have already made substan-
tial investments.
69. G. THOMPSON—I might add being part of a regulatory agency
people sometimes do tend to run with some of the research
results that we produce. We have to be very cautious that
what we produce is correct, is factual, and does the job.
Allow me to provide one more example in the wastewater area.
Some of the "lime and settle systems" I've seen in the United
States are not too effective in metals removal. There are
new technologies that have been developed in the past few
years that don't produce a sludge. Sludge may not be one of
your industry's problems, but it certainly is a problem with
other industries. If we don't advance the state-of-the-art
for certain existing technologies, we're not taking care of
the total problem, and we obviously won't be preparing for
the future.
70. S. CALVERT—I don't want to say too much because Jim Abbott
is going to give a paper on new technology, and I hope he
may even mention something we're doing. In general, we may
not have discovered any great new principles for particle
collection, but I think we have learned a lot about the way
particles behave and what can be done to improve collection.
If one is able to do a good engineering design nob, he 11
take advantage of the conditions that exist and use them to
improve collection efficiency rather than perhaps fighting
against it. There are many installations where the equip-
ment simply didn't use what was available to the best advan-
tage. Then, I'm struck with another thing in listening to
491
-------
the papers here, that in many cases the question wasn't
whether the scrubber could give the proper collection effi-
ciency but whether you could just keep it on line, and
whether it would be free from being plugged up by scaling or
otherwise being made inoperative. So putting things together
in different combinations, and different configurations that
are easier to operate and easier to maintain can offer some
real advances. And this is an area where people in industry
with the problem have to tell the people in R&D just what the
problems are so something can happen which will be beneficial.
71. L. C. TROPEA—Dick, I have one comment to pick on this water
pollution commentary that was just made. Pretty soon, say
sometime in the 1980's, we're going to have, for the first
time, the major legislation in air, water, and hazardous
wastes, all out and in final form. EPA's effluent guidelines
program is restudying everything in the world. Limitation on
the traditional pollutants, like the total suspended solids,
pH, and things like that, are things of the past. They are
going back now and looking at a whole bunch of substances
which many of us would have difficulty even spelling. In the
very near future, they are going to amend all these guide-
lines. Anybody who's got a water discharge from air pollu-
tion control systems is going to be faced with some real
challenges with some of the limits that they are going to be
imposing in this area. When you take the new water require-
ments and add the RCRA requirements, the hazardous waste
requirements, which will eventually require any sludges that
you produce at any waste treatment facilities or any bag-
house catch that you make, there will likely be some new
ground plowed in employing a system's approach to future air-
quality decision-making.
72. F. R. CULHANE—George, I might just make a comment that there
was a development on the horizon in the fifties that was
exciting. It was a sonic agglomerator. And those of us
that were doing the work in those days looked at it as an
opportunity to increase the throughput in fabric filters for
a given pressure drop or to reduce the energy requirements
of the wet scrubber, or to reduce the size of the precipita-
tor. The energy requirements of the sonic agglomerator
proved to be unattractive and the device faded and never
really became commercial. But it is something, perhaps, our
industry should reinvestigate. If we eliminate the fines
and have only coarse material to collect, we could reduce
the size and cost of the control equipment.
73. E. L. COE—I'd like to add a little on my previous comment
on combinations of effects to improve particulate collection.
There's been a tendency over the years to address particulate
collection in terms of one device, that is a baghouse, a pre-
cipitator, or a scrubber. I think that the nonferrous metals
492
-------
industry is probably the leader in handling their pollution
problems, air pollution problems that is, with a train of
devices. The hot precipitator, followed by the scrubber,
followed by the acid mist precipitator, followed by the sul-
furic acid train, is a tremendous combination of equipment.
But, it does illustrate one of the principals that is assist-
ing in this area and that's the application of a series of
devices each operating in its own best regime. The electro-
static enhancement of scrubbing, or, in terms of scrubbers
that have a problem with outlet emissions, use of scrubbing
followed by wet electrostatic precipitators is a way to
achieve levels of control although not necessarily levels of
cost that are highly attractive.
74. JOHN D. MARTIN—BHAS of South Australia-There has been no
mention tonight of venturi scrubbers, and I'd like to direct
a question to the panel or anybody else who's had experience
with them. At Port Pirie, we use long-throat venturi scrub-
bers in our sinter plant circuit, and we get quite good effi-
ciencies with these units. Our problems at the moment are
concerned with the out-of-balance on our fans due to carry
over of solids probably in our liquors. My question is, does
anybody here have any experience to show that the long throat
venturi is, in fact, more efficient than more commonly used
short-throat. We put these in at the recommendation of a
consultant and unfortunately haven't been able to compare
results. Secondly, I wonder if this buildup which we are
getting on the fan, which is giving us our out-of-balance
problems, isn't connected with our inefficient operation of
our cyclone and scrubber. Would someone like to comment on
it, just for interest?
