United States Office of Air Quality EPA 450/3-83-018a
Environmental Protection Planning and Standards March 1984
Agency Research Triangle Park NC 27711
Air
&EPA Review of
New Source
Performance
Standards for
Primary Copper
Smelters
Chapters 1 Through 9
-------
EPA 450/3-83-018a
Review of New Source Performance
Standards for Primary Copper Smelters
Chapters 1 Through 9
Emission Standards and Engineering Division
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
March 1984
. n Protection
;,! Pro t
IL 60604
-------
This report has been reviewed by the Emission Standards and Engineering Division of the Office of Air
Quality Planning and Standards, EPA, and approved for publication. Mention of tirade names or commercial
products is not intended to constitute endorsement or recommendation for use. Copies of this report are
avialable through the Library Services Office (MD-35), U.S. Environmental Protection Agency, Research
Triangle Park, North Carolina 27711, or, for a fee, from National Technical Information Services, 5285 Port
Royal Road, Springfield, Virginia 22161.
-------
ENVIRONMENTAL PROTECTION AGENCY
REVIEW OF NEW SOURCE PERFORMANCE STANDARDS
FOR
PRIMARY COPPER SMELTERS
Prepared by:
J^ckR. Farmer . .
Director, Emission Standards and Engineering Division
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
1 Existing standards of performance for primary copper smelters were
promulgated in 1976. Section 111 of the Clean Air Act (42 USC 7411),
as amended, directs that the Administrator periodically review promul-
gated standards.
2 Copies of this document have been sent to the following Federal depart-
ments- Labor, Defense, Interior, Health and Human Services, Agriculture,
Transportation, Commerce, and Energy; EPA Regional Administrators; and
other interested parties.
3. For additional information contact:
Dr. James U. Crowder
Industrial Studies Branch (MD-13)
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
Telephone: (919) 541-5601
4. Copies of this document may be obtained from:
U.S. EPA Library (MD-35)
Research Triangle Park, NC 27711
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
m
-------
TABLE OF CONTENTS
1. SUMMARY } ,
1 1 REGULATORY ALTERNATIVES f f
1 1.1 Reverberatory Smelting Furnace Exemption 1-1
1.1.2 Control of Reverberatory Furnace Particulate
Matter Emissions |~^
1.1.3 Expansion |"^
1.1.4 Fugitive Emissions |~~
1.2 IMPACTS 1~3
2. INTRODUCTION l~]
2 1 BACKGROUND AND AUTHORITY FOR STANDARDS £-1
2'2 SELECTION OF CATEGORIES OF STATIONARY SOURCES 2-5
2'3 PROCEDURE FOR DEVELOPMENT OF STANDARDS OF PERFORMANCE ... 2-7
2^4 CONSIDERATION OF COSTS 2-9
2 5 CONSIDERATION OF ENVIRONMENTAL IMPACTS 2-10
2.6 IMPACT ON EXISTING SOURCES ^\L
2.7 REVISION OF STANDARDS OF PERFORMANCE ^-^
3. THE PRIMARY COPPER SMELTING INDUSTRY: PROCESSES AND POLLUTANT
EMISSIONS 3~:
3.1 GENERAL ^ ,
3.2 PROCESS DESCRIPTION ^
3.2.1 Roasting and Drying •f ^
3.2.2 Smelting ~~r:
3.2.3 Converting ^"^"
3.2.4 Fire Refining :T^'
3.2.5 Continuous Smelting Systems *~**
3.3 EMISSIONS FROM PRIMARY COPPER SMELTERS 3-44
3.3.1 General *~ .
3.3.2 Process Emissions ^ ^
3.3.3 Fugitive Emissions ^"^°
3.3.4 Summary of Fugitive Emissions Data J" = /
3.4 EXPANSION OPTIONS FOR EXISTING FACILITIES 3-62
3.4.1 Multihearth Roasters 3-62
3.4.2 Fluid-Bed Roasters ;>-fa4
3.4.3 Reverberatory Furnaces ;T°^
3.4.4 Electric Furnaces ?~'*
3.4.5 Outokumpu Flash Furnaces ^~«j
3.4.6 Noranda Reactors 3-81
3.4.7 Converters
3.5 SUITABILITY OF ALTERNATIVE TECHNOLOGIES FOR PROCESSING
HIGH-IMPURITY FEEDS 3"83
-------
TABLE OF CONTENTS (con.)
3.5.1 Background 3-83
3.5.2 Impurity Behavior During the Smelting Process . . . 3-85
3.5.3 High-Impurity Feed Processing Experience with
Outokumpu Flash Furnaces 3-100
3.5.4 High-Impurity Feed Processing Experience with
Inco Flash Furnaces 3-103
3.5.5 High-Impurity Feed Processing Experience with
the Mitsubishi Process 3-104
3.5.6 High-Impurity Feed Processing Experience with
Noranda Reactors 3-104
3.5.7 Conclusions 3-107
3.6 BASELINE EMISSIONS 3-111
3.6.1 Process Sources 3-111
3.6.2 Fugitive Sources 3-117
3.7 REFERENCES 3-118
EMISSION CONTROL TECHNIQUES 4-1
4.1 GENERAL 4-1
4.2 SULFURIC ACID PLANTS 4-3
4.2.1 Summary 4-3
4.2.2 General Discussion 4-6
4.2.3 Design and Operating Considerations 4-8
4.2.4 Acid Plant Performance Characteristics 4-13
4.3 SCRUBBING SYSTEMS 4-20
4.3.1 Background 4-20
4.3.2 Calcium-Based Scrubbing Systems 4-22
4.3.3 Ammonia-Based Scrubbing Systems 4-44
4.3.4 Magnesium-Based Scrubbing Systems 4-58
4.3.5 Citrate Scrubbing Processes 4-68
4.3.6 Conclusions Regarding Flue Gas Desulfurization
Systems 4-84
4.4 INCREASING THE S02 STRENGTH OF REVERBERATORY FURNACE
OFFGASES 4-90
4.4.1 Elimination of Converter Slag Return 4-91
4.4.2 Minimizing Infiltration 4-92
4.4.3 Preheating Combustion Air 4-93
4.4.4 Operation at Lower Air-to-Fuel Ratio 4-94
4.4.5 Predrying Wet Charge 4-95
4.4.6 Oxygen Enhancement Techniques 4-95
4.4.7 Summary of Operating Modifications Useful for
Increasing Offgas SOo Concentrations 4-117
4.5 GAS BLENDING 4-120
4.5.1 Converter Scheduling as a Means of Facilitating
Gas Blending 4-120
4.5.2 Weak-Stream Blending as Applied to a New Smelter
that Processes High-Impurity Ore Concentrates . . . 4-120
-------
TABLE OF CONTENTS (con.)
4.5.3 Partial Weak-Stream Blending as Applied to
Existing Smelters 4-121
4.6 PARTICULATE MATTER CONTROL FOR REVERBERATORY FURNACES .... 4-123
4.6.1 Important Factors Governing the Specification
of a Particulate Control Device for Reverbera-
tory Furnace Offgases 4-123
4.6.2 Venturi Scrubbers 4" n
4.6.3 Fabric Filters 4'130
4.6.4 Electrostatic Precipitators 4-136
4.6.5 Conclusions Regarding Particulate Removal From
Reberberatory Furnace Offgases 4-143
4 7 CONTROL OF FUGITIVE EMISSIONS FROM PRIMARY COPPER
SMELTERS
4.7.1 General
4.7.2 Local Ventilation 4-14b
4.7.3 General Ventilation • 4~149
4.7.4 Control of Fugitive Emissions From Roasting
Operations • 4-150
4.7.5 Control of Fugitive Emissions From Smelting
Furnace Operations 4-153
4.7.6 Capture of Fugitive Emissions From Converter
Operations 4-161
4.7.7 Summary of Visible Emissions Data for
Fugitive Emissions Sources 4-181
478 Removal of Particulate Matter From Fugitive
Gases 4-193
4.8 REFERENCES 4~197
5. MODIFICATIONS AND RECONSTRUCTION S"1
5.1 SUMMARY OF 40 CFR 60 PROVISIONS FOR MODIFICATION AND
RECONSTRUCTION f"l
5.1.1 Modification 5"J
5.1.2 Reconstruction 5"2
5.2 APPLICABILITY TO PRIMARY COPPER SMELTERS 5-3
5.2.1 General 5-3
5.2.2 Modifications ^
5.3 REFERENCES 5'9
6. MODEL PLANTS AND ALTERNATE CONTROL TECHNOLOGIES 6-1
6.1 INTRODUCTION 6~J
6.2 REVERBERATORY FURNACE EXEMPTION 6-2
6.3 FUGITIVE EMISSION CONTROL 6-17
6.4 EXPANSION OPTIONS AND ALTERNATIVE CONTROL TECHNOLOGIES. . . . 6-22
6.5 REFERENCES 6"36
vii
-------
TABLE OF CONTENTS (con.)
7. ENVIRONMENTAL IMPACT 7-1
7.1 GENERAL 7-1
7.1.1 New Greenfield High-Impurity Smelters--
Process Emissions 7-1
7.1.2 New Greenfield High-Impurity Smelters—
Fugitive Emissions 7-3
7.2 AIR POLLUTION IMPACT 7-3
7.2.1 S02 Controls for Reverberatory Smelting Furnaces. . . 7-3
7.2.2 Fugitive Particulate Emissions 7-7
7.2.3 Expansion Scenarios 7-7
7.3 WATER POLLUTION IMPACT 7-9
7.3.1 Gas Cleaning and Conditioning Systems 7-11
7.3.2 FGD Absorbent Purges 7-11
7.4 SOLID WASTE IMPACT 7-16
7.4.1 Calcium Based FGD's 7-17
7.4.2 Gas Cleaning Purges 7-17
7.4.3 Particulate Control on Reverberatory Smelting
Furnaces 7-18
7.5 ENERGY IMPACT 7-20
7.5.1 New Greenfield Smelters—Process Emissions 7-20
7.5.2 New Greenfield Smelters—Fugitive Emissions 7-20
7.5.3 Expansion Scenarios 7-20
8. COSTS 8-1
8.1 INTRODUCTION 8-1
8.2 CONTROL OF WEAK S02 STREAMS FROM NEW REVERBERATORY
FURNACES 8-3
8.2.1 Capital Costs 8-5
8.2.2 Annual ized Costs 8-17
8.3 COSTS FOR FUGITIVE EMISSION CONTROL 8-29
8.3.1 Capital Costs 8-29
8.3.2 Annual ized Costs 8-33
8.4 COST OF CONTROLLING PROCESS PARTICULATE EMISSIONS
FROM REVERBERATORY FURNACES IF THE REVERBERATORY
EXEMPTION IS RETAINED 8-35
8.4.1 Capital Costs 8-35
8.4.2 Annual ized Costs 8-36
8.5 PROCESS COSTS 8-38
8.5.1 Capital Costs 8-38
8.5.2 Annual ized Costs 8-38
8.6 EXPANSION SCENARIOS 8-38
8.6.1 Incremental Capital and Annualized Process
Costs for Expansion Scenarios 8-40
8.6.2 Incremental Capital and Annualized Costs
for Control 8-46
8.6.3 Summary of Expansion Scenario Incremental Costs . . . 8-50
viii
-------
Page
TABLE OF CONTENTS (con.)
8.7 COST-EFFECTIVENESS 8-50
8.8 REFERENCES 8'59
9. ECONOMIC IMPACT 9-1
9.1 INDUSTRY ECONOMIC PROFILE 9-1
9.1.1 Introduction 9-1
9.1.2 The Copper Smelters—Ownership, Location,
Concentration 9-2
9.1.3 The Copper Refiners 9-7
9.1.4 Domestic Supply 9-9
9.1.5 Flow of Copper from Mines to U.S. Smelters 9-11
9.1.6 Copper Production Costs 9-17
9.1.7 U.S. Copper Resources 9-21
9.1.8 Smelter Capacity Growth 9-24
9.1.9 Trends in U.S. Productivity 9-26
9.1.10 U.S. Total Consumption of Copper 9-29
9.1.11 Demand by End Use 9-29
9.1.12 Copper Prices 9-33
9.1.13 Substitutes 9-44
9.1.14 World Production and Consumption of Copper 9-45
9.2 ECONOMIC IMPACT ASSESSMENT 9-48
9.2.1 Introduction 9-48
9.2.2 Methodology of Impact Analysis 9-49
9.2.3 Price Elasticity of Supply 9-53
9.2.4 The Price Elasticity of Demand 9-55
9.2.5 Analysis 9-57
9.2.6 Findings 9-71
9.3 SOCIOECONOMIC IMPACT ASSESSMENT 9-76
9.3.1 Executive Order 12291 9-76
9.3.2 Regulatory Flexibility 9-79
9.4 REFERENCES 9-79
APPENDIX A EVOLUTION OF THE BACKGROUND INFORMATION DOCUMENT A-l
APPENDIX B INDEX TO ENVIRONMENTAL IMPACT CONSIDERATIONS B-l
APPENDIX C EMISSION SOURCE TEST DATA C-l
APPENDIX D (Not Used)
APPENDIX E USE OF COAL IN THE OUTOKUMPU FLASH FURNACE AT THE
TOYO SMELTER E-l
APPENDIX F COST ANALYSIS TO ESTIMATE THE INCREMENTAL INCREASE IN
CAPITAL COST INCURRED BY INCREASING SULFURIC ACID PLANT
GAS-TO-GAS HEAT EXCHANGE CAPACITY F-l
TX
-------
TABLE OF CONTENTS (con.)
APPENDIX G ANALYSIS OF CONTINUOUS S02 MONITOR DATA AND
DETERMINATION OF AN UPPER LIMIT FOR THE INCREASE IN
S02 EMISSIONS DUE TO SULFURIC ACID PLANT CATALYST
DETERIORATION G-1
APPENDIX H SULFUR DIOXIDE EMISSION TEST RESULTS FOR SINGLE-STAGE
ABSORPTION SULFURIC ACID PLANTS PROCESSING METALLURGICAL
OFFGAS STREAMS FROM PRIMARY COPPER SMELTERS H-l
APPENDIX I ANALYSIS OF DUAL-ABSORPTION SULFURIC ACID PLANT
CONTINUOUS S02 MONITORING DATA 1-1
APPENDIX J EXAMPLE CALCULATIONS MODEL PLANT OPERATING PARAMETERS ... J-l
APPENDIX K MATHEMATICAL MODEL FOR ESTIMATING POSTEXPANSION
REVERBERATORY GAS FLOW AND S02 CONCENTRATION FOR OXYGEN
ENRICHMENT AND OXY-FUEL EXPANSION OPTIONS K-l
APPENDIX L METHODOLOGY FOR ESTIMATING SOLID AND LIQUID WASTE
DISPOSAL REQUIREMENTS L-l
APPENDIX M DETAILED COSTS FOR GREENFIELD SMELTERS M-l
APPENDIX N FUGITIVE EMISSION CONTROL COSTS N-l
APPENDIX 0 DETAILED COSTS FOR EXPANSION SCENARIOS 0-1
APPENDIX P METHODOLOGY UTILIZED TO DETERMINE THE COSTS ASSOCIATED
WITH SULFURIC ACID PLANT PREHEATER OPERATION P-l
-------
Number
FIGURES
Page
3-1 The conventional copper smelting process 3-5
3-2 Types of roasters ^
3-3 Reverberatory smelting furnace ^~JJ
3-4 Electric smelting furnace a-i»
3-5 Inco flash smelting furnace 6"^
3-6 Outokumpu flash smelting furnace *''?
3-7 Peirce-Smith Converter *"*"
3-8 Copper converter operation ^"^
3-9 Hoboken converter f"£?
3-10 Noranda continuous smelting ;T^
3-11 Mitsubishi continuous smelting • • • • • *-«
3-12 Fugitive emissions sources for primary copper smelters. . . J-^e
3-13 Methods of oxygen addition • • • 3"69
3-14 Converter elimination of arsenic as a function of
matte grade '•*'*•
3-15 Converter elimination of antimony as a function of
matte grade •
3-16 Converter elimination of bismuth as a function of
matte grade
4-1 Contact sulfuric acid processes 4-7
4-2 Calcium-based scrubbing processes • • • ^-^4
4_3 Effect of pH of calcium sulfite-bisulfite solution on S02
equilibrium vapor pressure 4~29
4-4 Flow diagram of the lime/gypsum plant at the Onahama
smelter 4"38
4-5 Ammonia scrubbing process with sulfuric acid
acidulation • • •„
4-6 Ammonia scrubbing process with ammonium bisulfite
acidulation ^ -
4-7 Magnesium oxide (MAGOX) scrubbing process 4-bU
4-8 Bureau of Mines citrate scrubbing process 4-/1
4-9 Flakt-Boliden citrate scrubbing process 4-/J
4-10 Typical absorber configuration £-88
4-11 • Methods of oxygen addition 4~y/
4-12 Conventional copper reverberatory smelting furnace that
has been converted to an oxygen sprinkle smelting
furnace c' ' ' V V '
4-13 Oxy-fuel burner locations in Reverberatory Furnace No. 3
at the Caletones smelter 4-102
4-14 Plan and elevation of Reverberatory Furnace No. 3
XI
-------
FIGURES (con.)
Number Page
4-15 Reverberatory furnace temperatures in the vicinity of
the furnace roofs with and without oxygen-undershooting
at Inco smelter 4-115
4-16 Typical collection efficiency curves for several types
of particulate removal devices 4-124
4-17 Venturi scrubber 4-129
4-18 Typical relationship between fractional collection
efficiency and particle size for venturi scrubbers 4-131
4-19 Baghouse with mechanical shaking 4-133
4-20 Baghouse with reverse flow cleaning 4-134
4-21 Baghouse with cleaning by jet pulse 4-134
4-22 Electrostatic precipitator 4-139
4-23 Illustration of null point formation 4-148
4-24 Spring-loaded car top and ventilation hood,
ASARCO-Hayden 4-152
4-25 Typical hooding for a matte tapping port 4-155
4-26 Schematic of a typical fugitive emissions control system
for matte tapping operations 4-156
4-27 Typical sectional launder covers 4-157
4-28 Launder hoods utilized at the Phelps Dodge-Morenci
Smelter for the capture of fugitive emissions generated
during matte tapping operations 4-158
4-29 Schematic of the matte tapping and ladle hoods at the
ASARCO-Tacoma Smelter 4-160
4-30 Schematic of the slag skimming (plan view) fugitive
emissions control system at the ASARCO-Tacoma Smelter . . . 4-162
4-31 Controlled airflow from a heated source 4-164
4-32 Uncontrolled airflow from a heated source 4-164
4-33 Inlet-outlet openings in converter building at. ASARCO-
El Paso 4-167
4-34 A typical fixed secondary converter hood 4-171
4-35 Retractable-type secondary hood as employed at ASARCO-
Hayden 4-172
4-36 Entrained flow diagram 4-175
4-37 Converter air curtain/secondary hooding system as employed
at the Onahama and Naoshima smelters 4-176
4-38 Schematic diagram of the converter housing/air curtain
system at the Tamano smelter 4-178
6-1 Model plant for new "greenfield" smelter processing
high-impurity materials 6-1
6-2 Model smelter converter operating schedule 6-8
6-3 Model Plant I for expansion of existing smelters 6-28
6-4 Model Plant II for expansion of existing smelters 6-29
6-5 Model Plant III for expansion of existing smelters 6-30
6-6 Model Plant IV for expansion of existing smelters 6-31
6-7 Model Plan V for expansion of existing smelters 6-32
-------
FIGURES (con.)
Page
Number
8-1 Capital cost of a DC/DA sulfuric acid plant 8-6
8-2 Capital cost of an MgO FGD system °
^ .. i ___j_ ^ .e — — ~H*mvii-i-i ^C^nex/ctom _ O J-£
8-3 Capital cost of an ammonia FGD system
8-4 Capital cost of a limestone FGD system . . ........ »
8-5 Capital cost of an SC/SA sulfuric acid plant ....... 8
9-1 Principal mining States and copper smelting and ^
refining plants, 1978 ................... _
9-2 U.S. sources and uses of copper ...... • • • .....
9-3 Comparison of copper price index and mine and mill ^_^
capital cost index ..................... ~yr
9-4 U.S. copper smelter production ...............
9-5 Quarterly price movements for copper wirebars
(1973 to 1981) ....................... gI3?
9-5 u s copper price ............... . • • • ' ' '
9-7 Annual recoverable copper available from domestic deposits
over a copper price range of $1.10 to $1.30/kg .......
9-8 Costs for smelting and refining in Japan vs. costs at
new smelters in the United States .............
9-9 Costs for smelting and refining in Japan vs. costs
at expanding smelters in the United States ........ »
xm
-------
Number
TABLES
Page
1-1 Expansion Scenarios Selected for Economic Analysis. ... 1-4
1-2 Impacts of S02 Regulatory Alternatives of a Typical
New Greenfield Smelter (Multihearth Roaster,
Reverberatory Smelting Furnace, Converter) Processing
High-Impurity Materials (All Impacts are Long Term
Unless Otherwise Noted)
1-3 Impacts of Particulate Matter Regulatory Alterna-
tives of a Typical New Greenfield Smelter (Multi-
hearth Roaster, Reverberatory Smelting Furnace,
Converter) Processing High-Impurity Materials
(All Impacts are Long Term Unless Otherwise Noted). ... i-b
3-1 Domestic Primary Copper Smelters 3-2
3-2 Major Copper-Bearing Minerals 6~tL
3-3 Emissions Factors for Uncontrolled Major Process
Sources f,"^
3-4 Potential Sources of Fugitive Emissions d-*/
3-5 Summary of Fugitive S0? Emissions Factors for Primary
Copper Smelting Operations 3~58
3-6 Summary of Fugitive Particulate Emissions Factors for
Primary Copper Smelting Operations 3~by
3-7 Maximum Acceptable Impurity Levels in Anode Copper, and
Corresponding Levels in Blister Copper Produced at the
ASARCO-Tacoma Smelter 3"86
3-8 Assays of Various High Impurity Materials Processed at
ASARCO-Tacoma •.-;••••
3-9 Distribution of Impurity Elements in Conventional
Smelting When Processing High-Impurity Feeds 3-90
3-10 Distribution of Impurity Elements in the Noranda
Process (Matte Production Mode) 3~95
3-11 Distribution of Impurity Elements in the Noranda
Process (Blister Copper Production Mode) 3-9b
3-12 Impurity Assays of Feed Materials Processed in the
Outokumpu Flash Furnace at the Kosaka Smelter 3-101
3-13 Maximum Impurity Levels Recommended for the Outokumpu
Flash Furnace 3"102
3-14 Range of Impurity Concentrations Tested in the Inco
Miniplant Flash Furnace • 3~105
3-15 Maximum Impurity Levels Processed in the Mitsubishi
Process 3"106
3-16 Maximum Impurity Levels Recommended for the Noranda
Process (Matte Production Mode) 3-108
XV
-------
TABLES (con.)
Page
Summary of Experience Processing High-Impurity Feeds
in Alternative Smelting Technologies 3-110
3-18 Sulfur/Sulfur Dioxide Emission Limitations by State . . . 3-113
3-19 Particulate Emission Limitations by State 3-115
4-1 Estimated Maximum Impurity Limits for Metallurgical
Offgases Used to Manufacture Sulfuric Acid 4-15
4-2 Composition of Scale From the Onahama Lime-Gypsum
Process 4-31
4-3 Major Domestic Utility-Related FGD Installations That
Use the Limestone-Scrubbing Process 4-33
4-4 Lime/Limestone FGD Systems That Have Achieved S02
Removal Efficiences of 90 Percent or Greater on
Offgases Generated by Coal-Fired Steam Generators .... 4-36
4-5 Summary of Emission Test Data for the Duval Sierrita
Lime Scrubbing System, 1977-1980 4-37
4-6 Performance Data on the Cominco-Type Ammonia-Based
Scrubbing Units at Trail, British Columbia 4-56
4-7 Flue Gas Desulfurization Processes Assessed for
Application to Reverberatory Furnace Offgases 4-85
4-8 Efficiency and Reliability Data for the FGD Processes
Being Considered in the NSPS Revision for Primary
Copper Smelters 4-86
4-9 General Specifications of the Type of Oxy-Fuel Burner
Employed at the Caletones Smelter 4-104
4-10 General Specifications of the Type of Oxy-Fuel Burner
Employed at the Onahama Smelter 4-106
4-11 Typical Reverberatory Furnace Operating Data Before
and After the Use of Oxy-Fuel Burners at the Onahama
Smelter 4-107
4-12 Summary of Experience Involving the Use
of Oxygen in Reverberatory Smelting Furnaces 4-118
4-13 Typical Fractional Collection Efficiencies of
Particulate Control Equipment 4-125
4-14 Summary of Particulate Test Data for the Spray
Chamber/Baghouse at the Anaconda Smelter 4-137
4-15 Summary of In-Stack/Out-of-Stack Particulate Matter
Test Results at Reverberatory Furnace ESP Outlets .... 4-142
4-16 Summary of Particulate Test Data for the Spray
Chamber/Roaster-Reverberatory ESP at the ASARCO-
El Paso Smelter 4-144
4-17 Function of Air Curtain and Secondary Hood System
During Various Modes of Converter Operation at Tamano
Smelter 4-179
4-18 Summary of Design Data for the ASARCO-Tacoma
Converter Secondary Hooding/Air Curtain System 4-182
xvi
-------
TABLES (con.)
Summary of Visible Emission Observation Data for
Capture Systems on Fugitive Emission Sources at
ASARCO-Tacoma
4-20 Visible Emission Observation Data for Reverberatory
Furnace Matte Tapping Operations at the Phelps Dodge-
Morenci Smelter 4"18b
4-21 Visible Emission Data for Reverberatory Furnace
Matte Tapping Operations at the Phelps Dodge-
Morenci Smelter
4-22 Visible Emission Observation Data for Reverberatory
Furnace Slag Skimming Operations at the Phelps Dodge-
Morenci Smelter 4"lbb
4-23 Visible Emission Observation Data for Converter
Secondary Hood System During Matte Charging at the
Tamano Smelter
4-24 Visible Emission Observation Data for Blister
Discharge at the Tamano Smelter 4-194
4-25 Summary of Emissions Testing Performed on the
Converter Building Evacuation Baghouse at ASARCO-
EI Paso ;;••'•'
4-26 Summary of Emissions Testing Performed on the Calcine
Discharge Baghouse at Phelps Dodge-Douglas 4-19b
6-1 Model Plant Charge Composition and Sulfur Elimination
for Greenfield High-Impurity Smelter °~b
6-2 Model Plant—Greenfield High-Impurity Smelter Repre-
sentative Converter Offgas Stream Profile 6-10
6-3 Model Plant, New Greenfield High-Impurity Smelter
Control Alternatives 6~1£
6-4 Parameters for Particulate Control Alternatives--
Primary Offgases from Dirty Reverberatory Furnaces. . . . 6-18
6-5 Summary of Fugitive Particulate Emissions Capture
and Control Systems °"^
6-6 Smelting Configuration/Expansion Scenarios °"<^
6-7 Model Plant Configurations and Existing U.S. Smelters . . 6-26
6-8 Model Plant Expansion Scenarios: Exit Gases,
Composition, and Flow Rate 6-33
6-9 Model Plants for Expansion Options: Representative
Feeds, Matte Grades, and Sulfur Elimination Rates .... 6-35
7-1 Evaluated Control Options for Control of Process S02
Emissions at a Greenfield Copper Smelter (Multihearth
Roaster-Reverberatory Smelting Furance-Converter)
Processing High-Impurity Materials '~2
xv n
-------
TABLES (con.)
Number Page
7-2 Evaluated Alternatives for Control of Fugitive Particulate
Emissions at a Greenfield Copper Smelter Processing
High-Impurity Materials (Multihearth Roaster-
Reverberatory Smelting Furnace-Converter) 7-4
7-3 Evaluated Alternatives for Control of Fugitive
Particulate Emissions at a Greenfield Copper Smelter
(Flash Furnace-Converter) 7-5
7-4 Air Pollution Emission Impact of S02 Control Alter-
natives for a New Greenfield Smelter, Multihearth
Roaster-Reverberatory Furnace-Converter 7-6
7-5 Air Pollution Fugitive Particulate Emission Impact
for Each Source and Control Alternatives—New
Greenfield Smelters 7-8
7-6 Air Pollution Fugitive Particulate Emission Impact for
Expansion at Existing Smelters 7-10
7-7 Estimated Production Rate of Solid and Liquid
Effluents Requiring Disposal From Gas Cleaning and
Conditioning Equipment, Greenfield Smelters 7-12
7-8 Estimated Incremental Increase in Effluents
Requiring Disposal From Gas Cleaning and Conditioning
Equipment, Expansion Options 7-13
7-9 Estimated Production Rate of Solid and Liquid
Effluents Requiring Disposal from FGD Systems
Associated With Greenfield Smelter Models 7-14
7-10 Estimated Production Rate of Solid and Liquid
Effluents Requiring Disposal from FGD Systems
Associated With Expansion Options , 7-15
7-11 Estimate of Emission Reduction Due to Particulate
Control of Reverberatory Smelting Furnace Primary
Offgases--High-Impurity Greenfield Smelter 7-19
7-12 Energy Impact—Process S02 Control Alternatives for
New Greenfield Smelter, Multihearth Roaster-
Reverberatory Furnace-Converter 7~21
7-13 Incremental Energy Impact—Fugitive Emission Control
Alternatives for New Greenfield Smelters 7-22
7-14 Energy Impacts—Expansion Scenarios for Existing
Primary Copper Smelters 7-23
8-1 Control Alternatives 8-2
8-2 Input Data to Cost Estimation, New High-Impurity
Smelter 8-4
8-3 Labor and Utility Unit Costs 8-18
8-4 FGD Raw Material and Utility Usage Rate 8-22
8-5 Evaluated Alternatives for Control of Fugitive
Particulate Emissions from a New Copper Smelter (Multi-
hearth Roaster, Reverberatory Furnace, Converter or
Flash Furnace-Converter) Processing High-Impurity
Materials 8-30
xvm
-------
TABLES (con.)
Number —^-
8-6 Model Plant Expansion Scenarios • 8-39
8-7 Input Data to Cost Estimations, Expansions Options. . . . 8-41
8-8 Summary of Incremental Costs Incurred Due to Acid
Plant Preheater Operation 8~bl
8-9 Expansion Costs (Includes Cost of Controlling S02
Emissions from New Roasters and Converters as Required
by Existing NSPS) 8'52
8-10 Cost-Effectiveness: Control of Reverberatory Furnace
S02 Emissions in a New Copper Smelter (Multihearth
Roasters, Reverberatory Furnace, Converter) Processing
High-Impurity Materials 8"53
8-11 Costs for Control of Fugitive Particulate Matter
Emissions by Source, New Greenfield Smelter 8-54
8-12 Cost-Effectiveness of Expansion Scenairos 8-55
8-13 Cost-Effectiveness, Fugitive Particulate Matter Control,
Expansion Scenarios -. • • • 8-56
8-14 Incremental Cost Data, Least Cost Expansion Scenarios . . b-b/
8-15 Incremental Cost Data, Fugitive Emission Control
Least Cost Expansion Scenarios 8-58
9-1 Smelter Ownership, Production and Source Material
Arrangements 9-5
9-2 U.S. Refining Facilities for Primary Copper 9-8
9-3 Flow of Copper From Mines to U.S. Smelters,
Mine Output 9-12
9-4 Flow of Copper From Mines to U.S. Smelters,
Smelter Sources 9-14
9-5 Smelting Cost Estimates 9-20
9-6 U.S. Copper Production by Mine (1977), Cents per
Kilogram and Production Capacity 9-22
9-7 Copper Resources of U.S. Companies 9-23
9-8 Productivity in the Copper Industry 9-27
9-9 Output and Productivity Indices 9-28
9-10 U.S. Copper Consumption 9-30
9-11 U.S. Copper Demand by Market End Uses 9-32
9-12 U.S. Shipments of Copper-Base Mill and Foundry
Products—Gross Weight 9-34
9-13 U.S. Copper Mine Capacity: Current and Potential .... 9-42
9-14 United States and World Comparative Trends in Refined
Copper Consumption, 1963-1979 9-46
9-15 United States and World Comparative Trends in
Copper Production: 1963-1979 9-47
9-16 Price Elasticity of Supply Estimates 9-54
9-17 Price and Income Elasticities of Demand Estimates .... 9-56
9-18 Cost Data for New High Impurity Greenfield Smelters . . . 9-58
9-19 Cost Data for New Greenfield Smelter Processing
Clean Concentrates Using a Flash Furnace 9-59
9-20 Smelter Cost Data for Expansion Scenarios 9-61
9-21 Maximum Percentage Price Increase 9-72
xix
-------
TABLES (con.)
Number Page
9-22 Maximum Percent Profit Reduction 9-74
9-23 Summary of Selected Cases 9-75
9-24 Number of Employees at Companies That Own Primary
Copper Smelters 9-78
xx
-------
1. SUMMARY
1.1 REGULATORY ALTERNATIVES
In response to petitions filed by the National Resources Defense
Council (NRDC) and the American Smelting and Refining Company (ASARCO)
after the existing standard for primary copper smelters was promulgated
January 15, 1975 (41 FR 2338), the U.S. Environmental Protection
Agency (EPA) entered into negotiations with NRDC and ASARCO that led
to a court-approved settlement of the petitions. Under the terms of
the settlement, EPA would review the standard and make whatever changes
were considered appropriate. This review coincided with the periodic
review of established standards required by the 1977 Clean Air Act
Amendments.
The review focused on four areas: (1) reexamination of the
current exemption for reverberatory furnaces processing high-impurity
materials, (2) reassessment of the feasibility of controlling particu-
late emissions from reverberatory furnaces processing high-impurity
materials, (3) revaluation of the impact of the existing standard on
the ability of existing smelters to expand production, and (4) assess-
ment of the technical and economic feasibility of controlling fugitive
emissions at primary copper smelters.
1.1.1 Reverberatory Smelting Furnace Exemption
The following seven alternatives were considered for controlling
S02 emissions from reverberatory smelting furnaces if the exemptions
while processing high-impurity materials were withdrawn:
I-A Partial blending of 45 percent of the reverberatory smelting
furnace stream with multihearth roaster and converter strong
streams followed by treatment in a double contact/double
absorption (DC/DA) sulfuric acid plant.
1-1
-------
I-B Treatment of the reverberatory furnace stream in a magnesium
oxide flue gas desulfurization (FGD) system, strong FGD S02
stream to a DC/DA sulfuric acid plant.
I-C Same as I-B except NH3 FGD system.
I-D Treatment of the reverberatory smelting furnace stream in a
limestone FGD system.
I-E Blending 100 percent of the reverberatory smelting furnace
stream with multihearth roaster and converter streams and
treatment in a DC/DA sulfuric acid plant.
I-F Blending of the oxygen-enriched reverberatory smelting
furnace stream with multihearth roaster and converter streams
and treatment in a DC/DA sulfuric acid plant.
I-G Blending of the oxy-fuel reverberatory smelting furnace
stream with multihearth roaster and converter streams and
treatment in a DC/DA sulfuric acid plant.
Control Alternative I-G, determined to be both the most cost-effective
and the most efficient, was subjected to the economic analysis.
1.1.2 Control of Reverberatory Furnace Participate Matter Emissions
Two alternatives were considered for controlling participate
matter emissions from reverberatory smelting furnaces if the exemption
while processing high-impurity materials is retained: (1) gas cooling
followed by fabric filtration and (2) gas cooling followed by treat-
ment by an electrostatic precipitator (ESP).
1.1.3 Expansion
Twenty-four expansion scenarios were analyzed to determine the
impact of the existing standard on the ability of existing smelters to
expand production capacity. The analysis focused on the reverberatory
smelting furnace because it is the rate-determining limitation. Each
scenario consists of a smelter configuration, an expansion option, and
a control alternative. The smelter configurations are representative
of the existing domestic smelter. Expansion options include oxygen
enrichment of reverberatory combustion air, retrofitting the reverbera-
tory furnace with oxy-fuel burners, conversion from green to calcine
charging, and replacement of the reverberatory smelting furnace with a
flash smelting furnace. Control alternatives, designed to reduce
1-2
-------
postexpansion reverberatory smelting furnace weak S02 stream emissions
to preexpansion levels include partial blending with converter and/or
roaster strong S02 streams followed by treatment in an acid plant, and
treatment of a slipstream from the reverberatory smelting furnace in
an FGD.
The scenarios shown in Table 1-1 were determined to be the most
cost-effective and were included in the economic analysis.
1.1.4 Fugitive Emissions
Control alternatives for capturing fugitive particulate matter
emissions from primary copper smelters were studied: larry car inter-
lock ventilation systems for multihearth roasters; local hooding at
slag skim and matte tap ports; ladle hoods and launder covers for
smelting furnaces; and air curtain/secondary hood or building evacuation
for converters. Particulate matter would be removed from captured
streams by fabric filtration.
1.2 IMPACTS
In matrix form, Tables 1-2, 1-3, 1-4, and 1-5 show environmental,
energy, and economic impacts of the control alternative described in
Section 1.1, above, for the reverberatory smelting furnace exemption,
for control of particulate matter, for expansion capability, and for
control of fugitive emissions, respectively.
1-3
-------
TABLE 1-1. EXPANSION SCENARIOS SELECTED FOR ECONOMIC ANALYSIS
Smelter configuration
Expansion option
Control alternative
weak S02 streams
Capacity
increase (%)
Low-impurity feed
Multihearth roaster, reverberatory
furnace, converter
Multihearth roaster, reverberatory
furnace, converter
Multihearth roaster, reverberatory
furnace, converter
Reverberatory furnace, converter
Reverberatory furnace, converter
Reverberatory furnace, converter
Fluid-bed roaster, reverberatory
furnace, converter
Fluid-bed roaster, reverberatory
furnace, converter
Electric furnace, converter
Flash furnace, converter
High-impurity feed
Multihearth roaster, reverberatory
furnace, converter
Oxygen enrichment
Conversion to flash smelting None
Conversion to flash smelting Nonec
Oxygen enrichment
Conversion to flash smelting
Conversion to flash smelting
Oxygen enrichment
Conversion to calcine charge
Oxygen enrichment
Oxygen enrichment
Partial blending and
acid plant
a
Conversion to flash smelting None
Partial blending and
acid plant
None3
None8
Partial blending and
acid plant
a
None
Nonec
Partial blending and
acid plant
20
50
100
20
50
100
20
60
40
20
20
Post expansion smelting furnace offgases are strong streams.
Post expansion smelting furnace S02 emissions are less than preexpansion.
-------
TABLE 1-2. IMPACTS OF S02 REGULATORY ALTERNATIVES OF A TYPICAL
NEW GREENFIELD SMELTER (MULTIHEARTH ROASTER, REVERBERATORY
SMELTING FURNACE, CONVERTER) PROCESSING HIGH-IMPURITY
MATERIALS (ALL IMPACTS ARE LONG TERM UNLESS OTHERWISE
NOTED)3
Regulatory Alternative
based on:
Control
Control
Control
Control
Control
Control
Control
Alternative
Alternative
Alternative
Alternative
Alternative
Alternative
Alternative
I-A
I-B
I-C
I-D
I-E
I-F
I-G
No changes to standard
Air
impact
+2
+3
+3
+3
+3
+3
+3
0
Water
impact
-1
-2
-1
-2
-1
-1
-1
0
Solid
waste
impact
-1
-2
-1
-2
-1
-1
-1
0
Energy
impact
-1
-1
-1
-1
-1
-1
+1
0
Economic
impact
N.
N.
N.
N.
N.
N.
P.
P.
P.
P.
P.
P.
-1
0
If exemption of reverberatory smelting furnace while processing high-
impurity materials is deleted.
Key: + Beneficial impact
- Adverse impact
0 No impact
1 Negligible impact
2 Small impact
3 Moderate impact
4 Large impact
N.P. Analysis not performed
1-5
-------
TABLE 1-3. IMPACTS OF PARTICIPATE MATTER REGULATORY ALTERNATIVES OF A
TYPICAL NEW GREENFIELD SMELTER (MULTIHEARTH ROASTER, REVERBERATORY
SMELTING FURNACE, CONVERTER) PROCESSING HIGH-IMPURITY MATERIALS
(ALL IMPACTS ARE LONG TERM UNLESS OTHERWISE NOTED)a
Regulatory Alternative
based on:
Gas cooling, fabric
filtration
Gas cooling, ESP
No change to standard
Air
impact
+2
+2
Ci
Water
impact
-1
-1
0
Solid
waste
impact
-1
-1
0
Energy
impact
-1
-1
0
Economic
impact
-1
-1
0
If exemption of reverberatory smelting furnace while processing high-
impurity material is retained.
Key: + Beneficial impact
- Adverse impact
0 No impact
1 Negligible impact
2 Small impact
1-6
-------
2. INTRODUCTION
2.1 BACKGROUND AND AUTHORITY FOR STANDARDS
Before standards of performance are proposed as a Federal regula-
tion, air pollution control methods available to the affected industry
and the associated costs of installing and maintaining the control
equipment are examined in detail. Various levels of control, based on
different technologies and degrees of efficiency, are expressed as
regulatory alternatives. Each of these alternatives is studied by the
U.S. Environmental Protection Agency (EPA) as a prospective basis for
a standard. The alternatives are investigated in terms of their
impacts on the economics and well-being of the industry, the impacts
on the national economy, and impacts on the environment. This document
summarizes the information obtained through these studies so interested
persons will be able to see the information considered by EPA in the
development of the proposed standards.
Standards of performance for new stationary sources are established
under Section 111 of the Clean Air Act (42 USC 7411) as amended,
herein referred to as the Act. Section 111 directs the Administrator
to establish standards of performance for any category of new station-
ary source of air pollution that ". . . causes or contributes signifi-
cantly to air pollution which may reasonably be anticipated to endanger
public health or welfare."
The Act requires that standards of performance for stationary
sources reflect "... the degree of emission reduction achievable
through the application of the best system of continuous emission
reduction which (taking into consideration the cost of achieving such
emission reduction, and any nonair quality health and environmental
impact and energy requirements) the Administrator determines has been
2-1
-------
adequately demonstrated for that category of sources." The standards
apply only to stationary sources, the construction or modification of
which commences after regulations are proposed by publication in the
Federal Register.
The 1977 amendments to the Act altered or added numerous provisions
that apply to the process of establishing standards of performance.
EPA is required to list the categories of major stationary
sources that have not already been listed and regulated
under standards of performance. Regulations must be promul-
gated for these new categories on the following schedule:
a. 25 percent of the listed categories by August 7, 1980,
b. 75 percent of the listed categories by August 7, 1981,
and
c. 100 percent of the listed categories by August 7, 1982.
A governor of a State may apply to the Administrator to add
a category not on the list or may apply to the Administrator
to have a standard of performance revised.
EPA is required to review the standards of performance every
4 years and, if appropriate, to revise them.
EPA is authorized to promulgate a standard based on design,
equipment, work practice, or operational procedures when a
standard based on emission levels is not feasible.
The term "standards of performance" is redefined, and a new
term, "technological system of continuous emission reduction,"
is defined. The new definitions clarify that the control
system must be continuous and may include a low- or nonpollut-
ing process or operation.
The time between the proposal and promulgation of a standard
under Section 111 of the Act may be extended to 6 months.
Standards of performance, by themselves, do not guarantee protec-
tion of health or welfare because they are not designed to achieve any
specific air quality levels. Rather, they are designed to reflect the
degree of emission limitation achievable through application of the
best adequately demonstrated technological system of continuous emission
reduction, considering the cost of achieving such emission reduction,
2-2
-------
any nonair-quality health and environmental impacts, and energy require-
ments.
Congress had several reasons for including these requirements.
First, standards with a degree of uniformity are needed to prevent
situations where some States may attract industries by relaxing stand-
ards relative to other States. Second, stringent standards enhance
the potential for long-term growth. Third, stringent standards may
help achieve long-term cost savings by eliminating the need for more
expensive retrofitting when pollution ceilings may be reduced in the
future. Fourth, certain types of standards for coal burning sources
can adversely affect the coal market by driving up the price of low-
sulfur coal or effectively excluding certain coals from the reserve
base because their untreated pollution potentials are high. Congress .
does not intend for New Source Performance Standards to contribute to
these problems. Fifth, the standard-setting process should create
incentives for improved technology.
Promulgation of standards of performance does not prevent State
or local agencies from adopting more stringent emission limitations
for the same sources. States are free under Section 116 of the Act to
establish even more stringent emission limits than those established
under Section 111 or those necessary to attain or maintain the National
Ambient Air Quality Standards (NAAQS) under Section 110. Thus, new
sources may in some cases be subject to limitations more stringent
than standards of performance under Section 111, and prospective
owners and operators of new sources should be aware of this possibility
in planning for such facilities.
A similar situation may arise when a major emitting facility is
to be constructed in a geographic area that falls under the prevention
of significant deterioration of air quality provisions of Part C of
the Act. These provisions require, among other things, that major
emitting facilities to be constructed in such areas are to be subject
to best available control technology. The term best available control
technology (BACT), as defined in the Act, means:
2-3
-------
... an emission limitation based on the maximum degree of
reduction of each pollutant subject to regulation under this Act
emitted from or which results from any major emitting facility,
which the permitting authority, on a case-by-case basis, taking
into account energy, environmental, and economic impacts and
other costs, determines is achievable for such facility through
application of production processes and available methods, systems,
and techniques, including fuel cleaning or treatment or innovative
fuel combustion techniques for control of each such pollutant.
In no event shall application of "best available control technol-
ogy" result in emissions of any pollutants which will exceed the
emissions allowed by any applicable standard established pursuant
to section 111 or 112 of this Act. (Section 169(3])
Where feasible, standards of performance are normally structured
in terms of numerical emission limits. However, alternative approaches
are sometimes necessary. In some cases physical measurement of emis-
sions from a new source may be impractical or exorbitantly expensive.
Section lll(h) provides that the Administrator may promulgate a design
or equipment standard in cases where it is not feasible to prescribe
or enforce a standard of performance. For example, hydrocarbon emis-
sions from storage vessels for petroleum liquids are greatest during
tank filling. The nature of the emissions—high concentrations for
short periods during filling and low concentrations for longer periods
during storage—and the configuration of storage tanks make direct
emission measurement impractical. Therefore, a more practical approach
to standards of performance for storage vessels has been equipment
specification.
In addition, Section lll(j) authorizes the Administrator to grant
waivers of compliance to permit a source to use innovative continuous
emission control technology. To grant the waiver, the Administrator
must find:
A substantial likelihood that the technology will produce
greater emission reductions than the standards require or an
equivalent reduction at lower economic, energy, or environ-
mental cost;
The proposed system has not been adequately demonstrated;
The technology will not cause or contribute to an unreasonable
risk to the public health, welfare, or safety;
2-4
-------
The governor of the State where the source is located con-
sents; and
The waiver will not prevent the attainment or maintenance of
any ambient standard. A waiver may have conditions attached
to ensure that the source will not prevent attainment of any
NAAQS. Any such condition will have the force of a perform-
ance standard. Finally, waivers have definite end dates and
may be terminated earlier if the conditions are not met or
if the system fails to perform as expected. In such a case,
the source may be given up to 3 years to meet the standards
with a mandatory progress schedule.
2.2 SELECTION OF CATEGORIES OF STATIONARY SOURCES
Section 111 of the Act directs the Administrator to list categor-
ies of stationary sources. The Administrator "... shall include a
category of sources in such list if in his judgment it causes, or
contributes significantly to, air pollution which may reasonably be
anticipated to endanger public health or welfare." Proposal and
promulgation of standards of performance are to follow.
Since passage of the Clean Air Amendments of 1970, considerable
attention has been given to the development of a system for assigning
priorities to various source categories. The approach specifies areas
of interest by considering the broad strategy of the Agency for imple-
menting the Clean Air Act. Often, these "areas" are actually pollutants
emitted by stationary sources. Source categories that emit these
pollutants are evaluated and ranked by a process involving such factors
as:
Level of emission control (if any) already required by State
regulations,
Estimated levels of control that might be required from
standards of performance for the source category,
Projections of growth and replacement of existing facilities
for the source category, and
Estimated incremental amount of air pollution that could be
prevented in a preselected future year by standards of
performance for the source category.
Sources for which new source performance standards were promulgated or
2-5
-------
under development during 1977, or earlier, were selected on these
criteria.
The Act amendments of August 1977 establish specific criteria to
be used in determining priorities for all major source categories not
yet listed by EPA. These are:
The quantity of air pollutant emissions that each such
category will emit, or will be designed to emit;
The extent to which each such pollutant may reasonably be
anticipated to endanger public health or welfare; and
The mobility and competitive nature of each such category of
sources and the consequent need for nationally applicable
new source standards of performance.
The Administrator is to promulgate standards for these categories
according to the schedule referred to earlier.
In some cases it may not be feasible to develop immediately a
standard for a source category with a high priority. This problem
might arise when a program of research is needed to develop control
techniques or because techniques for sampling and measuring emissions
may require refinement. In the development of standards, differences
in the time required to complete the necessary investigation for
different source categories must also be considered. For example,
substantially more time may be necessary if numerous pollutants must
be investigated from a single source category. Further, even late in
the development process the schedule for completion of a standard may
change. For example, inability to obtain emission data from well-con-
trolled sources in time to pursue the development process in a systema-
tic fashion may force a change i ri scheduling. Nevertheless, priority
ranking is, and will continue to be, used to establish the order in
which projects are initiated and resources assigned.
After the source category has been chosen, the types of facilities
within the source category to which the standard will apply must be
determined. A source category may have several facilities that cause
air pollution, and emissions from some of these facilities may vary
from insignificant to very expensive to control. Economic studies of
2-6
-------
the source category and of applicable control technology may show that
air pollution control is better served by applying standards to the
more severe pollution sources. For this reason, and because there is
no adequately demonstrated system for controlling emissions from
certain facilities, standards often do not apply to all facilities at
a source. For the same reasons, the standards may not apply to all
air pollutants emitted. Thus, although a source category may be
selected to be covered by a standard of performance, not all pollutants
or facilities within that source category may be covered by the stand-
ards.
2.3 PROCEDURE FOR DEVELOPMENT OF STANDARDS OF PERFORMANCE
Standards of performance must:
Realistically reflect best demonstrated control practice;
Adequately consider the cost, the nonair-quality health and
environmental impacts, and the energy requirements of such
control;
Be applicable to existing sources that are modified or
reconstructed as well as new installations; and
Meet these conditions for all variations of operating condi-
tions considered anywhere in the country.
The objective of a program for developing standards is to identify
the best technological system of continuous emission reduction that
has been adequately demonstrated. The standard-setting process involves
three principal phases of activity: information gathering, analysis
of the information, and development of the standard of performance.
During the information-gathering phase, industries are queried
through a telephone survey, letters of inquiry, and plant visits by
EPA representatives. Information is also gathered from many other
sources, and a literature search is conducted. From the knowledge
acquired about the industry, EPA selects certain plants at which
emission tests are conducted to provide reliable data that characterize
the pollutant emissions from well-controlled existing facilities.
2-7
-------
In the second phase of a project, the information about the
industry and the pollutants emitted is used in analytical studies.
Hypothetical "model plants" are defined to provide a common basis for
analysis. The model plant definitions, national pollutant emission
data, and existing State regulations governing emissions from the
source category are then used in establishing "regulatory alternatives."
These regulatory alternatives are essentially different levels of
emission control.
EPA conducts studies to determine the impact of each regulatory
alternative on the economics of the industry and on the national
economy, on the environment, and on energy consumption. From several
possibly applicable alternatives, EPA selects the single most plausible
regulatory alternative as the basis for a standard of performance for
the source category under study.
In the third phase of a project, the selected regulatory alterna-
tive is translated into a standard of performance, which, in turn, is
written in the form of a Federal regulation. The Federal regulation,
when applied to newly constructed plants, will limit emissions to the
levels indicated in the selected regulatory alternative.
As early as is practical in each standard-setting project, EPA
representatives discuss with members of the National Air Pollution
Control Techniques Advisory Committee (NAPCTAC) the possibilities of a
standard and the form it might take. Industry representatives and
other interested parties also participate in these meetings.
The information acquired in the project is summarized in the
review document. The review document, the standard, and a preamble
explaining the standard are widely circulated to the industry being
considered for control, environmental groups, other government agencies,
and offices within EPA. Through this extensive review process, the
points of view of expert reviewers are considered as changes are made
to the documentation.
A "proposal package" is assembled and sent through the offices of
EPA Assistant Administrators for concurrence before the proposed
standards are officially endorsed by the EPA Administrator. After
2-8
-------
they are approved by the Administrator, the preamble and the proposed
regulation are published in the Federal Register.
As a part of the Federal Register announcement of the proposed
standards, the public is invited to participate in the standard-setting
process. EPA invites written comments on the proposal and also holds
a public hearing to discuss the proposed standards with interested
parties. All public comments are summarized and incorporated into a
second volume of the review document. All information reviewed and
generated in studies in support of the standard of performance is
available to the public in a "docket" on file in Washington, DC.
Comments from the public are evaluated, and the standard of
performance may be altered in response to the comments.
The significant comments and EPA's position on the issues raised
are included in the "preamble" of a "promulgation package," which also
contains the draft of the final regulation. The regulation is then
subjected to another round of review and refinement until it is approved
by the EPA Administrator. After the Administrator signs the regulation,
it is published as a "final rule" in the Federal Register.
2.4 CONSIDERATION OF COSTS
Section 317 of the Act requires an economic impact assessment
with respect to any standard of performance established under Section 111
of the Act. The assessment is required to contain an analysis of:
Costs of compliance with the regulation, including the
extent to which the cost of compliance varies, depending on
the effective date of the regulation and the development of
less expensive or more efficient methods of compliance;
Potential inflationary or recessionary effects of the regula-
tion;
Effects the regulation might have on small business with
respect to competition;
Effects of the regulation on consumer costs; and
Effects of the regulation on energy use.
Section 317 also requires that the economic impact assessment be as
extensive as practicable.
2-9
-------
The economic impact of a proposed standard upon an industry is
usually addressed both in absolute terms and in terms of the control
costs that would be incurred as a result of compliance with typical,
existing State control regulations. An incremental approach is neces-
sary because both new and existing plants would be required to comply
with State regulations in the absence of a Federal standard of perform-
ance. This approach requires a detailed analysis of the economic
impact from the cost differential that would exist between a proposed
standard of performance and the typical State standard.
Air pollutant emissions may cause water pollution problems, and
captured potential air pollutants may pose a solid waste disposal
problem. The total environmental impact of an emission source must,
therefore, be analyzed and the costs determined whenever possible.
A thorough study of the profitability and price-setting mechanisms
of the industry is essential to the analysis so an accurate estimate
of potential adverse economic impacts can be made for proposed standards.
It is also essential to know the capital requirements for pollution
control systems already placed on plants so additional capital require-
ments necessitated by these Federal standards can be placed in proper
perspective. Finally, it is necessary to assess the availability of
capital to provide the additional control equipment needed to meet the
standards of performance.
2.5 CONSIDERATION OF ENVIRONMENTAL IMPACTS
Section 102(2)(C) of the National Environmental Policy Act (NEPA)
of 1969 requires Federal agencies to prepare detailed environmental
impact statements on proposals for legislation and other major Federal
actions significantly affecting the quality of the human environment.
The objective of NEPA is to build into the decisionmaking process of
Federal agencies a careful consideration of all environmental aspects
of proposed actions.
In a number of legal challenges to standards of performance for
various industries, the United States Court of Appeals for the District
of Columbia Circuit has held that environmental impact statements need
not be prepared by the Agency for proposed actions under Section 111
2-10
-------
of the Clean Air Act. Essentially, the Court of Appeals has determined
that the best system of emission reduction requires the Administrator
to take into account counterproductive environmental effects of a
proposed standard, as well as economic costs to the industry. On this
basis, therefore, the Court established a narrow exemption from NEPA
for EPA determinations under Section 111.
In addition to these judicial determinations, the Energy Supply
and Environmental Coordination Act (ESECA) of 1974 (PL-93-319) specifi-
cally exempted proposed actions under the Clean Air Act from NEPA
requirements. According to Section 7(c)(l), "No action taken under
the Clean Air Act shall be deemed a major Federal action significantly
affecting the quality of the human environment within the meaning of
the National Environmental Policy Act of 1969" (15 USC 793c[l]).
Nevertheless, the Agency has concluded that the preparation of
environmental impact statements could have beneficial effects on
certain regulatory actions. Consequently, although not legally required
to do so by section 102 (2)(C) of NEPA, EPA has adopted a policy
requiring that environmental impact statements be prepared for various
regulatory actions, including standards of performance developed under
Section 111 of the Act. This voluntary preparation of environmental
impact statements, however, in no way legally subjects the Agency to
NEPA requirements.
To implement this policy, a separate section in this document is
devoted solely to an analysis of the potential environmental impacts
associated with the proposed standards. Both adverse and beneficial
impacts in such areas as air and water pollution, increased solid
waste disposal, and increased energy consumption are discussed.
2.6 IMPACT ON EXISTING SOURCES
Section 111 of the Act defines a new source as ". . . any station-
ary source, the construction or modification of which is commenced . . ."
after the proposed standards are published. An existing source is
redefined as a new source if "modified" or reconstructed" as defined
in amendments to the general provisions of Subpart A of 40 CFR Part 60,
which were promulgated in the Federal Register on December 16, 1975
(40 FR 58416).
2-11
-------
Promulgation of a standard of performance requires States to
establish standards of performance for existing sources in the same
industry under Section 111 (d) of the Act if the standard for new
sources limits emissions of a designated pollutant (i.e., a pollutant
for which air quality criteria have not been issued under Section 108
or which has not been listed as a hazardous pollutant under Section 112).
If a State does not act, EPA must establish such standards. General
provisions outlining procedures for control of existing sources under
Section lll(d) were promulgated November 17, 1975, as Subpart B of
40 CFR Part 60 (40 FR 53340).
2.7 REVISION OF STANDARDS OF PERFORMANCE
Congress was aware that the level of air pollution control achiev-
able by any industry may improve with technological advances. Accord-
ingly, Section 111 of the Act provides that the Administrator ". . .
shall, at least every 4 years, review and, if appropriate, revise . . ."
the standards. Revisions are made to ensure that the standards continue
to reflect the best systems that become available in the future. Such
revisions will not be retroactive, but will apply to stationary sources
constructed or modified after proposal of the revised standards.
2-12
-------
3 THE PRIMARY COPPER SMELTING INDUSTRY: PROCESSES AND
POLLUTANT EMISSIONS
3.1 GENERAL
The primary copper smelting industry extracts copper from sulfide
copper ores by the pyrometallurgical process of smelting. A major
commodity metal, the refined copper ultimately produced has many uses
such as in electrical wire, electrical components, heat exchangers,
pipes, coins, and as a component of various alloys such as bronze and
brass.x
The domestic smelting industry is comprised of seven companies that
operate 15 smelters in the United States. The majority of plants are
located in the southwest near large deposits of copper ore. Arizona,
which has extensive ore deposits, has the greatest number of smelters—
a total of seven. Two smelters are located in New Mexico, and a
single smelter is located in each of the following States: Nevada,
Texas, Utah, Tennessee, Michigan, and Washington. Domestic smelters
range in production capacity from 13,600 Mg/yr (15,000 tons/yr) for the
Cities Service Company smelter in Copperhill, Tennessee, to 254,000 Mg/yr
(280,000 tons/yr) for the Kennecott Corporation smelter in Garfield,
Utah. The 15 smelters, their locations, and capacities are presented
in Table 3-1.
Raw copper ore is a natural mixture of copper-bearing minerals
and rock (gangue). The copper ores are distinguished generally as
either sulfides, oxides, or native, depending upon the copper-bearing
minerals they contain. Some 165 copper minerals are known, but only a
few are commonly found in ore deposits.2 Some of the more important
sulfide and oxide minerals from which copper is extracted are listed
with their chemical compositions in Table 3-2.
3-1
-------
TABLE 3-1. DOMESTIC PRIMARY COPPER SMELTERS
a
Company
ASARCO, Incorporated
Cities Service Company
Copper Range Company
Inspiration Consolidated
Copper Company
Kennecott Minerals Company
(SOHIO)
Magma Copper Company
Phelps Dodge Corporation
Location
El Paso, Texas
Hayden, Arizona
Tacoma, Washington
Copperhill, Tennessee
White Pine, Michigan
Miami, Arizona
Gar field, Utah
Hayden, Arizona
Hurley, New Mexico
McGill, Nevada
San Manuel , Arizona
Ajo, Arizona
Douglas, Arizona
Hidalgo, New Mexico
Morenci, Arizona
Annual capacity
Mg
91,000
182,000
91,000
13,600
52,000
136,000
254,000
71,000
73,000
45,000
181,000
64,000
115,000
163,000
191,000
Tons
100,000
200,000
100,000
15,000
57,000
150,000
280,000
78,000
80,000
50,000
200,000
70,000
127,000
179,000
210,000
Refer-
ence
4
4
4
5
6
7
7
8
7
7
7
7
7
9
9
Based on information in Reference 7 (1979 data), updated to 1982 where
.possible.
Production of "blister" copper (99 percent Cu).
TABLE 3-2. MAJOR COPPER-BEARING MINERALS
Type
Sulfide
Oxide
Mineral
Chalcopyrite
Bornite
Chalcocite
Coy/ellite
Malachite
Azurite
Chrysocol la
Cuprite
Formula
CuFeS2
Cu5FeS4
Cu2S
CuS
CuC03-Cu(OH)2
2CuC03-Cu(OH)2
CuSi03-2H20
Cu20
3-2
-------
The sulfide ores account for 85 to 95 percent of domestic copper
production,3 with the most common copper-bearing minerals contained in
these ores being chalcopyrite, bornite, and chalcocite. The oxide
minerals, which were formed from the weathering of the sulfides, are
generally found in the upper portions of the sulfide deposits. Native
copper, which consists of nearly pure metallic copper, occurs in small
amounts in most major copper deposits. However, it is found in suf-
ficient quantities to be of importance only in Michigan's upper peninsula.
Copper ores are mined from both underground and, more commonly,
open pit mines. The average tenor (copper content) of domestic ores
is low, less than 1 percent.3 Materials comprising the majority of
the ore include siliceous oxides, iron sulfides (pyrite FeS2, pyrrhotite
FeS), and various other impurity metals such as zinc, lead, arsenic,
antimony, and bismuth, which are typically in sulfide form. In addition,
the ores contain small quantities of gold and silver.
Sulfide ores have a low copper content and are not directly
smelted because of extensive energy requirements. Rather, these ores
are beneficiated at the mine. Beneficiation consists of crushing and
grinding to liberate individual mineral particles, followed by physical
separation using the froth flotation process. The product of flotation,
ore concentrates, typically contains copper, iron, and sulfur in similar
proportions (approximately 20 to 40 percent each). In comparison,
oxide ores typically are not concentrated. These ores are processed
hydrometallurgically by leaching with acid, and the dissolved copper
is recovered by chemical precipitation on scrap iron* followed by
smelting, or by solvent extraction coupled with electrowinning.
3.2 PROCESS DESCRIPTION
The pyrometallurgical processes used for extracting copper from
sulfide ore concentrates are based on the strong affinity of iron for
oxygen compared to the affinity of copper for this element. The
conventional copper smelting process, which has been in use since the
turn of the century, includes three fundamental operations:
*The precipitated copper-rich material is known as "precipitates."
3-3
-------
Roasting (optional) of the ore concentrates in the presence
of air to eliminate a portion of the sulfur. (Moisture is
eliminated coincidentally.)
Smelting of the roasted (calcines) or unroasted ore concen-
trates with fluxes to produce an iron-copper sulfide mixture
(matte) and an iron oxide slag.
Converting (oxidizing) of the matte to eliminate the remaining
iron and sulfur and yield blister copper (about 99 percent
pure copper).
Briefly, the smelting of copper concentrates, high- and low-impurity,
is accomplished by melting the charge and suitable fluxes in a smelting
furnace. Part or all of the concentrates and fluxes may receive a
partial roast before smelting to eliminate part of the sulfur and
essentially all of the moisture. In the smelting furnace, a portion
of the undesirable components combine with the fluxes and float to the
top as a slag to be skimmed off and discarded while the copper, most
of the iron and sulfur, and any contained precious metals form a
product known as matte, which collects and is drawn off from the lower
part of the furnace. The molten matte, ideally represented as a
mixture of the compounds FeS and Cu2S, is transferred to a converter,
where air blown through the matte burns off the sulfur, oxidizes the
iron for removal in a slag, and yields "blister" containing about
99 percent copper.
Typically the blister copper is fire refined in an anode furnace,
cast into "anodes," and shipped to an electrolytic refinery for further
impurity elimination. A schematic flowchart of the conventional
smelting process is presented in Figure 3-1. Offgases containing
particulates and sulfur dioxide (S02) in various concentrations are
emitted from each operation.
Four different smelting technologies are currently used by the
domestic industry: the traditional reverberatory furnace, the electric
furnace, the flash furnace, and the Noranda process. A more detailed
discussion of the operations involved in copper smelting and each of
the various technologies is presented in succeeding subsections.
3.2.1 Roasting and Drying
Currently, 7 of the 15 domestic smelters perform the roasting
step. Of the remaining smelters, five dry the charge before smelting,
3-4
-------
Ore Concentrates + Silica Fluxes
Fuel
Air
ROASTING
(Optional)
Converter Slag (4% Cu)
\
Slag
(0.
»- Offflas
Calcine
SMELTING
>• Off gas
Matte
(-40% Copper)
CONVERTING
Natural Gas -
Air-
»• Offgas
Blister Copper
(98.5+% Cu)
FIRE-REFINING
I
»»- Offgas
Anode Copper (99.5% Cu)
to Electrolytic Refinery
Figure 3-1. The conventional cooper smelting process.
3-5
-------
and three charge the ore concentrates directly to the smelting furnace.
Whether a smelter uses roasters or not is primarily determined by the
copper-to-sulfur ratio of the feed, as well as by the feed impurity
level. Drying is distinguished from roasting in that sulfur and trace
metals are not removed.
In roasting, copper sulfide ore concentrates are heated under
controlled conditions to a high temperature (but below the melting
point of the constituents) in the presence of air to fulfill two
primary objectives: (1) to dry and heat the furnace charge, which
results in considerable savings of energy in the smelting step and
increased smelting furnace throughput; and (2) to eliminate a portion
of the concentrate sulfur as S02 and oxidize a portion of the iron
sulfides to iron oxides. The latter objective leads to an increased
copper concentration in the Cu2S:FeS matte produced during smelting.
Roasting also serves to drive off a portion of volatile impurities,
especially arsenic and antimony, that are present in significant
amounts in some concentrates. This particular result is most important,
at custom smelters, which may process feeds with high impurity levels.
Numerous chemical reactions occur during roasting. Many result
in the elimination of a portion of the sulfur as S02. A large percent-
age of the emitted S02 results from reactions with iron sulfides (such
as iron pyrite, FeS2), which are present to some extent in all concen-
trates. Representative reactions include the following:
2CuFeS2 •* Cu2S + 2FeS + S
FeS2 -» FeS + S
S + Q2 -» S02
2FeS + 302 - 2FeO + 2S02.
The product of roasting is known as calcine. The exact composition
of calcine produced from a given feed composition is dependent upon the
degree of roast, i.e., the degree of sulfur removal. The degree of
roast achieved depends on the roaster temperature, the residence time,
and the air-to-concentrate ratio. Increasing the degree of roast
3-6
-------
leads to increases in the grade of matte produced in the subsequent
smelting step. However, the matte grade (and hence degree of roast)
is ultimately limited by the fact that adequate separation of copper
from iron can only be achieved if sufficient sulfur is present to
maintain all of the copper and a significant portion of the iron in
sulfide form during smelting. Other factors, discussed in succeeding
subsections, may further limit the degree of roast selected at certain
smelters. Domestic smelters generally eliminate between 15 and 50
percent of the sulfur in the charge during roasting.
Roasters in use by the industry are either of the multihearth type
or the fluid-bed type. An illustration of each is presented in Figure
3-2. Fluid-bed roasters are the more modern of the two designs.
Currently three smelters use fluid-bed roasters, and four smelters use
multihearth roasters.
3.2.1.1 Multihearth Roasters. Multihearth roasters are cylindri-
cal, refractory-lined vessels divided from top to bottom by (usually)
six or seven refractory hearths. The outer shell is made of steel and
has hinged access doors at each level. The moist concentrates enter
the roaster through an annular opening at the top, dropping to the
top-most or dryer hearth. Plows or rabble blades attached to a central,
rotating shaft and positioned directly above each hearth serve to
expose fresh surfaces of the feed to the oxidizing air. The rabble
arms are set at an angle and direct the charge alternatively to the
center of the hearth or to the periphery, where it drops through holes
to the next lower hearth. The roasted calcine is discharged through
the bottom of the roaster. The air required for the controlled oxida-
tion of the feed enters the vessel primarily through the bottom and
flows counter-currently against the descending charge. The gases exit
through a flue at the top.
Multihearth roasters are started by preheating to a temperature
at which the concentrates will be ignited by air. The temperature of
the calcines as discharged is typically 540° to 590° C (1,000° to
1,100° F).10 The principal roasting reactions are all exothermic.11
3-7
-------
Offgas
GO
I
co
Air
Fluid-Bed Roaster
Feed
Offgas
t Drying
Hearth
Hot Air
to Exhaust
Rabble
Arm
Rabble
Blade
Calcine
Natural
Gas
Multihearth Roaster
Figure 3-2. Types of roasters.
-------
Hence, if sufficient sulfur is removed, the operation is autogenous.
If the degree of sulfur removal is to be small (as in a light roast),
supplementary fuel is required.
Typically, multihearth roasters process from 180 to 360 Mg/day
(200 to 400 tons/day) of feed, although a high throughput unit process-
ing 730 Mg/day (800 tons/day) is in operation at the ASARCO-Hayden
smelter.12 Offgas flow rates from existing multihearth roasters used
in the domestic industry generally range from 400 to 480 Nm3/min
(14,000 to 16,000 scfm). The S02 concentration in the offgases from
existing U.S. multihearth roasters (after gas cleaning) varies from
0.9 to 2.4 percent, on a dry basis, depending upon the smelter. The
industry-wide average S02 concentration (weighted by the respective
plant flow rates) is 1.5 percent. The low S02 concentrations result
because substantial dilution air enters domestic units. Extensive
tests of multihearth roaster gases made in a cooperative study between
the United States and Yugoslavia show dry S02 concentrations up to
6.5 percent before gas cleaning,13 which corresponds to levels of up
to about 5.6 percent after gas cleaning. New multihearth roasters are
considered capable (conservatively) of producing 4.5 percent S02 after
gas cleaning.*
3.2.1.2 Fluid-Bed Roasters. Fluid-bed roasters are cylindrical,
refractory-lined vessels having a single diffuser plate in the bottom
containing tuyeres or bubble caps through which air is blown from the
bottom. Finely ground ore concentrates, charged continuously into the
vessel through the side, form a bed maintained in a turbulent suspension
by the air introduced—the mixture of air and solids having the flow
characteristics of a fluid. The finely ground feed is introduced
either as a slurry through a feed pipe or in relatively dry form (6 to
12 percent moisture) through a screw conveyor or drum feeder. Roasting
*Based on theoretical calculations14 that consider a unit processing
3 percent As charge. The sulfur removal considered (at 16.7 percent)
corresponds to a light roast, which is typical of the ASARCO-Tacoma
smelter. Five percent dilution air is assumed to enter the offgases
during gas cleaning.
3-9
-------
occurs as the solid particles come in contact with the fluidizing air.
Because of the large surface area of the finely ground material exposed
to the air stream, the roasting reactions occur very rapidly and the
residence time in the oxidizing atmosphere is kept short. As is the
case for multihearth roasters, the roaster operates autogenously if
sufficient sulfur is removed.
Roasted solids are discharged by overflow through a side port and
through entrainment in the offgas, which exits through the top of the
vessel. Typically, the offgas contains 75 to 90 percent of the roasted
calcine.15 Hence, cyclone collectors are an integral part of the
roasting operation.
The roasting temperature in domestic fluid bed roasters is similar
to that of multihearth roasters, typically 540° to 650° C (1,000 to
1,200° F). Due to the high chemical efficiency of the fluid bed
roasters, the feed tends to overheat from the exothermic roasting
reactions. Overheating leads to overoxidation of the product, which
results in the possible formation of magnetite (Fe304). This compound
in undesirable in the subsequent smelting step. Hence, the roaster
temperature must be carefully controlled. Cooling is performed by
adding water or inert fluxes (used in the smelting step) to the concen-
trates.
Offgases from fluid-bed roasters typically have S02 concentrations
of 12 to 15 percent on a dry basis, as measured at the roaster outlet.
The fluid-bed roaster at the Kennecott-Hayden smelter, for example,
gives an offgas S02 concentration of 13 percent.16 The offgas flow
rate from this unit (at the roaster outlet) has been reported at 880
NnrVmin (31,000 scfm), when processing 1,000 to 1,100 Mg/day (1,100 to
1,200 tons/day) of feed.16 The offgas S02 concentration from the
Kennecott-Hayden unit after gas cleaning is estimated to be 9.6 percent
on a dry basis. This concentration is considered representative for
both new and existing units.
The primary advantages offered by fluid-bed roasters lie in their
high offgas S02 concentrations and their high throughput rates as ecu-
pared to multihearth roasters. Multihearth roasters offer the advantage
3-10
-------
of providing multiple beds so that staged roasting with varying degrees
of roasting intensity at each stage (plus fuel additions if necessary)
can be used. Such flexibility is of notable importance if the feed
contains substantial quantities of volatile impurities. (This point
is discussed further in Section 3.5.2.1.)
3.2.1.3 Concentrate Dryers. Concentrate dryers are used to
reduce the moisture content of ore concentrates and other feed materials
before smelting. Typically, the ore concentrates as charged to the
smelter contain from 5 to 15 percent moisture. Dryers are used to
reduce the moisture content to less than 3 percent, and to as low as
0.1 to 0.3 percent in some cases. Drying differs from roasting in
that its only purpose is to reduce the moisture of process feed
materials. Because dryers operate at a relatively low temperature of
65° to 90° C (150° to 200° F), very little of the sulfur is driven off
as S02.
There are a number of systems that can be used to dry copper
concentrates, including both multihearth and fluid-bed dryers (roaster-
type dryers). Perhaps the most common type, however, is the rotary
dryer. The rotary dryer is a rotating cylinder inclined to the
horizontal with material fed into one end and discharged at the opposite
end.
In most types of dryers, air or combustion gases flow co-current
or countercurrent to the movement of the concentrate. Intimate contact
between the drying gases and the concentrate is usually permitted.
Dryers are always used at installations that use flash furnaces
for smelting. They may be used with the other types of smelting
furnaces. Also, dryers may be used upstream of fluid-bed roasters.
The use of a dryer leads to a net decrease in energy requirements
because less energy is required to remove moisture at a relatively low
temperature than at the high temperatures of a smelting furnace.
3.2.2 Smelting
Smelting is the pyrometallurgical process in which solid feed
materials are melted together with fluxing agents to form two or more
immiscible layers. The objective in copper smelting is the production
3-11
-------
of a metal sulfide mixture (matte) containing primarily FeS and Cu2S
and of a separate (oxide) slag layer containing primarily iron silicates.
During copper smelting, hot calcines from the roaster or raw,
unroasted concentrates are melted in a smelting furnace with siliceous
or limestone flux. Converter slag, collected dust, oxide ores, and
any other material rich in copper (including precipitates) may be
added to the furnace charge. Essentially all copper present in the
charge, independent of its chemical state, combines with sulfur also
present in the charge to form the stable compound, cuprous sulfide
(Cu2S).17 Sulfidic iron compounds such as FeS2 decompose, yielding
S02 and a comparatively stable compound, ferrous sulfide (FeS). The
mixture of primarily Cu2S and FeS is known as matte, which, due to its
high density, collects in the lower part of the furnace, from which it
is periodically removed for further processing in the copper converters.
Copper mattes produced by the domestic industry contain from 35 to
75 percent copper, with 40 to 45 percent being the most common. The
percentage copper present in the matte is referred to as the matte
grade. Matte also contains small amounts of other sulfides, such as
C°3S2» Ni3S2, pt>S, ZnS; impurities such as As, Sb, Se, and Te; and
precious metals such as Au, Ag, and Pt.18
The remainder of the molten mass, containing primarily metal
oxides and gangue materials, is known as slag. Since slag is of lower
density than matte, it floats on top of the matte layer, from which it
is periodically drawn off and generally discarded. Slags contain
generally low percentages of copper, which is present in the form of
both dissolved matte and entrained matte droplets.19 Since slags are
generally discarded directly, the copper content is a major cause of
copper loss. Because the concentration of copper in the slag increases
with increasing matte grade, matte grades produced in conventional
practice seldom exceed 50 percent copper.
The primary purpose of the added fluxing materials is to effect
the removal of iron oxides to the slag. A portion of the iron is
removed during smelting, and the balance is removed during the subse-
quent converting operation. Iron oxide (FeO) is produced readily
3-12
-------
during roasting and smelting because of the higher affinity for oxygen
of FeS compared to Cu2S. Molten iron oxides are highly miscible with
matte. The addition of silica, however, leads to the formation of
iron silicates, which are of lower density and immiscible in the
matte. Iron silicates are a major component of the slag, which is
ideally represented by the compound 2FeO-Si02.20 Slags also usually
contain small amounts of alumina and lime, which are present naturally
in the charge or are added to increase the fluidity of the slag.
Currently, four different smelting furnace technologies are
employed by the domestic industry. These include the reverberatory,
electric, Noranda, and Outokumpu (flash) furnaces. The conventional
reverberatory furnace is employed by the majority of the smelters.
However, three smelters plan or expect to retire their reverberatories
within the next few years in favor of other technologies. Two of
these smelters favor the Inco flash furnace technology while the third
is considering modifying its reverberatory furnaces for oxygen sprinkle
smelting (Section 3.4.3.5).
3.2.2.1 Reverberatory Furnaces. The conventional reverberatory
furnace as shown in Figure 3-3 is currently employed by 11 of the 15
domestic smelters. Reverberatory furnaces are long, rectangular
structures consisting of a hearth, side and end walls, and an arched
or suspended (flat) roof. Typical dimensions are about 11 M (36 feet)
in width and 40 m (130 feet) in length.21 Reverberatory furnaces
typically process from 800 to 1,200 Mg (900 to 1,300 tons) of charge
per day. In a reverberatory furnace, fossil fuels such as oil, gas, or
pulverized coal are burned above the charge being smelted. Furnace
burners are placed in one end wall, and hot gases exit the far end wall.
Flames from the burners may extend half the length of the furnace.
Temperatures at the burner end of the furnace exceed 1,500° C (2,730° F).
A portion of the heat radiates directly to the charge lying on the
hearth below, while a substantial part radiates to the furnace roof and
walls and is reflected down to the charge.
In addition to smelting the charge and allowing the settling of
matte and slag into separate layers, a major function of the conventional
3-13
-------
co
i
Green Feed or
Calcine Feed
Fuel
\
Converter
Slag
Air
Fettling Drag
Conveyor
Burners
Matte
Fettling Pipes
Offgas
Slag
Figure 3-3. Reverberatory smelting furnace.
-------
reverberatory furnace is to simultaneously recover copper from slag
produced in the converters. Molten converter slag is returned to the
furnace, whereupon matte and copper that are mechanically entrained in
this slag settle out by gravity. To some extent, copper oxides trapped
in the slag are reduced to Cu2S by reactions with iron sulfide and also
settle out. The copper content of the slag from the reverberatory
furnace is typically 0.3 to 0.6 percent.
Reverberatory smelting is a continuous process, although charging
is intermittent. In green-charged furnaces, the concentrates and fluxes
are typically charged along the sidewalls (side feeding), where they
form "banks" or piles that protect the sidewall refractory from burner
heat. These banks slowly melt into the bath. Charging is performed
using drop pipes ("fettling" pipes) that penetrate the furnace roof
and are spaced regularly along the length of the furnace on each side.
Charge is delivered to the drop pipes by enclosed drag chain conveyors
(see Figure 3-3) or by moveable charge bins atop the furnace that
direct the feed to one drop pipe at a time. Typically, most of the
charge is placed along 60 to 70 percent of the furnace length (from
the burner end), which constitutes the smelting zone. The remainder
of the furnace is termed the settling zone, in which matte entrained
in the slag settles to the matte layer. Green-charged furnaces are
also charged, in limited cases, by charge slingers. These devices are
high-speed conveyors that "throw" the charge into the furnace through
ports that are opened along the sidewalls. The result is to spread
the charge evenly over the bath. The use of three slingers on each
side of the reverberatory furnace at the Phelps Dodge-Ajo smelter has
been reported.22
Calcine-charged reverberatory furnaces are typically charged
using fixed or retractable gun-type feed pipes penetrating the furnace
sidewalls (Wagstaff guns) that spread the charge uniformly over one-half
to two-thirds of the surface of the molten bath. Typically, two such
feeders are present on each side of the furnace. Furnaces charged
with Wagstaff guns generally employ water-cooled panels on the outside
of the furnace at the slag line to protect the sidewall refractory
from burner heat. Calcine may also be charged through the roof arch
3-15
-------
at either the center or the sidewalls. Sidewall charging of calcines,
using a drag chain conveyor/drop pipe system similar to that used on
green-charged furnaces, is used in some cases. However, the resulting
charge banks are not as substantial as those produced with green
charge because of the comparatively free-flowing nature of hot calcine.
Reverberatory furnaces produce a large volume of offgases containing
a relatively low S02 concentration. Both the large offgas flow rate and
the low S02 concentration result because most of the heat is produced
from the combustion of fossil fuel in air, which is 79 percent inert
nitrogen. Offgas flow rates have been reported at 1,800 to 2,100 Nm3/
min (63,000 to 74,000 scfm)23 and may be substantially higher.
The principal mechanism for S02 formation within the reverberatory
furnace is the volatization and oxidation of pyritic (loosely bound)
sulfur. Typically, from 10 to 30 percent of the sulfur contained in
the original concentrate feed is eliminated in the furnace offgases.
The average S02 concentration in the offgases from domestic caicine-
charged furnaces (after gas cleaning) varies from 0.4 to 1.5 percent
S02 on a dry basis depending upon the plant, with an industry-wide
average (as weighted by respective flow rates) of 0.8 percent S02.
Domestic green-charged furnaces produce average S02 concentrations on
a dry basis ranging from 1.0 to 2.0 percent, depending upon the plant.
The industry-wide average for green-charged furnaces is 1.4 percent
S02. The variation in S02 concentration is largely due to the
difference in charge composition, i.e., the availability of pyritic
sulfur, although the infiltration of air into the gas handling equipment
downstream of the furnace is also a factor.
New calcine-charged reverberatory furnaces processing high-impurity
feed materials (see Section 3.5) are considered capable of producing
1.7 percent S02 in the offgases (on a dry basis) after gas cleaning.*
*This concentration is based on theoretical calculations using the
average charge composition for the (typical) year 1979 at the ASARCO-
Tacoma smelter. Calculations assume (1) 28.4 percent sulfur removal
in the furnace, (2) 40 percent matte grade, (3) natural gas fuel,
(4) 4.5 x io6 Btu required per ton calcine smelted, (5) 1 percent
oxygen in the gases entering the furnace uptake, and (6) 100 percent
dilution of the uptake gases.
3-16
-------
The primary advantage afforded by reverberatory furnaces over
most other smelting technologies is their versatility. Feed materials
that are wet or dry, lumpy, or fine may all be smelted readily. A
major disadvantage of these furnaces is the weak S02 stream produced.
Furthermore, since the reverberatory furnace uses little of the inherent
energy of the sulfide charge, its energy requirement is among the
greatest of the major pyrometallurgical processes.
3 2.2.2 Electric Furnaces. Electric smelting furnaces are
employed by two of the domestic smelters. As shown in Figure 3-4, the
heat required for smelting in these furnaces is generated by the
passage of electric current through the molten bath. Carbon electrodes
dip into the slag layer of the bath, forming an electrical circuit.
As electric current is passed through the circuit, the resistance of
the slag causes the generation of heat, which results in smelting
temperatures.
Typically, electric furnaces used for matte smelting are rectangular
and are about 35 m (115 ft) long by 10 m (33 ft) wide. A furnace of
this size uses six Soderberg-type electrodes, 1.8 m (6 ft) in diameter,
which are spaced uniformly along the long axis of the furnace.24 The
current flow and voltage between pairs of electrodes are on the order
of 30,000 amps and 500 volts, respectively.24 The electric furnace at
Inspiration Consolidated Copper Company is designed to process about
1,640 Mg/ day (1,800 tons/day) of feed.
Feed to electric furnaces is typically charged through the roof
of the furnace, near the electrodes. The feed generally consists of
dried concentrates or calcines. The charging of wet concentrates is
avoided because the moisture can cause steam explosions.25 The
unsmelted charge materials float on top of the molten bath. Heat is
thus transferred from the hot slag (where the heat is produced) to the
charge floating on its surface. Keeping the bath covered with charge
maximizes the rate of heat transfer to the charge. Also, the floating
charge reduces heat losses from the bath and prevents overheating of
the roof refractories.26
3-17
-------
Offgas
CO
I
co
Fettling Pipes DrV Feed
or
Calcine
Feed
Converter Slag \^ \
Launder
\
Electric
Power
Electrodes
Slag
Figure 3-4. Electric smelting furnace.
-------
Electric furnaces are similar to reverbs in that they offer
essentially the same degree of versatility. The physical and chemical
changes occurring in the molten bath of an electric furnace are similar
to those occurring in a reverberatory furnace. Furthermore, molten
converter slags are normally returned to electric smelting furnaces
for the recovery of entrained copper, as with reverberatory furnaces.
A major advantage of electric furnaces over reverberatory furnaces is
their ease of control, resulting from a high concentration of S02 in
the exhaust gases (on the order of 5 percent).27 Furthermore, since
combustion air is not required, the offgas flow rates are less than
half of those in reverberatory furnaces. The Inspiration furnace, for
example, produces approximately 850 NnrVmin (30,000 scfm) of offgas
containing from 4 to 6 percent S02 on a dry basis.28 Low offgas flow
rates minimize the size and cost of downstream gas handling equipment.
The primary disadvantage of the electric furnace is that, like
the reverberatory furnace, it makes limited use of the energy potentially
available from oxidizing the sulfide minerals of the charge. Further-
more, its operating costs tend to be high because of the high price
of electricity.29
3.2.2.3 Flash Furnaces. In contrast to reverberatory and electric
furnaces, flash furnaces use the heat evolved from the partial oxidation
of their sulfide charge to provide much or all of the energy required
for smelting. The result is that the energy required for flash furnace
operation is considerably less than that associated with reverberatory
and electric furnaces. Flash furnaces also produce offgas streams
containing high concentrations of S02, which may be efficiently recovered
as sulfuric acid or liquid S02. For these reasons, most of the world's
new smelting furnaces installed since 1965 have been of the flash type.30
Currently, one of the domestic smelters employs a flash furnace, and
two others intend to convert to flash smelting technology.
In flash smelting, dried ore concentrates and finely ground
fluxes are injected together with oxygen, preheated air, or a mixture
of air and oxygen into a hot furnace of special design. Within the
furnace, the sulfide particles react rapidly with the oxidizing gas,
releasing heat. Important reactions include the following:30
3-19
-------
4CuFeS2 + 502 -* 2(Cu2S-FeS) + 2FeO + 4S02
2FeS + 302 -» 2FeO + 2S02.
The melted droplets fall to the bath below, where the matte- and
slag-forming reactions are completed. The matte droplets settle
through the slag layer to form the matte layer.
The combustion reactions use essentially all of the oxygen
contained in the process atmosphere. Consequently, the regulation of
the oxygen/ concentrate ratio in the furnace controls the extent to
which the flash combustion reactions proceed and thus determines both
the grade of matte produced and the heat released for smelting the
furnace charge. Increasing the incoming temperature or oxygen content
of the combustion air also effectively increases the heat available
for smelting. As a result, in some cases it is possible for the flash
combustion and smelting reactions to occur autogenously. Under these
conditions, the heat released by the oxidation of iron and sulfur is
sufficent to smelt the furnace charge.
The charge to a flash smelting furnace must be fine grained, and
essentially "bone dry" to insure an even and homogeneous distribution
of the charge as it is injected into the furnace. The copper concen-
trates should be of a fineness corresponding to at least 50 percent
minus-200 mesh, and the fluxing material should be of a fineness
corresponding to at least 80 percent minus-14 mesh.31 Since most
concentrates are obtained from ores by flotation techniques, their
fineness normally meets these requirements without further grinding.31
The fluxing materials, however, usually require additional grinding
beyond that necessary for use in reverberatory or electric smelting
furnaces.
In most cases it is necessary to dry the charge to bone-dry
conditions (0.1 to 0.3 percent moisture) before smelting, as the
concentrates typically contain from 5 to 15 percent moisture.31 It is
common practice to use the drying facilities not only for drying the
charge, but also for blending the fluxing materials and the various
copper concentrates available to provide a charge of uniform composition.
3-20
-------
It should be noted that the charge is not roasted, as the flash
combustion process makes use of the roasting reactions to melt the
charge.
The principle advantages of flash furnaces lie in their low energy
requirements, their high S02 strength in the offgases, and their high
production rates. The productivities of flash furnaces are on the
order of 8 to 12 Mg of charge per day/m2 (0.8 to 1.2 tons of charge
per day/ft2) of hearth area, which is two to four times that of rever-
beratory furnaces.32
The principle disadvantage of flash smelting is that the copper
content of the furnace slag tends to be higher than that of reverbera-
tory and electric furnaces. As a result, the use of separate facilities
may be required to recover some of the copper from the flash furnace
and converter slags before discard.
Flash smelting technology has been developed by two different
companies: International Nickel Company (Inco) in Canada and
Outokumpu Oy in Finland. The major difference between the two techno-
logies is in the design of the smelting furnace and in the oxidizing
environment within the furnace. The Inco furnace uses pure oxygen,
while the Outokumpu furnace employs preheated air or oxygen-enriched
air as the oxidizing medium. Currently, 30 Outokumpu flash furnaces
are operating or are licensed to operate worldwide. Two Inco flash
furnaces are currently operating worldwide, and two additional units
are slated for construction in the United States. The larger number
of Outokumpu furnaces is attributed to the fact that Outokumpu has
been marketing its technology for a number of years, while Inco has
only recently offered its technology for license.
3.2.2.3.1 Inco flash furnaces. The Inco flash furnace, shown in
Figure 3-5, is the simpler of the two flash furnace designs. The furnace
used by Inco in Sudbury, Ontario, is about 24 m (80 ft) long, 7 m (24 ft)
wide, and 6 m (20 ft) high at the ends.33 The furnace uptake extends
the full width of the furnace at its center. For gas tightness, the
furnace is essentially totally enclosed in a shell made of mild steel,
1 cm (3/8 in.) thick. This particular furnace has a nominal capacity
3-21
-------
CO
I
ro
Dried Ore
flux Concentrates
Dried Ore
Concentrates
and Fluxes
Constant
Weight Feeders
Oxygen
Oxygen
Siag Matte
Figure 3-5. Inco flash smelting furnace.
-------
of 1,360 Mg/day (1,500 tons/day) of dry concentrate,33 although extended
operation at feed rates in excess of 1,630 Mg/day (1,800 tons/day) has
been demonstrated.34
Dried ore concentrates, fluxes, and commercial oxygen are blown
horizontally into the furnace through four water-cooled burners, two
at each end. This design produces a short flame and a uniform tempera-
ture over the entire hearth area. Water-cooled copper jackets faced
with refractory brick cover approximately 20 percent of the sidewalls,
mainly in the region below the gas offtake. Matte is tapped from the
central zone of one sidewall, while slag is skimmed from beneath the
burners at one end of the furnace.
Because pure oxygen rather than air is the oxidizing medium in
the furnace, the concentration of sulfur dioxide in the Inco furnace
offgases is very high, normally in the range of 70 to 80 percent.33
At this concentration, the gases are suitable for the economic produc-
tion of sulfuric acid or liquid S02. The use of oxygen rather than
air has the added advantage of providing a low off-gas volume per unit
of charge. The offgas flow rate from Inco's furnace in Sudbury when
operating at 1,360 Mg/day (1,500 tons/ day) of feed is 130 NmVmin
(4,600 scfm).33 Several benefits accrue from the low offgas volume.
The size and cost of downstream gas-cleaning equipment is reduced
substantially. Also, because of the relatively low volume and heat
content of the gases, Inco maintains that the use of a waste heat
recovery system downstream of the furnace is not justified.33 (The
heat content of the gases represents about 20 percent of the heat
generated in the furnace.) Finally, the low offgas flow rate results
in a low gas velocity leaving the furnace. As a result, the dusting
rate from the furnace is low, amounting to 2 to 3 percent of the feed
rate.33
The grade of matte produced at the Sudbury installation is reported
to varying between 40 and 50 percent copper, depending upon the through-
put rate and the amount of secondary materials charged to the furnace.33
The feed at this installation is a chalcopyrite concentrate analyzing
30 percent Cu, 30 to 31 percent Fe, and 33 percent S. Inco can return
3-23
-------
about 50 percent of the converter slag to the furnace for copper
recovery. The remainder has to be processed in other facilities
because converter slag is an important bleed for the impurity nickel,
which occurs at fairly high levels in the feed at this installation.
Inco flash furnaces appear to be useful, however, for processing all
of the converter slag generated. The flash furnace slag at Inco,
which contains 0.6 to 0.7 percent copper, is discarded without
additional treatment.33 This copper concentration is somewhat higher
than that typically encountered in slags produced by reverberatory
furnaces. The Inco furnaces to be installed at the ASARCO-Hayden
smelter will process all of the converter slag, and the flash furnace
slag will be discarded without additional treatment.3'1
The recycle of converter slag to the furnace requires that some
additional heat be provided to maintain the bath temperature. Heat
can be provided readily by increasing the oxygen-to-concentrate ratio,
which results in the flash combustion of additional sulfur and iron.
This scheme also leads to an increase in the matte grade. Tests made
by Inco indicated that the matte grade increased from 40.5 to 45.0
percent copper when 43 percent of the converter slag was reverted to
the furnace.33
Inco has reported that the furnace matte grade can be decreased
by the addition of coal to the feed.33 The possibility of decreasing
matte grade by this means was established when Thompson copper concen-
trates containing 4 to 12 percent carbon (as graphite) was processed
as a component of the feed. During the period from 1964 to 1969, a
total of 30,000 Mg (33,000 tons) of Thompson concentrates were processed
at a rate of up to 100 Mg/day (110 tons/day).36 The total feed rate
to the furnace was 1,400 to 1,600 Mg/day (1,540 to 1,760 tons/day);
hence the Thompson concentrates supplied up to about 1 percent carbon
to the total feed. All of the carbon was combusted in the furnace.
The result was a decrease in matte grade and a slight drop (of 2 to
5 percent) in the S02 concentration in the offgases.36 Inco has
tested the use of coal mixed with the feed in a bench-scale furnace
3-24
-------
and in pilot-plant tests.33 Inco reports that additions of less than
1 percent coal lead to decreases in matte grade of from 10 to 15
percentage points, depending on the composition of the feed.33
3.2.2.3.2 Outokumpu flash furnaces. The Outokumpu flash furnace,
shown in Figure 3-6, consists of three distinct sections: a reaction
shaft, a settler, and an uptake shaft. The dried copper concentrates
and fluxing materials are injected continuously down the reaction
shaft onto the slag surface through concentrate burners located at the
top of this shaft. In the burners, the charge is mixed with preheated
air (450° to 1,000° C) or oxygen enriched air, preheated or ambient.37
Large furnaces contain up to four burners.
Recent Outokumpu flash furnaces are 20 m (65 ft) long (inside),
7 m (23 ft) wide, and 3 m (10 ft) high (hearth to roof). The firing
towers are 6 m (20 ft) diameter and 6 m (20 ft) high (above the roof),
while the offtake towers are the width of the furnace (7 m [23 ft]),
3 m (10 ft) long, and 6 m (20 ft) high. This size furnace treats
1,200 Mg (1,320 tons) of dry charge per day.38
The S02 concentration in the offgases from Outokumpu furnaces is
high, typically 10 to 30 percent, allowing the efficient removal of
S02 as sulfuric acid.39 The S02 concentration varies depending upon
the copper concentrate analysis, the grade of matte produced, the
degree of oxygen enrichment, and the degree of combustion air preheat.
The furnace matte grade is typically in the range of 50 to 65 percent.40
The offgas flow rate for the Outokumpu furnace at the Phelps
Dodge-Hidalgo smelter has been reported at 2,200 Nm3/min (77,700 scfm),
with an S02 concentration of 13 percent, when operated at a feed rate
of 2,440 Mg/day (2,680 tons/day).41 This furnace uses preheated
combustion air that is not oxygen enriched.
Unlike the Inco flash furnace, Outokumpu furnaces are not
autogenous unless the ingoing air contains 40 percent or more oxygen
and oil burners are placed at the top of the combustion tower and in
the hearth (settling zone).38 At the Hidalgo flash furnace, additional
heat is provided by preheating the combustion air, firing oil burners,
and mixing finely ground coal with the feed. Extensive use is made of
3-25
-------
Preheated
Air
ro
CTl
Dried Ore
Concentrates
and Fluxes
Concentrate Burner
,— Oil
Offgas
Slag
Matte
Settler
Slag
Figure 3-6. Outokumpu flash smelting furnace.
-------
coal mixed with the feed to supply additional heat at the Toyo smelter in
Japan.42 A discussion of the Toyo experience is provided in Appendix E.
The concentration of copper in slags produced by Outokumpu flash
furnaces tends to be fairly high, typically 1 percent.43 As a result,
this type furnace cannot be used efficiently to recover copper from
converter slags. Also, the slags from the flash furnace must themselves
be retreated in a separate process to minimize the loss of copper.
Dust losses in the Outokumpu flash smelting process are also
fairly high, up to 10 percent of the feed,38 which is substantially
greater than the dust losses encountered with Inco flash furnaces.
High dust losses result because significant quantities of concentrate
particles do not settle from the gas/solid suspension during their
passage through the furnace.
3.2.2.4 Slag Cleaning Furnaces. Slag cleaning furnaces are
designed to recover a portion of the copper entrained in molten slags
from smelting furnaces and converters. These furnaces are typically
used at Outokumpu flash smelting installations because of the relatively
high copper content of the slag from this technology. An electric
slag cleaning furnace is currently operated at the Phelps Dodge-Hidalgo
smelter.
Slag-cleaning furnaces are typically small, low-powered electric
furnaces. Normal furnace operating temperature has been reported at
1,230° to 1,320° C (2,250° to 2,400° F).44 Slags charged to the
furnace are allowed to settle under quiescent and reducing conditions.
Reducing conditions are maintained by the addition of coke or iron
sulfide to the bath. The addition of iron sulfide serves to recover
copper that is in oxide form and return it to the matte phase according
to the following reaction:
FeS(£) + Cu20(£,slag) •* FeO(£,slag) + Cu2SU,matte).
The residence time of slags within the furnace is of the order of
5 hours.45 The slag discharged typically has a copper concentration
of 0.4 to 0.5 percent.45
Although smelting is not the primary objective of slag-cleaning
furnaces, these furnaces may be used to smelt some revert materials.
3-27
-------
It should be noted that an alternative to slag-cleaning furnaces
for recovering copper from smelting furnace and converter slags is the
froth flotation process, which yields somewhat better copper recovery
but requires that the slags be slow cooled, crushed, and ground before
processing. Froth flotation is employed for the recovery of slag-
contained copper at the Kennecott-Garfield smelter, which uses the
Noranda process (Section 2.3.5.1).
3.2.3 Converting
The converting operation is the final major step in the pyrometal-
lurgical extraction of copper from sulfide ore concentrates. The
purpose of converting is to eliminate the remaining iron and sulfur
present in the matte, leaving molten "blister" copper (98.b to 99.5
percent Cu). The blister copper product is subsequently fire refined
and electrorefined to produce high-purity copper (99.99 + percent Cu).