75. S. CALVERT—So far as the effect of the throat-length on the
effect on venturi scrubber efficiency goes, we have a mathe-
matical model to predict venturi scrubber efficiency, and it
indicates that throat-length is important up to a point.
There's no one number that indicates how long the throat
should be but, in general, probably lengths on the order of
two or three feet are adequate. For very large diameter
throats, this may be larger.- But there is a mathematical
model, and it's presented in an EPA publication entitled,
"Particulate Control Highlights, Performance, and Design
Models for Scrubbers" and we have another previous report
that deals particularly with the venturi scrubber model. So
to sum up, there's a theoretical reason for believing that
the longer throat is more efficient. To determine whether
this is really so, it would be necessary to look at your
collection efficiency data and compare them with other data
from Venturis and similar scrubbers with shorter throats.
We'd be very interested in getting some data for that pur-
pose. The problem of the fan going out of balance "very
likely due to liquid entrapment being carried into the fan
493
-------
and might require asking one more question. Do you have any
sort of mist eliminator between the venturi scrubber and the
fan?
76. J. D. MARTIN—No, we haven't. We have a cyclone separator to
trap the liquor, but we have not any mist eliminator at all.
That was something that I thought that we should probably try.
Initially the fans were made of mild steel and we had cor-
rosion problems and buildup. Then we changed to stainless
steel, and these are giving reasonably satisfactory perform-
ance. But we are still getting buildup which requires clean-
ing three times a week, which is excessive.
77. S. CALVERT—Well, a cyclone separator is one type of entrain-
ment separator or mist eliminator. They can be efficicent,
that is, they can be adequate. However, it is possible to
overload them by having too high an inlet velocity so there
can be reentrainment of the liquid drops that are deposited
on the cyclone walls. So it's a matter of looking into
specifics of the case. I think this problem of solids depo-
sition or entrainment carry-over is a very common one and
really was the basis for EPA funding several research con-
tracts which we performed to study entrainment separator
characteristics and to define them. Again, we've written
some reports, some EPA publications on that subject and if
you'd like afterward, I can give you some references to those.
78. E. L. COE—I have a slight additional comment with regard to
mist eliminators. Mist eliminators used with most scrubbers
in present practice are basically mechanical collectors.
They are designed to collect liquid droplets as particles.
They do have a limit in terms of the energy that is put into
them which limits what they can collect. If all the droplets
came through large, you would presumably collect them all
providing you didn't have a reentrainment problem. The prob-
lem is that nobody told the droplets that they had to stay
that nice large size. For any mist eliminator of a mechan-
ical collection type, inertial, impaction, or centrifugal
(and obviously some are better than others), there does
appear to be a limit as to how clean it can get the gas
stream in terms of the entrained particles, particularly
if a portion of the particles are of extremely small size.
In this connection, I refer back to the remark I made pre-
viously about putting another device behind the scrubber
to do a portion of the collection job. I'm not particularly
promoting wet precipitators, but that is one way of doing it
in the extreme.
79. F. E. TEMPLETON—I might just add, we've had this same prob-
lem with scrubbers and we've tried everything, such as coat-
ing the fan blades, and I guess you probably have done the
same sort of thing. We have one unit that has a fan on the
494
-------
hhotoi S±2e ln °rder to 5et enough pressure
The hot side fan doesn't have any problems at all
So one of the approaches we were going to ?akewas to pu^
lem on^h '"V?6 $Ot Side S° ™ W°n
-------
gotten essentially the same sort of performance from pipe
line contactors, from orifice contactors, and Venturis of
different shapes and sizes.
83. R. W. LEE—I'd like to take this question of cyclonic separa-
tors a bit further. For several years now, I've been vainly
trying toacquire design information on cyclonic separators,
and I have acquired a certain amount just by collecting sizes
of various separators. But, I've looked through quite a lot
of Dr. Calvert's work, and I haven't seen very much in the
way of design information about cyclonic separators. You've
certainly done a lot on mist eliminators, but I would be very
interested in any reference you can give on this particular
subject. But going on a bit from that, after this particular
problem about fan build-up, I think it depends on what effi-
ciency you're getting from the venturi itself. If you use
low pressure drop Venturis on fine size fume, you don't
really remove a great deal of it. And, if you add up over a
day or three days, or seven days, or whatever it is, the
amount of actual penetration you are getting through a
scrubber like this, you'll be surprised how much solid matter
the fan is actually handling. Depending on what type of fan
you're using, whether it's a paddle blade or whether it's a
more efficient fan on which fume tends to stick more easily
anyway, you can get these problems.