Upon reaching the converters, all of the rock (gangue) and a
portion of the iron and sulfur present in the original ore concentrates
have been eliminated. The matte charge consists of a Cu2S:FeS melt
containing small amounts of other elements and precious metals. The
batch-converting process serves to eliminate, sequentially, the FeS
component and the sulfur present in the remaining Cu2S component. As
mentioned previously, the separation is based on the high affinity of
iron for oxygen, as compared to the oxygen affinity of copper. The
quantity of sulfur eliminated during converting operations generally
amounts to 40 to 70 percent of the sulfur in the original ore concen-
trates.
The extraction of copper is accomplished by adding siliceous
fluxes to the molten matte and then blowing air through the mixture to
oxidize the iron sulfides to iron oxides. The iron oxides combine with
the silica fluxes to form a slag, which is removed from the converter.
The copper sulfide or "white metal" that remains is then oxidized to
blister copper through continued blowing. The oxidation reactions
occuring during converting are highly exothermic, and the entire
operation is autogenous. In fact, the converter gradually heats up
during the process. To prevent excessive temperatures, which lead to
3-28
-------
high refractory wear, substantial quantities of "cold dope" (revert
materials; copper scrap; and, in some cases, cast blister copper) are
charged during the converting cycle.
Aside from their role in the elimination of iron and sulfur,
converting operations are also very important in the elimination of a
number of other impurity elements. The role of converters in impurity
elimination is discussed in Section 3.5.2.3.
Generally within the domestic industry, two or three converters
(one of which is a standby) are associated with each smelting furnace.
3.2.3.1 Peirce-Smith Converters. Converting operations in both
the domestic and foreign industries are dominated by the Peirce-Smith
converter. Domestically, these vessels are used at 14 of the 15
smelters. As shown in Figure 3-7, the side-blown Peirce-Smith unit is
a horizontal, refractory-lined steel cylinder with a large opening or
"mouth" in the side. Typical (inside) dimensions are 4 m (13 ft)
diameter by 9 m (30 ft) long. A vessel of this size normally processes
from 350 to 450 Mg/day (380 to 500 tons/day) of matte. Compressed air
or oxygen-enriched air is supplied to the converter through a header
along the back of the vessel, from which a horizontal row of generally
40 to 50 tuyeres provide passages through the converter shell into the
interior. The vessel rotates about its major axis, swinging the
converter mouth through an arc of about 120° from the vertical. When
the converter is in the upright or blowing position, a large retractable
hood, referred to as the primary hood, is lowered over the mouth to
capture the escaping gases.
Molten matte produced in the smelting furnace is charged to the
converter through the mouth by ladles, using overhead cranes, filling
the vessel approximately half full. Silica fluxing materials are also
charged, either through the mouth or through one end of the converter,
as shown in Figure 3-7. During charging the converter is rotated to
bring the mouth to an angle of about 45° from the vertical, as shown
in Figure 3-8. With the converter mouth in the charging position, the
tuyeres are above the level of the matte. Following charging of the
matte and fluxing materials, air or oxygen-enriched air is supplied
3-29
-------
Offgas
OJ
OJ
o
Tuyere
Pipes
Siliceous
Flux
Pneumatic
Punchers
Figure 3-7. Peirce-Smith Converter.
-------
CO
I
CO
Charging
Blowing
Skimming
Figure 3-8. Copper converter operation.
-------
under pressure to the tuyere line, and blowing commences. Blowing
rates are generally between 425 and 740 Mm3/ min (15,000 to 26,000 scfm).
The converter is then rotated, as shown in Figure 3-8, swinging the
converter mouth to a vertical position and submerging the tuyeres to a
depth of 6 to 24 in. below the surface of the matte.46 The primary
hood is then lowered into position over the mouth.
As air blown through the tuyeres enters the molten matte, the
matte in the immediate vicinity of the tuyeres is cooled, forming
accretions which obstruct the tuyere openings and reduce the blowing
air flow rate. To prevent complete obstruction of these openings, the
tuyeres are mechanically cleaned every several minutes by a machine
that forces an iron bar through each tuyere passage.
As the tuyere air passes through the molten matte, the iron
sulfide is converted to iron oxide and S02 with the release of substan-
tial heat according to the following reaction:
2FeS + 302 = 2FeO + 2S02 .
The sulfur oxides are removed in the converter gases discharged through
the converter mouth. The oxidizing conditions also lead to the formation
of solid magnetite, according to the following empirical relation:
6FeO + 02 -> 2Fe304.
The iron oxide produced combines with the molten silicates to produce
an immiscible slag according to the usual slagging reaction:
2FeO + Si02 = 2FeO-Si02.
This stage of the converter cycle operation is termed the slag blow.
Blowing is continued until a substantial layer of slag is formed
in the converter. The vessel is then rotated (after raising the
primary hood) as shown in Figure 3-8, swinging the converter mouth
through an arc of about 120° from the vertical, and raising the tuyere
line above the surface of the molten bath. The air supply to the
tuyere line is shut off and the blowing discontinued. Slag is skimmed
or poured from the converter into a ladle and returned to the smelting
furnace or transferred to slag treatment facilities for the recovery
3-32
-------
of copper contained in the slag. The converter is then rotated to the
charging position, and fresh matte, fluxing materials, and cold supple-
ments (such as smelter reverts* and copper scrap) are added to bring
the converter charge back to the working level. Blowing is resumed
and the converter rotated to the working position. The primary hood
is lowered into position over the converter mouth.
The process of charging, blowing, and slag skimming is repeated
until a charge of copper sulfide is accumulated in the converter,
filling it to the working level. The vessel is then rotated to the
blowing position, and the copper blow or finish blow begins. During
this stage of the converter cycle, the copper sulfide (white metal) is
oxidized, forming S02 and copper. Following the copper blow, the
converter contains only metallic copper known as blister copper, which
is approximately 99 percent pure. The converter is rotated to the
pouring or skimming position and the blister copper poured into ladles
for transfer to refining facilities. The emptied converter is then
charged with fresh matte and fluxing materials, and the converting
cycle repeated.
A converter generally makes from one to three cycles in a 24-hour
period, with the actual blowing time comprising about 70 to 75 percent
of the cycle.47 The remainder of the cycle is spent in charging and
skimming operations, and holding (waiting) due to normal process
fluctuations within the smelter. The primary determinant of the time
required for a complete cycle is the grade of matte charged to the
converter, although the blowing rate is also an important factor.
Low-grade mattes, which have a larger percentage of FeS, require
longer cycle times than do high-grade mattes. At the ASARCOE1 Paso
smelter, for example, with a matte grade of 40 percent, the duration
of the slag blow has been reported at 5.8 h, while the copper blow
required 3.9 h.48 In contrast, at the White Pine smelter with a matte
grade of 65 percent, the duration of the slag and copper blows were
^Materials recycled from the smelting process, including accre-
tions, shells from ladles, solidified spillage from material handling,
and flue dust.
-------
0.5 h and 3.25 h, respectively.49 In addition to having a longer
cycle time, low grade mattes also generate more heat during the cycle.
Hence, smelters desiring to process large quantities of scrap materials
in their converters tend to produce lower-grade mattes.
The primary converter hood is designed to be relatively close-
fitting to the converter mouth (or to the converter body, depending
upon the design). However, some infiltration of air into the offgases
is both inevitable and necessary. A tight seal cannot be maintained
between the primary hood and the converter because of irregularities
(buildups) caused by pouring operations and bath splatter. Offgas
leaving the converter mouth typically has temperatures of 1,150° to
1,200° C (2,100° to 2,200° F).50 Even though many primary hoods are
water cooled, some cooling of these gases via infiltration is necessary
to prevent damage (i.e., warpage and buckling) to the primary hood, as
well as to dampers and flues. The volume of infiltration air typically
entering around the hood amounts to 100 to 200 percent of the true
converter offgas flow.
The offgas flow rate leaving the primary hood of converters
typically ranges from 850 to 1,270 NmVmin (30,000 to 45,000 scfm).
The average S02 concentration in these gases is normally in the range
of 4 to 5 percent during the slag blow and 7 to 8 percent during the
copper blow.51 Values of the overall average S02 concentration in
the offgases (after gas cleaning) from existing domestic converting
operations fall in the range of 1.6 to 6.5 percent on a dry basis.
The industry-wide average value, as weighted by the respective plant
flow rates, is 4.3 percent. New converting operations are capable of
producing, after gas cleaning, 5.5 percent S02 during the slag blow
and 7.9 percent S02 during the copper blow, with an overall average of
6.3 percent.*
*S02 strength for new converting operations are based on theoretical
calculations that consider 100 percent dilution of the offgas leaving
the converter mouth. The overalI average was derived for three opera-
ting converters. All values are on a dry basis.
3-34
-------
As indicated previously, converting is a batch operation, and up
to 30 percent of the cycle time of each converter is spent in charging,
skimming, and holding operations (during which essentially no offgas
is produced). Each operating converter in a plant may or may not be
in the same mode of operation. Obviously, these characteristics of
converter operation can cause problems in the emission control system
because of the erratic nature of the resulting flow rate and S02
concentration in the combined converter offgas stream. Fortunately,
however, the scheduling of converter operations can be used to eliminate
discontinuities in the S02 supply and stabilize (to some extent) the
combined offgas flow rate. A detailed analysis of converter scheduling
is discussed in Section 4.5.1.
3.2.3.2 Hoboken Converters. An alternative to the traditional
Peirce-Smith converter is the newer Hoboken or "siphon" converter,
first developed by Metallurgie Hoboken, N. V. in Belgium. The Hoboken-
type converter is currently used by one of the domestic smelters.
Although the Hoboken converter is essentially the same as a
conventional Peirce-Smith converter, this vessel is fitted with a side
flue located at one end of the converter and shaped as an inverted U,
as shown in Figure 3-9. The inverted, U-shaped flue, or "siphon,"
rotates with the converter and is fitted with a cylindrical duct, also
rotating with the converter, which leads to a fixed vertical flue. A
specially designed seal exists between the rotating duct and the fixed
flue. This flue arrangement permits siphoning of the converter gases
from the interior of the converter directly to the offgas collection
system.
The primary advantage of the Hoboken converter as designed lies
in emissions control. By maintaining a slightly negative pressure at
the converter mouth, it is possible to minimize or eliminate the
escape of S02, while maintaining a high S02 strength in the offgases.
The draft is maintained using variable-speed fans and dampers. With
two converters in operation, only one of which is blowing at any time,
personnel at Metallurgie Hoboken, N. V. indicate that converter off-gases
averaging 8 percent S02 can be expected.52 An additional advantage is
3-35
-------
Figure 3-9. Hoboken converter.
3-36
-------
that, since the mouth is freely accessible throughout the operations,
it is possible to charge large quantities of liquid or solid materials
during blowing.53
Personnel at Inspiration Consolidated Copper Company, the only
domestic smelter using siphon converters, have indicated that these
converters have given unsatisfactory performance.54 Problems result
because there is a tendency for molten particles swept from the bath
to solidify and accumulate in the siphon area. The result is severe
buildups, which plug the gas flow passage and prevent proper draft at
the converter mouth.54 As the buildups accumulate, the converter
gases vent increasingly through the converter mouth and not through
the siphon. The buildups are responsible for limiting the campaign
life (i.e., period of operation before major maintenance) of each
converter to approximately 3 months.55 To alleviate the problem,
Inspiration intends to modify the converters by eliminating the siphon
system.
3.2.4 Fire Refining
Fire refining operations serve to eliminate the gross impurities
from blister copper. The resulting product, anode copper, is further
refined electrolytically to remove remaining impurities and recover
gold and silver.
Although the majority of domestic copper produced by pyrometal-
lurgical means is destined for electrolytic refining, it is not so
processed directly because of small quantities of dissolved sulfur and
oxygen. Electrolytic refining operations require strong, thin copper
anodes with smooth surfaces. Blister copper is not suitable for the
production of anodes because, upon cooling, the sulfur and oxygen
combine to form S02, which leads to gas bubbles throughout the metal
and "blisters" on the surface.56
Fire refining is performed in rotary-type refining furnaces
resembling Peirce-Smith converters or in small hearth-type furnaces.
Among the domestic industry, the rotary-type furnace predominates.
Blister copper is charged to the vessel directly from the converters.
Dimensions of the rotary-type refining furnaces vary; however, a
3-37
-------
4- by 9-m (13- by 30-ft) furnace may be regarded as typical.57 The
hearth-type furnaces are used in limited cases where the melting of
solid charge is practiced.
Fire-refining operations are accomplished by blowing gases (air,
natural gas) through the molten metal. Gas flow rates are relatively
low to accurately control the metal composition.1"'8 In contrast to
converters, very little heat is provided by the refining reactions;
hence some combustion of fuel is necessary to maintain the temperature
of the furnace. The temperature of operation is about 1,130° to
1,150° C (2,070° to 2,100° F), which provides sufficient superheat for
the subsequent casting of anodes.59
Blister copper from converting contains about 0.05 percent
dissolved sulfur and 0.5 percent dissolved oxygen.56 The removal of
sulfur and oxygen is accomplished in two stages. During the first or
oxidation stage, air is blown through the blister copper to remove the
sulfur (as S02). The duration of the oxidation cycle is variable,
depending upon factors such as the mass and sulfur content of the
charge and the blowing rate. Times have been reported at 0.5 to
1.0 hour,b° and at 3 to 4 hours.61 The oxidation step is completed
when the sulfur content drops to a level of 0.001 to 0.003 percent.60
This stage may be ended with the skimming of a small amount of slag
from the bath surface.
The second or "poling" stage serves to remove oxygen, which has
dissolved in the copper both during converting and during the oxidation
stage of refining. Oxygen is removed by blowing natural gas, reformed
natural gas, or propane through the molten metal. The duration of the
poling stage is variable, having been reported at 2.5 to 3 hours.61
Gas addition to the bath is stopoed when the oxygen level in the anode
copper is 0.05 to 0.2 percent, wlich gives a "flat set." to the anodes
when they are cast.60 At this point, the surface of the batch may be
covered with a layer of low-sulfjr coke to prevent reoxidation of the
copper. The poling stage is so-called because of the older practice
of lowering green wood poles into the molten bath to supply the
necessary reductants. Poling by this means is still practiced at the
ASARCO-Tacoma and White Pine smelters.
3-38
-------
6.
The removal of some metallic impurities may also be achieved
during fire refining. In general, oxidation and slagging techniques
are employed.62 The concentration of lead may be reduced by the
addition of silica to the bath just before the end of the oxidation
stage.61 Continued blowing of air effects the removal of lead into
the slag. The addition of soda ash and lime to the charge at the end
of the oxidation stage is used to slag off arsenic and antimony. With
successive treatments, it has been reported that almost complete
removal of these two elements can be achieved.62 ASARCO has indicated,
however, that essentially complete removal cannot be achieved in a
practical sense because (1) successive treatments are quite time
consuming and may be cost prohibited and (2) soda ash is corros've to
anode furnace refractories.
It should be noted that some metallic impurities have a very high
affinity for blister copper. For these elements, fire refining is
virtually ineffective for reducing their concentration. A notable
example is the element bismuth.64 Fortunately, however, substantial
elimination of bismuth can be achieved during the slag blow of convert-
ing. (Impurity elimination during converting operations is discussed
in Section 3.5.2.3.)
3.2.5 Continuous Smelting Systems
In recent years, a number of foreign companies have initiated
development of continuous smelting systems. In such systems, a steady
stream of blister copper is produced on an uninterrupted basis from a
steady feed of ore concentrates. Continuous smelting systems are
designed to make maximal use of the inherent energy of the sulfide
charge, and hence are generally among the most energy-efficient of the
major pyrometal1urgical processes. The most notable of these smelting
technologies are the Noranda process and the Mitsubishi process.
3.2.5.1 Noranda Process
The Noranda process was developed by Noranda Mines, Ltd., of
Canada. As originally designed, the process allowed the production of
blister copper on a continuous basis in a single vessel, by effectively
combining roasting, smelting, and converting into one operation.
3-39
-------
Metallurgical problems, however, led to the operation of these reactors
for the production of copper matte. Presently, two installations
operate Noranda reactors: the Noranda Mines, Ltd., Home smelter and
the Kennecott Garfield smelter in Garfield, Utah. Both of these
installations produce a high grade (70 to 75 percent copper) matte,65
which is subsequently processed in the converters.
Noranda reactors, as shown in Figure 3-10, are cylindrical,
refractory-lined vessels 21 m (70 ft) long and 5 m (16 ft) in diameter
that closely resemble Peirce-Smith converters. Oxygen-enriched air is
introduced into the molten bath through a row of tuyeres along one
side of the vessel. The vessel may be rotated about its horizontal
axis to bring the tuyeres out of the bath for servicing. At the
Kennecott Garfield smelter, feed is introduced continuously to the
vessel through an opening at one end using a belt-driven slinger.
Matte is tapped intermittently through ports located on the cylindrical
side of the vessel. Slag is tapped intermittently from the end of the
vessel opposite the charging end. Offgases from the process contain
16 to 20 percent S0266 and exit from the mouth in the top. As the
mouth is used only for the removal of gas from the reactor (hence not
for charging and tapping), it can be closely hooded to prevent the
escape of S02 to the surroundings and reduce the influx of outside air
into the gas stream.
As in flash smelting, the Noranda process takes advantage of the
heat energy available from the charge itself. Air blown through the
tuyeres creates a turbulent, well-mixed zone in which iron sulfide
(FeS) is oxidized with the release of heat. The remaining thermal
energy required is supplied by coal, which is mixed with the ore
concentrates, and by two oil burners, one positioned at each end of
the vessel.
Due to the turbulence within the bath, slags from the furnace
contain high copper concentrations,(3 to 8 percent)66 and must be
further processed for copper recovery. Slags are cooled, ground, and
processed by froth flotation to produce a slag concentrate, which is
charged to the reactors, and a tailings product containing 0.4 percent
copper,66 which is discarded.
3-40
-------
Feeder
SO2
Off-Gas
Slag Settling
Concentrate
Pellets and Flux
Slag
Air Tuyere | Copper
Matte Slag
Figure 3-10. Noranda continuous smelting.
3-41
-------
The Noranda process is not used as originally designed to produce
blister copper directly because of problems encountered with impurity
elimination.65 67 Impurity elimination in the Noranda process is dis-
cussed in Section 3.5.
3.2.5.2 Mitsubishi Process. The Mitsubishi smelting process was
developed by Mitsubishi Metal Corporation of Japan. Early development
work was carried out in a pilot plant, constructed in 1968, processing
65 Mg/day (72 tons/day) of charge. Subsequent tests of the process
were made with a prototype scale unit, completed in 1971, processing
150 Mg/day (165 tons/day) of charge.68 Presently, Mitsubishi operates
a commercial plant rated at 650 Mg/day (720 tons/day) feed capacity,
which went into operation in 1974 at its Naoshima smelter in Japan.69
A second full-scale production plant is in operation at the Texasgulf
Canada smelter in Timmins, Ontario.70
As shown in Figure 3-11, the Mitsubishi process employs three
furnaces interconnected by a continuous flow of matte and slag. The
furnaces are connected in cascade fashion so that matte and slag flow
by gravity between them. In the first or smelting furnace, dried
copper concentrates are smelted and oxidized to form a high grade (60
to 65 percent Cu) matte. A mixture of slag and matte flows from the
smelting furnace to an electric settling furnace, in which the matte
and slag are separated. Slag (0.5 percent copper) flows from the
electric furnace, and is discarded. Matte from the electric furnace
flows into the final or converting furnace, where it is continuously
oxidized to blister copper. Due to its high copper content (15 percent
Cu) the slag from this furnace is solidified and recycled to the
smelting furnace for copper recovery.
Oxygen-enriched air is introduced into the smelting and converting
furnaces via vertical, stainless steel lances, which are installed in
the roof of each furnace.
A major advantage of the Mitsubishi process over other processes
is that materials handling operations are greatly reduced, since most
matte and slag transfer is by gravity flow. (The water-granulated
slag from the converting furnace is recycled to the smelting furnace
3-42
-------
Concentrates, Silica Converting Furnace Slag Air. Oxygen,
Flux, Air, Oxygen Granulation and Recycle CaC03 Flux
u — \
-b—•—^ Coke
T T and \
Burner Slag FeS2
Blister
Smelting Furnace
Electric Slag
Cleaning Furnace j Converting Furnace
Slag Granulation
and Discard
Exit Gas
Concentrates, SiO2, Air, 02
Granulated
Revert Slag
T Exit Gas
Air, O2, CaCO3
Smelting Furnace
Matte and
Slag
Blister Copper
Converting Furnace
Electric Slag
Cleaning Furnace
Discard Slag
Granulation
Figure 3-11. Mitsubishi continuous smelting.
3-43
-------
by means of a bucket conveyor system.) Also, because the process
makes use of the inherent energy content of the sulfide charge, its
energy requirement is among the lowest of the major pyrometallurgical
processes.
A possible disadvantage to this process is that, if the feed con-
tains high levels of certain impurities, adequate impurity elimination
may not be achievable. This point is discussed further in Section 3.5.
3.3 EMISSIONS FROM PRIMARY COPPER SMELTERS
3.3.1 General
S02 and particulate emissions from primary copper smelters can be
categorized as either process or fugitive emissions. Process emissions
include primary offgas emissions, from roasting, smelting, and converting.
Fugitive emissions include those escaping from material transfer
operations, leakage from process vessels, and leakage from primary
offgas flues. Fugitive emissions may be considered low-level emissions
because they usually escape at or near ground level. The process
emissions are typically discharged through a tall stack.
3.3.2 Process Emissions
If uncontrolled, process sources account for the majority of
primary copper smelter emissions. Uncontrolled emissions factors for
S02 and particulate matter from roasting, smelting, and converting
operations are shown in Table 3-3.
Under the existing new source performance standards (NSPS) regula-
tion, control equivalent to that attained using double-contact acid
plants is required for new roasters, converters, and smelting furnaces
processing a charge containing a low level of volatile impurities. At
this level of control, S02 emissions from these sources are reduced by
approximately 98.5 percent, while particulate emissions are reduced by
greater than 99 percent. However, reverberatory furnaces processing a
charge containing a high level of volatile impurities are not subject
to control of S02 or particulate matter. As noted in Table 3-3, these
furnaces represent a significant fraction of process emissions from
copper smelters.
3-44
-------
TABLE 3-3. EMISSIONS FACTORS FOR UNCONTROLLED MAJOR
PROCESS SOURCES
Mass S02 per unit Mass participate perb
of blister copper unit of blister copper
Operation type
Roasting
Smelting (reverberator
kg/Mg
780
y 460
Ib/ton
1,560
920
kg/Mg
90
40
Ib/ton
180
80
furnace)
Converting 1,160 2.320 120 240
aBased on an average of sulfur elimination data for the ASARCO-E1 Paso,
ASARCO-Hayden, ASARCO-Tacoma, and Phelps Dodge-Douglas smelters. All of
these plants use multihearth roasters.
bAdapted from Reference 71 and assumes feed contains 25 percent copper.
3-45
-------
3.3.3 Fugitive Emissions
Potential sources of fugitive emissions of particulate matter
and, in most cases, S02 are listed in Table 3-4 and shown in Figure 3-12.
The actual quantities of emissions from these sources depend to some
extent on the type and condition of the equipment and the operating
techniques employed by the smelter. Although emissions from many of
the sources are released into a building, they are ultimately discharged
to the outside. Each of the potential sources is discussed briefly
here. The discussion applies to sources without fugitive emissions
capture systems (hooding). It should be noted that a number of judg-
ments are presented, based on U.S. Environmental Protection Agency
(EPA) inspections and visible observations, where mass emissions data
are not available.
3.3.3.1 Roasters.
3.3.3.1.1 Charging. Fugitive emissions from the charging of
multihearth roasters are generally minimal. Particulate emissions are
slight because of the high moisture content (5 to 15 percent) of the
feed. The escape of S02-laden gases from the interior of the roaster
through the annular charging port is effectively prevented by the flow
of material cascading from the uppermost drying hearth to the first
roasting hearth and by the operation of the roaster under negative
pressure.
As with multihearth roasters, fugitive particulate emissions from
the charging of fluid-bed roasters are slight—both because of the
suppressive effect of the moisture contained in the feed materials and
because charging systems are generally totally enclosed.
3.3.3.1.2. Leakage. Fugitive emissions from multihearth roasters
may be emitted from leaks around the doors located at each one of the
hearth levels, from holes in the actual shell of the roaster, or from
leaks around the central drive shaft. Under normal operation, these
emissions are minimized by operating the roasters under a slight
negative pressure and by good maintenance practices.
The fluid-bed roaster is essentially a vertical cylinder of steel
plate lined with insulation and fire bricks. Because it operates
3-46
-------
TABLE 3-4. POTENTIAL SOURCES OF FUGITIVE EMISSIONS
Roaster
Charging
Leakage
Hot calcine discharge and transfer
Smelting furnace
Charging
Leakage
Matte tapping
Slag tapping
Converter slag return
Converters
Charging (matte, reverts, flux, lead smelter by-products, cold dope
or other)
Blowing (primary hood leaks)
Skimming
Holding
Pouring of slag and blister
Converter leaks
Anode furnace
Charging
Blowing (oxidation and poling modes)
Holding
Pouring
Mi seellaneous
Dust handling and transfer
Ladles
Slag dumping
3-47
-------
Concentrate Storage
i
-^
oo
Silica
Unloading
Reverb Matte Tapping
and Slag Skimming
Roaster
Leakage
Roaster
Unloading and
Concentrate Handling
Converter
Leakage
Anode Furnace
Converter
Charging
Calcine Transfer
Charge of Blister to
Anode Furnace and Slag
Skimming and Handling
Copper Tapping
and Casting
1. Larry Car
2. Conveyer
3. Wagstaff Feeder
Slag and
Blister
Tapping
Matte Transfer
to Converters
Slag to Dump
To Refinery
Reverberatory Furnace
Figure 3-12. Fugitive emissions sources for primary copper smelters.
-------
under a positive pressure, the unit is designed for containment of the
material, and leakage from the vessel proper is usually negligible
with proper maintenance. Because of their high S02 concentration, gas
leaks are readily detected if present.
Calcine is discharged from a fluid-bed roaster primarily by
entrainment in the gas exiting the top of the roaster. The material
is collected from the gas stream using a series of cyclones. The
gas-handling system is an integral part of the roaster. As the entire
system is under positive pressure, it should therefore be airtight and
free of leaks. However, as hot calcine is both corrosive and abrasive,
flue leakage can be a problem, resulting in some fugitive emissions if
proper maintenance is not applied. OSHA has indicated that substantial
leakage occurs from existing units at the expansion joints and cyclones.72
3.3.3.1.3 Hot calcine discharge and transfer. Fugitive emissions
may be generated during the discharge and transfer of hot calcine from
the roaster to the smelting furnace. Smelters with multihearth roasters
usually use larry cars (small rail cars) to transport calcines to the
furnace. When the calcine is dropped from the hopper located beneath
the roaster into the larry car through the feed opening, large quanti-
ties of dust are generated as a result of material movement and pressure
changes within the car.73 Fugitive emissions can also occur during
the transportation of the roaster calcines to the smelting furnace.
In the case of larry cars, the feed opening is usually covered to
minimize this effect.73
Calcine collected from the cyclones associated with a fluid-bed
roaster is generally fed by a closed system to a calcine storage bin
located in close proximity to the smelting furnace. Again, because
this system is totally enclosed, fugitive emissions are generally
negligible.
3.3.3.2 Smelting Furnaces. As mentioned previously, four basic
types of smelting furnaces are used by the industry: reverberatory,
electric, flash, and Noranda furnaces. The following is a discussion
of the fugitive emission sources associated with these furnaces.
3-49
-------
3.3.3.2.1 Charging. Fugitive emissions associated with the
charging of smelting furnaces may be substantial, depending upon the
type of furnace being used, the nature of feed materials charged, and
the charging technique.
When green or calcine charge contacts the molten furnace bath,
rapid reactions with the bath lead to gas formation. This in turn can
cause positive pressure surges within the furnace, which can result in
the release of fugitive emissions through all of the furnace openings.
In the case of side-charged furnaces using green feed, there is the
added possibility of a portion of the charge bank caving in or sloughing
into the molten bath, which results in a similar rapid gas release.
In such cases, the pressure surge can be great enough to damage the
furnace arch. Generally, however, it is believed that fugitive emis-
sions associated with charging calcine feed are greater than those
from charging green feed, because of the comparatively dusty and
free-flowing nature of hot calcine.
With reverberatory furnaces, the method of charging can vary
depending upon whether green feed or calcine feed is used. Green-feed
furnaces are most commonly charged using drop pipes that penetrate the
furnace roof. Green feed can also be charged through openings in the
furnace sidewalls using charge slingers (high-speed conveyors). With
the latter system, it is possible for fugitive emissions to be released
as the charge is thrown through the opening.
Reverberatory furnaces processing calcine feed can be charged by
(1) fixed or retractable Wagstaff guns, which penetrate the furnace
sidewalls; (2) drag-chain conveyors in conjunction with fettling pipes
along the sidewalls; and (3) feed pipes penetrating the arch of the
furnace near the centerline.
Electric furnaces are usually charged continuously through feed
pipes in the roof. The charge usually consists of calcine or dried
concentrates. Pressure surges can occur during charging in electric
furnaces as well as in reverberatory furnaces; however, generally
fewer openings are present on electric furnaces.
3-50
-------
The feed to flash furnaces is usually from a concentrate dryer.
Dried concentrates from the dryer are discharged to the feed storage
bin by a closed system. The feed is then conveyed to the flash furnace
using variable speed drag conveyors or screw conveyors where it is
injected with air into the furnace. Because the system is designed to
be gas tight, fugitive emissions are not normally emitted.
Noranda reactors are charged through an opening in one end by
means of a charge slinger (high speed conveyor). Some fugitive emis-
sions are emitted as the charge is thrown through the opening,67
although the amount is generally small since the charge is moist.
3.3.3.2.2 Leakage. Fugitive emissions, especially S02, can
result from leakage points on most types of smelting furnaces when the
pressure inside exceeds atmospheric pressure. Leakage points include
thermal expansion spaces between bricks and all other furnace openings.
Reverberatory furnaces have perhaps the greatest potential for
leakage if adequate draft is not maintained on these furnaces. However,
the draft cannot be excessive because outside air causes the furnace
temperature to drop. In addition to thermal expansion spaces and
charging ports, openings exist for admitting secondary combustion air
to the burners. On those furnaces having a roof constructed of silica
brick, ports are present along the furnace length to allow silica
slurry to be sprayed onto the arch for maintenance.
Electric furnaces generally do not have as many openings as do
reverberatory furnaces. Secondary air openings are not necessary
because burners are not required. Also, arch maintenance ports like
those used on reverberatory furnaces are not necessary because the
roof temperature is usually not a problem. Potential leakage points
on electric furnaces are the expansion spaces, spaces around the seals
where the electrodes enter the furnace, and charging ports.
Inco flash furnaces are totally encased in a mild steel shell.74
Hence, fugitive emissions associated with furnace leakage are unlikely.
Outukompu flash furnaces are also steel encased, except for their
roofs. Potential leakage points other than expansion cracks include
spaces around the plug in the uptake ceiling and the damper slot
between the furnace and the waste heat boiler.
3-51
-------
With Noranda reactors, which are encased in a steel shell, leakage
can occur only through openings such as the charge port or around the
hood over the offtake. However, emissions are slight if adequate
draft is maintained on the primary offtake hood.
3.3.3.2.3 Matte tapping. Matte tapping is a source of fugitive
S02 and participate emissions from smelting furnaces. Smelting furnaces
typically have from two to six matte-tapping ports with associated
launders. The launder directs the flowing matte to a location where
it is discharged into a ladle. Normally, only one tap port is used at
a time, although two ports may be used concurrently. Typically, a
single matte-tapping operation lasts from 5 to 10 minutes. Matte-
tapping frequency varies with furnace capacity. Matte is tapped from
reverberatory furnaces with a frequency of from five to eight times
per 8-hour shift. The Outokumpu flash furnace at the Phelps Dodge-
Hidalgo smelter is tapped with a frequency of from 10 to 20 times per
8-hour shift.75 During matte tapping, fugitive emissions are visible
at the tap port, along the launder, and at the launder-to-ladle discharge
point.
3.3.3.2.4 Slag skimming. Slag skimming is another source of
fugitive emissions from smelting furnaces. Smelting furnaces typically
have from one to three slag-skimming ports. As many as two may be
used concurrently. A single slag-skimming operation usually lasts
from 10 to 20 minutes. Slag is skimmed with a frequency of from 10 to
25 times per 8-hour shift. As with matte tapping, emissions are
evident at the skimming port, along the launder, and at the launder-to-
ladle discharge point.
3.3.3.2.5 Converter slag return. Reverberatory, electric, and
Inco flash furnaces generally have a single converter slag return port.
in the furnace wall. Converter slag is returned to the furnace using
a launder or chute leading to the opening. Fugitive emissions result
as the slag flows from the ladle to the furnace port. Also, some emis-
sions may escape from within the furnace through the open port. Fhe-e
emissions stem from pressure surges within the furnace, which are caused
by chemical reactions between the converter slag and the bath.
-------
In Outokumpu flash and Noranda furnaces, converter slag is usually
processed separately in slag cleaning furnaces or flotation plants.
3.3.3.3 Slag-Cleaning Furnaces.
Slag-cleaning furnaces are frequently used in conjunction with
flash furnaces for recovering matte entrained in the smelting furnace
and converter slags. Slag-cleaning furnaces are most commonly small,
low-powered electric furnaces. Potential fugitive emission points on
these furnaces are the same as those on smelting furnaces.
The charging of molten slag to the slag-cleaning furnace from the
smelting furnace is performed with the same frequency as this slag is
skimmed from the smelting furnace. At the Phelps Dodge-Hi rial go smelter,
slag from the flash furnace is transferred directly into the slag-cleaning
furnace via a launder leading to an open port in the wall of the
slag-cleaning furnace. Fugitive emissions can be observed during the
entire operation.
Fugitive emissions from furnace leakage can occur whenever insuffi-
cient draft is maintained. Leakage occurs primarily through the
furnace roof, because of the presence of expansion cracks and spaces
around the electrodes.
Matte is tapped from slag-cleaning furnaces in the same fashion
as from smelting furnaces. However, the frequency of tapping, at
about three times per 8-hour shift,75 is lower because only a small
fraction of the matte produced in the smelter is entrained in slag.
The grade of matte from a slag-cleaning furnace is slightly higher
than that produced in the smelting furnace, because some additional
sulfur is removed in the slag-cleaning furnace. Hence, fugitive
emissions per unit of matte tapped from slag-cleaning furnaces would
be expected to be slightly less than those liberated (per unit of
matte) from matte produced in a smelting furnace.
Slag skimming, like matte tapping, is also performed in much the
same fashion on slag-cleaning furnaces as on smelting furnaces. Hence,
fugitive emissions are evident at the skimming port, along the launder,
and at the launder-to-ladle discharge point. Ihe quantity of slag
o
-53
-------
skimmed from slag-cleaning furnaces is approximately the same as that
skimmed from smelting furnaces, because only a comparatively small amount.
of matte settles from the slag charged.
Converter slag is returned to slag-cleaning furnaces in the same
manner as it is returned to smelting furnaces. Fugitive emissions
result as the slag flows down the chute and into the furnace. The
frequency of tins operation on slag-cleaning furnaces would be somewhat
lower, as compared its frequency on reverberatory furnaces, however.
This decrease results from the higher matte grade (and consequent
decrease in converter slag production) in most smelters employing
slag-cleaning furnaces.
3.3.3.4 Converters.
The various stages of converter operation are charging, blowing,
slag skimming, and blister pouring. Each of these operations is a
potential source of fugitive emissions.
3.3.3.4.1 Charging. During charging, the converter is rotated
until its mouth is approximately 45 degrees from the vertical, and the
primary hood is raised. Emissions result for an instant as the converter
is rotated because of the need to maintain blowing air through the
tuyere lines until the tuyeres are above the level of the bath to
prevent plugging. An overhead crane lifts the ladle above the mouth
of the converter and pours the charge (matte or revert materials) into
the converter by tilting the ladle. During the pour, visible emissions
are heavy but of relatively short duration (15 to 20 seconds). When
charging is completed, blowing air through the tuyeres is resumed,
which results in a burst of emissions while the vessel is rotated to a
vertical position. Once in place, the primary hood is then lowered
into position.
3.3.3.4.2 Blowing. Most domestic smelters have attempted to
provide relatively close-fitting primary hoods over the converter-
mouth to contain and capture the offgases generated during blowing
operations. However, these hoods do not completely seal the opening
because of irregularities around the mouth. The irregularities are
caused by accretions formed by pouring operations and by bath splatter
3-54
-------
during blowing. Fugitive emissions escape from these openings.
Generally, the emissions are proportional to the blowing rate and the
condition of the primary hood.
3.3.3.4.3 Skimming. During skimming operations, the mouth of
the converter is rotated to a position between 65 and 85 degrees from
the vertical, depending upon the bath level. Some emissions occur
during the brief roll-out period because the blowing air to the tuyeres
is maintained until they are above the bath level. Slag is skimmed
from the converter mouth into a slag ladle. Fugitive emissions are
visible during the entire skimming operation, which typically lasts
from 2 to 3 minutes. Although the primary hood is not necessarily
retracted during this operation, it is usually isolated by dampers
from the main duct system to prevent dilution air from mixing with the
S02 gases being collected from other blowing converters. At the
completion of the skim, blowing is resumed, and the converter is
rotated back to the upright position. Once in place, the primary hood
is lowered.
3.3.3.4.4 Pouring. During blister copper pouring operations,
the converter is rotated downward until the mouth reaches a position
approximately 90 to 125 degrees from the vertical, depending upon the
volume of blister copper within the converter. Again, emissions are
discharged briefly during the roll-out, because blowing air is main-
tained until the tuyeres are above the liquid level. Steady fugitive
emissions are observed as the copper is poured into the ladle. After
its contents are emptied, the converter is rotated upward until the
mouth reaches a position approximately 45 degrees from the vertical to
await a new matte charge and the start of a new cycle.
3.3.3.4.5 Holding. At times during normal smelting operations,
slag or blister copper cannot be transferred immediately to the ladles.
This condition may occur for several reasons, including insufficient
matte in the smelting furnace, the unavailability of a crane, and
others. Under these conditions the converter is rolled out (rotated)
about 30 to 45 degrees to raise the tuyeres above the bath. An
3-55
-------
auxiliary burner may be fired through the mouth to keep the bath hot.
While in the holding mode, fugitive emissions from the molten bath
escape from the mouth into the converter building.
3.3.3.4.6 Converter leaks. Since the ends of most Peirce-Smith
converters are joined by bolts and springs, they occasionally leak at
the end joint. When this leakage is below the molten material surface,
it is usually repaired rapidly to prevent major erosion. However, in
cases where it occurs above the bath surface, it may not be repaired
in a timely fashion. Thus, fugitive emissions may occur at this
point.
3.3.3.5 Anode Furnaces. Refining of blister copper to anode
copper is performed in either rotary or hearth-type furnaces. Emissions
from these furnaces occur during all phases of operation: charging,
oxidizing, poling, skimming, holding, and pouring. In the hearth-type
furnaces, which are used at the ASARCO-Tacoma and Kennecott-Hurley
smelters, the primary offgases generated during the actual refining
operation (oxidation and reduction blows) are siphoned through a
furnace offtake and vented through a stack. In the case of the more
conventional rotary-type refining furnaces, which are similar to
Peirce-Smith converters, emissions during blowing operations vent
through the open mouth of the vessel.
3.3.3.6 Miscellaneous Sources.
3.3.3.6.1 Dust handling and transfer.76 Dust-handling and
transfer operations can generate fugitive emissions if carelessly
performed. However, most smelters take precautions to minimize fugitive
emissions from dust handling anc? transfer. Dust transfer from control
devices to storage facilities is usually performed by covered conveyors.
Dust transfer from storage bins is usually made through dust-tight
connections to surface transportation units such as tank trucks and
dumpsters. Cleaning and unloading of dust from flues and settling
chambers is performed by enclosed conveyors that feed into hoppers
provided at spaced intervals underneath the flues and settling chambers.
Both screw- and drag-type conveyors are used. These flue dusts are
3-56
-------
usually treated in a pugmill or pelletizing disc where moisture is
added. The wet dust is then transferred to a bedding area, blended
with other feed constituents, and recycled. Dust from waste heat
boilers and crossover flues is usually removed by manual methods,
which, if properly implemented, result in minimal fugitive emissions.
3.3.3.6.2 Ladles. Visible emissions can be observed from ladles
containing molten materials (matte, slag, or blister copper) as they
are transported between process stages within the smelter. Emissions
from a particular ladle are generally of short duration because the
time required for material transport is usually short. Normal process
fluctuations may require that ladles containing matte, slag, or blister
copper be temporarily set aside until needed. The emissions from
fuming are still short-lived, however, since the exposed surface of
the material cools rapidly, forming a solidified layer that greatly
limits fugitive emissions.
3.3.3.6.3 Slag dumping.77 Smelting furnace slag is disposed of
by water granulation or by transport in the molten state for dumping a
short distance from the smelter. Slag dumping is the more widely used
method. The slag is transported to the dump site by train or slag
hauler (Kress hauler). The slag train is usually comprised of a
number of slag pots or ladles on flat cars. Solidification at the
surface of the slag in the pots is fairly rapid. Fugitive emissions
during transportation to the dumping site are therefore limited.
However, during dumping of slag at the dumping site, substantial
fugitive emissions, although short in duration (less than 1 minute),
can be observed.
3.3.4 Summary of Fugitive Emissions Data
Emissions factors for fugitive S02 and particulate matter have
been developed78 for most of the major fugitive emissions sources at
primary copper smelters: roaster calcine discharge, matte tapping,
slag skimming, converter slag return, converter operations, and anode
furnaces. Summaries of these emissions factors are presented in
Tables 3-5 and 3-6. There are no known emissions data for sources not
included in Tables 3-5 and 3-6.
3-57
-------
I
CJ1
00
TABLE 3-5. SUMMARY OF FUGITIVE S02 EMISSIONS FACTORS FOR PRIMARY
COPPER SMELTING OPERATIONS79
S02 emission factors
Source
Calcine discharge
Matte tapping
Slag skimming
Mass per unit
of material processed
0.5 kg/Mg of calcine
(1.0 Ib/ton of calcine)
2.6 kg/Mg of matte
(5.2 Ib/ton of matte)
0.24 kg/Mg of slag
Mass per unit
Conventional
2.2
7.0
0.6
of blister cojpjper,
smelters
(4.4)
(14.0)
(1-2)
kg/Mg
(Ib/ton)
New smelters
4.0
0.8
(8-0)
(1-6)
Converters
Blowing segment
only
Total converter
cycle
Converter slag
return
Anode furnace
a
(0.48 Ib/ton of slag)
2,198 kg/blowing hour
(4,845 Ib/blowing hour)
2,060 kg/'h
(4,538 Ib/h)
0.14 kg/Mg of slag return
(G.?8 ih'ton of slag return)
241
320
(482)
(640)
0.2
0.15
(0.4)
(0.30)
130
190
0.15
Except for converters where mass emissions per hour are given.
Calcine or "green" feed reverberatory smelters.
'Flash furnace or Noranda reactor smelters.
Emissions data pertain to converters with primary hooas on'y (i.e., no secondary hoods)
(260)
(380)
(0.30)
-------
TABLE 3-6. SUMMARY OF FUGITIVE PARTICULATE EMISSIONS FACTORS FOR PRIMARY
COPPER SMELTING OPERATIONS80
IP
l£>
Source
Calcine discharge
Matte tapping
Slag skimming
Converters
Blowing segment
only
Total converter
cycle
Converter slag
return
Anode furnace
Mass per unit
of material processed
1.2 kg/Mg of calcine
(2.4 Ib/ton of calcine)
0.13 kg/Mg of matte
(0.26 Ib/ton of matte)
0.13 kg/Mg of slag
(0.20 Ib/ton of slag)
61 kg/blowing hour
(134 Ib/bl owing hour)
69 kg/h
(153 Ib/h)
N/A
Particulate emission
Mass per unit
Conventional
5.2
0.34
0.31
6.6
10
N/A
0.95
factors
of blister copper,
smelters
(10.4)
(0.68)
(0.62)
(13.2)
(20)
(1-95)
kq/Mq (Ib/ton)
New smelters
0.2 (0.4)
0.4 (0.8)
4.0 (8.0)
7.3 (14.6)
N/A
0.95 (1.95)
Except for converters where mass emissions per hour are given.
Calcine or "green" feed reverberatory smelters.
'F~!ash furnace or Noranda reactor smelters.
Emissions data pertain to converters with primary hoods only (i.e., no secondary hoods).
-------
The emissions factors developed for each source are based primarily
on emission tests conducted by EPA at several domestic primary copper
smelters. These tests were conducted in flues associated with the
local ventilation systems used to capture fugitive emissions from the
various sources investigated.
Using the emissions test data obtained and pertinent process
information, emissions factors were first computed in terms of mass
emissions per unit of material processed by the source.78 In the case
of sources for which more than one data point was available, an arith-
metic average was used to compute the emissions factor. Because the
capture systems tested were less than 100 percent effective, the
emissions factors were adjusted upward based on a subjective estimate
of the capture effectiveness of each capture system tested to account
for emissions escaping capture. The resultant uncontrolled fugitive
emissions factors, expressed in terms of mass emissions per unit of
material processed, were then normalized to mass emissions per unit of
end product (blister copper) by using representative plant material
balances.78 On this basis, the various fugitive emissions sources are
readily comparable as to relative significance.
The normalized emissions factors (mass emissions per unit of
blister copper) are distinguished in Tables 3-5 and 3-6 according to
smelter type--"conventional" smelters, representing green- and calcine-
charged reverberatory furnace operations, and "new" smelters, represent-
ing flash furnace and Noranda reactor installations.
Based on the normalized S02 emissions factors for conventional
smelters presented in Table 3-5, converters* are by far the greatest
source of fugitive S02 emissions at copper smelters. Fugitive emissions
associated with the entire converter cycle are more than 45 times
greater than those associated with the other fugitive sources. Con-
verter blowing accounts for the majority of the emissions during the
converter cycle. Second to converter emissions in relative signifi-
cance are fugitive emissions from matte tapping, followed by calcine
*Data pertain to converters without secondary hooding.
3-60
-------
discharge and slag skimming. Fugitive S02 sources having the lowest
relative significance among those for which data are available are
converter slag return and anode furnaces.
With regard to new smelters, the distribution of fugitive S02
emissions among the various sources is similar to that noted for
conventional smelters. Fugitive S02 emissions from converters are,
again, more than 45 times greater than those from the other sources.
It is noted that fugitive emissions of S02 from most sources at new
smelters are lower in magnitude, however, than those from the same
sources at conventional smelters. This decrease is due to the higher
matte grade produced in Noranda reactors and flash furnaces, which
corresponds to increased sulfur elimination in the process offgas
stream from these furnaces, and hence a lower availability of sulfur
for fugitive emissions.
Based on the normalized particulate matter emissions factors for
conventional smelters shown in Table 3-6, converters are the greatest
source of fugitive particulates at copper smelters. The blowing phase
of the converter cycle contributes the majority of the particulate
emissions. Calcine discharge represents another major source of
particulates, being approximately half as significant as converters.
The other sources—matte tapping, slag skimming, and anode furnaces—are
relatively minor by comparison. For the "new" smelters (which do not
normally employ roasters), converters are the major source of fugitive
particulate matter.
Data are not available to characterize fugitive* emissions during
matte tapping, slag skimming, and converter slag return operations on
slag-cleaning furnaces. However, fugitive S02 and particulate emissions
per unit of blister copper can be estimated based on the frequency at
which these operations are performed on slag-cleaning and smelting
*WHh regard to process offgas emissions from slag-cleaning fur-
naces, emissions data supplied by Phelps Dodge Corporation81 indicate
that S02 emissions are approximately 3 kg/Mg (6 Ib/ton) blister copper,
while particulate emissions are between 9.4 and 22.6 kg/Mg (18.7 to
45.3 Ib/ton) blister copper.
3-61
-------
furnaces. Based on previous discussions of relative frequencies,
fugitive S02 and particulate emissions from matte tapping on slag-
cleaning furnaces would be expected to be no more than about one-third
the magnitude of matte-tapping emissions from flash furnaces, while
fugitive S02 and particulate emissions from slag skimming are expected
to be approximately equal to those from the same operation on flash
furnaces. Fugitive S02 and particulate emissions per unit of blister
copper from converter slag return operations on slag-cleaning furnaces
are expected to be somewhat less than those from the same operation at
reverberatory furnaces.
3.4 EXPANSION OPTIONS FOR EXISTING FACILITIES
Most primary copper smelters have at least one rate-limiting
operation or "bottleneck." In most smelters, the bottleneck is the
capacity of the smelting furnace(s). If additional throughput is
needed from the smelter, additional capacity can be added either
through the installation of new process units (i.e., new roasters,
smelting furnaces, and converters) or through the expansion of the
existing rate-1imiting equipment.
In lieu of installing new process technology when additional
capacity is needed, the smelting industry, both worldwide and domestic,
has traditionally chosen to add capacity by expanding existing equip-
ment. It should be noted that the elimination of a particular bottle-
neck often creates another bottleneck, which must also be eliminated.
For example, if the rate-limiting operation is smelting furnace capa-
city and the furnace is expanded, additional converters may be required,
depending upon the magnitude of the increase in furnace capacity.
Various expansion options available for roasters, smelting furnaces,
and converters are discussed below. In general, S02 and particulate
emissions on an uncontrolled basis increase proportionately to increases
in capacity.
3.4.1 Multihearth Roasters
Increasing the capacity of multihearth roasters by increasing
their shaft rotation speed has been achieved at the Noranda Smelter in
Quebec, Canada.82 These roasters are not substantially different from
3-62
-------
those used at domestic smelters. The Wedge roasters used are 7.60 m
(25 ft) in diameter, with an external feed or drying hearth and se\fen
internal hearths. Two air-cooled arms are positioned directly above
each hearth, one arm carrying 10 and the other 11 rabble blades. Each
roaster is driven by a 15-hp motor belt-connected to a speed reducer.
The drive is provided with a shear-pin arrangement to prevent breaking
of the arms in case of overload or jamming in the roaster.
Rated at 136 Mg/day (150 tons/day) of concentrate, each roaster
is operated with a shaft rotation speed of 0.75 rpm. Roaster throughput
was increased to about 295 Mg/day (325 tons/day) after experimental
work was done to determine the maximum tonnage of feed that could be
accommodated with satisfactory sulfur elimination. At this capacity,
the roaster shaft speed was 1.09 rpm.
Increasing the throughput was found to raise the temperatures
throughout the roaster, with a maximum of 760° to 820° C (1,400° to
1,500° F) being achieved on the third and fourth hearths. The increased
temperatures combined with an increase in gas volume resulted in an
accumulation of primarily oxidized pyrite fines on the underside of
the second and fourth hearths. This problem was rectified by attaching
a plow to the top of the rabble arm on the next lower hearths to
mechanically eliminated any buildup.
The experience at Noranda Mines illustrates that increasing the
rotation speed of the shaft can result in a significant increase in
roaster capacity. Increasing the shaft rotation speed will also
decrease the residence time and increase the roaster temperature.
Both of these parameters are critical, because they influence the
volatilization of impurities and the degree of sulfur elimination
achieved. Most domestic smelters use the roaster speed as a means of
obtaining a desired degree of sulfur elimination. Hence, it is con-
cluded that changing the shaft rotation speed to increase multihearth
roaster capacity would not be a viable expansion option at most domestic
smelters.
Physical expansion of multihearth roasters is not considered
feasible because of the geometry of these units. It is concluded that
no viable expansion options are available for multihearth roasters.
3-63
-------
3.4.2 Fluid-Bed Roasters
Major components of fluid-bed roasters include the blower, the
roasting vessel, and the calcine recovery system (which may use up to
16 cyclones). The entire system is sized very closely to the rated
throughput.83 It is conceivable to attempt to increase roaster capacity
by increasing the capacity of the blower. However, increased throughput
might be precluded by (1) excessive wear in the cyclones due to abrasion
and (2) the capacity of the calcine recovery system.83 The potential
for these problems indicates that this scheme would have to be evaluated
on a case-by-case basis at domestic smelters. Such an evaluation is
considered beyond the scope of this analysis.
Oxygen enrichment of the fluidizing air has been used as a means
of increasing the capacity of sulfating fluid-bed roasters operated at
copper smelters in Australia and Zambia.83 Such roasters operate under
the conditions necessary to produce copper sulfate from the sulfide
charge. Based on this experience, oxygen enrichment could potentially
be used to increase throughput at domestic units by 20 to 25 percent.83
However, oxygen enrichment could lead to localized overheating within
the bed which causes incipient melting of the feed in some cases.83
Such melting can lead to bed defluidization and plugged cyclones.
Also, it should be noted that this expansion option has not been
demonstrated on fluid-bed roasters like those operated by the domestic
primary copper industry.83 For these reasons, it is concluded that
oxygen enrichment is not a viable expansion option for fluid-bed
roasters in general. Rather, the potential usefulness of this option
would have to be determined on a case-by-case basis.
In conclusion, it appears that no viable options are available
for increasing the capacity of existing fluid-bed roasters.
3.4.3 Reverberatory Furnaces
A number of options have been used in the past to increase rever-
beratory furnace capacity, including the conversion from green to
calcine charging, the addition of concentrate dryers, physical expansion
of the furnace, the elimination of converter slag return to the furnace,
and the use of various oxygen-enhancement techniques. Each of these
options is discussed below.
3-64
-------
3.4.3.1 Conversion from Green to Calcine Charging. Reverberatory
furnaces operating on green (unroasted) charge may increase throughput
by converting to calcine-charged operation through the addition of
roaster capacity. Throughput is increased because less heat per unit
of feed is required to smelt hot calcines than cold, moist concentrates.
Consequently, at a given level of heat input, a greater tonnage of
calcine can be processed in a given period of time. Considerable
savings in energy, as well as enhanced sulfur recovery, are also
afforded.
In 1969, the Kennecott Corporation modified its smelter at Hayden,
Arizona, by adding a Dorr-Oliver fluid-bed roaster.84 Use of the
roaster resulted in a 50-percent increase in reverberatory furnace
capacity over that achieved with green-charged operation. The roaster
is fed with a mixture of copper concentrates, copper precipitates, and
silica flux. The bulk of the roasted calcines is collected from the
exhaust gases using eight primary cyclones and eight secondary cyclones.
The collected calcine reports to calcine bins on each side of the
reverberatory furnace and is fed to the furnace by Wagstaff feeders.
During the conversion to calcine-charged operation, the reverberatory
furnace was altered by installing water jackets (water-cooled panels)
around the furnace perimeter at the slag line. These panels were
needed to protect the sidewall refractory from excessive temperatures
resulting from the close proximity of the burners. Before the conver-
sion, charge banks of green feed along the sidewalls provided the
necessary protection.
A similar conversion to calcine-charged operation performed at
the Cities Service Company smelter in Copperhill, Tennessee, in 1961
increased the capacity of a reverberatory furnace by 40 percent.85 In
addition, the conversion enabled the overall smelter sulfur recovery
to be increased from 63 to 85 percent because the roaster gases could
be processed in a sulfuric acid plant.
The Copperhill fluid-bed roaster, designed by Dorr-Oliver, operates
with slurry feed. By changing the percentage of water in the slurry,
sulfur elimination in the roaster can be controlled so that the matte
3-65
-------
grade from the smelting furnace can be readily varied from 30 to over
50 percent. Calcine is removed from the roaster offgas stream using
two primary cyclones and two secondary cyclones and is transferred to
a bin. From the roaster discharge bin, calcine is fed intermittently
to the reverberatory furnace using a single Wagstaff gun inserted
through a sliding door in the side of the furnace.
In this analysis, an increase in throughput of 40 percent is
assumed achievable when converting from green- to calcine-charged
operation. Based on the practice of Kennecott-Hayden and Cities
Service-Copperhil1, it is assumed that Wagstaff gun feeders would be
used for charging calcine to the furnace in lieu of side-charging of
calcines. It is further assumed that smelters making the conversion
would install water-cooled panels around the furnace perimeter at the
slag line to protect the sidewall refractory from excessive wear.84 86
3.4.3.2 Addition of Concentrate Dryers. Reverberatory furnaces
charged with green feed may also increase capacity to a limited extent
via the addition of dryer capacity. As with the conversion from green
to calcine charging, capacity is increased because less thermal energy
per unit of charge is required to smelt the dried, heated material.
Capacity increases are not as substantial, however, because concentrate
dryers do not increase the temperature of the feed material to the
same level as does a roaster. Kennecott has indicated that the use of
a concentrate dryer purchased for its McGill smelter will increase
furnace throughput by approximately 15 percent.8
3.4.3.3 Physical Expansion. Physically expanding a reverberatory
furnace is considered to be a technically feasible option for increasing
capacity. In the early 1970's, ASARCO increased the capacity of a
furnace at its Tacoma smelter by 20 percent by increasing the furnace
width.87 It should be noted, however, that the physical expansion of
a reverberatory furnace requires an extended furnace shutdown, with
possible adverse effects on smelter throughput. Also, this option may
not be useful at some smelters because of physical space limitations.
For these reasons, and because many industry representatives consider
this scheme to be impractical, physical expansion is not considered to
be a viable expansion option in this analysis.
3-66
-------
The quantity of sensible heat carried out of the furnace in
the combustion gases is decreased.
The concentration of S02 in the furnace offgases is increased.
This particular benefit is discussed further in Section 4.4.6.
The net result of these effects is an increase in furnace efficiency,
manifested primarily in an increase in the production rate. Other
benefits that can result with increased oxygen usage include a decrease
in the fuel requirements per unit of charge and a decrease in copper
losses in the slag. The latter benefit results from a decrease in
slag viscosity, which comes about through the increased slag tempera-
ture.*
3.4.3.5.2 Methods of oxygen introduction. Various methods of
oxygen introduction into reverberatory furnaces have been used to
date.
Enriching the primary combustion air with oxygen.
Undershooting the flame with oxygen or oxygen-enriched air.
Oxygen lancing through the roof.
Introducing oxygen directly with fuel in roof-mounted
oxy-fuel burners.
Introducing oxygen directly with dried feed in roof-mounted
oxy-sprinkle burners.
These schemes are illustrated in Figure 3-13.
The first scheme, oxygen enrichment of the primary combustion
air, refers to the addition of oxygen to the air supply of the existing
end-type burners. This method, because of its simplicity, requires
very little change to the furnace proper.
With the second scheme, undershooting the flame with oxygen or
oxygen-enriched air, oxygen lances (typically water-cooled) are
retrofitted into the end of the furnace. These lances are typically
positioned just below the existing burners.
*The decrease in copper loss via decreased slag viscosity could
be offset, in some cases, by (1) decreased residence time for settling
due to increased throughput and (2) the possibility of higher copper
solubilities in slag at higher temperatures.
3-68
-------
Oxygen Introduction to the
Primary Combustion Air
Primary Combustion Air for Burner
O O-
y////
Oxygen Enriched
Primary Combustion
Air for Burners
Oxygen-Enrichment of Primary Combustion Air
Undershooting of Flame with Oxygen
Fuel Burners
Oxygen Jets
Undershooting the Flame with Oxygen
Fuel Input
Oxygen Lances
Charge Banks
Oxygen Lancing of the Bath
, Oxy-Fuel Burners
Charge Banks
Oxygen-Fuel Burner Usage in the Furnace
Oxygen Sprinkle Burners.
1 1
1 1 1
y////////////////,
^— —
i
~— — ^
'//////,
Oxygen Sprinkle Smelting
Figure 3-13. Methods of oxygen addition.
3-69
-------
Oxygen lancing through the roof refers to the installation of
oxygen lances in vertical position on the roof of the furnace. This
scheme may use up to three rows of lances, which are spaced regularly
down the furnace length. The existing end-burners are used to intro-
duce a fuel-air mixture containing insufficient oxygen for complete
fuel combustion.
With the oxy-fuel scheme, fuel is mixed directly with commercial
oxygen in burners that are retrofitted onto the furnace roof. These
burners are positioned vertically, or nearly so, and are generally
spaced regularly down the length of the furnace in two rows. If a
sufficient number of oxy-fuel burners are used, the existing end-type
burners are not operated.
The oxygen-sprinkle scheme, as developed by Queneau and Schumann,yl
differs from the other schemes in that it operates on the same principle
as a flash furnace. Three specially designed burner^ positioned on
the furnace roof are used to introduce and disperse a mixture of
primarily dried concentrates and oxygen. A small percentage of ground
coal may be mixed with the feed. The heat required for smelting is
generated from the flash combustion of the mixture, and the molten
droplets fall to the hearth below.
3.4.3.5.3 Operating experience. One of the earliest investiga-
tions of the use of oxygen in reverberatory furnaces was described by
Saddington et al.92 of the Inco Copper Cliff smelter in Sudbury,
Ontario. Tests were made on a calcine-charged nickel reverberatory
furnace. Four water-cooled oxygen lances were installed in the end of
the furnace, one below each of the coal burners used to fire the
furnace. The lances were angled away from the furnace wall. By
introducing the oxygen below the burners, the hottest zone of the
flame was at the bottom next to the bath. The first test, made
primarily to examine fuel efficiency, indicated a 10-percent increase
in throughput could be accompanied by a 19-percent decrease in fuel
consumption. In the second test, an increase of 36 percent in through-
put was achieved with approximately the same fuel consumption rate as
occurred during standard operation (without oxygen enrichment).
3-70
-------
Although the oxygen was not added directly to the primary burner air,
the "equivalent" levels of oxygen enrichment may nevertheless be
defined. These levels were 26 percent and 27 percent during the two
tests, respectively.93
Achurra et al.89 have reported tests using oxygen enrichment in a
calcine-charged reverberatory furnace at the Caletones smelter. A
level of 26 percent oxygen in the primary combustion air yielded a
15-percent increase in capacity and reduced the fuel requirements per
unit of charge by 19 percent. Refractory consumption, however, was
noted to increase by 20 percent. In later tests, oxygen was injected
through the furnace roof using lances to make the temperature
distribution uniform within the furnace. An oxygen-enrichment level
of 30 percent in this test yielded a 20-percent increase in smelting
capacity, accompanied by a 19-percent reduction in the fuel requirement
per unit of charge. No mention was made of changes in refractory
wear.
Similar investigations have been made for green-charged furnaces.
Eastwood et al.94 have reported tests made at the Rokana smelter in
Zambia. In accordance with the practice of Inco, oxygen lances were
installed near the coal burners. Furnace throughput was increased by
18 percent. The level of oxygen enrichment was not specified. In a
later paper, Gibson95 reported that the furnace refractory wear had
not increased measurably with the use of oxygen enrichment. Furnace
operating campaigns were indicated to be longer than 30 months.
Extensive tests of various oxygen enhancement schemes and various
levels of enrichment have been made at the Almalyk smelter in the
Soviet Union.96 97 In the work of Kupryakov et al.,96 oxygen was
introduced (1) through water-cooled lances positioned beneath the
burners and (2) through the existing burners (primary air enrichment).
In the former scheme, an oxygen-enrichment level of 25 percent yielded
a production increase of 20 percent and a decrease in the specific
fuel requirement of 17 percent, while oxygen enrichment to the 30-percent
level provided a production increase of 45 percent, accompanied by a
30-percent reduction in the specific fuel requirement. When oxygen
3-71
-------
was mixed directly with the primary burner air, a slightly greater-
efficiency was noted. With a 25-percent level of enrichment, production
was noted to increase by 22 percent, with a 19-percent reduction in
the specific fuel requirement. Oxygen enrichment to the 30-percent
level yielded a 56-percent increase in capacity and a 36-percent
decrease in specific fuel requirements. With regard to roof refractory
wear, it was determined that the wear did not differ markedly from
conventional operation.
When oxygen-fuel burners are used in the roof of a reverberatory
furnace, the efficiency in terms of the heat utilization for smelting
is considerably higher than in the schemes discussed previously. Such
is the case because, primarily, the heat is directed upon the charge
and the heat is transferred to a large extent by convection. Also,
the flame temperature is much greater with pure oxygen than with
oxygen-enriched air, resulting in a higher thermal driving force.
Finally, Goto98 has pointed out that an additional heat transfer
mechanism results with oxygen-fuel burners. C02 and H20 are dissociated
when they are formed with oxygen-fuel combustion at temperatures in
the 2,000° to 2,900° C (3,600° to 5,300° F) range, thereby absorbing
heat. This dissociation heat is released when the gases are cooled
sufficiently to permit reformation of the molecules. By applying the
flame directly to or near the comparatively cool charge piles in the
reverberatory furnace, the reformation heat is applied directly to the
material to be smelted.
Extensive investigations on the use of oxy-fuel burners to increase
production in a green-charged reverberatory furnace have been made at
the Caletones smelter in Chile. As reported by Achurra et al.,89
throughput has been increased by up to 71 percent when using seven
oxy-fuel burners and only one of the three original conventional
burners. In this particular test, the fuel requirement per Mg of
charge showed a decrease of 40 percent, from its value with conven-
tional firing. In later tests, the furnace uptake was rebuilt to
correct for a design flaw, and production increases of up to 122 percent
were achieved with 12 oxy-fuel burners. This expansion yielded a
3-72
-------
56-percent reduction in the specific fuel requirements of the furnace
proper. With regard to roof refractory wear, the tests indicated that
the refractory consumption per unit of charge was the same or slightly
lower than that for a conventional reverberatory furnace.89 " More
detailed information on the experience at Caletones is summarized in
Section 4.4.6.1.1
Extensive experience with oxy-fuel burners on a calcine-charged
nickel reverberatory furnace has been accumulated by Inco at its
Copper Cliff smelter in Ontario, Canada. Blanco et al.100 have reported
increasing throughput by 45 percent through the use of 10 oxy-fuel
burners. The increase in throughput was accompanied by a 55-percent
decrease in fuel requirements per unit of charge. Calcine is charged
to the furnace along the sidewalls using a drag conveyor/fettling pipe
system. As of November 1980, some 13 months of operation were
achieved,100 and the use of oxy-fuel burners continues. Inco has
indicated that essentially no change in matte grade occurred with the
conversion from conventional firing to full oxy-fuel firing at Copper
Cliff.101 Furthermore, no changes were made in the degree of roast of
the feed.101 The implication is that essentially the same degree of
sulfur elimination (per unit of feed) occurred in the furnace during
conventional operation and full oxy-fuel firing. More detailed informa-
tion on Inco's experience with oxy-fuel firing is summarized in Section
4.4.6.1.3.
Small-scale tests of the oxygen-sprinkle smelting scheme have
been made by Phelps Dodge Corporation at its Morenci smelter.102 103
The tests indicated that furnace throughput could be increased by
100 percent, while the total energy requirement per unit of charge
could be reduced by two-thirds. No insurmountable problems with the
process were identified during the tests. Both the bath temperature
and slag copper content were found to be acceptable. The process was
noted to have a slightly higher dust generation rate, however, as
compared to conventional reverberatory furnace operation. At this
point, Phelps Dodge is undecided whether it will adopt oxygen-sprinkle
smelting technology.103 This decision will be made after additional
testing in commercial-scale facilities.
3-73
-------
3.4.3.5.4 Feasibility of oxy-fuel firing in furnaces charged with
Wagstaff guns. No documented experience has been identified with regard
to the combination of oxy-fuel firing in calcine-charged reverberatory
furnaces employing Wagstaff feeders. ASARCO, while acknowledging that
such a combination would not be infeasible, maintains that it would be
impractical.104 Upon charging such a furnace, the calcines would spread
more or less uniformly over a significant fraction of the bath. Those
calcines beneath the burners and in the immediate vicinity would most
likely smelt most rapidly, while those more distant from the burners
would be expected to smelt more slowly. According to ASARCO, unless
the furnace was charged again when charge beneath the burners had
smelted, the bath beneath individual burners could conceivably over-
heat,105 resulting in slag foaming, which hinders the separation of
matte and slag and increases copper losses. However, ASARCO maintains
that charging the furnace before all of the previous charge was smelted
would not be practical because the regions with partially smelted
charge would hinder the spreading of the new charge, causing charge
buildups—especially beneath the Wagstaff guns.105 If such a condition
occurred, the furnace firing rate would have to be reduced to avoid
overheating the furnace refractories while the piles of charge were
smelted. The implication is that smelting rates would decrease. It
has also been indicated that the retrofit of oxy-fuel burners to
existing furnaces would hinder the maintenance of roof refractories.
Most of the existing calcine-charged furnaces employ sprung-arch roofs
of silica brick, which are maintained via silica slurry patching.
ASARCO has indicated that difficulties are anticipated with regard to
accurately directing the slurry to roof areas in the vicinity of the
burners.104
ASARCO's comments on slag foaming beneath oxy-fuel burners are
based essentially on pilot tests with a 3- by 9~m (10- by 3Q-ft) dross
lead furnace fired with oxy-fuel burners.104 This furnace does not
use Wagstaft feeding. Observations of the test were used to estimate
what impacts would result if Wagstaff guns and oxy-fuel burners were
used on a calcine-charged copper reverberatory furnace.104 These
3-74
-------
observations104 were (1) an uneven heat distribution across the bath*
and (2) foaming slag beneath the burners due to overheating. Inco's
experience with the oxy-fuel-fired nickel reverberatory furnace also
indicates that a less uniform distribution of heat across the bath
surface results with oxy-fuel firing, as compared to conventional
operation36 (although no bath temperature data are available).
However, Inco has indicated that a less uniform heat distribution is
not expected to preclude the use of Wagstaff gun charging systems.36
With regard to slag foaming, Inco has reported that its oxy-fuel fired
furnace has been operated at various times without any charge banks
(although this operation was not deliberate). Such a condition is
similar to that which would occur with Wagstaff charging when the
charge beneath the burners is smelted. During these periods at Inco,
no slag foaming was observed.36 Slags from calcine-charged nickel
reverberatory furnaces at Inco have been reported as containing
(typically) 36.9 percent, Si02 and 36.9 percent Fe100 (which corresponds
to 47.5 percent FeO). Slags from various calcine-charged copper
reverberatory furnaces in use worldwide have been reported as containing
27.4 to 41.7 percent Si02 and 42.0 to 54.5 percent FeO.10b" Of special
note is the Mt. Isa Mines copper operation, which produces slags
almost identical in concentration of major constituents to Inco's
nickel furnace slags: 35.7 percent Si02 and 47.0 percent FeO.106
Because of the similarity in composition of nickel and copper smelting
slags, it appears that Inco's experience related to slag foaming can
be applied to copper reverberatory furnaces. Based on the data, it
appears unlikely that slag foaming would be a problem in oxy-fuel-fired
copper reverberatory furnaces charged with Wagstaff guns.
ASARCO's comments on the partially smelted charge's hindering the
spread of new charge are based primarily on the fact that calcine has
been noted to accumulate, under certain circumstances, beneath Wagstaff
guns and hinder the spreading of charge that continues to flow from
*No temperature measurements or profiles were developed during
the pilot test. Heat distribution conclusions were based on observing
the bath during and after charging.
3-75
-------
the gun.104 The circumstances alluded to are those in which calcine-,
which is sticky due to composition or excessive teinpet ature, tends to
"hang up" in the larry car and slowly discharge from the gun.10' Such
behavior has been noted previously by Weisenberg et cil.luf Undor
these conditions, the relatively small weight of calcines initially
entering the furnace starts to smelt directly beneath the gun discharge.
ASARCO reports107 that this mass of partially smelted material hinders
the spreading of additional calcines, which slowly discharge from the
larry car. It appears, however, that calcine spreading under these
conditions is hindered primarily by a low velocity of discharge from
the gun rather than by the presence of a partially smelted mass beneath
the gun. While it is acknowledged that hot calcine can stick together
and resist flowing, under normal conditions it is very free flowing.109
Hence, under normal conditions, calcine charged via Wagstaff guns is
imparted substantial horizontal velocity (due to the angle of the gun)
and is reported as flowing "almost like a liquid" over one-half to
two-thirds of the surface of the bath.110 In light of this fact, it
is difficult to envision how regions of partially smelted material on
the bath would substantially hinder the spreading of the next charge.
Insufficient information exists with regard to the potential
magnitude of difficulties with roof refractory maintenance in existing
silica-arch furnaces retrofitted with oxy-fuel burners. However, such
potential difficulties would not be expected in new furnaces employing
oxy-fuel firing. The Inco furnace employs a suspended roof of magnesite
refractory, which does not require hot patching. It is expected that
new furnaces would employ magnesite roofs because the trend has generally
been toward greater use of magnesite refractory.
It should be noted that sidewall overheating is not expected to
be a problem with the retrofit of oxy-fuel burners to Wagstaff-charged
furnaces. The furnace at Inco makes some use of the small calcine
charge banks to protect the sidewall refractory from excessive heat.
However, this particular furnace does not have sidewall cooling
panels.111 Most furnaces charged by Wagstaff guns employ sidewall
cooling by design, and Inco has -indicated that it is likely that such
3-76
-------
furnaces could rely on the existing sidewall cooling scheme to prevent
excessive refractory wear.111 It should also be noted that the sidewall
temperature is strongly affected by (1) the distance from each burner
to the wall and (2) the angular position of the burner relative to the
vertical. At the Inco furnace, some initial problems with sidewall
overheating were eliminated when the burners were repositioned. lfl° It
can be concluded that the combination of sidewall cooling and proper
burner positioning would eliminate any potential sidewall overheating
problems.
Overall, Inco has reported that Wagstaff gun charging systems are
believed to be technically feasible for use on reverberatory furnaces
with full oxy-fuel firing, although some development work would be
required.36 It is possible that the furnace would require more Wagstaff
guns and more frequent charging than is characteristic of conventionally
fired furnaces with Wagstaff feeding systems.36
3.4.3.5.5 Conclusions. The operating experience discussed
indicates that three methods of oxygen introduction to reverberatory
furnaces have been extensively tested: (1) the use of primary air
enrichment, (2) undershooting the flame with oxygen, and (3) the use
of roof-mounted oxy-fuel burners. The first two schemes are considered
demonstrated for green- and calcine-charged reverberatory furnaces,
irrespective of the means of charging.* Roof-mounted oxy-fuel firing
is considered demonstrated for green-charged reverberatory furnaces,
based on the extensive experience at the Caletones smelter. With
regard to the use of oxy-fuel firing on calcine-charged furnaces,
extensive experience has been accumulated by Inco on a side-charged
nickel reverberatory furnace. Nickel concentrates are similar to
copper concentrates in that both are sulfide materials containing a
substantial percentage iron that are produced via froth flotation
*The opinion of the domestic industry lends further support to
this conclusion. ASARCO has indicated that primary air enrichment and
oxygen undershooting are viable alternatives for increasing the capacity
of calcine-charged furnaces.112 Kennecott has reported that oxygen
enrichment is a technically feasible expansion option for green-charged
reverberatories at its McGill smelter.8
3-77
-------
after fine grinding of the ores. Hence, the calcines produced by
roasting have similar physical properties. On this basis, oxy-fuel
firing is considered viable for calcine-charged copper reverberatory
furnaces using sidewall feeding.
There appears to be no technical reason why full oxy-fuel firing
could not be used on reverberatory furnaces charged with Wagstaff
guns. However, some of the data indicate possible engineering and
production problems that may preclude such usage under some conditions.
Consequently, for the purposes of this analysis, the use of oxy-fuel
firing on Wagstaff-charged furnaces is not considered to be fully
demonstrated.
The literature indicates that some tests have been performed in
full-scale furnaces using oxygen lancing through the roof. This
technique of oxygen enhancement may be demonstrated.
Oxygen-sprinkle smelting technology is not considered fully
demonstrated at the present time in light of the hesitancy on the part
of Phelps Dodge to adopt it without further testing.
On the basis of Inco's experience, it is concluded that essentially
no changes in matte grade would be expected with the addition of
oxy-fuel burners to reverberatory furnaces. Similarly, it is expected
that no changes in matte grade would occur with oxygen enrichment and
oxygen undershooting because these schemes are less extreme in terms
of oxygen usage and flame temperature than oxy-fuel firing. A more
thorough discussion of the question of matte grade changes with oxygen
enhancement is addressed in Section 4.4.6.3.
Primary air enrichment and oxygen undershooting have generally
been employed to obtain production increases of less than 50 percent.
In this analysis, it is assumed that a 20-percent increase in production
is achievable when using either of these schemes in green- or calcine-
charged reverberatory furnaces. It is further assumed, based on the
extensive operating experience at the Rokana smelter and the fact that
maximum flame temperatures shift closer to the bath when undershooting
the flames with oxygen, that this scheme is preferable to oxygen
enrichment of the primary burner air for increasing furnace capacity.
3-78
-------
In light of Inco's experience with roof-mounted oxy-fuel burners,
a 40-percent increase in production is considered achievable with this
scheme in calcine-charged reverberatory furnaces.
Based on the experience at the Caletones smelter, a 50-percent
increase in furnace throughput will be considered achievable for
green-charged furnaces retrofitted with roof-mounted oxy-fuel burners.
3.4.3.6 Replacement of Reverberatory Smelting by Flash Smelting.
In lieu of expanding existing reverberatory furnaces to increase
capacity, smelters may elect to replace this technology with flash
furnaces and simultaneously increase throughput. This option would be
viable for smelters that could tolerate the increased matte grade
produced by the flash furnace. Depending on the matte grade produced
by the flash furnace, substantial increases in throughput can be
achieved without increasing converter capacity because converter cycle
time is reduced with a higher grade matte. For example, increasing
matte grade from 40 to 55 percent via installation of a flash furnace
allows furnace throughput (hence plant throughput) to be increased by
100 percent without adding additional converters. However, use of
flash smelting does require the installation of concentrate dryer
capacity.
The ASARCO-Hayden smelter is planning to convert from reverberatory
smelting to Inco flash furnaces.35 Kennecott Corporation is considering
a similar conversion for its Hurley smelter.113
3.4.4 Electric Furnaces
As with reverberatory furnaces, green-charged electric furnaces
may increase capacity by converting to calcine-charged operation. The
furnace capacity would increase because less heat is required to smelt
the roasted calcine at about 540° C (1,000° F) than is required to
smelt dried feed at a temperature of about 65° C (150° F). For this
analysis, an increase of 40 percent is assumed achievable. Inspiration
Consolidated Copper Company has indicated that the conversion to
calcine-charged operation would be its most likely electric furnace
expansion mode.54
3-79
-------
In contrast to reverberatory furnaces, the conversion to
calcine-charging in an electric furnace would not require the addition
of cooling panels at the furnace slag line because neither charge
banks nor fossil fuel burners are used in electric furnaces. Also, it
is riot likely that extensive feed system modifications would be required
in order to process calcine feed. Electric furnaces smelt dried
concentrates, and the handling of dried concentrates and calcines is
similar.
Both green- and calcine-charged electric furnaces can increase
capacity by eliminating converter slag return. An increase in produc-
tion similar to that achieved in reverberatory furnaces (25 percent)
would be expected.
It is conceivable to increase electric furnace capacity by installing
a larger transformer. However, up-powering the furnace would increase
slag temperatures, leading to increased refractory wear.114 Hence,
this option is not considered to be viable to the industry.
Physically expanding electric furnaces is also a conceivable
option. However, the transformer and electrode design parameters are
usually sized closely to the rated furnace capacity.114 As a result,
it appears that this option would be impractical, and it is considered
unfeasible in this study.
It should be noted that expansion modes employing oxygen enrich-
ment, such as those useful for reverberatory furnaces, would not be
feasible for electric furnaces. Oxygen enrichment offers no advantage
because no fuel is combusted.
In lieu of expanding existing electric furnaces to increase
capacity, smelters may elect to replace this technology with flash
furnaces and simultaneously increase throughput. This option would be
viable for smelters that could tolerate the increased matte grade
produced by the flash furnace. Depending on the flash furnace matte
grade, substantial increases in throughput can be achieved without
increasing converter capacity because converter cycle time is reduced
with a higher grade matte. For example, increasing matte grade from
3-80
-------
40 to 55 percent via installation of flash furnace allows plant
throughput to be increased by approximately 100 percent without adding
additional converters.
3.4.5 Outokumpu Flash Furnaces
Outokumpu flash furnaces can increase capacity readily by using
oxygen-enrichment of the process air. Physically expanding Outokumpu
furnaces is not considered technically feasible,115 primarily because
of the furnace geometry. No option exists for expanding furnace
capacity by eliminating converter slag return because this slag is
generally processed in other facilities by design.
Outokumpu Oy has reported increasing the capacity of its flash
furnaces at Harjavalta, Finland, by 60 to 70 percent when the oxygen
content of the process air was raised to 30 to 40 percent.116 The use
of oxygen yielded an S02 concentration in the offgases of 18 to 20 per-
cent and reduced the oil requirement of the furnace. The increased
capacity in the furnace resulted primarily from the decreased gas
volume afforded by oxygen enrichment.
Phelps Dodge-Hidalgo smelter personnel have indicated that their
Outokumpu furnace capacity could probably be increased by 25 percent
through the use of oxygen enrichment of the combustion air.117 Expan-
sions greater than 25 percent would lead to overheating of the reaction
shaft and increased logistics problems with respect to ancillary
material-handling systems.
3.4.6 Noranda Reactors
The primary expansion mode for Noranda reactors is through oxygen
enrichment of the blowing air. Kennecott Corporation has increased
reactor capacity by some extent at its Garfield facility by increasing
the oxygen-enrichment level from the design value of 30 percent to
34 percent.118 A further increase in throughput by about 6 percent
could possibly be achieved with an enrichment level of 36 percent
oxygen. A level of about 36 percent oxygen is considered to be the
upper limit because it represents the point of autogenous reactor
operation.118 Such operation is undesirable because reactor control
is difficult.
3-81
-------
It is conceivable to increase Noranda reactor capacity by increasing
the blowing rate via the installation of a larger blower. However,
this scheme is not considered to be a viable expansion option because
of offgas handling constraints.118
Physical expansion of Noranda reactors coupled with increasing
the number of tuyeres is also a conceivable expansion option. This
option is not considered viable in this analysis because of physical
space limitations, down time requirements, and offgas handling con-
straints.
In conclusion, no expansion options other than additional oxygen
enrichment are considered viable for Noranda reactors. The use of addi-
tional oxygen would produce only a slight increase in capacity, however.
3.4.7 Converters
ASARCO has reported increasing the capacity of a Peirce-Smith
converter at its Tacoma smelter through physical expansion.119 The
converter was lengthened by 5 feet, and 6 additional tuyeres were
added, increasing the number of tuyeres from 46 to 52. As a result,
the capacity of the converter increased by approximately 13 percent.
Converter capacity may be increased by increasing the air-blowing
rate alone, which would decrease the time required for matte conversion.
Such an expansion could require the installation of a. larger blower.
However, the upper limit on blowing rate is determined by the excessive
ejection of molten material from the converter. In this analysis, it
is assumed that domestic converters operate at or near their maximum
blowing rate. Hence, this expansion mode is not assumed to be a
viable one for the industry.
Oxygen enrichment of converter blowing air appears to be an
option for increasing converter capacity because it results in an
increase in the rate of conversion of matte to blister copper during
blowing. Oxygen usage in converters has been widespread in the past,
with 7 of the 15 domestic smelters having reported its use during the
slag and/or copper blows.1'0 The percentage of oxygen in the blowing
air has not exceeded 29 percent (with 24 to 26 percent being most
common) because higher levels can lead to refractory damage from
increased operating temperatures.
3-82
-------
It should be noted, however, that oxygen enrichment of converter
blowing air has been used, almost exclusively, to allow increased pro-
cessing of scrap materials. Indeed, scrap materials are required when
oxygen enrichment is employed, to reduce converter temperatures. Charg-
ing scrap materials to the converters serves to interrupt the blowing
cycles. Hence, although the actual blowing time is reduced with oxygen
enrichment, it is not clear that a decrease in total cycle time ensues.
Also, the use of oxygen enrichment would be contingent upon the availa-
bility of sufficient scrap materials. For these reasons, and because it
has not been employed for increasing the rate of matte throughput in the
industry, oxygen enrichment of converter blowing air is not considered a
viable option for increasing converter capacity in this analysis.
It is concluded that physical expansion coupled with increasing
the number of tuyeres is the most likely mode of increasing the matte
throughput rate of a converter. Only a limited expansion can be
achieved by this method, however.
3.5 SUITABILITY OF ALTERNATIVE TECHNOLOGIES FOR PROCESSING HIGH-
IMPURITY FEEDS
3.5.1 Background
Domestic smelters that process high-impurity feeds--!.e., those
defined in the present NSPS as containing more than 0.2 weight percent
arsenic, 0.1 weight percent antimony, 4.5 weight percent lead, or 5.5
weight percent zinc—generally employ the multihearth roaster-
reverberatory furnace-converter configuration. An exception is the
Phelps Dodge-Ajo smelter, which processes feeds containing 0.3 percent
arsenic with a green-charged reverberatory furnace.121 The industry
has stated the need to maintain this configuration because of the
flexibility afforded in terms of impurity elimination capability,
which allows product quality to be maintained, and the capability to
process secondary materials. The primary factor affording both these
advantages is the low (40 to 45 percent) matte grade produced by this
smelting configuration. Such a matte grade leads to long blowing
times in the converters, which are highly effective for eliminating
impurities. Furthermore, the heat released while blowing such a matte
allows substantial quantities of secondary materials, which can include
3-83
-------
some lead-smelter byproducts, to be processed in the converters. The
industry further endorses the multihearth roaster-reverberatory furnace-
converter configuration because its multistep treatment allows some
impurity segregation during impurity recovery.*
The rationale (under the original NSPS) for exempting reverberatory
furnaces from control when processing a charge containing a high level
of volatile impurities is multifaceted. At promulgation, the cost of
control of the weak S02 stream was considered unreasonable. Operation
of electric furnaces, which were considered technically capable of
processing such a charge, was not considered affordable in the Southwest.
The other alternative to the conventional reverberatory furnace evaluated
at that time was the Outokumpu flash furnace, which was in use worldwide.
This technology was dismissed on technical grounds. It had not been
used to smelt a charge containing more than a fixed level of certain
impurities and was not considered to have been demonstrated for smelting
feeds having compositions similar to those encountered at installations
such as ASARCO-Tacoma.t The maximum levels of impurities processed by
the Outokumpu furnace were used, however, as the basis for defining a
"high level of volatile impurities." No other technologies were
assessed as possible replacements for the reverberatory furnace since
the Outokumpu furnace was considered to be the most likely option in
lieu of an electric furnace.
It is the purpose of this section to reassess, in light of the
requirements of the industry, the suitability of alternative technologies
for processing HI feed materials.
*At the ASARCO-Tacoma smelter, for example, which processes feeds
having high levels of arsenic, impurities are separated into an arsenic-
rich stream (the roaster and reverb offgases) and a lead-antimony-rich
stream (the converter offgases). The arsenic-rich material is roasted
in a separate process to produce arsenic trioxide and to recover the
copper in the dust, while the lead-antimony-rich material is processed
in a lead smelter for the recovery of lead, zinc, bismuth, and antimony.
tThis installation, a custom smelter, has by far the greatest
impurity burden of all of the domestic copper smelters.
3-84
-------
3.5.2 Impurity Behavior During the Smelting Process
The ultimate objective in copper smelting and refining operations
is the production of a product containing less than specified maximum
levels of impurities. Impurity removal occurs at each stage of the
operation: roasting, smelting, converting, fire-refining, and electro-
lytic refining. If the impurity burden to the plant is high, care
must be taken during the first four operations to remove the bulk of
the impurities. Current practice is to maintain the impurity levels
of anode (fire-refined) copper below certain limits before electrolytic
refining. In Table 3-7, the maximum acceptable average impurity
levels in anode copper from the ASARCO-Tacoma smelter are presented.122
The typical average impurity levels in blister copper corresponding to
this particular anode composition are presented for comparison.122
High impurity feed materials include concentrates from various
sources, as well as various smelter byproducts—notably those from
lead smelters. Assays of various high impurity materials that have
been processed at the ASARCO-Tacoma smelter are presented in Table 3-8.
The Lepanto concentrate, produced in the Phillipines, shows high
levels of arsenic and antimony, at 11 percent and 0.75 percent, respec-
tively. North Peru concentrate shows high levels of arsenic (11.2 percent),
antimony (2.1 percent), and zinc (9.3 percent). With regard to the
lead smelter byproducts, both the lead matte and lead speiss contain
high levels of arsenic, antimony, and lead. In general, substantially
lower quantities of lead smelter byproducts are processed, as compared
to the high impurity concentrates. Also, it should be noted that
current practice involves blending high-impurity materials with other
materials containing low levels of impurities to produce a roaster
charge containing manageable impurity levels. ASARCO has indicated
(1976) that the anticipated maximum impurity levels in future (blended)
feeds to the Tacoma smelter are 6.3 percent arsenic, 0.78 percent
antimony, 3.3 percent lead, and 1.5 percent zinc.125
The behavior of any specific impurity element during the copper
smelting process is dependent upon thermodynamic, kinetic, and process
3-85
-------
TABLE 3-7. MAXIMUM ACCEPTABLE IMPURITY LEVELS IN ANODE
COPPER, AND CORRESPONDING LEVELS IN BLISTER COPPER
PRODUCED AT THE ASARCO-TACOMA SMELTER122
Maximum
acceptable Corresponding
average average
concentration level in
in anode copper, blister copper,
Element weight percent weight percent
Arsenic
Antimony
Lead
Zinc
Bismuth
Nickel
Selenium
Tellurium
0.20
0.15
0.07
N/A
0.02
0.19
0.05
0.06
0.35
0.20
0.15
N/A
0.03
0.20
0.06
0.06
3-86
-------
TABLE 3-8. ASSAYS OF VARIOUS HIGH IMPURITY MATERIALS PROCESSED
AT ASARCO-TACOMA
Material
Concentrates
Lepanto
North Peru
Lead Smelter By-Productsa
ASARCO Lead Matte
ASARCO Lead Speiss
% Cu % Fe % S % As % Sb % Pb % Zn Reference
31.0 11.9 35.2 10.94 0.75 0.21 1.0 123
30.2 7.7 29.8 11.19 2.07 4.00 9.3 123
37-49 5-11 6-15 0.5-7 0.2-1.2 7-10 2.67 123,124
52-58 1-2 0.1-1.4 16-20 4-7 10-13 0.10 123,124
aTypical analyses from the ASARCO E. Helena and El Paso lead smelters, 1979-81.
co
-------
parameters. The extent of removal of a given impurity obviously
depends, to some degree, upon its concentration. Impurities are
removed by volatilization and slagging.
Volatilization can occur for impurity elements in sulfide form,
in oxide form, or in the free state, depending upon the element. Mosf
impurities are present as sulfides in the charge to the smelter.
However, oxides of most impurities can form under oxidizing conditions.*
The sulfides of the impurities Sb, Pb, and Zn are more volatile than
the corresponding oxides. In tne case of As, both the sulfide and
oxide forms are extremely volatile. Overall, arsenic compounds are
perhaps the most volatile of the major impurities.
Impurity elimination through slagging occurs by the combination
of the metal oxides with silica, similar to the slagging of iron
oxide.
3.5.2.1 Impurity Elimination During Roasting. Conventional
roasting removes impurities as volatilized materials and as chemically
altered dusts. Hence, impurity elimination is affected by the tempera-
ture, residence time, and the roaster atmosphere. These same parameters
govern the elimination of sulfur in the roaster. As discussed previ-
ously, the sulfur elimination during roasting is a major determinant
of the matte grade produced during smelting, for a given charge
composition. Overall impurity removal is maximized in the conventional
multihearth roasting, reverberatory smelting, converting operation
concurrent with the production of a relatively low-grade matte (40 to
45 percent copper) in the smelting furnace. Because of this constraint
on matte grade, there is little latitude for direct control of impurity
elimination in the roasting operation itself.li;tl
Impurity elimination in a multihearth roaster is generally greater
than in d fluidized-bed unit for any level of sulfur elimination. In
a multihearth roaster, it is possible to vary oxidizing and reducing
conditions, temperature, and gas composition on each hearth. The
residence times of concentrate particles in a multihearth roaster
*The impurity elements As, Sb, Pb, Zn, and Bi all have a greater
affinity for oxygen than does copper.
3-88
-------
generally range from 1 to 2 hours.127 A single fluidized bed roaster
can provide only one set of conditions at a time, either oxidizing or
reducing. The temperature remains essentially constant throughout the
bed, although hot spots may be present. The average residence time of
concentrate particles in a fluidized bed roaster is much shorter
compared to a multihearth roaster and can differ by as much as a
factor of ten.1-8 It should be noted that this difference is suffi-
ciently great to permit the operation of two or more fluid-bed roasters
in series. This scheme would overcome the limitation of providing
only one set of conditions during fluid-bed roasting. While this
approach has been used in processing other materials, there is no
experience with two-stage fluid-bed roaster systems in the copper
industry.
The extent of elimination of impurities during roasting in multi-
hearth roasters for one particular feed impurity level is indicated in
Table 3-9, which provides information on impurity distributions during
roasting and reverberatory furnace smelting (based on the ASARCO-Tacoma
operation123 processing HI feed materials). It is noted that 25 percent
of the feed arsenic is eliminated during roasting, while only between
4 and 8 percent of other important impurity elements are eliminated.
The sulfidizing roast process,129 developed by Outokumpu Oy in
Finland, was developed specifically to eliminate volatile impurities
from ore concentrate feeds before smelting. This process, although
still in the pilot stage of development, yields substantially greater
impurity removal than does multihearth roasting.
In sulfidizing roasting, ore concentrates are dried and preheated
and fed to a rotary sulfidization kiln. The sulfidizing atmosphere is
provided by elemental sulfur, which is vaporized and transported by a
preheated nitrogen carrier gas into the kiln. The hot gas flows
countercurrently to the feed. Under the process conditions the complex
compounds decompose into base metal sulfides (iron and copper sulfides)
and volatile impurity sulfides containing arsenic, antimony, and
bismuth. The volatile impurity sulfides are carried out of the kiln
in the offgases and are condensed for recovery.
3-89
-------
TABLE 3-9. DISTRIBUTION OF IMPURITY ELEMENTS IN
CONVENTIONAL SMELTING WHEN PROCESSING HIGH-IMPURITY FEEDSa
Percentage reportinq in various streams123
Impurity
element
Arsenic
Antimony
Lead
Zinc
Bismuth
Tin
Nickel
Selenium
Tellurium
Multihearth
roaster
dust
25
6
5
5
4
5
0
6
8
Reverb
dust
52
25
20
21
20
18
2
18
22
Reverb
slag
10
50
19
65
1
47
6
2
2
Reverb
matte
13
19
56
9
75
30
92
74
68
a
Based on the ASARCO-Tacoma smelter. The levels of these impurities in
the feed are not available.
3-90
-------
Both laboratory and pilot-plant tests of the sulfidizing roast
process have been made. The laboratory tests established the following
impurity-elimination capability:12°
Arsenic removal (up to 10 percent As in feed), 99 percent
Antimony removal (tip to 1.5 percent Sb in fend), 50 to 80
percent
Bismuth removal (up to 0.2 percent Bi in feed), 20 to 30
percent
On the basis of the laboratory tests, Outokumpu established that
arsenic can be removed essentially completely, regardless of its
original concentration in the feed.120 Optimum arsenic elimination
occurs with operating temperatures in the range of 600° to 800° C
(1,1.10° to 1,470° F).
Pilot-scale tests of the process were made in a plant of 10 to
100 Kg/h (20 to 220 Ib/h) feed capacity. The tests with copper concen-
trates were made at feed rates of 10 and 15 kg/h (20 and 30 Ib/h).
The pilot plant was comprised of a concentrate preheats, sulfur
vaporizer with nitrogen carrier gas system, sulfur vapor preheater,
rotary sulfidization kiln, concentrate cooler, and volatile impurity
condenser. In the tests, the sulfidization kiln was heated indirectly
by an electric resistance heating system. Problems with this heating
scheme caused the operating temperature of the sulfidization kiln to
be too low. Alsu, the temperature at the discharge end of the kiln was
often too low, which caused arsenic sulfide vapor to condense back
onto the sulfide particles. The result of these problems was that
arsenic removal from copper concentrates in two pilot tests (at
92 percent and 98 percent) was lower than that achieved in the labora-
tory tests. The level of sulfur removal during these two tests was
32 percent and 37 percent, respectively.*
*The pilot test data also indicate that the degree of sulfur
elimination is a function of the degree of impurity elimination achieved.
3-91
-------
The next step in the development, of the process will be tests
with a larger pilot plant, which will process up to 1 Mg/h (1.1 tons/h).
The sulfidization unit in this plant will be a brick-lined, direct-heated
rotary kiln designed for the combustion of sulfur vapor and coal in
the reaction space.
Projections of process parameters achievable in full-scale operation
(10 Mg/h [11 tons/hj feed) of the sulfidizing roast process have been
made by Outokumpu. These projections show in excess of 99 percent
removal of arsenic from copper concentrate containing 11.4 percent As.
Sulfur removal is projected at 43 percent. Projections for the removal
of antimony and bismuth in full-scale operation are no1 given; however,
based on the elimination predicted for arsenic, it is reasonable to
assume that antimony and bismuth would be eliminated at approximately
the same level as occurred in the laboratory tests.
3.5.2.2 Impurity Elimination During Smelting. Impurity elimina-
tion in smelting furnaces occurs by volatilization and slagging. the
amount of a given impurity that is slagged or volatilized generally
varies from one furnace to another.
Data illustrating the distribution of various important impurities
during reverberatory furnace smelting are shown in Table 3-9. As
noted for roasting, arsenic is volatilized to the greatest extent. In
contrast, antimony and zinc report extensively to the slag. The
majority of the lead and bismuth present in the charge report to the
matte and must be eliminated during converting.
The behavior of the impurities arsenic, antimony, lead, and zinc
during electric furnace smelting has been investigated by the Bureau
of Mines.125 A total of 14 smelting tests were made in cooperation
with ASARCO, which was interested in evaluating electric furnace
smelting as a possible alternative for its facoma smelter. A small
(800 KVA) electric furnace was jsed in the investigation. The furnace
charges consisted primarily of blends of chalcopyrite concentrate and
smelter hy-products. These materials were blended by ASARCO to \iiiudate
the present and anticipated future material flow through the Facuma
smelter The feed materials contained the following range of elemeital
concentrations: 19 7 to ?2.b percent Cu, ib. / to 26 7 percent, 1- e,
3-92
-------
17.6 to 24.4 percent S, 0.94 to 6.3 percent As, 0.24 to 0.78 percent
Sb, 1.5 to 3.3 percent Pb, and 0.80 to 1.5 percent Zn. Smelting
parameters were changed during the tests to determine the effect of
various electric-furnace smelting conditions on the compositions of
the smelting products.
Of special note are tests in which changes were made in the air
sweep over the furnace bath. Investigations were made in which (1) the
flow rate of ambient air over the furnace bath was varied and (2) ambient
air was replaced by a stream of argon blown over the bath. Arsenic
analyses of the smelting products did not indicate any effect from
these changes. The Bureau of Mines concluded that changes in air
sweep through the furnace do not affect the arsenic distribution.125
Impurity analyses were made of the matte produced in essentially
all of the tests. It was further concluded that, with the exception
of arsenic, which reports more readily to the matte in an electric
furnace, the distribution of impurity and by-product elements (including
lead, zinc, and antimony) is essentially the same as in a reverberatory
furnace.125 No information was provided as to the extent the matte
arsenic levels were increased over those in reverberatory smelting.
ASARCO has indicated, however, that the conclusions reached in the
study with respect to matte arsenic levels pertain to a comparison of
green-charged electric furnaces and the combination of roasters and
reverberatory furnaces130 (such as is used at the Tacoma plant). Hence,
the Bureau of Mines study does not contradict previous conclusions to
the effect that electric furnaces are technically demonstrated for
processing high-impurity feeds.
Inco has investigated the extent of volatilization of arsenic,
lead, and zinc in its Copper Cliff flash furnace.131 When producing a
40- to 50-percent copper matte, the proportions of these elements
fumed were, respectively, 50 to 60 percent, 20 to 25 percent, and 5 to
10 percant. The investigation used Copper Cliff concentrate, which
typically contains 0.002 percent As, 0.05 percent Pb, and 0.17 percent
Zn i3i 132 Arsenic was indicated to fume within the flash furnace as
As406, a form consistent with the degree of exposure of the feed to
the oxidizing atmosphere in this process. Through tests made in a
3-93
-------
bench-scale flash furnace (Section 3.5.4), Inco established that the
proportion of the impurity fumed increased with matte grade.131
Furthermore, over the range of impurity concentrations examined in the
bench-scale furnace (Section 3.5.4), it was determined that the
proportion of the impurity fumed was independent of the level of the
impurity in the feed.131
Although complete data are not available, impurity elimination in
flash furnaces appears to be greater than in reverberatory furnaces.
The intimate contact between the feed and the hot, oxidizing atmosphere
may increase volatilization and should lead to increased impurity
removal via slagging. Inco has reported that some flash furnace
feasibility studies it has performed for various companies indicated
that blister copper impurity levels are lower with flash smelting.36
Mackey et al.133 have investigated impurity behavior in the
Noranda process when operated for the production of both high-grade
matte and blister copper. The levels of the impurities in the feed
were not reported, however. The investigations were made using a
730-Mg/day (800-tons/day) prototype reactor. The distribution of
impurities when making a 70-percent copper matte is shown in Table 3-10.
Extensive volatilization is noted to occur for arsenic, antimony,
lead, and bismuth—primarily because the reactor operates much like a
converter. Zinc, which oxidizes readily, reports primarily to the
slag.
When blister copper is produced in the reactor directly, the
distribution of impurity elements changes substantially, as shown in
Table 3-11, with substantial percentages of As, Sb, and Bi reporting
to the blister copper. Such behavior results since these elements are
quite stable in copper and can dissolve therein before volatilization
can occur. Mackey et al. concluded in their investigation that blister
copper produced directly in the reactor contained higher levels of
thess impurities than those produced by conventional reverberatory
smelting. Similar conclusions were reached by Kennecott when considering
the Noranda process for its Garfield, Utah, smelter. The concentrates
processed at this smelter have trace quantities of arsenic, antimony,
and bismuth.66 Increased levels of these impurities in the blister
3-94
-------
TABLE 3-10. DISTRIBUTION OF IMPURITY ELEMENTS IN THE NORANDA
PROCESS (MATTE PRODUCTION MODE)133
Impurity
element
As
Sb
Pb
Zn
Bi
Percentage reporting in
Offgas (dust) Slag
85 7
57 28
74 13
27 68
70 21
reactor streams
70% Cu matte
8
15
13
6
9
aThe levels of these elements in the feed were not reported.
3-95
-------
TABLE 3-11. DISTRIBUTION OF IMPURITY ELEMENTS IN THE NORANDA
PROCESS (BLISTER COPPER PRODUCTION MODE)133
Impurity
element
As
Sb
Pb
Zn
Bi
Percentage
Offgas (dust)
19
29
24
21
52
reporting in
Slag
27
36
74
78.9
30
reactor streams
Reactor copper
54
35
2
0.1
18
JThe levels of these elements in the feed were not reported.
3-96
-------
copper were considered to result in unacceptable levels in the anode
and cathode copper.66 For this reason, Kennecott opted to use the
Noranda process for the production of high-grade matte, with subsequent
oxidation to blister copper in Peirce-Smith converters.
With respect to the Mitsubishi process, no data are available
with which to characterize the distribution of impurity elements.
However, the possibility exists that the retention of As, Sb, and Bi
might be high because of the continuous contact between the matte and
blister copper in the converting furnace.134
3.5.2.3 Impurity Elimination During Converting. The converting
operation provides the greatest latitude for controlling the impurity
content of blister copper. The converter is considered to be a fairly
ideal vessel for impurity control because of the large effective area
between the gas phase and the liquid matte. Impurities are eliminated
during converting by both volatilization and slagging.
The removal of impurities is most effective during slag blows.
This is because, as discussed previously, impurities are quite stable
in blister copper (produced during the copper blow). Converting of a
low-grade matte (40 to 45 percent Cu) is desirable since the duration
of the slagging cycle is increased, which allows more time for impurity
removal. The effect of matte grade on the removal of the impurities
arsenic, antimony, and bismuth during converting has been investigated
by George et al.,135 and is illustrated graphically in Figures 3-14
through 3-16, respectively. The levels of these impurities in the
matte were relatively low, at 0.01 to 0.1 percent.136 For antimony
and bismuth, the percentage elimination decreases with increasing
matte grade. In the case of arsenic, the elimination is quite constant
until a matte grade of approximately 65 percent is attained, at which
point a rapid decrease ensues. It should be noted that a low degree
of impurity removal during converting may not be consequential if, as
in the case of Noranda reactors operated for the production of high-
grade matte, substantial impurity elimination occurs during the smelting
step.
3-97
-------
100
90
BO
X70
3 BO
tu
y
5 40
M
oc
30
20
10
0
o
20 25 30 35
40 45 50 55 60
MATTE: GRADE, %cu
66 70 75 80
Figure 3-14. Converter elimination of arsenic
as a function of matte grade.135
100
00
80
*
S70
<
560
60
"
30
20
10
20 25 30 36 40 45 50 55 60
MATTE GRADE, XCu
66 70 75 80
Figure 3-15. Converter elimination of antimony
as a function of matte grade.135
3-98
-------
100
90
80
ft 70
= 60
60
z
i
i
m
30
20
10
0
J L
' ' L
20 25 30 35
40 45 50 56 60
MATTE QRAOE, % Cu
66 70 75 80
Figure 3-16. Converter elimination of bismuth
as a function of matte grade.135
3-99
-------
With regard to the mechanism of impurity elimination during
converting, volatilization is more important for arsenic and bismuth,
while oxidation-slagging is more important for antimony and zinc.
Both mechanisms are essentially equally important for lead.137
3.5.3 High-Impurity Feed Processing Experience with Outokumpu Flash
Furnaces
The smelting of concentrates containing substantial quantities of
lead and zinc impurities in a flash furnace has been reported at the
Kosaka smelter in Japan.138 Assays of two such concentrates are
presented in Table 3-12. All of the impurities reported at Kosaka
have concentrations below those levels specified in the current
exemption. However, their experience processing feeds with these
impurity levels is pertinent to this analysis.
Because most impurities were concentrated in the flue dusts at
Kosaka, the dust production rates of both the furnace and the converters
were fairly high as compared to other smelters. Initially, all of the
dusts from the flash furnace were recycled. The high lead content,
however, resulted in excessive accretions in the flash furnace waste
heat boiler, causing low heat recovery. These problems led to the
elimination of dust recycle to the furnace. As an alternative, the
flash furnace dust was ultimately processed in a hydrometallurgical
treatment plant which afforded the separation and recovery of copper,
zinc, lead, and cadmium.
Outokumpu Oy has investigated the processing of high-impurity
feeds in its technology. Concentrates containing 10 percent zinc
and 5 percent lead have been smelted in a pilot plant operation.139
As a result of its investigations, Outokumpu Oy recommends the maximum
feed impurity limits in Table 3-13, which vary depending upon the mode
of operation of the smelter.139 Condition A pertains to operation
with the recycle of all dusts produced except those recovered from the
sulfuric acid plant. Condition B pertains to operation of the smelting
complex in the same manner as discussed for A, but with increased
purification of electrolytes in the electrolytic refining operation to
compensate for slightly increased impurity levels in the anode copper.
Condition C refers to operation with more outlets for impurities, such
3-100
-------
TABLE 3-12. IMPURITY ASSAYS OF FEED MATERIALS PROCESSED
IN THE OUTOKUMPU FLASH FURNACE AT THE KOSAKA SMELTER138
Impurity concentration, percent
concentrate As Sb Pb Zn Bi Cd^
Concentrate A oTlOI 2J3 sToo7(J4oToT
Concentrates 0.14 - 3.1 2.5 0-03 0-01
3-101
-------
TABLE 3-13. MAXIMUM IMPURITY LEVELS RECOMMENDED FOR THE
OUTOKUMPU FLASH FURNACE139
Maximum concentration of
impurity in concentrate feed, percent
Mode of operation As Sb Pb Zn Bi
A
B
C
0.25
1
5
0.025
0.1
0.5
1.5
3
5
5
10
10
0.03
0.2
1.0
3-102
-------
as separate treatment of converter dusts; electric slag-cleaning-furnace
flue dusts; anode furnace slag; all or part of the flash furnace flue
dust; and, possibly, increased purification of the electrolyte during
electrolytic refining. Where the arsenic content of the feed exceeds
5 percent, Outokumpu recommends pretreatment of the concentrate, such
as by sulfidizing volatilization, before smelting.139
3.5.4 High-Impurity Feed Processing Experience with Inco Flash Furnaces
Concentrates processed in the commercial flash furnace at Inco in
Canada are very "clean" with respect to impurities, containing about
0.002 percent As, 0.05 percent Pb, and 0.17 percent Zn.131 132 Other
impurities are present only in trace amounts.131 As indicated previously,
Inco has studied the elimination of the impurities As, Sb, Pb, and Zn
at higher levels in a bench-scale flash furnace.131 The objective of
these tests was to evaluate the applicability of Inco oxygen flash
smelting to concentrates having higher impurity levels than those
encountered at Copper Cliff.
The miniplant flash furnace employed in the tests has a flashing
space enclosed within a silicon carbide tube. The tube is covered by
a refractory lid. A vertical burner is used to inject the concentrate-
oxygen mixture into the furnace flashing space. Matte and slag collect
in a receiving crucible sitting in the lower part of the chamber.
Offgases escape the chamber via an exhaust port in the refractory lid.
The unit is capable of processing up to 15 kg (33 Ib) of concentrate
per hour. At this rate, the process cannot be operated autogenously
as does the commercial Inco flash furnace because the heat losses
greatly exceed the heat generated by combusting the feed. The additional
heat required for autogenous operation is supplied by burning natural
gas in an annular space between the silicon carbide tube and an outer
refractory shell. Hence, the furnace temperature may be controlled
independently of furnace operation.
The duration of a test is limited by the capacity of the matte-slag
receiving crucible. At the normal throughputs employed, a test lasts
for about 1.5 hours. Temperatures at various locations within the
3-103
-------
unit are monitored by means of thermocouples. The furnace is operated
at a slightly negative pressure during tests to prevent fugitive
losses of gases, fumes, and dust.
Most of the tests with increased impurity levels were made with
Copper Cliff concentrate, which was doped with PbS and PbO, ZnS and
ZnO, arseno-pyrite (As), and speiss (As and Sb). The range of impurity
concentrations examined is shown in Table 3-14. The tests were used
primarily to determine the distribution of the various impurities
between the matte and slag, as a function of matte grade.
Initial tests made in the mi nip!ant flash furnace showed that the
metallurgy of the commercial flash furnace could be totally reproduced
in the mi nip!ant furnace,* in terms of the state of oxidation of the
+3 +2
system (slag Fe /Fe ratio) and copper losses in the slag. Hence,
conclusions (mentioned previously) reached by Inco concerning the
proportion of impurities fumed from the miniplant furnace are expected
to apply equally well to commercial-scale furnaces.
3.5.5 High-Impurity Feed Processing Experience with the Mitsubishi
Process
The Mitsubishi process is currently in operation at two smelters
worldwide—the Naoshima smelter in Japan and the Texasgulf Canada
smelter in Timmins, Ontario. Mitsubishi Metals Company has provided
information on the maximum impurity levels that it has processed.140
These levels are presented in Table 3-15. Mitsubishi indicated that
there was no limitation on impurity-levels in their process and that
no impurity related problems would occur at high levels, provided that
considerations were made for gas handling and dust treatment.
3.5.6 High-Impurity Feed Processing Experience with Noranda Reactors
The Noranda process, operated for the production of high-grade
matte, is currently employed at the Kennecott-Garfield smelter and at
the Home smelter of Noranda Mines, Ltd., in Quebec. The Home smelter
*The only exception noted was a lower oxygen efficiency achieved in
the miniplant furnace as compared to the commercial furnace. This differ-
ence, however, results from the difficulty of obtaining an absolutely
uniform delivery of concentrates through the small flash burner.
3-104
-------
TABLE 3-14. RANGE OF IMPURITY CONCENTRATIONS TESTED
IN THE INCO MINIPLANT FLASH FURNACE131
Range of
Impurity concentration, percent
As 0.25 to 1.0
Sb 0.1 to 0.3
Pb 0.1 to 2.0
1.0 to 5.0
3-105
-------
TABLE 3-15. MAXIMUM IMPURITY LEVELS PROCESSED
IN THE MITSUBISHI PROCESS140
Impurity Maximum concentration, percent
As 0.329
Sb 0.097
Pb 0.96
Zn 6.10
Bi 0.034
3-106
-------
is a custom smelter and typically processes copper concentrates* from
about 25 different mines.141 These materials have a range of analyses
as follows:141 up to 1 percent As, up to 0.05 percent Sb, up to
8 percent Pb, up to 10 percent Zn, up to 0.05 percent Bi, 20 to
40 percent Cu, 15 to 35 percent Fe, 20 to 40 percent S, and up to 10
percent Si02. Other types of concentrates, precipitate copper, copper
and zinc refinery residues, smelter reverts, dust, and chopped scrap
are also treated in the reactor. Special blending of concentrates and
drying are not required. Noranda has indicated that the maximum levels
of impurity elements smelted are not limited by the process per se, but
rather by recycle practice, workplace considerations (primarily
pertains to arsenic exposures to personnel), and electrolytic refining
practice. Based on these considerations, Noranda has established two
sets of impurity limits, corresponding to two modes of operation of
the smelter (see Table 3-16). Condition I pertains to the removal of
some of the process dusts for separate treatment and to the use of
local ventilation systems on matte tapping and slag skimming locations.
Condition II pertains to the removal of anode furnace slag and addi-
tional dust from the process for separate treatment and also to
additional purification of the electrolyte during electrolytic refining.
For higher levels of As, Sb, and Bi, Noranda suggests concentrate
pretreatment before smelting.
3.5.7 Conclusions
Previous conclusions regarding the applicability of electric
furnaces for processing high-impurity feeds remain unchanged; i.e.,
electric furnaces are considered technically demonstrated for processing
high-impurity materials.
The Noranda process, operated for the production of blister
copper directly, has been found to lead to increased anode and cathode
*No documented experience has been identified with regard to
smelting calcines in Noranda reactors, although Noranda has indicated
that it believes it to be feasible.142
3-107
-------
TABLE 3-16. MAXIMUM IMPURITY LEVELS RFCOMMENDED
FOR THE NORANDA PROCESS (MATTE PRODUCTION MODE)141
Mode Of Maximum concentration of impurity in concentrate feed, percent
Operation As Sb Pb Zn Bi
I 0.25 0.04 1 to 2 5 to 7 0.04
II 1 0.1 to 0.2 10 10 0.1 to 0.2
3-108
-------
copper impurity levels when processing only trace quantities of arsenic,
antimony, and bismuth impurities. Hence, it is concluded that this
technology would not be a feasible alternative to reverberatory or
electric furnaces for processing feed materials containing elevated
impurity levels.
The remaining technologies—Outokumpu flash furnaces, Inco flash
furnaces, the Mitsubishi process, and Noranda reactors (matte-production
mode)--all show some experience processing high-impurity feed materials.
A summary of the experience reported by each respective company is
shown in Table 3-17.
Experience accumulated with the Mitsubishi process indicates that
the levels of two impurities, arsenic and zinc, are higher than the
limits specified in the current definition of high-impurity feeds.
However, the levels processed are not considered to differ substantially
from these limits. Also, as reported by Biswas and Davenport,134 the
process may not be suitable for high-impurity materials in general
because of the potential for increased impurity levels in the blister
copper.
The Inco flash smelting process shows experience processing high
levels of arsenic and antimony. However, because only bench-scale
tests have been made, this technology is not considered demonstrated
for processing high impurity materials.
The Outokumpu flash smelting process has been used to smelt high
levels of lead and zinc impurities (although the lead concentration,
at 5 percent, is not substantially different from the current high-
impurity threshold for this element). Because only pilot-scale tests
were made, however, the Outokumpu process is also not considered
demonstrated for processing high impurity materials.
The Noranda process operated for the production of high-grade
matte shows full-scale experience processing high levels of arsenic,
lead, and zinc. However, to process feeds containing impurity concen-
trations exceeding significantly the limits specified in the definition
of high-impurity feeds implies operation according to Operating Mode
II (see Table 3-16), which requires increased electrolyte purification
3-109
-------
TABLE 3-17. SUMMARY OF EXPERIENCE PROCESSING HIGHrlMPURITY FEEDS'
IN ALTERNATIVE SMELTING TECHNOLOGIES13
Company
Outokumpu
Inco
Mitsubishi
Noranda
Furnace/
process
type
Flash
Flash
Mitsubishi
Noranda
Nature
of test or
experience
Pilot-
scale
Bench-
scale
N/A
Commercial
scale
Maximum
in feed
As
N/A
1.0
0.33
1
level of
, weight
Sb Pb
N/A 5
0.3
c
8
impurity
percent
Zn
10
c __c
C 6.1
10
Refer-
ence
139
131
140
141
By definition, feeds containing more than 0.2 percent As, or 0.1 percent
Sb, or 4.5 percent Pb, or 5.5 percent Zn.
Other than reverberatory or electric furnaces.
cMaximum levels reported were below the "High-Irnpurity" limits.
Operated for the production of high-grade matte (70 to 75 percent Cu).
3-110
-------
during electrolytic refining. Also, use of the Noranda process would lead
to a mixed dust containing arsenic and lead (from the reactor itself),
which could complicate impurity recovery. As indicated previously,
current practice involves segregation of these two impurities during
the smelting process for subsequent recovery. For these reasons, the
Noranda process operated for the production of high grade matte is not
considered to be a viable option for processing high-impurity feed
materials in this analysis.
The combination of sulfidizing, roasting and flash-furnace smelting
shows potential as an option for processing feeds containing high
levels of arsenic and antimony, although the sulfidizing-roast process
is not yet demonstrated. The sulfidizing-roast process could conceivably
be used to reduce the volatile impurity level of the feed to those
levels processed by flash smelters. This process would allow the
segregation of arsenic into an arsenic-rich dust from which arsenic
could be recovered. Because sulfur removal by the process is fairly
high, however (30 to 40 percent), it is possible that, for some feeds,
insufficient sulfur would be present after roasting to produce a
low-grade matte during flash smelting.
3.6 BASELINE EMISSIONS
The baseline control level is that level of emission control that
is required at the source under consideration in the absence of a
revised NSPS. This level is determined from an examination of all
pertinent regulations. Baseline requirements are discussed here for
both process and fugitive sources.
3.6.1 Process Sources
The baseline control level for most new process sources is given
by the existing NSPS regulation. For example, new roasters, converters,
and smelting furnaces—with the exception of reverberatory furnaces
processing high-impurity feeds—are subject to control by double-contact
acid plants, or the equivalent. Under the NSPS, reverberatory furnaces
processing high-impurity feeds are exempted from control of both S02
and particulate matter.
3-111
-------
The SIP regulations for each state having active copper smelters
are pertinent to defining the regulatory baseline for new reverberatory
furnaces processing high-impurity feeds. Regulations governing sulfur
dioxide emissions are presented in Table 3-18, and those governing
particulate matter emissions are presented in Table 3-19.
In the case of both sulfur dioxide and particulate matter, many
States have different regulations for new and existing smelters.
Considering the regulations for new smelters alone, three States--
Arizona, which has 7 smelters; Washington, which has a smelter
processing high-impurity feeds; and Tennessee—adopted the existing
NSPS regulation. Nevada and New Mexico require the removal of
90 percent of the feed sulfur for all new smelters, which corresponds
to the requirement that approximately 82 percent of the S02 in the
offgases from a new reverberatory furnace be controlled.* S02 regula-
tions adopted by Utah and Texas appear to be tailored to the existing
smelters in these States. The State of Michigan does not specify
regulations for S02.
Since the majority of the existing smelters are located in States
which have adopted the NSPS for new smelters, the baseline for S02
control for new reverberatory furnaces processing high-impurity feeds
is assumed to be the NSPS (which in effect is no control) for the
purpose of this analysis.
^Assuming the greenfield smelter operated similarly to ASARCO-
Tacoma, the feed sulfur balance (uncontrolled) would be as follows:143
multihearth roaster offgases--20 percent; reverberatory offgases--
28 percent; converter offgases--47 percent; fugitives--4 percent; and
reverberatory slag--l percent. Under the existing NSPS, the roaster
and converter streams would likely be controlled by double-contact
acid plants, which have an efficiency of ~98-. 5 percent. Hence, the
level of feed sulfur removal achieved would be 1 + (0.985)(20) +
(0.985)(47) = 67 percent. To meet the State requirement, however, an
additional 90 - 67 = 23 percent of the feed sulfur must be removed.
Assuming that all of the additional requirement is supplied from the
reverberatory stream (i.e., none of the fugitive sulfur is controlled),
the level of control required on the reverberatory would be (23/28) x
100 = 82 percent.
3-112
-------
TABLE 3-18. SULFUR/SULFUR DIOXIDE EMISSION LIMITATIONS BY STATE
State
Arizona
co
CO
Michigan
Nevada
Sources addressed
by regulations
New primary copper
smelters
Existing primary cop-
per smelters
New copper smelters
Kennecott Copper
Corporation, White
Pine County
New Mexico New smelters
Existing nonferrous
smelters
Emission limitations
40 CFR 60, Subpart P
is adopted
10 percent of feed
sulfur
10 percent of feed
sulfur
10,150 Ib/hr S02
(6-hour average).
May be raised to
29,000 Ib/hr on
approval.
10 percent of feed
sulfur
3,550 Ib/hr (24-hour
average)
Notes
Reverberatory furnaces proc-
essing high impurity feeds
are exempted from S02
regulations.
Determined from weighted
average (by feed sulfur)
of feed sulfur removal
required for each of the
7 smelters by the
September 20, 1979,
Arizona Multipoint Roll-
back (MPR) SIP revision.
No regulation could be
found.
Effective date of this
regulation has been
indefinately delayed.
(Reference 2).
Phelps Dodge-Hidalgo is
considered to be a new
source in New Mexico
Effective 12/31/82.
References
Reference 144, Arizona
Air Regulations,
Part R9-3-814.
Reference 145.
Reference 144, Michigan
Air Regulations.
Reference 144, Nevada Air
Regulations, Article 8.
Reference 144, Nevada
Air Regulations,
Article 14, and
40 CFR 52.1475.
Reference 144, New Mexico
Air Regulations,
Part 652.
Reference 144, New Mexico
Air Regulations,
Part 652.
(continued)
-------
Table 3-18. (continued)
State
Sources addressed
by regulations
Emission limitations
Notes
References
Tennessee
CO
Texas
Utah
New primary copper
smelters
Existing copper
smelters
Primary copper
smelters
Kennecott copper
smelter
Washington New primary copper
smelters
ASARCO-Tacoma smelter
40 CFR 60, Subpart P
is adopted
100 ppm from copper
smelters, 500 ppm
from sulfuric acid
plants
6,000 ppm (b.v.) for
reverberatory fur-
naces, 650 ppm
(b.v.) for sulfuric
acid plants and all
other processes.
6,030 Ib/hr (6-hr
average)
40 CFR 60, Subpart P
is adopted.
10 percent of feed
sulfur
Reverberatory furnaces proc-
essing high impurity feeds
are exempted from S02
regulations.
Regulation evidently written
for Cities Service--
Copperhill, which is
atypical among copper
smelters.
Reverberatory furnaces proc-
essing high impurity feeds
are exempted from S02
regulations.
Reference 144, Tennessee
Air Regulations,
Part 1200-3-16-21.
Reference 144, Tennessee
Air Regulations,
Part 1200-3-19-19.
Reference 144, Texas Air
Regulations, Part
131.04.01.016.
40 CFR 52.2335(d).
Reference 144, Washington
Air Regulations, Part
WAC 173-400-115.
Reference 146.
-------
State
GO
I
en
Nevada
Sources addressed
by regulations
TABLE 3-19. PARTICULATE EMISSION LIMITATIONS BY STATE
Notes
Arizona New primary copper
smelters
Existing primary cop-
per smelters
Michigan General sources
New Mexico
Primary nonferrous
smelters
New nonferrous
smelters
Existing nonferrous
smelters
Emission limitations
40 CFR 60, Subpart P
is adopted
E = 3.59 P0-62 for
P<30 tons/hr
E = 17.31 P0-16 for
P>30 tons/hr
E = 4.10 P0-67 for
P<30 tons/hr
E = 55.0 P0-n-40 for
P>30 tons/hr
1,300 Ib/hr solid
particulate matter
2,100 Ib/hr total
particulate matter
0.03 gr/dcf
0.05 gr/dcf for rever-
beratory furnaces.*
0.05 gr/dcf for acid
plants and reverber-
atory feed dryers.
No particulate regulations
are specified for rever-
beratory furnaces proc-
essing high impurity
feeds.
P = process feed rate in
tons/hr, E = particulate
emissions in Ib/hr.
Units of P and E same as
above.
^Effective 12/31/82.
References
Reference 144, Arizona
Air Regulations, Part
R9-3-814.
Reference 144, Arizona
Air Regulations, Part
R9-3-515B.
Reference 144, Michigan
Air Regulations,
Part 3.
Reference 144, Nevada
Air Regulations,
Article 7.
Reference 144, New Mexico
Air Regulations,
Part 506.
Reference 144, New Mexico
Air Regulations,
Part 506.
(continued)
-------
TABLE 3-19. (continued)
State
Sources addressed
by regulations
Emission limitations
Notes
References
Tennessee
CTl
lexas
Utah
Washington
New primary copper
smelters
Existing sources
(general)
General
except certain
steam generators
Kennecott Copper
Corporation Smelter,
Salt Lake County
New primary copper
smelters
General process
sources
40 CFR 60, Subpart P
is adopted.
E = 4.10 P0-67 for
P<30 tons/hr
E = 55.0 P0-n-40
for P>30 tons/hr
E = 0.048q0.62
364 Ib/hr
40 CFR 60, Subpart P
is adopted
0.10 gr/dscf
No particulate regulations
are specified for rever-
beratory furnaces proc-
essing high impurity
feeds.
P = process feed rate in
tons/hr, E = particulate
emissions in Ib/hr.
q = stack effluent rate in
acfm, E = particulate
emissions in Ib/hr.
Annual average for smelter
main stack.
No particulate regulations
are specified for rever-
beratory furnaces proc-
essing high impurity
feeds.
Reference 144, Tennessee
Air Regulations,
Part 1200-3-16-21.
Reference 144, Tennessee
Air Regulations, Part
1200-3-7.
Reference 144, Texas Air
Regulations, Reg. I,
Chapter 3.
Reference 144, Utah Air
Regulations, Part
3.2.1.
Reference 144, Washington
Air Regulation,
Part WAC 173-400-115.
Reference 144, Washington
Air Regulations,
Part WAC 173-400-060.
-------
With regard to particulate matter, the existing NSPS regulation
does not address control requirements. However, new reverberatory
furnaces processing high-impurity feeds are expected to be subject to
regulations at least as stringent as those pertaining to existing
smelters. Based on the regulations in Table 3-19, particulate control
requirements for existing furnaces range from a low of about 56 percent
(in Nevada) to a high of about 98 percent (in New Mexico and Utah).
For purposes of this analysis, the baseline for particulate control
for new reverberatory furnaces processing high-impurity feeds is taken
to be the statewide average of about 91 percent.
3.6.2 Fugitive Sources
Regulations governing fugitive emissions of S02 and particulates
include the arsenic regulation of the Occupational Safety and Health
Administration (OSHA), and SIP's.
The OSHA arsenic standard limits occupational exposure to inorganic
arsenic to 10 ng/m3, averaged over an 8-hour period. As of February
1982, exposure in various job classifications at ASARCO-Tacoma, ASARCO-
El Paso, ASARCO-Hayden, Kennecott-McGill, and Kennecott-Garfield
exceeded the limit. To reduce exposures, OSHA is requiring that local
ventilation/hooding be installed (or upgraded, if already in place) on
the following fugitive emission sources: roaster calcine discharge,
matte tapping, and slag skimming.147 Converters are required to
install secondary hooding. Approximate compliance dates are April 1984
for converter secondary hoods, and June 1985 or later for all other
local ventilation systems.147
Of the five smelters affected, feeds processed by the Kennecott-
Garfield smelter are believed to contain the lowest quantity of arsenic--
approximately 0.14 percent. It is assumed for this analysis that all
new smelters processing feeds containing in excess of this amount will
thus be required to install local ventilation/hooding systems on
roaster calcine discharge, matte tapping and slag skimming operations,
and converters.
For all copper smelters, independent of arsenic levels, the SIP
regulations are believed to impose similar requirements for fugitive
3-117
-------
emissions control. The Multipoint Rollback (MPR) regulations adopted
by the State of Arizona pertain to stack emissions. However, each
source is required to demonstrate that fugitive emissions will not
lead to violations of the National Ambient Air Quality Standards
(NAAQS). As yet, existing smelters have not been able to make this
demonstration. In this analysis, it is assumed that these smelters
will be found to be in violation of the NAAQS as a result of fugitive
emissions and will be forced to implement capture systems (on roaster
calcine discharge, matte tapping, slag skimming, and converters)
coupled with dispersion to achieve compliance. Furthermore, it is
assumed that all other smelters will also be required to implement
similar fugitive controls as a result of future revision to their
SIP's. Hence, it is projected that new smelters would likewise be
subjected to such controls on the fugitive sources under consideration.
3.7 REFERENCES
1. Edwards, 0., and P. Robbins. Guide to Non-Ferrous Metals and
Their Markets. London, Kogan Page Limited, 1979. p. 96.
2. Pennebaker, E. N. Copper Minerals, Ores, and Ore Deposits. In:
Copper—The Science and Technology of the Metal, Its Alloys and
Compounds, Butts, A. (ed.). American Chemical Society Monograph
Series. New York, Reinhold Publishing Co., 1954. p. 21-62.
3. Field Surveillance and Enforcement Guide for Primary Metallurgical
Industries. U.S. Environmental Protection Agency. Research
Triangle Park, N.C. Publication No. EPA 450/3-73-002. December
1973. p. 175-180.
4. Letter and attachments from Henderson, J. M., ASARCO, to
Goodwin, D. R., U.S. Environmental Protection Agency. January 11,
1982. p. 1. Response to Section 114 letter on primary copper
smelters.
5. Letter and attachment from George, J. W., Cities Service Company,
to Goodwin, D. R., U.S. EPA. Feb. 4, 1982. Response to Section 114
letter.
6. Letter and attachment from Maksym, J. W., Copper Range Company,
to Goodwin, D. R., U.S. EPA. August 18, 1981. Response to
Section 114 letter.
3-118
-------
7. Arsenic Emissions from Primary Copper Smelters—Background Informa-
tion for Proposed Standards, Preliminary Draft. U.S. Environmental
Protection Agency. Research Triangle Park, North Carolina.
Publication No. EPA-450/ . p. 3-2.
8 Letter and attachments from Mai one, R. A., Kennecott Minerals
Co., to Goodwin, D. R., U.S. EPA. December 23, 1981. Response
to Section 114 letter.
9. Telecon. Clark, T. C., Research Triangle Institute, with Bennett,
K. C. , Phelps Dodge Corporation. October 13, 1982. Phelps Dodge
Plant Capacities.
10. Reference 4, p. 13.
11. Biswas, A. K., and W. G. Davenport. Extractive Metallurgy of
Copper. New York, Pergamon Press, 1980. p. 71.
12. Reference 4, p. 2.
13. Institute for Copper. Air Pollution Caused by Copper Metallurgy
in Bor, Volume II—Multilevel roasters. Bor, Yugoslavia. June 1975.
p. 115-118.
14. Carpenter, B. H. Nonferrous Smelter Studies: Theoretical Investi-
gation of Role of Multihearth Roaster Operations in Copper Smelter
Gas Blending Schemes for Control of S02. Environmental Science
and Technology. 12:57-62. January 1978.
15. Reference 11, p. 74.
16. Weisenberg, I. J., and R. C. Hill. Design, Operating, and Emission
Data for Existing Primary Copper Smelters (Draft). Pacific
Environmental Services. Santa Monica, CA. EPA Contract No. 68-02-
2606. March 1978. p. 2-91.
17. Reference 11, p. 83.
18. Reference 11, p. 83-84.
19. Boldt, J. R. Jr., and P. Queneau. The Winning of Nickel. Longmans
Canada Limited, Toronto, Canada, 1967. p. 230.
20. Reference 7, p. 3-11.
21. Reference 7, p. 3-12.
22. Weisenberg, I. J., and J. C. Serne. Design and Operating
Parameters for Emission Control Studies: Phelps Dodge Ajo Copper
Smelter. U.S. Environmental Protection Agency. Research Triangle
Park, N.C. Publication No. EPA-600/2-76-036f. February 1976.
p. 8.
3-119
-------
23. Reference 4, p. 16.
24. Reference 11, p. 139.
25. Reference 11, p. 141-142.
26. Reference 11, p. 142.
27. Reference 11, p. 138.
28. Reference 16, p. 2-144.
29. Reference 11, p. 140.
30. Reference 11, p. 156.
31. Background Information for New Source Performance Standards:
Primary Copper, Zinc, and Lead Smelters—Volume 1: Proposed
Standards. U.S. Environmental Protection Agency. Research
Triangle Park, N.C. Publication No. EPA-450/2-74-002a. October
1974. p. 3-49.
32. Reference 11, p. 157.
33. Antonioni, T. N., T. C. Burnett, C. M. Diaz, and H. C. Garven.
Inco Oxygen Flash Smelting of Copper Concentrate. Inco Metals
Company, Toronto, Ont. (Presented at the British Columbia Copper
Smelting and Refining Technologies Seminar. Vancouver. November 5,
1980.) 40 p.
34. Letter and attachments from Garven, H. C. , Inco Metals Company,
to Vervaert, A. E. , U.S. Environmental Protection Agency. May 18,
1982. Comments on draft Chapters 3-6 of the BID.
35. Letter and attachments from Cahill, L. G., ASARCO, to Bowerman,
L. J., U.S. Environmental Protection Agency. October 27, 1980.
Response to questions regarding changes to be made at the Hayden
smelter.
36. Telecon. Clark, T. C., Research Triangle Institute, with Garven,
H. C., Inco Metals Company. July 6, 1982. Response to questions
on oxy-fuel smelting and Inco oxygen flash smelting.
37. Reference 11, p. 161.
38. Reference 11, p. 164.
39. Reference 11, p. 159.
40. Reference 11, p. 167.
3-120
-------
41. Reference 16, p. 2-46.
42 Moriyama, K., T. Terayama, T. Hayashi, and T. Kimura. The Applica-
tion of Pulverized Coal to the Flash Furnace at Toyo Smelter.
In: Copper Smelting—An Update, George, D. B. and J. C. Taylor
(eds.) New York, American Institute of Mining, Metallurgical,
and Petroleum Engineers, Inc. 1981. p. 201-212.
43. Reference 11, p. 160.
44 Letter and attachment from Scanlon, M. P., Phelps Dodge Corpora-
tion, to Goodwin, D. R., U.S. Environmental Protection Agency.
October 18, 1978. Response to questions on the Phelps Dodge-
Hidalgo smelter, p. 9.
45. Reference 11, p. 210.
46. Reference 31, p. 3-81.
47. Reference 31, p. 3-82.
48. Johnson, R. E., N. J. Themelis, and G. A. Eltringham. A Survey
of Worldwide Copper Converter Practices. In: Copper and Nickel
Converters, Johnson, R. E., (ed.). New York, American Institute
of Mining, Metallurgical, and Petroleum Engineers. 1979. p. 12A.
49. Reference 48, p. 10A.
50. Reference 4, p. 17.
51. Reference 4, p. 18.
52. Reference 31, p. 3-89.
53. Lenoir, P. J., J. Thiriar, and C. Coekelbergs. Use of Oxygen
Enriched Air at the Metallurgie Hoboken-Overpelt Smelter. In:
Advances in Extractive Metallurgy and Refining, Jones, M. J.
(ed.). London, The Institution of Mining and Metallurgy. 1971.
p. 379.
54. Trip Report. Carpenter, B. H., J. Wood, and C. Clark, Research
Triangle Institute with Tittes, A. F., and T. Larsen, Inspiration
Consolidated Copper Company. February 19, 1981. Familiarization
Plant Visit.
55. Reference 16, p. 2-149.
56. Reference 11, p. 242.
57. Reference 7, p. 3-17.
3-121
-------
58. Reference 11, p. 244-245.
59. Reference 11, p. 243.
60. Reference 11, p. 245.
61. Trujillo, A. D, R. G. Kindel, 0. I., Gonzalez, and S. N. Sharma.
Fire Refining Practice at Chino Mines Division of Kennecott
Copper Corporation. The Metallurgical Society of AIME. Warren-
dale, PA. Paper No. A-79-36. 1979. p. 7.
62. Miller, H. J. The Fire Refining of Copper. In: Copper—The
Science and Technology of the Metal, Its Alloys and Compounds,
Butts, A. (ed.). New York, Reinhold Publishing Corporation.
1954. p. 297.
63. Letter and attachment from Nelson, K. W. , ASARCO, to Cuffe,
S. T., U.S. Environmental Protection Agency. May 19, 1982.
Comments on draft Chapters 3-6 of BID. p. 7-8.
64. Bigley, A. C., Jr. Impurity Distribution and Control in Copper
Smelters and Refineries. The Metallurgical Society of AIME.
Warrendale, PA. TMS Paper No. A78-67. 1978. P. 3.
65. Reference 11, p. 224.
66. Dayton, S. Utah Copper and the $280 Million Investment in Clean
Air. Engineering and Mining Journal. 180:72-83. April 1979.
67. Trip Report. Carpenter, B. H. and Wood, J. P., Research Triangle
Institute with Templeton, F., Heaney, R., and Taylor, S., Kennecott
Minerals Co. May 6, 1981. Familiarization Plant Visit.
68. Suzuki, T. The Mitsubishi Process—Operation of Semi-Commercial
Plant. In: The Future of Copper Pyrometallurgy, Diaz, C. (ed.).
Santiago, The Chilean Institute of Mining Engineers. 1974. p.
107.
69. Nagano, T., and T. Suzuki. Commercial Operation of Mitsubishi
Continuous Copper Smelting and Converting Process. In: Extractive
Metallurgy of Copper, Volume I, Yannopoulos, J. C., and Agarwal,
J. C. (eds.). The Metallurgical Society of AIME. Baltimore,
Port City Press. 1976. p. 439-457. 3.2.6
70. Amsden, M. P., R. M. Sweetin, and D. G. Treilhard. Selection and
Design of Texasgulf Canada's Copper Smelter and Refinery. Journal
of Metals. 30:16-26. July 1978.
71. Compilation of Air Pollutant Emission Factors. U.S. Environmental
Protection Agency. Research Triangle Park, N.C. Publication
No. AP-42. August 1977. p. 7.3-2.
3-122
-------
72 Cassady, M. E., OSHA Health Response Team. Comments on draft
Chapter 3 of the background information document for revision of
the NSPS for primary copper smelters. Received June 7, 1982. p.
3-47.
73. Reference 7, p. 3-32.
74. Reference 11, p. 161.
75 Letter and attachments from Judd, L. R., Phelps Dodge Corporation,
to Cuffe, S. T., U.S. Environmental Protection Agency. November 3,
1981. Transmittal of process information pertinent to emissions
testing.
76. Reference 7, p. 3-38.
77. Reference 7, p. 3-39.
78 Pacific Environmental Services. Development of Fugitive Emission
Factors for Primary Copper Smelters. U.S. Environmental Protection
Agency. Research Triangle Park, N.C. Contract No. 68-02-3060.
December 1981. 40 p.
79. Reference 78, p. 2.
80. Reference 78, p. 3.
81. Reference 44, p. 9, 18.
82. Boggs, W. B., and J. N. Anderson. The Noranda Smelter. American
Institute of Mining, Metallurgical, and Petroleum Engineers
Transactions. 106:165-201, 1933.
83. Telecon. Clark, T. C., Research Triangle Institute, with Lee,
L. V., Dorr-Oliver, Inc. December 18, 1981. Increasing fluid-bed
roaster capacity.
84 Mulholland, L. E. , and D. J. Nelson. Operation of the Fluo-Solids
Roaster at Kennecott's Ray Mines Division. In: Copper Metallurgy,
Erlich, R. P. (ed.). New York, The Metallurgical Society of
AIME. 1970. p. 141-145.
85. Blair, J. C. Fluo-Solids Roasting of Copper Concentrates at
Copperhill. In: Pyrometallurgical Processes in Nonferrous
Metallurgy, Anderson, J. N., and P. E. Queneau (eds.). New York,
Gordon and Breach. 1967. p. 55-65.
86. Telecon. Clark, T. C., Research Triangle Institute, with Johnson,
R. E. , Phelps Dodge Corporation. August 7, 1981. Industry
expansion options.
3-123
-------
87. Reference 4, p. 8.
88. Itakura, K. , T. Nagano, and J. Sasakura. Converter Slag Flotation-
Its Effect on Copper Reverberatory Smelting Process. Journal of
Metals. 21:30-34. July 1969.
89. Achurra H. , J. , R. Espinosa G. , and L. Torres J. Improvements in
Full Use of Oxygen in Reverberatory Furnaces at Caletones Smelter.
IMS Paper No. A77-91. Metallurgical Society of AIME. 1977.
90. Reference 11, p. 134.
91. Queneau, P. E. , and R. Schuhmann, Jr. Metamorphosis of the
Copper Reverberatory Furnace: Oxygen Sprinkle Smelting. Journal
of Metals. 31:12-15. December 1979.
92. Saddington, R., W. Curlock, and P. Queneau. Tonnage Oxygen for
Nickel and Copper Smelting at Copper Cliff. Journal of Metals
18(4):440-452. April 1966.
93. Reference 11, p. 135.
94. Eastwood, W. B., J. S. Thixton, and T. M. Young. Recent Develop-
ments in the Smelting Practice of Nchanga Consolidated Copper
Mines, Rokana Smelter. TMS Paper No. A71-75. Metallurgical
Society of AIME. 1971.
95. Gibson, N. The Application of Oxygen in a Copper Smelter.
Journal of the South African Institute of Mining and Metallurgy
(Transvaal). 74:303-311. March 1974.
96. Kupryakov, Y. P., et al. Operation of Reverberatory Furnaces on
Air-Oxygen Blasts. Tsvetnye Metally, (English Translation).
10:13-16. February 1969.
97. Kupryakov, Y. P., and N. I. Artemiev. The Use of Oxygen in
Reverberatory Furnaces in USSR Smelters. In: The Future of
Copper Pyrometallurgy, Diaz, C. (ed.). Santiago, The Chilean
Institute of Mining Engineers. 1974. p. 193-197.
98. Goto, M. Green-Charge Reverberatory Furnace Practice at Onahama
Smelter. In: Extractive Metallurgy of Copper, Volume I,
Yannopoulos, J. C., and J. C. Agarwal (eds.). The Metallurgical
Society of AIME. Baltimore, Port City Press. 1976. p. 154-167.
99. Schwarze, H. Oxy-Fuel Burners Save Energy at El Teniente's
Caletones Smelter. World Mining. 30:58-61. May 1977.
100. Blanco, J. A., T. N. Antonioni, C. A. Landolt, and G. J. Danyliw.
Oxy-Fuel Smelting in Reverberatory Furnaces at Inco's Copper
Cliff Smelter. Inco Metals Company, Copper Cliff, Ontario.
(Presented at 50th Congress of the Chilean Institute of Mining
and Metallurgical Engineers. Santiago. November 23-29, 1980.)
16 p.
3-124
-------
101. Telecon. Clark, T. C. , Research Triangle Institute, with
Garven, H. C., Inco Metals Company. July 9, 1982. Matte grade
considerations with oxy-fuel smelting.
102. Successful Tests Encourage Phelps Dodge to Modify its Copper
Smelter. Chemical Engineering. 89(3):18-19. February 8, 1982.
103. Telecon. Clark, T. C., Research Triangle Institute, with
Chen, W. J., Phelps Dodge Corporation. May 28, 1982. Oxy-sprinkle
smelting technology.
104. Telecon. Massoglia, M. F., Research Triangle Institute, with
Henderson, J. M., ASARCO. June 25, 1982. Comments on Chapters 3-6
of the BID.
105. Reference 63, p. 13.
106. Reference 11, p. 118.
107. Letter from Henderson, J. M., ASARCO, to Clark, T. C., Research
Triangle Institute. August 4, 1982. Clarification of comments
on Chapters 3-6 of the BID. p. 3.
108. Weisenberg, I. J., T. Archer, F. M. Winkler, T. J. Browder, and
A. Prem. Feasibility of Primary Copper Smelter Weak Sulfur
Dioxide Stream Control. U.S. Environmental Protection Agency.
Cincinnati, OH. Publication No. EPA-600/2-80-152. July 1980.
p. 28.
109. Newton, J., and C. L. Wilson. Metallurgy of Copper. New York,
John Wiley and Sons, 1942. p. 82.
110. Reference 107, p. 1.
111. Telecon. Clark, T. C., Research Triangle Institute, with Garven,
H. C., Inco Metals Company. March 30, 1982. Sidewall cooling
considerations with oxy-fuel burners.
112. Reference 4, p. 6.
113. EPA Action May Lead to New Copper Smelter at Chi no. Big Sky
Paydirt. #16:57-58.
114. Telecon. Clark, T. C., Research Triangle Institute, with Persson,
J. A., Lectromelt Corporation. August 6, 1981. Increasing
electric furnace capacity.
115. Trip Report. Carpenter, B. H., J. Wood, and C. Clark, Research
Triangle Institute, with Shaw, M. F., and A. S. Gillespie, Phelps
Dodge Corporation Hidalgo Smelter. February 17, 1981. Familiari-
zation plant visit.
3-125
-------
116. Juusela, J., S. Harkki, and B. Andersson. Outokumpu Flash Smelting
and Its Energy Requirement. In: Effic. Use Fuels Metal 1. Ind.
Symp. Pap. Chicago, Inst. Gas Technol. 1974. pp. 555-575.
117. Trip Report. Clark, T. C., Research Triangle Institute, Vervaert,
A. E. , and F. Clay, U.S. Environmental Protection Agency, with
Winslow, R. L, J. Brandt, and W. J. Chen, Phelps Dodge Corporation
Hidalgo Smelter. August 25, 1981. Pretest survey visit.
118. Telecon. Clark, T. C., Research Triangle Institute, with Weddick,
A. J., Kennecott Copper Corporation. August 19, 1981. Noranda
Reactors.
119. Reference 4, p. 3.
120. Reference 48, p. 1-32.
121. Reference 7, p. 3-5.
122. Letter and attachment from Henderson, J. M., ASARCO, to Clark, T. C.,
Research Triangle Institute. February 23, 1982. Response to
questions on blister and anode copper impurity levels.
123. Letter and attachments from Loughridge, K. D., ASARCO, to Goodwin,
D. R., EPA. October 9, 1975. Response to Section 114 letter on
primary copper smelters.
124. Reference 4, Attachment 1.
125. Paulson, D. L., W. Anable, W. L. Hunter, and R. S. McClain.
Smelting of Arseniferous Copper Concentrate in an Electric-Arc
Furnace. U.S. Bureau of Mines. Washington, D.C. Report of
Investigations 8144. 1976.. 30 p.
126. Reference 4, p. 21.
127. Reference 7, p. 3-24.
128. Reference 4, p. 27.
129. Tuovinen, H. , and P. Setala. Removal of Harmful Impurities from
Iron, Copper, and Cobalt Concentrates and Ores. The Metallurgical
Society of AIME. Warrendale, PA. TMS Paper No. A82-4. 20 p.
130. Letter from Henderson, J. M., ASARCO, to Clark, T. C., Research
Triangle Institute. August. 13, 1982. Response to questions on
impurity elimination, p. 4.
131. Victorovich, G. S., C. Diaz, and J. Raskauskas. Impurity Distribu-
tions, Dusting and Control of Matte Grade in Inco Oxygen Flash
Smelting. Inco Metals Company. Mississauga, Ontario. (Presented
at Fiftieth Anniversary Meeting of the Chilean Institute of
Mining Engineers. Santiago. November 1980..) 14 p.
3-126
-------
132. Trip Report. Clark, T. C., and B. H. Carpenter, Research Triangle
Institute, with Garven, H. C., D. C. Lowney, and C. M. Diaz Inco
Metals Company. April 28-29, 1981. Familiarization visit to
Inco Metals Company and Copper Cliff smelter.
133 Mackey P J., G. C. McKerrow, and P. Tarassoff. Minor Elements
in the Noranda Process. Noranda Mines Limited. Noranda, Quebec.
(Presented at the 104th Annual AIME Meeting. New York. February 16-
20, 1975.) 27 p.
134. Reference 11, p. 237.
135 George, D. B., J. W. Donaldson, and R. E. Johnson. Minor Element
Behavior in Copper Smelting and Converting. In: World Mining
and Metals Technology, Vol. I, Weiss, A. (ed.). Baltimore, Port
City Press. 1976. p. 534-549.
136 Telecon. Clark, T. C., Research Triangle Institute, with George,
D. B., Kennecott Minerals Company. July 27, 1982. Impurity
elimination during converting.
137. Reference 4, p. 22.
138 Mohri, E., and M. Yamada. Recovery of Metals from the Dusts of
Flash Smelting Furnace. In: World Mining and Metals Technology,
Vol. I., Weiss, A. (ed.). Baltimore, Port City Press. 1976.
p. 520-531.
139 Letter and attachments from Harkki, S., Outokumpu Oy, to Clark, T. C.,
Research Triangle Institute. April 22, 1981. Response to questions
on Outokumpu flash furnaces.
140. Letter and attachment from Sukekawa, I., Mitsubishi Metal Corporation,
to Clark, T. C., Research Triangle Institute. October 13, 1981.
Response to questions on the Mitsubishi process.
141 Letter from Mackey, P. J., Noranda Mines, Ltd., to Clark, T. C.,
Research Triangle Institute. June 1, 1982. High impurity feed
processing experience with the Noranda process.
142 Letter from Mackey, P. J., Noranda Mines, Ltd., to Clark, T. C.,
Research Triangle Institute. May 26, 1982. Comments on Chapters
3-6 of the draft BID.
143 Weisenberg, I. J., and J. C. Serne. Design and Operating
Parameters for Emission Control Studies: ASARCO, Tacoma, Copper
Smelter U.S. Environmental Protection Agency. Research Triangle
Park, NC. Publication No. EPA-600/2-76-036K. February 1976.
30 pp.
144. Sections 201-556, State Air Laws. Environmental Reporter.
Bureau of National Affairs, Inc., Washington, D.C.
3-127
-------
145. Memorandum from Crow, S. F. , EPA Region 9 Administrator, to
Bennett, K. M., EPA Assistant Administrator for Air, Noise, and
Radiation. October 29, 1981. Congressional Request for Smelter
Information.
146. Memorandum from Rathbun, R. A., EPA, to Pratapas, J., U.S. Environ-
mental Protection Agency. December 21, 1981. Smelter Information.
147. Telecon. Clark, T. C., Research Triangle Institute, with Cassady, M.
OSHA Health Response Team. February 3, 1982. Engineering control
requirements under the OSHA arsenic regulations.
3-128
-------
4. EMISSION CONTROL TECHNIQUES
4.1 GENERAL
This chapter describes and evaluates emission control techniques
applicable in the primary copper smelting industry to reduce sulfur
dioxide (S02) and particulate matter emissions to the atmosphere,
including both primary process effluents and fugitive emissions.
For primary process effluents, controlling "weak" S02 streams
from reverberatory smelting furnaces is the primary topic discussed.
Control techniques assessed for possible application include the
following:
Contact sulfuric acid plants of the "dry gas" type.
Calcium-based flue gas desulfurization systems.
Ammonia-based flue gas desulfurization systems.
Magnesium-based flue gas desulfurization systems.
Flue gas desulfurization systems based upon a citric
acid—sodium citrate buffer.
The discussion of contact sulfuric acid plants consists of a summary
(Section 4.2.1), a general discussion of the contact process for
producing sulfuric acid (Section 4.2.2), a discussion of sulfuric acid
plant design and operating considerations (Section 4.2.3), and a
summary of sulfuric acid plant performance capabilities (Section 4.2.4).
The assessments of the various flue gas desulfurization (FGD)
processes are presented in Sections 4.3.2 through 4.3.5, each of which
contains the following:
4-1
-------
A summary of the technical feasibility of applying each
process to reverberatory furnace offgases.
A general discussion of the unit operations involved in each
process.
A detailed discussion of the design and operating considera-
tions involved in each process.
A discussion of the operational problems known to be
associated with each process.
A survey of operating experience for each process.
A discussion of the applicability of each process to rever-
beratory furnace offgases.
Section 4.3.6 presents general conclusions regarding the reliability
and performance of the various FGD systems discussed.
Section 4.4 describes several methods for increasing the S02
strength of reverberatory smelting furnace offgases. These methods,
all of which involve modifying furnace operation, are summarized as
follows:
Elimination of converter slag return (Section 4.4.1)
Sealing points of leakage (Section 4.4.2)
Preheating combustion air (Section 4.4.3)
Operation at lower air-to-fuel ratios (Section 4.4.4)
Predrying wet charges (Section 4.4.5)
Oxygen enhancement techniques (Section 4.4.6).
Section 4.4.6.1 presents detailed discussions of a number of oxygen
enhancement techniques, and Section 4.4.6.3 presents conclusions
regarding the technical feasibility of these techniques for possible
domestic applications.
Section 4.5 discusses blending of reverberatory furnace offgases
with strong S02 streams from other smelter processes. Gas blending is
assessed as a means by which to allow autothermal processing of "weak"
streams in contact sulfuric acid plants. Gas blending is analyzed for
two scenarios:
4-2
-------
A new smelter that process high impurity ore concentrates
(Section 4.6.1).
Existing smelters that undergo physical or operational changes
to achieve a greater production capacity (Section 4.6.2).
The former necessarily assesses the technical aspects of partial and
total weak-stream blending, while the latter assesses partial blending
of the weak stream, i.e., blending enough of the weak stream to ensure
that postexpansion S02 emissions are at or below the preexpansion S02
emission level.
This chapter also assesses the control of particulate matter
emissions associated with reverberatory smelting furnace effluents.
Sections 4.6.2, 4.6.3 and 4.6.4 consider venturi scrubbers, fabric
filters (baghouses), and electrostatic precipitators (ESP's),
respectively.
Finally, Section 4.7 describes techniques used to control fugitive
S02 and particulate matter emissions from sources within primary
copper smelters. Both local and general ventilation techniques are
considered. Sections 4.7.4, 4.7.5, and 4.7.6 discuss control of
fugitive emissions from roasting, smelting, and converting operations,
respectively. Section 4.7.7 summarizes visible emissions data for
fugitive emissions sources within primary copper smelters.
4.2 SULFURIC ACID PLANTS
4.2.1 Summary
The contact sulfuric acid process involving the catalytic oxidation
of S02 to S03 is the most widely used process for removing S02 from
primary copper smelter effluent gases.1 Sulfuric acid plants can be
designed to process feed streams that contain only a fraction of a
percent of S02; however, practical limitations have usually restricted
application to gas streams that contain at least 3.5 percent S02.
Metallurgical single-stage and dual-stage absorption sulfuric acid
plants constructed in the past have commonly been designed to operate
autothermally on feed streams that contain 3.5 and 4.0 percent S02,
respectively. It is technically feasible, however, to design acid
plants that will operate autothermally on feed streams that exhibit
S02 concentrations below the 3.5 to 4.0 percent range. Estimates have
4-3
-------
indicated that lowering the autothermal requirement for a dual-stage
absorption acid plant from 4.0 to 3.5 percent S02 would increase the
plant's installed capital cost by 8 to 13 percent. Similarly, the
installed capital cost of a single-stage absorption plant would increase
4 to 7 percent if the autothermal requirement were lowered from 3,5 to
3.0 percent S02. (See Appendix F for supporting calculations.)
Large fluctuations in feed stream volumetric flow rate and S02
concentration may adversely affect sulfuric acid plant operation at
copper smelting facilities. Fluctuations of either type tend to lower
the conversion of S02 to S03, thus decreasing sulfur recovery and
increasing S02 emissions to the atmosphere. Generally, acid plants
are designed to accommodate the highest volume of gas anticipated as
well as the lowest expected gas stream S02 concentration. This is
done to facilitate autothermal operation while maintaining a high S02
conversion efficiency.
Acid plant operation on copper smelter effluent gases is not
possible without adequate gas cleaning before the gases are sent to
the contact section of the plant.2 Gas cleaning involves cooling and
removal of particulates. The degree of cooling is sufficient to
condense volatilized impurities with subsequent removal by the gas
cleaning system. Major problems will result if the gases are not
adequately cleaned. Thus, proper design and maintenance of the gas
purification system are necessary to minimize these problems.
It is evident that blending multihearth roaster offgases with
offgases from other smelter operations to provide an acid plant
feedstock can result in the production of contaminated "black" acid.
This occurs due to the trace amounts of organic flotation agents that
may remain unoxidized after the roasting process. This problem may
also exist in cases where fluid-bed roasters are used to effect a
partial roast. The presence of contaminants in the product acid
reduces its value when compared to the optically clear acid that can
be produced from a feed stream that does not contain trace contaminants.
Acid plant vendors typically guarantee maximum emission concentra-
tions of 2,000 ppm for single-stage absorption plants and 500 ppm for
dual-stage absorption plants.3 4 A conversion of approximately 98.5
percent is required to ensure the 500-ppm concentration in the
4-4
-------
acid plant effluent. The above-mentioned guarantees take into account
the large fluctuations in volumetric flow rate and S02 concentration
generally encountered at primary copper smelters; however, no allowance
is made for increased emissions due to catalyst deterioration that
occurs between regularly scheduled catalyst screening operations.4
During steady state operation, which is practically impossible due to
the transient nature of copper smelter effluent gas streams, conversions
of 99.8 percent and above are expected in dual-stage absorption plants,
while conversions of 98.5 percent and above are expected in single-stage
absorption plants.
U.S. Environmental Protection Agency (EPA) source tests indicate
that the S02 conversion efficiency of metallurgical sulfuric plants
depends upon the frequency and magnitude of fluctuations in gas-stream
flow rate and S02 concentration and upon catalyst deterioration. EPA
analyses have indicated, however, that an averaging time of 6 hours
and a reference emission level of 10 to 20 percent above the commonly
accepted vendor/contractor S02 emission guarantees effectively masks
normal, short-term fluctuations in S02 emissions that occur because of
fluctuations in the feed-stream flow rate and S02 concentration.
Consequently, once an additional allowance of 10 percent is added, to
account for catalyst deterioration between screenings, a reference
emission level 30 percent in excess of that typically guaranteed by
the vendors allows for all factors that tend to increase emissions
above the vendor's guarantee. This is supported by EPA statistical
analyses performed with the above-mentioned source test data. S02
emission data from a dual-absorption acid plant processing copper
converter offgases were obtained by EPA using a continuous monitoring
system. These data indicated that S02 emissions from the acid plant
can be limited to 650 ppm or less 98.8 percent of the time.
The use of sulfuric acid plants to control S02 emissions from
primary copper smelters is a well-demonstrated technology. Currently,
12 of the 15 active domestic primary copper smelters produce sulfuric
acid from process offgases.
4-5
-------
4.2.2 General Discussion
The basic steps involved in the contact process for producing
sulfuric acid from S02-laden gases are as follows:
Gas cleaning and conditioning
Gas drying
Catalytic oxidation of S02 to S03
Absorption of S03 into a sulfuric acid solution to form addi-
tional sulfuric acid.
These procedures are shown schematically in Figure 4-1. Adequate gas
cleaning before the gas stream enters the contact section of the acid
plant is essential. Particulate matter and volatilized metals (e.g.,
impurities present in the feed) must be removed to avoid costly
shutdowns and maintenance.4 5 Generally, the feed gases first enter a
weak acid scrubber, where they are cooled to approximately 55° C
(130° F) by water evaporation.2 3 5 The gases are then cooled to
about 30° C (85° F) to reduce their water content to the level required
to maintain the acid plant water balance. The subsequent cooling is
generally accomplished in an additional packed- or tray-type scrubber
with liquor coolers in the recirculated weak acid stream. Finally,
the gases are passed through electrostatic mist precipitators to
remove traces of dust and acid that may remain after cooling. The
types of equipment used in the gas purification section of an acid
plant may vary somewhat; however, typical installations use scrubbing
towers, coolers, and electrostatic mist precipitators as the primary
conditioning equipment.5
After the gases are cleaned, they must be dried before entering
the contact section of the plant. Cleaned gases are dried by contact
with 93 percent acid. The cooled, dried gases are then passed through
a series of gas-to-gas heat exchangers, where their temperature can be
raised to the optimum temperature for the conversion of S02 to S03.
Single-stage absorption acid plants use three or four catalyst stages,
which constitute the converter. The clean, dry gases pass through the
catalyst beds where the conversion of S02 to S03 takes place. Because
the conversion of S02 to S03 is exothermic, the heat of reaction
4-6
-------
GAS CLEANING
SO2 - Laden Gases
Electrostatic
Precipitator or
Baghouse
Dust
Cooling and
Scrubbing
Facilities
Electrostatic
Mist
Precipitators
Weak Acid
and Solids
ACID PRODUCTION
To Atmosphere
Drying
Tower
Heat
Exchangers
Converter
To Atmosphere
4
First
Absorption
Tower
93% Acid
Pump
Tank
98% Acid
93% Acid
r
_!
1
Second
Absorption
Tower
i
-1
1
1
I
98% Acid
Pump
Tank
H2O
93% Acid
98% Acid
Denotes Major Process Streams in the
Single-Stage Absorption Process
Denotes Additional Process Streams
Required in the Dual-Stage Absorption Process
Figure 4-1. Contact sulfuric acid processes.
-------
generated in each catalyst bed must be removed for the optimum con-
version temperature to be maintained. This is accomplished by routing
the exit gas stream from each catalyst bed through the tube side of
the gas-to-gas heat exchangers used to raise the temperature of the
incoming gases just prior to their entry into the contact section of
the plant. The resultant S03-laden stream that exits the converter is
then passed to an absorber where the S03 is^absorbed by strong (slightly
over 98 percent) sulfuric acid. Maintaining the absorbing acid at
just over 98 percent ensures that the strong acid stream exiting the
absorber has a concentration very close to 98 percent.
In a dual-stage absorption plant, the unabsorbed S03 and remaining
S02 in the gas stream are reheated and reintroduced into the converter,
where a portion of the remaining S02 is converted to S03. The gases
leaving this second stage of conversion are then passed to the final
absorption tower where the S03 is absorbed from the gases. The exiting
gases are then treated to remove acid mist prior to being vented to
the atmosphere.
4.2.3 Design and Operating Considerations
Proper design of the gas purification section of an acid plant is
essential in avoiding excessive shutdowns and maintenance. The presence
of high levels of solid or gaseous contaminants in copper smelter
offgases can present many difficulties in the design of the gas purifi-
cation system. Generally, these contaminants must be removed before
the gas stream enters the contact section of the plant. The offgases
contain varying amounts of entrained dust, as well as fumes formed by
the vaporization of volatile components.4 After cleaning, the composi-
tion of the acid plant fuel gas is independent of the impurity levels
in the concentrate. Contaminants in the offgas stream include compounds
of arsenic, cadmium, antimony, and Tiercury. Copper, lead, and zinc
dusts are also commonly entrained in copper smelter offgas streams.
Dust and fumes are generally reclaimed for their economic: value through
the use of cyclones, ESP's, and baghouses. However, in many cases,
additional cleaning is required to remove residual quantities of
contaminants that would otherwise hinder acid plant operation.
Special design considerations are necessary if the offgases
contain appreciable amounts of halogens. If fluorine is present in
4-8
-------
the gas stream, two scrubbing towers in series may be required to
achieve complete removal.2 3 Complete flourine removal is necessary
to prevent catalyst poisoning in the contact section of the plant. A
carbon brick lining must be used for the first scrubber because fluorine
will attack the more commonly used acid-proof refractory linings.
This scrubber is usually a venturi-style or open-spray-type tower in
which the gas stream is quenched to its saturation temperature. The
second scrubber is usually a packed- or tray-type tower in which the
gases must be cooled to achieve the proper water content for acid
production. If the dust and/or fluorine content of the gases is not
excessive, both gas cooling and scrubbing can be accomplished in a
single packed- or tray-type tower.
Because of the possible presence of halogens, construction materials
are important considerations in the proper design of a gas purification
system.3 If halogens are not present in significant quantities,
20-alloy stainless steel is suitable for pumps, valves, and the liquor
coolers used to cool the gases to achieve proper water content.2
However, if appreciable amounts of halogens are present in the gases,
this equipment must be constructed of higher alloy steels, graphite,
or glass. Another alternative would be to line the equipment with an
appropriate plastic. The scrubbing towers usually consist of a carbon
steel shell with an impervious membrane and an acid brick lining,
although plastic can be used for lining in some areas.
Although halogens and dust are almost completely removed in the
scrubbers, some acid mist will remain entrained in the gas stream.
Acid mist is formed when small amounts of S03 present in the gas
stream react with water vapor in the gases to form sulfuric acid mist.
Most of the acid mist is condensed in the scrubbers; however, a small
portion remains in the gas stream. Sulfuric acid mist generally
consists of particles less than 5 (jm in diameter that are very difficult
to remove from a gas stream except by electrostatic precipitation.6
Thus, gases exiting the scrubbers—commonly containing 3,530 mg/Nm3
(1.4 gr/scf) or more of H2S04 as acid mist3--are generally routed to
two electrostatic mist precipitators placed in series to improve
4-9
-------
efficiency and reliability. Mist precipitators, usually constructed
of lead, achieve a particulate removal efficiency commonly greater
than 99 percent. Gases entering the contact section of the plant
should contain only S02, 02, N2, and H20 vapor. If the gases are not
cleaned sufficiently, the following major problems may result:
Aggravated corrosion of heat exchangers and carbon steel ducts
Plugging of catalyst beds
Partial deactivation of the catalyst
Production of poor quality acid.
After the gases are dried in the contact section of the plant,
they become essentially noncorrosive. Hence, carbon steel ducts can
be used for the remainder of the plant. The total pressure drop
through a clean, well-maintained dual-stage absorption plant is usually
about 0.5 atm (50,660 Pa).2 Normally, an additional 0.060 to 0.075
atm (6,080 to 7,600 Pa) are added in the design of the main blower to
compensate for pressure buildup that may occur in the system.
When multihearth roasters are used, it is not uncommon for trace
amounts of organic contaminants to pass through both the gas cleaning
section of the plant and the catalyst beds. These contaminants will
result in the production of dark, discolored acid ("black" acid). The
presence of contaminants in the product acid reduces its value in
comparison to the optically clear acid that can be produced from
offgases generated by other sources. These trace organic: contaminants
are produced in multi hearth roasters when various organic: agents used
in the floatation process are vaporized and only partially oxidized.
Normally, within fluid-bed roasters, organic flotation agents are
completely oxidized, and thus the product acid is free of organic
contaminants.4 An exception occurs when fluid-bed roasters are used
to effect a partial roast. In this case, trace amounts of organic
flotation agents may not be oxidized due either to the relatively low
temperatures involved or to the low concentration of oxygen in evidence
during partial roasting. Techniques exist to purify or bleach discolored
acid, but they are usually costly and may produce undesirable side
4-10
-------
effects. For example, organic contaminants can be oxidized by hydrogen
peroxide, although the oxidation is accompanied by the release of
water, which dilutes the product acid. However, outlets exist for
sulfuric acid that are not sensitive to acid color, such as the produc-
tion of fertilizers or refinery alkylation processes.4 One operator
reports that some black acid has been sold, after equalization of
freight, for $4.00 per ton.7
A major consideration in the design of a metallurgical acid plant
is the concentration of S02 in the acid plant feed gases. Although
sulfuric acid plants can be designed to process feed streams that
contain only a fraction of a percent of S02, economic considerations
have limited applications to higher concentrations. Metallurgical
sulfuric acid plants constructed in the past were commonly designed to
operate autothermally on feed streams containing 3.5 to 4.0 percent
S02. Single-stage absorption plants are commonly designed to operate
autothermally at 3.5 percent S02, while dual-stage absorption plants
are designed to operate autothermally at 4.0 percent S02.3 4 However,
these autothermal operating requirements can be lowered by designing
the plants to operate autothermally at lower feed stream S02 concentra-
tions. This has not been the practice in the past because the incremental
cost of reducing the autothermal requirements rises quite rapidly when
S02 concentrations below the 3.5 to 4.0 percent range are considered.
Thus, a somewhat larger capital investment is required to build sulfuric
acid plants that will operate autothermally below the 3.5 or 4.0
percent S02 levels. As noted earlier, however, the incremental costs
of lowering autothermal requirements by 0.5 percent may not be excessive.
(See Appendix F for supporting calculations.)
If acid plants do not operate autothermally, supplemental heat
must be supplied to maintain the appropriate conversion temperature in
the catalyst beds, thus increasing operating costs. Supplemental heat
is generally supplied via gas-fired preheaters that heat the gas
stream indirectly prior to entry into the acid plant converter. Since
offgas streams from reverberatory smelting furnaces typically exhibit
S02 concentration in the 0.5- to 2.5-percent range, a dual-absorption
plant designed to operate autothermally at 4.0 percent S02 would have
4-11
-------
to operate its preheater continuously to process such streams, thus
increasing the total annual operating cost associated with the acid
plant.
The amount of gas cooling required in the gas purification section
of the plant depends upon the S02 concentration in the inlet gas
stream, the concentration of the product acid, and the plant eleva-
tion.2 3 4 Feed streams that contain less than 4 percent S02 fre-
quently require extra cooling to remove excess water vapor. Gas
streams containing less than 3 percent S02 may require refrigeration
to condense enough water to obtain an acceptable water-to-sulfur
ratio. Careful control of the gas stream water content is essential
in preventing dilution of the product acid below commercial-grade
strength.5
The oxygen content of the feed gases is also an important
consideration in metallurgical acid plant design. Oxygen is generally
fed in excess of the amount required by the reaction stoichiometry.
Most plants operate at an oxygen-to-sulfur-dioxide ratio (02/S02) of
not less than l.l.8 In some cases, when the gas stream has an S02
concentration of 9 percent or more, it does not contain sufficient
oxygen for the conversion of S02 to S03. When this occurs, the gas
stream must often be diluted with air or other offgases to enhance the
oxygen content of the stream.4
For maximum operating efficiency, metallurgical sulfuric acid
plants should operate on a gas stream of uniform flow rate and com-
position. Large fluctuations of either type tend to lower the conversion
of S02 to S03, thus decreasing acid production and increasing S02
emissions to the atmosphere. Metallurgical acid plants must therefore
be designed with the worst possible operating conditions in mind.
This objective requires the incorporation of adequate process control
technology in the design to allow operators to compensate for varia-
tions in S02 concentration. Variations in feed stream volume are
generally less of a problem than variations in S02 concentration and
can be tolerated within reasonable limits.4 The Norddeutsche Affinerie
of Hamburg, Germany, operates a dual-stage absorption plant that
processes offgases from a Outokumpu flash smelting furnace as well as
4-12
-------
offgases from converters. This particular plant was designed to
process a feed stream that varies from 76,000 Nm3/m (2,684,000 scfm)
to 200,000 NrnVm (7,063,000 scfm) while maintaining a stack concentration
of not more than 500 ppm S02. This demonstrates the ability of contact
sulfuric acid plants to handle rather large fluctuations in the feed-
stream volume. Generally, acid plants are designed to accommodate the
highest volume of gas anticipated as well as the lowest expected
gas-stream S02 concentration. This is done to facilitate autothermal
operation while maintaining a high S02 conversion efficiency.
Because capital and operating costs are directly related to the
gas volume to be handled, a design that will significantly reduce the
volume of gas in the plant will have a major effect on both cost
factors.8 9 The most cost-effective design will maximize the S02
concentration in the feed stream while simultaneously reducing the
total gas volume. Section 4.4 contains discussions of several methods
that can be used to increase the S02 concentration in smelter offgases.
4.2.4 Acid Plant Performance Characteristics
Metallurgical acid plant vendors currently guarantee maximum S02
emissions concentrations of 2,000 ppm for single-stage absorption
plants and 500 ppm for dual-stage absorption plants.3 4 10 li However,
these guarantees refer to emissions that occur during new plant per-
formance tests,4 which are conducted for 3 to 5 consecutive days while
the plant is operating, without any malfunctions, on gases that contain
the percentages of S02 specified in the design basis. In addition,
although these guarantees are for maximum S02 emissions and thus
include inherent allowances for increased emissions due to fluctuations
in the inlet S02 concentration, they do not include allowances for
increased S02 emissions due to catalyst or plant deterioration with
age. Furthermore, these guarantees are for acid plant performance
only and thus do not include emissions to the atmosphere during periods
of acid plant shutdown for catalyst screening or replacement or for
other acid plant maintenance. One domestic manufacturer of metallurgical
sulfuric acid plants guarantees maximum S02 emissions of 650 ppm from
its dual-stage absorption plants for a period of 1 year after startup.10
This guarantee assumes continuous operation with no shutdowns. This
4-13
-------
particular guarantee obviously includes allowances for increased
S02 emissions that occur during the 1-year period between startup and
the first scheduled catalyst screening.
The degree of catalyst deterioration over a given period of time
will depend largely on the levels of impurities present in the feed
stream. Table 4-1 presents estimates of the maximum levels of impurities
that can be tolerated in smelter offgases to be processed in a sulfuric
acid plant. The degree of catalyst deterioration experienced at these
various impurity levels can be accommodated by an acid plant that
shuts down once per year to screen the catalyst and repair equipment.4
Table 4-1 also contains the estimated upper levels of impurities that
can be removed by typical gas purification systems having prior coarse
dust removal. As discussed previously, more elaborate cleaning systems
must be designed if the feed stream contains contaminants, such as
halogens, that pose special problems. The details of the more elaborate
designs vary depending upon the contaminants involved, but, the designs
generally involve the use of more efficient dust or mist collectors
and the scrubbing of gases with liquids that absorb the contaminants.
Although complete removal of impurities from the feed stream is not
practical, 99.5 to 99.9 percent overall removal is considered to be
attainable.4
Because vendor guarantees on S02 emissions are commonly based on
emissions that occur during new plant performance tests, the effect of
catalyst deterioration on S02 emissions must be determined to accurately
assess acid plant performance capability over time. S02 emissions
data gathered by simultaneous EPA source testing of the No. 6 and No.
7 single-stage absorption acid plants at the Kennecott Garfield smelter
during the period of June 13-16, 1972, indicate that normal catalyst
deterioration and differences in acid plant design and technology can
result in a 30-percent increase in S02 emissions.13 A summary of this
analysis can be found in Appendix G.
At the time of the EPA source testing, the No. 6 (Parsons) plant
was in the second month of its 12-month catalyst cleaning cycle, and
the No. 7 (Monsanto) plant was in the twelfth and last month of its
catalyst cleaning cycle. A statistical analysis of the emissions data
4-14
-------
TABLE 4-1. ESTIMATED MAXIMUM IMPURITY LIMITS FOR METALLURGICAL
OFFGASES USED TO MANUFACTURE SULFURIC ACID12
Approximate limits (Mg/Nm3)3
Substance
Chlorides, as Cl
Fluorides, as F
Arsenic, as As203
Lead, as Pb
Mercury, as Hg
Selenium, as Se
Total solids
H2S04 mist, as 100% acid
Water, as H20
Acid plant inlet
1.2
0.25
1.2e
1.2
0.25
50.0
1.2
50.0
-
Gas purification
system inlet
125C
25d
200
200
2.5f
100
l.OOO9
-
400,000
aBasis: dry offgas stream containing 7 percent S02.
bFor a typical gas purification system with prior coarse dust removal.
°Must be reduced to 6.0 if stainless steel is used.
dCan be increased to 500 if silica products in scrubbing towers are
replaced by carbon; must be reduced if stainless steel is used.
eCan be objectionable in product acid.
fCan be increased to between 5 and 12 if lead ducts and precipitator bottoms
are not used.
9Can usually be increased to between 5,000 and 10,000 if weak-acid settling
tanks are used.
4-15
-------
shows that the 30-percent greater average emissions of the No. 7
plant, compared to the average emissions of the No. 6 plant, are
statistically significant at the 90-percent probability level. This
emissions difference reflects not only catalyst deterioration, but
also design or construction differences between the Parsons 1967 acid
plant technology and Monsanto 1970 acid plant technology. However,
the major portion of this difference is assumed to result from catalyst
deterioration.4
Although there are no analogous data available for dual-stage
plants, metallurgical acid plant vendors have agreed that the EPA
estimate of a 30-percent increase in S02 emission concentrations as
the upper limit for deterioration of catalyst performance between
catalyst screenings for single-stage acid plants is also a reasonable
estimate for dual-stage acid plants.5 In most cases, the frequency of
catalyst screenings is primarily a function of pressure drop rather
than conversion efficiency. Normally, the first catalyst bed is
constructed with a depth approximately 50 percent greater than the
theoretical design depth to compensate for the anticipated decrease in
conversion efficiency as the catalyst becomes partially plugged and
the pressure drop increases between catalyst beds.4 Catalysts are
guaranteed for various periods, although longer guarantees require the
use of more catalyst for a larger conversion efficiency over time.
Screening periods generally vary from 1 to 2 years, depending upon
blower capacity and the participate collection efficiency of the gas
purification system.4
Analysis of the previously mentioned EPA source tests (full text
in Appendix H) at Kennecott Garfield showed S02 emissions during
normal operations of less than 2,000 ppm when averaged for long periods,
such as a week.13 "Normal" operations were defined by analyzing acid
plant operating logs to ascertain when upset conditions, i.e., malfunc-
tions, startups, shutdowns, etc., were in evidence. Significantly,
the long-term average S02 concentration was considerably less than the
emission concentration (2,700 ppm) corresponding to the vendor guarantee
of 95 percent conversion at an inlet concentration of 5 percent S02.
As mentioned previously, metallurgical sulfuric acid plants
operate at maximum efficiency when the feed stream is uniform in flow
4-16
-------
rate and composition. However, offgas streams from some copper smelter
operations, most notably copper converting, exhibit large fluctuations
in both volumetric flow rate and S02 concentration. While fluctuations
in feed-stream flow rate and S02 concentration are widely believed to
adversely affect acid plant performance, few data exist to quantify
the effect on S02 emissions.4 Further analysis of the Kennecott
Garfield data showed that instantaneous S02 emission concentrations
varied greatly (<1,000 ppm to >7,000 ppm), depending upon fluctuations
in the feed-stream S02 content (<1 percent to >7 percent). Specifically,
when the data were averaged over 4-hour periods, 13 data periods were
evident during which the average S02 emissions exceeded the vendor's
guarantee (2,700 ppm).13 Increasing the averaging time to 6 hours
decreased to seven the number of periods that exceeded the reference
emission level (2,700 ppm). Increasing the reference emission level
from 2,700 ppm (the vendor's guarantee) to 3,000 ppm (approximately 10
percent greater) reduced the number of periods exceeding the reference
emission level by approximately 50 percent. Further increases in
either the averaging time or the reference emission level selected for
comparison did not significantly decrease the number of periods that
exceeded the reference concentration. Further analysis of the same
data, based on the actual time during which S02 emissions exceeded the
reference concentration level, led to the same conclusions. Thus,
this analysis, which does not consider catalyst deterioration, shows
that an averaging time of 6 hours and a reference emission level 10 to
20 percent above the commonly accepted vendor/contractor S02 emission
guarantees effectively masks normal, short-term fluctuations in S02
emissions.13
S02 emissions from a dual-absorption sulfuric acid plant operating
on copper converter offgases were monitored by EPA through the use of
a continuous monitoring system. This testing was performed at the
ASARCO, El Paso, Texas, copper/lead smelter from mid-May through
November 1973. The data were validated to ensure their accuracy and
then analyzed by EPA. The analysis showed that 6-hour averages effec-
tively mask the extreme fluctuations that are encountered with copper
4-17
-------
converter offgases. S02 emissions were limited to 250 ppm or less
95 percent of the time. However, the inlet S02 concentration was
rather low, and no catalyst deterioration was detected during the test
period. Based upon readings taken at 3-minute intervals, the inlet
gas stream S02 concentration averaged 3.8 percent S02 for the entire
test period. Further analysis that took into account a possible
10-percent increase in emissions due to catalyst deterioration and
that extrapolated the data to allow for the highest inlet S02 concen-
trations expected from copper, lead, or zinc primary smelting opera-
tions (~9 percent S02) showed that S02 emissions can be limited to 500
ppm or less 95 percent of the time and 650 ppm or less 98.8 percent of
the time. A complete analysis of these data is included in Appendix I.
Acid mist emissions from dual-stage absorption sulfuric acid
plants are normally less than acid mist emissions from single-stage
absorption plants because the mist loading of the second absorption
tower is usually less.4 However, an acid mist eliminator must be
installed following the first absorption tower in dual-stage acid
plants to protect downstream equipment from corrosion. Absorption
towers have inherent lags and are sensitive to many variables, in-
cluding inlet S03 concentration, absorbing acid strength, temperature,
and gas stream flow rate. However, currently available control tech-
nology is adequate to restrict acid mist emissions to low opacity
wisps, except for infrequent upsets.4 Such upsets are caused when the
absorbing acid concentration becomes greater than the azeotropic, thus
allowing S03 to remain unabsorbed in the gas phase and to eventually
create a visible acid mist plume as it combines with water after
leaving the stack. When an azeotropic condition exists, there is no
"driving force" for mass transfer, i.e., the transfer of S03 into the
sulfuric acid; thus, the S03 remains in the gas phase. Of course,
mist eliminators are particulate collection devices and thus cannot
prevent acid mist emissions that occur during this type of upset
condition because the mist is produced by S03 combination with water
after the gas stream exits the mist eliminator.
4-18
-------
Manufacturers of mist eliminators commonly guarantee collection
efficiencies of 99 percent or greater.14 However, these guarantees
are generally based only on the total weight of entrained particulate
removed from the gas stream. When guaranteed collection efficiencies
are restated as percentages of particles collected by size rather than
by weight alone, they generally decrease substantially. Therefore,
when a mist elimination device is chosen for a particular application,
collection efficiency by particle size and weight should be considered.
High-efficiency mist eliminators are available that provide high
collection efficiencies (by size and weight) over a large particle
size range. Monsanto1s E-S Type Mist Eliminator guarantees collection
efficiencies (by size and weight) of at least 99 percent over a particle
size range of 0.1 pm to 1.0 urn.14 Guarantees of this magnitude limit
maximum stack emissions to 35 mg/m3 (0.015 gr Ib/scf) or less.4 Lower
efficiency models generally ensure emissions of 70 mg/m3 (0.031 gr
Ib/scf) or less. Under the worst conditions, the 70 mg/m3 emission
value can represent a 20-percent opaque plume, but normally the emissions
from a high-efficiency mist eliminator are less than 10 percent opaque.4
The use of sulfuric acid plants to control S02 emissions from
primary copper smelters is a well-demonstrated technology. Currently,
12 of the 15 active domestic primary copper smelters produce sulfuric
acid from process offgases. Of these 12 facilities, 6 use dual-stage
absorption plants, and 10 produce acid from roaster and/or converter
offgases. The Phelps Dodge facility at Playas, New Mexico, uses a
dual-stage absorption plant to produce acid from gases that originate
in an Outokumpu flash smelting furnace; at the Inspiration Consolidated
copper smelter, offgases from an electric smelting furnace are processed
in a dual-stage absorption plant to manufacture acid.
EPA source tests such as those described earlier indicate that
the S02 conversion efficiency of metallurgical sulfuric plants depends
upon the frequency and magnitude of fluctuations in gas stream flow
rate and S02 concentration, as well as catalyst deterioration. Thus,
an S02 emission limitation that reflects these factors is appropriate
for primary copper smelters. Based upon analyses performed to determine
4-19
-------
the effect of fluctuations in gas stream flow rate and S02 concentration,
a reference emission level of 20 percent higher than that guaranteed
by the vendor/contractor and a 6-hour averaging time will effectively
mask normal, short-term fluctuations in S02 emissions. As noted
previously, the upper limit for increased S02 emissions due to catalyst
deterioration between screenings has been established at 30 percent.
However, noting that no catalyst deterioration was detected over a
SV-month testing period at ASARCO's dual-stage absorption plant in
El Paso, a 10-percent increase in emissions due to catalyst deteriora-
tion has been deemed reasonable for a dual-stage absorption plant over
the year between screening operations. Consequently, an emissions
limitation based upon a 6-hour average S02 emissions concentration
30 percent in excess of that guaranteed by the vendor will reflect
variations in gas stream flow rate and S02 concentration, as well as
the catalyst deterioration that occurs between annual catalyst screenings.
4.3 SCRUBBING SYSTEMS
4.3.1 Background
Historically, there has been little economic incentive to desul-
furize process offgases containing S02 in concentrations ranging from
0.05 to 3.5 percent. This category includes offgases generated by
reverberatory smelting furnaces, as well as offgases generated by
fossil-fuel-fired steam generatars, refinery sulfur recovery plants,
sulfuric acid plants, and lead sinter machines. Prior to the last
decade, the control of S02 emissions from these sources was prac-
tically nonexistent. However, techniques for removing S02 from "weak"
S02 streams have received a great deal of attention over the last 10
to 15 years. Consequently, numerous methods have been devised to
remove S02 from weak streams.
The approach most commonly used for weak-stream S02 control has
involved the use of scrubbing systems that chemically react the S02
with liquid phase absorbents to yield sulfur compounds that can be
either discarded, reprocessed, or sold for direct use in other
industries. The term "scrubbing system" is commonly used to describe
such systems, since a scrubber (absorber) is used to effect the
required pollutant mass transfer.
4-20
-------
The three major types of scrubbing systems can be summarized as
follows:
Noncvclic systems-These open-loop systems generate a throw-
away product.The scrubbing liquor makes only one pass
through the scrubber.
Cyclic nonregenerative systems-These closed-loop systems
produce a sulfur-containing compound that is either discarded
or sold. As much of the scrubbing liquor as possible is
recovered and recycled through the scrubber.
Cyclic regenerative systems—These closed-loop systems
recover S02 and have a relatively small waste product for
disposal. The absorbent is regenerated and the scrubbing
liquor is recycled through the scrubber.
For applications to weak streams from reverberatory furnaces,
consideration must be given to the fact that reverberatory furnace
effluents contain a wide variety of contaminants in addition to S02.
The presence of high concentrations of oxygen (relative to 02 concen-
trations commonly encountered in gases generated by fossil-fuel-fired
generating plants), particulates, acid gases, metallic fumes and high
gas temperatures must be carefully considered, especially where cyclic
absorption systems are being considered. In most scrubbing systems,
some type of offgas conditioning would be required prior to the absorp-
tion of S02 in the scrubbing media.
Over the past decade, several scrubbing schemes using various
absorbents have been applied to metallurgical weak streams on either a
pilot- or full-scale basis. Two types of scrubbing systems have been
applied to reverberatory furnace offgases on a full-scale basis. One
is a calcium-based system that uses a lime/limestone slurry as the
scrubbing medium, and the other is a magnesium-based system that uses
a magnesium oxide slurry as the scrubbing medium. The calcium-based
system is a cyclic nonregenerative type; the magnesium-based system is
a cyclic regenerative type. Both systems are located at the Onahama
smelter in Japan. Other scrubbing systems applied to metallurgical
weak streams on either a pilot- or full-scale basis are citrate-type
systems and ammonia-based systems. Discussions of the four above-
mentioned types of scrubbing systems are presented in the following
sections.
4-21
-------
4.3.2 Calcium-Based Scrubbing Systems
4.3.2.1 Summary. Calcium-based scrubbing systems have been
demonstrated for the control of S02 emissions from fossil-fuel-fired
power plants since the 1930's. During the last 10 years, applications
to metallurgical offgas streams have also been demonstrated. In the
United States, a calcium-based scrubbing system has been controlling a
weak S02 stream (<0.6 percent) from a molybdenum ore roaster since
1972.4 During the period 1977-1980, this system maintained a sulfur
removal efficiency in excess of 96 percent. Operation of this
particular system at the Duval Corporation near Tucson, Arizona, has
also demonstrated the feasibility of operating a calcium-based system
in an area where water is scarce.
Perhaps the most significant application of calcium-based scrubbing
technology to occur on metallurgical processes over the past few years
has been the system installed at the Onahama smelter in Japan in late
1972. This system controls a portion of a weak S02 stream that orig-
inates from three green-charged reverberatory furnaces. Both system
reliability and S02 removal efficiency are reported to be high.16
It has been demonstrated for some time that calcium-based
scrubbing systems are viable control methods for gas streams with low
S02 concentrations (500 to 5,000 ppm). However, the lime/limestone
system at Onahama was the first system to fully demonstrate control of a
weak S02 stream from reverberatory furnaces. Experience at Onahama has
also demonstrated that meticulous furnace control can maintain a
relatively steady S02 concentration in the weak stream.6 16 Also,
because the S02 removal efficiency of these systems increases as the
gas-stream S02 concentration decreases, effective (>90 percent) removal
of S02 from offgases generated by calcine-charged reverberatory furnaces
should be possible. However, to handle the fluctuations in gas stream
S02 concentration inherent in calcine-charged furnace operation, the
scrubbing system will have to include a well-designed process control
system capable of reacting to sudden changes in the gas-stream S02
concentration. Considering the range of S02 concentrations encountered
in calcine-charged operations (0.4 to 2.0 percent S02), providing the
adequate process control should be quite feasible.
4-22
-------
4.3.2.2 General Discussion. Calcium-based scrubbing systems may
be of two types: noncyclic or cyclic-nonregenerative. In the noncyclic
system, the absorbent passes through the scrubber on a once-through
basis. Early work on this type of system was conducted by the London
Power Company in the 1930's;4 17 alkaline Thames River water provided
the absorbent. The system removed a high percentage of the S02 (about
90 percent), but the acidic effluent from the process lowered the pH
of the Thames to an undesirably low level. The noncyclic system has
inherent water pollution problems in some situations that would preclude
its use on a large scale. This system also has the drawbacks of
requiring a very large amount of water and of cooling the gas to an
unduly low temperature.
Also in the 1930's, technology was developed on cyclic-nonregen-
erative scrubbing systems that employed calcium-based absorbents. The
two most widely accepted cyclic-nonregenerative, calcium-based scrubbing
systems employ either calcium carbonate (limestone) or calcium hydroxide
(slaked lime) as the absorbent. Simplified flow diagrams for the
lime/limestone slurry scrubbing processes are presented in Figure 4-2.
The principal process steps are as follows:
S02 absorption
Demisting
Liquor loop operation
Lime/limestone handling operations.
S02 absorption occurs when the S02-laden process offgases are vented
to a scrubber, where they are scrubbed countercurrently with the
absorbent slurry.
The chemistry of the absorption step is quite complicated. As
many as 28 chemical equations have been postulated by some authorities
to characterize the reactions involved. The following rather simple
mechanism is thought to be representative of the reactions that occur:5
4-23
-------
I. Scrubber Addition of Limestone
Cleaned Gas Stream
Effluent Gas
Stream
U
CO
CaCO.
Pump Tank
II. Scrubber Addition of Lime
Cleaned Gas Stream
Effluent Gas Stream
CaCO,
§
CO
Ca(OH),
Calciner
CaO
Pump Tank
Settler
Solid Wastes
(CaSO3 and CaSO4>
Settler
Solid Wastes
(CaSO3 and CaSO4)
Figure 4-2. Calcium-based scrubbing processes.
4-24
-------
S02 + H20 + H+
H+ + CaC03 j Ca+2
Ca+2 + HSOs + *sH20 + CaS03 • J§H20 + H+
H+ + HC03 + H20 + C02 .
Some of the calcium sulfite formed will oxidize to form calcium sulfate
as follows:
CaS03 + hQ2 -> CaS04 .
The reactions that take place within the absorbers are heterogeneous
in nature because they involve the gas, liquid, and solid phases
present. The scrubbed gases are then trapped in a chevron mist elimina-
tor, which is washed with clear water to prevent the escape of acidic
droplets to the atmosphere. The resultant gas stream is then vented
to the atmosphere. Liquor loop operation involves splitting the
S02-laden slurry, with a portion going to the pump tank and a portion
going to the settler. If the limestone scrubbing process is used,
calcium carbonate is added to the pump tank as makeup; if the lime
scrubbing process is used, calcium oxide is added as makeup. Effluent
from the pump tank and settler is recycled to the scrubber in both
types of processes. In most cases, solids in the pregnant slurry
(calcium sulfite [CaS03] and gypsum [CaS04]) are removed in the settler
and pass to disposal. It is possible, however, to oxidize the slurry
to create a solid product that consists almost entirely of gypsum.
This is normally done when the gypsum can be sold, as is the case at
the Onahama smelter in Japan. Limestone-handling operations consist
of field storage and transfer of mined limestone to a milling and
sizing plant for preparation. If the lime scrubbing process is to be
used, the limestone must be calcined onsite to calcium oxide, or lime
must be purchased directly instead of limestone.
Limestone is the absorbent chosen in most cases. Although not as
reactive as lime, it is less costly. In some areas the cost of
lime as CaO is twice or more than that of limestone.18 For applications
in the utilities industry, where high-sulfur coals are involved, the
4-25
-------
cost differential between the two absorbents becomes such a major item
that it usually outweighs the advantages commonly associated with
lime, namely slightly lower capital costs, higher S02 removal effi-
ciency, a lower liquid-to-gas flow (L/G) requirement, and probable
higher reliability.18 In the western part of the country, however,
where most primary copper smelters are located, relatively long shipping
distances reduce the cost differential, because limestone (CaC03)
weighs nearly twice as much as lime (CaO). Thus, for potential applica-
tions to smelters in the western United States, an economic analysis
would determine the most cost-effective absorbent.
4.3.2.3 Design and Operating Considerations. Several factors
may significantly affect the S02 removal efficiency of limestone
scrubbing systems. These may be summarized as follows:
Design of the scrubber proper
Type and size of the limestone used
Inlet gas temperature
pH of the absorbent slurry
Solids content of the absorbent slurry.
The design of the scrubber is critical to the limestone scrubbing
process. S02 removal efficiency must be maximized by improving the
S02 gas-phase mass transfer rate at the liquid interface, which is
dependent upon the scrubber type selected and its operating parameters.
L/G is important because of its effect of reducing the gas-phase
resistance to the mass transfer of S02. S02 removal efficiency is
favored by high L/G and low inlet S02 concentration.4 5 18 As the S02
concentration in the feed gases increases, removal efficiency will
decrease. A fairly high ratio, about 6,700 2/1,000 Nm3 (50 gal/1,000
scf), has been necessary in most power plant applications to achieve a
high degree of S02 removal (>80 percent). Generally speaking, the
higher the S02 concentration in the feed gases, the higher the L/G
will have to be to maintain an acceptable S02 removal efficiency.
Thus, application to reverberatory furnace offgases will require
higher L/G's than those in evidence at utility-related installations.
4-26
-------
The type and particle size of the limestone used as the absorbent
will affect S02 removal efficiency as well as determine how efficiently
the absorbent is used. The ability of carbonate stones to chemisorb
S02 varies greatly. For example, calcite (CaC03) stones have been
shown to be superior to dolomite (MgC03) stones as far a S02 removal
efficiency is concerned.4 Calcite stones maintain a consistently high
S02 removal efficiency until nearly exhausted. The degree to which
the limestone is ground will affect the S02 absorption capacity of the
absorbent slurry. Personnel at the Onahama smelter report that lime-
stone shows nearly as good an S02 absorption capacity as slaked lime,
Ca(OH)2, if it is ground to minus 325 mesh.16 Particle size does not
appear to be critical when lime is used, apparently because the particle
size of most slaked limes is inherently small.4
Studies conducted to determine the effect of gas stream temperature
on S02-removal efficiency indicate that S02 removal efficiency decreases
linearly as the temperature of the feed gas stream increases.4 The
equilibrium considerations that govern the degree of S02 absorption
depend upon the partial pressure of S02 in the inlet gas stream, which
in turn depends upon the inlet gas stream temperature. The partial
pressure of S02 increases by about 18 percent for a temperature increase
of 5.6° C (10° F). If the partial pressure of S02 is allowed to
increase too much, S02 removal efficiency will eventually become zero,
and the inlet gas stream will actually begin to strip S02 from the
absorbent slurry. Thus, in some cases, precooling of the inlet gases
may be necessary.
As mentioned above, the degree and rate of S02 absorption depend
upon the difference between the partial pressure of S02 in the gas
stream and the vapor pressure of S02 above the absorbent slurry. The
pH of the absorbent slurry has a distinct effect on the vapor pressure
of S02 above the slurry, and thus pH can affect the degree of S02
removal. The vapor pressure of S02 above the absorbent slurry must be
kept quite low to provide a driving force for absorption, especially
if the S02 concentration in the feed gases is low, as in reverberatory
4-27
-------
furnace offgases. The effect of slurry pH on the S02 equilibrium
vapor pressure, PS02, is shown in Figure 4-3. As indicated, the S02
vapor pressure is highest at low pH. Control of pH is normally accom-
plished by adding makeup lime/Iimestone slurry at the pump tank.6 17
High slurry solids loadings provide improved rates of solubility
for calcium, thus providing more effective replenishment of the calcium
ion.4 The most efficient S02 removal has consistently been obtained
with slurry solids loadings of 12 to 15 percent.
4.3.2.4 Operational Problems. Considerable progress has been
made in reducing the frequency of operational problems that have
tended to reduce the reliability of calcium-based scrubbing systems in
the past. Currently, the primary cause of forced outages is mechanical
failure of pumps and other operating equipment, while corrosion and
erosion are probably the next most significant causes.18 Chemical
difficulties, e.g., scaling, related to the complex system chemistry
have also been a source of trouble.
Because of the high L/G requirement, very large pumps are commonly
used in calcium-based scrubbing systems. Failures may occur due to
the large stresses that must be placed on the pumps. The most common
types of failure involve rotary parts and rubber linings.18 An impor-
tant aid in maintaining the best possible reliability is a well-planned
program of preventive maintenance. Also, spares should be provided
wherever feasible, especially for pumps. Simplicity of design is
important in areas where frequent maintenance is likely. For this
reason, simple spray-type scruboers are preferred even though they are
less efficient than packed- or tray-type absorbers in effecting mass
transfer.18
Lime/limestone scrubber slurry is both corrosive and erosive in
nature. As a result, some system components may require frequent
maintenance and/or replacement. Stainless steel is required to prevent
corrosion where metal contacts acidic solutions.5 While a high slurry
solids loading tends to enhance S02 absorption, it may also increase
the rate of erosion within process equipment. Erosion is reduced by
applying elastomer linings in pumps, piping, and other surfaces that
are subjected to heavy slurry impingement.
4-28
-------
Figure 4-3. Effect of pH of calcium sulfite-bisulfite
solution on SO2 equilibrium vapor pressure.17
4-29
-------
Solids deposition and the resultant plugging of equipment have
been major problems in the operation of calcium-based scrubbing systems.
These problems may be especially troublesome in limestone scrubbing
systems. The deposition can occur at several locations within the
system, namely, at wet/dry interfaces, on scrubber surfaces, in scrubber
packing, and in the mist eliminator.18 Generally, utilities do not
analyze the deposited solids; however, solids formed in the absorbers
at the Onahama smelter have been analyzed.16 The results of these
analyses are presented in Table 4-2.
Careful control of the slurry pH has been identified as the most
critical factor in the prevention of scaling.6 16 Operating experience
at the Onahama smelter has shown that the best operating pH is slightly
above 4.16 If the pH is less than 4, oxidation of calcium sulfite
occurs in the scrubbers, resulting in calcium sulfate scale formation.
Also, at low pH values, S02 removal efficiency may be adversely affected.
If the slurry pH is appreciably above 4, calcium sulfite will become
insoluble, resulting in calcium sulfite scale formation. A rapid
decrease in pH caused by S02 absorption followed by a rapid increase
in pH due to limestone addition can cause sulfite precipitation on the
limestone slurry particles and on the equipment surfaces. This will
result in blinding of the reactive surfaces and thus lead to inefficient
absorbent use. Consequently, where reverberatory offgases are involved,
it is desirable to maintain the slurry pH just above 4 without large
fluctuations.
The inlet gas stream temperature also has an effect on scaling.