84. S. CALVERT—With regard to design of cyclone separators, if
you're talking about designing for the primary collection
efficiency, by which I mean the efficiency with which the
drops are first removed from the gas, we presented a model
for that in the Scrubber Handbook which is an EPA publication
and the model is essentially the Leith and Licht model for
particle collection by cyclone separators. It takes into
account the turbulent mixing in the cyclone. We went on
from there into an experimental study of cyclone separators
as mist eliminators or entrainment separators in order to
investigate the reentrainment velocity which essentially
limits the capacity of the entrainment separator. We had
found some conflicting statements in the literature indicat-
ing that you could use anywhere from 45 ft/sec to over 100
ft/sec inlet velocity, which gives you quite a bit of room
for confusion. So we did experimental work on one type of
cyclone separator which actually was a design which we lifted
from Nonhebel's book. I believe it's a Stairmand design.
This is a cylindrical body with a top inlet with the anti-
creep skirt on the outlet and an anti-swirl baffle in the
bottom. We were unable to get reentrainment at inlet velo-
cities up to something a little over 100 ft/sec. So, we
have some limited data on the reentrainment characteristics
of a cyclone separator used as an entrainment separator.
And we have, as I said, presented design methods for comput-
ing the primary collection efficiency and indeed have
496
-------
repeated it and improved it in this little report on the
desxgn methods for scrubbers. report on the
P°int' Itve looked at these methods
efficiencies. But as far as I can
H a11 Jhe cyclonic separators that I've seen have
bottom inlets and top outlets whereas most of the work seems
to have been done on the dust cyclone which is rather differ-
S? I Jf°u.k?OWf there must be some significance to the fact
that the inlet is on the bottom instead of the top. This
must have some effect on its performance.
86. S. CALVERT— I don't think the drops know about it. And
there are cyclone separators that are used as entrainment
separators in which the gas inlet is at the top, and they
are essentially analogous to cyclones used for dust collec-
tion. The problems that we could get into with bottom inlet
and top outlet type cyclone separators are more of the type
related to operations. There can be solids depositions,
there can be very severe problems due to that. Where solids
can be deposited and can impair the efficiency of the cyclone
separator and can cause very serious maintenance problems.
But so far as the primary collection efficiency goes, there's
no reason to believe that it should be any different assuming
one has taken all the variables into account.
87. E. L. COE — I have an additional comment. I cannot add to
what Dr. Calvert has said about the design of the cyclonic
separators; however, if you wish to look at more literature
on the subject, a good deal of it is contained in the patent
literature on the separation of droplets from steam.
88. TOM C. SUNTER — Carborundum-Some comments were made this morn-
ing about the use of an extra compartment in a dust collector
for maintenance purposes and the possible use of opacity
meters after each compartment to detect a broken bag. Has
the panel any experience in the use of such units for detect-
ing a broken bag in a compartment and will that find the bag?
Our experience is that if you don't catch a broken bag right
away, you'll have five of them or more and the sooner you do
it the better job you have. Has the panel any experience?
89. E. L. COE — There is a unit at the Pennsylvania Power and
Light Company that has been in service now for about five and
a half years, and it has been found at that installation that
the compartment in which there is a perforated bag can be
located by correlation of the opacity reading on the outlet
and the timing of the cleaning cycle. This is very conven-
ient at this particular station because the opacity meter and
the cleaning cycle control panel happen to be side by side.
The indication of a failed or perforated bag appears when the
497
-------
compartment having the perforation is on the cleaning cycle
the opacity drops because that leaking compartment is then
off line. This is a reverse air baghouse. It has been found,
as recounted to me by the station personnel, possible to
locate a perforation equivalent to a 1/4-in. diameter hole in
one bag in one compartment. For reference purposes the bag-
house has approximately 5000 bags, and I do not recall the
exact number of compartments, it's on the order of 16, I
believe.
90. F. R. CULHANE—I'd just like to support that statement that
Lee has made. We have an opacity meter on a large 2 million
cfm baghouse. The opacity meter picks up a blip, it comes
on when a leaking compartment goes off the line and is no
longer filtering. You can't pick it up with your eye in the
stack, the stack still remains clear until we're down in the
order of magnitude maybe -004, .005 or something like that. So
the opacity detection system on this installation is extreme-
ly sensitive. Prior to opacity meters, we used to be able
to do this visually. In a carbon black plant you could see
a small plume coming off a stack, and when a leaking com-
partment would go off for cleaning, the stack would clear up.
Then you knew that you could isolate your problem in a com-
partment. But the opacity meters on the market today are
much more sophisticated and much more accurate than the naked
eye. It's a very practical approach.
91. E. L. COE—I would add a caution, however, (1) you have to
have a good opacity meter well maintained or you are going to
miss the effect, (2) you need to have a good correlation
between the opacity readings and the cleaning cycle. If you
do not have the opacity meter at the panel that indicates the
compartment being cleaned, you would need to, somehow or
other, log the two operations so that you could correlate
them and (3) if you get several such compartments simultane-
ously, you never will be able to find out which one it is.
Therefore, you need to catch them as they go.
92. E. S. GODSEY—The latest baghouse installation that we have
installed has a system very similar to the one described
except it carries it a little further. We have an opacity
meter that monitors the total output for the baghouse and
this one is rated for about 200,000 acfm. The instrumenta-
tion is arranged so that if the opacity reaches a pre-set
limit, the baghouse will automatically go into a shut-damper,
delay, open damper, and then continue on through the 16 com-
partments until the damper that is closed causes a reduction
in the opacity reading. If that happens, the damper will
automatically stay closed, the compartment is taken off the
line until the operators can get around to going in and find-
ing the broken bag. If the first compartment that has a
broken baq in it shows a decrease in opacity, but not to the
498
-------
set point, that damper will stay shut, the cycle will con-
tinue until it reaches the second compartment that has a
5f2?Lba9Qand.i? that °ase we'd have two compartments out of
service. So, it s a system similar to the one described
except that it's automatic, and it does locate by actively
seeking out the bag compartment that's causing the problem.
The instrumentation costs almost as much as the baghouse.
93. GEORGE KAZONICH—Phelps Dodge, Morenci, Arizona-Getting back
to Mr. Billings' original comment on the precipitator effi-
ciency versus maintenance, I'm one of the people who crawls
into precipitators and cleans them out and keeps them running,
and I think we had some of the highest and some of the lowest
points on his testing. And it comes down to the point of how
much effort are you going to expend to keep up with that 99%
efficiency? One of the highest points that he recorded was
after a two months rebuild of a unit with total plate
straightening, replacement of outlet plates, replacement of
virtually all wires in the outlet section, tuning up the
electronics by the manufacturer's representative, just on and
on and on through the entire unit. But obviously you can't
go through a two-month rebuild of the unit, put it on line
for a week, take if off and do the same thing again. There
has to be some economics put into this thing to say, well, we
can maintain a 90% capture or a 95% capture, or 98% effi-
ciency, whatever. But economics have to get into this some-
where or other. Because you can't just clean a precipitator
every day in order to keep it up at 99.9% all the time. And
one other comment, I'd like to side with Mr. Billings on
one thing, because he's requesting some guidance from EPA
and I know Arizona has been spanked on the hands several
times because the guidelines that they've come up with or
their presentations for control parameters are generally
declined as being unsuitable by EPA. So, I think the fact
that he is seeking guidance here is showing forethought on
his part and that it hasn't seemed to come from the EPA
people at this meeting is a bit sad.
94. R. L. MEEK—If I may comment on your question, Mr. Kazonich-
—The purpose of this forum is an interchange of technology
and information, not to respond or supply guidance from EPA
which should be available through other channels. I think
the answer to the first part of your question on the opera-
tion and maintenance of precipitators probably addresses
several members of the panel. Let's start with Dr. Templeton.