The evaporation of water from the slurry due to high inlet gas stream
temperatures will create wet-dry interfaces at which scaling tends to
occur.4
Prevention of scaling is essential in maintaining stable operation
and high S02 removal efficiency. Careful pH control coupled with
induced desupersaturation of the slurry at a point outside of the
scrubbing circuit have been shown to be extremely effective means of
minimizing scaling.16
4-30
-------
TABLE 4-2. COMPOSITION OF SCALE FROM THE ONAHAMA
LIME-GYPSUM3 PROCESS16
(Percent)
Sample origin CaO CaS04 CaS03
Lower zone of No. 1 absorber 33.0 43.8 2.6
Upper zone of No. 2 absorber 34.0 40.6 5.7
Upper zone of No. 2 absorber 34.9 35.7 8.3
Lower zone of No. 2 absorber 36.6 29.3 17.0
aLime was the absorbent used in this process at the time that these
analyses were performed.
Repetitive samples from the same origin.
Note: CaO, CaS04, and CaS03 were not the only compounds present in
the scale; thus the individual analyses do not add to 100 percent.
4-31
-------
4.3.2.5 Survey of Operating Experience
4.3.2.5.1 Domestic utility-related applications. The U.S.
utility industry has tried many FGD approaches over the past decade.
The relative simplicity and lower costs of limes tone-based scrubbing
systems have made these systems the most popular.19 Table 4-3 lists
the major domestic utility-related FGD installations that employ the
limestone scrubbing process. The technology currently preferred by
the U.S. utility industry has the following features:18
Simplicity of scrubber design. Single-stage spray scrubbers
with a minimum of interior parts are the most frequently
used type.
Low-pressure drop to conserve energy. This is another
reason for selecting spray scrubbers.
High L/G to get adequate mass transfer and to avoid scaling.
Direct entry of hot gas into the scrubber, with provision to
prevent splashback into the inlet duct to avoid wet/dry
interface deposition.
Adequate retention time in the slurry recirculation tank to
allow time for the limestone to neutralize sulfurous acid
picked up in the scrubber and for dissipation of the sulfate
and sulfite supersaturation developed during passage through
the scrubber.
High slurry solids content (12 to 15 percent) in the scrubber
loop to provide seed crystals on which dissolved sulfite and
sulfate can precipitate.
High use of limestone to prevent scaling in the mist elim-
inator. This requires a fairly low pH (4.0 to 5.5) in the
circulation tank, which has an adverse effect on S02-removal
efficiency. To offset this, means to increase the rate of
mass transfer are useo such as very high L/G, staged scrubbing,
or chemical additives.
Use of simple chevron-type mist eliminators mounted in a
horizontal position in the top of the scrubber and washed
intermittently with jets of fresh water at relatively high
pressure.
Commonwealth Edison's Will County No. 1 station, which started up
in February 1972, consists of two identical parallel wet scrubbing
systems.4 Each system consists of a venturi for particulate removal,
4-32
-------
TABLE 4-3 MAJOR DOMESTIC UTILITY-RELATED FGD INSTALLATIONS
THAT USE THE LIMESTONE-SCRUBBING PROCESS.19
Station/unit,
Power company
Cholla No. 1
Arizona Public Service
Duck Creek No. 1A
Central 111 inois Light
La Cygne No. 1
Kansas City Power & Light
Lawrence No. 4
Kansas Power & Light
Lawrence No. 5
Kansas Power & Light
Martin Lake No. 1
Texas Utilities
•** Sherhurne No. 1
do Northern States Power Company
CO
Sherburne No. 2
Northern States Power Company
Southwest No. 1
Springfield City Utilities
Widows Creek No. 8
Tennessee Valley Authority
Will County No. 1
Commonwealth Edison
Winyah No. 2
South Carolina Public Service
Size
(MW)
115
400
820
125
400
793
710
680
200
550
167
280
Startup
date
10/73
8/78a
2/73
12/68
11/71
10/77
3/76
9/77
4/77
5/77
2/72
7/77
New or
retrofit
Retrofit
New
New
Retrofit
New
New
New
New
New
Retrofit
Retrofit
New
Percent
sulfur in
coal
0.4-1
2.5-3
5.0
0.5
0.5
1.0
0.8
0.8
3.5
3.7
4.0
1.0
Design
S02 removal
efficiency
(percent)
58.5 overall
92 in the absorber
75
76
76
65
60
50
50
80
80
82
70
aOne module operated from September 1976 to April 1977.
Note: This table presents domestic installations as they existed in 1978.
-------
followed in series by a turbulent contact absorber (TCA) for S02
absorption. The S02 control system is guaranteed to achieve 80 to 85
percent S02 removal.4 19 This removal efficiency has been achieved;
however, operational problems have prevented continuous operation in
the past.4 These problems included demister plugging, nozzle plugging
by construction debris, power loss to the pond reclaim pumps, vibration,
loosened screens in the pump and in the recirculation tank, reheater
plugging, failure of expansion joints, and breakage of the paddle on
the slurry tank mixer. However, only the demister plugging problem
proved to be chronic, and the solution to the problem involved rede-
signing the demister washers. Scaling has not been a serious problem
with the system.
Kansas City Power and Light's La Cygne Unit No. 1, which began
operation in February 1973, has proved to be one of the most reliable
large domestic utility FGD systems.19 The system was designed to
achieve a 76-percent S02 removal efficiency. The La Cygne unit was
plagued with numerous startup problems, most of which were not due to
FGD system operation. However, despite the problems at startup, the
availability of the system improved steadily. Availability increased
from approximately 76 percent in 1974 to about 93 percent in 1977.
The actual S02 removal efficiency has varied from 70 "to 83 percent.19
The Northern States Power Company Sherburne County Station No. I
and No. 2 units have demonstrated excellent reliability. Availability
for Unit No. 1, which began operation in March 1976, averaged 85
percent during the first 4 months after startup.19 Reliability over
the subsequent 12-month period was in excess of 90 percent. Unit No.
2 demonstrated availabilities in excess of 95 percent during its first
4 months of operation. Both units were designed to achieve 50 percent
S02 removal. Actual S02 removal efficiencies have been in the 50 to
55 percent range.
Based on the operating experience of domestic utility-related
limestone scrubbing systems, sufficient evidence exists to suggest
these systems can operate at their design S02 removal efficiencies
while simultaneously maintaining a high degree of reliability,
generally in excess of 90 percent.19 As suggested by Table 4-3, the
4-34
-------
majority of the full-scale utility-related limestone FGD systems in
place are not designed to achieve overall S02 removal efficiencies in
excess of 90 percent. However, no evidence exists to suggest that S02
removal efficiencies of 90 percent or greater could not be achieved
with lime/limestone systems. In fact, two limestone FGD's have achieved
S02 removal efficiencies of 90 percent or greater during demonstration
runs, while seven lime FGD's have achieved S02 removal efficiencies of
90 percent or greater. These units are summarized in Table 4-4.
In summary, the domestic utility-related experience indicates
that both high S02 removal efficiencies and high system reliabilities
are achievable. The critical factors behind the successful operation
of these systems are proper system design and maintenance.
4.3.2.5.2 Applications to metallurgical offgases. Lime/limestone
scrubbing technology has been applied to metallurgical offgas streams
at both foreign and domestic facilities. The Duval Corporation,
located near Tucson, Arizona, is operating two four-stage model 500
TCA's to remove S02 from offgases generated in molybdenum sulfide
roasters.4 The units, designed by UOP, use lime slurry as the absorb-
ent and are rated at 1,420 NmVmin (50,000 scfm) each. The system
began operation in July 1971 and experienced extensive problems with
scaling and plugging. However, these problems have been overcome and
the system is reported to be working well.18 This system processes
offgases containing 0.35 to 0.75 percent S02 and generates an offgas
stream with an S02 concentration of less than 200 ppm.6 S02 removal
efficiencies are commonly in excess of 96 percent. Emissions test
data for this system are presented in Table 4-5. The successful
operation of this system demonstrates that scrubbing systems using
water recycle can be successfully operated in areas where water is
scarce.
In late 1972, the Onahama Smelting and Refining Company, Ltd.,
installed a lime-gypsum plant at its Onahama smelter for fixation of
S02 in green-charged reverberatory furnace offgases. This facility
was the first commercial-scale plant of this type in Japan designed to
treat smelter offgases containing up to 3 percent S02.16 A flow chart
of this system is presented in Figure 4-4. At the Onahama smelter,
4-35
-------
I
CO
en
TABLE 4-4. LIME/LIMESTONE FGD SYSTEMS THAT HAVE ACHIEVED S02 REMOVAL EFFICIENCIES OF 90 PERCENT
OR GREATER ON OFFGASES GENERATED BY COAL-FIRED STEAM GENERATORS19
Uti 1 i ty company
Arizona Public Service
Duquesne Light
Louisville Gas and Electric
Southern California Edison
U.S. Air Force
Kentucky u'ti 1 i Lies
Tennessee Valley Authority
Station
Cholla
Four Corners
Phillips
Cane Run
Paddy's Run
Mohave
Mohave
Rickenbacker3
Green River
Shawnee
Unit
number
1
5
1-6
4
6
1
2
1-9
1-3
10
Size
(MW)
115
160
110
175
65
170
170
20
64
10
Nature of
appl ication
Full-scale
Demonstration
Ful 1-scale
Ful 1-scale
Demonstration
Demonstration
Demonstration
Ful 1-scale
Full-scale
Prototype
Type
of process
Limestone
Lime
Lime
Lime
Lime
Limestone
Lime
Lime
Lime
Lime/Limestone
S02
removal
achieved
(%)
92
95
90+
90
99.5
95
95
99
yu+
95-99
Denotes a military base.
Note: Data in this table reflect the domestic FGD situation in 1977.
-------
TABLE 4-5. SUMMARY OF EMISSION TEST DATA FOR THE DUVAL SIERRITA
LIME SCRUBBING SYSTEM, 1977-1980
Source
Year
1977
1978
1979
1980
Molybdemum roaster
throughput rate (tons/h)
Particulate matter
emission rate (Ib/h)
Allowable particulate matter
emission rate (Ib/h)
3.50
3.83
7.81
3.55
3.50
7.87
3.65
4.01
11.92
3.50
4.01
7.81
S02 emission rate (Ib/h)
H2S04 emission rate (Ib/h)
Total sulfur emission
rate (Ib/h)
Sulfur removal
efficiency (percent)
11.00
2.31
6.26
99.80
52.60
9.03
61.63
97.80
25.38
1.91
27.29
99.00
88.84
0.93
89.77
96.50
aArizona State regulations require removal of 90 percent of the sulfur
that enters the process as a feed.
4-37
-------
No. 1 No. 2
ABSORBER ABSORBER
REVERBERATORY
FURNACE
MIST MIST
ELIMINATOR COTTRELL
TO OCEAN -«
TO WATER
STOCKYARD CENTRIFUGE
Figure 4-4. Flow diagram of the lime/gypsum plant at the Onahama smelter.16
-------
converter gases and a portion of the reverberatory furnace offgases
are routed to two acid plants, where approximately 30 Gg (33,000 tons)
of 98 percent sulfuric acid are produced per month. The remainder of
the reverberatory offgases serve as the feedstock for the gypsum
plant. The feed-stream flow rate varies from 1,800 to 2,400 dry
mVmin (63,550 to 84,750 dry ftVmin) at standard conditions.20
Initially, only quick lime was used as absorbent. Soon after startup,
however, a serious absorber scaling problem evolved. Scaling problems
were eventually solved by improving certain process characteristics
and operating conditions.6 16 One important improvement involved the
substitution of limestone for a portion of the more expensive quick
lime. To minimize scaling, the Onahama system presently uses two
stages of absorption in conjunction with the use of seed crystals.
Currently, system reliability is high (-99.3 percent) while an
S02 absorption efficiency of 99.5 percent is maintained. The average
S02 emission concentration from the plant is in the 40- to 60-ppm
range.6 Good-quality gypsum suitable for board or cement is produced
by the subsequent oxidation of the calcium sulfite produced.16
In addition to the full-scale applications discussed above, some
pilot-scale tests were conducted by the Smelter Control Research
Association, Inc. (SCRA), at the Kennecott smelter located in McGill,
Nevada. These tests were performed on reverberatory furnace offgases
to investigate the effects of parameters such as L/G and feed-stream
S02 concentration on S02 removal efficiency. Results indicated that
S02 removal efficiency was highest at low S02 inlet concentrations and
high L/G ratios.5
4.3.2.6 Applicability to Reverberatory Smelting Furnaces.
4.3.2.6.1 Transfer of utility-related scrubbing technology. To
assess the feasibility of applying utility-related scrubbing technology
to weak S02 streams generated by reverberatory smelting furnaces, feed
stream characteristics must be compared. There are several differences
between the waste gases from power plants and smelters in regards to
composition, flow rate, and other characteristics.18
4-39
-------
The average S02 content of reverberatory furnace offgas streams
would be higher than the average S02 concentration in power plant
effluents. The concentration of S02 in reverberatory furnace offgases
is typically in the range of 0.5 to 2.5 percent.4 Generally, 1.5
percent is taken to be the average value, which is about four times
the highest level found in the flue gas from coal-fired boilers.18
Utility-related scrubbing experience,18 along with pilot studies
performed on reverberatory furnace offgases,5 indicate that operation
at high S02 removal efficiencies becomes progressively more difficult
as the feed-stream S02 concentration increases. Problems arise at
higher S02 concentrations because the amount of sulfite and sulfate
formed per scrubbing cycle, the "make-per-pass," becomes higher because
L/G is almost never increased in proportion to the inlet S02 concen-
tration.18 Thus, a single-stage scrubbing system would soon be loaded
with sulfite and attendant scaling.5 At a feed-stream composition of
1.5 percent S02, achieving the required S02 removal is difficult
without exceeding the allowable make-per-pass, which would result in
higher than desired levels of supersaturation that would in turn cause
scaling problems. An extremely high L/G would be required to keep the
make-per-pass within acceptable limits; however, at such a high L/G,
flooding or excessive entrainment would occur.18 Therefore, multistaged
scrubbing for reverberatory furnace effluent desulfurization is indicated,
with each stage having its own reaction tank.5 18 This was confirmed
by the SCRA studies cited previously. At Onahama, two absorbers are
employed, each removing about 50 percent of the S02 contained in the
feed stream.6 Several general conclusions were developed from the
SCRA pilot studies conducted at the Kennecott-McGill smelter. Analysis
of the data obtained led to the following conclusions, assuming that
90 percent S02 removal from a reverberatory gas stream containing 1
percent S02 was desired:5
A venturi followed by several stages of absorption would be
required.
A high L/G would have to be maintained,
4-40
-------
A high-quality limestone ground to a moderate fineness would
be required.
The feed-stream S02 concentration would have to be stabilized
as much as possible.
In contrast to utility boiler operations that produce offgas
streams of a relatively steady S02 concentration, a reverberatory
furnace can produce offgas streams of varying S02 concentrations.
Immediately after charging, S02 concentrations may be greater than the
daily average.8 Such S02 concentration surges in reverberatory furnace
offgases can potentially cause serious problems if an adequate process
control system is not specified in the system design (see Section 4.3.6).
This is especially true in the case of calcine-charged reverberatory
furnaces because of their greater variability in S02 strengths. In
the case of green-charged reverberatory furnaces, meticulous furnace
control as practiced at the Onahama Smelter is a demonstrated means by
which a reasonably steady S02 concentration in the offgas stream can be
maintained. Fuel combustion and concentrate charging are instrument
controlled to produce a gas stream with a steady concentration of
S02.6 18 Furnace interiors are monitored by closed-circuit television
at all times so that needed adjustments in furnace operations can be
made almost instantaneously before adverse conditions can affect the
S02 concentration. Assays of slag, silica, mixed ore, and matte can
be obtained within an hour to further indicate and facilitate
adjustments.
Wide variations in feed stream flow rate could also cause problems
in system control; however, the gas flow from a reverberatory furnace
is fairly uniform.4 6 18
Some degree of feed gas pretreatment is necessary to process
reverberatory furnace offgases in calcium-based scrubbing systems.
Gases to be treated by slurry scrubbing for S02 removal would probably
be precleaned in an ESP.5 However, gas pretreatment as part of a
calcium-based scrubbing process would be primarily for cooling and
humidification at approximately 55° C (130° F) to achieve high S02
removal efficiency and to prevent evaporation and scaling in the
scrubber(s). If the solid product is to be of the throwaway type,
4-41
-------
exhaustive paniculate removal is not warranted. Cooling to this
extent is considered to be uneconomical in the utility industry because
the gas stream must be reheated to improve its buoyancy as it exits
the stack. Ordinarily, flue gas at about 150° C (300° F) is fed
directly into the scrubber.18 After undergoing waste heat recovery
and dust removal, reverberator./ furnace offgases generally have a
temperature of about 370° C (700° F). At Onahama, gases are cooled to
approximately 45° C (115° F) prior to entering lime/limestone absorbers.16
Onahama representatives have indicated that this degree of cooling is
necessary to maintain a high S02 removal efficiency.18
Disposal of solid wastes rnay also present a problem in the appli-
cation of calcium-based scrubbing technology to weak S02 gas streams
from reverberatory furnaces. Standard practice for sludge disposal in
the utility industry involves mixing ash with the sludge and then
landfilling the mix.18 Dry lime may also be added to the sludge prior
to disposal.21 Mixing the sludge with ash prevents the sludge from
maintaining a "swampy" consistency for long periods after disposal.
The addition of lime gives the sludge a harder and stronger consistency.
Because the dust content of reverberatory furnace offgases may be up
to 10 times greater than that of power plant effluents, one might
contemplate mixing these dusts with the sludge and then dumping the
dust/sludge mixture into abandoned mines or quarry pits.5 18 However,
the metals content of smelter dusts is always high, prompting operators
to recycle the dusts to recover metal values.
Several sources have reported that sludges composed mainly of
gypsum tend to have better disposal properties than sludges composed
primarily of calcium sulfite.5 B 21 Thus, forced oxidation of the
calcium-sulfite slurry may be a means of improving the disposal prop-
erties of the slurry. When forced oxidation is used, gypsum is formed
via the following reaction:
CaSOa + h 02 -> CaS04 .
At the Onahama smelter, oxidation of the slurry is a principal
feature of the process because gypsum is desired as a salable product.
4-42
-------
The oxidation is carried out in oxidation towers, where the slurry
reacts with oxygen to yield gypsum crystals.6 16
In conclusion, the higher S02 concentration and other differences
in reverberatory gas-stream characteristics do not appear to be serious
obstacles to the transfer of current utility-related lime/limestone
scrubbing technology. The most significant potential problem involves
the capability of the system to respond to fluctuations in the gas-stream
S02 concentration, especially where calcine-charged reverberatory
furnaces are involved. For calcine-charged furnaces, it is likely
that some "overdesign" would have to be incorporated into the FGD
system to ensure that the desired S02 removal efficiency is maintained.
This would consist of providing an excess of mass transfer area (larger
absorbers) in conjunction with excess solvent handling capability. A
well-designed process control scheme that minimizes response lags
would also be desirable (see Section 4.3.6). In the case of green-
charged reverberatory furnaces, careful furnace control in conjunction
with the proper scheduling of smelting operations are proven means by
which to maintain a relatively steady gas-stream S02 concentration.
Thus, the potential process-control-related problems can be minimized
or perhaps even eliminated.
4.3.2.6.2 Assessment of applicability based upon existing
scrubbing systems that process metallurgical effluents. As indicated
previously, the lime/limestone scrubbing system currently in operation
at the Onahama smelter is reported to be operating with high reliability
as well as high S02 removal efficiency.16 18 Therefore, application
of this technology for the control of weak S02 streams generated by
green-charged reverberatory smelting furnaces should be considered
technically demonstrated. The S02 concentration in the effluent from
the green-charged reverberatory furnaces at Onahama varies from 2.5 to
2.8 percent, demonstrating the ability to stabilize the S02 concen-
tration.6 Lowering the S02 concentration by diluting the feed stream
would actually improve the S02 removal efficiency, but gas-handling
equipment would have to be enlarged.6 The expected S02 removal
efficiency on domestic green-charged reverberatory furnaces would
4-43
-------
naturally be the demonstrated efficiency obtained at Onahama,
approximately 99.5 percent.16 Also, because the S02 removal efficiency
of these systems increases as the gas stream S02 concentration decreases,
effective (>90 percent) removal of S02 from offgases generated by
calcine-charged reverberatory furnaces should be possible. 'However,
to handle the fluctuations in gas-stream S02 concentration inherent in
calcine-charged furnace operation, the scrubbing system will have to
include a well-designed process control system capable of reacting to
sudden changes in the gas-stream S02 concentration. Considering the
range of S02 concentrations encountered in calcine-charged operations
(0.4 to 2.0 percent S02), providing the adequate process control
should be feasible.
The work at Duval indicates that the efficiency of a calcium-
based scrubbing system operating with turbulent contact absorbers can
be 92 to 96 percent with very little downtime.6 The system now in
place at Duval has also demonstrated another important point. The
operation of this system in an area where water is scarce has shown
that the proper use of water recycle techniques does allow operation
with high reliability in such areas. This is an important considera-
tion because the bulk of the domestic smelters are located in the
desert-southwest area of the United States.
4.3.3 Ammonia-Based Scrubbing Systems
4.3.3.1 Summary. Ammonia-based scrubbing systems of the Cominco
type (designed by the Consolidated Mining and Smelting Company of
Canada, Ltd.) have been demonstrated for the control of S02 emissions
from Dwight-Lloyd sintering machines, zinc roasters, and sulfuric acid
plant tail gases since the 1930's.4 The successful operation of these
units has demonstrated their ability to maintain high S02 removal
efficiencies while operating on metallurgical feed streams that exhibit
wide variations in both volumetric flow rate and S02 concentration.
S02 removal efficiencies in excess of 90 percent have been achieved
and sustained in these applications at Cominco's facility in Trail,
British Columbia. Cominco-type units can easily maintain high S02
removal efficiencies over the range of S02 concentrations encountered
4-44
-------
in reverberatory smelting furnace effluents.6 Thus, the Cominco-type
process is a technically viable option for the control of S02 emissions
from properly cleaned reverberatory furnace offgases.
Ammonia-based scrubbing systems of Cominco design, which use
sulfuric acid in the acidulation step, produce a strong S02 stream and
ammonium sulfate. If the ammonium sulfate produced is not marketable,
operation of these systems may not be economically justified. In
anticipation of this problem, investigations into an ammonia-based
scrubbing system that uses ammonium bisulfate for acidulation rather
than sulfuric acid have begun.22 This type of system produces ammonia
and ammonium bisulfate via thermal decomposition of the ammonium sul-
fite produced. The ammonia and ammonium bisulfate are then recycled
back to the absorbent makeup and acidulation steps, respectively.
The ammonium-bisulfate acidulation scheme has not achieved large-
scale application; however, a pilot study jointly sponsored by EPA and
the Tennessee Valley authority (TVA) and conducted at TVA's Colbert
Power Plant has provided some data on the operation of this type of
system. Extensive tests have been run on a 3,000-cfm unit to establish
important operating parameters and to develop solutions for process
difficulties. The pilot plant generally operated with S02 removal
efficiencies of 90 percent or higher, with a feed stream concentration
of 0.2 to 0.3 percent S02, and an exit-gas concentration of 200 to
300 ppm S02.5 Despite the extensive work accomplished to date, the
ammonium bisulfate acidulation scheme cannot be considered fully
developed for commercial application; however, this process is fore-
casted to be the S02 removal system of the future22 because it provides
ammonium bisulfate and ammonia for recycle while simultaneously elim-
inating the need to market ammonium sulfate.
4.3.3.2 General Discussion. Ammonia-based scrubbing systems
have received considerable attention in the history of S02 removal
from process offgases. The reasons for this include the relatively
high affinity of ammonia solutions for S02 and the ability to keep all
the compounds involved in solution, thereby avoiding scaling and
silting problems in scrubbers.4
4-45
-------
S02 removal from gas streams by ammonia-based scrubbing has been
studied intermittently by various groups since the 1880's.5 The earliest
reference found was a British patent issued to Ramsey in 1883.5 6 22 2S
The earliest studies were aimed at obtaining a regenerative system in
which the ammonium bisulfite solution from the absorber could be
regenerated by heat in a stripping column.22 However, during the
course of the regeneration cycle, the ammonium salts produced would
release ammonia that would subsequently contaminate the recovered S02.
Consequently, the S02 produced was unfit for sulfur, liquid S02, or
sulfuric acid production without further treatment. In addition,
these earlier systems consumed a great deal of process steam while
controlling S02 emission concentrations only to the .1,000- to 1,500-ppm
range.22 Oxidation of ammonium sulfite/bisulfite salts by oxygen in
the gas stream was another problem that hampered the early systems.
Oxidation reactions of this type were catalyzed by numerous contaminants,
mostly metallic oxides,5 transferred to the solution by the gas stream.
Additional investigation of the equilibria and chemistry involved in
ammonia-based processes was necessary to alleviate many of the short-
comings associated with the earlier systems.
Commercial and experimental modifications have evolved over the
years, with the greatest developmental emphasis being placed on methods
of regenerating the scrubbing liquor to reduce operating costs while
producing a variety of useful products. Among the most widely studied
methods of scrubber liquor regeneration are:5
Thermal stripping to yield primarily S02
Oxidation to yield primarily ammonium sulfate
Disproportionate to yield ammonium sulfate and elemental
sulfur
Acidulation to yield S02 and an ammonium salt.
Acidulation of the scrubber effluent has proven to be the most popular
route, having treated metallurgical gases as well as sulfuric acid
plant tail gases for over 50 years.22 Two basic variants of this
scheme exist; however, the basic unit operations involved are identical.
The principal process steps of the acidulation scheme are:
4-46
-------
Gas cleaning
S02 absorption
Acidulation of the absorber effluent
Stripping.
The most popular scenario involves absorber effluent acidulation
with sulfuric acid.4 5 22 A simplified flow diagram of this process,
more commonly known as the Cominco process, is presented in Figure 4-5.
The gas stream is first conditioned in the gas cleaning section of the
facility. Generally, the gases are cooled, fine solids are washed
out, and any S03 that may be present is absorbed in water, forming
weak sulfuric acid.4 The cool, clean gases then pass to the bottom of
an absorption tower where they are contacted countercurrently with an
ammoniacal solution. To improve S02 removal efficiency, absorption is
commonly performed as a staged operation. The effluent from the
absorption step consists of ammonium sulfite, ammonium bisulfite, and
ammonium sulfate, with ammonium bisulfite being the primary constituent.
All of these components are kept in solution, thus eliminating scaling
problems such as those associated with calcium-based systems. Off-
gases from the absorption step are vented to the atmosphere.
The acidulation step consists of pumping the absorber product
liquor to a continuous stirred tank reactor, where it is reacted with
sulfuric acid. The ammonium sulfite/bisulfite/sulfate liquor from the
absorbers reacts with the sulfuric acid to yield ammonium sulfate,
S02, and water. The reactions involved in this step are rapid at
ambient temperature and pressure.22
An S02 stripper completes the basic system. Liquor from the
acidulation step is fed to a packed column, where the S02 is stripped
from solution by contact with air or steam. The recovered S02 can be
used for the production of sulfuric acid, liquid S02, or elemental
sulfur. The liquid ammonium sulfate can be sold as is, fed to an
ammonium phosphate fertilizer plant, or crystallized to a dry crystal-
line material and marketed.22 If crystallization of the ammonium
sulfate is desired, the system may include a single-effect or multi-
effect evaporator-crystallizer, a crystal centrifuge, a dryer, screens,
conveyors, and a solid storage facility.
4-47
-------
^
I
SO2-
ESP or
Baghouse
Laden Ga
ses
»«2 noaunr I l\jn i MUIUULMI IUIM
1
Gases to Atmosphere •
i
— 1
Dust and Fume i
[7^
3
/^^
s
1
a
u
co
' ..
_^
•^
Electrostatic
Mist
Precipitator
H2O
^
Weak
./
k.
.8
S
.Q
u
8
Z
r*^ 5
t
LU
^
F
(D
0)
to
vt
^_
1
1
SlHimiMU
SO2
; H2so4
^»v
0)
J
<
^~-
T
' ^
•*
i
^
*^
CO
•S
^.
m
DC
3
E
UJ.
*l
r
Acid '
and Solids
1
1
1
|
:
^
3
t
III
o
1
L
i
I
1
1
1
1
|
^- —
r
(<*
\
\>
^—^
A
s I
L4
-~^~*
*
)
1
i
1
1
r i
i
i
i
1
Air or
/^
l-«v
XS02
Stripper
Steam
1
Ammonium Sulfate
(NH4)2S04
Make-up NH3 and H2O
Figure 4-5. Ammonia scrubbing process with sulfuric acid acidulation.
-------
Another variation of ammonia scrubbing that uses acidulation with
ammonium bisulfate rather than sulfuric acid might be used in areas
where ammonium sulfate is not marketable.5 22 This type of system
would use ammonium bisulfate to acidulate the absorber product liquor,
thus producing S02 and ammonium sulfate. A simplified flow diagram of
this process is presented in Figure 4-6. As indicated, the ammonium
sulfate produced in the acidulation step would be decomposed at
approximately 385° C (725° F) in an electrically heated or fuel-fired
furnace via the following reaction:
(NH4)2S04 -> NH4HS04 + NH3 .
The resulting ammonium bisulfate and ammonia are then recycled back to
the acidulation and absorbent makeup steps, respectively. Laboratory
tests have shown that ammonium bisulfate behaves very similarly to
sulfuric acid in the acidulation step and should give comparable
results.5
4.3.3.3 Design and Operating Considerations. Proper design and
operation of the gas cleaning system is necessary to ensure the reliable
operation of ammonia-based scrubbing systems. The primary purposes of
the gas cleaning step in ammonia-based processes are:
To cool and humidify the feed stream to prevent excessive
evaporation of the absorbent
To remove residual particulate matter that may catalyze the
oxidation of ammonium sulfite to ammonium sulfate in the
absorbers.
Waste gases from any process must be cooled to a reasonable
temperature prior to being fed to an absorption tower to avoid excessive
evaporation of the scrubbing solution. With regards to ammonia-based
scrubbing, the optimum feed stream temperature is established as a
tradeoff between mass transfer considerations and the costs of
preceding.5 Since the absorption of S02 is exothermic, interstage
cooling could also be required to maintain an acceptable temperature
in the absorbers (<55° C[<130° F]).5 However, if the feed stream is
sufficiently weak (<1 percent S02), precooling in a wet scrubber along
4-49
-------
on
O
SO2 ABSORPTION ACIDULATION
Gases to Atmosphere
STRIPPING
SO2
AMMONIUM SULFATE DECOMPOSITION
Electrostatic
Mist
Precipitator
Ammonium Sulfate
-------
with the humidification cooling that occurs naturally in the absorbers
should eliminate the need for interstage cooling.5 Operating experi-
ence at Cominco's acid plant tail gas cleanup system has shown that
humidification cooling of the 40° C (104° F) dry inlet gas stream
resulted in 25° C (75° F) absorption tower operation without prestage
or interstage cooling.5 At a TVA/EPA-sponsored pilot study on coal-
fired boiler gases at TVA's Colbert Power Plant, liquor temperatures
of up to 55° C (130° F) have been used in the absorber (after preceding)
while an acceptable S02 absorption efficiency was maintained.5 In
both of these cases, feedstocks were well below 1 percent S02. The
TVA/EPA pilot plant uses the ammonium bisulfate acidulation scheme to
process a feed stream that varies from 0.2 to 0.3 percent S02.
In applications to primary copper smelter effluents, gas pre-
cleaning would also be necessary to minimize S03 and particulate
concentrations in the absorber(s). S03 and certain metallic oxides
present in smelter effluents would tend to promote the oxidation of
ammonium sulfite to ammonium sulfate in the absorber proper.5 Parti -
culates tend to serve as nuclei upon which ammonium sulfate will
form.6
The most significant operating parameters in the absorption step
have been shown to be:5 6
Scrubbing solution temperature
Total concentration of S02 and NH3 in solution
Concentration of individual ammonium salts (sulfite,
bisulfite, and sulfate), which also determines pH
Ratio of liquid to gas flow
Type of internal absorber construction.
The important absorber operating parameters are necessarily related to
the vapor-liquid equilibria of the system and the associated approach
to equilibrium conditions. Studies have shown that the reactions
involved in the absorption of S02 by ammoniacal solutions are quite
rapid and do not affect the overall absorption rate.23 Thus, mass
transfer is the rate-determining process in the absorption step.
4-51
-------
The S02 gas-to-liquid mass transfer rate has been shown to be a
strong function of the system temperature.5 23 The transfer rate will
decrease as the system temperature increases. The S02 transfer rate
at 23° C (73° F) is approximately 4.4 times greater than the rate at
52.5° C (126.5° F).23 Thus, decreasing the solution temperature
enhances the equilibrium absorption of S02.
Minimizing the total S02 concentration in solution will enhance
S02 absorption efficiency, as will minimize the molar ratio of S02 in
solution as ammonium bisulfite to S02 in solution as ammonium sulfite
(S/C). This is because of the effects that these parameters have on the
S02 vapor pressure and the solution pH, both of which effect the
equilibrium absorption of S02. The system pH can be well correlated
over the range of operation (4.8 to 6.6) as a function of bisulfite-
sulfite ratio only.5
While decreasing the solution temperature enhances the equilibrium
absorption of S02, it also serves to minimize ammonia losses from the
system.5 6 Minimizing the total NH3 concentration in solution also
serves to minimize ammonia losses. However, while minimizing S/C
tends to favor S02 absorption, it also tends to increase ammonia
losses from the system. Thus, the choice of an optimum S/C is a
tradeoff between the equilibrium S02 absorption and the rate of ammonia
loss.
The solution pH must also be determined as a tradeoff between S02
absorption and the rate of ammonia loss. Experience at the TVA/EPA
pilot plant has indicated that a solution pH below 5.6 allows essen-
tially no S02 absorption while a pH above 6.8 results in unacceptable
rates of NH3 loss.5
Two primary factors that affect the approach to equilibrium in
the absorption column(s) are the internal absorber construction and
L/C. Absorption of S02 in ammoniacal solutions has been accomplished
in both packed- and tray-type columns.5 In the older units treating
Cominco's zinc roaster and acid plant waste gases at Trail, British
Columbia, wood-slat packing was used in one or more stages with L/G's
ranging from 2.1 to 4.3 £/m3/min (16 to 32 gal/1,000 cfm).4 5 The
4-52
-------
acid plant tail gas cleanup system at Olin-Mathieson's facility in
Pasadena, Texas, also employed this type of absorber.9 The more
recent TVA/EPA pilot plant work involved the use of multistage marble
bed and valve tray arrangements.5 S02 removal efficiencies in excess
of 90 percent have been achieved on a steady basis at each of these
installations.5
Although the L/G required to achieve a specified S02 absorption
efficiency will vary depending upon the type of absorption column
used, high L/G's are generally required to achieve acceptable S02
removal efficiencies. The benefits from the higher mass transfer
rates associated with high L/G's more than offset the cost of the
higher pressure drop.23
4.3.3.4 Operational Problems. No detailed information is
available on ammonia-based scrubbing system operation without prior
removal of particulate matter; however, the potential for problems is
evident. Excessive precipitation of solids in the absorber(s) could
be a problem if no gas-stream pretreatment were provided.5 Studies
performed in the U.S.S.R. have been the basis for claims that these
solids can be easily removed by filtration; however, domestic study
does not verify these claims. A yellow solid, identified as a homo-
geneous iron-ammonia-sulfur compound, was formed in the absorber
liquor at the TVA/EPA pilot plant, even though 90 percent or higher
particulate removal was accomplished prior to gas stream entry into
the absorber. In contrast to the Soviet claims, attempts to remove
the solids by filtration were unsuccessful because the precipitated
solids and flyash formed a gelatinous, thixotropic material that
blinded the filter media.5 However, in an application to smelter
offgases, precipitated solids may not behave in this manner.
As mentioned previously, the presence of particulate matter in
the feed stream may also cause undesirable reactions to occur in the
absorber(s), most notably sulfite oxidation. Particulates tend to
catalyze the oxidation of ammonium sulfite, resulting in ammonium
sulfate formation in the absorber(s). If smelter process offgases
were not cleaned prior to entry into the absorber(s), Cominco maintains
that problems could arise from fouling of cooler lines and other
4-53
-------
process equipment.4 Cominco precleans and conditions the offgases
from its Dwight-Lloyd sintering machines prior to SQ2 absorption.
It is generally accepted that the presence of particulate matter
in the absorbers has no effect on absorption rates;4 however, particulate
matter would have to be removed from the solution prior to acidulation
because it would interfere with the springing of S02 from solution.
Ammonia loss from ammonia-based scrubbing systems can be a problem
if the process is not controlled quite carefully. Decreasing the
absorbent temperature tends to reduce NH3 losses and enhance S02
absorption.5 6 Absorbent pH is also an important factor. At low pH,
there is no S02 absorption, while high pH results in an unacceptable
level of ammonia loss.6 Thus, the absorbent pH must be maintained
within a narrow range to result in good S02 absorption and reasonably
low NHs loss. This is achieved at Cominco by adding aqua ammonia to
each absorption tower and controlling the absorbent temperature as it
is fed to the towers.6
Scaling, erosion, and corrosion are generally not problems in
ammonia-based scrubbing systems.4 Scaling and erosion do not occur
because the absorbent is a solution rather than a slurry, and all
components tend to remain in solution. Corrosion does not cause
problems if the proper construction materials are used.
A serious problem encountered with most of the ammonia-based
scrubbing systems is the formation of an opaque fume in the exit gas
stream.4 5 6 The fume is partially attributed to gas-phase reactions
of ammonia, S02, and water-forming ammonium sulfite,5 6 which, due to
its small size, is not efficiently removed by conventional mist elimina-
tors.6 The fume is objectionable as an opacity problem. Wet ESP's
have been used at some installations to eliminate this problem.
Absorbent temperature and pH have been cited as having the most signifi-
cant effect on plume formation.4 Cominco reported adequate control of
the fume when operating with a liquor temperature of 25° C (77° F) and
a clean inlet gas stream.5 01 in Mathieson operated with a pH control
system that reportedly eliminated the plume opacity problem, but the
exact pH limitations were not reported.5 In the U.S.S.R., wet ESP's
are reportedly used at the top of ammonia absorbers to control the
4-54
fume.5
-------
At the TVA/EPA pilot plant, plume opacity was controlled to 5
percent or less when operating with an absorbent temperature of
approximately 50° C (120° F) by using a prewash to remove particulates
and S03, maintaining a low salt concentration on the top stage of the
absorber, and reheating the exit gas stream 6° to 11° C (10° to 20° F)
above the temperature required to dissipate the steam plume.5 However,
a visible plume would often reform outside the absorber on humid days.
4.3.3.5 Survey of Operating Experience. Scrubbing S02 from
waste gases with ammoniacal solutions has been practiced commercially
by Cominco in Trail, British Columbia, by Olin Mathieson in Pasadena,
Texas, and in Romania, Japan, France, Czechoslovakia, Germany, and the
U.S.S.R.5 The systems of Cominco-type design in place at Trail and
Pasadena use the sulfuric acid acidulation process. Details of foreign
systems are lacking, but it is unlikely that any of the foreign installa-
tions use the ammonium bisulfate acidulation process. No known ammonia-
based scrubbing processes employ ammonia bisulfate acidulation on a
large scale; however, the TVA/EPA pilot-plant study has provided data
on the ammonia bisulfate acidulation scheme.
The Cominco-type systems in place at Trail, British Columbia,
have been quite successful. These units have processed offgases from
a lead sintering plant, a zinc roaster, and a sulfuric acid plant.5 6
Performance data on these units are summarized in Table 4-6. As
indicated, these systems have achieved S02 removal efficiencies ranging
from 91 to 98 percent. The main problem with the Cominco-type systems
has proven to be ammonia loss.6 Presently, Cominco converts part of
the S02 into sulfuric acid while the remainder is converted into
ammonium sulfate.
While the ammonium bisulfate acidulation process has not achieved
commercial operational status, it merits consideration based upon the
fact that the need for an ammonium sulfate market is eliminated. In
addition, ammonia and ammonium bisulfate are regenerated for recycle
to the absorbent preparation and acidulation steps, respectively.
The TVA/EPA pilot plant has proven to be quite successful in
removing S02 from boiler offgases. This system has exhibited S02
4-55
-------
TABLE 4-6. PERFORMANCE DATA ON THE COMINCO-TYPE AMMONIA-BASED
SCRUBBING UNITS AT TRAIL, BRITISH COLUMBIA
Feed-
stream source
Lead sintering
plant
Zinc roaster
Sulfuric acid
plant
Feed-stream
volumetric
flow rate, scfm
150,000 to 200,000
0 to 45,000
50 to 95,000
Percent
S02 in
feed stream
0.3 to 2.5
0.5 to 7.0
0.9 to 1.0
Percent
S02 in
cleaned
gas b
stream
-0.10
-0.10
-0.09
S02
removal
efficiency,
percent
-96
-98
-91
Data taken from Reference 4.
DData taken from Reference 5.
"Estimated based upon the maximum feed-stream S02 concentration.
4-56
-------
removal efficiencies in excess of 90 percent, with a feed-stream
concentration of 0.2 to 0.3 percent S02 and an exit-gas S02 concen-
tration of 200 to 300 ppm.5 This experience indicates that desulfuri-
zation efficiencies of 90 to 95 percent are technically achievable.
As with Cominco-type systems, however, careful process control is
required to prevent excessive ammonia loss and the formation of an
opaque plume.
4.3.3.6 Applicability to Reverberatory Smelting Furnaces. The
Cominco-type process can achieve high S02 removal efficiencies over a
wide range of S02 concentrations. The range of S02 concentrations
over which these systems will operate efficiently easily encompasses
the range of S02 concentrations encountered in offgases from reverber-
atory smelting furnaces.6 In addition, applications of this process
to metallurgical effluents at Trail, British Columbia, have demonstrated
the system's capability to handle wide variations in feed-stream
volumetric flow rate and S02 concentration. However, ammonia loss
from these systems can be a problem. Ammonia volatility may limit
minimum the S02 emission concentration to the 200- to 300-ppm range
for practical applications.6
An additional factor that must be considered in assessing the
applicability of the Cominco-type process for control of reverberatory
furnace effluents is the possible need for interstage cooling as
discussed in Section 4.3.3.3. There would be no technical problems in
providing interstage cooling; however, this cooling must be evaluated
as an additional cost consideration.
In conclusion, with adequate process control provisions, the
Cominco-type system is a technically viable control option for the
control of reverberatory furnace effluents. Precise control of the
absorbent temperature and pH would be required to minimize ammonia
loss and eliminate the visible emissions problem; however, neither of
these problems appears to be chronic. Pretreatment of the feed stream
for the purposes of cooling, humidification, and particulate removal
would also be required, but this would present no technical problems
because adequate gas conditioning technology already exists.
4-57
-------
The feasibility of applying the ammonium bisulfate acidulation
scheme for reverberatory furnace effluent control cannot be assessed
accurately because no data exist concerning its application to metallurg-
ical offgases. However, because these systems are identical to the
Cominco-type systems except for the acidulation method and the sulfate
decomposition requirement, no evidence exists to suggest they would
not be technically feasible control options. This process is forecasted
to be the S02 removal system of the future22 because it provides
ammonium bisulfate and ammonia for recycle while simultaneously elimina-
ting the need to market ammonium sulfate.
4.3.4 Magnesium-Based Scrubbing Systems
4.3.4.1 Summary. Magnesium-based scrubbing systems have been
the object of a great deal of developmental work during the last
decade, especially in Japan, the U.S.S.R., and the United States.
Most of this developmental work has concentrated on the MAGOX process,
which uses magnesium sulfite/magnesium oxide slurries to effect S02
removal from gas streams. Systems of this type are especially attractive
because the absorbent can be regenerated.
Commercial demonstration runs on both coal- and oil-fired utility
boilers in the United States have demonstrated the ability of the
MAGOX process to achieve S02 removal efficiencies in excess of 90 per-
cent when operating on utility-related weak S02 streams.5 Experience
obtained from the earliest utility-related applications has also been
instrumental in the improvement and optimization of many of the design
features of the MAGOX slurry scrubbing process. In addition, the MAGOX
scrubbing system currently in place at the Onaharna smelter in Japan is
said to be a direct transfer of Japanese utility-related scrubbing
technology, which suggests that differences in the characteristics of
boiler and reverberatory furnace effluents do not constitute serious
obstacles to the transfer of utility-related technology.
Perhaps the most significant application of magnesium-based
technology to occur over the past few years has been the system installed
at the Onahama smelter in 1972. This system produces a concentrated
S02 stream from a portion of a weak S02 stream generated by three
green-charged reverberatory furnaces and shows a substantial ability
4-58
-------
to absorb fluctuations in the feed-stream S02 concentration.6 S02
removal efficiencies are typically in excess of 99 percent. Based
upon the successful operation of the Onahama system, a similar system
could provide S02 removal efficiencies of well over 90 percent in
applications to reverberatory furnace offgases of domestic origin,6
providing an adequate water supply is available to meet the process
water requirement. A system of this type would also provide product
flexibility because the concentrated S02 stream produced can be used
to manufacture liquid S02 and elemental sulfur as well as sulfuric
acid.
4.3.4.2 General Discussion. A number of magnesium-based scrubbing
systems provide effective S02 removal; however, U.S. developmental
work, as well as related Japanese and Russian work, has concentrated
on the use of magnesium sulfite/magnesium oxide slurries to effect S02
removal.5
A simplified flow diagram for the MAGOX process is presented in
Figure 4-7. This process involves operations associated with the
following primary areas:
Gas cleaning and conditioning
S02 absorption
Slurry handling
Solids drying and calcining.
Gas cleaning occurs when the feed stream is passed through a wet
scrubber, usually of the venturi type, where it is cooled and residual
particulate matter is removed. Undesirable halogens that may be
present are also removed in the wet scrubber. A bleed stream from the
wet scrubber is thickened to concentrate the particulate matter as a
slurry underflow, which is transported to a disposal area.
S02 absorption occurs when the cleaned gas stream is vented to a
scrubber, where it is contacted with the absorbent slurry. The absorp-
tion reaction takes place between S02 and magnesium oxide (MgO) and
results in the formation of magnesium sulfite (MgS03). Some of the
S02 may also react with MgS03 in the presence of water to form magnesium
4-59
-------
GAS CLEANING
AND CONDITIONING
i
CT>
O
SO2-Laden
Gases
SOo SLURRY HANDLING 1 SOLIDS DRYING
ABSORPTION | AND CALCINING
1
'
kener
1 !
ksh 1
to I
posal 1
1
1
Gases to Atmosphere /~)
V 1 St
/ | ' '--^ | Coke '
1 _^~ Makeup MgO
/ F~~7 \ and H2O
/ 1 ^"*— ' f^^ i '
I ' ^-^— —^^
/ Recycled „ , .
>- i Slurry Makeup Tank *- •• _. ualcmer
j. i wigu
1 I ''
i— ^ \
\ ' Centrate ' Dehydrated
\ ' ~ j Solids
\ *^ t
c 1
^ t Solids Cake
JH
° 1
JS '
< 1
1
1
1
1
1
1
Figure 4-7. Magnesium-oxide (MAGOX) scrubbing process.
-------
bisulfite, Mg(HS03)2, which immediately reacts with excess MgO to
yield additional MgS03. A small amount of MgS04 may be formed by in
situ oxidation of a small portion of MgS03 and/or by S03 absorption
into the slurry. The resulting aqueous slurry that exits the absorber
contains hydrated crystals of MgS03 and MgS04—e.g., MgS03 • 3H20,
MgS03 • 6H20, and MgS04 • 7H20; some excess MgO; and a solution saturated
with each of these components.
Slurry-handling operations involve splitting the slurry after it
exits the absorber, routing a portion to a centrifuge for partial
dewatering, and recycling the remainder to the absorber. At the
centrifuge, a moist cake consisting of crystals of MgS03 • 3H20,
MgS03 • 6H20, MgS04 • 7H20 and unreacted MgO is obtained. The clear
liquor centrate is then returned to the main recirculating slurry
stream together with makeup MgO slurry from the slurry makeup tank,
and the resulting stream is recycled to the absorber for further S02
recovery. The crystal cake is then conveyed to a dryer, where free
and bound moisture is removed by using a direct contact drying gas
under nonoxidizing conditions. The resulting anhydrous salts are then
calcined to MgO, which is reused in the absorption system after having
been slaked and slurried in the slurry makeup tank. Coke is added to
the calciner to convert any MgS04 present to MgO.
4.3.4.3 Design and Operating Considerations. The MAGOX process,
in common with most aqueous scrubbing systems, requires the feed
stream to be at least saturated with water vapor to minimize absorbent
evaporation and localized high salt concentrations in the absorber(s).
With initial gas-stream temperatures of between 150° and 315° C (300°
and 600° F), water quenching in a wet scrubber should provide accept-
able cooling and humidification.5 A venturi-type scrubber is usually
employed for this purpose. Also, because the MAGOX system uses the
closed system mode of operation, the introduction into the absorber of
particulate matter—most notably oxidation catalyzing metals and
sulfuric acid mist—must be minimized. An ESP in series with the wet
scrubber should provide adequate cooling, humidification, and particulate
matter removal in applications where copper smelter effluents are to
be processed.5
4-61
-------
The most significant operating parameters in the absorption step
have been shown to be:
Ratio of liquid to gas flow
pH of the absorbent slurry.
The S02 removal efficiency of the MAGOX system will increase as L/G is
increased;5 however, the L/G required to achieve a particular S02
removal efficiency will vary depending upon the type of absorber used.
Due to its effect on the S02 vapor pressure over the scrubbing
solution, the pH of the absorbent slurry has a distinct effect on the
S02 removal efficiency. Increasing the pH of the absorbent slurry
will tend to increase S02 removal efficiency5 because a high-slurry pH
will keep the vapor pressure of S02 above the solution small in compari-
son to the S02 partial pressure in the gas stream and thus promote
mass transfer.
Temperature exercises little adverse effect on the mass transfer
rates if the slurry pH is maintained at 6 or above due to the very low
S02 vapor pressures in evidence at this pH level.
Both venturi and mobile bed absorbers have been evaluated for use
in MAGOX systems, and both types have proven capable of attaining S02
absorption efficiencies in excess of 90 percent.5 However, venturi
absorbers require a higher operating L/G to achieve a given S02 removal
efficiency. In spite of this factor, the venturi absorber appears to
provide several operating advantages and has thus been selected for
use in domestic commercial installations. The Japanese have used
TCA's almost exclusively in their MAGOX systems; however, they do not
feel that TCA's provide any significant operational advantages and
have shown an interest in using venturi absorbers for future applica-
tions.6
The S02 concentration in the gas phase would not be expected to
adversely affect the S02 absorption rate until the concentration rises
above about 3.5 percent.5 Therefore, this would not be expected to
present any problems in applications to reverberatory furnace offgases,
which generally exhibit a maximum S02 concentration of 2.5 percent.4
4-62
-------
In the drying and calcining sections of a MAGOX scrubbing facility,
the use of fluid-bed equipment offers several advantages—namely,
lower capital and operating costs, lower heat loss, and better tempera-
ture control.5 Also, in the case of calciner operation, fluid-bed
operation would allow for more precise control of oxygen, which in
turn could minimize the requirement for coke to reduce the MgS04.
With regard to oxygen, where the gas stream is likely to contain high
levels of 02 such as those in reverberatory furnace offgases, the use
of organic inhibitors may be warranted to reduce the sulfite oxidation
rate.
4.3.4.4 Operational Problems. A great deal of information on
problems that have been encountered with MAGOX systems of Chemico
design was compiled during a commercial-scale test program at Boston
Edison's Mystic Power Station. This system processed boiler offgases
and produced solid MgS03, which was then sent to Essex Chemical's
facility in Rumford, Rhode Island, for calcination. The majority of
the problems encountered with this system occurred at startup and did
not prove to be chronic.
The majority of the problems encountered during startup centered
around the dryer.24 The initial problem encountered involved an
inability to transfer all of the dried product into the MgS03 storage
silo due to excessive entrainment of product fines in the countercur-
rent drying gas stream. Entrainment of this nature ultimately resulted
in overloading and plugging of the dryer cyclones. Efforts to alleviate
this problem by reducing the draft through the dryers were only parti-
ally successful, and the eventual solution involved modifications to
the dryer internals.
During the early stages of startup, the changing consistency of
the centrifuge cake led to some difficulty in dryer operation. Upon
startup of the absorption systems, the principal nucleation appeared
to result in the formation of crystals of MgS03 • 6H20. However, as
the system aged and reached an equilibrium temperature of 56° C (132° F),
nucleation involving the formation of MgS03 • 3H20 became more predomi-
nant. Because the mature trihydrate crystals were considerably smaller
4-63
-------
in size than the hexahydrate crystals, they were less efficiently
centrifuged and thus resulted in a wetter cake, which tended to agglom-
erate as it passed through the dryer. The formation of large nodules
caused shutdown of the conveying system by jamming and ripping the
rubber weigh belt. Larger agglomerates also caused some problems in
the pneumatic conveying operations. Ultimately, a pulverizer was
installed at the dryer exit to alleviate these problems.
The dryer was also unable to operate satisfactorily on high-sulfur
oil. Burner flameouts were frequent because of the formation of
coking deposits, which were later attributed to insufficient oil
preheating. With insufficient preheating, the oil did not attain the
proper viscosity for atomization. This problem was resolved by ensur-
ing that the oil was preheated to at least 125° C (260° F).
The calcining system also experienced some problems at startup,
the most significant being the difficulty in attaining a nonoxidizing
atmosphere in the calciner. This difficulty was overcome by reducing
air infiltration through the seals at the firing end of the calciner.
The venturi absorption system at this facility has operated
successfully with no evidence o'c scaling or silting at any points.24
Pluggage of small lines did occur occasionally; however, this problem
was easily remedied by back flushing.
The MAGOX process as now commercially defined imposes no unusual
problems in terms of corrosion or special processing equipment.5 As
is normal in slurry systems, the use of elastomer-1ined or specially
coated equipment and piping in the scrubbing and recirculation system
has provided acceptable service in the domestic commercial installa-
tions apart from some localized problems.
To avoid downtime for absorber maintenance, the Japanese included
a spare turbulent contact absorber in the Onahama system.6 Ball wear
in the absorbers is controlled by replacement every 2 to 3 months.
In conclusion, no chronic operational problems exist that would
hinder the operation and reliability of MAGOX systems that might be
employed to control weak S02 streams generated by reverberatory smelt-
ing furnaces. Although the earlier domestic systems of Chemico design
experienced numerous difficulties, all serious problems were eventually
alleviated.24
-------
While no single outstanding reason exists for the early diffi-
culties encountered, the development and operating experience in Japan
was greater than domestic experience at the time the first systems
were installed.6 The Japanese development work on these systems was a
direct transfer of utility-related technology and appears to have been
considerably more successful than the domestic development work on the
same type of system.25 However, domestic experience gained from the
earlier systems such as the Boston Edison system has led to numerous
improvements in domestic MAGOX system design.
4.3.4.5 Survey of Operating Experience. The MAGOX slurry scrub-
bing system has been under commercial-scale evaluation in the United
States since 1972.5 Construction of the first domestic commercial
system, the Chemico/Basic MgO sulfur recovery process, was completed
in April 1972. The scrubbing and recovery system at Boston Edison's
Mystic Station operated on offgases from a 150-MW, oil-fired boiler.
The regeneration facilities were located at Essex Chemical's facility
in Rumford, Rhode Island, where the S02 was used to produce sulfuric
acid. As noted in Section 4.3.4.4, several difficulties were encoun-
tered with this system; however, none proved to be chronic, and all
serious operational difficulties were eventually alleviated via design
modifications and/or changes in operating procedure. This unit was a
prototype trial installation built primarily to obtain operating data,
and its use was eventually terminated in June 1974. The unit demon-
strated availabilities in the range of 80 percent during the final
month of operation.
Another prototype Chemico/Basic unit was intalled at Potomac
Electric and Power Company's Dickerson Station No. 3 in 1973. This
scrubbing system processed offgases generated by a 95-MW, coal-fired
boiler. Performance testing of this system resulted in an average S02
removal efficiency of 88.9 percent.19 This unit, like the unit at
Boston Edison's Mystic Station No. 6, fulfilled its purpose by indi-
cating areas where improvement was needed. Demonstration runs on the
Dickerson unit were completed in August 1975.
Experience obtained from the systems at Boston Edison's Mystic
Station and at Potomac Electric and Power Company's Dickerson Station
4-65
-------
led to several design changes that helped to improve the reliability
of the Chemico/Basic system. Chemico subsequently installed a MAGOX
scrubbing system at the Chiba refinery in Japan, and this system is
apparently working satisfactorily.6 While this system provides a
concentrated S02 stream as feed to a Glaus plant, it is essentially
the same as the earlier Chemico system except for design modifications
to eliminate the previously mentioned operational problems and thus
improve system reliability.
In December 1972, the Onahama Smelting and Refining Company,
Ltd., began the operation of a MAGOX system designed to produce a
strong S02 stream from green-charged reverberatory furnace offgases.
The development and installation of the MAGOX S02 concentration system
at the Onahama smelter was carried out as a joint effort between the
Tsukishima Kikai Company (TSK) and the Mitsubishi Metals Corporation.6
The TSK-Mitsubishi MAGOX system is essentially the same as the Chemico/
Basic system as far as major design considerations are concerned, with
the only distinct difference being the type of absorber employed.
However, as mentioned in Section 4.3.4.3, the Japanese do not feel
that the TCA's they use have any distinct advantages over the venturi-
style scrubber employed in the Chemico/Basic design. The MAGOX system
was chosen to provide a concentrated S02 stream for direct acid-plant
processing because of low energy requirements in comparison to other
regenerative systems. This system processes approximately 1,600 dry
NmVmin (55,000 dry scfm) of reverberatory furnace off gas at an S02
concentration of approximately 2.5 percent.26 Although the feed
stream to the MAGOX system is maintained at about 2.5 percent,6 18 the
system has shown a considerable capability to handle the fluctuations
in the feed-stream S02 concentration that do occur.6 Also, no problems
exist with magnesium sulfate buildup because the turnover ratio and
makeup material are sufficient. A stream averaging 10 percent S02 is
produced by the calciner, while the gas stream that exits the absorber
generally exhibits an S02 concentration of about 20 ppm.6 The degree
of S02 removal associated with the resultant 20 ppm absorber offgas
stream is in excess of 99 percent. This system is of particular
4-66
-------
significance as it demonstrates the ability of a MAGOX system to
operate on a weak S02 stream generated by reverberatory smelting
furnaces.
4.3.4.6 Applicability to Reverberatory Smelting Furnaces.
4.3.4.6.1 Transfer of utility-related scrubbing technology.
Commercial demonstration runs on both coal- and oil-fired utility
boilers in the United States have demonstrated that the MAGOX slurry
scrubbing process is able to achieve S02 removal efficiencies of
96 percent or greater on boiler offgases.5 Considerable experimental
work in recent years has also served to establish, with some assurance,
the chemistry, kinetics, and mass transfer relationships that govern
the MAGOX process, suggesting that appropriate adjustment of L/G and
the slurry pH could achieve S02 removal efficiencies of 90 percent or
greater on weak S02 gas streams. In addition, the MAGOX scrubbing
system currently in place at the Onahama smelter is a direct transfer
of utility-related scrubbing technology, which suggests that differ-
ences in the characteristics of boiler and reverberatory furnace
offgases are not serious obstacles to the transfer of utility-related
technology.
4.3.4.6.2 Assessment of applicability based upon existing scrub-
bing systems that process metallurgical effluents. As discussed in
Section 4.3.4.5, a MAGOX system has been operated satisfactorily on
green-charged reverberatory furnace offgases at the Onahama smelter in
Japan. Therefore, application of this technology for the control of
weak S02 streams generated by green-charged reverberatory smelting
furnaces should be considered demonstrated. Although the MAGOX system
at Onahama has shown a considerable capability for absorbing fluctua-
tions in the feed stream S02 concentration, the S02 concentration in
the effluent from the green-charged reverberatory furnaces is kept
constant at about 2.5 percent,6 27 thus demonstrating the ability to
stabilize the S02 concentration. The system produces a concentrated
stream containing 10 to 13 percent S02 that can be processed directly
in a dual-stage absorption or single-stage absorption sulfuric acid
plant.6 Based upon the successful operation of the Onahama system, a
4-67
-------
similar system could undoubtedly provide S02 removal efficiencies of
well over 90 percent in applications to green-charged reverberatory
furnace offgases of domestic origin,6 providing an adequate water
supply is available to meet the process water requirement. Furthermore,
product flexibility is obtained as the concentrated S02 stream produced
can be used to make liquid S02 and elemental sulfur as well as sul-
furic acid.
Since the S02 removal efficiency of MAGOX systems will increase
as the gas-stream S02 concentration decreases, applications to offgases
generated by calcine-charged reverberatory furnaces should be tech-
nically feasible. A properly designed scrubbing system should be able
to accommodate the fluctuations in gas-stream S02 concentration that
occur with either sidewall or Wagstaff-charged furnaces that smelt a
calcine charge (see Section 4.3.6).
4.3.5 Citrate Scrubbing Processes
4.3.5.1 Summary. Scrubbing systems that use citrate-type absorbents
have been the subject of a great deal of developmental work, particu-
larly in the United States and Sweden. The U.S. Bureau of Mines, Salt
Lake City Metallurgy Research Center, has sponsored several studies on
the absorption of S02 in citrate-type absorbents. As a result of
these investigations, three pilot-scale studies that used process
technology developed initially by the Bureau of Mines were conducted.
The feedstocks involved in these studies were offgas streams
generated by green-charged reverberatory furnaces, coal-fired utility
boilers, and a Lurgi updraft lead sintering machine. The results of
all three studies were generally inconclusive because various opera-
tional problems proved to be chronic in nature, thus preventing long-term
reliable operation. However, most of the operational problems occurred
in the S02 reduction circuit and not the S02 absorption circuit.
Consequently, the pilot studies demonstrated that the system has the
ability to achieve high S02 removal efficiencies while processing weak
streams. Removal efficiencies in two of the three applications were
in excess of 93 percent.6 Thus, with further development and commercial-
scale demonstration, the Bureau of Mines process may prove to be a
4-68
-------
viable control option for weak S02 streams from reverberatory
furnaces.
Investigation into the absorption of S02 in citrate-type solutions
began in Sweden as an effort to improve the S02 absorption properties
of pure water. The Boliden AB Ronnskarsverken smelter at Skelleftehamn,
Sweden, has been removing S02 from smelter offgases by absorption in
cold water on a commercial scale since 1970. The economic success of
this operation depends upon the year-round supply of cold water (5° to
8° C [41° to 46° F]) near the smelter site. The investigations into
S02 absorption using citrate-type absorbents were prompted by the fact
that ample year-round supplies of cold water were not available in all
areas where water scrubbing might be used to effect S02 removal.
Thus, a means to improve the absorption properties of water at tempera-
tures greater than 8° C (46° F) was desired.
An absorption process that uses a brine of citric acid and sodium
citrate was eventually developed as a result of long-range work by the
Boliden Company of Sweden, the Norwegian Technical Institute SINTEF,
and Svenska Flaktfabrika (Flakt).6 A pilot-scale version of the
Flakt-Boliden process was used at the Ronnskar smelter to establish a
design background for the process. Gases from a number of various
lead and copper smelting operations were used as feedstocks for the
pilot facility. Flakt has reported that the test results look very
promising; however, the actual S02 removal efficiencies achieved
during the tests are not readily available.
The Flakt-Boliden process exhibits a very distinct advantage over
the Bureau of Mines process; i.e., it is relatively simple. The
process uses steam stripping of the absorber effluent to produce a
concentrated S02 stream, while the Bureau of Mines process employs a
very complex S02-reduction scheme to produce elemental sulfur. Conse-
quently, the Flakt-Boliden process is not likely to involve the numerous
operational problems that have been encountered thus far with the
Bureau of Mines process. Product flexibility is also obtained with
the Flakt-Boliden process because the concentrated S02 stream produced
may be liquefied to produce liquid S02, fed to a contact sulfuric acid
4-69
-------
plant to produce sulfuric acid, or fed to a Claus plant to produce
elemental sulfur.
Like the Bureau of Mines process, the Flakt-Boliden process has
only been applied on a pilot scale. Although the range of feed-stream
S02 concentrations processed by the Flakt-Boliden pilot plant easily
encompassed the range of S02 concentrations produced by reverberatory
smelting furnaces, the system cannot be considered to be a viable
control option for reverberatory furnace effluents until it is demon-
strated in a commercial-scale application that firmly establishes the
achievable range of S02 removal efficiencies.
4.3.5.2 General Discussion. The Bureau of Mines citrate scrubbing
process was developed as a result of investigations showing that
several organic acids—including acetic, citric, and lactic acid—had
a great affinity for S02. A mixture of citric acid, sodium citrate,
and sodium thiosulfate was finally selected for further development
because of its chemical stability. This system was designed to produce
elemental sulfur by using manufactured hydrogen sulfide (H2S) as a
reducing agent.
A simplified flow diagram of the Bureau of Mines citrate scrubbing
process is presented in Figure 4-8. The principal process steps are
as follows:
Gas cleaning and cooling
S02 absorption
S02 reduction and absorbent regeneration
Sulfate removal
Sulfur recovery
H2S generation.
Gas cleaning and cooling consists of removing particulate matter and
acid mist from the gas stream while cooling it to the 45° to 65° C
(115° to 150° F) temperature range. The cool, clean gas stream is
then vented to an absorber, where S02 absorption occurs. In the
absorber, the gas stream is contacted countercurrently with a sodium
4-70
-------
GAS CLEANING
AND COOLING
SO2- Laden / >
Gases
SO2 ABSORPTION
Gases to Atmosphere
S02 REDUCTION
AND
ABSORBENT SULPATE SULFUR
REGENERATION REMOVAL RECOVERY
H2S GENERATION
Kerosene or
SAE-10 Motor Oil
Steam
CH,
Figure 4-8. Bureau of Mines citrate scrubbing process.
-------
citrate/citric acid solution. Gases exiting the absorber are vented
to the atmosphere, while the loaded absorbent is pumped to a continuous
stirred tank reactor. In the reactor, the S02 is reduced to elemental
sulfur, and the citrate solution is regenerated by introducing H2S as
a reducing agent. Sulfur slurry from the S02 reduction reactor is fed
to the sulfur separation system. In the sulfur separation unit, the
sulfur is agglomerated by flotation with either kerosene or SAE-10
motor oil. The floated sulfur, at 35 to 40 percent solids, is fed to
a heater to produce molten sulfur. Liquid phases are then separated
in a decanter under a pressure of about 2.4 atm (243,200 Pa), resulting
in a bottom layer of high-quality molten sulfur and a top layer of
citrate solution. The molten sulfur is drawn off, and the citrate
solution is sent to a vacuum crystallizer, where sodium sulfate (Na2S04)
is removed. Effluent from the vacuum crystall izer i:; then recycled to
the absorber.
The H2S gas, which serves as the reductant in the S02 reduction
step, is produced from a portion of the molten sulfur drawn off the
decanter. Filtered molten sulfur from the decanter Is preheated by
indirect heat exchange with the H2S reactor effluent before being
vaporized and subsequently superheated to 650° C (1,200° F). A portion
of the resulting sulfur vapor is mixed with a hot natural gas/steam
mixture, and the resultant stream is routed to a querich-cooled catalytic
H2S-generation reactor. The remaining sulfur vapor is mixed with an
unpreheated natural gas/steam mixture and injected into the catalyst
bed to keep the bed temperature approximately constant. The hot
reactor effluent, consisting primarily of H2S and C02, is then used to
preheat the sulfur feed as mentioned above before being used in the
production of the steam required for H2S generation. The reactor
effluent is then cooled to approximately 60° C (140° F) before being
fed to the S02 reduction reactor. The portion of the molten sulfur
not used to produce H2S is generally cast into bricks.
The Flakt-Boliden system has resulted from long-range develop-
mental work carried out by several copper smelters and technical
institutes in Scandinavia.6 A flow diagram of the Flakt-Boliden
4-72
-------
GAS CLEANING AND COOLING
SO2 ABSORPTION
STRIPPING ANDSULFATE REMOVAL
SO2~Laden
Gases
Electrostatic
Precipitator
•vj
CO
Gas Cooler
Cleaned
Gases
1
Gases to Atmosphere
1
Absorbent
1
H2O
S02
Stripper
-Steam
Regenerator
Na2S04
Makeup Absorbent
Figure 4-9. Flakt-Boliden citrate scrubbing process.
-------
citrate scrubbing process is presented in Figure 4-9. The basic
process steps involved in the Flakt-Boliden process are summarized
below:
Gas cleaning and cooling
S02 absorption
Stripping and sulfate removal.
The gas stream is cleaned in a high-efficiency particulate collector
and then cooled to saturation by direct water injection. In addition,
any sulfuric acid mist present is removed in an electrostatic mist
precipitator prior to gas-stream entry into the absorber. The cool,
saturated gases are then vented to the bottom of an absorption tower,
where they are contacted countercurrently with citrate absorbent.
Gases exiting the absorber are passed through a mist eliminator and
then vented to the atmosphere, while the loaded absorbent is pumped to
a stripping tower. Stripping is accomplished by contacting the loaded
absorbent with steam. The concentrated S02 stream produced, which
contains a small amount of water vapor, is routed to a condenser,
where most of the water is condensed out. The condensate, which
contains only a small quantity of S02, is returned to the stripping
column. Absorbent solution containing a small amount of S02 is with-
drawn from the bottom of the stripper and pumped to a regenerator
unit. During regeneration, the sodium citrate and sodium sulfate are
separated from solution. This is accomplished by using seed crystals
and a cooling unit to recover sodium citrate and remove sodium sulfate.
Details of the regeneration system are presented in U.S. patent
No. 3,886,069.6 The sodium sulfate is discarded, and the recovered
sodium citrate solution is recycled back to the absorber.
The concentrated S02 stream, which can be as high as 95 percent
S02 with a water saturation temperature of approximately 30° C (85° F),
can be routed directly to (1) a Claus plant for the production of
elemental sulfur, (2) a contact sulfuric acid plant for the production
of sulfuric acid, or (3) a refrigeration/condensation system for the
production of liquid S02.
4-74
-------
4.3.5.3 Design and Operating Considerations. Proper design and
operation of the gas cleaning system is necessary to ensure efficient
S02 removal in citrate-type systems. As with other types of scrubbing
systems, the citrate systems require that the feed gas stream be
cooled and thus humidified to prevent excessive evaporation of the
absorbent solution. Because S02 absorption is favored by lower tempera-
tures, the optimum feed stream temperature is established as a trade-off
between mass transfer considerations and the costs of precooling.5
Pilot-plant work sponsored jointly by the Bureau of Mines and the
Magma Copper Company was conducted at temperatures between 42° C and
52° C (108° F and 125° F), with resulting S02 removal efficiencies in
the range of 93 to 99 percent.5 This particular pilot plant was
placed into operation in November 1970 at Magma's San Manuel, Arizona,
smelter and processed 8 mVmin (300 cfm) of gas from a green-charged
reverberatory furnace.
In the Bureau of Mines citrate process, adequate gas cleaning can
be effected by a baghouse, a packed scrubber, an electrostatic precip-
itator, or a venturi scrubber.6 In the Flakt-Boliden process, gas
cleaning and conditioning is accomplished by routing the gas stream
through a high-efficiency particulate collector such as an electro-
static precipitator and then cooling the stream to saturation by
direct-water injection.6 Passing the gas stream through an electro-
static mist precipitator prior to gas-stream entry into the absorber
is deemed advantageous in both types of processes because sulfuric
acid mist removal prior to absorption minimizes sodium sulfate forma-
tion in the absorber and thus minimizes the purge requirements.
The solubility of S02 in citrate-type absorbents is a function of
the S02 partial pressure in the gas phase, the hydronium ion concentra-
tion, and the ionization constants of sulfurous acid.5 The important
process variables in the absorption step can thus be identified as:
Absorbent solution pH
Absorbent solution composition
Feed-stream S02 concentration
4-75
-------
Feed-stream temperature
Ratio of liquid-to-gas flow.
S02 absorption is increased by maximizing the pH and buffer
content of the absorbent solution and minimizing the feed-stream
temperature. Under actual operating conditions, the citrate concen-
tration in the absorbent solution is established at the lowest level
compatible with the feed-stream S02 concentration and requisite solution
flows. The operating temperature selected balances the advantages of
higher absorbent S02 loadings at the lower temperatures against the
costs of feed-stream cooling. The actual absorbent recirculation rate
chosen to achieve a specified S02 removal efficiency is dependent upon
the absorbent citrate concentration required to effect the specified
removal efficiency at a given feed-stream S02 concentration. In a
Bureau of Mines pilot-plant program at the Bunker Hill lead smelter in
Kellogg, Idaho, the absorbent flow rate used to treat a 30 NmVmin
(1,000 scfm) stream from a lead sinter machine containing 0.5 percent
S02 was about 38 2/min (10 gal/min) of 0.5 M citrate solution.5 The
resultant S02 loading in the absorber offgas stream was approximately
107 mg/m3 (4,400 gr/ft3). With regard to the Bureau of Mines process,
the absorbent solution pH is reported to be limited by factors that
concern the S02 reduction procedure rather than by factors that govern
the absorption step.6 The pilot-scale work conducted by the Bureau of
Mines at the Magma and Bunker Hill smelters has used an absorbent pH
in the range of 4.0 to 4.6 and a 0.5 M citrate solution with a molar
ratio of NaOH to citric acid of approximately 2.0. Operating tempera-
tures at both installations ranged from 42° to 65° C (108° to 149° F).
L/G must be sufficiently high to ensure an adequate rate of S02
absorption. In a countercurrent, packed scrubber—the type of absorber
chosen in both the Bureau of Mines and Flakt-Boliden processes—the
L/G required is about 1.3 £/m3 (10 gal/1,000 ft3).6
The Bureau of Mines work has also shown that S02 absorption is
enhanced by the presence of thiosulfate. In practice, thiosulfate is
introduced into the system as the regenerated absorbent is returned to
4-76
-------
the absorption loop.5 In the pH range of 4.0 to 4.5--the usual pH
range of citrate absorbent solution—thiosulfate, trithionate, tetra-
thionate, and polythionate are all believed to be formed in the reduc-
tion/regeneration step of the Bureau of Mines process. These species,
once formed, are thought to react with H2S to yield elemental sulfur.
Studies of the chemistry involved in the reduction/regeneration step
have indicated that the reaction of H2S with thiosulfate is the rate-
determining process in this step.5 Thus, by allowing the thiosulfate
to build up in the system, the reaction rate is increased, which in
turn causes a decrease in the size of the reactor required to provide
the necessary residence time. For this reason, the presence of the
thiosulfate ion in the absorbent solution is deemed essential in the
Bureau of Mines process.
As with all closed-loop scrubbing systems, the accumulation of
oxidation products, particulates, and other soluble impurities must be
controlled by purging a portion of the recirculated absorbent solution.
In the Bureau of Mines process, this purge consists primarily of
sodium sulfate, and is taken from the vacuum crystal!izer as mentioned
in Section 4.3.5.2. Soda ash, Na2C03, which is added to the absorbent
recirculation loop, produces the sodium sulfate via reaction with the
sulfate ion is present in the system. The resulting sodium sulfate is
then readily removed after crystallization and filtration.
In the Bureau of Mines process, the sulfur slurry produced in the
reduction/regeneration step is vented to a sulfur separation unit,
where the sulfur is agglomerated by flotation with either kerosene or
SAE-10 motor oil. While the use of kerosene is quite effective in
this operation, it can be quite expensive. The use of kerosene and
the resultant losses that occur as presently estimated by the Bureau
of Mines constitute a major part of the total raw materials cost.5
However, laboratory-scale tests have indicated that the use of SAE-10
motor oil should provide results equivalent to those obtained by the
use of kerosene. In addition, predicted oil consumption would only be
one-fourth that of kerosene.
Very little experience has been gained in the Bureau of Mines
pilot-plant studies regarding onsite generation of H2S for use as a
4-77
-------
reductant. A reliable source of H2S must be available to ensure
uninterrupted operation of the Bureau of Mines process. However, due
to the locations of the domestic primary copper smelters, onsite H2S
generation would probably be preferred based upon economical considera-
tions and could conceivably be achieved by using a portion of the
recovered sulfur, steam, and a reducing agent such as methane. Production
of H2S via this route is exothermic, and the hot product stream that
exits the H2S generation reactor can be used to preheat the sulfur
feed to the reactor as well as to generate a portion of the steam
required for H2S generation.6
The Flakt-Boliden process uses steam stripping to produce a con-
centrated S02 stream from the absorber effluent. Thus, as shown in
Figure 4-9, the absorber effluent is routed directly to a stripping
column, where it is subjected to steam treatment in a countercurrent
fashion. Due to the equilibrium behavior of the S02/H20 system, the
stripping process is best conducted at reduced pressure (down to
approximately one-tenth of an atmosphere). The fact that stripping is
conducted under a vacuum is favorable with regard to steam consump-
tion, and low pressure steam or hot water can be used.6 Equilibrium
considerations also dictate the amount of steam required in the stripping
process. Consequently, it is highly desirable to operate the stripper
at the lowest possible temperature to reduce the steam requirement.
Steam consumption in the stripper is also directly related to the
feed-stream S02 concentration. In general, the higher the feed-stream
S02 concentration, the lower the specific steam consumption.
There is little information in the literature regarding the
design and operating considerations of the absorbent regeneration
system used in the Flakt-Boliden process. Regeneration is accom-
plished by using seed crystals and a cooling unit to effect separation
of sodium citrate and sodium sulfate. The sodium citrate is recycled
to the absorber, and the sodium sulfate is discarded. The details of
the regeneration system are contained in a U.S. patent, as mentioned in
Section 4.3.5.2.
4-78
-------
4.3.5.4 Operational Problems. The pilot-scale demonstrations
have revealed a number of actual and potential problems in citrate-
type scrubbing systems. The Bureau of Mines process is rather complex,
and more development work is needed to increase the reliability of the
process, especially with regard to the H2S generation and sulfur
handling systems.Q As currently defined, the Bureau of Mines process
uses natural gas, normally methane, in the production of H2S. WHh
the future availability of natural gas in doubt, its use is not very
desirable. This consideration is also applicable to the use of kerosene
or SAE-10 motor oil in the sulfur flotation step.
Mechanical problems have plagued every pilot-scale effort to date
involving the Bureau of Mines process. The majority of these problems
have occurred in the S02 reduction circuit. The problems experienced
include gas cleaning system failures, frequent pump failures, and flow
lines plugged with precipitates and melted sulfur.6 Due to the frequency
of the problems encountered, achieving complete steady-state operation
has been difficult. Many problems did prove chronic and were never
eliminated to the extent that system reliability could be increased to
acceptable limits. However, the Bureau of Mines is looking into steam
stripping of the loaded absorbent as an alternative to sulfur precipita-
tion by H2S,6 which would, if adopted, transform the Bureau of Mines
process into a process that would be essentially identical to the
Flakt-Boliden process. This alternative would simplify the process
significantly and would probably eliminate most of the aforementioned
problems.
As suggested above, the Flakt-Boliden process has not involved
the numerous problems on the pilot scale that the Bureau of Mines
process has encountered. This is primarily due to the fact that the
steam stripping procedure used by Flakt-Boliden is not nearly as
complex as the S02 reduction/H2S generation scheme employed in the
Bureau of Mines process. As mentioned previously, the bulk of the
problems that were encountered in the Bureau of Mines pilot studies
occurred in the S02 reduction circuit. Pilot plant experience with
the Flakt-Boliden process at the Ronnskar smelter in Sweden has been
more successful in terms of demonstrating system reliability.6
4-79
-------
4.3.5.5 Survey of Operating Experience. Neither the Bureau of
Mines citrate scrubbing process nor the Flakt-Boliden citrate scrubbing
process has been applied on a commercial scale. However, three pilot-
scale studies that involved the Bureau of Mines process have been
conducted in the United States, and pilot-scale testing of the Flakt-
Boliden system has been conducted at the Ronnskar smelter in Sweden.
The Bureau of Mines, Salt Lake City Metallurgy Research Center,
began research on FGD, with particular emphasis on the application of
scrubbing technology for control of S02 emissions from the nonferrous
smelting industry. Absorption in an aqueous solution of citric acid
and sodium citrate was selected for intensive study due to the chemical
stability, low vapor pressure, and buffering capacity of the absorbent.6
The purity and physical character of the precipitated sulfur were also
considered advantageous.
After promising bench-scale results were obtained, the Bureau
of Mines, in conjunction with the Magma Copper Company, constructed
and operated a pilot plant to remove S02 from reverberatory furnace
offgases generated at Magma's facility in San Manuel,, Arizona. This
pilot plant, constructed in 1970, treated approximately 8.5 NmVmin
(300 scfm) of gas from the reverberatory furnace containing 1.0 to
1.5 percent S02 and consistently removed 93 to 99 percent of the S02
from the gas stream. This system was consistently plagued by several
problems, most of which occurred in the S02 reduction circuit, as
mentioned in Section 4.3.5.4. Pump breakdowns and plugged flow lines
were the most frequently encountered problems.6 Failure of the gas
cleaning system was also quite frequent. Because of the chronic
nature of several of the problems encountered, useful data on the
consumption of citric acid and other reagents were not obtained.
However, as noted above, the system did achieve S02 removal effi-
ciencies of 93 to 99 percent while in operation.6
Another pilot plant was constructed by the Bureau of Mines and
operated jointly by the Bureau of Mines and the Bunker Hill Company at
Bunker Hill's lead smelter in Kellogg, Idaho. This pilot plant had a
nominal capacity of 28 NmVmin (1,000 scfm) of gas containing 0.5 percent
S02, with a corresponding sulfur production of about 1/3 ton of sulfur
4-80
-------
per day. The feed stream was taken as a slipstream from a Lurgi
updraft sintering machine. Problems with the S02 reduction system
plagued this pilot facility as well. Problems in maintaining a steady-
state feed stream were also encountered. Many problems did prove to
be chronic, thus preventing conclusive system tests.
A third pilot plant demonstrating the Bureau of Mines process was
independently built and operated by Arthur G. McKee and Company,
Peabody Engineering Systems, and Pfizer, Inc., at Terre Haute, Indiana,
in 1972.6 This facility treated approximately 57 NmVmin (2,000 scfm)
of stack gas with an S02 concentration ranging from 0.1 to 0.2 percent
from a coal-fired steam generating station. After several modifica-
tions to arrive at a final equipment configuration, the pilot plant
operated from March 15 to September 1, 1974. Although operational
difficulties prevented the steady-state operation of the entire system,
S02 removal efficiencies were consistently in the range of 95 to
97 percent. The longest sustained run was 180 hours.
The Flakt-Boliden process has been applied on a pilot scale at
the Ronnskar smelter in Sweden. The major reasons for installing this
system were to establish a design background for the absorption/stripping
process and to investigate the influences of various components in the
raw gas on the oxidation of S02 in the absorbent.
The raw gases that comprise the feed stream to the scrubbing
system originate in the multihearth roasters, electric furnaces,
converters, and various lead smelting operations at the Ronnskar
works. Because most of the metallurgical processes involved are not
continuous, the S02 concentration in the effluent streams fluctuates
between 0.2 and 6.0 percent by volume; however, the S02 concentration
does remain approximately constant at one level or another for long
enough periods to allow steady-state observations.
Flakt has reported that the test results look very promising,
with oxidation rates at least an order of magnitude lower than those
found in the Bureau of Mines process. Consequently, the requirement
for makeup chemicals should be relatively low for the Flakt-Boliden
process.6 The removal efficiency associated with this pilot facility
has not been published extensively in the literature; however, it is
4-81
-------
thought to be in excess of 90 percent. Similarly, information on
operational problems that may have been encountered is scarce.
Flakt is reportedly responsible for a new process development
that significantly decreases steam consumption in the stripping step.6
This could be especially important where gas streams with low S02
concentrations are to be processed because stripper steam consumption
is not a linear function of feed-stream S02 concentration and can
increase rapidly at low concentrations. As of this writing, however,
no details regarding the specifics of this new development are available.
4.3.5.6 Applicability to Reverberatory Smelting Furnaces. As
mentioned in Section 4.3.5.4, the Bureau of Mines process has experi-
enced numerous problems, most of which occurred in the S02 reduction
circuit; thus, the system was never able to demonstrate adequate
reliability. However, when operable, the system did demonstrate the
ability to effect S02 removal efficiencies of 93 to 99 percent from a
1.0 to 1.5 percent S02 stream that originated in green-charged reverb-
eratory smelting furnaces at the Magma smelter. In addition, the
system has proven its ability to handle extremely weak streams as a
result of the 95 to 97 percent removal efficiencies achieved by the
McKee-Peabody-Pfizer system while operating on coal-fired boiler
offgases with an S02 concentration of 0.1 to 0.2 percent. Thus, if
system reliability could be improved via the elimination of chronic
operational problems, the system could be a viable option for the
control of reverberatory furnace offgases. Nothing has indicated that
the problems encountered thus far represent fundamental flaws in the
application of the theory behind the system operation; however, reli-
able full-scale operation on either reverberatory furnace effluents or
other comparable streams will have to occur before the system can be
said to be fully demonstrated for the control of weak streams generated
by reverberatory furnaces.
Perhaps the most significant drawback of the Bureau of Mines
process is its requirement for kerosene (or SAE-10 motor oil) and
natural gas. It has been estimated that the costs of these commodities
alone would constitute 25 to 30 percent of the total annual direct
4-82
-------
operating costs.5 In addition, the uncertainty regarding the future
availability of these commodities tends to make their use undesirable.
When the absence of the S02 reduction/H2S generation steps and of
other related sulfur handling equipment are considered, the Flakt-Boliden
process is relatively simple compared to the Bureau of Mines process.
In lieu of the problems associated with the aforementioned operations,
steam stripping of the absorber effluent may be an attractive alterna-
tive. The supply of steam at a smelter should not be a critical item
because it is normally advantageous to strip at reduced pressure,
which allows the use of low-quality steam or even hot water. However,
this system has not been demonstrated in areas where the supply of
water may be a critical item.
Full-scale applications of the cold water scrubbing process at
the Ronnskar smeHer have resulted in S02 removal efficiencies in the
range of 98 percent.28 Therefore, because the addition of the citric
acid/sodium citrate buffer is known to increase S02 absorption effi-
ciency, it would be natural to assume that a full-scale application
using the citrate absorbent would be at least equally efficient.
However, full-scale application of the system to either reverberctory
furnace offgases or other comparable streams would have to be demon-
strated with adequate reliability before the system could be considered
to be a technically viable control alternative for weak streams from
reverberatory furnaces.
In summary, both the Bureau of Mines process and the Flakt-Boliden
process require further development before they can be considered
technically viable weak-stream control options. Additional develop-
mental work will be required to eliminate the numerous problems associated
with the Bureau of Mines process before a commercial-scale application
would be feasible. The Flakt-Boliden process, due to its relative
simplicity, has great potential for weak-stream control; however, a
commercial-scale application that uses the citrate-type absorbent will
be required to demonstrate that the system can achieve the required
level of S02 removal efficiency while maintaining high reliability.
Furthermore, with the Flakt-Boliden system, product flexibility is
obtained because the concentrated S02 stream produced can be liquefied
4-83
-------
to produce liquid S02, fed to a Claus plant to produce elemental
sulfur, or fed to a conventional sulfuric acid plant to produce sulfuric
acid.
4.3.6 Conclusions Regarding Flue Gas Desulfurization Systems
Sections 4.3.2 through 4.3.5 present discussions of six FGD
processes based upon four different types of chemical systems. These
are summarized in Table 4-7.
Although primary, ammonia scrubbing with ABS acidulation and both
types of citrate scrubbing systems are not considered demonstrated
because of a lack of full-scale demonstration. Of the three systems
that remain, two (lime/limestone and magnesium oxide scrubbing) have
been applied to green-charged reverberatory furnace offgases on a
full-scale basis, while the other (ammonia scrubbing with sulfuric
acid acidulation) has been applied to metallurgical offgases from
several sources on a full-scale basis. These systems, along with
their reported S02 removal efficiencies and reliabilities, are summar-
ized in Table 4-8.
All of these systems have proven capable of achieving their
design S02 removal efficiencies. Operating experience suggests that
the limiting factor is system reliability. However, while all three
types of systems have experienced operational problems that hindered
system reliability, the most severe problems have been eliminated or
minimized to acceptable levels. Consequently, the reliability of
these systems should be acceptable. The only possible exception might
be the MAGOX system, for which reliability data are scarce.
Because none of these systems has been applied to a calcine-charged
reverberatory furnace that uses Wagstaff charging, concern has been
indicated regarding their ability to handle fluctuations in the gas-
stream S02 concentration. The following paragraphs summarize the
capability of these systems to handle fluctuations in the feed-stream
S02 concentration.
Consider the typical absorber (scrubber) illustrated in Fig-
ure 4-10. To ensure that the absorber S02 removal efficiency remains
constant at a specified level over the range of possible gas-stream
4-84
-------
TABLE 4-7. FLUE GAS DESULFURIZATION PROCESSES ASSESSED FOR
APPLICATION TO REVERBERATORY FURNACE OFFGASES
Process
Type of
absorbent used
1. Lime/limestone scrubbing
2. Ammonia scrubbing with sulfuric acid
acidulation (the "Cominco" process)
3. Ammonia scrubbing with ammonium bisulfite
(ABS) acidulation
4. Magnesium oxide scrubbing (the MAGOX
process)
5. Bureau of Mines citrate scrubbing process
6. Flakt-Boliden citrate scrubbing
Calcium-based
Ammonia-based
Ammonia-based
Magnesium-based
Based upon a citric acid-
sodium citrate buffer
Based upon a citric acid-
sodium citrate buffer
4-85
-------
TABLE 4-8. EFFICIENCY AND RELIABILITY DATA FOR THE FGO PROCESSES BEING CONSIDERED
IN THE NSPS REVISION FOR PRIMARY COPPER SMELTERS
Type of FGD
system
Origin of the
feed gas stream
Reported S02
removal
efficiency, 2
Reported
reliability
Data source(s)
Lime/1impstone
Molybdenum ore roaster
(0.35 to 0.75% S02)
92 to 96
Good. Problems with
plugging and scaling
overcome.
i
CO
CTl
Primary copper smelter 99.5
reverberatorv furnace
(-2.5% S02) "
Approximately 99.3%.
Problems with
seal ing overcome.
Coal-fired utility boilers 90+
(0.04 to 0.4% S02)
Coniinco Ainiiioiiid Dwiyhl-Lloyd sintering 90 to 9»
machines, zinc roasters
and sulfuric acid plant
tail gases (0.3 to 7.0%
S02)
In excess of 90%.
bood. No plugging or
scaling since the
absorbent is a
solution rather
than a slurry
Weisenberg, I. J., T. Archer, F. M. Winkler, and
A. Prem. Feasibility of Primary Copper Smelter Weak
S02 Stream Control. Prepared for IERL, U.S. Environ-
mental Protection Agency, Cincinnati, Ohio, under
EPA Contract No. 68-03-2378. Publication No.
EPA-600/2-80-152. July 1980.
Background Information for New Source Performance
Standards: Primary Copper, Lead, and Zinc Smelters,
Volume I: Proposed Standards. U.S. Environmental
Protection Agency. Research Triangle Park, N.C.
Publication No. EPA 450/2-74-002a. October 1974.
Kohno, H., and Y. Sugawara. S02 Pollution Control with
the Lime-Gypsum Process at the Onahama Smelter.
(Presented at the AIME Annual Meeting. Chicago.
February 22-26, 1981.)
Slack, A. V. Application of Flue Gas Desulfurization
in the Non-Ferrous Metals Industry. (Presented at the
AIME Annual Meeting. Chicago. February 22-26, 1981.)
Background information for New Source Performance
Standards: Electric Utility Steam Generating Units,
Background Information for Proposed S02 Emission
Standards. U.S. Environmental Protection Agency,
Research Triangle Park, N.C. Publication No.
EPA 450/2-78-007a. July 1978.
Background information for New Source Performance
Standards: Primary Copper, Lead, and Zinc Smelters,
Volume I: Proposed Standards. U.S. Environmental
Protection Agency. Research Triangle Park, N.C.
Publication No. EPA 450/2-74-002a. October 1974.
Weisenberg, I. J., T. Archer, F. M. Winkler, and
A. Prem. Feasibility of Primary Copper Smelter
Weak S02 Stream Control. Prepared for IERL, U.S.
Environmental Protection Agency, Cincinnati, Ohio,
under EPA Contract No 68-03-2378. Publication No.
EPA-600/2-80-152. July 1980.
Matthews, J. C., F. L. Bellegia, C. H. Gooding, and
G. E. Meant. S02 Control Processes for Nonferrous
Smelters. Research Triangle Institute. Research
Triangle Park, N.C. Publication No. EPA-600/2-76-008.
January 1976.
-------
TABLE 4-8 (continued)
Type of FGD
system
Reported S02
Origin of the removal
feed gas stream efficiency, %
Reported
reliability
Data source(s)
MAGOX
I
CO
Primary copper smelter 99+
reverberatory furnace
(-2.5% S02)
Coal-fired utility 90+
boilers
(0.04 to 0.4% S02)
17 to 80%C
Located at the Onahama smelter in Japan.
No reliability data given.
GBaspd on data presented for Boston Edison's Mystic No. 6.
weisenberg, I. J. , T. Archer, F. M. Winkler, and
A. Prem. Feasibility of Primary Copper Smelter
Weak S02 Stream Control. Prepared for IERL, U.S.
Environmental Protection Agency, Cincinnati, Ohio,
under EPA Contract No. 68-03-2378. Publication No.
EPA-600/2-80-152. July 1980.
Background Information for New Source Performance
Standards: Electric Utility Steam Generating Units,
Background Information for Proposed S02 Emission
Standards. U.S. Environmental Protection Agency,
Research Triangle Park, N.C. Publication No.
EPA 450/2-78-007a. July 1978.
Maxwell, M. A., and G. R. Koehler. The Magnesia Slurry
S02 Recovery Process Operating Experience with a
Large Prototype System. " (Presented at the AICHE
Annual Meeting. New York. November 26-30, 1972.)
System reliability improved steadily as operational problems were alleviated.
-------
Absorber Offgas Stream
Inlet Gas Stream
Absorbent Pump
La, Xa
Pregnant Liquor Pump
LEGEND
Y = gas phase pollutant mol fraction
X = liquid phase pollutant mol fraction
G = gas stream molar flow rates
L = liquid stream molar flow rates
a = process streams entering or exiting at the top of the absorber
b = process streams entering or exiting at the bottom of the absorber
Figure 4-10. Typical absorber configuration.
4-88
-------
S02 concentrations, proper measures must be taken in the design stage.
To qualitatively determine how fluctuations in the gas-stream S02
concentration will effect the performance of a given scrubber, one
must examine the equilibrium and operating relationships that are
involved. Fluctuations in S02 concentration below that specified in
the design basis should actually enhance the S02 removal efficiency.
Fluctuations above the design basis will, however, cause the S02
removal efficiency to decrease if L/G is held constant. Thus, to
ensure that the specified S02 removal efficiency can be maintained,
provisions must be made during the design stage- to ensure that the
absorbent flow rate to the column, and thus L/C, can be increaseo when
fluctuations on the high side of the design basis are encountered.
This would involve sizing the absorbent pump to handle the expected
maximum pumping duty, while ensuring that all of the equipment downstream
of the absorber(s) could operate efficiently at the higher absorbent
flows. Consequently, it is reasonable to assume that any designer,
given the gas-stream profile, would account for anticipited fluctuations
in the gas-stream S02 concentration by providing the necessary provisions
in terms of solvent handling capability and process control. In
actual practice, this is normally accomplished by incorporating some
"overdesign" into the systems. For instance, if a 90-percent S02
removal efficiency is desired at all times, the system should be
designed to achieve a somewhat higher S02 removal efficiency so that
fluctuations in the inlet gas stream S02 concentration will not cause
the efficiency to drop below 90 percent at any time. This would
involve providing more area for mass transfer (using larger absorbers)
as well as the items mentioned above. Since, in reality, process
control systems do have inherent time lags, some overdesign would
ensure that these systems could maintain the desired S02 removal
efficiency. Based upon the actual operating data presented in Section
4.3.2 through 4.3.5, as well as on engineering judgment, EPA has
concluded that any of the systems presented in Table 4-8 designed to
operate with an S02 removal efficiency of 99 percent (at the highest
anticipated S02 concentration) should not exhibit S02 removal efficiencies
4-89
-------
below 90 percent as a result of the short-level fluctuations inherent
in offgases from calcine-charged reverberatory furnaces. Thus, 90 percent
has been specified as a reasonable level of $62 control for these FGD
systems in light of the anticipated maximum reverberatory furnace S02
concentration (~2.0 percent).
4.4 INCREASING THE S02 STRENGTH OF REVERBERATORY FURNACE OFFGASES
The quantity of sulfur emitted from reverberatory furnaces as S02
averages about 22 percent (by weight) of the total sulfur entering the
smelter. However, this sulfur quantity can be as high as 34 percent
for green charge or as low as 9 percent for calcine charge produced by
a fluid-bed roaster.6 A survey29 30 31 of the domestic reverberatory
furnace installations indicates that the S02 concentration in the
offgases is typically in the range of 1.0 to 2.0 percent for green-
charged furnaces and 0.4 to 1.5 percent for those using calcine feed.*
As noted previously, these concentrations are too low for economically
processing in conventional contact sulfuric acid plants.
Options to be considered for facilitating the control of reverber-
atory furnace offgases fall into two major categories. The first
group includes control systems that are applied directly to the weak
S02 stream. These are discussed in Section 4.3. The second group
includes furnace operating modifications, which can lead to an increase
in the concentration of S02 in the offgases and thus facilitate acid
plant control. Such operating modifications include the following:
Elimination of converter slag return
Minimizing infiltration
Preheating combustion air
Operation at lower air-to-fuel ratio
Predrying wet charge
Oxygen enhancement techniques.
Each of these techniques is discussed in subsequent subsections.
Although these techniques have been used at various smelters, in most
*S02 concentrations after gas cleaning, dry basi
4-90
-------
cases they have been applied not to obtain control-related benefits
but to improve operating conditions and/or increase production.
Discussion of the first five operating techniques is based primarily
on information from Reference 6.
4.4.1 Elimination of Converter Slag Return
As discussed in Section 3.2.2.1, converter slag is typically
returned to reverberatory furnaces to recover copper. These slags are
charged while molten by pouring them into a chute or launder leading
to an opening in the furnace wall. Fifty or more ladles of converter
slag can be returned in a 24-hour period, depending upon the number of
converters and the smelter throughput rate.6 Converter slag return is
employed by most domestic smelters using reverberatory furnaces.
In contrast to reverberatory furnace slag, converter slag is high
in magnetite content—normally 17 to 35 percent.6 To some extent, the
magnetite reacts with iron sulfide in the bath to release S02. However,
each time converter slag is returned, a large volume of air enters the
furnace (which is kept under a slight negative pressure). The net
effect is a reduction in the average S02 concentration, although no
data are available with regard to the extent by which the S02 concentra-
tion is reduced. An additional disadvantage of converter slag return
to reverberatory furnaces is that adding converter slag can lead to
magnetite buildups on the furnace bottom.
An alternative to processing converter slag in reverberatory
furnaces directly is to use the flotation process to recover the
copper content of this material. The flotation process requires slow
cooling of the slag to allow the growth of crystalline particles of
sufficient size to be amenable to flotation. Cooling requirements are
on the order of 1 to 2 days. Following cooling, the slags are ground
to liberate the copper-containing mineral particles. Grinding to at
least 85 percent minus 200 mesh is not uncommon.6 It should be noted,
however, that the elimination of converter slag return may not be
feasible at some smelters due to inadequate capacity in the flotation
circuit.
4-91
-------
4.4.2 Minimizing Infiltration
The general objective in the operation of the reverberatory
-5
furnace is to maintain a slight negative pressure (1.2 x 10 atm to
1.2 x io"4 atm [1.22 to 12.2 Pa] in the furnace. This practice prevents
gases from escaping through any openings and draws some outside air
into the furnace. Of course, the lower the pressure, the greater the
amount of air entering will be. Excessive infiltration into the
furnace proper is undesirable because the air must be heated to furnace
temperatures, which wastes fuel. With regard to possible acid plant
control of the offgases, infiltration is undesirable because it reduces
the S02 concentration.