95. F E.. TEMPLETON—I think I would have to agree with the idea
that you certainly have to bring economics into the picture
on any piece of equipment, not just precipitators. Our
experience with our precipitator is perhaps a little differ-
ent than what Mr. Billings had in mind because our precipita-
tor is not on the terminal end of the control string. It's
499
-------
on the front end of it and so there are other pieces of
equipment after the precipitator that take other materials
out. We haven't had any particular problems with our precip-
itator from a maintenance standpoint. Ours is designed with
a spare capacity in terms of the number of sections to it and
so we can take a section out and still run the precipitator
and do maintenance work. That's one approach that is useful
in keeping it on line. I'd go back to a point I made earlier,
that is if you are going to choose a specification for
emissions limitations for a plant, you have to look at that
particular plant and I think that you will find that that
specification really should not be a single number but a
number that reflects the operation of the plant itself, the
process, as well as the fluctuations and efficiency of the
equipment, that takes into account the changes that take
place over time and the efficiency because of the degradation
of the collection efficiency over time. Again, I don't think
it's correct to sample when you just finished doing a clean-
ing job on the precipitator. I think you need to take enough
samples over a long enough period of time that the overall
performance is evaluated and can be looked at in light of
what it really does, not just what it does in ideal condi-
tions which never really exists except for a particular
instance of time.
96. E. L. COE—One of the questions that is being raised here,
at least by implication, is the performance of control
equipment over time, and one of the problems in accessing
this factor is the fact that we have only a cumbersome,
relatively instantaneous method of measuring mass emissions,
yet we are characterizing everything that we are talking
about in terms of these mass emissions. And my comment is
this, that it would behoove everyone who is working on meas-
urement devices, Bruce Harris included, to look for a way of
instantaneously or continuously measuring mass emissions
over time.
97. C. H. BILLINGS—I'd like to respond to some of the things
that were said. In the first place, I didn't bring the
question up to seek guidance and go right home to Arizona and
write a regulation and that would be it. We do get along
pretty well with EPA although we have our differences; I
think all the states do. I didn't want to cause any dis-
sention in the ranks as it were. The reason I brought the
question up is that I thought it would be an interesting
discussion, and it was germane to what we were talking about
tonight. Some of the things that we have to consider on
any dust collection device in determining what the efficiency
is and should be are based on certain theoretical precepts
and are also based on certain- practical precepts. For
instance, in electrostatic precipitators dust buildup on the
electrodes and configuration of the electrodes, the wires,
500
-------
reentrainment, resistivity, that sort of thing, once your
tS Snt and in Place' ^ose ar^fiSTpSL-
1Sn ^ much you can do about them. It can be
Ve; in/act' to correct some of them. The
aspects of this are that you might have poor
maintenance, or inadequate voltage control, improper condi-
niSo^S9 °x Jhe material going in, etc. So I think we have to
consider both the practical and the theoretical aspects of
any dust collection device. What I was seeking tonight was
to get some guidance as to what the electrostatic precipita-
tor on a copper plant was capable of. I didn't exactly want
one number but I just wanted to know some sort of guidance
as to what such a device is capable of. As I said before,
I'm sure it's better than 74 and a half.
98. R. L. MEEK—Any other comments?
99. D. B. HARRIS—Dick, could I just get in a commercial while
I'm here? Since everybody has been participating so well in
the symposium, I just wanted to mention a couple that we
sponsor that might be of interest to you on the measurement
of particles. One that will be coming up in October in
Daytona Beach, Florida on October 7th through the 10th is
titled "Advances in Particle Measurement and Sampling" and
a call for papers is going out. If anybody has been doing
any unique efforts in this area, we'd be glad to hear from
you. We had a very successful one last May in Asheville, N.C.
The proceedings for that are now available if anyone was not
able to attend and would like to receive a copy of that, get
in touch with me back at IERL or with our reports distribu-
tion center. We also sponsored one on fugitive emissions,
the last one was in October in San Francisco and hopefully
we'll have the proceedings of that soon. We're planning
on May of 1980 as being the time of the next meeting on that,
but we have yet to pick a site.
100. R. L. MEEK—No further questions? Well, I think we owe the
panel a round of applause for their participation. And
gentlemen, I thank you. This session is adjourned.
501
-------
INDEX TO PANEL COMMENTS
Abbott, J. H. (USEPA, IERL-RTP)
2. (p. 470), 21. (p. 479), 25. (p. 480), 28. (p. 482), 34. (p. 483),
37.(p. 484), 49.(p. 485), 62. (p. 489)
Billings, C. H. (Arizona Bureau of Air Quality Control)
16.(p. 477), 22.(p. 479), 31.(p. 482), 97.(p. 500)
Burckle, J. 0. (USEPA, lERL-Ci)
29.(p. 482)
Calvert, S. (A.P.T. Inc.)
3. (p. 470), 51. (p. 485), 56. (p. 486), 70. (p. 491), 75. (p. 493),
77.(p. 494), 84.(p. 496), 86.(p. 497)
Coe, E. L. , Jr. (Joy Manufacturing Co.)