Although the volume of infiltrated air is strongly affected by
the furnace pressure, the number and the size of the openings are also
major factors. Sources of infiltrated air to the furnace proper
include, in addition to the converter slag return port, charge feed
system openings, furnace repair ports, spaces around burners, and
expansion spaces between bricks. Infiltration also occurs through
expansion spaces in the furnace uptake and through waste heat boilers
and ESP's downstream of the furnace. Infiltration from the latter
sources does not affect operation of the furnace but does serve to
decrease the S02 concentration of the offgases. Minimizing the infil-
tration occurring from all sources has the advantage of reducing the
offgas volume, which allows for the use of smaller gas handling equip-
ment and reduces gas handling energy requirements.
It appears that, due to the energy requirement for heating air
entering the furnace, only sufficient air to achieve complete combustion
and to allow the oxidation of labile sulfur is permitted to enter the
furnace proper under current operating practice. For example, at
ASARCO smelters, which use natural gas fuel, air inleakage is regulated
such that approximately 1 percent oxygen* is maintained in the offgases
*This low oxygen level may not be achievable with other types of
fuels, which require more excess air for combustion (see Section
4.4.4).
4-92
-------
entering the furnace uptake.30 However, indications are that substantial
infiltration typically occurs downstream of the furnace,32 in the
waste heat boiler and ESP, primarily because there has heretofore been
no need for eliminating such dilution to enhance S02 control.
The most extensive work performed to reduce infiltration in a
reverberatory furnace and associated gas handling equipment specifically
to facilitate S02 control was conducted at the Onahama smelter in
Japan. At this furnace, it was indicated that the draft was rigidly
controlled to prevent the intake of any excessive air, although particu-
lars of the draft control system were not described. Also, an extensive
sealing effort was employed to eliminate air leaks through crevices,
burner clearances, openings in the furnace roof, sidewalls, fettling
chutes, damper slots, expansion joints, peep holes, cleaning doors,
and, especially, dust-discharging hoppers of the boilers and the ESP.6
The result was a reduction in the total air infiltrated into the
furnace and flue from a level of approximately 50 percent of the
furnace gas at the uptake to less than 15 percent. The reduction in
infiltration led to an increase in the S02 concentration of approxi-
mately 0.4 percentage points.27 Ongoing maintenance at Onahama includes
repairing and replacing portions of the furnace roof every 6 months to
minimize inleakage.
4.4.3 Preheating Combustion Air
Preheating the furnace combustion air can increase the concentra-
tion of S02 in the exhaust gases. The increase results because the
sensible heat content of the air serves to decrease the fuel require-
ment, which in turn results in a smaller volume of combustion products.
The use of preheated air for natural gas combustion is usually
required to increase the flame temperature, which, in turn, increases
the rate of heat transfer to the melt by radiation and convection.
The use of preheated air for oil and coal combustion can, however,
produce furnace control and durability problems because of the higher
flame temperatures.6
At the Onahama smelter, preheating of the secondary air to the
oil burners has increased the smelting rate. No operational problems
4-93
-------
were reported by Onahama.33 The resulting increase in S02 strength
has been estimated at 0.1 to 0.2 percentage points.6
4.4.4 Operation at Lower Air-to-Fuel Ratio
The quantity of combustion air supplied to the furnace burners
has a direct effect on the S02 concentration in the furnace offgases.
Introducing less air will result in higher offgas S02 concentrations.
The theoretical amount of air required to combust a given amount
of fuel is that necessary to complete the oxidation of all of the
carbon, hydrogen, and sulfur contained in the fuel. In most cases,
however, excess air (more than the theoretical amount) is supplied to
ensure complete combustion because the air and fuel are not perfectly
mixed. The quantity of excess air supplied should not exceed the
level at which the heat lost to the additional air exceeds the heat
gained from the combustion of additional fuel.
The quantity of excess air used in firing reverberatory smelting
furnaces varies depending upon the type of fuel burned. Because
gaseous fuels mix with air easily, natural gas burners can achieve
essentially complete combustion using a relatively small amount of
excess air (between 0 and 10 percent). Thorough mixing of liquid
fuels with air is more difficult to achieve, and, as a result, oil
fuels require up to 18 percent excess air. Solid fuels, such as
ground coal, are the most difficult to combust completely and require
from 12 to 50 percent excess air.
The average S02 concentration of 2.6 percent in the offgases from
the Onahama reverberatory furnaces is partially attributed to operation
at a low air-to-fuel ratio.6 These furnaces, which are fired with
oil, operate at approximately 10 percent excess air.6 The air entering
at the burner end of each furnace is actually less than the theoretical
requirement—the balance being supplied by infiltration.
While operation of the Onahama furnace with a fuel-rich mixture
does not provide maximum heat release within the furnace, it does (in
addition to increasing the S02 concentration) increase furnace durability
and reduce the formation of NO . Originally, 500 ppm NO resulted
X X
when operating on the oxygen-rich side. When the air-fuel ratio was
reduced to the fuel-rich side, the NO concentration dropped to 100 ppm.6
4-94
-------
4.4.5 Predrying Wet Charge
Ore concentrates processed in green-charged reverberatory furnaces
typically contain from 5 to 15 percent moisture. As the moisture is
eliminated in the furnace offgases, it contributes to the dilution of
S02.
Processing green charge in concentrate dryers allows some increase
in the concentration of S02 from the reverberatory furnace by eliminat-
ing the dilution effect of the moisture (although no data on the
magnitude of the increase are available). Other advantages also
result. Concentrate drying results in an overall decrease in fuel
consumption because the temperature of the moisture is raised to only
about 80° C (180° F) in the dryer, as compared to 1,290° C (2,350° F)
in the furnace. Also, by eliminating moisture, the possibility of
steam-induced pressure surges or destructive steam explosions in the
furnace is greatly reduced. Perhaps the major disadvantage to use of
a dryer is the capital outlay required for the unit and associated
feed handling equipment.
4.4.6 Oxygen Enhancement Techniques
Oxygen enhancement involves tiu- substitution of pure oxygen for
all or a portion of the combustion -.: 'vd to a piece of smelting
equipment by a variety of separate l~;».>:ques. A number of these
techniques are currently used by the soffiting industry worldwide. Use
of oxygen enhancement results in an increase in the S02 concentration
in the exhaust gases and could lead to the production of an acid-plant-
strength gas. A inajor cause of the increased concentration is the
elimination of the nitrogen associated with air when it is used for
combustion. Because air contains 79 percent nitrogen and only 21 per-
cent oxygen, the dilution effect associated with the nitrogen in the
air is evident. The decrease or elimination of nitrogen also results
in a reduction in the size and cost of all downstream gas handling and
processing equipment. Another factor contributing to increased S02
concentrations in the case of some techniques (e.g., oxygen-sprinkle
smelting and roof oxygen lancing) is increased sulfur removal. Other
4-95
-------
major advantages associated with oxygen-enhancement include increased
furnace capacity, or a decrease in fuel usage for a fixed furnace
capacity. These aspects are discussed in Section 3.4.3.5.
4.4.6.1 Survey of Experience with Oxygen Enhancement. The
advantages of oxygen enhancement techniques in reverberatory furnaces
from the production and pollution control standpoints have been empha-
sized by numerous authors.26 33 49 All of the theoretical and experi-
mental studies on oxygen enhancement in reverberatory furnaces show
that the S02 concentration in the flue gas stream increases with the
use of oxygen. Thus, the use of oxygen enhancement techniques to
alleviate the weak stream control problem is a real possibility.
Various methods of introducing oxygen into the reverberatory
furnace have been used to date. These may be categorized as follows:
Oxygen introduced directly with fuel in oxygen-fuel burners
Oxygen mixed with primary air and introduced into the
existing burner system (oxygen enrichment)
Undershooting the flame with oxygen or oxygen-enriched air
Oxygen lancing of the molten furnace bath
Oxygen sprinkle smelting
Illustrations of these alternative methods of oxygen introduction are
presented in Figure 4-11.
Itakura et al.35 and Goto33 report using oxygen-fuel burners at
the Onahama smelter and refinery. The result was an increase of
0.3 percentage points in the flue gas S02 concentration per oxy-fuel
burner used,33 or a total increase of 0.6 percentage points for the
2-burner configuration.
The work at the Caletones smelter41 with oxy-fuel burners indicated
that S02 concentrations between 5.7 and 7.3 percent (on a dry basis)
could be attained with a full-scale, green-charged reverberatory
furnace. This was accomplished by burning all of the fuel with indus-
trial grade oxygen (97 percent pure) in individual oxygen-fuel burners
located in the roof of the furnace. Processing of a nickel calcine
4-96
-------
Oxygen Lances
Fuel Input
Oxygen Lancing of the Bath
,Oxy-Fuel Burners
Oxygen-Fuel Burner Usage in the Furnace
Undershooting of Flame with Oxygen
Undershooting the Flame with Oxygen
Oxygen Introduction to the
Primary Combustion Ajr_____ — Primary Combustion Air for Burner
i
Oxygen-Enrichment of Primary Combustion Air
. Oxygen Sprinkle Burners
Oxygen Sprinkle Smelting
Figure 4-11. Methods of oxygen addition.
Charge Banks
Charge Banks
Fuel Burners
Oxygen Jets
Oxygen Enriched
Primary Combustion
Air for Burners
4-97
-------
charge by the same technique at the Inco Copper Cliff smelter indicated
that offgas S02 concentrations of 5 to 6 percent, as measured in the
furnace offtake, could be achieved.48
A number of smelters have conducted studies involving oxygen
enrichment of the combustion air. Pluzhnikov et al.37 report oxygen
enrichment of primary combustion air using roof-mounted burners at
Noril'sk Ore-Mining Combine. Kupryakov et al.38 report using oxygen-
enriched primary air in burners at the Almalyk copper smelter. Achurra
et al.41 report on limited tests made with oxygen enrichment of the
primary burner air at the Caletones smelter. Wrampe et al.39 report
Linde's experience with oxygen enrichment of the combustion air in
domestic reverberatory furnaces.
As mentioned previously, all of the theoretical and experimental
studies on oxygen enrichment indicate that the S02 concentration in
the flue gas stream increases with oxygen enrichment. Wrampe et al.39
present a model that expresses the flue gas S02 concentration in terms
of the smelting rate and the degree of oxygen enrichment. The relation-
ship between the flue gas SQ2 concentration and the degree of oxygen
enrichment as given by the model39 and the results of studies conducted
by Kupryakov were compared and found to be in reasonable agreement.
At a constant fuel rate with 21 and 25 percent oxygen content in the
combustion air, the model projects flue gas S02 concentrations of 3.6
and 5.5 percent, respectively. For the same levels of enrichment,
Kupryakov1s investigations yielded offgas S02 concentrations of 3.4
and 5.2 percent. Kupryakov's results appear somewhat high because
measurements were made at the furnace outlet, i.e., upstream of the
gas-handling equipment. The data are assumed to be on a dry basis.
The figure of merit with regard to Kupryakov's work is the percentage
increase in the S02 concentration with oxygen enrichment, which amounts
to 53 percent. As reported by Eastwood et al.,36 theoretical investi-
gations at the Rokana smelter have predicted that, with 30 percent
oxygen in the primary combustion air, reverberatory furnace offgas S02
concentrations could be increased from an initial value of 1.1 percent
to 2.0 percent on a dry basis, which amounts to an 82-percent increase
4-98
-------
in S02 concentration. The theoretical concentrations appear to pertain
to the furnace outlet. The difference in initial S02 concentrations
(before oxygen enhancement) reported by Kupryakov and Eastwood could
be attributed to differences in the pyritic sulfur content of the
charge and/or weaknesses in the theoretical model used at the Rokana
smelter.
Undershooting the flame with oxygen or oxygen-enriched air has
been studied at several smelters. Saddington34 and Kupryakov38 report
that studies of this nature have been performed at the Inco Copper
Cliff and Almalyk smelters, respectively. Eastwood et al.36 report on
experience accumulated at the Rokana smelter.
Beals et al.40 report S02 concentrations as high as 18 percent
with oxygen lancing of the bath. These data resulted from tests
conducted in a pilot-scale furnace. Beals currently has a patent
assigned to Kennecott for using oxygen lancing of the melt to increase
production and to obtain a high S02 concentration in the flue gases.
Achurra et al.41 report on limited investigations made with roof-mounted
oxygen lances at the Caletones smelter.
Oxygen sprinkle smelting has been largely developed by Queneau
and Schuhmann.47 This method involves retrofitting existing reverbera-
tory furnaces with oxygen sprinkle burners as illustrated in Figure 4-12.
In this process, which operates on the same principle as flash smelting,
dry concentrate charge is burned with commercial oxygen, and the
molten droplets of charge fall to the hearth below. The S02 concentration
produced with oxy-sprinkle smelting is expected to be in the 20 to 30
percent range.49
Detailed discussions of the most significant applications of
oxygen-related technology are presented in the following sections.
4.4.6.1.1 Use of oxygen at the Caletones smelter. Development
work involving the use of oxygen in a reverberatory furnace began at
the Caletones smelter in Chile during 1971.41 The use of oxy-fuel
firing began in 1974, with the retrofit of a single burner onto the
roof of the green-charged No. 3 furnace. By mid-1976, the furnace was
operated with 12 oxy-fuel burners and none of the conventional burners.
4-99
-------
Copper Concentrate,
Flux and Oxygen
O
o
Converter Slag
Return Launder
Alternate Matte
Launder
Oxygen Sprinkler Burners
Copper Concentrate,
Coal and Oxygen
Matte
Offgases
Matte
Launder
Slag
Launder
Figure 4-12. Conventional copper reverberatory smelting furnace that has been converted to an
oxygen sprinkle smelting furnace.47
-------
The positions of the oxy-fuel burners in the furnace roof are shown in
Figure 4-13. A top view of the furnace is illustrated in Figure 4-14.
Oxy-fuel burner dimensions and operating data are presented in Table 4-9.
The technique of full oxy-fuel firing is considered successful by
Caletones42 and apparently was in use as of May 1977. More up-to-date
information on oxy-fuel firing at Caletones is not available.
The matte grade is approximately 49 percent with full oxygen
usage (all fuel burned with commercial oxygen), while the copper
content of the slag is 0.7 percent. The use of calcium carbonate as
flux could be discontinued due to higher slag temperatures that resulted
from the use of 100 percent oxygen. The increase in matte temperature
also minimized magnetite buildup on the furnace bottom. In addition,
refractory wear on a per-ton-of-copper-produced basis was either less
or the same as previously encountered.
The production rate of Reverberatory Furnace No. 3 at the Caletones
smelter was increased from 686 dry Mg/day (755 dry tons/day) to 1,520
dry Mg/day (1,670 dry tons/day) with the use of the oxy-fuel burners.
Overall energy usage (including oxygen manufacture) was reduced from
1.6 x 106 kcal/Mg (5.9 x 106 Btu/ton) of material smelted without
oxygen to 1.1 x 106 kcal/Mg (4 x 106 Btu/ton) with full oxy-fuel
firing.42 The latter value represents a net decrease in energy usage
of 32 percent. The oxygen rate at the elevated production level was
345 Mg/day (380 tons/day).
The most significant result from a pollution control standpoint
is the S02 concentration in the offgases, which was measured at 5.7
and 5.8 percent on a dry basis for the 12-burner configuration.* The
location of measurement was not reported; however, it was indicated
that the gas is suitable for acid plant control.42 The S02 concentra-
tion in the offgases before oxy-fuel burners were used was likewise
not reported.
*A concentration of 7.3 percent S02 on a dry basis was measured for a
10-burner configuration. The increase in concentration, when compared
to the 12-burner configuration, was attributed to a change in the
feed composition.41
4-101
-------
Figure 4-13. Oxy-fuel burner locations in Reverberatory Furnace No. 3 at the Caletones smelter.
4-102
-------
o
GO
Oxy-Fuel Burners
Slag Skim
4-
Slag Return Launder fj Matte Taj Holes ' ^Matte Tap Syphon
_____ Old design
. New design
Figure 4-14. Plan and elevation of Reverberatory Furnace No. 3 (Oxy-fuel burner locations are shown in plan view).
-------
TABLE 4-9. GENERAL SPECIFICATIONS3 OF THE TYPE OF OXY-FUEL
BURNER EMPLOYED AT THE CALETONES SMELTER4
Burner dimensions
Length 94 cm (37 in.)
Outside diameter 15.2 cm (6 in.)
Fuel rateb 12 £/min (3.17 gpm)
Oxygen rate 35.0 NmVmin (1,236 ncfrn)
Ox>gen pressure 2.04 atm (30 psi)
Oil pressure 6.80 atm (100 psi)
Cooling water rate 11.5 1/min (3.04 gpm)
Noise level 90 dB
Data pertain to maximum firing rate.
Fuel type: ENAP-6 oil. with a heating value of approximately 10,200
kcal/kg.
4-104
-------
4.4.6.1.2 Use of oxygen at the Onahama smelter. In 1971, the
first of two oxygen-oil burners was installed in a green-charged
reverberatory furnace at the Onahama smelter in Iwaki City, Japan.33
Table 4-10 presents the general specifications of the type of burner
employed at Onaharoa.35 Two burners are presently being used when
furnace capacity nust be increased. The burners penetrate the roof
vertical "ty.
Table 4-11 summarizes reverberatory furnace operating data for
two time periods--one prior to the installation of the oxy-fuel burners
and one during operation 01 two oxy-fuel burners. As indicated, the
charging rate was increases from 22,520 Mg (24,770 tons) per month to
27,200 Mg (29,920 tons) per month with the use of the two oxy-fuel
burners—a gain of about 21 percent in throughput. The corresponding
oxygen consumption rate was 667,060 NmVmonth (2.4 x 10V scf/month).
Fuel consumption during conventional operation (without oxygen) was
approximately 170.0 £/Mg (^1.0 gal/ton) charge, while consumption at
the increased production rate (with 2 oxy-fuel burners) decreased to
146.0 £/Mg (35.0 gal/ton) charge in the furnace proper.
The use of oxy-fuel burners at Onahama was reported to resu'lt in
an increase in S0? concentration of 0.3 percentage points per oxy-fuel
burner, which amounts to an increase of 0.6 percentage points for the
2-burner configuration. The initial S02 concentration (before oxy-fuel
burner operation) was 1.5 percent.27 This value appears to have been
measured after gas cleaning, and is assumed to be on a dry basis.
It appears that S02 concentration data obtained at Caletones and
Onahama are fairly consistent, when examined on the same basis. If
the initial S02 concentration (before oxygen usage) at Caletones is
assumed to be approximately 1.5 percent, then each burner in the
12-burner configuration increases the S02 concentration by about
0.35 percentage points.
4.4.6.1.3 Use of oxygen at Inco's Copper Cliff smelter. The
Inco Copper Cliff smelter in Ontario, Canada, processes nickel calcines
in two reverberatory furnaces fitted with roof-mounted oxy-fuel burners.44
4-105
-------
TABLE 4-10. GENERAL SPECIFICATIONS OF THE TYPE OF OXY-FUEL BURNER
EMPLOYED AT THE ONAHAMA SMELTER35
Burner dimensions
Length 193 cm (76 in.)
Diameter 15.2 crn (6 in.)
Maximum fuel rate 6.6 £/min (1.8 gpm)
Maximum oxygen rate 20 NmVmin (707 cfm)
Oxygen pressure 4.8 atm (71 psi)
Type of cooling system Water jacket
aFuel type: Bunker C oil.
4-106
-------
TABLE 4-11 TYPICAL REVERBERATORY FURNACE OPERATING DATA BEFORE
AND AFTER THE USE OF OXY-FUEL BURNERS AT THE ONAHAMA SMELTER33
Parameter
Parameter
Time period
December 1970
(Without
oxy-fuel burners)
January 1972
(With 2
oxy-fuel burners)
Concentrate smelted
Silicious flux smelted
Limestone smelted
Reverts smelted
Total solid charge
Fuel oil consumed
Oxygen consumed
Matte produced
Matte grade
Slag produced
Slag copper content
22,470 Mg
2,966 Mg
1,868 Mg
518 Mg
27,822 Mg
4,723 x 103£
18,465 Mg
34.5 percent Cu
18,546 Mg
0.46 percent
27,142 Mg
3,663 Mg
1,970 Mg
707 Mg
33,482 Mg
4,870 x 103£
667,058 Nm3
22,843 Mg
34.4 percent Cu
23,530 Mg
0.47 percent
4-107
-------
Fluxed nickel concentrate feed containing 22 percent sulfur,
8.7 percent nickel, 2.5 percent copper, and 30.1 percent iron is first
roasted in Herreshoff multihearth roasters. Conveyors then deliver
calcine and crushed nickel converter slag to the reverberatory furnaces.
The calcine is sidewall charged using a drag conveyor/fettling pipe
feed system. Both furnaces are fired with Bunker C fuel oil and
measure 35 m (114 ft) long by 9 m (30 ft) wide. One of the furnaces
burns all of the fuel oil in 10 roof-mounted oxy-fuel burners, while
the other burns fuel oil in 2 conventional air/fuel end-wall burners
and 4 roof-mounted oxy-fuel burners. The grade of matte produced by
the Inco furnaces is about 25 to 30 percent Ni + Cu + Co.
Without the use of oxygen, the offgas flow rates from each furnace
varied from 1,400 to 1,700 NmVmin (50,000 to 60,000 scfm) and contained
1.0 to 1.5 percent S02.48 With full oxy-fuel operation, the furnace
offgas flow rate varies from 570 to 850 NmVmin (20,000 to 30,000 scfm)
and contains 5 to 6 percent S02 (measured at the furnace offtake).48
After passing through the uptake and waste heat boiler, these gases
are typically in the 1 to 2 percent S02 range. Also, with full oxy-fuel
operation, the charging rate increased from 1,270 to 1,830 Mg (1,400
to 2,020 tons) concentrate feed per day, and the fuel requirement per
unit of charge was reduced to 45 percent of the operating value in
evidence during conventional operation without oxygen. When the
overall energy requirement, including oxygen manufacture, is considered,
only 67 percent as much energy is used with oxygen enhancement, result-
ing in a 33-percent savings.
The increase in the S02 concentration in the reverberatory furnace
offgases was attributed to the following:48
A 50-percent reduction in the offgas volumetric flow rate as
a result of eliminating the nitrogen present when air was
used.
A decrease in the amount of air infiltration, because the
front burner wall is totally sealed.
4-108
-------
Besides increasing the S02 concentration in the offgases, oxy-fuel
firing has also increased NOX production as a result of the higher
flame temperatures and increased availability of oxygen. Inco has
indicated that the NO,, levels from the oxy-fuel-fired furnace are
X
substantially higher than those from conventional furnaces.50
Tonnage oxygen has been used in the Copper Cliff smelter since
1945, when Inco began piloting the flash smelting of copper and nickel
concentrates.44 Commercial application of this technology started in
1952. Air blast to the nickel converters has been enriched to 30 per-
cent oxygen by weight since 1958; by the mid-50's, the reverberatory
furnace combustion air was being enriched with oxygen, and chalcocite
concentrates were smelted in Peirce-Smith converters using an oxygen-
enriched blast. Development work on the oxy-fuel burners continued
until June 1978. The resulting Inco-designed oxy-fuel burners produce
a good flame pattern and acceptable noise levels operating at firing
rates of 3.8 kg (8.4 Ib) of oil per minute and 7 to 27 kg (15 to
33 Ib) of oxygen per minute. Oxygen is supplied to the burner for 70
to 100 percent of stoichiometric requirements. The burners are fed
oxygen at 25 psi and oil at 500 psi pressure. The oil is introduced
at 116° C (240° F).
For the furnace using only oxy-fuel burners, smelting started at
the beginning of October 1979 using 12 oxy-fuel burners and continued
for the next 13 months with only minor interruptions. During this
time, 652,000 Mg (718,700 tons) of dry solid charge were smelted, and
only minor repair delays were experienced. Overall, refractory consump-
tion and fuel efficiency compared favorably with those for conven-
tional operation; roof refractory consumption of about 0.76 kg/Mg of
dry solid charge was indicated to have been lower than initially
anticipated. The only major problem after startup involved difficulty
with slag temperature control. This problem was resolved by removing
two of the burners at the skimming end, resulting in the current
10-burner operation.
The higher heat transfer capability and improved heat distribution
made possible with oxy-fuel burners has led at Inco to better control
4-109
-------
of furnace bottoms and minimized the height of the charge banks. As a
result, the matte-holding capacity of the oxy-fuel furnace is now
about 400 Mg (440 tons) instead of the 150 Mg (165 tons) during conven-
tional operation. The increase in active furnace volume coupled with
the higher heat transfer rates has substantially increased the ability
of the furnaces to adapt to changes in feed rates and compositions.
Improvement in the ability to recover from furnace slow-downs arising
from maintenance or from environmental requirements has also been
accomplished.
The two reverberatory furnaces equipped with oxy-fuel burners
have over 27 combined months of smelting experience (as of November
1980) and over 1.27 million Mg (1.4 million tons) of charge smelted.44
Slag compositions, metal losses, and dust carryovers have remained
essentially constant over 16 months of operation and are the same for
the conventional and oxy-fuel furnaces.
4.4.6.1.4 Use of oxygen at the Phelps Dodge-Morenci smelter.
Phelps Dodge Corporation is examining the oxygen sprinkle smelting
approach as an option for its Morenci and Ajo smelters. Development
work is being performed at the Morenci smelter in Arizona. Small-scale
tests of the oxygen sprinkle smelting scheme have been completed to
date. The results of these tests are described in Section 3.4.3.5.3.
Oxygen sprinkle smelting is expected to produce S02 concentrations in
the 20 to 30 percent range.49
The oxygen sprinkle process involves closing all unnecessary
openings in an existing reverberatory furnace and fitting the furnace
with three or more sprinkle burners. These burners are designed to
accomplish several important functions within the confines of the
existing structure. The two most critical functions of the burners
are the following:
To provide intimate mixing of finely divided concentrates
with the oxygen-rich gas phase.
To sprinkle the feed uniformly over most of the molten bath
surface.
4-110
-------
Queneau and Schuhmann47 report that the use of sprinkle burners trans-
forms existing reverberatory furnaces into oxygen flash smelting
units. Figure 4-12 shows the configuration of a reverberatory furnace
as it would be after conversion to oxy-sprinkle smelting via the
installation of three oxy-sprinkle burners.
4.4.6.2 Comparison Between Calcine and Green Charged Oxy-fuel
Burner Operations. Location of the burner flame at the base of the
charge bank at the calcine charged Inco furnace is a modification of
the technique used at the Caletones and Onahama smelters with a green
feed.
Green charged furnace operation with oxy-fuel burners has been
established to impinge the burner flame directly on the charge banks
to (1) protect the furnace wall; (2) produce concentrated heat directly
on the smelting material; and (3) increase the available heat by
molecular recombination; and of course (4) eliminate the nitrogen heat
loss.
The end of the flame from the burner just touches the base of the
calcine charge banks in the Inco furnaces. Earlier tests resulted in
the flame actually hitting the bottom of the furnace causing an increase
in refractory wear as well as dust generation. The optimum flame
pattern was adjusted by the amount of oxygen supplied to the burner
which is currently 90 percent of stoichiometric. The additional
oxygen required enters through leaks in the sidewall as a result the
negative pressure within the furnace.
While charge banks are established in the Inco furnace, the angle
of repose is considerably lower than with green charge. This still
allows some protection of the sidewalls by the charge banks. However,
Inco has indicated that at times the charge banks have disappeared
completely. The operator must maintain proper feed distribution along
the full length of the furnace to prevent localized wall heating and
severe refractory erosion, as this furnace does not employ sidewall
cooling.
Localized heating trends at Inco are monitored with 12 thermocouples
inserted to within about 23 cm (9 in.) of the hot face and about 45 cm
(18 in.) above the slag line. In addition, Inco has a novel sidewall
4-111
-------
construction arrangement wherein castable refractory is installed in
small sections in metal cans to allow easy and rapid replacement.
Replacement of the entire sidewall in the upper section can be accom-
plished in 24 to 48 hours and has been done at 6-month intervals.
This technique has been used with conventional smelting as well as with
oxy-fuel smelting.
4.4.6.3 Effect of Oxygen Enhancement on Matte Grade. The use of
tonnage oxygen in reverberatory furnaces leads to higher oxygen concen-
trations within the furnace and increased flame temperatures. These
effects prompt an examination of the effect of oxygen usage on desul-
furization within the furnace and hence on the matte grade.
Furnace matte grades were examined at the Caletones smelter
concurrent with the increase in oxygen usage in the green-charged
No. 3 furnace.41 This furnace was started up in April 1974 and retro-
fitted with two oxy-fuel burners by August 1974. A typical matte
grade for the year 1974 was determined to be 53.6 percent Cu. This
value is assumed to apply to operation with 2 oxy-fuel burners.
Oxygen usage was increased during 1975, with a maximum of seven oxy-
fuel burners in use by August. The typical matte grade for 1975,
which is taken to apply to operation with seven oxy-fuel burners, was
reported at 55.6 percent copper. Oxygen usage was further increased
in 1976, with a maximum of 12 oxy-fuel burners in use by June. The
typical matte grade for 1976 (assumed to pertain to operation with 12
oxy-fuel burners) was determined to be 48.7 percent, a decrease from
the 1974 and 1975 values. Typical charge analyses reported for the
period 1974 to 1976 showed little variation in copper contents, which
ranged from 38.5 to 39.7 percent.
Onahama Smelting and Refining Company examined the effect of
oxy-fuel burner operation on desulfurization in both pilot tests and
full-scale operations. The pilot tests were made with a single oxy-fuel
burner. The results of the tests showed that the oxidation of sulfur
and the matte grade are comparable, under appropriate conditions, to
those in a conventional green-charged operation.33 Data collected
from full-scale operations support the pilot test conclusions. Typical
4-112
-------
operating data (see Table 4-11) indicate a matte grade of 34.5 percent
during conventional operation and a value of 34.4 percent with 2 oxy-fuel
burners.33
Inco has reported that the use of full oxy-fuel firing in its
calcine-charged nickel reverberatory furnace led to essentially no
changes in matte grade.51 Furthermore, no changes were made in the
degree of roast of the feed.51 The implication is that essentially
the same degree of sulfur elimination (per unit of feed) occurred in
the furnace during conventional operation and full oxy-fuel firing.
Beals et al.40 of Kennecott investigated changes in matte grade
occurring with roof oxygen lancing in pilot plant tests. Significant
increases in matte grade were noted with this technique, because the
oxygen was blown directly into the bath and oxidized a portion of the
sulfur in the matte. Matte grades of 41.8 to 48.8 resulted with lance
operation, as compared to a value of 38.0 for conventional furnace
operation.
Queneau and Schumaun,47 developers of oxygen sprinkle smelting,
have made projections (based on theoretical calculations) of matte
grade resulting with this technology. The calculations were performed
for a feed yielding a matte grade of 35 percent Cu during conventional
green-charged furnace operation. The projections indicate that sub-
stantial increases in matte grade may result, as the process operates
autogenously at matte grades of between 60 and 65 percent copper. It
should be noted, however, that increases in matte grade of this
magnitude would not occur if additional heat is provided via the
combustion of ground coal mixed with the feed.
In conclusion, it is noted that extensive tests of oxy-fuel
firing indicate essentially no changes in matte grade. It is expected
that no changes in matte grade would occur with oxygen enrichment and
oxygen undershooting because these schemes are less severe in terms of
oxygen usage and flame temperature than oxy-fuel firing.
Pilot-plant tests of roof oxygen lancing indicate that this
technique leads to increases in matte grade of from 4 to 10 percentage
points.
4-113
-------
Oxygen sprinkle smelting, which operates on the same principle as
a flash furnace, may lead to substantial increases in matte grade
because sulfur and iron in the concentrates are consumed to provide all
or a significant portion of the heat necessary for smelting.
4.4.6.4 Refractory Wear with Oxygen Usage In Reverberatory Furnaces.
A possible constraint on the use of oxygen in reverberatory smelting
furnaces is the increased potential for excessive refractory wear. To
some extent, the method of oxygen introduction to the furnace dictates
the amount of increase in the roof and side wall temperatures. For
example, when oxygen-fuel burners are used, they can release the heat
close to the charge, and the roof temperature is not significantly
increased. Inco experience44 with the calcine-charged reverbera-
tory furnace indicated that, when the burners were placed in the roof
but close to the side wall, deterioration of the wall occurred. When
the burners were moved toward the center line of the furnace the
refractory deterioration was better controlled.
When undershooting the flames with oxygen, the hottest zone of
the flame is at the bottom next to the bath.34 Preliminary investiga-
tions by Saddington et al.34 indicate that refractory cost per unit of
output in a reverberatory furnace should not increase when oxygen is
introduced by undershooting. Figure 4-15 shows the roof temperature
variation along the length of two different furnaces at Inco with and
without oxygen undershooting. These furnaces have the same dimensions
and (without oxygen enhancement) the same capacity.34 The maximum
temperature at the furnace roof did not increase significantly with
oxygen undershooting. However, the temperatures were higher along the
entire length of the furnace.
The recent work with oxy-fuel burners at Inco44 indicates that
overall refractory consumption compared favorably with conventional
operation. The roof refractory consumption with oxy-fuel firing was
about 0.76 kg/Mg DSC (dry solid charge) (1.52 Ib/ton DSC). Operation
of the furnaces with oxy-fuel firing has required an increased operator's
awareness of proper feed distribution along the full length of the
4-114
-------
2,700
2.600
LL
o
£ 2,500
3
re
E 2,400
05
re
Temperature Measurements
Taken 4' Below Top of
Furnace Roof
No. 6 Reverb
No Oxygen
85 TPD Coal
115 MCFH Natural Gas
Burner Nozzles
1
10 20 30 40 50 60
Feet from Burner End
70
No. 5 Reverb
with Oxygen-
Undershooting
80 TPD Oxygen
70 TPD Coal
85 MCFH Natural Gas
80
90
Figure 4-15. Reverberatory furnace temperatures in the vicinity of the furnace roofs
with and without oxygen-undershooting at Inco smelter.
4-115
-------
furnace. Improper fettling can easily lead to localized wall heating
and severe refractory erosion. Localized heating trends are monitored
for this reason.
Kupryakov et al.38 indicated that during a development period at
the smelter, where operating conditions varied markedly with various
concentrations of oxygen, the average wear of the furnace roof reached
a rate of 0.2 mm/24 h (0.008 in./24h). Taking into account the accuracy
of measurement and the frequent changes in the operating conditions
within the furnace, the wear of the furnace roof during reverberatory
furnace operation with an oxygen-enriched blast did not differ markedly
from wear when operating with conventional air blast.
Itakura et al.35 reported that using oxygen-fuel burners in a
reverberatory furnace at the Onahama smelter did not increase the
lining wear. They postulated that this probably was due to the heat
being released close to the charge rather than the lining of the
furnace.
Wrampe and Nollman39 reported during their work on oxygen enrich-
ment at reverberatory furnaces at various U.S. smelters that the
overall refractory temperatures increased by approximately 0.28° C
(0.5° F) per 28.0 NmVmin (1,000 scfm) of oxygen. During their test,
excessive roof temperatures did not become a problem until production
increases exceeded 50 percent.
Caletones smelter experience42 indicated that it was necessary
to use a bottom ventilation system. Details of this system were not
provided. Overall, refractory consumption on a per-unit-of-feed basis
was reported to be no more than during conventional operation.42
It is concluded that with proper implementation, oxygen enhancement
technology would probably not lead to increases in refractory consump-
tion on a per-unit-of-feed basis. Furthermore, based on the extensive
use of oxygen enhancement techniques by the industry worldwide, refrac-
tory wear considerations are not expected to hinder or limit the
adoption of oxygen enhancement techniques by the domestic industry.
4-116
-------
4.4.7 Summary of Operating Modifications Useful for Increasing Offgas
S02 Concentrations
Operating modifications that lead to increases in the S02
concentration from reverberatory furnaces include (1) elimination of
converter slag return, (2) minimizing infiltration, (3) preheating
combustion air, (4) operation at a lower air-to-fuel ratio, and
(5) using oxygen enhancement techniques. Extensive data are generally
lacking with regard to industry experience using the first four techniques
to increase S02 concentrations. Based on the available data and
engineering judgment, each of these techniques appear useful for
increasing S02 concentrations on the order of up to a few tenths of a
percentage point. Because inadequate data exist and because gains in
S02 strength achievable generally appear relatively small, the first
four techniques mentioned are not considered as possible control
strategies in subsequent analyses.
In contrast to the first four techniques, oxygen enhancement
techniques have been used extensively by the copper industry worldwide
and can lead to substantial increases in S02 concentration. Other
benefits, such as increased furnace throughput and decreased energy
consumption, also result. Hence, these techniques are considered as
possible control schemes in subsequent analyses.
A summary of experience involving the use of oxygen in reverbera-
tory furnaces is presented in Table 4-12. The largest reported increases
in S02 concentration and productivity (in full-scale demonstrations)
have resulted from the use of 10 to 12 oxygen-fuel burners that combust
the fuel with industrial grade (97 percent pure) oxygen. Increases in
S02 concentration by a factor of 4 to 5 are noted. Oxy-fuel burners
can be adapted to both green-charged furnaces, and to calcine-charged
furnaces employing sidewall charging.* The use of full oxy-fuel
*Based on the discussion in Section 3.4.3.5.4, there appears to be no
technical reason why roof-mounted oxy-fuel burners could not be used
on reverberatory furnaces charged with Wagstaff guns. However, some
of the data indicate possible engineering and production problems
that may preclude such usage under some conditions. Consequently,
for the purposes of this analysis, the use of oxy-fuel firing on
Wagstaff-charged furnaces is not considered fully demonstrated.
4-117
-------
TABLE 4-12. SUMMARY OF EXPERIENCE INVOLVING THE USE OF OXYGEN IN REVERBERATORY SMELTING FURNACES
Type of
technology applied
Oxy-fuel burners
(2 burners)
Full oxy-fuel firing
(12 burners)
Full oxy-fuel firing
(10 burners)
Oxygen-enriched pri-
mary combustion air
(25 percent 02)
Undershooting the
air-fuel flame with
pure oxygen or
oxygen-enriched air
! i Undershooting the
i — ' air/ fuel flame
0° with oxygen
>0xygen lancing of
the furnace bath
Oxygen-sprinkle
srnel ti ng
Smelter
Onahama
(green charge)
Caletones
(green charge)
Inco Copper Cliff
(Ni calcine
charge)
Almalyk
(green charge)
Inco Copper
Cliff (Ni
calcine
charge)
Rokana
(green charge)
Kennecott Copper
Corporation
(green charge)
Phelps Dodge-
Morenci
(green charge)
Nature of
application
Full-scale
demonstration
Full-scale
demonstration
Full-scale
demonstration
Full-scale
demonstration
Full-scale
tests
Full-scale
demonstration
Pilot-scale
tests
Smal 1-scale
tests
Reported S02 concentrations
Reported increase SO measurement
in production Without 02 enhancement With 02 enhancement location Reference
~ 21 percent 1.5 percent --b —c 27, 33, 35
~ 122 percent - 5.7 to 5.8 percentd — e 41, 42
(dry basis)
~ 45 percent 1.0 to 1.5 percent 5 to 6 percent Furnace offtake 44, 48
1 to 2 percent Downstream of
waste-heat
boiler
22 percent 3.4 percent 5.2 percent Furnace outlet 38
Up to 36 percent -- — — 34
18 percent --f --f -- 34
340 percent 1.9 percent 18 percent Furnace uptake 40
100 percent -- 20 to 30 percent -- 49
"Where possible, S02 concentrations are characterized as to a wet or dry basis of measurement. Values not explicitly categorized are assumed to be on a dry
basis.
Reported increase in S0? concentration was 0.3 percent per burner, or 0.6 percent overall. This implies a final nas-strpam concentration of 2.1 percent SO,,.
cNot reported explicitly, but was apparently after gas cleaning.
S02 concentrations of 1.3 percent, dry basis, were recorded for a 10-burner configuration smelting a different type feed.
Not reported, but Caletones indicated that the offgas is suitable for acid plant control.
Actual data not reported; however, a theoretical model developed at Rokana predicts 1.1 percent S02 (dry basis) in the furnace flue gases without oxygen
enhancement, and 2.0 percent S02 with 30 percent oxygen.
-------
firing has been reported to produce substantially increased levels of
NO in the furnace offgases but produces essentially no change in the
/\
furnace matte grade.
The use of the oxygen enrichment and oxygen undershooting schemes
leads to smaller increases in S02 concentration as compared to full
oxy-fuel firing because less oxygen is used. Gains of a factor of 1.5
to 2.0 in S02 concentration have been achieved. These two schemes can
be used on both green- and calcine-charged furnaces, irrespective of
the means of charging. It is expected that no changes in matte grade
would occur with oxygen enrichment and oxygen undershooting, because
these schemes are less extreme in terms of oxygen usage and flame
temperature than oxy-fuel firing.
Results from a pilot-plant study with roof oxygen lancing indicate
that S02 concentrations of 18 percent can be achieved. This value
represents roughly a 10-fold increase in S02 concentration. This
technique has been observed to increase significantly the furnace
matte grade. The increase in matte grade (i.e., desulfurization)
accounts in part for the high S02 concentration.
The oxygen-sprinkle smelting scheme is expected to produce the
highest S02 concentration of all of the oxygen enhancement techniques--
20 to 30 percent. These values represent a 15 to 20 fold increase in
concentration. Such increases are expected because oxygen sprinkle
smelting operates on the same principle as a flash furnace. Conse-
quently, substantial increases in matte grade may result with this
technique.
A review of literature pertaining to the well-established oxygen
enhancement techniques indicates that, with proper implementation,
refractory consumption on a per-unit-of-feed basis would not likely
increase. Furthermore, based on the extensive use of oxygen enhance-
ment techniques worldwide, refractory wear considerations are not
expected to hinder or limit the adoption of these techniques by the
domestic industry.
4-119
-------
4.5 GAS BLENDING
4.5.1 Converter Scheduling as a Means of Facilitating Gas Blending
The cyclic nature of converter operations must be considered in
the evaluation of any gas blending scheme involving converter offgases.
Because copper converting is a batch process, the total converter offgas
flow and associated composition will vary greatly over time unless
converter scheduling is adopted to ensure a continuous flow of S02
bearing offgas with as high an S02 concentration as possible. Thus,
when considering gas blending as a means by which to facilitate the
control of weak S02 streams as well as the autotherimal operation of
metallurgical sulfuric acid plants, the need for converter scheduling
becomes quite evident.
Appendix J presents the converter scheduling analysis that was
used to determine acid plant feed-stream flow and composition profiles
under various gas blending scenarios. The schedule is based on five
converter cycles daily using three converters.
4.5.2 Weak-Stream Blending as Applied to a New Smelter that Processes
High-Impurity Ore Concentrates
A smelter of this type would have the multihearth roaster-reverbera-
tory furnace-converter (MHR-RV-CV) configuration. Gas stream blending
as applied to this scenario might involve any of the following cases:
Case 1: Blending a portion of the weak stream with the
strong streams produced by the roasters and converters with
subsequent treatment in a dual-stage absorption sulfuric
acid p.lant.
Case 2: Blending the entire weak stream with the strong
streams produced by the roasters and converters, with subse-
quent treatment in a dual-stage absorption sulfuric acid
plant.
Case 3: Implementing oxygen enrichment of the primary
combustion air and blending the entire weak stream with the
roaster and converter strong streams, with subsequent treatment
in a dual-stage absorption sulfuric acid plant.
Case 4: Implementing oxy-fuel burners and blending the
entire weak stream with the roaster and converter strong
streams, with subsequent treatment in a dual-storage absorption
sulfuric acid plant.
4-120
-------
Case 5: Treating the entire weak stream in a limestone FGD,
while multihearth roaster and converter are blended for
processing in a dual-stage absorption sulfuric acid plant.
Case 6: Treating the entire weak stream in a Cominco NH3
FGD, with subsequent blending of the strong S02 stream from
the FGD, the roaster offgases, and the converter offgases
for processing in a dual-stage absorption sulfuric acid
plant.
Case 7: Same as Case 6, except a MAGOX FGD is used.
Case 1 (partial weak stream blending) is considered at a level
that would ensure autothermal acid plant operation at 3.5 percent S02)
at all times during which the converters are active. Thus, the acid
plant feed stream will not exhibit S02 concentrations below 3.5 percent
as long as there was some converter activity. However, at times when
there is no converter activity, supplemental heat will have to be
provided by the acid plant preheater since the S02 concentration in
the acid plant feed stream will be below the autothermal limit of
3.5 percent.
Cases 2 through 7 involve controlling the entire weak stream via
various blending schemes. Each scheme must be assessed in light of
the converter schedule presented in Appendix J in order to determine
its technical implications in terms of providing an acid plant feed
stream. Acid plant preheater operation will be required, however,
during any periods when the acid plant feed stream fails to exhibit an
S02 concentration in excess of 3.5 percent.
4.5.3 Partial Weak-Stream Blending as Applied to Existing Smelters
Existing facilities that undergo physical or operational changes
to achieve greater production capacity would not become subject to new
source performance standard (NSPS) requirements provided that postchange
or postexpansion emission levels at an existing facility do not exceed
preexpansion emission levels. For expanded reverberatory smelting
furnaces that produce a weak S02 offgas stream, one approach available
to maintain post-expansion reverberatory furnace S02 emissions at
preexpansion levels is partial weak-stream blending. Partial weak-stream
blending would consist of blending a sufficient portion of the post-
4-121
-------
expansion weak S02 stream with the strong S02 streams produced by the
roasters and/or converters, as well as the subsequent, treatment of the
resultant blended stream in a sulfuric acid plant. Depending upon the
scale of the expansion, the upgrading of existing acid plant capacity
or the installation of new acid plant capacity may be required.
Partial weak stream blending as applied to expanded reverberatory
furnaces might involve any of the following cases:
Case 1: Expansion at the reverberatory furnace (via adoption
of oxygen enrichment) for a smelter that processes a calcine
charge produced by multihearth roasters. A sufficient
portion of the reverberatory furnace offgas stream is blended
with the strong streams produced by the roasters and converters,
with subsequent treatment in a single-stage absorption
sulfuric acid plant.
Case 2: Expansion at the reverberatory furnace (via adoption
of oxygen enrichment) for a smelter that processes a calcine
charge produced by a fluid bed roaster. A sufficient portion
of the reverberatory furnace offgas stream is blended with
the strong streams from the fluid bed roaster and converters,
with subsequent treatment in a single-stage absorption
sulfuric acid plant.
Case 3: Expansion at the reverberatory furnace (via adoption
of oxygen enrichment) for a green-charged smelter. A suf-
ficient portion of the reverberatory furnace offgas stream
is blended with the strong stream from the converters, with
subsequent treatment in a single-stage absorption sulfuric
acid plant.
Case 4: Same as Case 3, except expansion is accomplished
via adoption of oxy-fuel burners, and a dual-stage absorp-
tion sulfuric acid plant is used to treat the blended stream.*
As indicated, all of the expansion options that might involve partial
weak stream blending involve oxygen enhancement. Since there is some
doubt as to the applicability of oxy-fuel burners to calcine-charged
reverberatory furnaces, only application to green-charged furnaces is
considered. The type of acid plant specified in each case was determined
from a survey of domestic primary copper smelter configurations so as
*Since a new converter (subject to NSPS) is required under this
scenario, a dual-stage absorption acid plant is required.
4-122
-------
to ensure that the cases outlined above would be typical of domestic
expansions involving oxygen enhancement.
4.6 PARTICULATE MATTER CONTROL FOR REVERBERATORY FURNACES
4.6.1 Important Factors Governing the Specification of a Particulate
Control Device for Reverberatory Furnace Off gases'
The nature of reverberatory furnace offgases as well as specific
considerations involved in the smelting process itself must be care-
fully considered before a particular class or type of control device
can be specified for application to reverberatory furnace offgases.
Among the more important considerations are the following:
The particle size distribution involved
The operating temperature of the control device and the
quantity of volatilized condensible material present in the
furnace offgas stream
The need to reprocess copper bearing dusts.
Reverberatory furnace offgases are known to contain substantial quan-
tities of particulate matter less than 10 pm in diameter at both
in-stack and out-of-stack temperatures.52 Some of the volatile com-
ponents that condense at out-of-stack temperatures form submicron-sized
fumes that are difficult to remove from the gas stream.53 Consequently,
a control device that efficiently removes particles in the submicron
range is required. For removal of particles in this size range,
high-energy scrubbers such as venturi scrubbers, fabric filters, and
dry ESP's provide the most efficient removal. This fact is illustrated
by Figure 4-16, a plot of typical collection efficiency versus particle
size for several particulate control devices, and Table 4-13, tabular
data relating particulate removal efficiency and particle size for
particles less than 10 urn in diameter.
The operating temperature of the control device at the point(s)
where particulate collection is to be effected is an extremely important
consideration as far as the control of particulate emissions from
reverberatory furnaces is concerned. Gases generated in a reverberatory
furnace exit the furnace at approximately 1,300° C (2,400° F). Generally,
the offgases are then passed through waste heat boilers to effect heat
4-123
-------
100
0
0
LEGEND:
A = HIGH-THROUGHPUT CYCLONE
B = HIGH-EFFICIENCY CYCLONE
C = SPRAY TOWER
6 8 10
PARTICLE DIAMETER, /urn
D = DRY ELECTROSTATIC PRECIPITATOR
E = VENTURISCRUBBER
F = FABRIC FILTER
12
Figure 4-16. Typical collection efficiency curves for several types of particulate removal devices.
-------
TABLE 4-13. TYPICAL FRACTIONAL COLLECTION EFFICIENCIES OF
PARTICULATE CONTROL EQUIPMENT54
Efficiency at given size (percent)
5 urn 2 |jm 1 urn
Medium-efficiency cyclone 30 15 10
High-efficiency cyclone 75 50 30
Electrostatic precipitator 99 95 85
Fabric filter 99.8 99.5 99
Spray tower 95 85 70
Venturi scrubber 99.7 99 97
4-125
-------
recovery, which in turn decreases the gas stream temperature to the
315° to 430° C (600° to 800° F) range. Gas streams in this temperature
range would obviously have to be cooled prior to being routed to a
venturi scrubber or fabric filter. While ESP's can undoubtedly with-
stand the temperatures involved., evidence indicates that, due to the
presence of condensibles that remain in the vapor phase in the 315° to
430° C (600° to 800° F) temperature range, ESP's that operate at or
near this temperature range will not remove the material that will
become particulate matter at out-of-stack temperatures. This is due
to the fact that ESP's will only remove materials that exist as solids
at the temperature of the gas stream. Some species, most notably
metallic oxides, may remain in the vapor phase until the gas stream is
vented to the atmosphere, at which point they condense and become
particulate matter. Thus, control efficiencies may be lowered sub-
stantially when the escape of condensible materials downstream of the
control device is considered. This fact is supported by in-stack/
out-of-stack test data obtained at several primary copper smelters.55
Thus, due to the presence of condensible substances, the temperature at
which the control device operates must be sufficiently low (90° to 110° C
[195° to 230° F]).53 The gas stream must be cooled from the 315° to
430° C (600° F to 800° F) temperature range down to the 90° to 110° C
(195° to 230° F) range if collection of the condensible material is to
be achieved. This degree of gas cooling may be accomplished by several
means, including:53
The use of radiative cooling towers
The use of spray towers or chambers (evaporative cooling)
Dilution with ambient air
The use of scrubbers and/or wet ESP's designed to accomplish
gas cooling and particulate removal.
Radiative cooling towers are typically used in lead smelter zinc
fuming operations as a cooling step between the waste heat boiler and
a baghouse. These towers normally consist of large U-shaped tubes
(about 1 m in diameter) that extend as high as 20 to 30 m.
4-126
-------
Spray towers that use water are employed in numerous industries
to cool and condition gas streams prior to gas cleaning in baghouses
and ESP's. Ambient air dilution involves mixing relatively cool
ambient air with the process stream to affect cooling. This approach
may increase the total volume of gases to be handled by a factor of 2
to 6, depending upon the temperatures of the gases involved. Consequently,
this form of gas-stream cooling is not recommended for cooling reverbera-
tory furnace offgases. Both gas cooling and particulate removal could
be effected in properly designed wet scrubber and/or wet ESP systems;
however, when a wet system for particulate removal is considered, the
difficulty involved in reclaiming the dusts from the liquid effluent
must be assessed.
The quantity of condensible material (feed impurities removed in
the furnace) will dictate the configuration of the particulate matter
control system. The amount of condensible material in the furnace
offgas stream will determine whether or not hot and cold control
devices placed in series will be required in order to separate copper-
bearing dusts from the condensibles. For instance, with low levels of
condensibles in the furnace offgases, it would not be necessary to
separate the copper-bearing dusts from the condensibles. Consequently,
cooling of the gas stream followed by a cold control device would be
adequate. However, for reverberatory furnace processing high-impurity
materials, a different situation could exist. The presence of relatively
large amounts of condensibles in the furnace offgases may necessitate
the use of a hot control device in series with a cold control device.
This configuration could be required so that an adequate portion of
the condensibles could be purged prior to dust recycle. The possible
need to purge a portion of the captured condensibles involves maintain-
ing favorable conditions for impurity elimination in the smelting
furnace. The impurity level above which hot-cold control is required
could be determined through a rigorous thermodynamic assessment of the
smelting process; however, such an analysis is beyond the scope of
this work.
The need to reclaim and reprocess the dusts in reverberatory
furnace offgases is another important factor that must be considered.
Recycling of dusts is generally practiced to recover valuable metals
4-127
-------
contained in the dusts. Consequently, particulate control methods
that facilitate dust reclamation are preferable.
While the gas-stream particle size distribution is always an
important consideration in the specification of a particulate control
device, the control device operating temperature at the point of
particulate collection and the need to reprocess recovered dusts are
two critical factors that must be considered in as far as potential
applications to reverberatory furnace offgases are concerned. The
following discussions on venturi scrubbers, fabric filters, and ESP's
evaluate each of these alternative technologies with respect to the
special considerations involved in the removal of particulate matter
from reverberatory furnace offgases.
4.6.2 Venturi Scrubbers
4.6.2.1 General Discussion. Like other wet collectors, venturi
scrubbers operate at variable collection efficiencies directly propor-
tional to the amount of energy expended.4 However, particles in the
0.1 to 20 urn range can be effectively removed with venturi scrubbers.54
Venturi scrubbers use a rectangular or circular flow conduit that
converges to a narrow throat section and diverges back to its original
cross-sectional area. When the gas stream enters the convergent
section, its linear velocity begins to increase and eventually reaches
a maximum in the throat area. The high-velocity gas stream tends to
atomize the liquid (usually water in the case of particulate removal),
which is injected into the stream via nozzles in the throat area. The
liquid droplets produced serve as targets for inertia! impaction of
the particles. Thus, good atomization is essential in providing sites
for inertial impaction. A typical venturi scrubber is illustrated in
Figure 4-17.
The typical water circulation rate to a venturi scrubber varies
from 0.3 to 1.6 £/m3/min (2 to 12 gal/103 cfm) of gas treated,54 which
is comparable to the water circulation rate required for other wet
particulate collectors. However, as compared with other dry and wet
collectors, pressure losses in venturi scrubbers can be quite large.
Venturi scrubbers are capable of attaining high removal efficiencies
4-128
-------
Dirty Gas Stream —.
Water
Clean Gas Stream
Figure 4-17. Venturi scrubber.
4-129
-------
of submicron particles at the expense of high pressure drops. Pressure
drops vary from 0.007 to 0.25 atm (3 to 100 in H20), depending upon
the collection efficiency desired;54 however, collection efficiencies
as high as 99 percent for submicron particles may be attained. Figure
4-18 depicts a typical relationship between fractional collection
efficiency54 and particle size for venturi scrubbers.
4.6.2.2 Application of Venturi Scrubbers to Reverberatory Smelt-
ing Furnaces. Properly designed venturi scrubbers could be used to
provide gas-stream cooling and efficient particulate matter removal in
applications to reverberatory furnace offgases. However, due to the
submicron size of some of the solid species involved, high pressure
losses will be required to effect efficient removal. Consequently,
since energy requirements are proportional to the pressure losses
incurred, energy costs may be high.
The treatment of the liquid effluent from venturi scrubbers must
also be considered. When dust reclamation is considered, venturi
scrubbers, by virtue of being wet collectors, would not prove to be as
convenient as dry collectors. This fact, coupled with potentially
high energy requirements, is probably the major reason why venturi
scrubbers have not found wide-spread application in the smelting
industry. Although the Phelps Dodge facility at Playas, New Mexico,
does use a venturi scrubber to remove particulate from gases originat-
ing in an electric stag cleaning furnace, most smelter applications of
Venturis have involved the removal of particulate from gas streams
bound for sulfuric acid plants. Venturi scrubbers are not currently
applied at any domestic smelter for the control of reverberatory
furnace particulate matter emissions.
4.6.3 Fabric Filters
4.6.3.1 General Discussion. Fabric filtration is one of the
oldest and most commonly used methods of effecting particulate removal
from gas streams.4 54 Fabric filters are typically used for high
efficiency (>99 percent) particulate removal.
Filters must be constructed of materials compatible with character-
istics of both the carrier gas and the particulate to be collected. A
4-130
-------
99.9
0.1
0.2
0.3 0.4
0.6 0.8 1
PARTICLE SIZE,
5678 10
Figure 4-18. Typical relationship between fractional collection efficiency and
particle size for venturi scrubbers.54
-------
wide range of filtering materials, including woven or felted fabric,
is used. Fabrics may be natural or synthetic, depending upon the
nature of the gas stream to be treated. Among the materials currently
in common use are cotton, nylon, fiberglass, polyesters, and aromatic
polyamides.54
Particulate matter is removed from the gas stream by impingement
on or adherence to the fibers. The filter fibers are normally woven
with relatively large open spaces, sometimes 100 |jm or larger across.
Consequently, the filtering process is not one of simple fabric sieving,
as evidenced by the fact that high collection efficiencies have been
achieved for dust particles with a diameter of 1 pm or less. Small
particles are initially captured and retained on the fiber of the
fabric by direct interception, inertial impaction, diffusion, electro-
static attraction, and gravitational settling. Once a mat or cake of
dust is accumulated, further collection is accomplished by mat or cake
sieving and, to a small extent, by the above mechanisms. Periodically,
the accumulated dust is removed, but some residual dust remains and
serves as an aid to further filtering.
One of the major operating characteristics of fabric filters is
the requirement that they be cleaned frequently to prevent excessive
pressure drops. Several means of cleaning the filter bags have been
devised, and filters are generally designed with ease of cleaning in
mind. The most common methods of cleaning are mechanical vibration or
"shaking," pulse jets, and reverse air flow. Figures 4-19, 4-20, and
4-21 present schematics of baghouses with each of the above-mentioned
cleaning systems. Cleaning is responsible for a major portion of the
filter degradation that occurs over time. Thus, the frequency of
cleaning must be determined as a tradeoff between higher operating
costs resulting from increased pressure drops and filter replacement
costs resulting from more freque^ cleaning.
In general, fabric filters must provide a large surface area per
volume of gas to be cleaned. The reverse of this ratio is known as
the air-to-cloth ratio. Optimum values normally range from 0.5 to
1.5 cm/s (1 to 3 ft/min) for shaker or reverse-air baghouses and from
4-132
-------
Shaker
Clean Gas
Stream
Particu late-Laden
Gas Stream
Filters
Collection
Hopper
*- Dust to Disposal or Recycle
Figure 4-19. Baghouse with mechanical shaking.
4-133
-------
[Flapper valve
Clean gas rN^T^/n
stream
Springs
I
5
I
a
F
a
Reverse
air fan
Particulate-laden-
gas stream
XT
Ash to disposal
or recycle
Figure 4-20. Baghouse with reverse flow cleaning.54
Clean gas
stream
Venturi
Compressed
air header
Particulate-laden
gas stream
\
TT
Ash to disposal
or recycle
Figure 4-21. Baghouse with cleaning by jet pulse."
4-134
-------
about 1.5 to 4.0 cm/s (3 to 8 ft/min) for pulse-jet baghouses. The
air-to-cloth ratio required for a given application may in turn require
that the baghouse be quite large.
4.6.3.2 Application of Fabric Filters to Reverberatory Smelting
Furnaces. With regard to smelters, baghouses are generally chosen as
the control device when the S03 concentration and chloride content of
the effluent gases are low.4 High S03 concentrations are known to
produce corrosion and deterioration of both the baghouse structure and
the filter fabric. If chlorides are present in the effluent gases,
they may tend to produce hygroscopic effects on the fabric filters.
Copper, zinc, and lead chloride act as desiccant materials and may
produce a sticky material that tends to blind and eventually tear the
filter fabric. Reverberatory furnace effluents may contain any of the
above-mentioned chemical species. The gas-stream temperature is the
primary factor that governs the extent to which these species are
formed in the gas stream.4 These species would likely exist at the
lower temperatures (90° to 110° C [195° to 230° F]) required for
effective particulate matter removal. Consequently, the design of the
baghouse and ancillary equipment will have to account for the increased
corrosion potential. Corrosion can be greatly reduced by the use of
appropriate construction materials and proper insulation of all flues,
baghouse structures and hoppers. While the presence of metal chlorides
creates the potential for blinding of the filter media, experience
with similar offgas streams from other sources (including an electric
smelting furnace) within copper smelters has not produced problems of
this type.53
It is important to note that, while the gas stream must be cooled
to the 90° to 110° C (195° to 230° F) temperature range prior to
particulate removal, this degree of cooling will probably not cause
the gas stream to reach its dew point. The dew point of reverberatory
furnace offgas has been reported to be in the 28° to 46° C (82° to
115° F) range,56 which is substantially lower than the temperature
range required for efficient particulate removal.
4-135
-------
Fabric filtration has never been used by the domestic primary
copper industry to control particulate matter emissions from reverbera-
tory furnaces. However, fabric filters have been used to control
particulate matter emissions from gases that originate in fluid-bed
roasters, multihearth roasters, electric furnaces, and converters.
EPA has tested a fabric filter used to remove particulate matter from
a gas stream composed of offgases from a fluid-bed roaster, an electric
furnace, and several converters. These tests were conducted at the
Anaconda Smelter in Anaconda, Montana, which subsequently shut down in
1980. Evaporative cooling was used to cool the gases prior to their
entry into the baghouse. This was accomplished by routing the gases
through a spray chamber. The fabric filter was tested at approx-
imately 100° C (215° F) and processed about 4,670 dscnt/min (-164,920
dscf/min) of gas.55 As shown in Table 4-14, the test results indicated
that this device was achieving a particulate removal efficiency of
99.7 percent, which resulted in a mass emission rate of 13.1 kg/h
(~29 Ib/h) to the atmosphere. (See Appendix C for details.)
There is no doubt that the offgases processed by the aforementioned
fabric filter at the Anaconda smelter contained significant quantities
of metallic oxides and SQ3, thus creating the potential for the formation
of metal chlorides. This was particularly true for the Anaconda
smelter since "dirty" concentrates were processed. Thus, when factors
that can hinder baghouse performance are considered, the parallel
between the previously mentioned offgas stream at Anaconda and the
typical reverberatory furnace offgas stream is evident. Consequently,
because the application at Anaconda was quite successful, fabric filters
are considered technically viable means by which to remove particulate
matter from reverberatory furnace offgases. The anticipated overall
particulate matter removal efficiency would be the demonstrated effi-
ciency of 99.7 percent.
4.6.4 Electrostatic Precipitators
4.6.4.1 General Discussion. Electrostatic precipitation has
played an important role in industrial gas cleaning since the original
development work by F. G. Cottrell in 1910. ESP's are capable of
4-136
-------
TABLE 4-14. SUMMARY OF PARTICULATE TEST DATA FOR THE SPRAY CHAMBER/BAGHOUSE
AT THE ANACONDA SMELTER
Run Temperature, °C (°F)
1
2
3
Average
281
288
302
290
(538)
(550)
(575)
(554)
Inlet
Outlet
Grain loading,
mg/Nnt3 (gr/dscf)
14,735
13,636
14,048
14,140
(6.44)
(5.96)
(6.14)
(6.18)
Mass
kg/h
4,069
3,751
3,827
3,882
rate, Grain
(Ib/h) Temperature, °C (°F) mg/Nm3
(8,951)
(8,253)
(8,419)
(8,541)
50.3
37.1
52.2
46.5
loading,
(gr/dscf)
(0.0220)
(0.0162)
(0.0228)
(0.0203)
Mass
kg/hr
14.6
10.0
14.6
13.1
rate,
(Ib/h)
(32.1)
(22.0)
(32.2)
(28.8)
Removal
efficiency,
percent
99.64
99.73
99.62
99.7
I
1—>
GO
-------
achieving high collection efficiencies on particles that range from
0.05 to 200 urn in diameter;54 thus, use of ESP's would be effective in
applications where a substantial portion of the particles to be col-
lected are in the submicron range. Generally, ESP's can be shown to
be more financially attractive (in comparison to fabric filters) as
the dust resistivity approaches the optimum range (10* to 1010 ohm-cm),
or as the volume of gas to be handled increases.
Particulate matter collection by electrostatic precipitation is
based upon the fact that particles of one electrical charge experience
an attraction to an electrode of opposite polarity. Separation of
suspended particulate matter by electrostatic precipitation requires
three basic steps:4
Electrical charging of the suspended matter
Collection of the charged particles on a grounded surface
Removal of the collected matter to an external receptacle.
A charge may be imparted to the particulate matter prior to
gas-stream entry into the ESP by either flame ionizat^on or friction;
however, the bulk of the charge is applied by passing the suspended
particles through a high-voltage, direct-current corona. The corona
is established between an electrode maintained at high voltage and a
grounded collecting surface. Particulate matter that passes through
the corona is subject to an intense bombardment of negative ions that
flow from the high-voltage electrode to the grounded collecting surface.
The particles thereby become highly charged within a fraction of a
second and migrate toward the grounded collection surface. A typical
ESP is illustrated in Figure 4-22.
After the particulate matte1^ deposits on the grounded collecting
surfaces, adhesive, cohesive, and primary electrical forces must be
sufficiently strong to resist any action or counter-electrical forces
that would cause reentrainment o~c the particulate matter. The particu-
late matter is dislodged from the collecting surfaces by mechanical
means such as vibrating with rappers or flushing with liquids. The
collected materials fall to a hopper, from which they can be reclaimed
for recycle or disposal.
4-138
-------
Clean
Gas Stream
High Voltage
Electrodes
-Fi
I—»
oo
Participate-Laden
Gas Stream
Grounded
Plates
Collection
Hopper
Figure 4-22. Electrostatic precipitator.
-------
Perhaps one of the most important properties of the participate
matter in relation to electrostatic precipitation is the electrical
resistivity of the material to be collected. The resistivity of
industrial dusts may vary from 10 3 to 1014 ohm-cm.54 If the resis-
tivity of the material to be collected is too low (<104 ohm-cm),
collected particles may not retain an electrostatic charge suffi-
ciently high to keep them firmly attached to the collecting surfaces,
thus allowing some of the collected material to become reentrained in
the gas stream. On the other hand, particulate matter with a resistivity
of greater than 1010 ohm • cm can cause precipitator collection effi-
ciency to suffer.54 When dust resistivity is high, a large portion of
the total voltage drop between the high-voltage electrodes and the
collecting plates actually occurs across the dust layer, which in turn
reduces the total corona power available to ionize and to charge the
particles in the gas stream. Electrostatic precipitation is most
effective in collecting dusts that exhibit resistivities in the range
of 104 to 1010 ohm-cm.
When the resistivity of the dusts to be controlled is not appropri-
ate for electrostatic precipitation, means exist by which to alter the
resistivity in such a way that the dusts become amenable to removal by
electrostatic precipitation. Resistivity is a strong function of both
gas stream temperature and humidity; thus, by appropriate manipulation
of these parameters, the resistivity of some dusts can be altered so
that efficient removal by electrostatic precipitation becomes feasible.
Another means by which to achieve electrical resistivities in the
desired range is the addition of conditioning agents to the gas stream.
Currently, S03 and NH3 are the only conditioning agents that are
technically and economically feasible in commercial practice.54
Ammonia or S03, when added to a gas stream in small amounts, act as
electrolytes when adsorbed on the deposited dust particles. This in
turn causes a marked reduction in resistivity.
4.6.4.2 Application of Electrostatic Preci'pitators to Reverbera-
tory Smelting Furnace Effluents. ESP's have been used by the domestic
primary copper industry for several years to control offgases from
4-140
-------
reverberatory furnaces as well as offgases from roasters and converters.
ESP's have several characteristics that make them particularly attrac-
tive for the reclamation of smelter dusts. ESP's are capable of
handling very large volumes of gas and can also easily reclaim valuable
dusts. The ability of ESP's to exhibit high collection efficiencies
on fine particles is also quite attractive, especially where the
control of reverberatory furnace offgases is considered. However, as
mentioned in Section 4.5.1, ESPs will only remove materials that exist
as solids at the gas stream temperature. In-stack/out-of-stack test
data presented in Table 4-15 support the fact that ESP removal effi-
ciencies are lowered substantially when the escape of condensible
species is considered. For example, in-stack measurements on the
reverberatory furance ESP at Phelps Dodge-Ajo showed that, at 315° C
(600° F), the ESP was achieving a 96-percent particulate removal
efficiency. However, out-of-stack measurements obtained at 120° C
(250° F) indicated that this ESP is less than 50 percent efficient
because of the vaporized metallic oxides that pass through the ESP and
subsequently condense. The data in Table 4-15 indicate that a sub-
stantial portion of the emissions generated remain in the vapor phase
at the temperature of the gas stream and thus are not removed by the
ESP. Thus, the need for control device operation in the 90° to 110° C
(195° to 230° F) temperature range is substantiated. ESP operation in
this temperature range is feasible, although the potential for certain
problems does exist. Some of the compounds that exist in the gas
streams in the 90° to 110° C (195° to 230° F) temperature ranges could
form a sticky mass that might adhere to ESP plates and thus hinder
efficient performance. As noted previously, however, experience with
similar smelter offgas streams has not produced this problem. Another
problem may occur if the resistivity of the particles involved becomes
too high at the lower temperatures involved. Tests have indicated,
however, that reverberatory furnace dusts do not exhibit high resis-
tivities at low ESP operating temperatures.53 Thus, dust resistivity
should not preclude dry ESP operation in the 90° to 110° C (195° to
230° F) temperature range.
4-141
-------
TABLE 4-15. SUMMARY OF IN-STACK/OUT-OF-STACK PARTICULATE MATTER
TEST RESULTS AT REVERBERATORY FURNACE ESP OUTLETS55
Smelter/test contractor
Magma, San Manuel, AZ/
EPA, NEIC
Phelps Dodge,6 Ajo, AZ/
Radian
Phelps Dodge, Ajo, AZ/
Aerotherm
Kennecott, Hayden, AZ/
Aerotherm
No. of
test
runs
4
3
2
1
Temper-
ature,
°C
280
315
290
145
In-stack
partic-
ulate
loading,
mg/dscm
230
95
70f
30f
Out-of-
stack
partic-
ulate ,
loading
at
120° C,
mg/dscm
1,030
2,335
1,7309
30g
Total
partic-
ulate
loading
at
120° C,
mg/dscm
1,260
2,430
1,800
60
Includes nozzle wash and in-stack filter.
Includes probe wash, back half of the in-stack filter holder wash, front
half of the out-stack filter holder wash, and the out-stack filter.
Includes in-stack and out-of-stack particulate (does not include impinger
catch).
Single-point sample.
Incomplete traversing.
Does not include nozzle wash. Therefore, these results are somewhat lower
than the actual loading.
^Includes nozzle wash. Therefore, these results are somewhat higher than
the actual loading.
4-142
-------
While there is no cold ESP that processes reverberatory furnace
offgases only, one situation does exist, at the ASARCO-E1 Paso Smelter,
where multihearth roaster and reverberatory furnace offgases are first
cooled by evaporative cooling in a spray chamber and then routed to an
ESP. EPA has tested this ESP to ascertain overall particulate matter
removal efficiency.57 The test data are presented in Table 4-16.
Analysis of the test results yielded an average overall particulate
matter removal efficiency of 96.6 percent. The resultant emissions to
the atmosphere were approximately 37.2 kg/h (81.9 Ib/h).
There is no doubt that the reverberatory furnace offgases at
ASARCO's El Paso smelter contain significant quantitates of metallic
oxides and S03, thus creating the potential for the above-mentioned
problems. However, these problems were not encountered. Consequently,
cold ESP operation on reverberatory furnace offgases is considered to
be technically feasible with an expected overall particulate matter
removal efficiency of 96.7 percent.
4.6.5 Conclusions Regarding Particulate Removal from Reverberatory
Furnace Offgases
Since a substantial portion of the particles to be removed from
reverberatory furnace offgases are in the submicron range,53 the
consideration of control devices must be limited to those which can
effectively remove particles in this size range. For this reason, the
discussions in Section 4.5 have been limited to high-energy (venturi)
scrubbers, fabric filters, and ESP's.
Venturi scrubbers could be used to effect the required gas cooling
and particulate removal; however, large pressure drops would be required
to remove the smaller particles. Consequently, energy requirements
would probably be high. In addition, because copper smelter dusts are
generally recycled, the use of venturi scrubbers would necessitate the
recovery of the dusts from the liquid effluent. In light of these
factors, venturi scrubbers are dismissed from consideration for the
control of particulates in reverberatory furnace offgases.
Because the control device must operate in the 90° to 110° C
(195° to 230° F) temperature range, fabric filters appear to be quite
4-143
-------
TABLE 4-16 SUMMARY OF PART1CU1ATF TF.SI DATA FOP THE SPRAY CHAMRER/ROASFFR-REVLRRFRA10RY
ESP AF IliE ASARCO-a. PASO SMELTER
Inlet ^utjet
— " Removal
Grain loading, Mass rate, firain loading, Miss rate efficiency
Run lompefature. °C (°F) ing/Mm" (gr/dscf) kg/h (Ib/h) Temperature, °C ("F) mg/Nm-' (gr/dscf) kg/hr (Ib/h) percent
~l 22G (439)4~530 (1.98) U>2F (?,2n7) 104 (220) 111 (0.0486) 40.8 (89.7) 96.0
2 ?31 (448) 5,651 (2.47) 1,236 (2,719) 104 (220) 85 (0.0372) 33.7 (74.1) 97.3
3
Averaqc 229 (444) 5,091 (2.23) 1,129 (2,483) 104 (220) 98 .^.^^JZlLJ-™----^96'7-
-------
feasible for removing participates from reverberatory furnace offgases.
Construction material would have to be judiciously selected, however,
to minimize corrosion that might occur due to species that may condense
at the low temperatures involved. Nonetheless, cooling to the 90° to
110° C (195° to 230° F) temperature range should not cause the gas
stream to reach its reported dew point of 28° to 46° C (82° to 115° F).
Significantly, a "cold" fabric filter has indeed been used to control
a gas stream composed of gases from a fluid-bed roaster, an electric
smelting furnace, and several converters (see Section 4.5.3.2). This
device was tested by EPA and exhibited a 99.7-percent overall capture
efficiency.55 Corrosion and blinding of the filters were not identified
as extensive problems in this application which occurred at the Anaconda
Smelter in Anaconda, Montana. Also, as suggested by the data presented
in Figure 4-16 and Table 4-13, fabric filters would provide the greatest
collection efficiencies for particles less than 10 urn in diameter.
Cold ESP's are also technically viable means by which to remove
particulate from reverberatory furnace offgases. EPA tests indicate
that the resistivity of the dusts involved in the 90° to 110° C (195°
to 230° F) temperature range is not too high for effective ESP operation.
While no test data exist that characterize cold ESP operation on
reverberatory furnace offgases alone, inlet and outlet ESP test data
are available for a cold ESP used to treat offgases from the multihearth
roasters and reverberatory furnaces at the ASARCO-E1 Paso smelter.
Test results based on total particulates indicate that this ESP was
operating with a 96.7 overall capture efficiency.58 The temperature
in evidence at the ESP outlet was 104° C (220° F)—well within the
range required for the condensation of metallic oxides. Consequently,
it would be reasonable to assume that ESP operation on reverberatory
furnace gases alone in the 90° to 110° C (195° to 230° F) range can
result in a overall efficiency of about 96.7 percent.
4.7 CONTROL OF FUGITIVE EMISSIONS FROM PRIMARY COPPER SMELTERS
4.7.1 General
Fugitive emissions may be characterized as emissions from material
transfer operations, process vessel leakage, and primary flue leakage
4-145
-------
that escape directly to the atmosphere. Capture of fugitive emissions
is effected by either local or general ventilation techniques. Once
captured, fugitive emissions may be routed directly to a control
device, or they may be combined with primary process offgases prior to
treatment in a control device. In many cases, they are simply routed
to a stack for dispersion without any type of prior particulate or
vapor removal. However, even though captured fugitive emissions may
not be subjected to particulate matter or S02 removal, high-level
dispersion of these emissions will result in lower ambient concentra-
tions of these pollutants at the ground level near the smelter.
Consequently, the ambient air quality would be improved even though
mass emissions of particulate matter and S02 to the atmosphere would
remain the same.
Fugitive emissions from some sources may be minimized or eliminated
by minor process changes and/or good operating and maintenance practices.
In other cases, add-on controls are required. Section 3.3 contains a
detailed discussion of fugitive emissions sources within a primary
copper smelter.
4.7.2 Local Ventilation
Local ventilation systems consist of localized hoods or enclosures
designed to confine and capture fugitive emissions at the source.
These systems use induced air currents to divert fugitive emissions
into an exhaust duct.
In this discussion, the term "hood" is used in a broad sense to
include all suction openings, regardless of shape or physical character-
istics. The design of a local exhaust hood involves the specification
of its shape and dimensions, its position relative to the emission
point, and its rate of air exhaust. In the design of local exhaust
hoods, an attempt is made to create a controlled air velocity that
will prevent the escape of fugitive emissions from the controlled area
to the surrounding atmosphere. The exhause rate at the hood entrance
is dependent upon the air velocity required to prevent the escape of
emissions. The air velocity that will just overcome the dispersive
motion of the contaminant(s), plus a suitable safety factor, is called
4-146
-------
the "capture velocity." The capture velocity must be high enough so
that particles at the most distant null point will be captured.
Emissions are generally released from the source with a considerable
velocity; however, momentum is soon lost, which results in a rapid
decrease in velocity. The position at which the fume velocity is
approximately zero is called the null point. Figure 4-23 illustrates
the formation of null points as the fume rises from the emission
source. If an adequate velocity toward the hood is provided at the
most distant null point from the hood, the majority of the fume will
be captured.
Wark and Warner54 report that a velocity of less than 30 m/min at
a null point seldom can be tolerated without a marked loss in capture
hood effectiveness. The optimum capture velocity depends upon the
following factors:54
The size and shape of the hood
The position of the hood relative to the emission source
The nature and quantity of the fume to be captured.
For an exhaust hood to be effective, the exhaust rate across the
space between the emission source and the hood must be sufficient to
entrain all of the emissions. Proper hood design must incorporate
allowances for indoor air currents that could deflect the emissions
away from the hood. A capture hood designed to work in a still
atmosphere may be completely ineffective in the presence of indoor air
currents. In addition, the design should be such that the exhaust
rate is as uniform as possible over the entire plane of the hood
inlet.58
In the case of fugitive emissions from hot sources associated
with primary copper smelters, the hood design and the specified ventila-
tion rate must account for the thermal draft that results from heat
transfer from the source to the surrounding air. The hood design must
accommodate not only the volume of fugitive gases to be collected but
also surrounding air set into motion by convective currents. In
addition, hoods should be placed as close to the emission source as
practical to enhance pollutant recovery.
4-147
-------
Null Points
Capture Hood
» ! / /
Emission Source
Figure 4-23. Illustration of null point formation.
4-148
-------
Air curtains can be used to complement local ventilation systems.
A detailed discussion of air curtains is presented in Section 4.7.6.3.
Several fugitive emission sources associated with primary copper
smelters can be controlled by local ventilation methods. Among these
sources are:58
Calcine discharge and transfer from multihearth roasters
Matte tapping
Slag tapping
Converter operations
Anode furnace operations.
4.7.3 General Ventilation
General ventilation is normally required when it is not possible
or expedient to use local exhaust hoods. Local hoods may handicap the
operation, maintenance, or surveillance of a process or a piece of
process-related equipment, in which case general ventilation would
become the preferred method of fugitive emissions control.
General ventilation has historically taken the form either of
natural air changes caused by wind and, possibly, convective air
currents or of mechanically assisted air change. Natural air changes
throughout a building can occur by either of two mechanisms:
The force of natural wind currents through windows or other
openings in the building
The force of convective air currents that occur due to
temperature gradients that exist between the inside of the
building and the surrounding environment.
Mechanical ventilation is induced by motor-driven fans and is used
when the emissions cannot be removed by natural ventilation.
Ventilation requirements for buildings are generally defined in
terms of the number of air changes required per unit of time. Although
essential in determining the ventilation requirements, the air change
rate, also referred to as the ventilation rate, is not the only factor
that must be considered. The evolution rate of emissions within a
4-149
-------
building must also be considered, as must other site-specific charac-
teristics such as heat sources, building configuration (number of
spans and the form and shape of the roof), and arrangement of venti-
lation openings (windows, roof bays, etc.). Proper ventilation entails
a delicate balance between mechanical and naturally occurring forces
within a structure.
4.7.4 Control of Fugitive Emissions From Roasting Operations
As suggested by the discussion of fugitive emission sources in
Section 3.3, calcine discharge and transfer are the only significant
sources of fugitive emissions from multihearth roaster operations.
Fluid-bed roasters are designed in such a manner that fugitive emissions
are virtually eliminated;58 therefore, this discussion will primarily
address the control of fugitive emissions from calcine discharge and
transfer operations associated with multihearth roasters.
Four domestic copper smelters currently use multihearth roasters:
ASARCO-E1 Paso
ASARCO-Hayden
ASARCO-Tacoma
Phelps Dodge-Douglas.
Calcine produced at these smelters is normally discharged from
the bottom of the multihearth roaster into a hopper, which in turn
distributes the calcine to a transfer vehicle (larry car) for trans-
portation to the smelting furnace(s). Calcine hoppers are discharged
intermittently rather than continuously. More than one hopper may be
discharged during the transfer of calcine to the larry cars for trans-
port to the smelting furnace(s). The frequency of discharge for any
one hopper may vary from zero to 30 times per hour.58 Typically, the
duration of discharge is approximately 30 to 60 seconds per hopper.
Large quantities of dust are generated as a result of material move-
ment and pressure changes within the transfer vehicle. Consequently,
local ventilation is needed to control emissions at the transfer
point, and the transfer vehicle feed opening must be covered to prevent
the escape of emissions during transport.
4-150
-------
The design and effectiveness of these systems vary from smelter
to smelter. Generally, the captured fugitive emissions are routed to
existing process control systems for particulate removal.
Figure 4-24 presents a schematic of the calcine transfer/fugitive
emissions control system at ASARCO-Hayden. A similar system is used
at ASARCO-Tacoma. A continuous, flat apron strip nearly 0.6 m (2 ft)
wide is mounted directly below the row of multihearth roasters at the
hopper discharge gate. In the apron below each hopper two ports are
connected to a duct 0.5 m (1.5 ft) wide. A matching leaf spring-loaded
flat apron is mounted on the larry car. When the larry car is driven
beneath the roaster, it is positioned so that the matching apron on
the car is directly aligned with the apron on the bottom of the roaster
hopper. Consequently, the ports in the hopper apron are perfectly
aligned with ports on the larry car apron. One port is used for
transferring the calcine to the larry car, while the other two ports
are used for ventilation. Each port has its own individual fan* rated
at approximately 140 NirrVmin (5,000 scfm). Emissions captured by this
system are routed to the roaster primary offgas flue.
At ASARCO-E1 Paso, calcine transfer occurs in a long shed. The
shed is open at one end for larry car entry, and an exhaust duct is
located at the opposite end. The captured emissions are exhausted
into the spray chamber system that handles the primary process emissions
from the reverberatory smelting furnaces. A visual inspection of this
facility indicated that about 50 percent of the visible emissions are
captured.58
At Phelps Dodge-Douglas, hooding with canvas flaps is provided
around the roaster discharge area into which the larry cars are driven.
Captured emissions are routed to a baghouse for particulate removal.
The average volumetric flow rate at the inlet of the baghouse used to
treat these emissions has been measured to be approximately 990 NmVmin
(35,000 scfm). A visual inspection indicated that about 70 percent of
the visible emissions are captured by this system.58
"Only a single fan is used at ASARCO-Tacoma.
4-151
-------
01
ro
12%"
NOTE: Car Top and Hood - 18 Ga. C.R.S.
I-KONTELEV.
SIDE ELEV.
Figure 4-24. Spring-loaded car top and ventilation hood, ASARCO-Hayden.58
-------
From an alternative control standpoint, the ASARCO system applied
at Tacoma and Hayden seems to be the most viable because it is very
effective in controlling visible fugitive emissions. This determi-
nation has been made based upon visual observations of the various
systems now in use for the control of fugitive emissions from calcine
discharge operations.
Although visible emissions from calcine discharge operations are
effectively controlled as discussed above, emissions from calcine
transfer operations are generally poorly controlled. Visible emis-
sions from moving larry cars at ASARCO-E1 Paso and ASARCO-Hayden were
evident.59 60 In addition, the odor of S02 was easily detected from
the larry car emissions observed at the El Paso smelter.59 Larry car
covers should be used while the cars are in transit to minimize fugitive
emissions.
4.7.5 Control of Fugitive Emissions From Smelting Furnace Operations
A complete discussion of the fugitive emissions sources associated
with various types of smelting furnaces is presented in Section 3.3.
From this discussion, it becomes evident that matte tapping and slag
skimming operations are the most significant easily controllable
sources of fugitive emissions from smelting furnaces. Consequently,
this discussion addresses the control of fugitive emissions from
tapping and skimming operations only.
4.7.5.1 Control of Fugitive Emissions from Matte Tapping Operations.
Matte tapping is the operation by which matte is removed from the
smelting furnace for transport to the converters. Matte is removed
(tapped) from the furnace through tapping ports, which are generally
located on the sides of the furnace. The number and location of the
tapping ports will vary from furnace to furnace depending upon the
size and type of smelting furnace; however, the tapping procedure is
generally the same. Normally, matte is tapped from one port at a time
and conveyed through launders into ladles. The ladles vary in volume
from 5 to 9 m3 (175 to 325 ft3). A single matte tap may last from
9 to 15 minutes, with emissions in evidence from the point at which
the matte exits the furnace to the point where it settles into the
ladle.
4-153
-------
Most smelters employ local exhaust hoods at the points where the
matte exits the furnace. In addition, launder covers are used at most
facilities. Tap port exhaust hoods may be of any shape as long as
they are designed to evacuate as much of the emission area as possible.
Hoods of this type are generally affixed to the side of the furnace.
Observations indicate that, when operated properly, hoods of this
nature are quite effective in capturing fugitive emissions generated
in the area of the tapping port. In addition, these hoods are quite
effective in capturing heavy emissions that occur during the lancing
required to open the ports. A typical hood of this type is illus-
trated in Figure 4-25.
A schematic of a matte tapping fugitive emissions control system
is presented in Figure 4-26. The system illustrated Is employed at
the ASARCO-Tacoma smelter. The actual matte tap hoods are 1.2 m by
1.2 m (4 ft by 4 ft) in cross section and are located less than 0.9 m
(3 ft) above the tapping port.58 Each matte tap hood is connected to
the main fugitive emissions duct as shown. The ducts, which connect
each hood to the main fugitive emissions duct, are 0.6 m (2 ft) in
diameter; the main duct is 1.2 m (4 ft) in diameter. During a tap,
approximately 280 NmVmin (10,000 scfm) are exhausted from a given
matte tap hood.
Launder covers are usually made of metal and are mounted on the
launders in sections to allow manual removal for launder cleaning. A
typical section of a launder hood is 1.2 to 1.5 m (4 to 5 ft) long.58
Launder covers of this type are depicted in Figure 4-27. A great deal
of the emissions that are generated by the molten matte as it flows
down the launder can be captured by launder covers if the covers are
well maintained and in place during tapping operations. Due to the
incline of the launder, hot fumes captured by the launder covers will
generally rise back to the tapping port area where they are captured
by the tap port hood.
An effective type of launder hood has been developed by the
Phelps Dodge Corporation and is currently in use at Phelps Dodge's
Morenci smelter. This type of hood, as depicted in Figure 4-28, is
4-154
-------
cn
en
»* To Fugitive Gas
Handling System
Front View
Side View
Figure 4-25. Typical hooding for a matte tapping port.
-------
en
cr>
To Baghouse
50"
dia
11' dia - Ladle Hood
(Movable)
40' dia X 3/16" Thick
7
Matte Tap
Hood
<3'8" X 3'8")
50" dia X 3/16"
Reverberatory
Furnace
35'
Figure 4-26. Schematic of a typical fugitive emissions control system
for matte tapping operations.58
-------
Launder •
Front View
Launder Cover
4-5'
J_
•-T-
I
Sections of Launder Cover
Top View
Figure 4-27. Typical sectional launder covers.
Tapping
Port Hood
4-157
-------
Rail
Support
en
CO
To Main
Fugitive Emissions
Duct
Launder Hood
Support
Observation
Port
P
Launder
Figure 4-28. Launder hoods utilized at the Phelps Dodge—Morenci Smelter for the
capture of fugitive emissions generated during matte tapping operations.
-------
movable and is placed over the launder during tapping operations. The
ventilation rate for each hood is approximately 150 NmVmin (5,400 scfm).
Visual observations yielded an estimated capture efficiency of 90 percent.
This type of hood can be effective if the surrounding atmosphere is
calm; however, if the air currents within the furnace building are
strong, a major portion of the emissions will be blown out from under
the hood.
As suggested by Figure 4-26, movable hoods can be employed to
capture emissions that occur at the ladle. Figure 4-29 illustrates
the type of ladle hood that is used at the ASARCO-Tacoma smelter. The
dimensions of the matte tap hoods used at the Tacoma facility are also
presented in Figure 4-29. Emissions that are captured by the ladle
hooding are evacuated to the main fugitive emissions duct via the
102-cm (40-in.) diameter offtake, as illustrated. The ventilation
rate at each ladle hood is about 570 NmVmin (20,000 scfm). The ladle
hood is retractable and is lowered into place over the ladle just
prior to tapping. The ladle hood is lowered and raised via use of a
cable and winch.
As indicated by Figure 4-26, the matte tapping fugitive emissions
capture system at the ASARCO-Tacoma facility employs both tap port
hoods and a ladle hooding system. Emissions testing and visual observa-
tions (see Section 4.7.7.1) conducted by the EPA at the Tacoma smelter
indicate that this capture system is quite effective in capturing
fugitive emissions generated during matte tapping operations. Visual
observations indicate that the capture efficiency of this system is
probably in excess of 90 percent.58 Consequently, from an alternative
control standpoint, the ASARCO system applied at Hayden and Tacoma
seems to be a viable alternative for the capture of fugitive emissions
generated during matte tapping operations.
4.7.5.2 Control of Fugitive Emissions From Slag Skimming Operations.
Slag skimming is the process by which molten slag is removed from the
smelting furnace for disposal. Slag is skimmed from the furnace
through skim bays that may be located on the sides or at one end of
the furnace. As with matte tapping ports, the number and location of
4-159
-------
cr>
o
Cable to Winch
Launder
26"
3'8"
Matte Tap Hood
6"
Figure 4-29. Schematic of the matte tapping and ladle hoods at the ASARCO-Tacoma Smelter.
-------
the skim bays will vary depending upon the size and type of furnace.
Normally, slag is skimmed from one bay at a time and conveyed through
launders into one or several slag pots (ladles). Slag pots may range
in capacity from 3 to 17 m3 (100 to 600 ft3). A single slag skim
may last from 10 to 20 minutes.
Because slag skimming is very similar to matte tapping, the
fugitive emissions capture technology used for both operations is
quite similar. Local exhaust hoods are employed over skim bays and
slag launders are either partially or completely covered. Design
rates for skim bay hoods vary from smelter to smelter; however, they
normally range from 566 to 850 NnrVmin (20,000 to 30,000 scfm).58
A schematic of the slag skimming fugitive emission control system
used at the ASARCO-Tacoma smelter is presented in Figure 4-30. The
slag skim hoods are pyramid!cal in shape, with a 1.2-m by 2.4-m (4 ft
by 8 ft) rectangular cross section, and they are less than 0.9 m
(3 ft) above the skim port. A larger exhaust hood with a 2.4-m by
4.3-m (8-ft by 14-ft) rectangular cross section is situated directly
above the slag pot transfer point. In addition, each launder is
covered with a fixed hood. During skimming, the slag skim hood operates
at about 142 NmVmin (5,000 scfm), while the hood above the slag pot
operates at approximately 560 NmVmin (20,000 scfm). Emissions that
are captured by the launder hooding are vented to either the slag skim
hood or the slag pot hood.
From visual observations made by the EPA at the Tacoma smelter,
it was determined that this type of hooding system achieves a capture
efficiency of approximately 90 percent.58 Consequently, the Tacoma
system is considered to be a viable alternative for the control of
fugitive emissions generated during slag skimming operations.
4.7.6 Capture of Fugitive Emissions From Converter Operations
Primary converter hoods capture the majority of the process
emissions generated during converter blowing operations with the
exception of some emissions that escape due to primary hood leakage.
However, during converter charging, skimming, and pouring operations,
the mouth of the converter is no longer under the primary hood, thus
4-161
-------
I
I—'
en
To Baghouse
Reverberatory
Furnace
NOTE:
1. All Ducts - 3/16" thick C.S.
2. Hoods - 3/16" thick C.S.
18
Ter V\
I y 36" dia
40" dia X 3/16"
65'
36" dia
NOTE: The dimensions indicated are those of
a system of this type currently utilized at
the ASARCO-Tacoma facility.
Figure 4-30. Schematic of the slag skimming (plan view) fugitive emissions control system
at the ASARCO-Tacoma Smelter.58
-------
significant quantities of fugitive emissions escape capture by the
primary hood. Fugitive emissions are also quite extensive during the
converter holding mode as discussed in Section 3.3.2.4.
Previously conducted emissions testing by EPA has indicated that
converters are the most significant sources of fugitive particulate
and S02 emissions within primary copper smelters (see Tables 3-5 and
3-6). There are three basic approaches being applied to capture
fugitive emissions from converter operations: (1) general ventilation,
i.e., building evacuation; (2) secondary mechanical hoods; and (3) second-
ary mechanical hoods coupled with air curtains.
4.7.6.1 General Ventilation as a Means of Capturing Fugitive
Emissions Generated by Converter Operations. Building evacuation has
historically taken the form of either natural air changes due to wind
and atmospheric density differences or mechanically assisted air
changes. However, some engineering considerations must be made before
applying a simple rate-of-air-change method for designing an industrial
ventilation system. The rate-of-air-change method estimates are based
on room volume only and do not consider the rate of evolution of the
contaminant, the number of heat sources, or the natural draft due to
building configuration. For example, a general ventilation installation
designed by the rate-of-air-change method can, under some conditions,
actually cause the contaminant to be spread throughout the building,
thus increasing the volume of dilution air required to maintain hygienic
conditions. This situation occurs when the distribution of the ventila-
tion air supply is poorly controlled. Uncontrolled airflows into a
building due either to negative pressure in the building or to poorly
designed air supply distributors may not only cause recirculation of
the contaminant but also may upset the local ventilation systems. It
is therefore important that the amount of air, its location of entry
into the building, and its direction be controlled. For example,
Figure 4-31 shows a convective flow from a heat source (such as a
ladle of molten metal) rising to be exhausted through a roof ventilator.
Figure 4-32 shows an uncontrolled air supply that results in a disrupted
rising plume and recirculation of the contaminant throughout the
building.
4-163
-------
Figure 4-31. Controlled airflow from a heated source.58
Figure 4-32. Uncontrolled airflow from a heated source:
58
4-164
-------
Natural air changes take place when hot air from the ground level
heat sources rises due to its buoyancy. If, however, a point exists
within the building where the temperature of the surrounding air is
equal to that of the rising column of hot air, buoyancy is lost.
Therefore, natural air changes will take place only if the temperature
of the rising column of hot air is high enough to maintain the buoyancy
of the column until it is discharged through the roof monitors.
However, in most hot metal workshops, this is not the case. Hot pools
of contaminated air are formed under the building roofs. As a result,
clean air entering the building will at times mix turbulently with
pools of contaminated air and transport it downward to the occupied
levels near the floor.
Increasing the ventilation rate is the most commonly used method
of correcting such an air contamination problem. Increasing the rate
of ventilation through the building will have the effect of raising
the column of hot air. Increased ventilation can be obtained by
increasing the area of supply openings and roof openings or by using
mechanical means such as exhaust fans.
Building evacuation by means of natural air changes or by a
combination of natural and mechanical air changes is used at most
domestic copper smelters. The captured fugitive emissions are usually
vented to the atmosphere either through roof monitors or through roof
stacks.
At the ASARCO-E1 Paso smelter, the concept of controlled ventilation
is being used to capture and collect the emissions from the converter
aisle. Controlled ventilation is accomplished by controlling the
airflow patterns within the building and determining the flow of air
to be handled. Control of airflow in the ventilated area is obtained
by isolating it from other areas and by the proper design and placement
of inlet and outlet openings. A well-contained and isolated area
results in the handling of a minimum volume of air. Proper location
and sizing of inlet and outlet openings provide effective airflow
patterns so that the fugitive emissions cannot escape to adjacent
areas or recirculate within the area. The configuration of the con-
4-165
-------
verier building inlet and outlet openings at the ASARCQ-E1 Paso smelter
is shown in Figure 4-33.
Isolation of the ventilated area at the ASARCO-E1 Paso smelter
was accomplished by installing additional sheeting in the roof truss
area along the converter building column lines to enclose the flow
from the smelting furnace area. The tuyere punching platforms east of
the three converters were enclosed. The larry car rail line at the
west side of the converter building was enclosed to contain the dust
and S02 released when the larry cars were emptied. The location and
sizes of these enclosures were selected to provide maximum feasible
containment without interference with metallurgical operations.
Partitions within the roof of the converter building were provided to
prevent lateral migration of fume into adjacent areas.
The air velocity through the inlet openings was controlled to
provide directional flow control and supply an adequate volume of air
into locations where needed. This was achieved by using adjustable
louvers on air inlet openings through building walls. Ventilating
outlets are located at the ridge line of the converter building roof
in the center of each partitioned area.
Inlet air is admitted to the converter aisle through louvers and
permanent openings in the east and west walls of the building. Makeup
air for the zinc holding and reverberatory furnace matte tap exhaust
systems, which operate periodically, enters through adjustable louvers
along the west wall. Inlet air for the reverberatory furnace charge
area enters through a permanent opening in the south wall for the
elevated-line railroad train and through adjustable louvers along the
west wall. Inlet air for the extremes of the converter building is
admitted through egress openings and louvers along the east and west
walls and through a permanent opening in the south wall. The latter
opening is permanent to accommodate the fairly frequent lead matte car
railroad traffic.
During winter operation, a majority of the gravity roof ventila-
tors in the extremes of the converter building are closed to keep heat
in the building. This requires partial closing of inlet air louvers.
4-166
-------
A H G
I
i—'
CTl
f
\
V
f
;
3
t
IV
I
3
k
T=_
V
/I
l
:T
N
L
/I
X
1
^ c
r 1
3
I
t
i I
11
J
V
L t
E
(
[
i i
. J
D
r
T]
U
G
1 ' 1 '
I
r
L
\
Vi
i
w
i i
o
1
J
-ill I—I
1 L 1
F C
1 C
L
llr
11
2
I
1
)
f
Cl
B J C
' i
r|
nllr-
K
[.
1
1
i
> J i
C
•"1
1
B
' 1
1
r I
[1
4
; i
f 1
i
r
t
r-
A
I
J
1
A
, ,
f l
4
•* -
-*- -
NOTE: See key on following page.
H O A
O J
Figure 4-33. Inlet-outlet openings in converter building at ASARCO-EI Paso
58
-------
INLET (AIR INTAKE) SCHEDULE (NEW
KEY
NO. TYPE
SIZE
BAGHOUSE SYSTEM)
NO. EACH
REQD. (scfm)
TOTAL
(scfm)
PERMANENT OPENINGS
Egress Openings
A in Walls
Converter Wall
Openings
_ Anode Casting
** Wheel Opening
_ Hi-Line Track
° Entrance
_ Zinc Fuming
fc Wall Opening
Skull Breaker
h Wall Opening
7'-6"H x 3'-8"W
5'-0"H x 16'-0"W
12'-0",H x 37'-0"W
(20 ft^ Net Opening!
12'-0"H x 8'-0"W
3'-2"H x 4'-6"W
5'-8"H x 5'-8"W
7 13,500
3 40,000
1 60,000
1 45,000
1 7,000
1 16,000
94,500
120,000
60,000
48,000
7,000
16,000
LOUVERED OPENINGS
_ Tuyere Puncher
Housing
.. Hi-Line Track
Enclosure, #3 Conv.
n Skull Breaker
u Wall Openings
1 #2 & 3 Converters
J Zinc Fuming Wall
K #1 Converter
1'-8"H x 1'-0"W
3'-2"H x 2'-3"W
5'-3"H x 6'-3"W
4' 5"H x 5'-6"W
3' 2"H x 5'-0"W
4' 5"H x 6'-3"W
9 1,300
3 3,000
3 16,000
3 16,000
2 6,500
1 12,000
TOTAL
11,700
9,000
48,000
48,000
13,800
12,000
485,000
AUXILIARY LOUVERS
_ Matte Tap Hood
^ Makeup Air
. Slag Tap Hood
L Makeup Air
5' 8"H x 6'-3"W
6'-11"H x 8'-3"W
2 15,000
1 25,000
TOTAL
30,000
25,000
55,000
. Egress Openings
A in Walls
Lead Slag Track
" Entrance
J Louvered Openings
M Louvered Openings
OUTLET
Existing
Hood
Systems
7'-6"H x 3'-8"W
12'-0"H x 10'-0"W
3'-2"H x 5'-6"W
5'-8"H x 8'-0"W
4 13,500
1 60,000
2 7,000
2 21,000
TOTAL
(EXHAUST) SCHEDULE
Proposed Roof
Evacuation
(New Baghouse System)
KEY scfm KEY scfm
[Tj 25,000
[2] 30,000
TOTAL 55,000
\J7 130,000
\§7 200,000
^7 136,000
\^7 16,000
488,000
Gravity Draft at
Extremes of Building
(Variable)
KEY scfm
^7 68,000
^7 102,000
1 70.000
54,000
60,000
14,000
42,000
170,000
Figure 4-33. (continued).
4-168
-------
The gravity roof ventilators over the portion of the building that is
vented to the building evacuation system are normally closed; however,
they automatically open in the event of power failure to provide
emergency ventilation. The ventilators over the extreme ends of the
converter building are normally open; they remove heated air from
areas where emissions are slight.
Railroad doors are kept closed when cars and slag haulers are not
entering or leaving the building. Allowing these doors to remain open
can adversely affect the building evacuation system, with a design volume
of 5,660 actual mVmin (200,000 acfm), depending on wind conditions; e.g.,
(I) airflow is reduced through the normal inlets, resulting in poor inlet
air distribution; (2) the air moving through the shop forms eddies that
pick up fume from the furnaces, entrain it, and spread the fume through-
out the converter building; and (3) a strong wind through the railway
car door entering the building results in air volume exceeding the
total ventilating capacity of the building evacuation system. A
positive pressure within the building could occur, and fume could be
forced out of the building through the normal inlet openings.
The present volume evacuation rate of this system at the ASARCO-
El Paso smelter is 16,800 NmVmin (600,000 scfm)--equivalent to 18 air
changes per hour.58 Supplementary air is provided when the zinc slag
holding furnace or reverberatory furnace matte launder local exhaust
systems are in use. The average exhaust gas temperature from the
converter building after the exhaust gases from the four building
exhaust ducts mix is 55° C (130° F). Nominal duct design gas veloc-
ities are 1,500 m/min (5,000 ft/min) from the converter building to a
baghouse and 900 m/min (3,000 ft/min) from the baghouse to the annul us
of the main stack. The building exhaust gases contain negligible
water vapor and S03. Ideally, 100 percent capture could be achieved
by this system. However, due to the need for egress and entry openings,
some losses are to be expected. Thus, performance is set at 95 per-
cent.58 It is evident, however, from visual observations, that a
higher air change rate than that applied at El Paso is needed to
alleviate exacerbated worker exposure.
4-169
-------
The baghouse receiving building ventilation gases at the ASARCO-
El Paso smelter contains 4,800 bags, each 15 cm (6 in.) in diameter by
9 m (30 ft) long. It is sized with a nominal air-to-cloth ratio of 3
to 1. Three backward curved air foil fans are used on the clean-air
side of the baghouse and are sized on the basis of 110 percent of the
estimated airflow. The results of tests that EPA performed on this
baghouse are summarized in Section 4.7.7.
4.7.6.2 Secondary Mechanical Hoods as Means of Capturing Fugi-
tive Emissions Generated by Converter Operations. In normal practice,
primary converter hoods are used when the converters are in the blowing
mode. However, fugitive emissions that occur as a result of primary
hood leaks are uncontrolled unless some type of additional capture
equipment is employed. Secondary mechanical hoods can be used to
capture emissions that result due to primary hood leaks. It is impor-
tant to note, however, that these hoods are not designed to capture
fugitive emissions that occur during the charging, skimming, and
holding modes; thus, only emissions that escape the primary hood
during blowing are captured. Currently, there are several domestic
primary copper smelters that use some form of secondary mechanical
hooding to capture fugitive emissions that result from primary hood
leaks. The types of hoods currently in use are:
Fixed type—Attached to the primary hood; currently in use
at Phelps Dodge-Ajo, Phelps Dodge-Hidalgo, Phelps Dodge-Morenci,
and Kennecott-Magna.
Retractable—Supported by walls on either side of the con-
verter; currently in use at ASARCQ-Hayden.
The fixed-type hood, illustrated in Figure 4-34, is approximately
3 m (10 ft) long, 6.4 m (21 ft) wide, and 1.7 m (5.5 ft) high and is
affixed to the upper front side of the converter primary offtake
hoods. Visual observations indicate that this type of hood is only
marginally effective in capturing fugitive emissions that are generated
during blowing.58 Therefore, this discussion will concentrate on the
retractable-type hood (shown in Figure 4-35).
4-170
-------
To Secondary 4
Hooding
Main Duct I
Figure 4-34. A typical fixed secondary converter hood:
58
4-171
-------
Retractable
Secondary
Hood
Primary
Hood
Wing Wall
Figure 4-35. Retractable-type secondary hood as employed at ASARCO-Hayden.
4-172
-------
Subjective visual observation of a retractable-type secondary
hood at the ASARCO-Hayden smelter yielded estimated capture effi-
ciencies of up to 75 percent for fugitive emissions that escaped the
primary hood during blowing operations. Thus, although the hood is
not intended to capture emissions that occur while the converter is
"rolled out," it is reasonable effective in capturing fugitive emissions
that occur during blowing operations.
The secondary hood, as shown in Figure 4-35, is retracted during
charging, pouring, and skimming operations so that the crane operator
can have access to the converter mouth as well as the area directly in
front of the converter. Then, as blowing begins, both the primary and
secondary hoods are lowered into position over the mouth of the rolled-
in converter. The most significant source of emissions that escape
capture by this hooding system during blowing are slots in the top of
the secondary hoods through which the cables for primary hood retraction
are passed.
In summary, the ASARCO-Hayden hood is judged by EPA to be reasonably
effective in capturing fugitive emissions generated by converters
during the blowing mode. The primary shortcoming of this type of hood
is its inability to capture fugitive emissions that occur during
operations other than blowing. The following discussion dealing with
fixed enclosure/air curtain systems will address the capture of fugitive
emissions generated during all converter operating modes.
4.7.6.3 Air Curtains and Fixed Enclosure Hoods as Means of
Capturing Fugitive Emissions Generated by Converter Operations.
Another method of controlling fugitive emissions from copper smelter
converting operations involves the use of an air curtain system along
with a secondary hood system. Although air curtains for the control
of fugitive emissions are not currently being used in the domestic
primary copper smelting industry, they are being used abroad and in
other U.S. industries.58
An air curtain is a suitably shaped air jet with sufficient
momentum to resist the forces of fugitive gas streams working against
it and to maintain its continuity across the opening it protects.
4-173
-------
Consideration in air curtain design must also be given to secondary or
entrained flows that start forming as the air curtain jet stream
leaves its slot or nozzle. As the entrained flows become fully mixed
with the air curtain jet stream some distance from the nozzle, the hot
or cold secondary flows are carried from one side of the air curtain
jet stream to the other where they are ducted for suitable discharge
(see Figure 4-36). The greater the entrained flow, the greater the
energy loss. To minimize energy loss, a thick, slow-moving jet stream
with a large air volume is required. A basic rule in the design of
air curtains is to project the thickest and lowest velocity air stream
possible across the shortest dimension of the opening.
The design of air curtains is quite complex because of the curving
pattern of the airflow from the air curtain jet nozzle or slot. Also,
the presence of secondary flows further complicates the design. Air
curtain design methods are discussed in References 61, 62, and 63.
The type of air curtain system being used at the Onahama and
Naoshima primary copper smelters in Japan is shown in Figure 4-37.
The capture/shielding device includes two steel plate partitions, one
on each side of the converter month. The air jet is blown from a slot
at the top of one of the plates across the opening to provide a sheet
or curtain of air that prevents fugitive emissions from escaping. The
other plate is equipped with an exhaust hood. The opening allows the
crane cables to move into position above the converter mouth.
A propeller fan is used to push the air through an elongated slot
on one side and a backward inclined fan provides suction on the opposite
side to pull in both the fugitive gases and the push air. Captured
gases pass through steel duct work to a baghouse. The combined tempera-
ture of the converter fugitive gases with 100 percent of the push air
entering the duct work on the pull side of the air curtain is approxi-
mately 80° C (100° F), which makes gas cooling unnecessary before
gas-stream entry into the control device.
The inlet air forming the air curtain above the converters at the
Naoshima smelter has a flow rate of approximately 600 NmVmin (21,000
scfm). The exhaust hood on the opposite side pulls in approximately
4-174
-------
Air Jet Source
Jet Width
Primary Flow
4-Jet Widths
Figure 4-36. Entrained flow diagram.58
4-175
-------
Converter
SCALE: 1"= 10'
Figure 4-37. Converter air curtain/secondary hooding system as employed at the
Onahama and Naoshima smelters.58
4-176
-------
1,000 Nm3/min (35,000 scfm) of gas to the main system. The capacity
of the total pull system at this smelter is three times this value or
3,000 Nm3/min (105,000 scfm) to allow for the operation of three hoods
at a time. According to the Naoshima authorities, the overall collection
efficiency of these hoods for fugitive emissions is approximately
90 percent.58 An air curtain system of similar design is currently
being installed at the ASARCO-Tacoma smelter. The Tacoma system will
eventually handle fugitive emissions from three converters.
The Tamano copper smelter in Japan uses a differently designed
air curtain system along with a fixed hood, which is essentially a
total enclosure equipped with front doors and a retractable roof, for
controlling fugitive emissions from each of its three converters.
(Usually one converter is operated at a time.) A sketch of the air
curtain system being used at the Tamano smelter is shown in Figure 4-38.
The enclosure has two front doors and a movable roof that is slightly
inclined toward the front. The air curtain ducts are located at the
top of the enclosure level at a position to push air from one side of
the converter to the other side. Ambient air is supplied by a ground
fan rated at 70,000 NmVh (41,000 scfm).
The typical functions of the air curtain system and secondary
hood system during each mode of a converter cycle at the Tamano smelter
are summarized in Table 4-17 and are described in the following para-
graphs.
For charging matte and other material to the converter, the
secondary hood doors and the movable roof are opened and the air
curtain system is turned on. When opened, the movable roof slides
toward the side away from the air curtain ducts. The air curtain
system is turned on only during charging of the converters when the
movable roof is kept open. The movable roof is closed during all
other modes of converter operation. A ladle containing charge material
is brought inside the secondary hood enclosure by an overhead crane.
The converter is rolled down to an inclined position. The ladle is
lifted up by the crane, and the material is charged into the converter.
4-177
-------
Roof Opening
for Fugitive Gas
Fugitive Gases
to Baghouse
Fugitive Gas
to .4
Desulfurization
Plant
End Flue for
X. Air Curtain
Offtake
Primary Offgases
to Acid Plant
Waste Heat Boiler
Movable Roof
Mouth of
the Converter
Front Door
Converter
Figure 4-38. Schematic diagram of the converter housing/air curtain system
at the Tamano smelter.58
4-178
-------
TABLE 4-17. FUNCTION OF AIR CURTAIN AND SECONDARY HOOD SYSTEM
DURING VARIOUS MODES OF CONVERTER OPERATION AT TAMANO SMELTER58
Configuration of primary hood and
Mode of operation secondary hood
Material charging (matte, A, C
or cold dope)
Slag blow B, D
Slag discharge A, D
Copper blow B, D
Blister discharge A, D
aPrimary hood position:
A = damper closed.
B = damper open.
Secondary hood position:
C = doors and roof open and air curtain on.
D = doors and roof closed and air curtain off.
4-179
-------
Usually three ladles of matte and one boat of cold material are charged
to the converter within a 10- to 30-minute period. Actual charging of
each ladle lasts 1 to 1% minutes. At the completion of material
charging, the ladle is brought out from the enclosure. The converter
is moved to its upright position and its mouth is contained in the
primary hood. The secondary hood housing doors and roof are closed.
During charging and discharging of material, two fugitive gas
streams are generated. A relatively high concentration gas up to
30,000 Nm3/h (18,000 scfm) in volume, generated in the vicinity of the
converter mouth, is captured by the duct work located at the lower
inside wall of the enclosure. The captured gas stream is continuously
pulled into a lime desulfurization plant for treatment. The larger
portion of fugitive gases is mixed with air generated by the air
curtain system. The combined total gas flow of up to 190,000 m3/h
(6.7 x 106 cfm) is captured by a duct work located at the upper level
of the enclosure. The captured gas is sent to a baghouse for particulate
matter removal.
Any gases escaping from the air curtain system are recirculated
through a building roof hood.
During the slag blow and copper blow, the converter mouth is
housed under the primary duct, and the secondary housing doors and
roof are closed. The primary offgases, which range between 65,000 and
75,000 NmVh (38,000 to 44,000 scfm) from each converter are treated
along with offgases from the flash furnace smelter in a 156,000-Nm3/h
(92,000-scfm) capacity acid plant for S02 removal. Any fugitive
emissions generated from the converter due to primary hood leaks will
pass to the baghouse and then to the stack.
During slag skimming at the end of each slag blow and during
blister discharge at the end of copper blow, a ladle is brought into
the secondary hood enclosure by the overhead crane and placed on the
ground in front of converters. The secondary hood doors and roof are
closed. The mouth of the converter is rolled down and the slag or
blister is poured into the ladle. After the slag or blister discharge
is completed, the converter mouth is moved up, the housing doors and
roof are opened, and the ladle is moved out.
4-180
-------
High S02 concentration gases and low S02 concentration gases are
passed through the corresponding ducts to the lime desulfurization
plant and the baghouse system, respectively.
Subjective evaluation of the air curtain and fixed enclosure
system by visible observation at the Tamano smelter in Japan indicate
the system to be at least 90 percent effective in controlling fugitive
emissions.58
ASARCO's Tacoma facility is currently installing an air curtain
system on one of their converters. The design volume for the system
is the anticipated possible maximum, which would occur when two
converters require 2,800 actual mVmin (100,000 acfm) each.64 Thus,
the total design volume is 5,700 actual mVmin (200,000 acfm) at 66° C
(150° F). Design data for the ASARCO system are summarized in
Table 4-18. ASARCO has estimated that the following overall capture
efficiencies will be in evidence after the system is placed into
operation:65
~98 percent during blowing
~60 percent during roll-in and roll-out
~85 percent during skimming
~85 percent during charging
~90 percent during holding.
The gases that are captured by the air curtain system will be sub-
jected to particulate removal before being passed to the atmosphere.
Based upon data supplied by ASARCO, EPA has tentatively concluded
that the ASARCO air curtain system will be able to achieve the capture
efficiencies outlined above. EPA does feel, however, that the addition
of doors and a retractable roof may merit consideration as a means of
enhancing capture efficiency.
4.7.7 Summary of Visible Emissions Data for Fugitive Emissions Sources.
4.7.7.1 Local Ventilation Techniques Applied to Calcine Discharge,
Matte Tapping, and Slag Skimming. The performance of local ventilation
techniques used at the ASARCO-Tacoma smelter for the control of fugitive
emissions from calcine discharge, matte tapping, and slag tapping
operations was evaluated.58 These techniques were previously described
4-181
-------
TABLE 4-18. SUMMARY OF DESIGN DATA FOR THE ASARCO-TACOMA CONVERTER
SECONDARY HOODING/AIR CURTAIN SYSTEM64
Mode Air curtain push rate, Main offtake evacuation
of operation actual mVmin rate, actual m3/min
Matte charging 510 (18,000 acfm) 2,322 (82,000 acfm)
Blowing -a 1,700 (60,000 acfm)
Slag skimming 510 (18,000 acfm) 2,322 (82,000 acfm)
Holding 510 (18,000 acfm) 850 (30,000 acfm)
Worst conditions'3 1,020 (36,000 acfm) 4,644 (164,000 acfm)
aAir curtain will not be used during the blowing mode,
Worst conditions would consist of either (1) two converters being
charged simultaneously or (2) one converter being charged while
another was being skimmed.
4-182
-------
in Sections 4.7.4 and 4.7.5. Visual observations were made using
either EPA Method 22 or EPA Method 9, depending on whether the emissions
observed were intermittent or continuous. Method 22 is used to determine
the occurrence of visible emissions and Method 9 is used to determine
the opacity of emissions. A summary of the visible emissions data
obtained is presented in Table 4-19.
Thirteen calcine transfer operations, averaging about 2 minutes
in duration each, were observed. The visual observations were made
using EPA Method 22 at the opening of the tunnel-like structure used
to house the calcine hoppers and larry cars during the calcine trans-
fer (discharge) operations. As the data indicate, no visible emissions
were observed at any time.
Visible emission observations during reverberatory furnace matte
tapping were also made at ASARCO-Tacoma using EPA Method 22. Simul-
taneous but separate observations were made both at the furnace tap
port and at the launder-to-ladle transfer point. Sixteen taps, averag-
ing approximately 5.5 minutes in duration, were observed. Out of the
16 observations made at the matte tap port, no visible emissions were
observed 100 percent of the time during 14, with only slight emissions
ranging from 1 to 3 percent of the time for the remaining 2. No
visible emissions were observed at any time from the launder or launder-
to-matte- ladle transfer point during all 16 observations.
Slag skimming emissions were observed using both EPA Methods 22
and 9. As with matte tapping, separate observations were made at the
furnace skim bay location and at the slag-launder-to-slag-pot transfer
point. Results obtained using EPA Method 22 for 8 observations at the
slag tap port showed that visible emissions were observed about 5 percent
of the time on the average, with the highest single observation showing
the presence of visible emissions 15 percent of the time. Visual
observations made at the slag-launder-to-pot transfer point indicated
very poor performance, with visible emissions being observed 72 to
99 percent of the time over 11 slag taps. Additional data obtained
using EPA Method 9 showed significant emissions, with opacities as
high as 50 percent. Conversations with smelter personnel revealed
4-183
-------
TABLE 4-10 SUMMARY OF VISIBLE EMISSION OBSERVATION DATA™ FOR CAPTURE
SYSTEMS OH FUGITIVE EMISSION SOURCES AT ASARCO-TACOMA
EPA Method 22
EPA Method 9
Operation
Average
percent
time Range of Average
Number Average emissions percent Number observ- Ranqe of
of observation observed time of ation Average opacity
readings time on all emissions readings time, opacity, percent.
taken' miirsec readings observed taken m in: sec percent observed
4^
i — i
05
-pi
Ca'rine transfer system
Matte tapping
i\\ mai i,e tip port and
launder
At matte disrhanjp
into ladle
Slag skimming
At s lag skim L>ay ind
launder
A! S ' -H! d i 'iCh'lt fjp
into pots
13 1.55 0 0
16 5:28 U.? 0 to 3
15 5:21 0 0
8 11:38 5.3 0 to 15 2 13:45 6 0 to 30
11 15.77 88 72 to 99 7 14-32 12 0 to 50
_ _ _.-_----- - —
''visible emission observations made June ?'i through 26,
-------
that the ventilation hood at the slag launder discharge point has been
damaged when hit by a truck. Although an inspection of the ventilation
hood and ancillary duct work showed no apparent damage, ventilation at
this location was concluded to be inadequate to handle the volume of
emissions and fume generated.
Visible emissions data for matte tapping and slag skimming operations
were also obtained at the Phelps Dodge-Morenci smelter. EPA Methods 9
and 22 were employed. Observations of matte tapping operations involved
the specific type of local capture hood discussed in Section 4.7.5.1
(see Figure 4-30). The results of the Method 9 observations for matte
tapping, presented in Table 4-20, indicate that the Morenci hood does
not achieve 100 percent capture. Average opacities ranged from 2 to
45 percent over 22 observation periods that ranged from 4 to 11 minutes
in duration. Method 22 results, presented in Table 4-21, showed
visual emissions to be in evidence 100 percent of the time during
three of the four observation periods. Both types of observations
(Methods 9 and 22) were made at the tapping port. Slag skimming
observations involved an evacuated "doghouse" type enclosure. Method
9 results, presented in Table 4-22, indicated that this type of hood
is reasonably effective in capturing emissions generated at the skim
bay. Average opacities ranged from 0 to 11 percent over six observation
periods that ranged from 6.25 to 33.00 minutes in duration. A subjective
evaluation yielded an approximate 90 percent capture efficiency for
this type of hood. One Method 22 observation of 30 minutes duration
was made (see Table 4-22). Emissions were in evidence approximately 3
percent of the time.
In conclusion, the fugitive emissions control system for calcine
discharge at the ASARCO-Tacoma smelter represents a most effective
means by which to capture emissions generated during calcine discharge
operations. Visual observations indicate that over 90 percent capture
is possible.
The ASARCO-Tacoma matte tapping fugitive emissions capture system
(see Section 4.7.5.1 for a complete description) also appears to
represent the most effective system for the capture of fugitive emissions
4-185
-------
TABLE 4-20. VISIBLE EMISSION OBSERVATION DATA FOR REVERBERATORY
FURNACE MATTE TAPPING OPERATIONS AT THE PHELPS DODGE-
MORENCI SMELTERa
Average opacity
Duration of observation for observation Range of
period, min period, percent individual readings
8.75
8.50
6.50
8.50
5.00
6.50
9.00
11.00
9.50
4.00
9.50
6.50
9.50
8.00
5.00
7.75
5.00
7.50
5.00
9.25
6.50
3.75
8.57
2.06
8.85
8.09
7.25
7.31
11.39
15.68
16.71
10.00
14.20
18.46
47.06
17.34
6.88
18.23
17.75
14.50
7.00
24.86
7.50
6.67
5 to 25
0 to 25
5 to 20
5 to 30
5 to 10
5 to 20
5 to 20
5 to 30
10 to 20
5 to 10
5 to 30
10 to 30
10 to 60
10 to 40
5 to 25
10 to 30
10 to 30
5 to 35
0 to 30
10 to 70
0 to 30
0 to 30
aBased on visual observations made in accordance with EPA Method 9.
4-186
-------
TABLE 4-21 VISIBLE EMISSION DATA FOR REVERBERATORY FURNACE MATTE
TAPPING OPERATIONS AT THE PHELPS DODGE-
MORENCI SMELTER3
Duration of observation
period, min
6.0
7.0
5.0
5.0
Percent of time
emissions observed
100
100
82
100
Light reading,
350
175
350
88b
lux
aBased on visual observations made in accordance with EPA Method 22.
bNot a valid observation since the light was less than 100 lux.
4-187
-------
TABLE 4-22. VISIBLE EMISSION OBSERVATION DATA FOR REVERBERATORY
FURNACE SLAG SKIMMING OPERATIONS AT THE PHELPS DODGE-
MORENCI SMELTER
Reference Method 9 results
Duration of observation
period, min
30.00
30.00
33.00
6.25
27.00
30.00
Average opacity
for observation
period, min
0.00
0.00
2.72
11.00
0.00
0.79
Range of
individual readings
_a
_a
0 to 5
5 to 30
_b
5 to 10
Reference Method 22 results
Duration of observation
period, min
Percent of time
emissions observed
Light reading, lux
30.00
175
No opacity readings above 0.0 were observed.
4-188
-------
generated during matte tapping and the associated launder-to-ladle
transfer operation. Visual observations indicate that a system of
this type can consistently achieve a 90-percent capture efficiency.
Visual observations of slag skim bay hoods at both ASARCO-Tacoma
and Phelps Dodge-Morenci indicate that local hooding can also be quite
effective in capturing fugitive emissions generated at the skim bay.
Subjective evaluations of both systems yielded approximate capture
efficiencies of 90 percent. Although the hooding used to capture
emissions at ASARCO's slag launder-to-ladle transfer points was observed
to be ineffective, local hooding should not be dismissed as a viable
capture scheme for emissions generated at slag launder-to-ladle transfer
points. Similar hooding at the matte launder-to-ladle transfer points
was judged to be effective, suggesting that a properly designed and
operated local hooding system should operate effectively at slag
launder-to-ladel transfer points. Thus, a 90-percent capture efficiency
would be reasonable at such points.
4.7.7.2 Fugitive Emission Controls for the Converter Fixed
Enclosure/Air Curtain Hood Capture System at the Tamano Smelter.
Visible emission observations were made at the Tamano smelter in Japan
for the fixed enclosure/ air curtain system operated on the No. 3
converter during day shifts on March 12 and March 13, 1980. The
converter is of conventional Fierce-Smith design, measuring about
9 meters in length and 4 meters in diameter. Observations were made
using EPA Method 22 and EPA Method 9, depending on whether the emissions
observed were intermittently or continuously, for the different modes
of converter operation comprising a converter cycle. Discussions of
the results obtained during each mode of converter operation are
presented in the following sections.
4.7.7.2.1 Visible emissions from matte charging. Usually three
ladles of matte are brought to the converter and charged in a 10- to
30-minute period. Actual matte charging from each ladle lasts for 1
to lh minutes. The fixed enclosure doors and roof are opened, the air
curtain system is turned on, and the ladle of matte is brought into
the secondary hood by an overhead crane. The converter is rolled down
4-189
-------
to an inclined position, the matte ladle is lifted up by the crane,
and matte is charged into the converter. At the completion of matte
charge, the ladle is moved out of the enclosure and, if needed, another
ladle of matte is brought in. After the matte additions are completed,
the converter is rotated into the primary converter hood, the roof and
doors are closed, and slag blowing commences.
Three separate matte charges were observed using both EPA Methods 9
and 22 simultaneously, and one matte charge was observed using EPA
Method 9 only. Visual observations for each matte charge observed
were made only during the period when the matte was actually flowing
into the converter. Results of the visual observations obtained are
summarized in Table 4-23.58
As shown in Table 4-23, visible emissions were observed for three
individual matte charges. The observations ranged from 44 to 77 percent
of the time (EPA Method 22). Although somewhat continuous in nature,
the opacity results indicate that these emissions were generally
slight, typically ranging from 0 to 10 percent opacity, with the
highest average opacity recorded for a single matte charge being
5 percent. When present, the emissions appeared as small puffs that
penetrated the air curtain stream.
4.7.7.2.2 Visible emissions during slag and copper blowing.
During slag blowing and copper blowing, the converter mouth is contained
in the primary duct, and offgases are directed to the acid plant. The
converter secondary hood doors and roof are closed, and the air curtain
system is turned off. Fugitive emissions generated during blowing as
a result of primary hood leaks are captured inside the converter
housing and are vented to a baghouse for collection. The slag blow,
which is divided into three segments, lasts for about 150 minutes per
converter cycle and the copper blow for about 200 minutes per cycle.
Visible emissions observations were made using EPA Method 9 for
the converter hood system for 30 minute, during the slag blow and for
27 minutes during the copper blow. No visible emissions (zero percent
opacity) were observed at any time.
4-190
-------
TABLE 4-23. VISIBLE EMISSION OBSERVATION DATA FOR CONVERTER
SECONDARY HOOD SYSTEM DURING MATTE CHARGING
AT THE TAMANO SMELTER58
Sample
run
Method 22
Percent
Observation of time
period, emissions
min observed
Obser-
vation
period,
min
Method 9
Average
opacity for
observation
period,
percent
Range of
individual
readings
1
2
3
4
Total
1.5
1.25
1.75
--
4.50
44
56
77
--
60
1.5
1.25
1.75
1.5
6.25
5.0
4.0
3.0
0
2.8
0 to 25
0 to 10
0 to 10
--
0 to 25
4-191
-------
4.7.7.2.3 Visible emissions during converter slag discharge. At
the end of each of the three slag blow phases, slag is skimmed into a
ladle and transported to a sand bed area for cooling. Because of the
quantities involved, slag is discharged from the converter two times
after the first slag blow and once after the second and third. Each
slag skim lasts for about 10 minutes. During each skim, an empty
ladle is brought into the enclosures by an overhead crane and placed
on the ground in front of the converter. The crane is moved out, and
the enclosure doors and roof are closed. The converter is rolled down
and slag is poured into the ladle. After the slag skimming is completed,
the converter is rotated upward slightly, the enclosure doors are
opened, and the slag ladle is moved out.
Only two skims were observed. The first, which lasted 11 minutes,
was observed using EPA Methods 22 and 9. The second slag skim, lasting
9 minutes, was observed using EPA Method 22 only. Each observation
period began as the converter started rolling down to pour the slag
into the ladle and lasted until the pouring was completed and the
converter started rolling up. During the first slag skim observed, no
visible emissions were observed at any time. In contrast, during the
second slag skim, visible emissions were observed 100 percent of the
time. The majority of time, however, these emissions were slight,
ranging from 5 to 10 percent opacity and consisting of small puffs
that escaped from the enclosure through a narrow opening between the
front doors and the enclosure roof.58
4.7.7.2.4 Visible emissions during converter blister discharge.
At the end of copper blow, blister copper is discharged into a ladle
and transported to a refining furnace. Usually four ladles of blister
are poured per converter cycle. Each of the first three blister pours
lasts about 12 to 14 minutes, with one final blister pour lasting
about 4 minutes. The time between each blister pour is about 8 to
15 minutes.
At the end of a copper blow, the secondary hood doors and roof
are opened. An empty ladle is brought into the secondary hood by the
overhead crane and placed in front of the converter. The crane is
4-192
-------
moved out, and the secondary hood doors and roof are closed. The
converter is rolled out, and blister is poured into the ladle. After
the blister pour is completed, the converter is rolled in slightly,
the hood doors are opened, the blister ladle is taken to the refining
furnace by the crane, and the hood doors and roof are closed.
Four blister discharges were observed. Both EPA Methods 22 and 9
were used. A summary of the results obtained are presented in
Table 4-24.58 Although the observations periods used in obtaining the
EPA Method 22 data were inconsistent (i.e., different start and end
times), the results nonetheless indicate that visible emissions during
blister discharge were generally continuous. The EPA Method 9 data,
which were obtained only during periods when the blister copper was
actually being poured, show that the visible emissions observed were
somewhat more substantial than those observed during either matte
charging or slag skimming. As shown in Table 4-24, the highest average
opacity recorded for a single blister pour was 13 percent, with individ-
ual opacity readings ranging from 0 to 35 percent. Again, as with
slag skimming, the emissions observed generally appeared above the
narrow opening between the front doors of the enclosure and the enclos-
ure roof.
4.7.8 Removal of Particulate Matter From Fugitive Gases
Currently, the building evacuation baghouse at ASARCO's El Paso
facility and the calcine discharge baghouse at Phelps Dodge-Douglas
are the only control devices for which emission data exist that control
a gas stream composed only of fugitive gases. Gases from the converter
building evacuation system at El Paso are routed to a fabric filter
for particulate matter removal. The baghouse that receives the 16,800
NmVmin (600,000 scfm) of building evacuation gases contains 4,800 bags,
each 15 cm (6 in) in diameter by 9 m (30 ft) long. It is sized with a
nominal air-to-cloth ratio of 3 to 1 (cfm/ft2). The baghouse consists
of 12 compartments; however, only 10 are used at any given time. The
total net cloth area is approximately 19,730 m2 (212,400 ft2). The
baghouse was designed to treat 15,280 mVmin (540,000 acfm) at 54° C
(130° F). Mechanical shaking is used as the cleaning mechanism.
4-193
-------
TABLE 4-24. VISIBLE EMISSION OBSERVATION DATA FOR BLISTER
DISCHARGE AT THE TAMANO SMELTER58
Sample
run
1
2
3
4
Total
Method
Observation
period,
min
25a
--
15b
6C
15.3
22
Percent
of time
emissions
observed
42
--
86
19
49
Obser-
vation
period,
min
--
15.0
12.0
3.5
30.5
Method 9
Average
opacity for
observation
period,
percent
--
6.2
13.0
3.2
8.5
Range of
individual
opacity
readings
--
0 to 30
0 to 35
0 to 25
0 to 35
Observations started when secondary hood doors opened 12 minutes prior to
the blister discharge, during which time the converter body was hit by a
vibrating ram.
Observations started with the blister discharge and continued for 3 min-
utes after completion of the blister discharge.
'Observations started with the blister discharge and continued for 2h min-
utes after completion of the blister discharge.
4-194
-------
TABLE 4-25. SUMMARY OF EMISSIONS TESTING PERFORMED ON THE
CONVERTER BUILDING EVACUATION BAGHQUSE AT ASARCQ-EL PASO57 58
Particulate mass rate,
Grain loading, mg/Nm3 kg/h
Run no.
1
2
3
Average
Inlet
60.3
53.3
70.5
61.3
Outlet
11.6
2.5
1.1
5.1
Inlet
44.7
46.3
61.2
50.7
Outlet
10.4
0.92
6.4
3.9
efficiency, %
76. 7b
98.0
99.3
91.3
Calculated based upon inlet and outlet particulate mass rates.
bThis result is not considered typical of normal operation.
4-195
-------
TABLE 4-26. SUMMARY OF EMISSIONS TESTING PERFORMED ON THE
CALCINE DISCHARGE BAGHOUSE AT PHELPS
DODGE-DOUGLAS
Run No.
1
2
3
Average
Grain
Inlet
4,040
6,740
6,150
5,643
loading, mg/Nm3
Outlet
7.1
112.1
--
59.6
Participate
mass rate kg/h
Inlet
211
348
346
302
Outlet
0.37
1.92
.._
1.15
Collection
efficiency,
percent
99.7
99.5
--
99.6
Calculated based upon inlet and outlet particulate mass rates.
4-196
-------
Emissions testing has been performed by EPA to determine the
performance of this baghouse.57 The results of these tests are sum-
marized in Table 4-25. As indicated, the test results yield an average
collection efficiency of 91.3 percent based upon the three test runs
that were made. (See Appendix C for details.) It is evident, however,
that Run 1 does not exemplify normal operation of this baghouse.
Comparing the collection efficiencies estimated from Runs 2 and 3 with
the efficiency estimated from Run 1 supports this conclusion. Conse-
quently, for the purpose of determining the expected level of performance
in such an application, only the data from Runs 2 and 3 are utilized,
yielding an average collection efficiency of 98.7 percent. Because of
the amount of dilution that occurs as the fugitive gases are being
collected, no gas cooling is necessary prior to gas entry into the
baghouse. Dilution also causes the inlet grain loading to be quite
low, averaging 61.3 Mg/m3 over the three test runs.
The baghouse that collects particulate matter from captured
calcine discharge fugitive gases at the Phelps Dodge-Douglas smelter
has also been tested by EPA. The results of these tests are summarized
in Table 4-26. (See Appendix C for details.) As indicated, based
upon the two completed test runs, the average collection efficiency
was 99.6 percent. Consequently, a baghouse should be capable of
achieving a 99.6 overall particulate matter collection efficiency in
applications involving fugitive gases captured at the point of calcine
discharge.
4.8 REFERENCES
1. U.S. Bureau of Mines Staff. Control of Sulfur Oxide Emissions in
Copper, Lead, and Zinc Smelting. U.S. Bureau of Mines Informa-
tion Circular 9527. Washington, D.C. May 1971.
2. Rinckhoff, J. B., and W. R. Parish. Double Catalysis Sulfuric
Acid Plants for Copper Converter Gas. Presented at the 78th
AICHE National Meeting. Salt Lake City, Utah, August 18-21,
1974, AICHE Paper 45-D, 9 p.
3. Rinckhoff, J. B. Sulfuric Acid Plants for Copper Converter Gas.
In: Advances in Chemistry Series. Vol. 139. 1975. p. 48-59.
4-197
-------
4. Background Information for New Source Performance Standards:
Primary Copper, Lead, and Zinc Smelters, Volume I: Proposal
Standards. U.S. Environmental Protection Agency. Research
Triangle Park, N.C. Publication No. EPA 450/2-74-002a. October
1974.
5. Matthews, J. C., F. L. Bellegia, C. H. Gooding, and G. E. Weant.
S02 Control Processes for Nonferrous Smelters. Research Triangle
Institute. Research Triangle Park, N.C. Publication No. EPA-600/
2-76-008. January 1976.
6. Weisenberg, I. J. , T. Archer, F. M. Winkler, T. J. Browder, and
A. Prem. Feasibility of Primary Copper Smelter Weak S02 Stream
Control. Prepared for IERL, U.S. Environmental Protection Agency,
Cincinnati, Ohio, under EPA Contract No. 68-03-2378. Publication
No. EPA-600/2-80-152. July 1980.
7. Letter from Henderson, J. M., Consultant for ASARCO, Inc. to
Clark, T. C., Research Triangle Institute, August 4, 1982.
Clarification of comments on Chapters 3-6 of the BID.
8. Friedman, L. J. Production of Liquid S02) Sulfur, and Sulfuric
Acid from High Strength S02 Gases. In: Sulfur Dioxide Control
in Pyrometallurgy, T. D. Chatwin and K. Kikumoto (eds.). The
Metallurgical Society of AIME, Warrendale, Pennsylvania. 1981.
p. 205-20.
9. Burckle, John 0., and Christian A. Worrel. Comparison of Environ-
mental Aspects of Selected Nonferrous Metals Production Technologies.
(Presented at AIME Annual Meeting. Chicago. February 22-26,
1981.)
10. Letter from Wood, J. P. Research Triangle Institute, to Shafer,
J. , Monsanto Enviro-Chem Systems, Inc., St. Louis,, MO. July 1,
1981. Emission guarantees for acid plants.
11. Davenport, W. G. Copper Smelting to the Year 2000. CIM Bulletin.
73(813):152-158. January 1980.
12. Donovan, J. R., and P. J. Stuber. Sulfuric Acid Production from
Ore Roaster Gases. Journal of Metals. November 1967. p. 45-50.
13. Porter, F. L., and G. H. Wood. Analysis of Continuous S02 Monitor
Data for the Kennecott Copper Smelter at Garfield, Utah, and the
Determination of an Upper Limit for Sulfuric Acid Plant Catalyst
Deterioration. U.S. Environmental Protection Agency. Research
Triangle Park, N.C. April 18, 1973.
14. Monsanto Enviro-Chem Systems, Inc., Brochure No. 78-MON-7254.
Mist Eliminators. 1978.
4-198
-------
15. Letter and attachments from Billings, C. H., Arizona Dept. of
Health Services, to Wood, J. P., Research Triangle Institute.
May 12, 1982. p. 2. Response to letter requesting current
information on Duval scrubbing system.
16. Kohno, H. , and Y. Sugawara. S02 Pollution Control with the
Lime-Gypsum Process at the Onahama Smelter. In: Sulfur Dioxide
Control in Pyrometallurgy, T. D. Chatwin and K. Kikumoto (eds.).
The Metallurgical Society of AIME, Warrendale, Pennsylvania.
1981. p. 103-119.
17. Sulfur Oxide Removal from Power Plant Stack Gases: Use of Limestone
in Wet Scrubbing Processes. Prepared for the National Air Pollu-
tion Control Administration by the Tennessee Valley Authority.
NTIS No. PB 183908. 1969.
18. Slack, A. V. Application of Flue Gas Desulfurization in the
Non-Ferrous Metals Industry. In: Sulfur Dioxide Control in
Pyrometallurgy, T. D. Chatwin and K. Kikumoto (eds.). The Metal-
lurgical Society of AIME, Warrendale, Pennsylvania. 1981.
p. 91-101.
19. Background Information for New Source Performance Standards:
Electric Utility Steam Generating Units, Background Information
for Proposed S02 Emission Standards. U.S. Environmental Pro-
tection Agency. Research Triangle Park, N.C. Publication
No. EPA 450/2-78-007a. July 1978.
20. Argenbright, L. P. Smelter Pollution Control — Facts and Problems.
Mining Congress Journal. 57(5):24-28. May 1971.
21. Economic and Design Factors for Flue Gas Desulfurization Technol-
ogy. Prepared for the Electric Power Research Institute by
Bechtel National, Inc. Publication No. EPRI CS-1428. April 1980.
22. Friedman, L. J. Analysis of Modern Sodium and Ammonia Based SO
Scrubbing Systems. In: Sulfur Dioxide Control in Pyrometallurgy,
T. D. Chatwin and K. Kikumoto (eds.). The Metallurgical Society
of AIME, Warrendale, Pennsylvania. 1981. p. 189-203.
23. Sulfur Oxide Removal from Power Plant Stack Gases: Ammonia
Scrubbing. Prepared for the National Air Pollution Control
Administration by the Tennessee Valley Authority. NTIS No. PB
196804. 1970.
24. Maxwell, M. A., and G. R. Koehler. The Magnesia Slurry S02
Recovery Process Operating Experience With a Large Prototype
System. (Presented at the 65th AICHE Annual Meeting. New York.
November 26-30, 1972.) 36 p.
4-199
-------
25. Weisenberg, I. J. , A Prem, and F. Winkler. Primary Copper Smelter
Reverberatory Furnace S02 Control. (Presented at the 71st Annual
APCA Meeting. Houston. June 25-30, 1978.) 16 p.
26. K. Itakuma, H. Ikeda, and M. Goto. Double Expansion of the
Onahama Smelter and Refinery. The Metallurgical Society of AIME.
Warrendale, Pennsylvania. IMS Paper No. A74-11. 1974.
27. Niimura, M., T. Konada, and R. Kojima. Sulfur Recovery from
Green-Charged Reverberatory Offgas at the Onahama Copper Smelter.
Met. Society of AIME Paper No. A73-47. 1973. 26 p.
28. Petersson, S. A. The Production of Sulphuric Acid and Liquid
Sulfur Dioxide at the Ronnskar Works of Boliden Aktiebolag,
Skelleftehanm, Sweden. (Presented at the AIME Environmental
Control Process Symposium. 1972.)
29. Weisenberg, I. J., and R. C. Hill. Design, Operating, and Emission
Data for Existing Primary Copper Smelters (Draft). Pacific
Environmental Services. Santa Monica, California. EPA Contract
No. 68-02-2606. March 1978.
30. Letter and attachments from Henderson, J. M., ASARCO, to
Goodwin, D. R., U.S. Environmental Protection Agency. January 11,
1982. Response to Section 114 letter on primary copper smelters.
p. 15.
31. Weisenberg, I. J. , and J. C. Seme. Design and Operating Parameters
for Emission Control Studies: Phelps Dodge, Morenci Copper
Smelter. U.S. Environmental Protection Agency. Research Triangle
Park, N.C. Publication No. EPA-600/2-76-036g. February 1976.
20 p.
32. Reference 30, p. 31.
33. Goto, M. Green Charge Reverberatory Furnace Practice at Onahoma
Smelter. In: Extractive Metallurgy of Copper. J. C. Yannopoulos
and J. C. Agarwal (eds.). Port City Press, Baltimore, Maryland.
1976. p. 154-67.
34. Saddington, R., W. Curlock, and P. Queneau. Tonnage Oxygen for
Nickel and Copper Smelting at Copper Cliff. Journal of Metals.
18(4): 440-452. April 1966.
35. Itakura, K., H. Ikeda and M. Goto. Double Expansion of Onahoma
Smelter and Refinery. TMS Paper No. A74-11. Metallurgical
Society of AIME. Warrendale, Pennsylvania. 1974.
36. Eastwood, W. B., J. S. Thixton, and T. M. Young. Recent Develop-
ments in the Smelting Practice of Nchanga Consolidated Copper
Mines, Rokana Smelter. TMS Paper No. A71-75. Metallurgical
Society of AIME. 1971. 32 p.
4-200
-------
37. Pluzhm'kov, A. I., et al. The Possibility of Using Roof Firing
Reverberatory Furnaces. Tsvetnye Metally. 12(10). UDC 669.332.2.
October 1971. p. 7-11.
38. Kupryakov, Y. P., et al. Operation of Reverberatory Furnaces on
Air-Oxygen Blasts. Tsvetyne Metally, July 1972.
39. Wrampe, P., and E. C. Nollman. Oxygen Utilization in the Copper
Reverberatory Furnace: Theory and Practice. TMS Paper No. A74-25B.
Metallurgical Society of AIME. 1974. 18 p.
40. Beals, G. C. , J. Kocherhans, and K. M. Ogilvie. Reverberatory
Matte Smelting Process, U.S. Patent Office, Patent No. 3,222,162,
December 7, 1965.
41. Achurra, J. H. , R. G. Espinosa, and L. J. Torres. Improvements
in Full Use of Oxygen in Reverberatory Furnaces at Caletones
Smelter. TMS Paper No. A77-91. Metallurgical Society of AIME.
1977. 16 p.
42. Schwarze, H. 0., G. B. Vera, and F. 0. Pino. Use of New Technol-
ogies at Caletones Smelter. TMS Paper No. A77-90. Metallurgical
Society of AIME. 1977.
43. Niimura, M., T. Konado, and R. Kojima. Control of Emissions at
Onahama Copper Smelter. Onahama Smelting and Refining Co. (Presented
at Joint Meeting of MMIJ-AIME. Tokyo, Japan. 1972.) 14 p.
44. Blanco, J. A., T. N. Antonioni, C. A. Landolt, and G. J. Danyliw.
Oxy-Fuel Smelting in Reverberatory Furnaces at Inco's Copper
Cliff Smelter. Inco Metals Company, Copper Cliff, Ontario.
(Presented at the 50th Congress of the Chilean Instititue of
Mining and Metallurgical Engineers. Santiago. November 23-29,
1980. 16 p.
45. Biswas, A. K. , and W. G. Davenport. Extractive Metallurgy of
Copper. Oxford, England, Pergamon Press, 1980.
46. Queneau, P. E. Copper Making in the Eighties—Productivity in
Metal Extracting from Sulfide Concentrates. Journal of Metals.
33(2): February 1981.
47. Queneau, P. E. , and R. Schuhmann. Metamorphosis of the Copper
Reverberatory Furnace: Oxygen Sprinkle Smelting. Journal of
Metals. 31(12):12-15. December 1979.
48. Trip Report. Weisenberg, I. J., Del Green Associates, with
Garven, H.C., and T. N. Antonioni, Inco Metals Company. July 20,
1981. Oxygen enrichment in a calcine charged reverberatory
furnace at the Inco Copper Cliff Smelter.
4-201
-------
49. Successful Tests Encourage Phelps Dodge to Modify its Copper
Smelter. Chemical Engineering. 89(3):18-19. February 8, 1982.
50. Telecon. Clark, T. C., Research Triangle Institute, with G
arven, H. C., Inco Metals Company. July 6, 1982. Response to
questions on oxy-fuel smelting and Inco oxygen flash smelting.
51. Telecon. Clark, T. C., Research Triangle Institute, with
Garven, H. C., Inco Metals Company. July 9, 1982. Matte grade
considerations with oxy-fuel smelting.
52. Nichols, G. B., J. P. McCain, J. E. McCormack, and W. B. Smith.
Evaluation of an Electrostatic Precipitator for Control of Emis-
sions from a Copper Smelter Reverberatory Furnace. Prepared for
IERL, U.S. Environmental Protection Agency, Cincinnati, Ohio,
under EPA Contract No. R804762. Publication No. EPA-600/2-80-151.
June 1980.
53. Environmental Protection Agency. Response to Petitions for
Reconsideration and Revision of the Process Weight Regulation [40
CFR 52.126(b)] filed by the Phelps Dodge Corporation and Magma
Copper Company in October and November 1975. Prepared by the
U.S. EPA, Region IX Enforcement Division, San Francisco, California.
September 1978.
54. Wark, K., and C. F. Warner. Air Pollution; Its Origin and Control.
2nd Ed. New York. Harper and Rowe. 1981. p. 341-351.
55. Bowerman, L. J. Particulate Matter Emissions From Selected
Arizona Copper Smelters. (Presented at the APCA Annual Meeting.
Houston. June 25-30, 1978.)
56. Wet Scrubber System Study. Vol. 1, Scrubber Handbook. Prepared
for the U.S. Environmental Protection Agency by ATP, Inc. July
1972.
57. TRW Environmental Engineering Division. Emission Testing of
ASARCO Copper Smelter, El Paso, Texas. U.S. Environmental Protec-
tion Agency, Research Triangle Park, North Carolina. EMB Report
78-CUS-7. April 1978. 150V
58. Arsenic Emissions from Primary Copper Smelters—Background Informa-
tion for Proposed Standards (Draft). U.S. Environmental Protection
Agency. Research Triangle Park, N.C. November 1980.
59. Trip Report. B. H. Carpenter, T. C. Clark, and J. P. Wood,
Research Triangle Institute. J. Richardson, ASARCO-E1 Paso.
February 16, 1981. Familiarization Visit.
60. Trip Report. B. H. Carpenter, T. C. Clark, and J. P. Wood,
Research Triangle Institute. P. Bhargava, Arizona Bureau of Air
Quality Control. W. Cummins, C. Guptill, and W. Marczeski,
ASARCO-Hayden. February 20, 1981. Familiarization Visit.
4-202
-------
61. Batroun, V. V. Fundamentals of Industrial Ventilation. Trans-
lated by 0. M. Blunn. Pergamon Press. Third Edition. 1972.
62. Grassmuck, G. Applicability of Air Curtains as Air Stopping and
Flow Regulators in Mini Ventilation. C.I.M. Bulletin. No. 691.
62:1175-1185. November 1969.
63. Powlesland, J. W. Air Curtains in Controlled Energy Flows—To
Stop or Regulate Air Flows—To Contain and Convey Airborne Contami-
nants. Presented at the 22nd Annual Industrial Ventilation
Conference. February 1973.
64. ASARCO Design Report. Converter Secondary Hooding for the Tacoma
Plant. Prepared by ASARCO Central Engineering Dept., Salt Lake
City, UT. January 22, 1981.
65. Trip Report. B. H. Carpenter and J. P. Wood, Research Triangle
Institute. ASARCO, Salt Lake City, Utah, May 5, 1981. Familiar-
ization visit notes for discussion at ASARCO1s Central Engineering
Department, Salt Lake City.
4-203
-------
5. MODIFICATION AND RECONSTRUCTION
Under the authority of Section 111 of the Clean Air Act of 1970,
the new source performance standards (NSPS) are applicable to newly
constructed facilities and to existing facilities that have undergone
modification or reconstruction. A facility to which the standards
apply is termed an affected facility. An existing facility is one of
the type for which standards have been promulgated and for which the
construction or modification was begun prior to the proposal date of
the applicable standards.
The criteria used to identify modifications and reconstructions
are summarized in the Code of Federal Regulations (40 CFR 60), U.S.
Environmental Protection Agency (EPA) Standards of Performance for New
Stationary Sources, under Subpart A, "General Provisions," Sections
60.14 and 60.15.
5.1 SUMMARY OF 40 CFR 60 PROVISIONS FOR MODIFICATION AND RECONSTRUCTION
5.1.1 Modification
A modification is defined to be, with certain exceptions, "any
physical or operational change to an existing facility which results
in an increase in the emission rate to the atmosphere of any pollutants
to which a standard applies." However, a facility that undergoes such
a change is considered to be modified only if the cost of the change
as a percentage of the original facility cost is greater than the
annual guideline repair allowance percentage specified in the latest
edition of Internal Revenue Service Publication 534. Merely increasing
production to a higher level when adequate capacity exists is not
considered a modification. Other items that are not considered as
modifications include (1) maintenance, repair, or replacement that is
judged by the Administrator to be routine; (2) an increase in the
5-1
-------
hours of operation; (3) use of an alternative fuel or raw material if,
prior to the standard, the existing facility was designed to accommodate
that alternative use; and (4) the addition or use of any system or
control device whose primary function is the reduction of air pollutants,
except when an emission control system is removed or replaced by a
system considered to be less efficient.
5.1.2 Reconstruction
While a modification refers to relatively minor changes to an
existing facility, major changes are deemed reconstructions. Upon
reconstruction, an existing facility becomes a new source, irrespective
of any change in emission rate. Generally, a reconstruction occurs
when components of an existing facility are replaced to such an extent
that the fixed capital cost of the new components exceeds 50 percent
of the fixed capital cost that would be required to construct a compar-
able entirely new facility, and it is economically and technically
feasible to comply with the applicable standards.
The Administrator provides the final judgment that indicates
whether a replacement constitutes a reconstruction and if it is tech-
nologically and economically feasible to comply with the applicable
standards. The Administrator's final determination will be based upon
(1) a comparison of the fixed capital costs of the replacement compon-
ents with the costs of a comparable new source; (2) the estimated life
of the source after the replacements compared to the life of a compar-
able entirely new source; (3) the extent to which the components being
replaced cause or contribute to the emissions from the source; and
(4) any economic or technical limitations on compliance with applicable
performance standards which are inherent in the proposed replacements.
The purpose of the reconstruction provision is to ensure that an
owner or operator does not perpetuate an existing source by replacing
all but vestigial components, support structures, frames, and housing
rather than totally replacing the source in order to avoid being
subject to applicable new source standards.
5-2
-------
5.2 APPLICABILITY TO PRIMARY COPPER SMELTERS
5.2.1 General
Most of the changes to existing primary copper smelting facilities
that could be considered as possible modifications relate to increasing
production capacity. These changes are discussed under modifications.
Changes that would qualify as reconstructions are expected to be
relatively few. Major activities that might be considered as recon-
struction include rebuilding/relining of affected facilities such as
roasters, smelting furnaces, and converters. Rebuilding is an inherent
consequence of extracting copper by existing methods and is practiced
on a routine, though relatively infrequent, basis. Major reverberatory
furnace rebuilds, for example, have been reported as occurring every 5
to 8 years1 and every 20 years.2 Converter relining, either partial
or complete, typically occurs one to three times per year.3 Other
activities that might be considered reconstructions include physical
expansions of existing process equipment. These changes are discussed
in succeeding paragraphs.
5.2.2 Modifications
Various options that have been employed by the industry to increase
or expand production capacity are discussed here and are described in
greater detail in Section 3.4. Most of these options cause an increase
in particulate and S02 emissions proportional to the increased
production. A notable exception is the conversion from green- to
calcine-charged furnace operation. In this case S02 emissions will
not necessarily increase because some sulfur removal occurs in the
newly installed roaster.
5.2.2.1 Multihearth Roasters. As discussed in Section 3.4.1,
increasing the shaft rotation speed has been used as a means of
increasing the throughput rate of multihearth roasters by over 100
percent.4 However, most domestic smelters use the shaft rotation
speed as a means of altering residence time in order to control the
degree of sulfur elimination.5 Because these units are designed for
such operation, increasing shaft rotation speed to gain increased
throughput would not be considered a modification.
5-3
-------
Physical expansion of multihearth roasters is not considered
feasible because of the geometry of these units.
5.2.2.2 Fluid-Bed Roasters. Fluid-bed roasters can potentially
expand capacity by increasing blower capacity or by oxygen enrichment
of the fluidizing air. Increasing the blower capacity may be unfeasi-
ble because of constraints associated with existing calcine recovery
systems6 (see Section 3.4.2). As a result, this potential modification
is considered unlikely. Oxygen enrichment of the fluidizing air could
possibly be used to increase roaster throughput by 20 to 25 percent.6
However, some melting of the feed may result, which would lead to
operational problems (see Section 3.4.2). Therefore, this expansion
option is also considered unlikely.
It should be noted that the cost of an oxygen plant, if constructed,
would not be included in the determination of possible modification
because it is not a part of the affected facility.
5.2.2.3 Reverberatory Furnaces. Five different methods of
increasing reverberatory furnace production capacity are described in
Section 3.4.3. These methods are (1) conversion from green- to calcine-
charging, (2) addition of concentrate dryers, (3) physical expansion
of the furnace, (4) elimination of converter slag return, and (5) oxygen
enhancement techniques.
5.2.2.3.1 Conversion from green to calcine charging. Reverbera-
tory furnace capacity has been increased by up to 50 percent with the
conversion from green- to calcine-charged operation.:' Alterations to
the furnace would likely include the installation of water-cooled
panels around the furnace perimeter and changes to the feed system.8
Emissions of S02 from the altered furnace would not necessarily increase,
depending upon the extent of sulfur removal in the new roaster.
Particulate emissions from the expanded furnace should increase at
least in proportion to the increase in throughput achieved.
5.2.2.3.2 Addition of concentrate dryers. Capacity increases of
13 percent are feasible for green-charged reverberatory furnaces that
install concentrate dryers.9 Emissions of S02 and participates should
increase approximately in proportion to the increase in capacity achieved.
5-4
-------
In contrast to the conversion from green to calcine charging, no
changes would be anticipated at the furnace proper in order to process
the dried feed.9 Hence, this expansion option is not foreseen to be a
modification.
5.2.2.3.3 Physical expansion of the furnace. Furnace capacity
has been increased by 20 percent in the past by physical expansion.10
Emissions of S02 and particulates should increase in proportion to the
expansion obtained. Costs associated with physical expansion may be
sufficiently great to qualify it as a reconstruction. Other problems,
such as a substantial down-time requirement and possible physical
space limitations, may preclude the use of this option at some smelters.
Hence, physical expansion of reverberatory furnaces is considered an
unlikely expansion option.
5.2.2.3.4 Elimination of converter slag return. Processing
converter slag in separate facilities (e.g., flotation plants) rather
than returning it to the reverberatory furnace has been reported to
increase furnace capacity by up to 25 percent.11 Emissions of S02 and
particulates should increase in proportion to the increase in capacity
achieved. Changes to the furnace proper, if any, would be expected to
be minimal. Hence, this expansion option is not foreseen as a modifi-
cation.
5.2.2.3.5 Oxygen enchancement techniques. Reverberatory furnace
expansion options involving the use of oxygen are as follows: (1) the
enrichment with oxygen of the combustion air fed to existing burners,
(2) the injection of oxygen or oxygen-enriched air into the furnace
through lances positioned beneath the existing burners, (3) the addition
of roof-mounted oxy-fuel burners, (4) oxygen lancing through the roof,
or (5) the addition of roof-mounted oxy-sprinkle burners. These
options are discussed in detail in Section 3.4.3.5.2.
The technique of oxygen enrichment of the combustion air has
produced increases of up to 56 percent in furnace production capacity.12
Particulate and S02 emissions should increase approximately in propor-
tion to the increase in production. Costs inherent with this scheme
would be those associated with instrumentation and control for the
5-5
-------
system. It should be noted that, if an oxygen plant were required,
its construction cost would not be included for the determination of
possible modification because it is not part of the affected facility.
This statement also pertains to the other oxygen enhancement schemes
described below.
The injection of oxygen into the furnace through separate ports
or lances beneath the existing burners—undershooting—has produced a
36-percent increase in throughput with a calcine-charged furnace13 and
a 45-percent increase in throughput for a green-charged operation. 12
Particulate and S02 emissions should increase approximately in
proportion to the increase in furnace production. Costs inherent with
this scheme would be those associated with the lances and their
instrumentation/ control.
Large increases in capacity have been reported with oxy-fuel
burners installed in the furnace roof. In the case of green-charged
furnaces, production increases of more than 100 percent have been
reported,14 and a 45-percent increase in furnace capacity has been
reported for a calcine-charged furnace.15 The costs associated with
this scheme are expected to be those associated with the burners and
their instrumentation/control. Particulate and S02 emissions should
increase approximately in proportion to increases in production.
Production increases of 20 percent have been reported with the
use of roof-mounted oxygen lances.14 The costs inherent with this
scheme would be those associated with the lances and their instrumen-
tation/control. Particulate and S02 emissions should increase
approximately in proportion to increases in production.
The oxy-sprinkle process is an option that could be adopted by
the industry, although it is still in the development stage. Testing
of oxy-sprinkle burners on a reverberatory furnace is currently being
performed, and production increases of 100 percent have been reported.16
This scheme differs from the other techniques in that it is based on
the principle of flash furnace operation. Because considerable sulfur
in the feed is combusted, S02 emissions would be expected to increase
somewhat more than linearly with the increase in production achieved.
5-6
-------
Because of the nature of charging, particulate emissions would be
expected to increase somewhat more than in proportion to the increase
in capacity. Costs associated with this scheme would include, in
addition to burner and instrumentation/control costs, the costs of
additional feed handling equipment.
5.2.2.4 Electric Furnaces. Electric furnaces can increase
capacity by converting from green to calcine charging and by eliminat-
ing converter slag return. Other conceivable expansion options include
installing a larger transformer and physically expanding the furnace.
Additional discussion on each of these options is provided in Section
3.4.4.
As in the case of reverberatory furnaces, green-charged electric
furnaces may also expand capacity through the conversion to calcine
charging, because less time is required to smelt hot, roasted calcine
than dried concentrates. Production increases of 40 percent are
believed to be achievable. Costs associated with making the conversion
on electric furnaces are believed to be lower than those for reverbera-
tory furnaces because side-wall cooling and extensive feed system
changes would probably not be required. S02 emissions from the expanded
furnace throughput would not necessarily increase, depending upon the
extent of sulfur removal in the roaster. Particulate emissions should
increase in proportion to the increase in capacity.
Based on experience achieved with reverberatory furnaces, production
increases of 25 percent are believed to be achievable when converter
slag is processed in other facilities rather than returned to the
electric furnace. Emissions of S02 and particulates should increase
in proportion to the increased capacity achieved. Changes required at
the furnace proper appear to be minimal; hence, this expansion option
is not foreseen as a modificiation.
Electric furnaces can conceivably increase capacity by installing
a larger transformer to increase smelting rate by increasing power
input. Emissions of S02 and particulates should increase in proportion
to the increase in capacity achieved. However, this potential modifica-
tion would cause increased refractory wear17 and is considered unlikely.
5-7
-------
Physical expansion of an electric furnace is a conceivable option
but would require extensive changes to the furnace proper. In addition
to enlarging the furnace, a larger transformer and electrodes would be
required.17 Costs may be sufficiently great to qualify as a reconstmo-
tion. Emissions of S02 and particulates should increase in proportion
to the increase in capacity. Because substantial changes would be
required, this expansion option is considered unlikely.
5.2.2.5 Outokumpu Flash Furnaces. As discussed in Section
3.4.5, the primary expansion mode for Outokumpu flash furnaces is via
oxygen enrichment of the combustion air. Increases of 60 to 70 percent
have been reported.18 Emissions of S02 should increase in proportion
to the increased capacity achieved. Particulate emissions should
increase somewhat less than proportionately with the expansion because
of the lower offgas volume per unit of charge afforded by oxygen
enrichment. As discussed previously, oxygen plant costs would not be
included in the determination of possible modification.
Physical expansion is not considered to be a feasible expansion
option because of furnace geometry.19 The elimination of converter
slag return is not an available option for increasing capacity because
converter slag is not returned directly to the furnace in present
operating design.
5.2.2.6. Noranda Reactors. As discussed in Section 3.4.6, the
primary expansion mode for Noranda reactors is via oxygen enrichment
of the blowing air.20 Emissions of S02 and particulates should increase
approximately in proportion with the increase in capacity.
Other conceivable expansion modes for Noranda reactors include
increasing the blowing rate via the installation of a larger blower,
and physically expanding the vessel coupled with increasing the number
of tuyeres. Emissions of S02 and particulates should increase approxi-
mately in proportion to increases in capacity achieved using either of
these schemes. However, because of offgas handling constraints and—with
respect to physical expansion only-physical space limitations and
downtime requirements, neither scheme is considered likely.
5.2.2.7 Converters. Converter capacity can be increased by
physical changes to the vessel. Other conceivable expansion modes
5-8
-------
include increasing blower capacity and oxygen enrichment. All of
these options are discussed in further detail in Section 3.4.7.
Increases in converter capacity by approximately 13 percent have
been attained by physical expansion coupled with increasing the number
of tuyeres.21 Emissions of S02 and particulates should increase in
proportion to the increase in capacity achieved.
Converter throughput can potentially be increased by increasing
the size of the blower. Emissions of S02 and particulates should
increase in proportion to the increase in blowing rate. However,
excessive ejection of molten material from the vessel may occur, and
this expansion option is considered unlikely.
Oxygen enrichment of converter blowing air may increase converter
capacity, although any increase may be offset by interruptions in the
cycle arising from the need of additional "cold dope" materials.
5.3 REFERENCES
1. Telecon. Clark, T. C., Research Triangle Institute, with Johnson,
R. E. , Phelps Dodge Corportion. August 7, 1981. Industry expan-
sion options.
2. Discussion. Clark, T. C., Research Triangle Institute, with
Parker, D. J., ASARCO, Inc. October 1, 1981. Reconstructions of
process equipment.
3. Johnson, R. E., N. J. Themelis, and G. A. Eltringham. A Survey
of Worldwide Copper Converter Practices. In: Copper and Nickel
Converters, Johnson, R. E. (ed.). New York, American Institute
of Mining, Metallurgical, and Petroleum Engineers. 1979. p.
1-32.
4. Boggs, W. B., and J. N. Anderson. The Noranda Smelter. American
Institute of Mining, Metallurgical, and Petroleum Engineers
Transactions, Vol. 106, 1933. pp. 187-188.
5. Letter and attachments from Henderson, J. M., ASARCO, to Goodwin,
D. R., U.S. Environmental Protection Agency, January 11, 1982.
Response to Section 114 letter on primary copper smelters, p.
21.
6. Telecon. Clark, T. C., Research Triangle Institute, with Lee,
L. V., Dorr-Oliver, Inc. December 18, 1981. Increasing fluid-bed
roaster capacity.
5-9
-------
7. Mulholland, L. E., and D. J. Nelson. Operation of the Fluo-Solids
Roaster at Kennecott's Ray Mines Division. In: Copper Metallurgy,
Erlich, R. P. (ed.). New York, The Metallurgical Society of
AIME. 1970. pp. 141-145.
8. Telecon. Clark, T. C. , Research Triangle Institute, with Johnson,
R. E., Phelps Dodge Corporation. August 7, 1981. Industry
Expansion Options.
9. Telecon. Clark, T. C., Research Triangle Institute, with Malone,
R. A., Kennecott Minerals Company. August 11, 1982. Proposed
use of a concentrate dryer at the McGill smelter.
10. Reference 5, p. 8.
11. Itakura, K., T. Nagano, and J. Sasakura. Converter Slag Flotation--
Its Effect on Copper Reverberatory Smelting Process. Journal of
Metals. (Vol.):30-34. July 1969.
12. Kupryakov, Y. P., et al. Operation of Reverberatory Furnaces on
Air-Oxygen Blasts. Tsvetnye Metally. July 1972 (Vol )
p. 13-16. '
13. Saddington, R., W. Curlook, and P. Queneau. Tonnage Oxygen for
Nickel and Copper Smelting at Copper Cliff. Journal of Metals,
April 1966. p. 445.
14. Achurra H. , J. , R. Espinosa G., and L. Torres J. Improvements in
Full Use of Oxygen in Reverb Furnaces at Caletones Smelter. The
Metallurgical Society of A.I.M.E., Paper number A77-91, 1977.
15. Blanco, J. A., T. N. Antonioni, C. A. Landolt, and G. J. Danyliw.
Oxy-Fuel Smelting in Reverberatory Furnaces at Inco's Copper
Cliff Smelter. 50th Congress of the Chilean Institute of Mining
and Metallurgical Engineers, Santiago, Chile. November 1980.
16. Successful Tests Encourage Phelps Dodge to Modify its Copper
Smelter. Chemical Engineering. 8_9(3): 18-19. February 8, 1982.
17. Telecon. Clark, T. C., Research Triangle Institute, with Persson,
J. A., Lectromelt Corporation. August 10, 1981. Increasing
Electric Furnace Capacity.
18. Juusela, J., S. Harkki, and B. Andersson. Outokumpu Flash Smelting
and Its Energy Requirement. Effic. Use Fuels Metall. Ind. Symp.
Pap., Inst. Gas Techno!., Chicago. 1974. pp. 555-575.
19. Trip Report. Carpenter, B. H. , J. Wood, and C. Clark, Research
Triangle Institute, with Shaw, M. F., and A. S. Gillespie, Phelps
Dodge Corporation Hidalgo Smelter. February 17, 1981. Familiari-
zation plant visit.
5-10
-------
20. Telecon. Clark, T. C., Research Triangle Institute, with Weddick,
A. J., Kennecott Copper Corporation. August 19, 1981. Noranda
Reactors.
21. Reference 5, p. 3.
5-11
-------
6. MODEL PLANTS AND ALTERNATE CONTROL TECHNOLOGIES
6.1 INTRODUCTION
This study focuses on three principal issues:
The possible deletion of the existing exemption of new or
modified reverberatory furnaces from the new source perform-
ance standard (NSPS) when the smelter processes a high-
impurity charge.
The possible establishment of emission standards for fugitive
emission sources at primary copper smelters.
The effect of the NSPS on future capacity additions or
expansions at existing smelters.
In considering the deletion of the existing exemption, candidate
demonstrated technologies for the control of weak S02 process gases
from reverberatory furnaces are evaluated for their environmental,
cost, economic, and energy effects. A model plant representative of a
new "greenfield" smelter capable of processing high-impurity materials
is used for the analysis. Baseline conditions are the current NSPS.
The candidate technologies and model plant parameters are discussed in
Section 6.2.
Candidate demonstrated technologies for the control of fugitive
emissions from selected sources are also analyzed for their environ-
mental, cost, economic, and energy effects on new "greenfield" smelters
and modified or reconstructed existing facilities. Fugitive emission
sources selected for control are described in Section 6.3, along with
candidate technologies and model plants. Baseline conditions for
these sources are identified.
Expansion options available for each existing smelting configura-
tion are identified along with alternative control technologies capable
6-1
-------
of reducing emissions from expanded facilities to preexpansion levels.
Cumulative costs for the expansion are estimated, including the cost
associated with the physical or operational process change(s) resulting
in the capacity increase and the cost of added control to achieve
preexpansion emission levels, and a determination is made of whether
or not the total cost for a particular expansion scenario is prohibitive.
Model plants representative of existing U.S. smelters are developed.
Baseline conditions for these models reflect existing control levels.
Expansion options and alternative control technologies are discussed
in Section 6.4, along with the model plants selected as representative
of existing smelters. Procedures, used in developing model plant
parameters are described in Appendix J.
6.2 REVERBERATORY FURNACE EXEMPTION
Under the existing NSPS,1 new, modified, or reconstructed rever-
beratory smelting furnaces are exempt from S02 control when the total
smelter charge contains more than 0.2 weight percent arsenic, 0.1
weight percent antimony, 4.5 weight percent lead, or 5.5 weight percent
zinc. Total smelter charge includes all copper sulfide ore concentrates
plus all other solid materials introduced into the roasters and smelting
furnaces except calcine. Smelter charges containing a higher percentage
than that specified for one or more of the four impurities cited in
the exemption are high-impurity charges. When the existing standard
of performance was promulgated, the cost for controlling the weak S02
gases discharged from a reverberatory furnace directly with either a
sulfuric acid plant or then-available flue gas scrubbers was concluded
to be exorbitant. The U.S. Environmental Protection Agency (EPA)
stated at that time that it would continue to investigate means of
controlling S02 emissions from reverberatory furnaces, including the
use of oxygen enhancement of the furnace combustion air and the blending
of gases from reverberatory furnaces with those from multihearth
roasters and converters as a means of producing a gas stream suitable
for S02 control at reasonable cost. An additional! issue requiring
investigation is the need for a particulate standard for reverberatory
furnaces in the event that the SO^, control exemption for smelters that
process high-impurity feeds is not deleted.
6-2
-------
A model smelter representative of a new "greenfield" smelter
capable of processing high-impurity materials to be used in this study
is shown in Figure 6-1. The model is sized at a capacity of 1,364 Mg/day
(1,500 tons/day) of dry metal-bearing material, which is equivalent to
312 Mg/day (343 tons/day) of blister copper. This capacity is considered
to be representative of the smallest smelter that is likely to be
built today and is essentially equivalent to the median capacity of
existing domestic primary copper smelters. Procedures for determining
model plant parameters are described below. Example calculations are
shown in Appendix J.
As shown in Figure 6-1, the baseline case model smelter consists
of five multihearth roasters, one reverberatory smelting furnace, and
four conventional Pierce-Smith converters. The combination of multi-
hearth roasters and a reverberatory furnace is considered necessary by
the industry for processing materials containing a high level of
volatile impurities. In accordance with existing standards of perform-
ance, gases discharged from the multihearth roasters and converters
are treated in a double contact/double absorption (DC/DA) sulfuric
acid plant for S02 removal. Gases discharged from the reverberatory
furnace are treated in a hotside electrostatic precipitator (ESP) (not
shown) and subsequently are discharged to the atmosphere without S02
control. Fugitive emissions, produced by calcine discharge operations,
matte and slag tapping, and converter operations, are captured but not
controlled (not shown in Figure 6-1). Captured fugitive emissions are
assumed to be discharged to the atmosphere through a 100-meter stack.
The charge composition assumed for the model plant is shown in
Table 6-1. The weight percent presented for each charge component
listed is based on the average smelter charge processed at ASARCO's
Tacoma smelter in 1979.2 Also shown in Table 6-1 are values assumed
for the sulfur elimination achieved during roasting, smelting, and
converting at the model plant. The degree of sulfur elimination
achieved during converting (52 percent) was computed based on the
following assumptions:
6-3
-------
CONCENTRATES & FLUXES
I 6
MULTIHEARTH
ROASTER
(5)
AIR CALCINE
t t
REVER-
BERATORY
FURNACE
(1)
t I
i
2
SLAG
MATTE
SLAG
I
(4)
BLISTER
COPPER
DUAL-STAGE
ACID PLANT
100%
H2SO4
Hours/day Nm3/min (scfm) SO2 (%) O2 (%) Mg/day (tons/day)
1
2
3
4
5
6
7
24.0
24.0
4.0
5.2
9.5
2.3
1.7
1.3
4.0
5.2
9.5
2.3
1.7
1.3
24.0
830 (29,200)
3,315(117,000)
2,555 (90,100)
1,720 (60,600)
1,695 (59,800)
860 (30,300)
835 (29,500)
0(0)
3,380(119,400)
2,550 (89,900)
2,520 (89,000)
1,690(59,500)
1,665 (58,700)
830 (29,200)
2,235 (78,600)
4.5
1.7
6.3
5.5
6.7
5.5
7.9
0.0
5.9
5.2
6.0
5.0
6.2
4.5
0.106
16.5
11.9
13.2
13.3
13.2
13.3
13.1
0.0
14.0
14.3
14.3
14.9
14.8
16.5
-
136 (1,500)
798 (877)
Figure 6-1. Model plant for new "greenfield" smelter processing high-impurity materials.
6-4
-------
TABLE 6-1 MODEL PLANT CHARGE COMPOSITION AND SULFUR ELIMINATION
FOR GREENFIELD HIGH-IMPURITY SMELTER
Component
Copper
Iron
Sulfur
Arsenic
Antimony
Lead
Zinc
Other
Operation
Roasting
Smelting
Converting
aBased upon
Weight percent
22.9
19.6
27.0
3.9
0.38
0.88
1.33
24.01
a charge of 1,364 Mg/day
Amount contained in raw feed,
Mg/day (tons/day)
312.3 (343.5)
267.3 (294.0)
368.2 (405.0)
53.2 (58.5)
5.2 (5.7)
12.0 (13.2)
18.2 (20.0)
327.4 (360.2)
Sulfur eliminated,
Mg/day (tons/day)
71.1 (78.2) (19.3%)
104.7 (115.0) (28.4%)
192.3 (211.8) (52.3%)
(1,500 tons/day) of dry material to
the roaster.
6-5
-------
All the copper charged to the smelting furnace reports to
matte.
The matte produced contains 40 percent copper, 50 percent
iron and sulfur, and 10 percent impurities.
The copper, iron, and sulfur in the matte are in the form of
cuprous sulfide (Cu2S) and ferrous sulfide (FeS).
All the sulfur contained in the matte is eliminated by
oxidation to S02 during converter blowing.
The values shown for smelting and roasting are based on information
provided by ASARCO on its Tacoma smelter and indicate that approxi-
mately 60 percent of the sulfur not eliminated during converting is
eliminated during smelting and 40 percent during roasting.2 ASARCO/
Tacoma deliberately processes high-impurity concentrates for their
arsenic and other impurity content. Recovery of these impurities from
smelter offgases is by gas cooling and recovery of the condensed material
in an arsenic processing plant. Thus, the clean gases entering the
acid plant are essentially the same as those from a smelter processing
low-impurity materials.
The roaster offgas parameters shown in Figure 6-1 are based on
achieving an S02 concentration in the roaster offgases of 4.5 percent
while eliminating 71.1 Mg (78.2 tons) of sulfur per day. The 4.5-percent
S02 concentration selected is conservative when compared to the 5.0-per-
cent level considered achievable for new multihearth roasters (see
Section 3.2.1.1). Offgas parameters for the reverberatory smelting
furnace were computed based on the following assumptions:
The reverberatory furnace is fired with natural gas with a
heating value of 37 MJ/Nm3 (1,000 Btu/scf).
The energy required per megagram of material smelted is
5.2 thousand GJ (4.5 million Btu/ton).
All the sulfur eliminated in the furnace, 104.5 Mg/day
(115 tons/day), is by oxidation to S02.
The total volume of air entering the furnace is controlled
to maintain a 1-percent oxygen concentration in the furnace
uptake.
Dilution due to infiltration of air into the waste-heat
boilers, electrostatic precipitator, and flues downstream of
the furnace is 100 percent.
6-6
-------
Based on these assumptions, the aggregate flow rate—the components of
which include combustion products, S02, excess air, and dilution
air--for tne model furnace were calculated along with the S02 and 02
concentrations.
Distribution of sulfur elimination between the multihearth roaster
and reverberatory furnace, while maintaining a relatively constant
matte grade and S02 concentration in offgases, is dependent to a large
extent on the composition and nature of the concentrate; the amount of
pyritic sulfur and size distribution of the concentrate being the
principal parameter of interest. It is recognized that these param-
eters vary, in some cases from day to day and particularly at a custom
smelter, with commensurate changes in offgas S02 concentrations.
Admittedly, this creates difficulties in the design of a new smelter.
Although it also poses similar problems in developing model plant
parameters, EPA is mandated to assess the environmental, energy, and
economic aspects as an essential part of the standard development
process. Model plant analysis has been determined to be a reasonable
and cost-effective means of making these assessments. Model plants
are not intended to represent what an individual plant should look
like, but rather to form the bases for analyses of the impact of the
regulation on the industry as a whole. Therefore the average concen-
trate composition used at ASARCO-Tacoma in 1979 is used as the feed
for the new greenfield smelter model plant.
As shown in Figure 6-1, the offgas parameters associated with
converting are variable both in terms of flow and S02 concentration.
As discussed in Sections 3.2.3.1 and 4.5.1, this is due to the cyclic
nature of converter operations and the need to schedule multiple
converters to maintain a continuous flow of S02-bearing offgas of
sufficient S02 concentration to operate a DC/DA sulfuric acid plant
autothermally. The converter schedule adopted for the model plant is
presented in Figure 6-2. The schedule is based on the operation of
three converters to perform five converter cycles daily. The duration
of each cycle is 11 hours, with 6 hours required for slag blow, 3 hours
for the copper or finish blow, and a total of 2 hours for charging and
skimming. Intercycle time for each of the three converters is about
3% hours. The charge per cycle consists of 12 ladles of matte, which
6-7
-------
Converter 1
Converter 2
Converter 3
CT>
oo
Hours 5 10 15 20 *
t , i , , I , , I , , I , . I , i I i . I r i I • • I • • I ' • I • • ' • • ' • ' ' • • ' ' ' ' ' ' - I ' ' I i ' I ' ' I ' ' I ' ' * ' ' '
— . ,
1 I 1 I 1 1
, ^ , , , h- -T 1
Figure 6-2. Model smelter converter operating schedule.
-------
is added intermittently in a 5-2-2-2-1 sequence. Each ladle contains
13 Mg (14.3 tons) of matte.
Table 6-2 summarizes the six sets of conditions encountered and
the duration of each condition in hours per day resulting from the
converter schedule cited above. As indicated, the conditions range
from having all three converters idle for an aggregate of 1-1/3 hours
per day to having two converters on slag blow and one converter on
copper blow for an aggregate of 4 hours per day. The flow rates and
S02 and 02 concentrations shown in Figure 6-1 were computed based on
the estimated flow and sulfur elimination produced by an individual
converter during the slag blowing and copper blowing portions of a
converter cycle. The estimates were based on the following assumptions:
A slag blow lasts 6 hours and a copper blow lasts 3 hours.
A total of 156 Mg (172 tons) of matte composed of 50 percent
Cu2S, 40 percent FeS, and 10 percent impurities are processed
per converter cycle.
During slag blowing, all the FeS contained in the matte is
oxidized to FeO and S02.
During copper blowing, all the Cu2S contained in the matte
is oxidized to Cu and S02.
The impurities present in the matte and cold dope additions
result in no additional oxygen demand.
Air used for blowing is not enriched.
Oxygen efficiency during both blowing periods is 75 percent.
Dilution due to infiltration is 100 percent.
There are five converter cycles per day.
It should be noted that converter flows based on these assumptions
are calculated using mean flow rates for slag and copper blows (see
Appendix J for an example calculation). It is recognized that these
flows exhibit a variability due to the variability in blowing rates.
During a study of this variability, standard deviations of offgases
from copper and slag blows were determined to be approximately 15 per-
cent of the mean flows.3 Changes to acid plant flows shown in
Figure 6-1 and Table 6-3 that would result by applying one and even
two standard deviations to determine maximum acid plant flows for acid
6-9
-------
I
h-1
O
TABLE 6-2. MODEL PLANT—GREENFIELD HIGH-IMPURITY SMELTER REPRESENTATIVE CONVERTER
OFFGAS STREAM PROFILE
Converter status
Slag blow Copper blow Idle
2 1
2 0
1 1
1 0
0 1
0 0
Average
0
1
1
2
2
3
Hours/
day
4
5.
9.
2.
1.
1.
2
5
3
7
3
Total gas flow from Vol
converters, NmVmin (scfm) SO
2,550
1,720
1,690
860
840
0
1,610
(90
(60
(59
(30
(29
(56
,100)
,600)
,800)
,300)
,500)
,800)
6.
5.
6.
5.
7.
-
6.
. %
2
3
5
7
5
9
-
3
Vol
0
13
13
13
13
13
13
. %
2
.2
.3
.2
.3
.1
--
.2
5Based upon a three-converter operation performing five converter cycles daily.
""Assumes 100 percent dilution after the offgas offtake.
-------
plant sizing would be within the study grade estimates required for
subsequent analyses and would not materially change the results of
these analyses.4
The model plant controls the S02 from roasters and converters
with a DC/DA acid plant, consistent with the existing NSPS, which is
the baseline for new "greenfield" smelters. The acid plants for this
model are equipped with burners to heat the gases at startup and
during operation, if necessary. As indicated in Figure 6-1, the acid
plant yields 798 Mg/day (877 tons/day) of sulfuric acid.
Control technologies for the control of weak S02 gas from
reverberatory furnaces are described in Sections 4.3 and 4.4. The
technologies selected for analysis and their corresponding performance
levels are discussed below. Included are seven alternative control
techniques corresponding to three levels of control ranging from 44 to
98 percent.
Table 6-3 shows the model plant process parameters estimated for
the application of the S02 control alternatives. The first and least
stringent control alternative is partial blending--i.e., blending a
portion of the reverberatory offgas stream with the roaster and
converter offgas streams and treating the combined streams in a DC/DA
sulfuric acid plant. The level of blending selected, 45 percent,
reflects the maximum quantity of reverberatory furnace gas that could
be blended and maintain the blended gas-stream S02 concentration over
the autothermal limit (3.5 percent S02) of the acid plant at all
times, except when no converters are in the slag or copper blowing
mode. The maximum flow to the acid plant is 4,870 NnrVmin (172,000
scfm), and the minimum S02 concentration treated is 2.7 percent.
Control effectiveness achieved on the reverberatory furnace is 44 per-
cent. The remaining 55 percent of the reverberatory stream passes
through a hot ESP (315° C [600° F]) before discharge to the atmosphere.
Alternatives 2 and 3 are similar. In both cases the reverberatory
furnace gases are initially treated in a regenerative flue gas desul-
furization (FGD) process to upgrade the S02 concentration of the
gases. The resultant strong S02 gas stream discharged from the FGD is
6-11
-------
TABLE 6-3 MODEL PLANT, NEW GRfctNUtLU Hiun-
I-A
45% blending of
Stream Hours /day
Roaster gas
Reverberatory gas
Converter gas
Acid plant feed stream
FGD effluent, NmVmin (scfm)
4
5.2
9.5
2.3
1.7
1 -3
4
5.2
9.5
2 3
1.7
1 3
% S02 removal
Acid plant effluent,3 NmVmin (scfm)
NmVmin (scfm) % S02
830
3,315
2,555
1,720
1,695
860
835
0
4,870
4,035
4.010
3,180
3,150
2,320
3,680
% S02 removal
S02 control led,a'b Mg/day (tons/day)
S02 emissions, 3>C Mg/day (tons/day)
inn* ariri,a'd Mq/day (tons/day)
Supplemental .heat required in
acid plant, GJ/day (Btu/day)
—
( 29 200)
(117,000)
( 90,100)
( 60,600)
( 59,800)
( 30,300)
( 29,500)
( o)
(172,000)
(142,500)
(141.700)
(112,200)
(111,400)
( 81,900)
(130,100)
615
125
940
1
4.5
1.7
6.30
5.54
6.69
5.54
7.87
4.6
3.9
4.4
3.5
4.1
2.7
0.076
98 3
(677)
(138)
(1,037)
(7 x 106)
% 0?
16.5
11.9
13.2
13.3
13.2
13.3
13.1
13.4
13.5
13.4
13.5
13.4
13.6
-
Control Alternative
I-B
Treatment of reverberatory
gases with an MgO FGD system
followed by a DC/DA acid plant
NmVmin (scfm)
830 ( 29,200)
3,315 (117,000)
2,555 ( 90,100)
1,720 ( 60,600)
1,695 ( 59,800)
860 ( 30,300)
840 ( 29,500)
0 ( 0)
3,890 (137,300)
3,050 (107,700)
3,030 (106,900)
2,190 ( 77,400)
2,170 ( 76,600)
1,330 ( 47,100)
3 260 (115 200)
2,660 ( 94,000)
710
33.4
1,089
0
* S02
4.5
1.7
6.30
5.54
6.69
5.54
7.87
6.4
6.0
6.7
6.2
7.1
6.6
0.017
90.0
0.123
98.3
(781)
(36.7)
(1,198)
(0)
% 02
16.5
11.9
13.2
13.3
13.2
13.3
13.1
12.2
12.0
11.9
11.4
11.3
10.2
I-C
Treatment of reverberatory
gases with an NH3 system
followed by a DC/DA acid plant
NmVmin (scfm)
830 ( 29,200)
3,315 (117,000)
2,555 ( 90,100)
-,720 ( 60,600)
1,695 ( 59,800)
860 ( 30,300)
840 ( 29,500)
0 ( 0)
3,580 (126,500)
2,750 ( 97,000)
2,720 ( 96,200)
1,890 ( 66,700)
1,870 ( 65,900)
1,030 ( 36,400)
3,260 (115,200)
2,360 ( 83,300)
710
33.4
1,089
0
% S02
4.5
1.7
6.30
5.54
6.69
5.54
7.87
6.9
6.7
7.4
7.2
8.2
8.5
0.017
90.0
0.137
98.3
(781)
(36.7)
(1,198)
(0)
% 02
16.5
11.9
13.2
13.3
13.2
13.3
13.1
13.2
13.3
13.2
13.3
13.2
13.3
(rnntinued)
-------
TABLE 6-3 (continued)
I-D
Treatment of reverberatory gases
with a limestone FGD system,
Hours/ '
Stream day Nnr/min (scfm) % S02 % 02
Roaster gas 830 ( 29,200)
Reverberatory gas 3,315 (117,000)
Converter- gas 4 2,555 ( 90,100)
52 1,720 ( 60,600)
9.5 1,695 ( 59,800)
23 860 ( 30,300)
17 840 ( 29,500)
1.3 0 ( 0)
Acid plant feed stream 4 3,380 (119,400)
5.2 2,540 ( 89,800)
9.5 2,520 ( 89,000)
2 3 1,690 ( 59,500)
1 7 1,660 ( 58,700)
1.3 830 ( 29,200)
FGD effluent, NnvVmin (scfm) 3,250 (115,200)
I % S02 removal
£ Acid plant effl.,3 NnrVmin (scfm) 2,240 ( 78,600)
% S02 removal
S02 controlled,3 Mg/day (tons/day) 710
S02 emissions,3 Mg/day (tons/day) 28.0
100% acid,3'd Mg/day (tons/day) 798
Supplemental .heat required in 0
acid plant, GJ/day (Btu/day)
4.5
1.7
6.30
5.54
6.69
5.54
7.87
5.9
5.2
6.0
5.0
6.2
4.5
0.017
90.0
0.106
98.3
(781)
(30.8)
(877)
(0)
16.5
11.9
13.2
13.3
13.2
13.3
13.1
14.0
14.3
14.3
14.9
14.8
16.5
=^-=
Control Al
I-E
100% blending of
reverberatorv stream
NnvVmin (scfm)
830 ( 29,200)
3,315 (117,000)
2,555 ( 90,100)
1,720 ( 60,600)
1,695 ( 59,800)
860 ( 30,300)
840 ( 29,500)
0 ( 0)
6,690 (236,300)
5,860 (206,800)
5,830 (206,000)
5,000 (176,500)
4,980 (175,700)
4,140 (146,200)
5,460 (192,800)
724
12.
1,120
78
?Based on average converter flows and S02 concentrations.
Total S02 controlled by FGD and acid plant.
% S02
% 02
4.5 16.5
1.7 11.9
6.30 13.2
5.54 13.3
6.69 13.2
5.54 13.3
7.87 13.1
3 8 13.0
3.2 13.0
3.6 12.9
2.8 12.9
3.2 12.9
2.3 12.8
0.061
98.3
(796)
5 (13.8)
(1,230)
(7.5 x 107)
• •-
ternatlve
I-F
Oxygen enrichment
and 100% blending
of reverberatory stream
Nm3/min (scfm)
830
2,330
2,555
1,720
1,695
860
840
0
5,710
4,870
4,850
4,010
3,890
3,150
4,460
( 29,200)
( 82,200)
( 90,100)
( 60,600)
( 59,800)
( 30,300)
( 29,500)
( 0)
(201,500)
(172,000)
(171,200)
(141,700)
(140,900)
(111,400)
(161,100)
724
12.5
1,120
6
^Total S02 emissions from FGD and
Double contact/double absorption
% S02 % 02
4.5 16.5
2 4 11.9
6.30 13.2
5.54 13.3
6.69 13.2
5.54 13.3
7.87 13.1
4.5 12.5
3.9 12.4
4.3 12.3
3.5 12.2
4.0 12.1
3.0 11.9
0.074
98.3
(800)
(13.8)
(1,230)
(6 x 106)
acid plant.
acid plant.
I-G
Oxy-fuel and 100% blending
of reverberatory stream
NmVmin (scfm)
830 ( 29,200)
1,285 ( 45,300)
2,555 ( 90,100)
1,720 ( 60,600)
1,695 ( 59,800)
860 ( 30,300)
840 ( 29,500)
0 ( 0)
4,660 (164,600)
3,830 (135,100)
3,800 (134,300)
2,970 (104,800)
2,940 (104,000)
2,110 ( 71,500)
3,440 (121,200)
724
12.5
1,120
0
% S02
4.5
4.3
6.30
5.54
6.69
5.54
7.87
5.4
4.9
5.4
4.7
5.4
4.4
0.096
98.3
(800)
(13.8)
(1,230)
(0)
% 02
16.5
11.9
13.2
13.3
13.2
13.3
13.1
13.6
13.7
13.7
13.9
13.8
14.0
-------
then blended with the roaster and converter gases and subsequently
treated in a DC/DA sulfuric acid plant with an autothermal limit of
4.5 percent S02. Alternative 2 employs magnesium oxide (MgO) scrub-
bing, which upgrades the S02 concentration to 10 percent; Alternative 3
uses ammonia (NH3) scrubbing, which upgrades the S02 concentration to
25 percent. Both FGD processes are considered to be 90 percent
effective in recovering S02 fed to them. This, coupled with the
98.3 percent S02 recovery achievable in the DC/DA sulfuric acid plant,
results in a net control effectiveness for these alternatives of
88.5 percent for the reverberatory furnace alone and 95 percent
smelter-wide. As shown in Table 6-3, maximum flow to the DC/DA acid
plant is 3,890 NmVmin (137,300 scfm) and 3,580 NmVmin (126,500 scfm)
for Alternatives 2 and 3, respectively.
Alternative 4 is the application of nonregenerative lime/limestone
scrubbing, which produces a throwaway end product, for the control of
the reverberatory furnace offgases and the continued use of a DC/DA
sulfuric acid plant for the control of roaster and converter gas
streams. As discussed in Section 4.3.2, lime/limestone scrubbing
applied to a reverberatory furnace should achieve 90 percent control
of the reverberatory furnace S02 emissions.
Alternative 5 consists of blending 100 percent of the reverbera-
tory furnace gases with the roaster and converter gases and treating
the blended gas stream in a DC/DA sulfuric acid plant. Again, the
acid plant is designed for an autothermal limit of 3.5 percent S02.
The S02 control effectiveness achieved on the reverberatory furnace is
98.3 percent. The maximum flow to the acid plant is 6,690 NmVmin
(236,300 scfm). For this alternative, the S02 concentration of the
blended stream will be moderately below the 3.5 percent considered
necessary to operate the acid plant autothermally for a total of
10% hours. Although somewhat conservative, it is assumed for this
analysis that supplemental heat requirement for this alternative is
estimated at 78 GJ (7.5 x 107 Btu/day).
Alternatives 6 and 7 both use oxygen to enhance the S02 concentra-
tion of the reverberatory furnace gases prior to blending them with
6-14
-------
the roaster and converter gases, followed by treatment of the blended
gas stream in a DC/DA sulfuric acid plant with autothermal limits of
3.5 and 4.5 percent S02, respectively. As with Alternative 5, gas
blending, 98.3 percent S02 control is achieved with either alternative.
In the case of Alternative 6, oxygen enrichment, sufficient oxygen
(105 Mg/day [116 tons/day]) is introduced to upgrade the oxygen
concentration of the primary combustion air to the reverberatory
furnace to 25 percent. The result is a 40-percent increase in the S02
concentration of the reverberatory furnace gases and a 30-percent
decrease in the gas flow from the reverberatory gases against that
indicated on the model plant under baseline conditions. As discussed
in Section 4.4.6, the increase in S02 concentration and decrease in
gas flow are a result of a reduction in nitrogen dilution due to the
substitution of oxgyen for air and a reduction in the volume of
combustion gases generated due to a reduction (about 18 percent) in
the quantity of fuel required per ton of material smelted.
In the case of Alternative 7, oxyfuel burners are substituted for
the conventional air-fuel burners. Then, industrial grade oxygen
(199 Mg/day [219 tons/day]) is used to provide 100 percent of the
primary combustion air (see Section 4.4.6). This results in an increase
in S02 concentration (150 percent) and a decrease in gas flow (60 per-
cent) from the reverberatory furnace. Again, this is due to a decrease
in nitrogen dilution and a decrease in smelting fuel requirements
(40 percent).
As with gas blending (Alternative 5), the S02 concentration of
the blended roaster, reverberatory furnace, and converter gases under
Alternative 6 (oxygen enrichment) is projected to be below the auto-
thermal limit of the DC/DA sulfuric acid plant (3.5 percent) for an
aggregate of about 1-1/3 hours. As shown in Table 6-3, supplementary
heat required under Alternative 6 is 6 GJ/day (6 x 106 Btu/day).
Although the autothermal limit for the acid plant applied under
Alternative 7 (oxyfuel) is higher, no supplementary heat is needed
because the blended gas stream never falls below the autothermal
6-15
-------
limit. The maximum gas flow treated by the DC/DA sulfuric acid plants
under Alternatives 6 and 7 is 5,710 NmVmin (201,500 scfm) and
4,660 NmVmin (164,600 scfm), respectively.
As noted previously, in the event that the exemption of the
reverberatory smelting furnace from the S02 standard continues, a
separate standard for particulate control will be considered. Candidate
particulate matter control alternatives for reverberatory smelting
furnaces are identified and evaluated in Section 4.6. Those selected
for further analyses are:
Spray cooling followed by an ESP
Spray cooling to be followed by a fabric filter (baghouse).
In both alternatives the furnace offgas is cooled to about 110° C
(230° F) after leaving the hot ESP included in the baseline case.
As discussed in Section 4.6, the level of emission control or
emission reduction achievable with the application of a particulate
matter control device, as measured by EPA Method 5, is dependent on
the quantity of volatile, condensible particulate matter present in
the furnace offgases and the operating temperature of the control
device. The former is proportional, for the most part, to the quantity
of volatile impurities contained in the smelter charge. At a very
high-impurity smelter, such as that represented by the model plant,
the effectiveness of a hot ESP in controlling both the nonvolatile and
volatile constituents of the particulate matter emitted is estimated
to be about 50 percent. In contrast, a cold control device or a hot
and cold control device in series is demonstrated to achieve 97 to
99 percent control for both constituents. For a new smelter processing
high-impurity materials at or near the levels specified in the rever-
beratory furnace exemption, a cold control device alone would be
suitable. For a new smelter processing materials with a high level of
one or more volatile impurities as shown for the model plant in Table
6-1, a cold control device alone would not be suitable. Rather, two
control devices in series, one hot and one cold, would be required.
The hot control device (e.g., a hot ESP) would collect copper-rich
6-16
-------
dusts. The cold control device (e.g., spray chamber followed by a
baghouse) would collect the condensable constituents passing through
the hot device. Prescribing a particulate matter standard for an
exempted reverberatory furnace would require installation of a cold
device in addition to the hot ESP included in the baseline case.
Control parameters for each of the two particulate control alterna-
tives are shown in Table 6-4. The ESP has a design efficiency of
97 percent, a specific collection area of 980 m2 per actual cubic
meters per minute (AmVmin [300 ftVacfm]), and a design treatment
velocity of 9 to 12 cm/s (3 to 4 ft/s). Particulate matter from
reverberatory smelting furnaces is considered to have a low resistivity.
The fabric filter is multicompartmented, operates at an air cloth
ratio of 2.5 to 1, is equipped with a mechanical shaker, and uses
Dacron or Orion bags.
6.3 FUGITIVE EMISSION CONTROL
Fugitive emissions include those that escape from material transfer
operations, leakage from process vessels, and leakage from offgas
flues. Their control includes capture and collection. Capture is
accomplished by hoods or enclosures into which the emissions are drawn
by induced or natural draft.
A hot ESP for the control of particulate matter in reverberatory
furnace primary offgas streams is included in the baseline. It must
be recognized, however, that hot ESP control prior to cold control may
not be necessary in all cases. The amount of condensible material
(feed impurities removed in the smelting furnace) will dictate the
configuration of the particulate control system. For instance, with
low levels of condensibles in the furnace offgases, it would not be
necessary to separate the copper-bearing materials from the condensibles.