4. (p. 471), 17. (p. 478), 64. (p. 489), 73. (p. 492), 78. (p. 494),
87. (p. 497), 89.(p. 497), 91.(p. 498), 96.(p. 500)
Craig, A. B. (USEPA, lERL-Ci)
55.(p. 487)
Culhane, F. R. (Wheelabrator-Frye Inc.)
5.(p- 472), 13.(p. 475), 15.(p. 477), 18.(p. 478), 39.(p. 484),
41. (p. 484), 59.(p. 488), 63.(p. 489), 72.(p. 492), 90.(p. 498),
Czuchra, P. Z. (FMC Corp.)
81. (p. 495)
Harris, D. B. (USEPA, IERL-RTP)
7. (p. 473), 20. (p. 479), 33. (p. 483), 36. (p. 483), 45. (p. 485),
99.(p. 501)
Hooper, M (USEPA, Seattle)
60.(p. 488)
Kazonich, G. (Phelps Dodge Corp.)
93.(p. 499)
Lee, R. W. (Imperial Smelting Processes Ltd.)
12.(p. 475), 54. (p. 486), 83.(p. 496), 85.(p. 497)
Martin, J. D. (Broken Hill Associated Smelters Pty.)
74.(p. 493), 76.(p. 494)
502
-------
Meek, R. L. (Southern Research Inst.)
1. (p. 469), 11. (p. 475), 19. (p. 479), 43. (p. 484), 48. (p. 485),
94.(p. 499), 98.(p. 501), 100.(p. 501)
Orem, S. R. (Industrial Gas Cleaning Inst.)
8.(p. 473), 65.(p. 490), 80.(p. 495)
Semrau, K. (SRI International)
30.(p. 482), 58.(p. 488), 82.(p. 495)
Sunter, T. C. (Carborundum)
88.(p. 497)
Templeton, F. E. (Kennecott Copper Corp.)
9. (p. 474), 24. (p. 480), 26. (p. 481), 44. (p. 485), 46. (p. 485),
66.(p. 490), 68.(p. 491), 79.(p. 494), 95.(p. 499)
Thompson, G. (USEPA, lERL-Ci)
61. (p. 488), 67. (p. 490), 69.(p. 491)
Tropea, L. C., Jr. (Reynolds Metals Co.)
10.(p. 475), 53.(p. 486), 71.(p. 492)
503
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
REPORT NO.
EPA-600/2-79-211
I. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Control of Particulate Emissions in the Primary
Nonferrous Metals Industries - Symposium Proceedings
5. REPORT DATE
December 1979 issuing date
6. PERFORMING ORGANIZATION CODE
. AUTHOR(S)
R. L. Meek, Editor
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Southern Research Institute
2000 Ninth Avenue South
Birmingham, Alabama 35205
10. PROGRAM ELEMENT NO.
1AB604
11. CONTRACT/GRANT NO.
R-804955
12. SPONSORING AGENCY NAME AND ADDRESS
Industrial Environmental Research Laboratory
Office of Research and Development
U. S. Environmental Protection Agency
Cincinnati, Ohio 45268
13. TYPE OF REPORT AND PERIOD COVERED
Final; March 1979
14. SPONSORING AGENCY CODE
EPA/600/12
15. SUPPLEMENTARY NOTES
Project Officer: John 0. Burckle
16. ABSTRACT
The purpose of the symposium on "Control of Particulate Emissions in
the Primary Nonferrous Metals Industries" was to provide a forum for the
exchange of knowledge and new ideas on particulate control technology with
emphasis on industrial applications of environmental particulate control
technology in the primary nonferrous industries. The symposium held at
Monterey, California, March 18-21, 1979 was sponsored by the U. S. Environ-
mental Protection Agency, Office of Research and Development, Metals and
Inorganic Chemicals Branch, Industrial Environmental Research Laboratory-
Cincinnati .
The symposium included presentations on foreign and domestic technology
applicable to copper, lead, zinc, and aluminum as well as discussions on
other advanced technology, measurement techniques, and current EPA programs
that are pertinent to particulate control in the nonferrous metals industry.
Speakers from England, Canada, Sweden, Japan, Australia, and the
United States discussed recent developments and technology for particulate
control in nonferrous operations.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. cos AT I Field/Group
Exhaust Emissions
Smelting
Trace Elements
Pollution
13B
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
512
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (Rev. 4-77) PREVIOUS EDITION is OBSOLETE 504
ft U.S. GOVERNMENT PRINTING OFFICE: 1980 -697-146/5553
------- |