Consequently, evaporative cooling of the gas stream followed by a cold
ESP or fabric filter is adequate. However, for reverberatory furnaces
processing high-impurity materials, a different situation could exist.
The presence of relatively large amounts of condensibles in the offgases
may necessitate the use of a hot ESP in series with either a cold ESP
or a cold fabric filter. This configuration may be required so that
6-17
-------
TABLE 6-4. PARAMETERS FOR PARTICULATE CONTROL ALTERNATIVES--
PRIMARY OFFGASES FROM REVERBERATORY FURNACES
Control device Design bases
Electrostatic Specific collection area--980 mVAmVmin (300 ft2/acfm)
precipitator Treatment velocity--9 to 12 cm/s (3 to 4 ft/s)
Fabric filter Air-to-cloth ratio--2.5 to 1.0
Cleaning mechanism—mechanical shaking
Filter material--Dacron or Orion
6-18
-------
an adequate portion of the condensibles could be purged prior to dust
recycle. The possible need to purge a portion of the recycled condens-
ibles involves maintaining favorable conditions for impurity elimination
in the smelting furnace. The impurity level above which hot/cold
control is required could be determined through a rigorous thermodynamic
assessment of the smelting process; however, such an analysis is
beyond the scope of this study.
This study is limited to the control of particulate fugitive
emissions. Collection of these emissions may be accomplished by
fabric filters, ESP's, or scrubbers. Although the fugitive emissions
contain S02, their capture results in dilution to very low concentra-
tions, which are considered impractical to control. The capture and
collection of particulate emissions will, however, result in the
capture of the fugitive S02 emissions and their dispersal through a
stack with the results invariably being a reduction in the ambient S02
levels near the smelter.
Sources selected for possible regulation and their particulate
emission rates in kilograms per megagram of blister copper produced
are as follows:
Multihearth roasters: Calcine discharge, 5.20
Smelting furnaces: Matte tapping, 0.34; slag skimming,
0.31.
Converters: Blowing, 6.60; charging, skimming, and pouring,
3.34.
The ranking of these sources is detailed in Section 3.3.4.
Control techniques for these sources, including both capture and
collection, are described in Sections 4.7 and 4.6. Those selected for
evaluation along with their performance capability and design parameters
are presented in Table 6-5. The capture technique selected for calcine
discharge operations associated with multihearth roasters is the larry
car interlock ventilation system detailed in Section 4.7.4. The
ventilation rate applied over the duration of each larry car charge is
140 Mm3 (5,000 scfm). Capture effectiveness achieved is estimated at
90 percent. The model plant assumes that two larry cars are charged
6-19
-------
TABLE 6-5. SUMMARY OF FUGITIVE PARTICIPATE EMISSIONS CAPTURE AND CONTROL SYSTEMS
I
ro
o
Source
Roasters
Calcine discharge
Smelting furnaces
Matte tapping and
slag skimming
Converters
Blowing, charging,
skimming and
Capture system
Larry car interlock
ventilation system
Tap port and skim bay hoods,
ladle hoods, and close-
fitting launder covers
1. Air curtain and fixed
enclosure
2. Building evacuation
Desi9n Collection
ventilatTon _ ^ device,
Capture rate, Tempei ature,
efficiency, % NmVmin (scfm) °C (°F) ESP FF
90a 280 ( 10,000) 49 (120) -b X
90a 1,840 ( 65,000) 38-79 (100-175) -b X
90e 5,660 (200,000) 66 (150) -b X
95f 21,240 (750,000) 54 (130) - X
Collection
device
control
efficiency, %
99. 6C
499. Od
899. 6d
98.79
Overall
system
efficiency, %
=89.6
>89.1
689.6
593.8
aBased upon visible emissions data obtained at the ASARCO-Tacoma smelter.
bNot assessed due to lack of demonstration on fugitive sources.
C8ased on emissions test data obtained at Phelps Dodge-Douglas smelter.
dNo actual test data exist; however, analysis of the particulate size distribution involved indicates that removal efficiency should be at
as the indicated value.
eEstimate based upon visual observations made at ASARCO-Tacoma.
fBased upon visible emissions data obtained at the ASARCO-E1 Paso smelter.
9Based upon emissions test data obtained at the ASARCO-E1 Paso smelter.
least as high
-------
simultaneously five times per hour, or once every 12 minutes, for a
duration of 1 to 2 minutes per charge. Thus, the maximum flow to the
control device is 280 NmVmin (10,000 scfm) at 30° C (85° F).
Matte tapping and slag skimming controls consist of applying
local hooding and ventilation both at the tap port and at the launder
to ladle or slag pot transfer point. As noted in Section 4.7.5,
90 percent capture effectiveness is considered achievable at both
locations. For matte tapping, a total ventilation rate of 1,130 NmVmin
(40,000 scfm) is assumed--280 NmVmin (10,000 scfm) applied at the tap
port and 850 NmVmin (30,000 scfm) at the matte ladle. For slag
skimming, it is assumed that 140 NmVmin (5,000 scfm) is applied at
the tap port and 560 NmVmin (20,000 scfm) at the slag pot. Maximum
flow to the control device is set at 1,840 NmVmin (65,000 scfm) at
50° C (120° F). This assumes that one matte tap and one slag skim
would occur simultaneously.
Alternative techniques for the control of fugitive emissions from
copper converters are discussed in Section 4.7.6. These include both
local ventilation and general ventilation techniques. The local
ventilation techniques range from the application of fixed secondary
hoods—which achieve only marginal capture efficiencies, especially
during periods when the converter is rolled out for charging and
skimming—to the application of an air curtain/fixed enclosure hood
capable of achieving an estimated capture effectiveness of 90 percent
or better (see Section 4.7.6.2). The general ventilation technique
evaluated consists of building evacuation (see Section 4.7.6.1).
Under this technique, the structure housing the converters is enclosed,
and the resultant building volume ventilated at a rate sufficient to
prevent out-leakage from openings in the building and to maintain a
reasonable worker environment within the building. As noted in Sec-
tion 4.7.6.1, if properly applied, a capture efficiency of 95 percent
or better should be achievable using this capture technique. Both
building evacuation and the air curtain/fixed enclosure hood will be
evaluated as alternative bases for the possible regulation of fugitive
particulate matter emissions from copper converters.
6-21
-------
As noted in Table 6-5, the design flow rate for the building
evacuation system is 25,500 actual cubic meters per minute (AmVmin)
[900,000 acfm] at 55° C (130° F). This flow is based on an assumed
volume for the model plant converter building of 51,000 AmVmin (1.8 mil-
lion acfm) and an air change rate of 30 changes per hour. The air
change rate selected is nearly two times that applied at the ASARCO-E1 Paso
smelter to alleviate problems encountered at El Paso concerning elevated
worker exposure to heat and pollutants within the building.
Operating parameters assumed for the air curtain secondary hood
for various modes of converter operation are the same as those presented
in Table 4-18 for the proposed ASARCO-Tacoma secondary hood. The
maximum condition encountered is during charging or skimming, when
510 AmVmin (18,000 acfm) is provided to the air curtain slot and
2,300 AmVmin (82,000 acfm) is ventilated at the exhaust hood. The
design flow rate to the control device is 5,660 AmVmin (200,000 acfm)
at 65° C (150° F), which accommodates the worst-case situation antici-
pated for the model plant—i.e., two converters blowing and one converter
being charged or skimmed. Applicable control devices for the collection
of captured fugitive particulate emissions include ESP's, fabric
filters, and scrubbers. However, since performance data available are
limited to the application of fabric filters, fabric filters alone
will be analyzed. As noted in Section 4.6.3, although inlet conditions
may result in relatively dilute inlet concentrations, 98 to 99 percent
particulate matter control is demonstrated to be achievable. For the
purpose of this analysis, the fabric filters applied will be mechanical
shaker types equipped with multiple compartments, Orion or Dacron
bags, and operated at an air-to-cloth ratio of 2.5 to 1.
6.4 EXPANSION OPTIONS AND ALTERNATIVE CONTROL TECHNOLOGIES
For the purpose of analyzing smelter expansions, the following is
assumed:
Owners/operators will expand existing facilities.
Owners/operators will consider reducing emissions from each
expanded facility to levels at or below preexpansion levels
so that the expanded facility does not become subject to the
provisions of 40 CFR 60 for modified sources.
6-22
-------
The fixed capital cost of the expansion of an existing
facility would not exceed 50 percent of that required to
replace the existing facility entirely so that the expanded
facility does not become classified as a reconstruction.
Thus, insofar as expansion options are concerned, the objective of the
analysis is to determine if it is economically feasible to increase
production at existing smelters by increasing the capacity of existing
equipment and reducing emissions to preexpansion levels. Generally,
emissions from an expanded piece of process equipment will increase
proportionately to the increase in capacity achieved as a result of
the expansion.
As discussed in Section 3.2, there are seven distinct smelting
configurations used in the United States. Options available for
effecting a production capacity increase across each of these configura-
tions are discussed in detail in Section 3.4. As noted, the production-
rate- limiting process step at most existing U.S. primary copper smelters
is the smelting furnace. Thus, the expansion options evaluated focused
primarily on increasing production through the smelting furnace.
A listing of the smelting configurations and the expansion options
selected for analysis are shown in Table 6-6. As indicated in this
table, expansion options available for existing reverberatory furnaces
(Configurations I, II, and III) include 20-percent expansions at
green- and calcine-charged furnaces by oxygen enrichment of furnace
combustion air; a 40-percent expansion at green-charged furnaces by
conversion to a calcine charge; a 40-percent expansion at green-charged
furnaces by replacing conventional end-wall burners with roof-mounted
oxy-fuel burners; 50- and 100-percent expansion at Configurations I
and II smelters and 60-percent expansion at Configuration III smelters
by replacing the reverberatory furnace/roaster combination with an
oxygen flash furnace. Green-charged electric furnaces (Configuration
IV) may also be expanded by 40 percent by converting to calcine charge
and up to 100 percent by conversion to flash smelting. A 20-percent
expansion option is available for existing flash furnaces (Configura-
tion V) using oxygen enrichment of the flash furnace combustion air.
6-23
-------
TABLE 6-6. SMELTING CONFIGURATIONS/EXPANSION OPTIONS
Configuration
I. MHR-RV-CV
II. RV-CV
III. FBR-RV-CV
IV. EF-CV
V. FF-CV
Expansion options
Convert to Convert to
Percent Oxygen Oxy-fuel calcine flash
expansion enrichment burner charge smelting
20 X
50 X
100 X
20 X
50 X
40 X
50 X
100 X
20 X X
60
40 X
24 20 X
MHR = Multihearth roaster
RV = Reverberatory smelting furnace
CV = Converter
FBR = Fluid-bed roaster
EF = Electric furnace
FF = Flash furnace
6-24
-------
Additional options discussed in Section 3.4 but not considered
for analysis include physical expansion of the existing furnace,
elimination of converter slag return, and the use of oxy-fuel burners
on calcine-charged reverbera^ory smelting furnaces. Physical expansion
is currently regarded by the U.S. industry as an option unlikely to be
exercised. Elimination of converter slag return will require the
addition of slag treatment facilities, with necessary pollutant con-
trols. The extra investment, plus the additional operating costs of
another processing unit, would tend to exclude this alternative. Use
of oxy-fuel burners on calcine-charged furnaces using Wagstaff gun
charging is not considered demonstrated.
In projecting expansion of furnace capacity, the need for addi-
tional roaster and converter capacity must be considered. Sufficient
excess roaster and converter capacity is considered to be available at
most existing smelters to accommodate throughput increases up to
20 percent. Except for expansion options in which a flash furnace
replaces the smelting furnace, a new roaster and/or converter are
required for expansion options resulting in a more than 20 percent
capacity increase. Because of the higher matte grade after expansion,
no additional converter capacity is required for options involving
replacement of the reverberatory furnace by a flash furnace.
Model plants are used to assess the effect of the standards of
performance on future capacity expansions at existing smelters employing
any of the five smelting configurations for which expansion options
are available. The model plant configurations and the existing smelters
they represent are shown in Table 6-7. The Kennecott Garfield smelter,
which uses Noranda reactors, is not represented in the models because
no viable expansion option is available other than the addition of a
new Noranda reactor controllable using a DC/DA acid plant. The Cities
Service smelter--which uses a fluid-bed roaster, electric furnace, and
converter—is not represented because this smelter operates primarily
for the production of sulfuric acid. The White Pine smelter also is
not represented because it processes native copper rather than sulfide
ore concentrates.
6-25
-------
TABLE 6-7. MODEL PLANT CONFIGURATIONS AND EXISTING
U.S. SMELTERS
Existing smelter
model Smelters represented
I. MHR-RV-CV ASARCO-Tacoma
ASARCO-Hayden
ASARCO-E1 Paso
Phelps Dodge-Douglas
II. RV-CV Kennecott-Hurley
Kennecott-McGill b
Phelps Dodge-Morenci
Phelps Dodge-Ajo
Magma
III. FBR-RV-CV Kennecott-Hayden b
Phelps Dodge-Morenci
IV. EF-CV Inspiration
V. FF-CV Phelps Dodge-Hidalgo
a CV = Converter.
EF = Electric furnace.
FBR = Fluid-bed roaster.
FF = Flash furnace.
MHR = Multihearth roaster.
RV = Reverberatory furnace.
bHas both green-charge and calcine-charge RV.
6-26
-------
Each model plant is sized at 1,364 Mg/day (1,500 tons/day) to
allow for analyses of alternative expansion options on a common basis.
Flow charts showing the model plant configurations are presented in
Figures 6-3 through 6-7. Converter S02 emissions from Model Plants I
and II and roaster and converter S02 emissions from Model Plant III
are treated in a single contact/single absorption (SC/SA) sulfuric
acid plant. Converter and smelting furnace S02 emissions from Model
Plants IV and V are treated in a DC/DA sulfuric acid plant. Process
offgases from the reverberatory smelting furnaces in Model Plants I,
II, and III are passed through wasteheat boilers and then through a
hot ESP (not shown in Figures 6-3, 6-4, and 6-5) to remove particulate
matter prior to stack discharge. Fugitive emissions are neither
captured nor controlled. Assumed feeds, matte grades, and sulfur
elimination ratios for each of the five model plants are shown in
Table 6-8.
The data shown in Figures 6-3 through 6-7 are average gas flows
and concentrations. Acid plants are sized to accommodate the maximum
flows that occur when two of the three active converters are on slag
blow and one on copper blow (Model Plants I through IV) and one active
converter on slag blow and one on copper blow (Model Plant V). These
maximum flows and associated S02 and 02 concentrations are shown in
Figures 6-3 through 6-7.
Several control alternatives are considered for each expansion
option. Each combination of an expansion option and a control alterna-
tive is considered an expansion scenario. A total of 26 expansion
scenarios will be analyzed. The control alternatives considered
include:
Blending with strong streams and treatment in a sulfuric
acid plant.
Treatment in a nonregenerative FGD, i.e., lime/limestone.
Treatment in a MAGOX regenerative FGD, i.e., blending the
FGD gas with other strong streams, and treatment in a
sulfuric acid plant.
6-27
-------
CONCENTRATES & FLUXES
Is
MULTIHEARTH
ROASTER
(5)
AIR CALCINE
1
i
SLAG
I
« • _____ '
REVER- 2
RERATORY
FURNACE
I MATTE
SLAG I
3
CONVbHTER
(4)
BLISTER
COPPER
I
Nm3/min (scfm) SO2 (%
1 2.000 (75,000) 1.5
2 5,980(211,200) 0.8
3 2,300 (81,300) 4.3
4 2,160 (76,300)
5
6
Maximum 3,660 (129,200) 4.3
flow to
acid plant
* ACID PLANT
100%
H2SO4
I 6
) O2 (%) Mg/day (tons/day)
18.7
11.4
15.4
—
1,364(1,500)
551 (606)
15,7
— _
Figure 6-3. Model Plant I for expansion of existing smelters.
6-28
-------
AIR CONCENTRATES & FLUXES
4
1 I
SLAG
1
1
REVERBERATORY
FURNACE
(D
I MATTE
SLAG |
I I
2
(4)
I
BLISTER
COPPER
1
Nm3/min (scfm) S02 (%)
1 5,940 (209,900) 1.4
2 2,770 (98,000) 4.3
3 2,600 (92,000)
4
5
Maximum 4,510 (159,200) 4.3
flow to
acid plant
SINGLE-STAGE
ACID PLANT
I
100%
H2SO4
I-
O2 (%) Mg/day (tons/day)
11.0
15.4
—
1,364(1,500)
667 (734)
15.5
Figure 6-4. Model Plant II for expansion of existing smelters.
6-29
-------
CONCENTRATES & FLUXES
FLUID-BED
ROASTER
(1)
AIR CALCINE
SL
AG
t t
REVER-
BERATORY
FURNACE
(1)
| MATTE
SLAG I
J
CONVERTER
(4)
BLISTER
COPPER
SINGLE-STAGE
ACID PLANT
100%
H2S04
I,
1
2
3
4
5
6
7
Maximum
flow to
acid plant
Nm3/min (scfm)
1,040 (36,700)
5,010 (176,900)
1,330(46,800)
2,370 (83,500)
2,100 (74,200)
3,160 (111,700)
S02 (%)
9.6
0.4
6.5
7.9
7.5
02 (%)
11.4
11.4
11.7
11.6
12.4
Mg/day (tons/day)
1,364(1,500)
1,034(1,138)
Figure 6-5. Model Plant III for expansion of existing smelters.
6-30
-------
SLAG
CONCENTRATES & FLUXES
I 5
ELECTRIC
FURNACE
(1)
MATTE
CONVERTER
(4)
DUAL-STAGE
ACID PLANT
1
2
3
4
5
6
Maximum
flow to
acid plant
Nm3/min (scfm)
1,160 (40,800)
3,330(117,600)
4,490 (158,400)
4,170 (147,300)
6,540 (230,800)
SO2 (%)
6.0
4.3
4.7
-
4.6
f\ in/ i
Uj \f<*t
19.7
11.3
13.5
-
16.2
Mg/day (tons/day)
1,364(1,500)
1,224(1,346)
Figure 6-6. Model Plant IV for expansion of existing smelters.
6-31
-------
AIR CONCENTRATES & FLUXES
11
1
SLAG
r
SLAG
FURNACE
(1)
SL
AG
FLASH
FURNACE
(D
1
I
MATTE
I
SLAG
MATTE
CONVERTER
(4)
BLIS
COP
TER
PER
2
' 3 DUAL-STAGE
" ACID PLANT
I
100%
H2SO4
i
1
2
3
4
5
6
Maximum
flow to
acid plant
Nm /min (scfm)
1,570(55,500)
1,180(41,700)
2,810 (99,200)a
2,490 (88,100)
3,400 (120,700)
SO2 (%)
10.3
4.3
7.7
-
7.7
02 (%)
5.0
16.1
10.0
-
11.0
Mg/day (tons/day)
1,364(1,500)
1,202(1,322)
Includes air to maintain 1.1 02 to SO2 ratio.
Figure 6-7. Model Plant V for expansion of existing smelters.
6-32
-------
TABLE 6-8. MODEL PLANT EXPANSION SCENARIOS. EXIT GASES, COMPOSITION AND FLOW RATE
a.b.c
I
U>
CO
SaeUing furnace
Model olant
I MHR-RV-CV
1
2
3
4
5
6
II RV-CV
7
8
9
10
11
12
13
14
15
16
17
III FBR-RV-CV
18
19
20
21
22
IV EF-CV
23
24
25
V FF-CV
26
Applicability Feed
Entire plant 1,364 (1,500)
Entire plant 1,636 (1,800)
Entire plant 1,636 (1,800)
Entire plant 1,636 (1,800)
Entire plant 1,636 (1,800)
New flash furnace 2,045 (2,250)
and old converter
New flash furnace 2,727 (3,000)
and old converter
Entire plant 1,364 (1.500)
Entire plant 1,636 (1,800)
Entire plant 1,636 (1,800)
Entire plant 1,636 (1,800)
Entire plant 1,636 (1,800)
Old converter 1,364 (1,500)
New converter and 682 (750)
expanded reverb
Old converters 1,364 (1,500)
New converter and 682 (750)
expanded reverb
Old converters 1,364 (1,500)
New converter and 682 (750)
expanded reverb
Old converters 1,364 (1,500)
New converter and 682 (750)
expanded reverb
Old converter and 1,909 (2,100)
reverb
New FBR and new 1,909 (2,100)
converter
New flash furnace 2,045 (2,250)
and old converters
New flash fi/rrace 2,727 (3,000)
and old converters
Entire plant 1,364 (1,500)
Entire plant 1,636 (1,800)
Entire plant 1,636 (1,800)
Entire plant 1,636 (1,800)
Entire plant 1,636 (1,800)
New flash furnace 2,182 (2,400)
and old converters
Entire plant 1,364 (1,550)
Entire plant 1,909 (2.100)
New flash furnace 2,046 (2,250)
and old converters
New flash furnace 2,728 (3,000)
and old converters
Entire plant 1,364 (1,500)
Entire plant 1,636 (1,800)
Roaster
HW« (scf») X SOj % 02 X SO,
1,995 (75,000) 1.5 18.7 0.8
2,400 (84,600) 1.5 18.7 1.1
2,400 (84,600) 1.5 18.7 1.1
2,400 (84,600) 1.5 18.7 1.1
2,400 (84,600) 1.5 18.7 1.1
1.4
2.0
- 2.0
2.0
2.0
- 3.0
3.0
3.0
3.0
0.8
1,040 (36,700) 9.6 11.4 0.4
1,250 (44,100) 9.6 11.4 0.6
1,250 (44,100) 9.6 11.4 0.6
1,250 (44,100) 9.6 11.4 0.6
1,250 (44,100) 9.6 11.4 0.6
840 (29,700) 9.6 11.4 1.0
XOZ
11.4
13.1
13.1
13.1
13.1
11.0
13.1
13.1
13.1
13.1
15.6
15.6
15.6
15.6
11.4
11.4
13.2
13.2
13.2
13.2
20.0
To
5,980
4,210
4,250
4,100
4,110
5,945
4,195
4,145
4,100
4,100
2,755
2,640
2,595
2,595
15.600
5,010
3,515
3,475
3.435
3,430
- -
hWn (scfl)
air
(211,200)
(148,600)
(150,100)
(145,100)
(145,100)
(209,900)
(148,200)
(146,400)
(144,700)
(144,700)
(97,200)
(93.200)
(91,600)
(91,600)
(198.000)
(176,900)
(124,100)
(122.600)
(121,200)
(121,200)
- -
To control
895 (31,600)
945 (33,400)
995 (35,100)
995 (35,000)
245 (8,620)
325 (11,500)
895 (31,500)
940 (33,400)
990 (35,000)
990 (35,000)
1,440 (50,900)
1,555 (54,900)
1,600 (56,500)
1,600 (56,500)
280 (9,870)
375 (13,200)
750 (26,400)
790 (27,900)
830 (29,300)
830 (29,300)
305 (10,700)
1,155 40,800
1,620 (57,200)
295 (10,400)
390 (13,800)
1,570 (55,500)
1,420 (50,000)
N»V>
2,305
2,760
2,760
2,760
2,760
1,845
2,465
2,775
3,330
3,330
3,330
3,330
2,770
970
2,775
970
2,775
970
2,775
970
2,775
780
2,105
2,810
1,330
1,590
1,590
1,590
1,590
2,275
3,330
(old) 3.330
(new) 935
2,220
2,965
1,180
1,415
Converter
(scf«)
(81,300)
(97.600)
(97,600)
(97,600)
(97,600)
(65,200)
(87,000)
(98,000)
(117,600)
(117,600)
(117,600)
(117,600)
(98,000)
(34,200)
(98,000)
(34,200)
(98,000)
(34,200)
(98,000)
(34,200)
(98.000)
(27,600)
(74,400)
(98,100)
(46,800)
(56,200)
(56,200)
(56,200)
(56,200)
(80,400)
(117.600)
(117.600)
(33,100)
(78,500)
(104,600)
(41.700)
(50,000)
X SO,
4.3
4.3
4.3
4.3
4.3
4.3
4.3
4.3
4.3
4.3
4.3
4.3
4.3
4.3
6.2
4.3
6.2
4.3
6.2
4.3
6.1
4.3
4.3
6.5
6.5
6.5
6.5
6.5
4.3
4.3
4.3
6.2
4.3
4.3
4.3
4.3
X02
15 4
15.4
15.4
15 4
15 4
16.3
16 3
15 4
15.4
15.4
15.4
15.4
15.4
13.3
15.4
13.3
15.4
13.3
15.4
13.3
15.4
13.3
16.3
16.3
11.7
11.7
11.7
11.7
11.7
16.3
15.4
15.4
13.2
16.5
16.5
16.1
16.1
Material balances nay not close due to rounding
bAU flows are at I atn, 70° F.
""Based on average flows.
Includes air to dilute flash furnace offgas to 11 percent S02
-------
TABLE 6-9 MODEL PLANTS FOR EXPANSION OPTIONS: REPRESENTATIVE
FEEDS, MATTE GRADES, AND SULFUR ELIMINATION RATES
Model Plant
I II III IV V
Component, wt. %
Copper 21.9 22.4 19.6 28.7 21.2
Iron 20.4 24.8 23 25.3 24.3
Sulfur 24.8 28.4 28.9 29.8 31
Cu/S 0.88 0.79 0.68 0.96 0.68
Matte grade 40 36 40 40 55
Sulfur eliminated,
Mg/100 Mg feed
Roaster 4.2 14
Smelter 6.7 11.7 2.8 9.7 21.4
Converter 13.9 16.7 12.1 20.1 9.6
6-35
-------
Treatment in a NH3 regenerative FGD, i.e., blending the FGD
gas with other strong streams, and treatment, in a sulfuric
acid plant.
Direct treatment in a DC/DA sulfuric acid plant.
Table 6-8 provides the exit gas parameters developed as the basis
for analyzing each of the 26 expansion scenarios. The five model
plants are shown together with their expansion scenarios. Exit gas
flows for the models shown are based on S02 and oxygen concentrations
reported by the actual smelters represented. Pertinent information on
matte grades and sulfur eliminations assumed for each model is shown
in Table 6-9. For oxygen enrichment and oxy-fuel scenarios, the
furnace gas flows are based on maintaining a 1-percent oxygen concen-
tration at the smelting furnace offtake (equivalent to approximately
10 percent excess oxygen). Flows to the atmosphere and sulfuric acid
plant are based on 100 percent dilution from all sources; i.e., smelting
furnace, waste heat boiler, and gas treatment system. Postexpansion
smelting furnace flows and S02 concentrations for the oxygen enrichment
and oxy-fuel expansion options are estimated using the mathematical
model shown in Appendix K.
6.5 REFERENCES
1. Standards of Performance for New Stationary Sources: Primary
Copper, Zinc, and Lead Smelters. Federal Register. January 15,
1976. pp. 2332-2341.
2. Letter, from Henderson, J. M., ASARCO, to Goodwin, D. R., U.S.
Environmental Protection Agency, January 11, 1982. Response to
Section 114 letter on primary copper smelters.
3. Letter from Varner, M. 0., ASARCO, to Vervaert, A. E. , U.S.
Environmental Protection Agency, January 13, 1983. Comments on
draft BID Chapters 6-8.
4. Memorandum, Massoglia, M. F., Research Triangle Institute, to
Vervaert, A. E., U.S. Environmental Protection Agency, January 20,
1983. Subject: ASARCO comments on revised BID Chapters 6-8.
6-36
-------
7. ENVIRONMENTAL IMPACT
7.1 GENERAL
This chapter identifies the beneficial and adverse environmental
impacts associated with the application of the alternative control
technologies selected in Chapter 6 for the control of (1) S02 and
particulate matter at a new greenfield smelter processing high impurity
materials and (2) fugitive particulate matter emissions from new
multihearth roasters, smelting furnaces, and converters. The impacts
addressed include those on air, water, solid waste, and energy. As
outlined in Sections 6.2 and 6.4, impact assessments will be based on
model plants considered representative of the domestic industry as it
exists today. Detailed procedures used in estimating impacts are
described in Appendix J.
7.1.1 New Greenfield High-Impurity Smelters—Process Emissions
As discussed in Chapter 6, seven control alternatives were selected
for control of reverberatory furnace process streams. These represent
three distinct levels of control ranging from 89 percent to 98 percent.
A summary of each control alternative and corresponding performance
level is shown in Table 7-1.
The baseline against which the control alternatives are compared
includes a double contact/double absorption sulfuric acid plant (DC/DA)
for control of multihearth roaster and converter strong S02 streams as
required by the current NSPS, no S02 control on the reverberatory
smelting furnace offgas streams, and hot electrostatic precipitator
(ESP) control of particulate matter in the reverberatory furnace
process gas stream.
In addition, two alternatives for further control of reverberatory
furnace process particulate matter are assessed in the event that
7-1
-------
TABLE 7-1 EVALUATED CONTROL OPTIONS FOR CONTROL OF PROCESS S02
EMISSIONS AT A GREENFIELD COPPER SMELTER (MULTIHEARTH ROASTER-
REVERBERATORY SMELTING FURNACE-CONVERTER) PROCESSING
HIGH-IMPURITY MATERIALS
Control level (%)
Control alternative RV only Smelterwide
I Baseline
0 70
I-A Blending 45 percent of RV offgas 44 83
with MHR and CV off-gases followed
by treatment in a DC/DA
I-B Treatment of RV off gas in a 89 95
MgO FGD. Blending of strong FGD
stream with MHR and CV off-gases,
followed by treatment in a DC/DA
I-C Same as I-B except on NH3 FGD is used 89 95
I-D Treatment of RV offgases in a lime- 90 96
stone FGD. MHR and CV stream to a
DC/DA
I-E Blending 100 percent of RV offgas 98 98
with MHR and CV offgases followed
by treatment in a DC/DA
I-F Same as I-E except that oxygen 98 98
enrichment is used to enhance
S02 concentration of RV offgas
I—G Same as I-E except that oxy-fuel 98 98
burners are used in RV to enhance
S02 concentration of off gas ==
MHR = Multihearth roaster.
RV = Reverberatory smelting furnace.
CV = Converter.
DC/DA = Double contact/double absorption sulfuric acid plant.
7-2
-------
subsequent analysis indicates that the current exemption of reverberatory
furnace processing high impurity materials should be retained. These
alternatives are (1) gas cooling and cold ESP with an overall control
effectiveness of 99.0 percent and (2) gas cooling and a baghouse with
an overall control effectiveness of 99.7 percent.
7.1.2 New Greenfield High-Impurity Smelters—Fugitive Emissions
Control alternatives and their performance levels for fugitive
emission sources at a new greenfield smelter are shown in Table 7-2
for multihearth roaster-reverberatory furnace-converter (MHR-RV-CV)
smelters and Table 7-3 for flash furnace-converter (FF-CV) configura-
tions. As discussed in Chapter 6, the sources selected for possible
regulation include multihearth roaster calcine discharge operations,
smelting furnace matte tapping and slag skimming operations, and
converter operations. Capture systems analyzed include both local and
general ventilation, with collection of the captured particulate
matter by fabric filters.
Baseline conditions with regard to fugitive emissions are the
capture and venting to the atmosphere of fugitive emissions from the
operations listed in the preceding paragraph.
7.2 AIR POLLUTION IMPACT
7.2.1 SO, Controls for Reverberatory Smelting Furnaces
Table 7-4 summarizes the S02 emission impact of each of the
control alternatives for new reverberatory smelting furnaces in the
MHR-RV-CV configuration.
As shown in Table 7-4, the annual reduction in S02 emissions from
the baseline case for reverberatory furnaces in the MHR-RV-CV configura-
tion ranges from 32,450 Mg (35,700 tons) per year for Alternative IA
(45 percent blending) to 72,110 Mg (79,320 tons) per year for Alterna-
tives IE (100 percent blending), IF (oxygen enrichment), and IG (oxy-
fuel). Corresponding emission reductions on the reverberatory furnace
only are 44.2 and 98.3 percent, respectively, and for the entire
smelter, 82.9 and 98.3 percent, respectively. Also shown in Table 7-4
are annual emission reductions expressed as kilograms S02 per megagram
7-3
-------
i
-Pa
TABLE 7-2 EVALUATED ALTERNATIVES FOR CONTROL OF FUGITIVE PARTICIPATE EMISSIONS AT A
' GREENFIELD COPPER SMELTER PROCESSING HIGH-IMPURITY MATERIALS
(MULTIHEARTH ROASTER-REVERBERATORY SMELTING FURNACE-CONVERTER)
Control
alternative
II baseline
IIA
IIB
IIC
IID
HE
IIP
Converter: blowing,
charging, skimming and pouring
Capture by air curtain and
secondary enclosure, no
collection
Capture, baghouse
Capture, baghouse
Capture, baghouse
Building evacuation,
baghouse
Building evacuation,
baghouse
Building evacuation
baghouse
Multihearth roaster:
calcine discharge
Capture by larry car
interlock and local
hooding, no collection
Capture, no collection
Capture, baghouse
Capture, baghouse
Capture, no collection
Capture, baghouse
Capture, baghouse
Reverberatory
furnace: matte
and slag skimming
Capture by local
hooding, ladle hoods,
launder covers, no
collection
Capture, no collection
Capture, no collection
Capture, baghouse
Capture, no collection
Capture, no collection
Capture, baghouse
-------
TABLE 7-3 EVALUATED ALTERNATIVES FOR CONTROL OF FUGITIVE PARTICULATE EMISSIONS
AT A GREENFIELD COPPER SMELTER (FLASH FURNACE-CONVERTER)
Control
alternative
Converter blowing, charging, skimming, and pouring
Flash furnace matte
tapping and slag skimming
III baseline
IIIA
IIIB
IIIC
HID
Capture by air curtain and secondary enclosure.
No collection
Capture, baghouse
Capture, baghouse
Building evacuation, baghouse
Building evacuation, baghouse
Capture by local hooding, ladle
hoods, launder covers, no collection
Capture, no collection
Capture, baghouse
Capture, no collection
Capture, baghouse
I
tn
-------
TABLE 7-4 AIR POLLUTION EMISSION IMPACT OF S02 CONTROL ALTERNATIVES FOR A NEW
GREENFIELD SMELTER, MULTIHEARTH ROASTER-REVERBERATORY FURNACE-CONVERTER
Blister copper
(Mg/day)
RV only
S02 emissions
(Mg/year)a
S02 reduction
(Mg/year)a
S02 reduction
(kg/Mg blister
copper)
Control (%)
RV only
MHR-RV-CV
S02 emissions
(Mq/year)a
S02 control (%)
Control al
I
(Baseline) I-A I-B I-C
312 312 312 312
73,360 40,910 8,460 8,460
0 32,450 64,900 64,900
0 296 593 593
0 44.2 88.5 88.5
76,600 44,050 11,590 11,590
70.3 82.9 95.5 95.5
ternati ves
I-D I-E I-F I-G
312 312 312 312
7,340 1,250 1,250 1,250
66,020 72,110 72,110 72,110
603 659 659 659
90.0 98.3 98.3 98.3
10,470 4,380 4,380 4,380
95.9 98.3 98.3 98.3
MHR = Multihearth roaster.
RV = Reverberatory smelting furnace.
CV = Converter.
Controlled emissions based on 350 days operation per year.
-------
of blister copper. Using these data, estimates of emission reductions
resulting from different size smelters can be made.
7.2.2 Fugitive Particulate Emissions
Table 7-5 summarizes the air pollution fugitive particulate
emission impacts of each of the control alternatives for fugitive
emissions from the new MHR-RV-CV and FF-CV smelter configurations.
Reduction in fugitive particulate matter emissions ranges from 0.6 kg/Mg
of blister to 9.4 kg/Mg with reduction for converter operations at the
high end of this range and smelting furnaces at the low. Because of
the higher capture efficiency, systems employing building evacuation
result in a slightly higher reduction for converters than do air-curtain
systems. Reductions in fugitive particulate matter emissions per unit
of blister produced are lower for smelters employing flash smelting
technology because of lower fugitive emissions prior to control.
7.2.3 Expansion Scenarios
There is no change in S02 emissions to the atmosphere from smelting
furnaces under any of the expansion scenarios because the control
alternatives are designed to reduce postexpansion process emissions
from the smelting furnace to preexpansion levels. However, there is
an increase in S02 and particulate process emissions from the roasters
and/or converters and for scenarios in which expansion occurs by
conversion to flash smelting. For the 20-percent expansion options
(Scenarios 1-10, 16-22, 26), the increased emissions from the roasters
and/or converters would not require control under the modification
provisions because the increased throughput is considered to be within
the design characteristics of the equipment and therefore not subject
to NSPS. For the expansion scenarios requiring a new roaster and/or
converter (Scenarios 11-15, 23), the S02 process emissions from the
new roasters and/or converters are subject to the limitations of the
existing NSPS and must be controlled to NSPS level. For the scenarios
involving conversion to flash smelting (Scenarios 5, 6, 16, 17, 22,
24, and 25), the new flash furnaces are subject to the limitations of
the existing NSPS and must be controlled to NSPS level.
7-7
-------
TABLE 7-5. AIR POLLUTION FUGITIVE PARTICULATE EMISSION IMPACT FOR
EACH SOURCE AND CONTROL ALTERNATIVES-NEW GREENFIELD SMELTERS'1
Blister copper (Mg/day)
Multi hearth Roasters
Number
Fugitive particulate
emissions (total) b
Baseline (Mg/yr)
Controlled (Mg/yrr
Reduction (Mg/yr)
Control (%)
Reduction (kg/Mg blister)
Smelting Furnace
Number
Fugitive particulate
emissions (total) d
Baseline (Mg/yr)
Controlled (Mg/yr)e
Reduction (Mg/yr)
Control (%)
Reduction (kg/Mg blister)
Converters
Number
Fugitive particulate
emissions (total) f
Baseline (Mg/yr)
Air Curtain
Controlled (Mg/yr)9
Reduction (Mg/yr)
Control (%)
Reduction (kg/Mg blister)
Building Evacuation
Controlled (Mg/yr)"
Reduction (Mg/yr)
Control (%)
Reduction (kg/Mg blister)
MHR-RV-CV
configuration
312
6.0
568
62
506
89
4.6
1.0
71
8
63
89
0.6
4.0
1,092
119
973
89
8.9
64
1,028
94
9.4
FF-CV
configuration
312
On
.0
NA
NA
NA
NA
NA
1A
.0
66
12
54
89
0.5
4.0
798
n~j
87
711
89
6. Ei
47
751
94
6.9
MHR = Multihearth roaster.
RV = Reverberatory smelting furnace.
CV = Converter.
FF = Flash furnace.
aBased on 350 days operation per year. Capture and venting of fugitive
particulate emissions to the atmosphere through a stack will result
in reduction of local ambient SOa concentration.
bCapture by larry car interlock and local hooding. No collection.
€Capture by larry car interlock and local hooding. Collection by
baghouse.
dCapture by local hooding, ladle hoods, launder covers. No collection.
eCapture by local hooding, ladle hoods, launder covers. Collection by
baghouse.
^Captured by air curtain and secondary hood. No collection.
9Capture by air curtain and secondary hood. Collection by baghouse.
hCapture by building evacuation. Collection by baghouse. No reduction
in fugitive S02 emissions.
7-8
-------
As with process emissions from the 20-percent expansion options,
any increase in fugitive S02 and particulate emissions from roasters
and/or converters would not require control that may be recommended
under a revised NSPS. Under the 20-percent expansion scenarios, the
increased throughput is considered to be within the design characteris-
tics of the equipment and therefore not subject to the NSPS. However,
the increased capacity of the smelting furnace is not within the
design characteristics and would therefore be subject to a revised
NSPS including limits on furnace tapping operation. Converters for
the expansion scenarios involving conversion to flash smelting would
not be subject to any fugitive emission limitations that might be
included in a revised NSPS because no capital expenditure is required
and no operational or physical change in the converting process occurs.
For the expansion scenarios requiring the addition of new equipment,
the new roaster and/or converter as well as the expanded smelting
furnace would be subject to any fugitive emission limitations that
might be included in the revised NSPS.
Reduction in fugitive particulate emissions for the expansion
scenarios are shown in Table 7-6. These reductions are from the
baseline, which has no capture or collection system for fugitive
emissions.
7.3 WATER POLLUTION IMPACT
Potentially significant sources of water pollution associated
with the control of weak S02 offgases from reverberatory smelting
furnaces are the following:
Gas cleaning and conditioning systems associated with contact
sulfuric acid plants and flue gas desulfurization (FGD) systems.
Absorbent purges taken from FGD systems in order to prevent
the buildup of impurities in absorbent recycle streams.
Water pollution impacts are based on neutralizing gas cleaning scrubbing
water with sulfuric acid and determining stoichiometrically the volume
of liquid (calculated as water) requiring disposal. A detailed methodology
is presented in Appendix L.
7-9
-------
TABLE 7-6.
AIR POLLUTION FUGITIVE PARTICULATE EMISSION IMPACT FOR EXPANSION AT EXISTING SMELTERSC
Qnpltinn furnace
Expan-
sion
scenario
1-4
5
6
7-10
11-14
15
16
17
18-21
22
23
24
25
26
Added
blister
capacity
(Mg/yr)
20,910
52,270
104,550
21,380
53,450
42,760
53,470
106,940
18,710
56,HO
54,790
68,510
137,010
20,240
Emissions (Mq/yr)
Baseline
80
95
125
85
105
100
100
130
75
90
115
125
165
75
Controlled0
10
10
15
10
10
10
15
15
10
10
15
15
20
10
Reduction
70
85
110
75
95
90
85
115
65
80
100
110
145
65
Reduction
(kg/Mg
blister)
3.5
1.6
1.1
3.5
1.7
2.0
1.6
1.1
3.5
1.4
1.9
1.6
1.1
3.2
Converter
Emissions (Mq/yr)
Baseline
f
f
f
f
535
435
f
400
f
f
Controlled6
f
60
45
f
T
45
f
T
Reduction
f
f
T
f
T
f
t
475
390
f
f
T
f
1
•f
355
f
Reduction
(kg/Mg
bl ister)
f
f
f
T
f
8.9
8.9
f
f
f
f
6.5
f
i
•e
f
aBased on 350 days operation per year.
Capture by local hooding at tap ports, skim ports, and ladle and launder covers.
Capture by local hooding at tap ports, skim ports, and ladles an launder covers.
dCapture by air curtain and secondary hood. No collection.
eCapture by air curtain and secondary hood. Collection by baghouse.
fNot applicable; no new units required.
No collection.
Collection by baghouse.
-------
7.3.1 Gas Cleaning and Conditioning Systems
Gas cleaning and conditioning systems normally employ one or more
wet (water) scrubbers, which serve to remove particulate matter from
the gas stream as well as to cool and humidify the gas stream. Since
the gas streams contain S03, the recirculated scrubbing water actually
begins to form a sulfuric acid solution that could theoretically reach
50 to 60 percent sulfuric acid if a portion was not purged and fresh
makeup water added. Therefore, a portion of the scrubbing water is
purged from the recirculation loop, creating a liquid effluent that
must be disposed of. Common procedure involves neutralizing this
liquid with limestone, which in turn produces an effluent that consists
primarily of calcium sulfate (CaS04) and water. Estimates of the
calcium sulfate (a solid waste) and liquid effluent (primarily water)
produced by this neutralization process are presented in Table 7-7 for
the greenfield smelter and in Table 7-8 for the expansion scenarios.
7.3.2 FGD Absorbent Purges
While the liquid effluent that exists after neutralization of the
gas cleaning system purge will consist primarily of water, it is quite
likely that the liquid will be saturated with calcium sulfate. Thus,
the potential for groundwater pollution becomes evident. Adequate
technology exists, however, to prevent liquids of this nature from
seeping into underground and surface water supplies. Ponding has been
a technique favored by a number of industries for waste disposal
problems of this type and should provide for containment of the liquid
effluent produced by the neutralization process. The use of impermeable
pond liners—such as polyvinyl chloride, polyethylene, polypropylene,
and nylons--should prevent seepage into groundwater supplies, while
proper closed-loop operation of the ponds will prevent pond liquor
overflow that could contaminate surface water supplies.
Similar situations would exist where the liquids generated by FGD
absorbent purges are involved. In the case of the nonregenerative
lime/limestone scrubbing system, the liquid effluent produced would be
similar in nature to that produced by the neutralization procedure
discussed above. Estimates of the amount of liquid generated by this
source are presented in Table 7-9 for the greenfield smelter and in
7-11
-------
TABLE 7-7 ESTIMATED PRODUCTION RATE OF SOLID AND LIQUID EFFLUENTS REQUIRING DISPOSAL FROM
GAS CLEANING AND CONDITIONING EQUIPMENT, GREENFIELD SMELTERS
Volume of gas|s to
acid plant ,
Nm3/min, (scfm)
Effluent production
rate associated
with gas cleaning
and conditioning
in acid plant, Mg/yr
Effluent production
rate associated
with gas cleaning
and conditioning
FGD system employed in FGD systems, Mg/yr
Total effluent
production rate, Mg/yr
Incre-
mental increase in
in effluent produc-
tion relative to
the base case, Mg/yr
Design flow
rate
Base Ca?e
I-A
I-B
I-C
1-0
I-E
I-F
1-G
2,440 (86,200)
3,920 (138,400)
2,935 (103,600)
2,635 (93,100)
2,440 (90,000)
5,730 (202,300)
4,470 (168,400)
3,725 (131,500)
CaS04
2
4
3
2
2
5
4
3
Liquid
14
23
17
16
14
34
28
22
Type
NA
NA
MgO
NH3
Lime-
stone
NA
NA
NA
HmVmin. (scfm) CaS04
NA
NA
3,315
3,315
3,315
NA
NA
NA
NA
NA
(117,000)
(117,000)
(117,000)
NA
NA
NA
0
0
3
3
2
0
0
0
Liquid
0
0
19
19
16
0
0
0
CaS04
2
4
6
5
4
5
4
3
Liquid
14
23
36
35
30
34
28
22
CaS04
--
2
4
3
2
3
2
1
Liquid
--
9
22
21
16
20
14
8
NA = Not Applicable
aBased on average converter offgas flows.
-------
TABLE 7-8.
ESTIMATED INCREMENTAL INCREASE IN EFFLUENTS REQUIRING DISPOSAL FROM GAS CLEANING AND
CONDITIONING EQUIPMENT, EXPANSION OPTIONS
Volume
aci
Effluent
production rate associated
of gases to be treated in with gas cleaning and condi-
d plants, NmVmin (scfm) tioning in acid plants, Mg/yr
Preexpansion
Base Case I
1
2
3
4
5
6
— 1
, . Base Case II
OJ 7
8
9
10
11
12
13
14
15
16
17
Base Case III
18
19
20
21
22
Base Case IV
23
24
25
Base Case V
26
2,305
2,305
2,305
2,305
2,305
2,305
2,305
2,775
2,775
2,775
2,775
2,775
2,775
2,775
2,775
2,775
2,775
2,775
2,775
2,365
2,365
2,365
2,365
2,365
2,365
4,485
4,485
4,485
4,485
2,750
2,750
(81,300)
(81,300)
(81,300)
(81,300)
(81,300)
(81,300)
(81,300)
(98,000)
(98,000)
(98,000)
(98,000)
(98,000)
(98,000)
(98,000)
(98,000)
(98,000)
(98,000)
(98,000)
(98,000)
(83,500)
(83,500)
(83,500)
(83,500)
(83,500)
(83,500)
(158,400)
(158,400)
(158,400)
(158,400)
(97,100)
(97,100)
Preexpansion
Postexpansion CaS04
2,305 (81,300)
3,660 (129,180)
2,765 (97,600)
2,860 (101,100)
2,800 (99,000)
3,510 (124,000)
4,685 (165,200)
2,775 (98,000)
4,220 (149,100)
3,330 (117,600)
3,505 (123,700)
3,400 (120,000)
5,185 (183,100)
3,745 (132,200)
4,175 (147,300)
3,915 (138,200)
4,365 (154,200)
4,010 (141,700)
5,350 (187.800)
2,365 (83,500)
3,585 (126,700)
2,840 (100,300)
2,880 (101,700)
2,855 (100,800)
4,345 (153,400)
4,485 (158,400)
6,730 (237,000)
3,240 (114,500)
4,325 (152,700)
2,810 (99,200)
2,890 (102,100)
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
3
3
2
2
2
2
2
2
4
4
4
4
3
3
Liquid
' 14
14
14
14
14
14
14
16
16
16
16
16
16
16
16
16
16
16
16
14
14
14
14
14
14
26
26
26
26
16
16
New FGD requirements
Postexpansion
CaS04
2
3
3
3
3
3
4
3
4
3
3
3
5
3
4
4
4
4
6
2
3
3
3
3
4
4
6
3
4
3
3
Li quid
14
22
16
17
17
21
28
16
25
20
21
20
31
22
25
23
26
23
30
14
21
17
17
17
26
26
40
19
26
16
17
Type
NA
NA
CaC03
MgO
NH3
NA
NA
NA
NA
CaC03
MgO
NH,
NA'
CaCO,
MgO
NH,
NA'
NA
NA
NA
NA
CaC03
MqO
NH3
NA
NA
NA
NA
NA
NA
NA
Design
flow rate,
Nm3/min (scfm)
NA NA
NA NA
940 (33,400)
990 (35,000)
990 (35,000)
NA NA
NA NA
NA NA
NA NA
940 (33,400)
990 (35,000)
990 (35,000)
NA NA
1,555 (54,900)
1,600 (56,500)
1,600 (56,500)
NA NA
NA NA
NA NA
NA NA
NA NA
790 (27,900)
830 (29,300)
830 (29,300)
NA NA
NA NA
NA NA
NA NA
NA NA
NA NA
NA NA
Effluent
product! on
rate associ-
ated with
gas clean-
ing and condi-
tioning in FGD
systems, Mg/yr
CaS04
0
0
1
1
1
0
0
0
0
1
1
1
0
1
1
1
0
0
0
0
0
1
1
1
0
0
0
0
0
0
0
Total pre-
expansi on
effluent
production
rate, Mg/yr
Liquid CaS04
0
0
4
5
5
0
0
0
0
5
6
6
0
8
9
9
0
0
0
0
0
4
5
5
0
0
0
0
0
0
0
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
3
3
2
2
2
2
2
2
4
4
4
4
2
2
Total
post-
expansi on
effluent
production
rate, Mg/yr
Liquid CaS04 Liquid
14
14
14
14
14
14
14
16
16
16
16
16
16
16
16
16
16
16
16
14
14
14
14
14
14
26
26
26
26
16
16
2
3
4
4
4
3
4
3
4
4
4
4
5
6
7
7
7
4
6
2
3
4
4
4
4
4
6
3
4
3
3
14
22
20
2?
22
21
28
16
25
25
27
26
31
30
34
32
26
23
30
14
21
21
22
22
26
26
40
19
26
16
17
Incremental
increase in
effluent
produc-
tion, Mg/yr
CaS04 Liquid
--
1
1
1
1
1
2
--
1
1
1
1
2
3
4
4
4
1
3
--
1
2
2
2
2
--
2
(1)
0
--
1
--
6
6
8
8
7
14
--
9
9
11
10
15
14
18
16
10
7
14
--
7
7
8
8
12
--
14
(7)
0
--
1
NA = Not applicable
-------
TABLE 7-9. ESTIMATED PRODUCTION RATE OF SOLID AND LIQUID EFFLUENTS
REQUIRING DISPOSAL FROM FGD SYSTEMS ASSOCIATED WITH GREENFIELD SMELTER MODELS
Control alternative
Type of FGD
system employed
Solid waste
production rate,
Mg/yr (tons/yr)
Liquid waste
production rate,
Mg/yr (tons/yr)
I-B
I-D
MgO
Limestone
6,300a ( 6,930)
406,500° (447,200)
332,600b ( 365,800)
2,402,300d C, 642, 500)
Calculated assuming the solids consist only of MgS03. In reality, some MgS04 as well as other
magnesium species will exist in the solids.
Estimated assuming that the total effluent requiring ponding is 2-percent solids by weight.
Estimated based upon a sludge (CaS03) production rate of 6 to 7 kg per kg S02 absorbed.
dEstimated assuming that the total effluent requiring ponding is 15-percent solids by weight.
-------
TABLE 7-10. ESTIMATED PRODUCTION RATE OF SOLID AND LIQUID EFFLUENTS REQUIRING
DISPOSAL FROM FGD SYSTEMS ASSOCIATED WITH EXPANSION OPTIONS
Type of FGD
Expansion option system employed
2
3
8
9
12
13
19
20
Limestone
MgO
Limestone
MgO
Limestone
MgO
Limestone
MgO
Solid waste
production rate,
Mg/yr (tons/yr)
77,000a (84,700)
1,250C (1,380)
133,200a (146,500)
2,170C (2,380)
333,800a (367,200)
5,300C (5,850)
31,900a (35,100)
520C (570)
Liquid waste
production rate,
Mg/yr
455,000b
62,600d
787,100b
108,400d
l,972,600b
265,900d
188,400b
25,900d
(tons/yr)
(500,500)
(68,900)
(865,800)
(119,200)
(2,169,900)
(292,500)
(207,300)
(28,500)
aEstimated based upon a sludge (CaS03) generation rate of 6 kg per kg S02 absorbed.
bEstimated assuming that the total effluent requiring ponding is 15 percent solids by
weight.
Calculated assuming the solids consist only of MgS03. In reality, some MgS04 as well as
other magnesium species will exist in the solids.
dEstimated assuming that the total effluent requiring ponding is 2 percent solids by
weight.
-------
Table 7-10 for the expansion scenarios. Although this liquid is
recycled to the absorbent makeup tank, it must be ponded temporarily
for the associated sludge to settle out. The lined pond required for
sludge disposal (see Section 7.4) will provide adequate containment of
any dissolved species that might adversely affect water quality.
The purge take i from the MAGOX scrubbing system will consist of a
liquid in equilibrium with several magnesium species. Thus, since
dissolved magnesium salts are known to affect water quality in terms
of increasing its "hardness," the potential for certain problems does
exist. Disposal of this liquid by ponding, with the use of impermeable
liners, should provide acceptable containment, however. Estimates of
the amount of liquid requiring disposal due to the purging of the
MAGOX FGD systems are presented in Tables 7-9 and 7-10.
Since no absorbent purge is required from the Cominco NH3 scrubbing
process, purging will not be a potential source of pollution. A
portion of the loaded absorbent stream is continuously bled off to
provide a feedstock for the acidulation reactor. This practice, in
effect, provides the required purge for the absorbent recycle circuit.
In summary, ponding is an acceptable means of disposal for all of
the above mentioned liquid effluents. There is, however, a solid
waste product associated with each of the liquid effluents discussed
above. Disposal of solid wastes is discussed in Section 7.4.
7.4 SOLID WASTE IMPACT
The significant sources of solid waste are the following:
Sludge produced by the calcium-based nonregenerative scrubbing
system.
Sludge produced by the neutralization of purges from gas
cleaning and conditioning systems.
Solids in the purge taken from the MAGOX regenerative FGD
system.
Baghouse and ESP dusts not recycled.
Methodology used in estimating solid waste impacts is presented in
Appendix L.
7-16
-------
7.4.1 Calcium Based FGD's
In terms of magnitude, the sludge (primarily CaS03) produced by
the calcium-based FGD's would represent the most significant solid
waste disposal problem. Thus sludge is produced at the rate of 6 to
7 kg/kg of S02 absorbed, and generally has poor disposal properties.
The calcium sulfite sludge tends to maintain a "swampy" consistency
for long periods of time following ponding. Consequently, settling
can be quite difficult and the space requirements for settling may be
quite large. Ponding in lined ponds is, however, an acceptable means
by which to dispose of the calcium sulfite sludge. The pond must be
lined because of the potential for groundwater pollution by leachates.
Estimates of solid waste disposal requirements for the calcium-based
FGD's are presented in Table 7-9 for the high-impurity greenfield
smelter and in Table 7-10 for the expansion scenarios.
7.4.2 Gas Cleaning Purges
The neutralization procedure involved in processing the weak acid
purge from gas cleaning and conditioning systems produces calcium
sulfate (CaS04). This material is similar in nature, however, to the
calcium sulfite sludge produced by the calcium-based (limestone)
scrubbing system. Consequently, this material could also be ponded.
The addition of the calcium sulfate sludge to the calcium sulfite
sludge produced by the scrubbing systems may, in fact, improve the
disposal properties of the calcium sulfite, which, as noted previously,
tends to maintain a "swampy" consistency for long periods of time
after ponding. Estimates of the amount of calcium sulfate generated
by the neutralization procedure are presented in Tables 7-7 and 7-8.
At this point, it should be noted that the particulate matter
contained in the liquid effluent taken from gas cleaning systems may
also be ponded in the event that the dusts are not reclaimed. Imperme-
able pond liners may be required, of course, to prevent the transport
of harmful species such as As203 into the groundwater. Some operators
may, however, wish to recover these dusts from the liquid effluent in
order to recover metal values.
The purge taken from the absorbent recirculation circuit of the
MAGOX scrubbing process also contains a solid waste that must be
7-17
-------
disposed of. These solids consist primarily of hydrated crystals,
e.g., MgS03 • 3H20, MgS03 • 6H20, and MgS04 • 7H20. MgO may also be
present in the solids. Magnesium species such as these will adversely
affect water quality if allowed to enter the groundwater.. However,
once again, a lined pond would provide an acceptable means of disposal.
7.4.3 Particulate Control on Reverberatory Smelting Furnaces
As discussed in Section 6.3, a hot ESP is included in the baseline;
cold particulate matte removal systems are included in the control
alternatives. Thus, the following control alternatives are considered:
Gas-stream temperature reduction by evaporative cooling with
a cold ESP for particulate collection
Gas-stream temperature reduction by evaporative cooling with
a cold fabric filter for particulate collection.
Copper-bearing dusts captured by hot ESP's are recycled to recover
metal values; however, at smelters that use cold control only, it is
possible that recycling all of the dusts captured by the cold control
device to the smelting furnace may prove to be impractical because of
the impurities content, as discussed above. Consequently, a potential
solid waste disposal problem becomes evident. The metallic species in
these dusts are very toxic to humans;* thus, the dusts are considered
to be hazardous wastes. If dusts of this nature were not recycled or
reprocessed, they would have to be disposed of in a manner that would
prevent entry into the atmosphere or the groundwater. Ponding or
burial, however, using impermeable liners should provide adequate
containment of the toxic species. The incremental impacts of reverbera-
tory furnace particulate matter control with cold (90 to 100° C [195°
to 212° F]) control devices (ESP or fabric filter) have been estimated,
based on a material balance, for the high-impurity greenfield smelter
and are presented in Table 7-11.
*Arsenic, antimony, lead, cadmium, zinc and their compounds are listed
on the Hazardous Waste List compiled by U.S. Environmental Protec-
tion Agency (EPA).
7-18
-------
TABLE 7-11 ESTIMATE OF EMISSION REDUCTION DUE TO PARTICULATE CONTROL OF REVERBERATORY SMELTING FURNACE
PRIMARY OFFGASES--HIGH-IMPURITY GREENFIELD SMELTER
Gas stream volumetric flow rate
Basel ine
ESPd
Fabric filter6
Before cooling,
NnrVnrin (scfm)
3,290 (116,200)
3,290 (116,200)
3,290 (116,200)
After cooling,
NmVmin (scfm)
NA
4,145 (146,400)
4,145 (146,400)
Cooling water
rate, mVmin
(gal/min)
NA
0.73 (190)
0.73 (190)
Participate
mass rate to .
control device,
Mg/yr (tons/yr)
17,080 (18,790)
17,080 (18,790)
17,080 (18,790)
Total particulate
emission rate.
to atmosphere,
Mg/yr (tons/yr)
12,470C (13,710)
560 (620)
50 (55)
Emission reduction
Mg/yr (tons/yr)
11,910 (13,090)
12,420 (13,655)
aBased upon the use of evaporative cooling to reduce the reverberatory furnace offgas temperature from about 400° C to about 100° C.
Measured by reference method 5.
cBased on impurities in feed as shown in Table 6-1, distribution of impurities as shown in Table 3-10, and net ESP efficiency of
96.7 percent of particulate matter in solid state at 400° C.
dJudged to be capable of achieving a 96.7-percent overall collection efficiency.
eJudged to be capable of achieving a 99.7-percent overall collection efficiency.
-------
7.5 ENERGY IMPACT
The incremental energy requirements for the control alternatives
are the result of increased fan requirements for moving the offgases
through the emission control systems, increased electricity require-
ments for ESP's used in gas cleaning, additional heat requirements in
the emission control system (e.g., supplemental heat for acid plants),
and regeneration o4 MgO in the MAGOX FGD. For the expansion scenarios,
additional energy is required for these activities to accommodate the
increased throughput resulting from expansion of existing smelter
capacity.
7.5.1 New Greenfield Smelters—Process Emissions
Total and incremental energy requirements for the alternatives
for controlling S02 streams from reverberatory smelting furnaces are
shown in Table 7-12. For the baseline, the process energy requirement
is 1,510 x 103 GJ/yr while the energy required for the control of
roasters and converters is 585 x 103 GJ/yr. Thus, the total energy
requirement is 2,095 x 103 GJ/yr, as indicated in Table 7-12. Incre-
mental energy requirements range from a reduction of 469 x 103 GJ/yr
for Alternative I-G to an increased requirement of 654 GJ for Alterna-
tive I-B. These incremental energy requirements result in total
energy requirements of 1,626 x 103 GJ/yr and 2,749 x 103 GJ/yr for
Alternatives I-G and I-B, respectively. Alternative I-G requires
906 x 103 GJ/yr of process-related energy and 720 x 103 GJ/yr of
control-related energy, while Alternative I-B requires 2,100 x 103 GJ/yr
of process-related energy and 649 x 103 GJ/yr of control-related
energy.
7.5.2 New Greenfield Smelters—Fugitive Emissions
Incremental energy requirements for the control of fugitive
emissions from new greenfield smelters are shown in Table 7-13.
7.5.3 Expansion Scenarios
Incremental energy requirements for each of the expansion scenar-
ios, expressed as gigajoules per increased blister copper capacity,
are shown in Table 7-14. Expansion scenarios that include a MAGOX FGD
require more energy than other scenarios, primarily because of the
high energy requirements for regenerating the MgO.
7-20
-------
TABLE 7-12. ENERGY IMPACT—PROCESS S02 CONTROL ALTERNATIVES FOR NEW GREENFIELD
SMELTER, MULTIHEARTH ROASTER-REVERBERATORY FURNACE-CONVERTER
Control alternative
I
(Baseline) I-A
I-B
I-C
I-D
I-E
I-F
I-G
Control system
Total energy requirement 2,095
(103 GJ/yr)
Incremental energy
requirement (103 GJ/yr)
Incremental energy
requirement (GJ/Mg
blister copper)
45 percent
blending
2,342
247
2.26
MgO
FGD
2,749
654
5.99
NH3
FGD
2,294
199
1.82
Limestone
FGD
2,118
23
0.21
100
percent
blending
2,657
562
5.15
Oxygen
enrich-
ment
2,130
35
0.32
Oxy-fuel
burners
1,626
(469)
(4.29)
Based on blister copper production of 312 Mg/day, 350 days operation per year.
Includes process and control system energy requirement.
Note: Numbers in parentheses represent decreases in energy requirements.
-------
TABLE 7-13. INCREMENTAL ENERGY IMPACT-FUGITIVE EMISSION
CONTROL ALTERNATIVES FOR NEW GREENFIELD SMELTERS"
MHR-RV-CV FF-CV
configuration configuration
Blister copper (Mg/yr) 312 312
Multihearth Roasters
Number 6.0 0
Energy requirements.
Base Case (GJ/yr)° 2.4 NA
Controlled (GJ/yr)c 4.5 NA
Incremental (GJ/yr) 2.1 NA
Incremental (103 J/Mg 19.5 NA
blister)
Smelting furnace
Number 1.0 1.0
Energy requirements^
Base Case (GJ/yr)° 3.1 2.1
Controlled (GJ/yr)e 5.9 *.l
Incremental (GJ/yr) 2.8 2.0
Incremental (103 J/Mg 25.4 17.6
blister)
Converter
Number 4 4
Air curtain
Energy requirements,:
Base Case (GJ/yr)1' 9.5 9.5
Controlled (GJ/yr)g 18.0 18.0
Incremental (GJ/yr) 8.5 8.5
Incremental (103 J/Mg 78.0 69.0
blister)
Building evacuation
Energy requirements.
Base Case (GJ/yr)T, 9.5 9.5
Controlled (GJ/yr)n 67.'5 67.5
Incremental (GJ/yr) 58.4 58.4
Incremental (103 J/Mg 530.6 530.6
blister)
MHR = Multihearth roaster.
RV = Reverberatory smelting furnace.
CV = Converter.
FF = Flash furnace.
alncludes only energy requirements for capture and collection.
bCapture by local hooding and larry car interlock. No collection.
cCapture by local hooding and larry car interlock. Collection by
baghouse.
dCapture by local hooding, tap and skim ports, ladle and launder covers.
No col lection.
eCapture by local hooding, tap and skim ports, ladle and launder covers.
No col lection.
^Capture by air curtain and secondary hood. No collection.
9Capture by air curtain and secondary hood. Collection by baghouse.
hCapture by building evacuation. Collection by baghouse.
7-22
-------
TABLE 7-14. ENERGY IMPACTS—EXPANSION SCENARIOS FOR EXISTING
PRIMARY COPPER SMELTERS
ro
CO
Expansion
scenario
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Expanded
capacity
blister copper
(Mg/yr)
20,910
20,910
20,910
20,910
52,270
104,540
21,380
21,380
21,380
21,380
53,450
53,450
53,450
53,450
42,760
53,450
106,910
18,710
18,710
18,710
18,7-10
56,310
54,790
68,510
137,010
20,240
Incremental energy
requirements over baseline (103
Electricity
142
90
91
123
471
621
140
98
100
158
403
223
275
399
263
536
705
98
55
47
67
579
414
479
640
100
Natural
Gas
23
0
0
0
0
0
5
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Fuel oil
54
54
169
54
(2,326)
(2,326)
(25)
(25)
(183)
(25)
(157)
(157)
347
(157)
606
(2,439)
(2,439)
(241)
(241)
(146)
(241)
(2,144)
580
0
0
(25)
GJ/yr)
Total
219
144
260
177
(1,855)
(1,705)
120
73
283
133
246
66
622
242
869
(1,957)
(1,734)
(143)
(186)
(99)
(1,734)
(1,565)
994
479
640
75
Incremental
energy
requirement
(GJ/Mg
blister
copper)
10.5
6.9
12.5
8.5
(35. b)
(16.3)
5.6
3.4
13.2
6.2
4.6
1.2
11.6
4.5
20.3
(36.6)
(16.2)
(7.7)
(10.0)
(5.3)
(9.3)
(27.9)
18.1
7.0
4.7
3.7
Note: Numbers in parentheses represent decreases in energy requirements.
-------
8. COSTS
8.1 INTRODUCTION
This chapter presents capital, operating, and annualized costs
for controlling (1) weak S02 streams from greenfield reverberatory
smelting furnaces processing high-impurity materials; (2) fugitive
emissions from new, modified, or reconstructed roasters; smelting
furnaces; and converters; (3) particulate matter emissions from rever-
beratory smelting furnaces processing high-impurity materials if the
existing reverberatory exemption is retained; and (4) S02 emissions
from expanded reverberatory smelting furnaces at existing smelters.
These cost data are used to determine cost-effectiveness of the control
alternatives identified in Chapter 6. Based on the cost-effectiveness
determinations, selected control alternatives are carried forward for
determining economic impact in Chapter 9. The control options are
summarized in Table 8-1.
Process and control installed capital costs, operating costs, and
annualized costs are estimated and effectiveness of the control
alternatives determined as part of the cost analysis.* For new
greenfield smelters, capital, operating, and annualized costs are
estimated for the entire smelter for the Baseline Case and each of the
control alternatives. Incremental costs for each of the control
alternatives are then determined by subtracting the baseline costs
from the control alternative costs. For fugitive particulate matter
control, particulate matter control for exempted reverberatory furnaces,
and expansion scenarios, a different approach is followed. For these
cases, incremental process and control costs are estimated directly.
^Annualized costs, the sum of operating costs and capital recovery
costs, are computed with the methodology described in Section 8.2.2.
8-1
-------
TABLE 8-1. CONTROL ALTERNATIVES
Alternative
Description'
Baseline Case
I-A
I-B
I-C
I-D
I-E
I-F
I-G
Processing MHR and CV gases in a DC/DA sulfuric acid
plant and venting the reverberatory furnace offgas
stream to the atmosphere.
Blending 45 percent of the RV offgas stream with the
MHK and CV offgas streams, processing this blended
stream in DC/DA sulfuric acid plant, and treating
the remaining RV offgas stream with a coldside ESP.
Processing the weak RV offgas stream with a MgO
regenerative FGD system, blending the strong S02
stream from the MgO FGD system with the MHR and CV
offgas streams, and processing the total blended
stream in a DC/DA sulfuric acid plant.
Identical to Alternative I-B except a NH3 FGD system
is used in place of a MgO FGD system.
Controlling the RV offgas stream with a limestone
FGD system and processing the MHR and CV offgases in
a DC/DA sulfuric acid plant.
Blending 100 percent of the RV offgas stream with
the MHR and CV offgas streams and processing the
blended stream in a DC/DA sulfuric: acid plant.
Identical to Alternative I-E except that oxygen
enrichment is used to enhance the S02 concentration
of the RV offgas stream.
Identical to Alternative I-E except that oxy-fuel
burners are used to enhance the S02 concentration of
the RV offgas stream.
1 CV = Converter.
DC/DA = Dual contact/dual absorption sulfuric acid plant.
ESP = Electrostatic precipitator.
MHR = Multihearth roaster.
RV = Reverberatory furnace.
8-2
-------
Section 8.2 describes the cost analyses of the S02 control
alternatives selected in Section C.2 for greenfield reverberatory
smelting furnaces, and Section 8.3 discusses alternatives tor control
of fugitive emissions from roasters, smelting furnaces, and converters.
Section 8.4 presents the coat analysis of the p^rticulate matter
control alternatives for reverberatory smelting furnaces, and Sec-
tion 8.5 contains br.se case costs for new greenfield smelters,
Section 8.6 discusses cost analyses of expansion scenarios for existing
smelters, and Section 8.7 presents cost-effectiveness summaries.
The cost estimates given in this chapter' are based on cost models
developed with information available in the literature. The details
of each cost model and the basic sources of cost factors and similar
data are discussed in detail in Section 8.2. Capital and operating
costs developed herein are study tirade estimates.
8.2 CONTROL OF WEAK S02 STREAMS FROM NEW REVERBERATORY FURNACES
This section presents the estimated costs of producing copper and
controlling S02 in the gaseous process stream from a new reverberatory
smelting furnace processing high-impurity materials. The control
alternatives considered are presented in Table 8-1. The selection of
these alternatives is discussed in Section 6.2.
The Baseline Case shown in Table £-1 was used to determine the
incremental cost of a revised new source performance standard (NSPS).
This Baseline Case includes control of S02 emissions from the multi-
hearth roaster and converter, which are covered under the existing
NSPS, and control of participate matter from the reverberatory furnace,
as required by existing State Implementation Plans (SIP's). The
percent S02 recovery for each alternative was calculated based on the
application of a double contact/double absorption (DC/DA) sulfuric
acid plant to the multihearth roasters and converters, with S02 control
on the reverberatory smelting furnace.
Table 8~2 shows the flow rates, S0? concentrations, and supple-
mentary heat requirements used to calculate the costs of the control
alternatives for a l,364~Mg/day (1,500-tonb/day) model plant. The acid
plant costs were calculated using the maximum gas flow rate and a design
S02 concentration based on the acid plant flow profiles given in Table 6-3.
8-3
-------
TABLE 8-2. INPUT DATA TO COST ESTIMATION, NEW HIGH-IMPURITY SMELTER
oo
Acid
Control
alternative
Baseline
Case
I-A
I-B
I-C
I-D
I-E
I-F
I-G
plant data3
Maximum
Nm3/min (scfm)
3,380
4,870
3,890
3,580
3,380
6,690
5,710
4,660
(119
(172
(137
(126
(119
(236
(201
(164
,400)
,000)
,300)
,500)
,400)
,300)
,500)
,600)
Design
% S02
4.
3.
4.
4.
4.
3.
3.
4.
5
5
5
5
5
5
5
5
New FGD requirements
Type
NA
NA
MgO
NH3
Limestone
NA
NA
NA
Input
NmVmin (scfm)
NA
NA
3,315
3,315
3,315
NA
NA
NA
NA
NA
(117,000)
(117,000)
(117,000)
NA
NA
NA
% S02
NA
NA
1.7
1.7
1.7
NA
NA
NA
02 usage
rate,
tons/day
NA
NH
NA
NA
NA
NA
116
219
Supplementary
heat required,
GJ/24 hours
NA
23
NA
NA
NA
360
50
NA
NA - Not applicable.
scfm = Standard cubic feet per minute, measured at 0° C. 1 atm.
All sulfuric acid plants are double contact/double absorption (see Table 6-3 for the acid plant flow
profile).
-------
8.2.1 Capital Costs
The capital costs developed in this section for each of the
control alternatives shown in Table 8-1 are based on published data to
which engineering judgment is applied. The design criterion for which
the published costs had been developed was evaluated to ensure its
applicability to the process or control alternative under analysis and
its consistency with the technology presented in Chapter 4. All
capital costs are escalated to June 1981 dollars using the Chemical
Engineering Plant Cost Index (June 1981 = 298.2).
8.2.1.1 Baseline Case Control Costs Under the Existing NSPS.
For the Baseline Case of Table 8-1, a DC/DA acid plant is required
under the existing NSPS for control of the converter and multihearth
roaster process offgases. As indicated in Figure 6-1, the maximum
flow rate to the DC/DA acid plant for the Baseline Case is 3,380 Nm3/
min (119,400 scfm, measured at 20° C and 1 atm), and the design S02
concentration is 4.5 percent. A hotside ESP controls particulate
matter in furnace offgases.
Figure 8-1 shows a comparison of several published DC/DA acid
plant cost estimates that have been updated to June 1981 dollars. The
solid lines of Figure 8-1, obtained from Reference 1, were used to
determine costs for all DC/DA sulfuric acid plants. Flow rates are at
1 atm and 0° C as used in the source material. Of the costs available
for this study, those from Reference 1 are the only ones calculated
over a range of flow rates and concentrations and are thus the most
comprehensive available. These costs include gas cooling and con-
ditioning, acid making, gas preheating equipment, and all ancillary
facilities (electrical and utility supplies, control room, etc.). A
detailed equipment list is given in Appendix D-3 of Reference 1.
Reference 2 also reports costs for DC/DA acid plants. This reference
shows the same trends in capital costs versus gas flow rate. Based on
the capital costs reported in Reference 1, the following mathematical
expression was developed relating the capital cost of an acid plant to
S02 concentration and gas flow rate:
In Cost =I+a+blnQ,
8-5
-------
100
90
80
70
60
= 50
•o
t—
03
5 40
c
3
—3
1-
| 30
o
u
I 20
CO
U
4.5% SO,
40 60 80 100 120 140
Flow Rate, 103 scfm
Note: Solid lines represent costs from Reference 1 and include gas cleaning and conditioning,
absorption and acid production, auxiliary preheating, storage facilities, materials
handling, and control equipment.
Legend
Costs are calculated for 4.5 percent SC<2 from data given by Reference 2.
A Cost is updated from Reference 3, p. 32.
Figure 8-1. Capital cost of a DC/DA sulfuric acid plant.
8-6
-------
where
Cost is the total capital cost in millions of June 1981 dollars
I is a constant accounting for inflation
a and b are functions of the S02 concentration
Q is the gas flow rate in thousands of scfm (1 atm and 70° F).
Using a regression analysis of In Q and In Cost to calculate values of
"a" and "b" and to convert from 0° C to 21° C and from scfm to NmVmin
gives the following expression:
In Cost = 0.1325 + 0.378 (% S02)°-177 + 0.783 (% S02)"0'104 [In Q - 3.418]
where % S02 is expressed as percent; i.e., 4% S02 is input as 4.0.
Based on this expression, the capital cost for a DC/DA sulfuric acid
plant for the Baseline Case control of the roaster and converter
streams is $43.8 million. The costs of a hot ESP for particulate
matter control of the reverberatory furnace weak stream results in a
total baseline control cost of $46.3 million (see Section 8.4.1 for
estimation of ESP capital costs).
8.2.1.2 Capital Cost of Alternative I-A. Alternative I-A consists
of blending 45 percent of the reverberatory furnace offgas stream with
the roaster and converter streams and processing the blended stream in
a DC/DA acid plant.
From Table 8-2, the maximum process flow rate for Alternative I-A
is 4,870 NmVmin (172,000 scfm), and the design S02 concentration is
3.5 percent. Using this information in the equation developed in
Section 8.2.1.1 yields an acid plant capital cost of $59.8 million.
The cost of a hot ESP for particulate control of the remaining 55 percent
of the reverberatory furnace offgas stream results in a total control
cost of $61.3 million.
8.2.1.3 Capital Cost of Alternative I-B. Alternative I-B consists
of an MgO regenerative flue gas desulfurization (FGD) system to process
the weak reverberatory gas stream. The strong S02 stream from the MgO
system is then blended with the roaster and converter streams and the
total blended stream processed in a DC/DA sulfuric acid plant.
8-7
-------
Figure 8-2 shows the reported capital costs of the MgO FGD system.
The solid lines in this figure are the reference costs used to calculate
costs herein and were obtained from Reference 4. There are differences
in the MgO systems for whir*- costs are developed in References 1, 5,
and 6. Specifically, Reference 1 reports costs for two flow rates and
S02 concentrations, and costs for spare equipment are included. The
costs given by Reference 5 include a dedicated wastewater treatment
plant, retrofit items specific to a certain site (stacks, piping,
etc.), and primary and reclaimed storage and associated materials
handling. This cost is somewhat higher than that developed for
1 percent S02 in Reference 4. The costs reported in Reference 6 are
based on data that have been updated over a considerable time (approx-
imately 12 to 15 years). The scope of the costs, however, is somewhat
more limited than that of Reference 4 (see Figure 8-2, Note 1). For
example, a baghouse is not included on the MgO storage silo and rotary
drier offgas, nor is a lined cooling pond or venturi gas scrubber (the
scrubber of Reference 6 is a somewhat less expensive spray tower).
The costs provided by the computer program documented in Refer-
ence 4 were used to develop the MgO system costs shown as solid lines
in Figure 8-2. The following mathematical expression was developed to
represent the solid lines shown in Figure 8-2:
In Cost = In [1.564 (% S02)°-°926] + [0.546 + 0.0252 (% S02)]
[In Q - 3.418].
where
Cost is the total capital cost in millions of June 1981 dollars,
% S02 is expressed as a percent, not a decimal (i.e., 4% S02
would be input as 4.0)
Q is in NmVmin.
Using the flow rate and S02 concentration shown in Table 8-2 for
Alternative I-B, namely, 3,315 NmVmin (117,000 scfm) and 1.7 percent
S02, the capital cost of the MgO system was calculated to be $26.0
million.
8-8
-------
100
80
60
40
« 30
to
•5
•a
r- 20
00
0>
g 15
3
O
10
8
« 6
o
o
"(5
i 4
03
o
10
3% SO2
2% SO2
1.5% SO2
1% SO2
.5% SO
j L
20
40 60 80 100 150 200 400
Flow Rate, 103 scfm
Notes: 1. Costs include materials handling and storage, gas cleaning and conditioning,
SO2 absorption, regeneration of scrubbing liquor, new stack, and stack gas
reheating. Acid plant cost not included.
2. Solid lines represent costs calculated using a computer program documented
in Reference 4.
Legend
A Costs are updated from Reference 1 for 1 percent SO2.
Costs are updated from Reference 6 for 1 percent S02.
• Costs from Reference 5 for 1 percent SO2 include a wastewater treatment plant,
reclaimed storage, and retrofit costs.
A Cost is updated from Reference 1 for 1.4 percent SO2.
Figure 8-2. Capital cost of an MgO FGD system.
8-9
-------
Using the equation developed in Section 8.2.1.1 and the acid
plant inlet conditions given in Table 8-2, 3,890 NmVmin (137,300 scfm)
and 4.5 percent S02 for the combined MgO outlet, roaster, and converter
streams, yields a DC/DA acid plant cost of $48.0 million. Thus, the
total capital cost for Alternative I-B is $74.0 million.
8.2.1.4 Capital Cost of Alternative I-C. Alternative I-C consists
of treating 100 percent of the reverberatory offgases in an NH3 regenera-
tive FGD system. The strong S02 stream from the NH3 FGD then goes to
a DC/DA acid plant after being blended with the roaster and converter
streams.
Costs for the Cominco NH3 system, discussed in Section 4.3.3, are
not available in the published literature. A search of the literature
and discussions with Cominco indicate that only two such systems
exist: the original facility built in 1936 at Trail, British Columbia,
and a plant built in 1952 under license from Cominco in Pasadena,
Texas, for Olin-Mathieson. The Olin-Mathieson plant was sold to Mobil
Oil in 1980. No information is as yet available on the costs for an
NH3 scrubbing system.
A closely related system for which costs are available is the
ammonia bisulfate (ABS) acidulation process (discussed in Section 4.3.3).
A comparison of Figures 4-5 (Cominco system) and Figure 4-6 (ABS
system) shows that the two processes are essentially the same, except
for the acidulation step (the Cominco process uses H2S04, and the ABS
process uses ammonia bisulfate, NH4HS04) and the final step (the
Cominco process uses the (NH4)2S04 as fertilizer, and the ABS process
reacts the (NH4)2S04 in a furnace to generate NH3 and NH4HS04). In
the Cominco process, NH3 is not recirculated as it. is in the ABS
process. Because NH3 is expensive, the Cominco system will generally
be economically feasible only when the ammonium sulfate, (NH4)2S04, is
sold as fertilizer. Whether this can be done at a given smelter is a
highly site-specific question involving the local fertilizer market,
transportation costs to other markets, and the necessary product
quality. For the Cominco process, the ammonium sulfate that leaves
the S02 stripper (see Figure 4-5) must be dried, washed to remove
8-lp
-------
discoloration, granulated, and packaged. This involves additional
capital cost beyond that for the equipment shown in Figure 4-5. As an
approximation, it is assumed herein that the capital cost of this
additional equipment is roughly equal to the capital cost of the
additional furnace required for the ABS process. Thus, the capital
cost of the ABS process is assumed to be comparable to that for the
Cominco process.
In the same manner as for the MgO FGD systems, Figure 8-3 shows
reported capital costs for ABS and other NH3 scrubbing processes. In
addition, Figure 8-3 also shows a cost reported by the Tennessee
Valley Authority7 for a process very similar to the Cominco process.
The cost for this system, which is used for scrubbing coal-fired power
plant stack gases, has been escalated (for inflation) and adjusted
(for capacity differences) using the "0.6 rule" to determine the cost
for an NH3 scrubbing system processing 3,050 NmVmin (108,000 scfm) at
0.358 percent S02. Moreover, Figure 8-3 also shows the cost of a
second NH3 FGD system,8 which includes H2S04 acidulation of the scrubber
effluent and fertilizer production. The cost of this system was
estimated with the same escalation and adjustment procedure to determine
the costs for a 3,050-Nm3/min (108,000-scfm) system at 0.27 percent
S02. Because these costs are consistent with those of Mathews6 for
the ABS process (note that they are somewhat lower due to the lower
S02 concentration, as expected), the costs developed by Mathews (shown
as solid lines) are used to determine the NH3 scrubbing costs for this
report.
The equation for the solid lines in Figure 8-3 is:
In Cost = -1.949 + 0.346 In (% S02) + 0.57 In Q ,
where
Cost is in millions of June 1981 dollars
%S02 is an absolute value, not expressed as a decimal (i.e., 4%
S02 would be input as 4.0)
Q is in NmVmin.
8-11
-------
100
80
60
40
30
,- 20
oo
O)
« 15
c
3
2 10
I 8
1
*J 6
6O
O
o
! 4
a.
ca
O
\ 1—i
T 1 I I
10
4% SO2
2% SO,
1%SO,
i i i
20
40 60 80 100 150 200
Flow Rate, 103 scfm
400
Notes: 1. Solid lines represent costs, including materials handling and storage, gas cleaning
and conditioning, SO2 absorption, and acidulation. An acid plant is not included.
2. Costs are for ammonium bisulfate acidulation of the scrubbing liquor from Reference 6.
Legend
A Costs have been scaled for a process similar to Cominco [(IMH4)2SO4 fertilizer production,
H2SO4 acidulation] for 0.27 percent SO2 from Reference 8.
• Costs have been scaled for a process similar to Cominco [(NH4)2SO4 fertilizer production,
H2SO4 acidulation] for 0.36 percent SO2 from Reference 7.
Figure 8-3. Capital cost of an ammonia FGD system.
8-12
-------
Based on this equation, the cost in Alternative I-C of the NH3 scrub-
bing system to process 3,315 NmVmin (117,000 scfm) of reverberatory
furnace gas at 1.7 percent S02 (see Table 8-2) is $17.4 million.
Based on the equation developed in Section 8.2.1.1, the cost of the
DC/DA acid plant to handle ti,e combined gases from the NH3 scrubber
outlet, roasters, and converters at a flow rate of 3,580 NnrVmin
(126,500 scfm) at a design S02 concentration of 4.5 percent is
$45.4 million. Thus, the total capital cost for Alternative I-C is
$62.8 million.
8.2.1.5 Capital Cost of Alternative I-D. Alternative I-D consists
of controlling the reverberatory offgases with a limestone FGD system.
The roaster and converter offgases are processed in a DC/DA acid
plant.
Figure 8-4 shows the capital costs of a limestone FGD system from
various published sources. These cost vary primarily due to the
difference in scope among the reported values. The reference costs,
shown as solid lines in Figure 8-4, are used in this report and represent
costs developed for EPA with a computer program based on post-1978
original detailed cost estimates.4 These costs, which have been
judged to be more comprehensive than other reported costs, include
material handling and storage, gas cleaning and conditioning using a
venturi scrubber, a lined cooling pond, disposal pond, a new stack,
and a stack gas reheater. Figure 8-4 shows an additional cost curve
with a more limited scope, although costs are reported over a wide
variety of conditions.6 For example, cooling and disposal ponds and a
stack are not included. Additional results of a later, though less
detailed, study are also shown.2 Costs for 1 and 1.4 percent S02--which
lie fairly close to the reference costs (shown as solid lines in
Figure 8~4)--are given in Reference 1.
The solid lines shown in Figure 8-4 were developed based on the
following expression, which was derived for the relationship between
costs and both S02 concentration and gas flow rate:
In Cost = In [1.191 - 0.104 ln(% S02)] + 0.628(% S02)°-117 [In Q - 3.418]
8-13
-------
100
80
60
40
S2 30
03
o
•a
r~
co
9)
•3
ta
O
O
o
a
ra
O
20
15
10
8
10
—T~l—'—' ' '
' 4% SO9
3% SO2
2% SO2
1.5% SO2
1.0% SO2
0.5% SO
2 _
J L
j i.
i i i
20
40 60 80 100 200
Flow Rate, 103 scfm
300 400
Notes: 1. Costs include materials handling and storage, feed preparation, gas cleaning and
conditioning, scrubbing, lined cooling pond, disposal pond, new stack, and reheating
of stack gases.
2. Solid lines represent costs calculated using a computer program documented in
Reference 4.
Legend
Costs are updated from Reference 6 for 1 percent SO2-
A Cost is updated from Reference 1 for 1 percent SOa-
A Cost is updated from Reference 1 for 1.4 percent SO2-
• Costs from Reference 5 for 0.6 percent SO2 include, in addition to Mote 1, a dedicated wastewater
treatment plant and special retrofit items.
Costs are updated from Reference 2.
Figure 8-4. Capital cost of a limestone FGD system.
8-14
-------
where
Cost is in millions of June 1981 dollars
Q is NmVmin
% S02 is expressed es a percent, not a decimal (i.e., 4% S02
would be input as 4.0).
Using this express-on with the values shown in Table 8-2, 3,315 NmVmin
(117,000 scfm) and 1.7 percent S02 yields a capital cost of $26.1 million
for controlling the reverberatory furnace gases with a limestone FGD.
The capital cost for control of the roaster and converter streams
is the same as that for the Baseline Case since, for both alternatives,
only the roaster and converter streams are treated in the DC/DA acid
plant (see Table 8-2). From Section 8.2.1.1, the cost of the acid
plant is $43.6 million. Thus, the total capital cost for control
under Alternative I-D is $69.7 million.
8.2.1.6 Capital Costs of Alternative I-E. Alternative I-E calls
for blending 100 percent of the reverberatory gas stream with the
multihearth roaster and converter streams and processing the combined
stream in a DC/DA sulfuric acid plant. From Table 8-2, the maximum
flow rate is 6,690 NmVmin (236,300 scfm), and the design S02 concen-
tration is 3.5 percent, which is the lower autothermal operating limit
for new DC/DA sulfuric acid plants. Table 6-3 shows that for 13.2 hours
of the day, the S02 concentration of the combined stream is less than
3.5 percent. Supplementary heat is required. The capital cost of the
equipment necessary for preheating the inlet gases is included in the
scope of the capital costs given by the mathematical expression developed
in Section 8.2.1.1. Thus, no additional capital expense is incurred
due to the low S02 concentration. Using the expression developed
above, the capital cost of the DC/DA plant to handle 6,690 NmVmin
(236,300 scfm) at 3.5 percent S02 is $74.4 million.
8.2.1.7 Capital Cost of Alternative I-F. Alternative I-F consists
of oxygen enrichment of the reverberatory furnace combustion air with
100 percent of the resulting reverberatory gas stream going to a DC/DA
acid plant.
8-15
-------
The use of oxygen enrichment is outlined in Sections 3.4.3.4 and
4.4.6. A central question is whether the smelter should have oxygen
shipped in from offsite, provide it onsite from facilities owned by a
company specializing in the field, or provide it by the owner's facili-
ties onsite. The first option might be chosen by a smelter in a
metropolitan area. However, the demands of a large smelter would
probably swamp any supply of oxygen distilled in the vicinity of the
smelter. The second and third choices are thus more likely, and the
realities of the marketplace are such that the two are comparable on
purely economic grounds; i.e., the annualized costs of oxygen in
either case are roughly the same. Because air separation plants
require an expertise and a replacement-parts inventory unlike those of
the smelter, smelter owners would probably be inclined to contract for
onsite oxygen. As an example, the Utah Copper Division of Kennecott
has a contracted supply of oxygen..9 A contract supply of oxygen is
assumed herein, and there is thus no additional capital cost for an
oxygen plant.
The total installed capital cost of oxygen enrichment hardware to
fit a reverberatory furnace using oxygen "undershooting" (see Sections
3.4.3.4 and 4.4.6) is estimated at $510,000. This figure is based on
a total equipment cost of $27,300, which includes four water-cooled
oxygen lances, a control panel, oxygen analyzer, regulator, and a
control valve. The cost of a single oxygen lance is assumed to be
roughly equal to the cost of an oxyfuel burner, $1,800.9 The costs of
the control panel, oxygen analyzer, regulator, and control valve are
approximately $12,000, $4,500, $2,400, and $1,200, respectively.10
The total installed cost was derived by multiplying the total equipment
cost of $27,300 by a factor of 3.0, which accounts for installation
labor, site preparation, and all indirect costs.11 These costs were
then added to the purchased equipment cost to obtain the total installed
capital cost of $110,000. The cost of piping and oxygen instrumenta-
tion, valving, and control equipment from the oxygen source to the
furnace is estimated at $400,000.
The capital cost of the DC/DA acid plant is based on a flow rate
of 5,710 NmVmin (201,500 scfm) and a design S02 concentration of
8-16
-------
3.5 percent. Based on the expression developed in Section 8.2.1.1,
the acid plant cost is $66.7 million. Thus, the total capital cost
for control equipment is $67.2 million.
8.2.1.8 Capital Cost of Alternative I-G. Alternative I-G consists
of adding oxyfuel burners to the reverberatory furnace and blending
100 percent of the resulting offgas stream with the roaster and
converter streams. This blended stream is then processed in a DC/DA
acid plant. The capital costs for this option include the cost of the
oxyfuel burners and the acid plant. The total installed capital cost
of the oxyfuel burners and associated equipment (piping, controls,
instrumentation, etc.) is approximately $1 million, based on actual
costs reported by Inco for an oxyfuel installation at its Copper Cliff
smelter.12 In addition to the cost of the burners and their installa-
tion, there are significant costs associated with sealing the burner
end of the furnace, installing thermocouples in the sidewalls, and
installing piping from the oxygen source to the furnace. The cost of
the DC/DA acid plant, based on a flow rate of 4,660 NnrVmin (164,600
scfm) and a design S02 concentration of 4.5 percent (from Table 8-3)
is $54.2 million. Thus, the total capital cost for Alternative I-G is
$55.2 million.
8.2.2 Annualized Costs
Annualized costs are the sum of total operating and capital
recovery costs. Annual ized costs for the control systems discussed in
Section 8.2.1 were developed using the methodology shown below. The
usage rate of raw materials, utilities, and direct operation labor
hours (person-hours per year) have been developed for various process
and control equipment. The annual ized costs were then calculated as
follows:
Raw materials and utility costs were calculated by multiply-
ing the usage rate (Ib/yr or kWh/yr, e.g.) by the
unit cost ($/lb or $/kWh, e.g.).
Direct operating labor costs were calculated by multiplying
the labor requirement (person-hours/yr) by the labor rate
($/person-hour).
8-17
-------
TABLE 8-3.
Item
Operating labor
Electricity
Process water
Natural gas
Cooling water
Limestone
MgO
NH3
Solids disposal
(trucking)
Bunker C fuel oil
Silica flux
Refractories
(for greenfield,
MHR-RV-CV)a
Additional operating
supplies
LABOR AND UTILITY UNIT COSTS
Unit cost
$10. 61/person-hour
$0.042/kWh
$0.465/1,000 gal
$2.77/1,000 ft3
$0.23/1,000 gal
$7.75/ton
$200/ton
$175/ton
$2.60/yd3
$.63/gal
$4.00/ton
$0.30/lb
$0.45/ton concentrate
Reference
5
5
5
5
5
5
5
5
5
13
1
1
1
CV = Converter.
MHR = Multihearth roaster.
RV = Reverberatory furnace.
8-18
-------
Supervision costs were calculated as 20 percent of direct
labor costs.
Maintenance costs (labor and materials combined) were calcu-
lated as 4 percent of total capital cost.
Maintenance supervision costs were calculated as 15 percent
of maintenance costs (i.e., 0.6 percent of total capital
cost).
Overheau costs were calculated as 50 percent of the sum of
operation labor cost (direct operating labor plus super-
vision) and maintenance cost (labor, materials, and super-
vision).
Taxes, insurance, and administration costs were calculated
as 4.0 percent of total capital cost.
Capital recovery costs were calculated as 16.275 percent of
the total capital costs (corresponding to a capital recovery
rate of 10 percent for 10 years).11
Raw material, labor, and utility unit costs used in this study are
shown in Table 8-3. All labor costs are based on 8,760 hr/yr (365
days); raw material and utility costs are based on 8,400 hr/yr (350
days). Appendix M contains a detailed listing of the capital, total
operating, and annualized costs of S02 control for the Baseline Case
and each control alternative of Table 8-1.
8.2.2.1 Annualized Costs for the Baseline Case. Annualized
control costs for the Baseline Case consist of the costs associated
with the DC/DA sulfuric acid plant controlling the roaster and converter
streams and with the hot ESP. The DC/DA acid plant and the hot ESP
are designed to handle 3,380 NmVmin (119,400 scfm) at 4.5 percent
S02. Direct operating labor for a DC/DA acid plant within the range
of sizes shown in Table 8-2 is 3 persons per shift.1 Utility and raw
material requirements depend on the plant size, however. These
requirements consist of cooling and process water, electricity, and
limestone (for fixation of gas cleaning water). Based on data
provided in Reference 1, mathematical relationships were developed for
computing the water and electricity usage rates. Similarly, a mathe-
matical relationship was developed from information given in Reference 6
8-19
-------
for computing the limestone usage rate. The relationships developed
are as follows:
_ .,. . fr. 1ori TA^wnA thousands of gallons
Cooling water usage = (0.120 x 103)(Q) ,——a
- . ,0, rn wriM-o-7 T i -?o f -i-i -m4\/n\ kilowatt-hours
Electrical usage = (4.11 x 104)(Q) —
Limestone usage = 0.761 (Q)
year
tons
year '
where
Q is in NmVmin
% S02 is expressed as an absolute number, not a decimal (i.e., 4%
S02 would be input as 4.0).
Development of annualized control costs for the baseline case is shown
below:
$106
Capital costs 46.3
Operating costs
Raw materials 3,385 x 0.761 x Z^| = 0.019
Cooling water: (3,385) (0.12 x 103) (°^3) = 0.093
Process water: (3,385)(4.5)[27.1-1. 78(4.5)] ^-^ = 0.135
Electricity: 6.477
Acid plant: (3,385)(4.11 x 104) (QffJ- = (5.843)
ESP: (0.634)
Labor, direct operating: 0.310
Acid plant: 3 x 3 x 365 x 8 x $1°^1 - (0.279)
ESP: 1/3 x 3 x 365 x 8 x ^P = (0.031)
Labor, supervision: 0.310 x 0.2 = 0.062
Maintenance, labor, and material: 0.04 x $46.3 = 1.852
Maintenance: supervision: 0.006 x $46.3 = 0.278
Overhead: 0.50 (0.310 + 0.062 + 1.852 + 0.278) := 1.251
Taxes, insurance, administration: 0.04 x $46.3 := 1.852
Total operating costs 12.323
Capital recovery cost: (0.16275) (46.3) = 7.535
Annualized cost: 12.310 + 7.535 = 19.858
8-20
-------
Note: Differences between these figures and those in Appendix M are
due to rounding.
After the labor, raw material, and utility usage rates were determined,
the annualized costs were calculated with the factored estimate method
discussed in Section 8.2.2. No credit was taken for the sulfuric acid
produced. The results are a total operating cost of $12.3 million per
year and a total annual ized cost of $19.9 million per year:
Baseline Case
Capital Total
recovery costs Total operating costs annualized costs
7.5 (0.16275 x 46.3) $12.3 million $19.9 million
8.2.2.2 Annualized Cost of Alternative I-A. Annualized costs
for Alternative I-A were calculated with the methodology discussed in
Section 8.2.2.1. The DC/DA acid plant for this alternative must
handle 4,870 NmVmin (172,000 scfm) at a design S02 concentration of
3.5 percent. Supplemental heat is required (see Tables 6-3 and 8-2).
The resulting additional cost, assuming that the heat of combustion
for natural gas is 106 Btu/103 ft3 and that the cost of natural gas is
$2.77/103 ft3, is $21,300. The total operating cost for Alternative I-A
is estimated at $16.8 million per year, and the annualized cost is
estimated at $26.8 million per year:
Alternative I-A
Capital Total
recovery costs Total operating costs annualized costs
10.0 (0.16275 x 61.3) $16.8 million $26.8 million
8.2.2.3 Annualized Cost of Alternative I-B. Annualized costs
for the MgO regenerative FGD and its associated acid plant consist of
labor, utilities, and raw materials. For the MgO FGD, usage rates
were derived from published data by multiplying the values given in
Table 8-4 by the appropriate quantities. Efficiencies for various FGD
systems, including the MgO, are given in Table 4-8. For example, to
calculate the annual usage rate of the MgO, the value (given in Table 8-4)
of 0.0322 Ib MgO/lb S02 absorbed was multiplied by the pounds of S02
absorbed per year. The result is as follows:
8-2}
-------
TABLE 8-4. FGD RAW MATERIAL AND UTILITY USAGE RATE
Process
Limestone
Limestone
Process water
Electricity
Usage Rate
1.9 Ib/lb S02 removed
8.90 Ib/lb S02 removed
201 kWh/yr-scfm
Reference
6
1
1
MgO
Solid generation
(for disposal)
MgO
Coke
Process water
Electricity
Fuel oil
0.0019 ydVlb S02 removed
0.0322 Ib/lb S02 removed
0.0105 Ib/lb S02 removed
7.12 Ib/lb S02 removed
107 kWh/yr-scfm
0.0397 gal/lb S02 removed
6
6
1
1
13
NH3
H2S04
Process water
Electricity
1 mol NHs/mol S02 removed
h mol H2S04/mol S02 removed
0.06 gal/lb S02 absorbed
4 kW/1,000 scfm, plus
0.39 kWh/lb S02 absorbed
6, 7
6, 7
6
6
8-22
-------
KW1'000) ft3@STP]
Ib S02 absorbed/yr = 0.0322 Ib 1b sobsorbed
) ( 0.178 lga eSTP ) ( 60 ) ( 8,400
( o.99 Ib SO, absorbed } = (2 86 x 104) (Q) (% So2) .
ID bi 2 Tea
The following are the usage rates derived:
Labor: 3 persons/shift
MgO: (9.47 x 102)(Q)(% S02) Ib/yr
Process water: (25.1)(Q)(% S02) thousands of gal/yr
Electricity: (3.63 x 103)(Q) kWh/yr
Coke: (3.09 x 102)(Q)(% S02) Ib/yr
Fuel oil: (1.65 x 103)(Q)(% S02) gal/yr,
where
Q is in NmVmin
% S02 is expressed as a percent (i.e., 4% S02 would be input as
4.0).
For this alternative, the MgO FGD must handle 3,315 NmVmin (117,000 scfm)
at 1.7 percent S02. The total operating and annual ized costs were
calculated with these values and the methodology discussed in Section
8.2.2. The total operating costs are $8.22 million, and the total
annual ized costs are $12.5 million.
The DC/DA acid plant handles the combined MgO, roaster, and
converter streams, a total flow rate of 3,890 NmVmin (137,300 scfm)
at a design S02 concentration of 4.5 percent; operating and annual ized
costs were calculated with the methodology of Section 8.2.2.1. The
total operating and annual ized costs computed for this alternative are
$21.0 million per year and $33.0 million per year, respectively:
8-23
-------
Alternative I-B
Total
Control Total operating costs annualized costs
MgO $ 8.2 million $12.5 million
DC/DA $12.7 million $20.1 million
Total $21.0 million $33.0 million*
^Includes a capital recovery cost of $12.0 million.
8.2.2.4 Annualized Cost of Alternative I-C. For the NH3 FGD,
the labor and utility rates were calculated with published references.6 7
The following were calculated as shown in Section 8.2.2.3 (see also
Table 8-4):
Labor: 3 persons/shift
NH3: (3.51)(Q)(% S02) tons/yr
Electricity: Q[1.10 x 103 + 1.04 x 104 (% S02)] kWh/yr
H2S04: (5.57)(Q)(% S04) tons/yr
Process water: (1.60)(Q)(% S04) thousands of gal/yr
Cooling water: (3.4 x 102)(Q)(% S02) thousands of gal/yr ,
where
Q is in NmVmin
% S02 is expressed as a percent, not a decimal (i.e., 4% S02
would be input as 4.0).
The total operating and annualized costs were calculated with these
expressions and the flow rate of 3,890 NmVmin (117,000 scfm) and a
S02 concentration of 1.7 percent. The results are total operating
costs of $10.4 million and total annualized costs of $13.3 million.
The DC/DA acid plant operating and annualized costs were calculated
with the methodology of Section 8.2.2.1. Thus, the total operating
and annualized costs for this alternative are $22.3 million per year
and $32.6 million per year, respectively:
9-24
-------
Alternative I-C
Total
Control Total operating costs annualized costs
NH3 $10.4 million $13.3 million
DC/DA $11.9 million $19.3 million
Total $22.3 million $32.6 million*
^Includes capital recovery cost of $10.3 million.
8.2.2.5 Annualized Cost of Alternative I-D. Alternative I-D
consists of limestone scrubbing of the reverberatory furnace stream
with the paster and converter streams going to a DC/DA acid plant.
The reverberatory furnace stream is 3,315 NmVmin (117,000 scfm) at
1.7 percent S02.
The labor, raw material, and utility usage rates for the limestone
FGD are taken from Reference 4. The labor requirement is five persons
per shift. Raw material, utility, and solid waste disposal rates are
presented in Table 8-4. Based on these rates, the following mathemati-
cal relationships were developed:
Limestone: (28.0)(Q)(% S02) tons/year
/oi cwnw
-------
The DC/DA acid plant operating and annualized costs were calculated
with the methodology outlined in Section 8.2.2.1. Total operating
costs for this alternative are $17.7 million per year, and total
annualized costs are $29.1 million per year:
Alternative I-D
Total
Control Total operating costs annualized costs
Limestone_ $ 6.38 million $10.6 million
DC/DA $11.4 million $18.5 million
Total $17.7 million $29.1 million*
^Includes capital recovery costs of $11.4 million,
8.2.2.6 Annualized Cost of Alternative I-E. Annualized costs
for Alternative I-E, which involves blending of 100 percent of the
reverberatory furnace stream with the roaster and converter streams,
were calculated in the same manner as those for the Baseline Case
described above. From Table 8-2, the DC/DA acid plant size is 6,690
NmVmin (236,300 scfm), with a design S02 concentration of 3.5 percent.
As noted in Section 8.2.1.2 and shown in Table 6-3, there is a period
of 1.3 hours during which the S02 concentration falls below the auto-
thermal operating limit of 3.5 percent. This requires supplementary
heat of 4.7 x 107 Btu/24 hours (see Tables 6-3 and 8-2), which is
supplied by natural gas burners and heat exchangers and represents an
additional operating cost. This additional operating cost, assuming
that the heat of combustion of natural gas is 106 Btu/103 ft3 and that
the cost of natural gas is $2.77/103 ft3, is $330,600.
Based on the expressions developed in Section 8.2.2.1, the cooling
water, process water, electricity, and limestone usage rates for the
DC/DA acid plant were calculated to be 8.01 x 108 gal/yr, 4.88 x 108
gal/yr, 2.74 x 108 kWh/yr, and 3.22 x 104 tons/yr, respectively.
Based on these values, the total operating costs for the acid plant
are $21.0 million per year, and the corresponding annualized costs are
$33.1 million per year:
8-26
-------
Alternative I~E
Capital Total
recovery costs Total operating costs annualized costs
$12.1 million $21.0 million $33.1 million
(0.46275 x 74.4)
8.2.2.7 Annualized Cost of Alternative I-F. Alternative I-F
consists of oxygen enrichment of the reverberatory furnace with the
resulting gas stre? i being treated in a DC/DA acid plant after blending
with the roaster and converter streams. Annualized costs for Alterna-
tive I-F include the costs associated with both the oxygen enrichment
and the DC/DA acid plant. It is assumed herein that no additional
labor or utility usage is required by the oxygen enrichment equipment.
The only costs attributable to this equipment are thus maintenance and
capital-related costs. Supplementary heat is required (see Tables 6-3
and 8-2) and, based on methodology discussed in Sections 8.2.2.2 and
8.2.2.6, the resulting annual cost is $45,600. These costs were
calculated with the methodology of Section 8.2.2.1, with the labor and
utility usage rates equal to zero.
The oxygen usage rate is 116 tons/day (see Table 8-2). The unit
cost of oxygen under contract was said to be $36/ton in 1979.14 This
cost must be escalated to June 1981 dollars. Equipment and plant cost
indexes are somewhat inaccurate when applied to oxygen, an energy-
intensive product.15 For example, although the Chemical Engineering
Plant Cost Index has increased 25 percent from 1979 to June 1981, the
increase in contract oxygen cost is probably closer to 30 percent, to
about $47/ton. Based on this value, the cost of oxygen is $1.91 million
per year.
The use of oxygen enrichment results in an 18-percent fuel oil
savings in the reverberatory furnace relative to the Baseline Case.
Thus, process costs for this alternative are less than the Baseline
Case due to the fuel savings associated with the use of oxygen. This
is reflected in the annualized fuel oil costs shown in Appendix M.
The DC/DA acid plant for this option must handle 5,710 NmVmin
(201,500 scfm) at a design S02 concentration of 3.5 percent. The
8-27
-------
total operating and annual!zed costs for the acid plant were calculated
with the methodology of Section 8.2.2.1. The total operating cost for
this alternative is $20.0 million, and the total annualized cost is
$30.9 million:
Alternative I-F
Total
Control Total operating costs annualized costs
Oxygen enrichment $ 2.0 million $ 2.1 million
DC/DA $18.0 million $28.8 million
Total $20.0 million $30.95 million*
^Includes a capital recovery cost of $10.9 million.
8.2.2.8 Annualized Cost of Alternative I-G. Alternative I-G
consists of adding oxyfuel burners to the reverberatory furnace and
blending the resulting offgases with the multihearth roaster and
converter streams and processing the blended stream in a DC/DA acid
plant. Annualized costs for this alternative include those for both
the oxyfuel system and the acid plant. This oxyfuel option results in
a 40-percent fuel oil savings relative to the Baseline Case. This
cost saving is reflected in the costs given in Appendix M and below.
The oxyfuel system is assumed to require no additional labor
costs. Thus, as for the oxygen enrichment option above, its annual
operating costs consist of maintenance and capital-related costs.
The oxygen usage rate for Alternative I-G is 219 tons/day (see
Table 8-2). At $47/ton, the annual cost for oxygen is thus $3.61 mil-
lion. The DC/DA acid plant for this option must handle 4,660 NmVmin
(164,600 scfm) at a design S02 concentration of 4.5 percent.
The total operating costs for this alternative are $18.5 million
per year, and the total annualized costs are $27.5 million per year
(based on the methodology of Section 8.2.2.1):
Alternative I-G
Total
Control Total operating costs annualized costs
Oxyfuel burners $ 3.2 million $ 3.9 million
DC/DA $15.3 million $23.6 million
Total $18.5 million $27.5 million*
^Includes capital recovery costs of $10.0 million.
8-28
-------
8.3 COSTS FOR FUGITIVE EMISSION CONTROL
Fugitive control costs are developed herein for two 1,364-Mg/day
(1,500-ton/day) model plants: one consisting of five multihearth
roasters, one reverberatory furnace, and three converters and a second
consisting of one flash furnace and three converters and an electric
slag cleaning furnace. Alternatives evaluated for the control of
fugitive particulate matter for a smelter processing high-impurity
materials and the percent reduction in emissions from the Baseline
Case are given in Table 8-5.
Costs for control of fugitive emission sources include the costs
associated with the capture of fugitive S02 and particulate matter
emissions, as well as the costs associated with the subsequent collec-
tion of the captured particulate matter. The cost of a new stack is
not included because a stack of sufficient size to handle the fugitive
emission controls is included in the Baseline Case. Capital and
annualized costs were developed for the control of fugitive emissions
from each of the sources selected in Section 6.4 for possible regula-
tion. See Appendix N for a detailed listing of capital and operating
costs for fugitive emission control.
The design ventilation rates for each type of fugitive control
system are identified in Chapter 4 based on the most effective systems.
Assumptions are made with regard to the number of operations (slag
skimming, matte tapping, etc.) that would occur simultaneously, taking
into account what may happen even under somewhat abnormal operation.
Consequently, system design ventilation rates were determined for a
"worst case" situation to ensure that adequate capacity would be
available. Specific values used for each source of fugitive emissions
are discussed below.
8.3.1 Capital Costs
8.3.1.1 Calcine Discharge. For the multihearth roaster, a larry
car interlock system such as the one employed at ASARCO-Hayden has
been judged to be effective in capturing emissions that occur during
calcine discharge. As discussed in Section 4.7.4 and shown in Figure
4-24, this system employs three ports at the point where the calcine
8-29
-------
TABLE 8-5. EVALUATED ALTERNATIVES FOR CONTROL OF FUGITIVE PARTICULATE EMISSIONS FROM A NEW
COPPER SMELTER (MULTIHEARTH ROASTER, REVERBERATORY FURNACE, CONVERTER OR FLASH
FURNACE-CONVERTER) PROCESSING HIGH-IMPURITY MATERIALS
Emissions captured and collected, %
Baseline case
Alternative II-A
Alternative II-B
u>
o
Roaster calcine discharge"
Smelting furnace matte
tapping and slag skimming
Converter, blowing,
charging, skimming,
pouring
Capture only
0
Capture only
0
Capture only
0
Larry car interlock and
ventilated enclosure
to baghouse
89.1
Tap port and skim bay
hoods, ladle hoods,
launder covers to baghouse
89.1
Building evacuation to
baghouse
94.1
Same as II-A
89.1
Same as II-A
89.1
Air curtain and fixed
enclosure to baghouse
89.1
Not applicable to flash furnace-converter configuration.
-------
hopper-larry car connection is made. One of the ports is used to
transfer the calcine from the hopper to the car; the other two
ports are used to ventilate the fugitive emissions generated. The
ventilation rate at each port is 140 Nm3 (5,000 scfm), for a total of
280 NmVmin (10,000 scfm) pcr hopper.
The total gas evacuation flow rate from the possible simultaneous
discharge of two hoopers is 565 NmVmin (20,000 scfm). The cost of a
fabric filter to control these emissions can be calculated from the
following equation, adapted from Reference 17, for an air/cloth ratio
of 2.5 ft/min and a design pressure drop across the fabric filter
system of 4 inches H20:
In Cost = -3.84 + 0.841 In [Q (T)] ,
where
Cost is in thousands of June 1981 dollars
Q is in NmVmin
T is the actual gas stream temperature, °K.
At a flow rate of 565 Nm3 (20,000 scfm) and a temperature of 50° C
(120° F) (based on information given in Reference 6), the baghouse
cost is $571,000 based on the above equation. The larry car capture
system equipment has a capital cost estimated at about $95,000. This
is based on a cost of $19,000 per hood5 for each of the five hoods
required (one hood per roaster). The capital cost of a fan for the
capture system that must handle 280 NnrVmin (10,000 scfm) at 120° F at
a pressure drop across the hoods, ducting, and stack of 4 inches H20
was calculated from information given in Reference 17 to be $21,000.
The total cost for the larry car system and the baghouse is thus
$687,000.
8.3.1.2 Matte Tapping and Slag Skimming. For the reverberatory
furnace, both matte tapping and slag skimming fugitives are to be
controlled. Hooding systems such as the one used at ASARCO-Tacoma
have been judged to be extremely effective in capturing these emissions
(see Section 4.7.5.1 and Figure 4-26). This system uses local hooding
over tapping ports and skim bays and ladle hoods for both matte ladles
8-31
-------
and slag pots. A typical furnace configuration involving four matte
tapping ports and two slag skimming bays was used as a basis for the
cost estimates. The evacuation rate for the matte tapping hoods is
280 NmVmin (10,000 scfm) per hood (Section 4.7.5.1) and the temperature
is about 80° C (175° F).5 The matte ladle hood is evacuated at 850
NmVmin (30,000 scfm) (see Figure 4-25), and the temperature is about
40° C (100° F) (temperature based on slag ladle hood gas temperature
given in Reference 6). For slag skimming, each of the two skim bay
hoods is evacuated at 140 NmVmin (5,000 scfm) at a temperature of
40° C (100° F),5 while each of the slag pot hoods is evacuated at
565 NmVmin (20,000 scfm) at a temperature of 40° C (100° F) (see
Figure 4-27). In determining the total evacuation rate for the
reverberatory furnace fugitive emission control system, it is assumed
that one matte tap and one slag skim could occur simultaneously,
although the frequency of such occurrences would not be great. Whenever
a tap or skim occurs, both the local and ladle hoods for the port or
bay involved must be evacuating. Other hoods are assumed to be closed.
The total evacuation rate for one matte tap plus one slag skim is
1,840 NmVmin (65,000 scfm) per reverberatory furnace. A design based
on this evacuation rate should ensure that there is adequate fan
capacity to capture fugitive emissions during all possible reverberatory
furnace operations.
Based on the baghouse cost equation above, a temperature of 50° C
(120° F), and a flow rate of 1,840 NmVmin (65,000 scfm), the cost of
the baghouse for the reverberatory furnace is $1.54 million. The
capture system cost is estimated at $298,000, based on information
given in References 5, 17, 18, and 19. The total cost of the system
is thus $1.84 million.
8.3.1.3 Converters. Two technologies are currently being used
for control of fugitive emissions generated by copper converters.
These are the building evacuation system and the air curtain/fixed
enclosure system (see Sections 4.7.6.1 and 4.7.6.3). For building
evacuation, the primary element upon which the evacuation rate is
based is the number of converters present within the structure. A
design evacuation rate for a building evacuation system that handles
8-32
-------
four converters is 25,500 NnrVmin (900,000 acfm), as discussed in
Section 6.3. Based on a temperature of 55° C (130° F) for the evacu-
ated air,18 this corresponds to 21,250 NmVmin (750,000 scfm). Since
all of the base case smelters that represent greenfield smelters have
four converters, 25,500 actual cubic feet per minute (AmVmin)
(900,000 acfm) has been used as the design evacuation rate for the
purpose of estimating the cost of building evacuation fugitive emis-
sion control. For t'.ie air curtain/fixed enclosure system, 2,830 NmVmin
(100,000 scfm) per converter is selected as the maximum design ventila-
tion rate.19 It was assumed that only two of the four converters are
blowing at any one time, and thus a total evacuation rate of 5,665 NnrVmin
(200,000 acfm) at 65° C (150° F) was used to calculate the fugitive
control costs for the air curtain/fixed enclosure.
The cost of the air curtain/secondary hood was taken from data
provided by ASARCO.19 For this case, the capital cost is $6.17 million
for all four converters. The balance of the air capture system cost
(primarily ducting) was estimated from Reference 19 to be $1.53 million.
When an air curtain is used, the baghouse cost is $4.13 million, based
on a flow rate of 5,665 NmVmin (200,000 scfm) at 65° C (150° F) and
the cost equation given in Section 8.3.1.1. Thus, the total system
cost for the air curtain is $11.83 million. For the capture system
using building evacuation, the baghouse cost is $12.2 million, based
on a flow rate of 21,250 NnrVmin (750,000 scfm) at 55° C (130° F) and
the cost equation given in Section 8.3.1.1. The cost of the ducting
and hoods required for building evacuation was estimated from Reference 18
to be $4.6 million, to which a fan cost of $900,000 is added based on
costs reported in Reference 17. Thus, the total system cost for
building evacuation is $17.5 million.
8.3.2 Annualized Costs
Annualized costs were developed for fugitive control equipment
based on the same methods described in Section 8.2.2. For the purposes
of estimation, all labor and utility costs were assigned to the bag-
house. That is, it was assumed that there are no operating costs for
the air capture system.
8-33
-------
Electrical usage was calculated assuming that the major electrical
consumption is by the main fans. A value of 23 kWh/yr-acfm can be
calculated assuming a 60-percent fan efficiency, a total pressure drop
from source to stack of 14 inches H20, and 8,400 hr/yr operation.
Additional electrical consumption by the screw conveyor, instrumenta-
tion, reverse air fans, and shakers was calculated as 10 percent of
the fan consumption. Thus, the total electrical usage rate is
25 kWh/yr-acfm.
Based on these inputs, the total annualized costs were calculated
as discussed in Section 8.1.1. Labor requirements were assumed to be
zero for the capture system, 1/3 person per shift for the roaster and
the converter-air curtain, and 1/4 person per shift for the reverbera-
tory furnace and the converter-building evacuation. These labor rates
correspond to those used in calculating the annual ized costs for the
ASARCO-Tacoma and Anaconda smelters in Reference 18. A value of 1/4
person per shift was used in Reference 18 for the matte and slag
tapping fugitive emission control and for building evacuation. The
annualized costs were calculated with the methodology of Section 8.2.2.1.
The results are shown below:
Capture system
(hood, duct, fan)
Collection
system (baghouse)
Source
Multi hearth
roaster
Reverberatory
furnace
Converter--BE
—AC
Total ,
capital cost
116
298
5,300
6,170
Total
annual ized
cost6
39
103
1,710
2,240
Total b
capital cost
571
1,540
12,200
4,130
Total
annual ized
cost6
234
533
4,185
1,400
BE = Building evacuation.
AC = Air curtain.
aThe cost given is for the total model plant. Specifically, the
multihearth roaster costs are for control of five roasters, the
reverberatory furnace costs are for one furnace, and the converter
.costs are for four converters.
In thousands of June 1981 dollars. Annualized costs include capital
recovery costs, 16.275 percent of the total capital cost.
8-34
-------
8.4 COST OF CONTROLLING PROCESS PARTICULATE EMISSIONS FROM REVERBERATORY
FURNACES IF THE REVERBERATORY EXEMPTION IS RETAINED
8.4.1 Capital Costs
If the S02 emission exemption for reverberatory furnaces processing
high-impurity material is retained, process particulate matter from
the reverberatory furnace must still be controlled. The cost of
controlling the emissions will be the cost of a direct contact spray
chamber immediately uownstream of the hot ESP and either a coldside
ESP or a baghouse for the model plant shown in Figure 6-1. The design
of the spray chamber is based on information given in Reference 6.
Design parameters for the baghouse are discussed in Section 6.2. The
flow rate from the reverberatory furnace was calculated at 3,315 NmVmin
(117,000 scfm) at a temperature of 110° C (230° F) from the material
balance for the model plant given in Figure 6-1. The installed capital
cost of the spray chamber is estimated, from cost information given in
reference 6 (p. 286) after adjusting to June 1981 dollars, to be $2.92
mi 11 on for this flow rate.
Based on published information on the sparkover voltage and
average corona current for reverberatory gas streams,20 the resistivity
of the reverberatory dust is judged to be relatively low. Based on
this judgment, a specific collection area (SCA) of 300 ft2/thousand
acfm is selected. The costs of hot and cold ESP designed for this SCA
can be estimated with the following equations adapted from Viner and
Ensor:16
In Cost = -7.559 + 0.627 In [Q (T)] (cold ESP),
In Cost = -9.065 + 0.634 In [Q (T)] (hot ESP),
where
Cost is in millions of June 1981 dollars
Q is the gas flow rate in NmVmin
T is the actual temperature of the gas in °K.
For a flow rate of 1,480 NmVmin (52,300 scfm) (45 percent blending
alternative), the temperature is 120° C (250° F). The ESP cost is
$2.1 million.
8-35
-------
Based on the above cost equation for the coldside ESP and the
cost equation developed in Section 8.3.1.1 for the baghouse, the
capital costs of these control systems are $3.4 million (coldside ESP)
and $2.7 million (baghouse)
8.4.2 Annualized Costs
The operating cost for the coldside ESP was calculated as indicated
in Section 8.2.2.1 based on an annual electrical usage rate calculated
from data given in Reference 17. The electrical usage rate derived
from these data is 0.474 (QT) kWh/yr, where Q is the flow rate in
NmVmin and T is the temperature of the gas stream in °K, For this
case, Q is 1,480 NmVmin (52,300 scfm) and T is 110° C (230° F), as
discussed in Section 8.2.1.2. Thus, the electrical usage rate of the
ESP is 2.76 x 105 kWh/yr.
The annualized costs for the coldside ESP, baghouse, and spray
chamber were calculated with the methodology described in Section 8.2.2.
No labor was assigned to the spray chamber. The labor rate associated
with both control systems is 1/3 person per shift based on information
in Reference 21 (see Sections 8.2.2.2 and 8.3.2). Cooling water,
limestone (for neutralization), and electricity are required for the
spray chamber. The only utility required for the ESP and baghouse is
electricity.
The limestone usage rate may be calculated from an assumed S03
content of the inlet gas of 0.3 percent,6 the stoichiometry of the
limestone neutralization reaction (1 mole Ca(03) per mole S03), and a
20-percent excess to be (0.83)(Q)^|^. For a flow rate of 3,315 NnrVmin
(117,000 scfm), the limestone usage rate is thus 2.75 x lo3 tons/year.
The cooling water used in the spray chamber may be calculated
assuming adiabatic humidification of an inlet gas stream at 315° C
(600° F)6 to saturation at 110° C (230° F). The cooling water required
for this operation can be calculated by standard engineering procedures
to be 1.0 gallon cooling water per standard cubic foot of gas. Assuming
a 10-percent loss in the cooling tower and other miscellaneous losses,
the total cooling water requirement can be expressed as (1.82 x
1Q4)(Q)thousandsj3f gallons For g f]ow rate Qf 3>315 Nn,3/min
8-36
-------
(117,000 scfm), the total cooling water requirement is 6.03 x 107
thousands of gallons/year.
The electrical usage rate for the spray chamber is calculated,
using information given in reference 6 and adding 10 percent as a
design contingency, to be (1.59 x 103)(Q)kWh/yr. For a flow rate of
3,315 NmVmin (117,000 scfm), the electrical usage rate is 5.27 x
106 kWh/yr.
For the coldside ESP, the total electrical usage rate is composed
of both the usage by the ESP--given as 0.474 (QT) in Section 8.2.2.2--
and by the fan (which can be calculated to be 16 kWh/yr-acfm, as in
Section 8.3.2, assuming a pressure drop of 10 inches H20, which is
less than that for the baghouse). For a flow rate of 3,315 NmVmin
(117,000 scfm) at 110° C (230° F) or 4,205 ArnVmin (148,300 acfm), the
total electrical usage rate for the ESP is 3.23 x io6 kWh/yr. For the
baghouse, the total electrical requirement is given as 25 kWh/yr-acfm
in Section 8.3.2. Thus, for 4,205 AnrVmin (148,300 acfm), the total
electrical usage rate is 3.71 x io6 kWh/yr. The total operating cost
for the baghouse is $0.50 million, and the total annualized cost is
$0.94 million. Based on the methodology of Section 8.2.2.1, the total
operating cost for the spray chamber is $0.59 million, and the total
annualized cost is $1.07 million. The total operating cost for the
ESP is $0.60 million, and the total annualized cost is $1.16 million.
For the Baseline Case (i.e., hot ESP), the installed capital costs are
$2.28 million, operating costs are $0.34 million, and annualized costs
are $0.71 million.
Thus, the total costs for a spray chamber and either an ESP or a
baghouse may be summarized as follows:
Total Total Total
Option capital cost operating cost annualized cost
Spray chamber/ $6.4 million $1.2 million $2.2 million
ESP
Spray chamber/ $5.6 million $1.1 million $2.0 million
baghouse
Baseline/hot $2.3 million $0.3 million $0.7 million
ESP
8-37
-------
8.5 PROCESS COSTS
8.5.1 Capital Costs
The capital costs of a model copper smelter with a multihearth
roaster-reverberatory furnace-converter (MHR-RV-CV) configuration
designed to process 1,364 Mg/day (1,500 tons/day) of concentrate were
developed with published information.1 The costs include concentrate
receiving and storage, flux handling and preparation, roasting and
dust recovery, reverberatory smelting and heat recovery, matte convert-
ing, anode furnace smelting and casting, and fugitive particulate
matter capture. The total capital cost for this smelter is $208 mil-
lion. This includes all process equipment and all pollution control
equipment required under the existing NSPS.
The capital costs of a smelter with a flash furnace-converter
(FF-CV) configuration using Outokumpu technology was also developed
with reported costs.22 23 Based on a concentrate throughput rate of
1,364 Mg/day (1,500 tons/day), the estimated total capital cost for
this smelter is $180 million. This includes all process equipment and
all pollution control equipment required under the existing NSPS.
8.5.2 Annualized Costs
Operating costs for base case greenfield smelters include the
same elements as for the Baseline Cases (see Section 8.3) except for
costs associated with operation of fugitive capture systems. The
estimated annualized process cost (which includes control costs) is
$43.4 million for the MHR-RV-CV smelter and $36.9 million for the
FF-CV smelter.
8.6 EXPANSION SCENARIOS
Expansion scenarios considered are discussed in Section 6.4.
Table 6-6 shows the smelter configuration, expansion option configura-
tion, percent expansion achieved, and applicable control alternatives.
This table is reproduced herein for convenience as Table 8-6.
There are four expansion options given in Table 8-6: oxygen
enrichment, oxyfuel burners, conversion to calcine charge, and conver-
sion to flash smelting. Control alternatives for the expanded
reverberatory furnaces selected for analysis include the following
(see also Section 6.2):
8-38
-------
TABLE 8-6. MODEL PLANT EXPANSION SCENARIOS
Scenario
Model
Percent
capacity
Unit expanded increase
Expansion option
Control option, expanded
smelting furnace
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
MHR-RV-CV
MHR-RV-CV
MHR-RV-CV
MHR-RV-CV
MHR-RV-CV
MHR-RV-CV
RV-CV
RV-CV
RV-CV
RV-CV
RV-CV
RV-CV
RV-CV
RV-CV
RV-CV
RV-CV
RV-CV
FBR- RV-CV
FBR-RV-CV
FBR-RV-CV
FBR-RV-CV
FBR-RV-CV
EF-CV
EF-CV
EF-CV
FF-CV
RV
RV
RV
RV
c
c
RV
RV
RV
RV
RV+New CV
RV+New CV
RV+New CV
RV+New CV
RV+New FBR+CV
c
c
RV
RV
RV
RV
c
EF+New FBR-'-CV
e
e
FF
20
^0
20
20
50
100
20
20
20
20
50
50
50
50
40
50
100
20
20
20
20
60
40
50
100
20
Oxygen enrichment
Oxygen enrichment
Oxygen enrichment
Oxygen enrichment
Flash furnace
Flash furnace
Oxygen enrichment
Oxygen enrichment
Oxygen enrichment
Oxygen enrichment
Oxy-fuel burners
Oxy-fuel burners
Oxy-fuel burners
Oxy-fuel burners
Conversion to calcine charge
Flash furnace
Flash furnace
Oxygen enrichment
Oxygen enrichment
Oxygen enrichment
Oxygen enrichment
Flash furnace
Conversion to calcine charge
Flash furnace
Flash furnace
Oxygen enrichment
PB-SC/SA
LL
MgO-SC/SA
NH3-SC/SA
DC/DA°
DC/DA0
PB-SC/SA
LL
MgO-SC/SA
NH,-SC/SA
PB-DC/DA
LL
MgO-DC/DA
NH3-DC/DA ,
Not required
DC/DA°
DC/DAa
PB-SC/SA
LL
MgO-SC/SA
NH3-SC/SA
DC/DAa
DC/DA
DC/DA°
DC/DAa
DC/DA
aControl option covers only smelting furnace stream. Existing roasters and converters controlled by SC/SA
acid plant under Scenarios 1 through 17. Existing roasters, smelting furnaces, and converters controlled
by DC/DA acid plant under Scenario 18.
all scenarios.
New roasters and converters controlled by DC/DA acid plant under
No control required; existing plant postexpansion emissions are less than preexpansion.
and roaster controlled by DC/DA.
cExisting reverberatory furnace replaced by a flash smelting furnace.
For control of new flash furnace.
Existing electric furnace replaced by a flash smelting furnace.
New converter
CV = Converter.
DC/DA = Double contact/double absorption acid plant.
EF = Electric furnace.
FBR = Fluid-bed roaster.
FF = Flash furnace.
LL = Limestone FGD on partial converter stream.
MgO = Magnesium oxide FGD on partial converter stream.
MHR = Multihearth roaster.
NH3 = Cominco NH3 FGD on partial converter stream.
PB = Partial blending.
RV = Reverberatory furnace
SC/SA = Single contact/single absorption acid plant.
8-39
-------
Partial blending of the reverberatory furnace gas stream
with the roaster and converter streams and processing the
combined stream in an SC/SA acid plant. A DC/DA acid plant
is used to process those gas streams that result from the
addition of a new roaster or converter.
Processing the reverberatory furnace gas stream in a lime-
stone FGD.
Processing the reverberatory furnace gas stream in a MgO or
NH3 FGD, blending the resulting strong S02 stream with the
roaster and converter streams, and processing the combined
stream in an SC/SA acid plant. A DC/DA acid plant is also
used when oxyfuel burners are used to expand the reverbera-
tory furnace capacity.
For expanded flash furnaces or electric furnaces (Scenarios 23 and 26)
and conversion to flash smelting (Scenarios 5, 6, 16, 17, and 22), S02
control is achieved with expanded DC/DA acid plant capacity. For
conversion of an electric furnace to a flash furnace, no additional
DC/DA and plant capacity are required.
Incremental capital and annualized costs are presented in
Section 8.6.1 for the physical and operational changes that result in
the increased production capacity. Incremental costs for reducing the
increased emissions associated with each expansion option to preexpan-
sion levels are presented in Section 8.6.2.
Table 8-7 summarizes the input data used to calculate the
incremental capital, operating, and annualized costs for each expan-
sion scenario.
8.6.1 Incremental Capital and Annualized Process Costs for
Expansion Scenarios
8.6.1.1 Oxygen Enrichment. Expansion Scenarios 1 through 4, 7
through 10, 18 through 21, and 26 require the addition of oxygen
enrichment to an existing smelting furnace to attain a 20-percent
expansion.
The cost for adding oxygen enrichment to an existing reverberatory
furnace is considered to be approximately the same as for a new
reverberatory furnace because the hardware and control systems would
be identical. No additional process equipment is required for these
expansion scenarios involving oxygen enrichment. Thus, the capital
cost for oxygen enrichment is $510,000, as developed in Section 8.2.1.7.
8-40
-------
TABLE 8-7. INPUT DATH TO COST ESTIMATIONS, EXPANSION OPTIONS
Acid plant data
Expansion
option
Base
Base
Base
Base
Base
Case I "
1
2
3
1:
b
Case II
7
8
9
10
12
13
14
15
17 a
Case III
18
19
20
21a
22a
Case IV
23
Case V
24
25
26
Existing plant
Type Nm3/min scfm)
SC/SA
SC/SA
SC/SA
SC/SA
SC/SA
SC/SA
SC/SA
SC/SA
SC/SA
SC/SA
SC/SA
SC/SA
NA
NA
NA
NA
NA
SC/SA
SC/SA
SC/SA
SC/SA
SC/SA
SC/SA
SC/SA
SC/SA
DC/DA
DC/DA 1
DC/DA 2
DC/DA
DC/DA
OC/OA
DC/DA
3,660
3,660
3,660
3,660
3,660
3,660
3,660
4,510
4,510
4,510
4,510
4,510
NA
NA
NA
NA
NA
4,510
4,510
3,170
3,170
3,170
3,170
3,170
3,170
6,450
6,450
2,460
3,400
6,450
6,450
3,400
(129,200)
(129,200)
(129,200)
(129,200)
(129,200)
(129,200)
(129,200)
(159,000)
(159,000)
(159,000)
(159,000)
(159,000)
NA
NA
NA
NA
NA
(159,000)
(159,000)
(112,000)
(112,000)
(112,000)
(112,000)
(112,000)
(112,000)
(231,000)
(231,000)
(27,000)
(121,000)
(231,000)
(231,000)
(121,000)
S02 (%)
3.5
3.5
3.5
3.5
3.5
3.5
3.5
3.5
3.5
3.5
3.5
3.5
NA
NA
NA
NA
NA
3.5
3.5
3.5
3.5
3.5
3.5
3.5
3.5
4.0
4.0
3.5
6.0
4.0
4.0
6.0
Expanded plant
NmVmi n
NA
5,285
4,390
4,490
4,490
2,585
3,450
NA
6,315
5,430
5,580
5,490
NA
NA
NA
NA
NA
2,835
3,780
NA
4,120
3,390
3,420
3,420
2,135
NA
NA
NA
c
c
NA
(scfm)
NA
(187,000)
(155,000)
(159,000)
(159,000)
(92,000)
(122,000)
NA
(223,000)
(198,000)
(197,000)
(194,000)
NA
NA
NA
NA
NA
(100,000)
(134,000)
NA
(146,000)
(120,000)
(121,000)
(121,000)
(76,000)
NA
NA
NA
c
c
NA
S02 (%)
NA
3.5
3.5
3.5
3.5
3.5
3.5
NA
3.5
3.5
3.5
3.5
NA
NA
NA
NA
NA
3.5
3.5
NA
5.5
7.0
7.0
7.0
3.5
NA
4.0
NA
c
c
6.0
NnrVmin
NA
NA
NA
NA
NA
1,665
2,190
NA
NA
NA
NA
NA
2,715
1,250
1,680
1,425
11,765
1,905
2,540
NA
NA
NA
NA
NA
2,070
NA
NA
NA
NA
NA
NA
New DC/DA
(scfm)
NA
NA
NA
NA
NA
(59,000)
(78,000)
NA
NA
NA
NA
NA
(96,000)
(45,000)
(60,000)
(51,000)
(63,000)
(68,000)
(90,000)
NA
NA
NA
NA
NA
(73,000)
NA
NA
NA
NA
NA
NA
S02 (%)
NA
NA
NA
NA
NA
10.0
10.0
NA
NA
NA
NA
NA
3.5
3.5
5.0
5.0
6.0
10.0
10.0
NA
NA
NA
NA
NA
10.0
NA
NA
NA
NA
NA
NA
NA = Not applicable.
aNew flash furnace as replacement for reverberatory furnace/roaster.
bAfter dilution to 11 percent S02.
cExisting acid plant has sufficient capacity to handle expanded smelter S02 streams.
8-41
-------
The incremental annualized costs for oxygen enrichment include
the cost of oxygen, given as $47/ton (see Section 8.2.2.4), and main-
tenance and capital charges related to the hardware (no additional
labor is required). The oxygen usage rates for expansion scenarios
requiring oxygen are shown in Table 8-7 and are used to calculate the
oxygen cost. The maintenance and capital charges associated with the
oxygen enrichment and oxyfuel hardware are calculated with the method-
ology of Section 8.2.2. Based on these values, the incremental
operating and annualized costs were calculated with the results shown
below:
Expansion
scenario
through 4
Expansion
option
Oxygen
Incremental
capital costs
$0. 5 million
Incremental
operating
costs
$3.0 million
Incremetal
annuali zed
costs
$3.1 million
enrichment
7 through 10 Oxygen $0.5 million $2.6 million $2.7 million
enrichment
18 through 21 Oxygen $0.5 million $0.6 million $0.7 million
enrichment
26 Oxygen $0.5 million $1.4 million $1.4 million
enrichment
8.6.1.2 Oxyfuel Burners. Expansion Scenarios 11 through 14
require the addition of oxyfuel burners to greenfield reverberatory
furnaces to achieve a 50-percent expansion. In addition, a new
converter capable of processing 391 Mg/day (430 tons/day) is required.
The capital cost of the hardware required for oxyfuel burner
installation is estimated to be $1.0 million, as developed in Section
8.2.1.8. This hardware is identical to that required for a new rever-
beratory furnace, and the cost is thus the same. Based on information
given in Reference 1, the capital cost of a converter for these expan-
sion scenarios was estimated at $31.8 million.
The annualized costs were developed with the methodology of
Section 8.2.2 and the oxygen usage rates shown in Table 8-7. Additional
raw materials and operating labor are also required for the new convert-
er and were estimated from information given in Reference 1. Based on
this information, the following results were obtained:
8-42
-------
Expansion
scenario
.1 through 14
Expansion
option
Oxyfuel
Incremental
capital costs
$1.0 million
Incremental
operating
costs
$6.0 million
Incremental
annual i zed
costs
$6.2 million
burners
Converter $31.8 million $3.8 million $9.0 million
8.6.1.3 Conversion to Calcine Charge. Expansion Scenarios 15
and 23 require conversion to calcine charge to achieve a 40-percent
expansion. Expansion Scenario 15 involves the conversion of a green-
charged reverberatory smelting furnace to calcine charge, while Expan-
sion Scenario 23 involves the conversion of an electric furnace to
calcine charge. For both scenarios, a new fluid-bed roaster to handle
1,910 Mg/day (2,100 tons/day) of feed is required. A new converter is
also required for both scenarios. For Scenario 15, the new converter
handles 340 Mg/day (373 tons/day), and for Scenario 23, the new converter
handles 392 Mg/day (431 tons/day).
For Expansion Scenario 15, the capital costs of conversion to
calcine charge consist of modifications to the reverberatory furnace
and the costs of the fluid-bed roaster and converter. Modifications
to the reverberatory furnace include (see Section 3.4.4) adding a
solids handling system between the roaster and reverberatory furnace
(Wagstaff feeders) and installing in the wall a line of water-cooled
copper plates 12 to 18 inches high because the calcine has a smaller
angle of repose and does not dependably protect the refractories at
the slag line. The capital cost for these changes is estimated to be
$350,000, based on a rebuild of a reverberatory furnace, which, while
not designed specifically to convert a reverberatory to calcine charge,
did include a comparable scale of labor and materials.24 Based on
information given in Reference 1, the capital costs of a fluid-bed
roaster and a converter for this scenario were estimated at $15.8 mil-
lion and $27.8 million, respectively.
For Expansion Scenario 23, there is no modification necessary to
the electric furnace to convert to calcine charge (see Section 3.4.4).
The capital costs thus consist of the cost of the new fluid-bed roaster
and new converter. Based on information given in Reference 1, the
8-43
-------
cost of the new fluid-bed roaster is $15.8 million and the cost of the
new converter is $36.2 million.
The annualized costs for Expansion Scenarios 15 and 23 were
computed with the methodology of Section 8.2.2. For the fluid-bed
roaster, fuel oil is used to preheat the fluidizing air and during
startup. Electricity is used for the fluidizing fans with a much
smaller amount needed for solids handling and other operations. Elec-
tricity and silica flux are used in the converter. Flux is added to
form the matte layer, and electricity is required for the air compressors
used to blow the converter charge. Slight differences in the annualized
costs for Scenarios 15 and 23 are due to differences in feed composi-
tion. Additional labor is required for both the fluid-bed roaster and
the converter. The additional labor, raw material, and utility usage
rates above the preexpansion levels are shown below as estimated from
published information:1
Expansion Expansion
Scenario 13 Scenario 18
Silica flux 95,700 tons/year 95,700 tons/yr
Additional operating $0.45/ton cone. $0.45/ton cone.
supplies*
Direct operating labor 5 persons/shift 5 persons/shift
Bunker C fuel oil 6.1 x 106 gal/yr 5.8 x 106 gal/yr
Electricity 14.2 x io6 kWh/yr 13,. 9 x io6 kWh/yr
*This includes various additives, operating materials, and miscellan-
eous expenses. For both expansion scenarios, 545 Mg/day (600 tons/day)
additional concentrate on a dry basis are processed.
The results of the calculation of capital and annualized costs
for Expansion Scenarios 15 and 23 are shown below:
8-44
-------
Expansion
scenario
15
23
8.6.1.
Expansion
option
Reverb modi-
fications
Fluid-bed
roaster
Converter
Fluid-bed
roaster
Converter
4 Conversion
capital costs
(106 $)
0.4
15.8
27.8
15.8
30.8
to Flash Smelting
Annual i zed
Annuali zed operating Annual i zed
costs costs
UP6 $) dO6 $)
-0- 0.1
6.7 9.3
3.8 8.3
6.7 9.3
3.8 8.8
16, 17, 22, 24, and 25 involve replacement of a green-charged reverberatory
furnace or a roaster and calcine charge reverberatory furnace with an
oxygen flash furnace. For scenarios 5, 16, and 24, the flash furnace
is sized to result in a 50-percent expansion; for scenarios 6, 17, and
25, a 100-percent expansion; and for scenario 22, a 60-percent expansion.
Because of the higher matte grade (55 percent), the existing converter
capacity should be adequate to handle the increased throughput without
modification. Process capital costs include the cost of the flash
smelting furnace and additional dryer capacity. The capital cost of
the new oxygen flash furnace and additional dryer capacity, based on
information contained in Reference 25, is $36.9 million. Operating
costs were developed with the methodology described in Section 8.2.2
and the oxygen usage rates shown in Table 8-7. The following process
cost estimates were developed on this basis.
Conversion from reverberatory to flash smelting results in a
considerable reduction in process operating costs because of the
elimination of the fuel requirements for reverberatory furnaces. For
a reverberatory furnace processing 1,364 Mg/day (1,500 tons/day) of
concentrate, this is estimated to be about 69.5 million using 4.5 x
106 Btu per ton of charge. As a result, for scenarios 5, 16, and 22,
operating costs are less than baseline costs, hence the net reduction
in incremental operating costs as shown in the following table (figures
in parentheses represent an incremental reduction in costs).
8-45
-------
Expansion
scenario
5
6
16
17
22
24
25
Expansion
option
New flash
smelter
New flash
smelter
New flash
smelter
New flash
smelter
New flash
smelter
New flash
smelter
New flash
smelter
i.6.2 Incremental Capital
8.6.2.
Incremental
capital costs
($106)
31.0
36.9
31.0
36.9
32.2
31.0
36.9
and Annual i zed
Incremental
operating
costs
($106)
(2.0)
7.8
(1.3)
8.6
(3.. 3)
7,1
19.8
Incremental
annual ized
costs
($106)
3.1
13.8
3.8
14.6
2.0
12.1
25.8
Costs for Control
, 1 Capital Costs. Capital costs for SC/SA
sulfuric acid
plants were obtained from data in Reference 1 and are shown graphically
in Figure 8-5. Based on this information, the following mathematical
expression relating capital costs to flow rate and S02 concentration,
was developed:
In Cost = 0.1325 + 0.351 (% S02)°-152 + 0.731 (% S02)"0'087 [In Q - 3.418]
where
Cost is total capital cost in millions of June 1981 dollars
% S02 is expressed as a percent, not a fraction (i.e., 4% S02
would be input as 4.0)
Q is in NmVmin.
In estimating the incremental capital costs for expansion scenarios
involving an expanded SC/SA acid plant (Scenarios 1 through 4, 7
through 10, and 18 through 21) to treat expanded roaster and/or con-
verter streams and the reverberatory furnace slipstream, the costs for
an entire new SC/SA acid plant, based on postexpansion flow rates and
8-46
-------
100
80-
$2 60
o
•o
T—
00
O)
r-
0)
3
w>
O
w>
O
u
50
40
30
« 20
'a
(3
10
3.5% SO2
.0% SO2
.5% SO2
40 60 80
Flow Rate, 103 scfm
100 120 140
Note: Costs include gas cleaning and conditioning, absorption and acid production,
auxiliary preheating, storage facilities, materials handling, and control equipment.
Source: Reference 1.
Figure 8-5. Capital cost of an SC/SA sulfuric acid plant.
8-47
-------
S02 concentrations (shown in Table 8-7), were calculated with the
equation developed above. From this value was subtracted the cost,
calculated with the same equation, for an entire new SC/SA acid plant
using the preexpansion f.ow rates and S02 concentrations. The differ-
ence between these two values was then designated as the cost of
expanding the existing acid plant. It is recognized that the cost of
expanding an existing acid plant is likely to be site-specific. For
example, if the existing acid plant were old, it might be infeasible
to expand it. In such a case, the control costs could be the cost of
a new plant at postexpansion flows. In most cases, plant data upon
which individual site analyses could be made are not available to EPA.
However, EPA is mandated to assess the environmental, energy, and
economic impacts as an essential part of the standard development
process. Model plant analysis has been determined to be a reasonable
and cost-effective manner in which to perform these assessments.
Model plants are not intended to represent what an individual plant
should look like, but rather to form the basis for subsequent analyses
of the impact of the regulation on the industry as a whole.
The same procedure was followed in estimating incremental capital
costs using the cost relationship in Section 8.2.1.1 for scenario 24
that requires expansion of an existing DC/DA acid plant to process the
increased flash furnace and converter flows.
New DC/DA acid plants are required for expansion scenarios 5, 6,
11 through 17, 22, and 23. For scenarios 5, 6, 16, 17, and 22, in
which flash furnaces replace reverberatory furnaces, the new DC/DA
acid plants are required to process the strong S02 offgas streams from
the new flash furnaces as required by the existing NSPS. For scenarios
7 through 11, the new DC/DA plants are required to treat the strong
S02 offgas streams from new roasters and/or converters as well as the
slipstream to reduce reverberatory furnace S02 emissions to preexpan-
sion levels. For scenario 23, the total strong stream flows after
expansion are large enough to justify a new DC/DA acid plant in
addition to the existing plant. The new plant will treat the new
fluid-bed roaster and expanded electric furnace offgas streams; the
8-48
-------
existing plant, both new and old converter flows. Costs of these new
DC/DA acid plants are estimated as described in Section 8.2.11.
The incremental capital costs for the MgO, NH3, and limestone FGD
systems (Scenarios 2, 3, 4, 8, 9, 10, 12, 13, 14, 19, 20, and 21) for
the expanded plants were calculated with the equations developed in
Sections 8.2.1.3, R.2.1.4, and 8.2.1.5, respectively. The input flow
rates and S02 concentrations to these equations are taken from Table 8-7
under the column labeled "New FGD requirement." It can be seen that
some of the flow rates are small and fall out of the range of the
costs given in Figures 8-2 (MgO), 8-3 (NH3), and 8-4 (limestone).
Nonetheless, the equations given in Sections 8.2.1.3 through 8.2.1.5
were used to calculate FGD capital costs because they are considered
to be reasonable approximations of the costs for systems in this low
flow range.
8.6.2.2 Incremental Annualized Costs. Incremental annualized
costs for the DC/DA acid plants were estimated using the procedures
described in 8.2.2. Annualized costs for the SC/SA acid plant were
based on the following raw material, labor, and utility usage rates:
Labor1 : 3 persons/shift
Process water1: {[11,769-1,051 In [(Q)(% S02)]} (0.0115)(Q)(% S02)
thousands of gal/yr
Cooling water1: (.12 x 103)(Q) thousands of gal/yr
Electricity1 : [3.64 x 104 - 1.370(Q)] (Q) kWh/yr
Limestone6 : (0.482) (Q) tons/yr,
where
Q is in Nm3/min
% S02 is expressed as a percent, not a fraction (i.e., 4% S02
would be input as 4.0).
The incremental annualized costs for the expanded SC/SA acid
plants were calculated in the same manner as the capital costs; i.e.,
the annual ized costs for the "existing plant" were subtracted from the
annualized costs for the expanded plant.
8-49
-------
The incremental annualized costs for the FGD systems were calcu-
lated as discussed in Sections 8.2.2.3 (MgO), 8.2.2.4 (NH3), and
8.2.2.5 (limestone). The input labor, raw material, and utility usage
rates for the annualized costs are the same as those given in these
sections. The input flow rates and S02 concentrations are taken from
Table 8-7.
Supplemental heat is required for the acid plants required in
Expansion Scenarios 1 and 7 (see Table 8-7). Table 8-8 is the result
of a calculation of the supplemental heat required for each of these
scenarios using the methodology discussed in Section 8.2.2.
A detailed itemized list of the process and control costs for
each expansion scenario, except those involving conversion to flash
smelting, is given in Appendix 0. Costs for the flash smelting option
are detailed in Section 8.6.1.4.
8.6.3 Summary of Expansion Scenario Incremental Costs
Table 8-9 summarizes the incremental capital annualized costs for
each expansion scenario. It should be noted that the incremental
costs presented include the incremental costs both for the process
change and for controlling S02 emissions from all offgas streams.
For subsequent analyses, the cost of control of new facilities
producing strong S02 offgas streams (i.e., new roasters, flash furnaces,
and converters) is in the baseline. Costs for weak stream control,
(i.e., reverberatory furnace offgas streams) are the incremental costs
attributable to the NSPS. Only the costs of S02 control on the rever-
beratory furnace slipstream are attributable to the standard. All
other costs are included in the baseline.
8.7 COST-EFFECTIVENESS
Cost-effectiveness of the alternative techniques for the control
of weak S02 streams at new greenfield smelters is shown in Table 8-10.
Cost-effectiveness ranges from $53 per Mg of S02 removed for Option I-G
(oxyfuel) to $203 for Option I-B (MgO FGD plus acid plant). Option I-G
is also the least costly, requiring a capital investment of $55.2 million,
or an incremental increase over the baseline case of $8.9 million.
8-50
-------
TABLE 8-8 SUMMARY OF INCREMENTAL COSTS INCURRED DUE TO
ACID PLANT PREHEATER OPERATION
Heat Estimated
deficiency per increase
Type of 24-hour cycle in annual
Plant3 acid plant (Btu) operating cost
Expansion Scenario 1 SC/SA 2.0 x 107 $60,000
Expansion Scenario 7 DC/DA 5.2 x 106 $14,300
aSee Table 8-7.
8-51
-------
TABLE 8-9. EXPANSION COSTS
(INCLUDES COST OF CONTROLLING S02 EMISSIONS FROM NEW ROASTERS AND CONVERTERS AS REQUIRED BY EXISTING NSPS)
CO
I
on
ro
Expansion
capability
scenario Configuration (%)
1
2
3
4
b
b
7
8
9
10
11
12
13
14
Ib
16
I/
18
19
20
21
22
23
24
2b
26
Oxygen enrichment
Option 1 plus limestone FGD
Option 1 plus HgO FGO
Option 1 plus NH3 FGD
Conversion to flash smelting
Conversion to flash smelting
Oxygen enrichment
Option 1 plus limestone FGD
Option 1 plus MgO FGD
Option 1 plus NH3 FGD
Oxyfuel burners, new CV, new DC/DA
Option 11 plus limestone FGD
Option 11 plus MgO FGD
Option 11 plus NH3 FGD
Converter to calcine charge, new
FBR, new CV, new DC/DA
Conversion to flash smelting
Conversion to flash smelting
Oxygen enrichment
Option 18 plus limestone FGD
Option 18 plus MgO FGD
Option 18 plus NH3 FGD
Conversion to flash smelting
Conversion to calcine charge, new
FBR, new CV
Conversion to flash smelting
Conversion to flash smelting
Oxygen enrichment
20
20
20
20
50
100
20
20
20
20
50
50
50
50
40
50
100
20
20
20
20
60
40
50
100
20
(Mg
copper)
20,910
20,910
20,910
20,910
52,260
104,520
21 , 380
21,380
21,380
21,380
53,450
53,450
53,450
53,450
42,760
53,450
106,910
18,710
18,710
18,710
18,710
56,310
54,790
68,490
136,980
20,240
Incremental
capital costs ($103)
(June 1981 dollars)
Process
31
36
32
32
32
32
44
31
36
32
46
31
36
510
510
510
510
,010
,860
510
510
510
510
,800
,800
,800
,800
,000
,010
,860
510
510
510
510
,240
,600
,010
,860
510
Control
10,910
15,560
17,590
13,460
23,810
28,350
11,360
17,510
20,200
15,580
40,010
41,320
47,460
37,420
26,840
25,890
30,910
2,940
5,700
6,970
2,910
26,650
37,500
0
0
5,090
Total
11,420
16,070
18,100
13,970
54,820
65,210
11,870
18,020
20,710
16,090
72,810
74,210
80 , 260
70,220
70,840
56,900
67,770
3,450
6,210
7,480
3,420
58,890
84,100
31,010
36,860
5.600
Incremental .
operating costs ($103)
(June 1981 dollars)
Process
3,040
3,040
3,040
3,040
(1,970)
7,750
2,590
2,590
2,590
2,590
9,790
9,790
9,790
9,790
10,520
(1,300)
8,630
620
620
620
620
(3,290)
10,470
7,080
19,800
1.360
Control
2,960
3,670
4,370
4,760
2,910
5,540
2,960
4,300
5,450
6,210
9,740
9,940
13,23U
14,520
6,630
3,810
7,280
1,430
1,860
2,230
2,370
5,230
9,030
0
0
2.270
Total
6,000
6,710
7,410
7,800
940
13,290
5,550
6,890
8,040
8,800
19,530
19,730
23.U20
24,310
17,150
2,510
16,910
2,050
2,480
2,850
2,990
1,940
19,500
7,080
19,800
3,630
Total
incremental
annualized
($103)
dollars total)
7,860
9,330
10,350
10,080
9,860
23,900
;,4to
9,82
11,420
11,420
31,380
31,800
36,080
35,740
28,680
11,780
26,940
2,610
3,490
4,070
3,240
11,530
33,180
12,130
25,800
4,540
Basea on a Base case feed of the following composition:
Copper content
Scenario of feed (%)
1-6 21.9
6-17 22.4
17-22 19.6
23-25 28.7
26 21.2
Based on 350 day/year operation.
Based on 10 percent discount rate over 10 years or 16.275 percent of capital costs
FGD = Flue gas desulfurization
CV = Converter.
FBR = Fluid-bed roaster
DC/DA = Double contact/double absorption sulfuric acid plant
FF = Flash furnace.
-------
OD
I
cn
OJ
TABLE 8-10 COST-EFFECTIVENESS: CONTROL OF REVERBERATORY FURNACE S02 EMISSIONS IN A NEW COPPER SMELTER (MULTIHEARTH ROASTERS,
REVERBERATORY FURNACE, CONVERTER) PROCESSING HIGH-IMPURITY MATERIALS'1
Emissions (Mg/yr)
Reverber- Reduction Capital costs ($103)
atory from Base
Operating costs ($103)
Incre-
mental Cost-
annu- effec-
Annualized costs ($103)c alized tiveness
costs ($/Mg S02
Control option furnace ""case"6 Processd Control Total Processd Control Total Processd Control Total ($103) removed)
73,360
162,000 46,280 208,280 31,110 12,320 43,430 57,480 19,850 77,330
Baseline Case
(no control on
RV, MHR and CV
to DC/DA)
Option I-A 40,910
(Blending 45% of RV
stream with MHR and
CV, DC/DA)
Option I-B (RV to 8,460 64,900 162,000 74,000 236,000 31,110 20,950 52,060 57,480
MgO FGD; strong
stream, MHR, and CV
to DC/DA)
Option I-C (RV to 8,460 64,900 162,000 62,840 224,840 31,110 22,340 53,450 57,480
NH3 FGD; strong
stream, MHR, and
34,450 162,000 61,190 223,190 31,110 16,870 47,980 57,480 26,830 a4,310 6,980
33,000 90,480 13,150
32,570 90,050 12,720
CV to DC/DA)
Option I-D (RV to 7,340 66,020 162,000 69,740
lime/limestone FGD,
MHR and CV to DC/DA)
Option I-E 1,250
(Blending 100% RV
stream with MHR and
CV, DC/DA)
Option I-F (02 1,250
enrichment, blending
100% RV with MHR and
CV, DC/DA)
Option I-G (Oxyfuel, 1,250
blending 100% RV with
MHR and CV, DC/DA)
231,740 31,110 17,730 48,840 57,480 29,080 86,560 9,230
72,110 162,000 74,230 236,230 31,110 20,640 51,750 57,480 32,720 90,200 12,870
72,110 162,000 67,200 229,200 29,390 20,000 49,390 55,760 30,930 86,690 9,360
72,110
162,000 55,220 217,220 27,290 18,520 45,810 53,660 27,510 81,170 3,840
aMultihearth roasters and converters are covered by existing NSPS.
Based on 350 days per year operation.
GBased on 10 percent discount rate over 10 years or 16,275 percent of capital costs.
dCost of DC/DA to control multihearth roaster and converter included in Baseline Case.
RV = Reverberatory furnace
MHR = Multihearth roaster
CV = Converter
DC/DA = Double contact/double absorption sulfuric acid plant
202
203
196
140
179
130
53
-------
TABLE 8-11. COSTS FOR C NTROL OF FUGITIVE PARTICULATE MATTER EMISSIONS BY SOURCE,
NEW GREENFIELD SMELTER
(All costs in thousands of June 1981 dollars)
Source
Emissions (Mg/yr)
Reduction
b *rom
Total base case
Capital
costs
($103)
(one cost
per source)
Operating
costs
($103)
Incremental
annual! zed
costs
($103)
Cost-
effec-
tiveness
($/Mg)
I MHR
Calcine discharge 568
II RV
Matte tapping 37
Slag skimming 34
III FF-ESCF
FF-matte tapping 22
FF-slag skimming 44
ESCF-matte tapping 8
ESCF-slag skimming 49
IV CV-BE 1,094
V CV-AC 1,094
VI CV (w/FF) BE 798
AC 798
571
1,539
506
33
30
19
39
7
42
1,028 12,213
975 4,133
750 12,213
711 4,133
141
283
415
234
462
533 8,460
849 7,935
1,850 4,185
728
1,401
1,850 4,185
728 1,401
4,071
1,437
5,580
1,970
Based on 350 day/year operation.
bSee Section 3.3.4, emissions are based on 1,360 Mg/day feed and 352 Mg/day blister copper.
cBased on 90 percent capture, except for building evacuation that uses 95 percent, and
99 percent collection efficiency.
dBased on 10 percent discount rate over 10 years or 16.275 percent of capital costs, see
Appendix N.
MHR = Multihearth roaster.
RV = Reverberatory furnace.
FF = Flash furnace.
ESCF = Electricity slag cleaning furnace.
CV = Converter.
BE = Building evacuation.
AC = Air curtain.
8-54
-------
TABLE 8-12. COST-EFFECTIVENESS OF EXPANSION SCENARIOS
00
I
CJ1
Base
case
Smelters
Ia
II
III
IV
V
Incremental
annual i zed
costs ($103)
Expansion Existing
scenario NSPS
processing
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
clean materials
7,860
9,330
10,350
10,080
9,860
23,900
7,480
9,820
11,420
11,420
31,380
31,800
36,080
35,740
28,680
11,770
26,940
2,610
3,490
4,070
3,240
11,530
33,180
12,130
25,800
4,540
Cost-effectiveness SO
Production
(Mg blister/yr)
Total
125,430
125,430
125,430
125,430
156,820
209,110
128,290
128,290
128,290
160,300
160,300
160,300
160,300
160,300
149,670
160,410
213,880
112,250
112,250
112,250
112,250
149,720
191,770
205,520
274,030
121,420
Incremental
20,910
20,910
20,910
20,910
52,270
104,550
21,380
21,380
21,380
53,450
53,450
53,450
53,450
53,450
42,760
53,470
106,940
18,710
18,710
18,710
18,710
56,140
54,790
68,510
137,010
20,240
S02 removed
(Mg/yr)
47,340
47,340
47,340
47,340
59,180
118,360
54,220
54,220
54,220
135,500
135,500
135,500
135,500
135,500
108,440
67,770
135,540
55,170
55,170
55,170
55,170
82,760
113,780
71,110
142,220
59,180
$/Mg S02
removed
166
197
219
213
167
202
138
181
211
211
232
235
266
264
264
174
199
47
63
74
59
139
292
171
181
77
Existing NSPS
$/Mg total
blister
63
74
83
80
63
114
58
77
89
89
196
198
225
223
191
73
126
23
31
36
29
77
173
59
94
37
o removal
$/Mg
incremental
blister
376
446
1QS
482
189
229
350
459
534
534
587
595
675
669
671
220
252
140
187
218
173
205
606
177
188
224
Applicable to both clean and high-impurity materials.
-------
TABLE 8-13. COST-EFFECTIVENESS, FUGITIVE PARTICIPATE MATTER CONTROL, EXPANSION SCENARIOS8
CO
i
en
CTl
-_ --__-_:=!_::.—
Base
case
Ib
II
III
IV
V
Expansion
scenarios
1-4
5
6
7-10
11-14
15
16
17
18-21
22
23
24
25
26
Annual ized
control
costs
($103)
.
_
-
579
579
-
-
579
-
-
Converter
Particulatt-
removed
(Mg/yr)
_
_
-
485
390
-
-
355
_
-
-
Smelting furnace
Incremental
cost-
effective-
ness
($/Mg)
_
_
-
1,194
1,485
_
-
-
1,631
-
-
-
Annual ized
control
costs
($103)
640
640
640
640
800
746
640
640
640
640
746
640
640
640
Particulate
removed
(Mg/yr)
70
85
110
75
95
90
85
110
65
80
100
110
145
65
Incremental
cost-
effective-
ness
($/Mg)
9,140
7,530
5,8?0
8,530
8,420
8,290
7,530
5,820
9,850
8,000
7,460
5,820
4,410
9,850
aNo new multihearth roaster required for any expansion scenario.
bApplicable to smelters processing clean or high-impurity material.
-------
TABLE 8-14. INCREMENTAL COST DATA, LEAST COST EXPANSION SCENARIOS
(June 1981 dollars)
oo
i
en
Smelter Expansion Percent
configuration Scenario expansion
Smelters Processing
I (MHR-RV-CV)
II (RV-CV)
III (FBR-CV-RV)
IV (EF-CV)
V (FF-CV)
Smelters Processing
I (MHR-RV-CV)
Reverberatory
Baseline furnace
annualized annualized
costs costs
($103) ($103)
Total
annualized
costs
($103)
Incremental
blister production
Mg/yr (tons/yr)
Clean Materials
1
5
6
7
11
15
16
17
18
22
23
24
25
26
High- Impurity
1
20
50
100
20
50
40
50
100
20
60
40
50
100
20
Material
20
6,700 1,160
9,860
23,900
6,460 1,020
22,670 8,710
28,680
11,780
26,940
2,210 400
11,530
33,180
12,130
25,800
4,540
67,000
7,860
9,860
23,900
7,480
31,380
28,680
11,780
26,940
2,610
11,530
33,180
12,130
25,800
4,540
6,700
20,910
52,270
104,550
21,380
53,450
42,760
53,470
106,940
18,710
56,340
54,790
68,510
137,010
20,240
20,910
(23,000)
(57,500)
(115,000)
(23,520)
(58,800)
(47,040)
(58,820)
(117,630)
(20,580)
(61,970)
(60,270)
(75,360)
(150,710)
(22,260)
(23,000)
-------
TABLE 8-15. INCREMENTAL COST DATA, FUGITIVE EMISSION CONTROL
LEAST COST EXPANSION SCENARIOS
June 1981 dollars
Incremental annual ized costs
Smelter
configuration
I (MHR-RV-CV)3
II (RV-CV)
III (FBR-RV-CV)
IV (EF-CV)
V (FF-CV)
Expansion
Scenario
1
5
6
7
11
15
16
17
18
22
23
24
25
26
Converter
0
0
0
0
580
580
0
0
0
0
580
0
0
0
Smelting
furnace
640
640
640
640
800
750
640
640
640
640
750
640
640
640
Also applicable to smelters processing high-impurity materials.
3-58
-------
Cost-effectiveness ratios for fugitive controls are shown in
Table 8-11. Cost-effectiveness of fugitive controls ranges from $462
per Mg of particulate matter removed for the multihearth roaster
calcine discharge operation to $7,935 per Mg for smelting furnace
matte tapping and slag skimming operations. For converter fugitive
controls, the use of an air curtain and secondary hood is more cost
effective than building evacuation, $1,437 per Mg versus $4,071 for
the MHR-RV-CV smelter configuration and $1,970 per Mg versus $5,580
for the FF-CV configuration. The primary reason for these differences
is the larger volume of air that must be moved in building evacuation
systems, more than quadruple that moved with air curtain/secondary
hood systems.
Cost-effectiveness ratios for the expansion scenarios analyzed
are shown in Tables 8-12 and 8-13 for weak-stream S02 and fugitive
particulate matter control, respectively. Incremental cost data for
selected expansion scenarios are shown in Table 8-14. Baseline costs
include control costs associated with strong stream control. Incre-
mental annualized costs shown in Table 8-12 include control of both
strong and weak S02 streams. These costs are partitioned into weak
S02 stream costs, shown in Table 8-14, and strong S02 stream costs,
included in baseline costs in Table 8-14 by the ratio of reverberatory
furnace slipstream flow rates to acid plant average flow rates shown
in Table 6-8. Fugitive emission control costs are shown in Table 8-15
for each of the least cost expansion scenarios.
8.8 REFERENCES
1. Weisenberg, I. J., and T. Archer, "Feasibility of Primary Copper
Smelter Weak S02 Stream Control Relative to Reverberatory Furnace
NSPS Exemption," Draft Final Report, July 1978.
2. Agarwal, J. C., and M. L. Loreth, "Preliminary Economic Analysis
of S02 Abatement Technologies," Presented at the AIME Meeting,
Chicago, February 22-26, 1981.
3. Hayashi, M., "Cost of Producing Copper from Chalcopyrite Concen-
trate as Related to S02 Emission Abatement," Bureau of Mines
report 7957, U.S. Department of Interior, 1974, p. 30.
8-59
-------
4. PEDCo Environmental, Inc., "User's Guide, Computerized Approach
to Estimating S02 Scrubber Costs at Nonferrous Smelters," EPA
Contract 68-03-2924, April 1982.
5. ASARCO, Response to EPA Section 114 letter, January 11, 1982.
6. Mathews, J. C., F. L. Belligia, C. H. Gooding, and G. E. Weant,
"S02 Control Processes for Nonferrous Smelters," EPA-600/2-76-008,
January 1976.
7. Williamson, P. C., and Puschaver, E.J., "Ammonia Absorption/
Ammonium Sulfate Regeneration Pilot Plant for Flue Gas Desulfuri-
zation," Tennessee Valley Authority, Muscle Shoals, Alabama,
Bulletin Y-116, August 1977.
8. Mcllvaine Co., Inc., "The Mcllvaine Scrubber Manual," The
Mcllvaine Company, Northbrook, Illinois, 1974, Ch. IX, Sec-
tion 4911-900, p. 176.
9. Telecon, Clark, T. C., Research Triangle Institute, with
Hoffman, D., Union Carbide Corporation. April 26, 1982. Reverb-
eratory Furnace Oxy-Fuel Burner Costs.
10. Telecon, Clark, T. C., Research Triangle Institute, with Zullo, A.,
Hauck Manufacturing Company, April 23, 1982. Reverberatory
Furnace Burner Systems Costs.
11. Neveril, R. B., "Capital and Operating Costs of Selected Air
Pollution Control Systems," EPA-450/5-80-002, December 1978.
12. Telecon, Clark, T. C., Research Triangle Institute, with H. C.
Garven, Inco, July 6, 1982'.
13. Anderson, K. D. et al., "Definitive SO Control Process Evalua-
tions: Limestone, Lime, and Magnesia ?GD Processes," TVA ECDP B-7,
January 1980.
14. Biswas, A. K., and W. G. Davenport, "Extractive Metallurgy of
Copper," 2nd edition, New York, Pergammon Press, 1980.
15. Kerry, F. G., "Technical and Economic Considerations of Very
Large Oxygen Plants, TMS/AIME, paper A75-61.
16. Viner, A. S. and D. S. Ensor, "Computer Programs for Estimating
the Cost of Particulate Control Equipment," prepared for U.S.
Environmental Protection Agency, Research Triangle Park, North
Carolina, August 1981.
17. Viner, A. S. and D. S. Ensor, "Computer Programs for Estimating
the Cost of Particulate Control Equipment," prepared by the
Research Triangle Institute for EPA-IERL, Research Triangle Park,
N.C., August 1981.
8-60
-------
18. U.S. Environmental Protection Agency, "Arsenic Emissions from
Primary Copper Smelters—Background Information for Proposed
Standards—Preliminary Draft," February 1981.
19 ASARCO Design Report: Converter and Secondary Hooding for the
Tacoma Plant, A. 0. March, Jr., Central Engineering Report (Salt
Lake City), January 22, 1981.
20. Strauss, W., "Air Pollution Control, Part I," Wiley-Interscience,
New York, 1971, p. 253.
21. Campbell, K. S., et al., "Economic Evaluation of Fabric Filtration
Versus Electrostatic Precipitation for Ultrahigh Particle Effi-
ciency," Electric Power Research Institute Report FP-775, June
1978, p. 111-18.
22. Securities and Exchange Commission, Washington, D.C., Form 10-K
for Phelps Dodge Corporation for the years 1974-1976.
23. Harkki, S., Outokumpu, letter to T. C. Clark, Research Triangle
Institute, April 22, 1981.
24. Telecon, Clark, T. C., Research Triangle Institute, with ASARCO-
Hayden, October 13, 1981.
25. Antonioni, T. N. et al. Inco Oxygen Flash Smelting of Copper
Concentrates. Inco Metals Company, Toronto, Ontario (Presented
at the British Columbia Copper Smelting and Refining Technologies
Seminar, Vancouver, November 5, 1980), 40 pp.
26. Perry, R. H., and C. H. Chilton, editors, Chemical Engineering
Handbook. 5th Edition, McGraw Hill, New York, 1973, p. 25-16.
8-61
-------
9.0 ECONOMIC IMPACT
This chapter describes the primary copper industry (Section 9.1)
using data that are subsequently used in an economic impact analysis
of the industry (Sections 9.2 and 9.3).
The environmental impacts of the primary copper industry have
been the subject of considerable study recently. This report uses
findings from these studies and others as appropriate. New or addi-
tional material not reported elsewhere is also included, such as the
copper flows from each mine to each smelter and various production
cost estimates.
The economic profile focuses on several primary copper industry
characteristics: number and location of smelters and refiners, copper
supplies, prices and costs, value of shipments, competition, growth,
and employment. The majority of data available for developing the
industry profile are from the years 1979 and 1980.
9.1 INDUSTRY ECONOMIC PROFILE
9.1.1 Introduction
Copper's utility stems from its qualities of electrical and
thermal conductivity, durability, corrosion resistance, low melting
point, strength, malleability, and ductility. Principal uses are in
transportation, machinery, electronics, and construction.
The Standard Industrial Classification (SIC) Code definition of
the primary copper industry is the processes of mining, milling,
smelting, and refining copper. Total 1980 employment, according to
Bureau of Labor Statistics figures, averaged 29,400 workers. Copper-
bearing ore deposits and substantial amounts of copper scrap provide
the raw materials for these processes.
9-1
-------
In addition to producing copper, the industry markets by-product
minerals and metals that are extracted from the ore deposits, such as
silver, gold, zinc, lead, molybdenum, selenium, arsenic, cadmium,
titanium, and tellurium. Many of the companies that own primary
copper facilities also fabricate copper. Many of these same companies
are also highly diversified and produce other metals, minerals, and
fuels.
The standard under consideration directly affects only one of the
four primary copper processes—smelting. However, the other three
related processes are an integral part of the ownership and economic
structures of copper smelters and therefore must be examined in deter-
mining industry impact. Mining and milling processes supplying a
smelter will be impacted secondarily by a smelter impact because
transportation costs to an alternate smelter will add a sizable business
cost. Transportation costs for concentrate are significant because
only 25 to 35 percent of the concentrate is copper and the remaining
75 to 65 percent of the material being transported is waste material.
The same interdependence between smelter and refiner is not as critical
since the copper content at this stage is typically 98 percent.
Even if there were no business dependencies among the processes,
the available financial data for smelters are aggregated in consoli-
dated financial statements, which makes smelter data difficult to
isolate. Thus, an impact analysis on copper smelters must address the
economic connection backward to the mines and forward through the
refining stage.
9.1.2 The Copper Smelters—Ownership, Location, Concentration
There are 15 pyrometallurgical copper smelters in the United
States as shown in Figure 9-1. Copper is also produced in limited
amounts by various hydrometallurgical methods which by-pass the smelting
stage. These hydrometallurgical facilities are not being considered
in the standard-setting process. In 1980 the 15 copper smelters had a
production capacity* of 1,693,000 Mg^ of copper. The hydrometallurgical
^Capacity is not a static measure of a smelter since capacity can
vary; for example, according to the grade of copper concentrates
processed.
Mg = 1.1 short tons.
9-2
-------
UD
OJ
Principal copper mining states
\ £-,
^
Figure 9-1. Principal mining States and copper smelting and refining plants, 1978.l
-------
processes in 1979 had a capacity of 189,000 Mg of copper or 10.1 percent
of the copper smelters' capacity. According to the Copper Development
Association, Inc. (CDA), actual smelter production in 1979 was
1,395,600 Mg and the Bureau of Mines estimated the hydrometallurgical
production at 98,800 Mg.
The 15 U.S. copper smelters are owned by seven companies. Although
this is a small number of companies, the 'industry is somewhat less
concentrated than in the early 1950's when Kennecott, Phelps Dodge,
Anaconda, and ASARCO dominated the industry share. All seven companies
are integrated in that they own some mining and milling facilities
that produce copper concentrates for the smelters. On the other hand,
several smelters buy concentrates from other mining and milling pro-
ducers, smelt and refine the copper, and then sell it. This practice
is referred to as custom smelting. Other smelters process (smelt and
refine) the concentrates and return the blister copper to mine owners
for them to sell, a practice referred to as tolling. Some smelters
perform both toll and custom smelting. It is important, in determining
the smelter's ability to absorb or to "pass back" potential pollution
control costs, to know whether the concentrates came from the smelter-
owned mines or from other sources on a tolling or custom basis.
It is general industry practice for companies to operate their
smelters as service centers at low profit margins to the owned mines.
This acts to shift profits of an integrated operation to the mines
where depletion allowances exist, thus maximizing profit to the overall
operation. An implication of this policy is that the impact on profits
from swings in copper prices frequently is manifested at the mines
more than at the smelters.
Table 9-1 lists the smelters, their corporate owners, capacities,
1979 and 1980 production amounts, and the distribution of integrated,
custom, and toll smelting. The total production figures and the
corresponding operating rates are shown in Table 9-1 for 1979, one
compiled from corporate reports and one reported by the Copper
Development Association, Inc. For 1979, the table shows an 83-percent
operating rate for corporate-reported amounts and a 74.7-percent rate
9-4
-------
TABLE 9-1. SMELTER OWNERSHIP, PRODUCTION, AND SOURCE MATERIAL ARRANGEMENTS
Smelter name
and location
Anaconda, MT
Tacoma, WA
hayden, AZ
El Paso, TX
Copper-hill, TN
White Pine, MI
Miami, AZ
McGill, NV
Garfield, UT
Hayden, AZ
Hurley, NM
Magma (San
Manuel , AZ)
Douglas, AZ
A jo, AZ
Morenci , AZ
Hidalgo, NM
Total production
(operating rate)
CDA production
(operating rate)
Ownership
ARCO
ASARCO, Inc.
Cities Service Co.
Copper Range Co,
Subsidiary of the Louisiana
Land Exploration Co.
Inspiration Consolidated
Copper Company
Kennecott Corp.
Newmont Mining
Phelps Dodge Corp.
1978 rated3
capacity
(Mg)
180,000
91,000
182,000
91,000
13,600
52,000
136,000
45,000
254,000
71,000,
73.0001
181,000
115,000
64,000
191,000
163,000
1,722,600°
1,693,000°
1979
Production
(Mg)
138,000C
61,000f
96,000
85,000
12.9009
39.7701
124, 070 J
If
296, 000 K
56,360
145, 9001"
283,000"
91,000
1,429,000
(83. OX)
1,399,100
(74.7%)
1980
Production
(Mg)
d
42,700f
59,000
47,300
10.0009
32, 500 1
107, 400 J
tf
259,300*
46,100
98,000m
n
324,100n
1,026,420
(59.6%)
Material arrangements (%)
Integrated -
primarily
Custom
Integrated -
Custom
Toll
Integrated*1 -
Integrated -
Integrated -
Toll
Integrated -
Toll
Integrated -
Toll
Integrated -
Custom
Toll
Integrated -
Custom
Toll
1976
74
26
31
21
48
100
100
35
65
100
81
19
79
5
16
72
8
20
1979
-
-
31
25
44
100
100
35
65
100
100
-
-
-
1980
~
-
20
44
36
100
100
58
42
90
10
100
75
25
73
6
21
Source: 114 letter responses (see Table 3-1).
Information primarily from corporate 10-K reports to the Securities and Exchange Commission.
cReference 2.
Anaconda smelter closed in 1980.
Estimate based on U.S. Bureau of Census data and Bureau of Mines data, which show concentrates coming from Canada and going
to Anaconda but very little refined copper going back to Canada.
Reference 3.
^Reference 4.
Estimate based on total copper sales for Cities Service minus the sales of its Arizona mines.
Reference 5.
•^Reference 6.
Reference 7.
Estimated to expand to 110,000 tons.
mReference 8.
"Reference 9.
°Rated capacity excluding Anaconda.
9-5
-------
based on CDA figures. CDA figures will be used throughout the report
for comparison to historical data. The total production figure for
1980 was compiled from corporate annual reports. For 1980, the table
shows that the industry operated at 59.6 percent of capacity. Produc-
tion was down for 1980 due to an industry strike.
Data in Table 9-1 indicate that the vast majority (approximately
89 percent) of smelting capacity is in Utah, Nevada, New Mexico,
Arizona, and Texas, near copper mines. The location is largely dic-
tated by the need to minimize shipping distances of concentrates,
which are normally 25 to 35 percent copper.
The three largest companies account for 68 percent of the entire
smelting capacity (1,273/1,873 = 68, prior to Anaconda closure),
69 percent of 1979 production, and 76 percent of 1980 production.
Kennecott Corporation has the largest smelting capacity, followed by
Phelps Dodge and then ASARCO. The remaining four companies each have
one smelter and in order of size are Magma (Newmont), Inspiration,
Copper Range, and Copperhill (Cities Service).
The table also shows that 72 percent of total 1976 smelter
production was from concentrate outputs from integrated arrangements.
Of the remaining concentrate, 20 percent was smelted on a toll basis
and 8 percent smelted on a custom basis. Three of the eight companies
process only their own copper concentrates.
The 1979 smelter production rate was 1,399,100 Mg of copper, down
15.6 percent from the 1973 high of 1,656,200 Mg. From 1975 through
1979, the average annual production was 1,379,100 Mg; the average for
the previous 5 years was 1,548.200 Mg, a decrease in average annual
production for the two periods of 12.2 percent.
In 1979, 60,400 Mg of copper scrap were resmelted by primary
copper smelters, amounting to 4.3 percent of total smelter production.
The average percent for the years 1975 to 1979 was 3.6 percent and for
the previous 5 years it was 4.4 percent.10 Although it appears that
the percentage trend in scrap use at the smelting stage is decreasing,
in recent years it has returned to previous levels.
9-6
-------
Foreign input of ore to U.S. smelters in 1979 was 1.6 percent of
smelter output. The 1975 to 1979 average was 3 percent; for the
5 previous years the percentage was 2.1 percent.11
9.1.3 The Copper Refiners
Following the smelting stage, the 98-percent-pure copper resulting
from the smelting operation is refined by either electrolytic or
firing processes to a product of greater than 99 percent purity.
Historically, refiners were much closer to the market than were
smelters. New Jersey and New York once had extensive refining
facilities. Today refiners are found throughout the country as can
be seen from Table 9-2.
Table 9-2 lists the refiners, their 1978 production capacities,
and 1978, 1979, and 1980 production of refined copper. Additional
scrap and blister and refined imports are important sources of total
refined copper supply. In 1979, scrap comprised 11.7 percent of refinery
output (excluding scrap that went directly to the smelter). Scrap
contributions to refined output in the 1975 to 1979 period are approxi-
mately 0.6 percent less than the average of the previous 5 years.18
In recent years, the contribution of refined imports to the total
supply of copper has increased. Imports of blister and refined copper
during 1979 totaled 273,700 Mg or 12.2 percent of refined copper
consumed. The average percentage for the 1975 to 1979 period was
17.9 percent and for the previous 5 years the percentage was 17.3.11
There has been a significant change in the amount of imports of
refined copper as compared to blister. For the 1970 to 1974 period,
refined imports accounted for only 56.6 percent of the total
blister and refined imports, whereas in the 1975 to 1979 period,
refined imports increased to 82.4 percent of the total blister and
refined imports.19 This trend toward importing copper in a more
processed form means that there is less demand for work at domestic
facilities. The U.S. Copper Industry is concerned about the consider-
able increase in the net imports of fabricated copper.
9-7
-------
TABLE 9-2. U.S REFINING FACILITIES FOR PRIMARY COPPER3
(megagrams)
Name and location
Amax-Carteret, NJ
Anaconda-Great Falls, MT
ASARCO-Tacoma, WAh
ASARCO-Amarillo, TX
Kennecott-Magna, UT
Kennecott-Anne Arundel
City, MD
Kennecott-Hurley, NM
Phelps Dodge-El Paso, TX
Phelps Dodge-Laurel Hill,
Magma-San Manuel, AZ
Copper Range-White
•\ • mtT"'
Pine, MI
Inspiration- Inspiration,
AZ
Southwi re-Carroll ton, GA
1978 .
Capacity
236,000
228,000
142,000
382,000
169,000
251,000
94, 000 1
382,000.
23.0001
NY 65,000
18,000
182,000
82.0001
64,000
65,000
1978
Production
180,325c>d
129, 8319
58,000
246,000
154,625
80,534
54,138
461,727J
146,000k
36, 950 ]
35,617n
22,000°
1979
Production
I71,436c>e
138,293s
1,000
300,000
187,210
108,950
56,362
490, 000 J
(34% ca-
pacity)
145,900k
39.7701
43,161n
12,900°
1980
Production
116,310c'e
204,000
181,900
77,400
46 , 100
421,800J
98,000k
32.5201
62,850n
10,000°
Total 2,383,000
Copper Development
Association Total
1,606,147 1,594,982 1,250,880
1,910,400 2,064,600
aA few refineries exist that process only secondary, or scrap, materials.
bReference 12.
clncludes copper from primary and secondary smelting facilities.
Reference 13.
Reference 14.
Anaconda refinery closed in 1980.
^Reference 15.
hASARCO-Tacoma refinery closed in 1980.
Yake or fire refining, otherwise all other facilities electrorefining.
•'Reference 16.
kReference 17.
Reference 5.
mln 1980 Copper Range announced plans to build a 54,530-Mg capacity
electrolytic refinery at White Pine, MI, to be completed in late 1982.
"Reference 6.
°0nly that portion of total company's production which came from Copper-
Hill, Tennessee, smelter facility.
9-8
-------
The 1979 value of shipments from refiners was $5.7 billion. This
is an increase over the 1976 level, which was $3.4 billion.
9.1.4 Domestic Supply
The refined copper output described above does not depict the
entire supply of copper that is consumed in the United States. A
large portion of copper scrap does not need to be resmelted or
re-refined and is readily available for consumption. Copper is a
durable material and if it is unalloyed or unpainted it can be reused
readily. Otherwise it is resmelted or re-refined as described earlier.
The ready availability of scrap as a secondary source of copper tends
to be a stabilizing influence on producers' copper prices.
Total copper consumption in any one year is therefore comprised
of refined U.S. production, scrap not re-refined, net imports, and any
changes in inventory of primary refined production from one year to
the next. The sources and uses of copper are shown in Figure 9-2.
The refined copper production in 1979 comprised 62.7 percent of
total copper consumed in the United States; scrap not re-refined
accounted for 31.0 percent and net refined imports 6.8 percent.19
Between 1969 and 1978, 67 percent of U.S. copper demand, excluding
stock changes, was met from domestic mine production; 21 percent was
from old scrap; and 12 percent from net imports. During these years,
total U.S. demand for copper averaged 2,012,000 Mg/yr. Of this amount,
1,337,000 Mg was from domestic production, 427,000 Mg from scrap, and
248,000 Mg from net imports.
Another method of describing the importance of scrap is to total
the three stages (smelting, refining from scrap, and reuse of scrap)
at which scrap can enter the production process and to compare the
figures to total copper consumption. In 1979 the percentage of total
consumed copper from scrap was 48.1, slightly higher than in more
recent years.
The 1979 refined copper production level was 2,064,600 Mg.
However, there is a slight increasing trend in refined production;
production has increased at a 1.1-percent annual rate for the last
5 years compared to the previous 5, which showed a decline of 1.0
percent annually.
9-9
-------
*
v>
E
o
o»
a
USES%
-I DOMESTIC MINES ^^^.^-^-^<^^^-^^
-• — —• — — — —————— -^*.j~T—T-T"^"'!^*.-^- ^~i r~ *™ — — —- —• — — — v- .— ^.^—.^^...^„ _rt.
" * """ ~" ~" "~~ — — — -^ —• — ~* — « »_ » ^ _ . . ^ '^r\iT"^~* "~r " "" "~ ^ *~ ~* ~~ — — — — — « ..
-j^r^TL~*^~^ ~7-T~L~* *" *~_ *^-T*v*^-T" ™~ ™ ""•"'— — — — — — -^^—™--^^--^™-r^-^—™i^z_r^«r^—"Lr^rt.j^.^'
-^~^^T-~~J_~'| ,7^ ~~r~ ~ '— — — *- — — — — ™ _r^. »J^-.~T.!~i-r^_i~Lr^-_r"r_~^_~r^'~ _^TTJ^_~^,*T. *~^^^*~ —~ ** ~* —
I— —r^-T"—r*—r*^rt-jrT_r^-j~!_r",^~^~L ~^_r^-^r-T"-.^~_r" — — — — •'— — — _ «. •J^r^j"_r^_~i_r^_*~^_r^^~^*~[ ~ r~^~
1950 1955 I960 1965 1970 1975 '78 CURRENT
8
19
58
OTHER
TRANSPORTATI
INDUSTRIAL
MACHINERY
CONSTRUCTION
ELECTRICAL
*Teragram«l.I million «horr ton*
BUREAU OF MINES, U.S. DEPARTMENT OF THE INTERIO*
(fmporttxport data from Bureau of the Cen»us)
Figure 9-2. U.S. sources and uses of copper.20 21
9-10
-------
9.1.5 Flow of Copper from Mines to U.S. Smelters
In subsequent analyses of how copper smelters are likely to
handle control costs, it is important to know the sources of copper
concentrates. Passing the cost of controls back to the mines and
mills is one option that will be considered. It is currently a more
likely option than passing the cost forward to the consumers, given
the competition from substitute products and imports and depressed
demand. Another option to be considered is absorption of control
costs by the smelters.
Tables 9-3 and 9-4 identify the estimated flows of copper from
the 35 domestic copper mines to the 16 domestic smelters in existence
in 1976 (before the Anaconda closure). The construction of such a
table involves considerable investigation and requires making many
assumptions since many companies do not normally reveal such informa-
tion in their corporate 10-K or annual reports. The tables are for a
specific year--1976. Some relationships can change from year to year,
especially for toll and custom smelting; however, although the figures
are estimates for a single year, they should provide a reasonable
indication of the flows within the industry.
Considerable information was obtained from corporate annual
reports and 10-K reports to the Securities and Exchange Commission.
Information also was obtained from the American Bureau of Metal Statis-
tics and from discussions with industry analysts and trade publications.
Import and export figures for ore and concentrates were also obtained
for reconciliation with national production figures and as inputs or
outputs to specific mines or smelters. Mine output going to hydrometal-
lurgical facilities and not to smelters was also identified in the
figures. There are limits to this type of analysis: the effects of
changes in mine and smelter inventory are not always apparent, and
information on intracompany flows often must be estimated. For example,
Duval reports exactly what percentage of total mine output goes to
ASARCO but not to which one of ASARCO's three smelters.
As an overall description, brief summaries of the sources of
copper are provided for each of the eight smelting companies. During
9-11
-------
TABLE 9-3. FLOW OF COPPER FROM MINES TO U.S. SMELTERS, MINE OUTPUT
(1976, gigagrams of copper)
Mines
Parent company,
mine identifying no. ,
mine location
Phelps-Dodqe
1 Bisbee, AZ
2 Metcalf, AZ
3 Morenci , AZ
4 Ajo, AZ
5 Tyrone, NM
Newmont
~~o San Manuel , AZ
7 Superior, AZ
Kennecott
~S" Ray , AZ
9 Chino, NM
10 Utah
11 Nevada
Anaconda
12 Berkeley, MT
13 Yerington, NV
14 Victoria, NV
Inspiration
15 Inspiration, AZ
16 Ox Hide, AZ
17 Christmas, AZ
Cyprus
18 Pi ma, AZ
19 Bagdad, AZ
20 Johnson, AZ
21 Bruce, AZ
ASARCO
22 San Xavier, AZ
23 Mission, AZ
24 Silver Bell, AZ
25 Sacaton, AZ
Cities Service
26 Pinto Valley, AZ
27 Copperhill , TN
Anamax
28 Twin Buttes, AZ
Copper Ranae
29 White Pine, MI
1976
Capacity
5
55
109
45
86
132
36
86
54
205
36
91
32
9
50
5
9
73
11
5
3
10
41
23
20
59
18
109
68
1976
Production
4
72
95
45
84
99
36
80
52
172
11
74
26
5
34
4
6
55
16
5
3
10
32
20
20
65
16
87
42
Smelter
Identifying
letter0
A
C
C
B
A
D
B
I
I
G
H
F
E
J
J
J
K
-
K
A
I
A
-
A
0
N
N
N
0
N
K
P
N
K
A
I
J
L
destinations Other destinations
Gg
4
72
95
45
33
35
16
99
36
57
52
172
11
67
26
5
34
-
6
42
14
10
-
3
4
6
32
16
4
20
63
16
8
13
16
9
15
42
Identifying
Source code
c
d
d
d
d
d
c,d
e
e
f HM
f
f
f
g HM
g
g
i
HM
i
g,j,k
g>k
g,k HM
HM
g,k
c, 1
c,l
c,l
c,l
c, 1
c,l
g HM
m
n HM
n
c
c
h
P
Gg Source
~
-
-
-
~
— —
-
-
"
23 f
-
~
~ ~
6 h
~ "~
"
~ ""
4 i
~ ~
-
-
6 k
5 k
"" ~
-
-
-
~
-
-
3 m
"
26 o
-
-
-
-
See footnotes at end of table.
(continued)
9-12
-------
TABLE 9-3 (continued)
Mines
Parent company,
mine identifying no. ,
mine location
Ouval
TOattle Mtn. , NV
31 Mineral Pk. , AZ
32 Esperanza, AZ
33 Sierrita, AZ
UV Industries
34 New Mexico
Hecla-El Paso
35 Lakeshore, Ai
Subtotals
Imports (IMP)
Total to Smelters
1976
Capacity
14
13
18
91
23
50
1,700
1976
Production
14
14
15
89
21
13
1,436
Smelter
Identifying
letter"
M
M
N
0
M
0
I
J
0
N
destinations
Gg
14
7
6
15
26
34
8
9
12
4
1.333
65
IJjg
Source
q
q
q
q
q
q
c
c,s
c,s
c,t
Other destinations
Identifying
code Gg Source
_ -
-
.
HM 15 r
INV 5 r
_ -
.
~ — —
HH 9 t
HM=102
HM = Hydrometallurgy.
INV = Inventory.
1 gigagram =1.1 thousands short tons.
bSee first column of Table 9-4.
GEstimate of destination and/or production split and/or amounts.
"Werence 22.
eReference 23.
Reference 24.
Reference 25.
hEstimate based on 1976 Official Statement of Anaconda $31.9 million Pollution Control Revenue Bond.
Reference 26.
•^Reference 27.
Reference 28.
Reference 29.
"Reference 30.
"Assuming other mine inputs and total smelter productions are correct for smelters K and N, the amount to
each of these smelters from this mine is estimated by difference.
°0utput will go to Phelps-Oodge's Hidalgo smelter in the future.
pReference 31.
qBased on ASARCO-supplied figures for 1971-1974.
r!6 percent of production stayed within the company, 5 percent to inventory, and 11 percent to new CLEAR
process.
^Reference 32.
Information from company 10-K report.
9-13
-------
TABLE 9-4.
FLOW OF COPPER FROM MINES TO U.S. SMELTERS, SMELTER SOURCES
(1976, gigagrams)
Smelters
Parent company,
smelter identifying letter,
mine location
Phe Ips -Dodge
"A Douglas, AZ
B A jo, AZ
C Morenci , AZ
D Hidalgo, NM
Kennecott
E McGill , NV
F Garfield, UT
G Hayden, AZ
H Hurley, NH
Newmont
~I 5an Manuel , AZ
Anaconda
~3 Anaconda, MT
Inspiration
K Miami, AZ
Copper Range
L White Pine, MI
ASARCO
M Tacoma, WA
Copper sources
1976
Capacity
115
64
161
91
45
254
73
73
182
180
136
82
91
1976
Produc-
tion
115
62
167
35
11
172
57
52
166
165
115
42
75
Identi-
fying,
number
SCR
1
18
19
21
28
5
4
3
3
2
5
11
10
8
9
28
6
7
33
18
IMP
12
13
14
28
34
15
17
26
28
29
SCR
IMP
30
31
33
Gq
8
4
42
10
3
16
33
45
16
95
72
35
11
172
57
52
9
99
36
8
14
43
67
26
5
15
9
34
6
63
13
42
6
22
14
7
26
Source
b
b
c,d
c,e
c
b
b,f
f
b
f
f
f
g
g
g
g
b
h
h
b
c
i
c
c
c
b,j
b,j
k
k
c,k
k,l
m
n
i
0
0
0
(continued)
9-1/1
-------
TABLE 9-4 (continued)
Smelters
Parent company,
smelter identifying letter,
mine location
N Hayden, AZ
*
0 El Paso, TX
Cities Service
P Copperhl 1 1 , TN
Total smelted
Copper sources
1976
Capaci ty
164
105
20
1,836
1976 Identi-
Produc- fying
tion number
93 22
23
24
25
28
31
35
75 22
24
32
33
34
SCR
- f i\^
16 27
im
Gg
6
32
16
20
8
6
4
4
4
15
35
12
7
•* £
16
Source
b.P
b,P
b.P
b.P
1
0
b.p
b.p
b.p
0
0
b.P
n
q
SCR = Scrap.
IMP = Import.
aSee first column of Table 9-3.
bEstimate of destination and/or production split and/or amounts.
Reference 25.
dReference 27.
e
!0utput will go to Phelps-Oodge's Hidalgo smelter in the future.
Reference 22.
^Reference 24.
Reference 23.
Bureau of Mines data.
JInformation from company 10-K report.
Reference 26.
Assuming other mine inputs and total smelter productions are correct for
smelters K and N, the amount to each of these smelters from this mine is
estimated by difference.
""Reference 31.
"Estimate equal to same proportion as 1971-74 period for which ASARCO-
supplied data available.
°Based on ASARCO-supplied figures for 1971-1974.
pReference 29.
qReference 30.
9-15
-------
1976, Anaconda's Montana smelter received most of its copper concen-
trates from its own mines in Nevada and Montana, but it also received
large imports from Canada according to Bureau of Mines figures.
Anaconda also may have received small amounts of copper concentrates
from UV Industries of New Mexico in 1976.
During 1976 (and in previous years), ASARCO was the largest toll
and custom smelting company. Thirty-one percent of ASARCO's blister
copper output was from its own mines; forty-eight percent was toll
smelted; and twenty-one percent custom smelted. Duval Corporation is
the largest supplier of concentrates to ASARCO. The Duval mines sent
77 percent of their 1976 output to ASARCO's smelters. ASARCO also
received imports at its Tacoma smelter; the amount was a significant
portion (approximately 30 percent) of its input.
During 1976 the Copper Range Company's White Pine smelter received
input from its own mines in Michigan.
The Copperhill, TN, smelter of the Cities Service Company received
its concentrates from its own nearby mines during 1976. More recently,
during May of 1981, Cities Service announced that it is "testing the
market" for the possible sale of its copper operations.
The Inspiration, AZ, smelter, received an estimated 35 percent of
its input from its own mines during 1976. Its largest external source
of copper was Cities Services' Pinto Valley, AZ, mine. It is believed
that Inspiration also received copper concentrates on a tolling basis
from Anamax.
Kennecott Corporation obtained concentrates for its four smelters
from its own mines during 1976. More recently, during December of
1980, Mitsubishi Corporation of Japan became a one-third partner with
Kennecott at Kennecott's Chino operation.
The Magma (Newmont) smelter primarily received copper concentrates
from its own mines during 1976, accounting for 80 percent of its
input, while approximately 20 percent was received from Cyprus and
others on a tolling basis.
Phelps-Dodge has three smelters that received copper concentrates
from its own mines during 1976. The Douglas smelter was estimated to
9-16
-------
have received over one-half of its copper input from external sources,
primarily Cyprus. On a total company basis, Phelps-Dodge received
79 percent of its smelter input from its own mines.
It should be noted that the Anaconda smelter and refinery were
closed in 1980. Anaconda copper concentrates are currently sent to
Japan for smelting and refining. The Japanese consortium that treats
the Anaconda concentrates has two options—to purchase the concentrates
or to treat them under a tolling arrangement and return them to Anaconda.
ASARCO closed its Tacoma, WA, copper refinery in late 1980. The
company's Amarillo, TX, refinery currently is sufficient to handle
ASARCO's smelter output.
9.1.6 Copper Production Costs
Copper production costs vary for a number of reasons including
location and physical characteristics of ore deposits. However, the
Bureau of Mines has estimated the costs for a representative large
open pit copper operation. In terms of the price of copper, it is
estimated that 30 percent is for mining; 20 percent is for ore benefi-
ciation; and 20 percent is for freight, smelting, and refining. The
balance of 30 percent is required for such items as discovery,
development, taxes, marketing, and general overhead including profit.
The copper industry is capital-intensive requiring over $7,000/Mg of
new annual capacity for facilities from mining through refining.
According to the Bureau of Mines, the cost for expanding an existing
facility is $5,000/Mg of expanded annual capacity.33
One major reason for increasing production costs is the long-term
declining yield of copper from copper ores. In the United States, the
average yield has dropped from 8.2 kg of copper per megagram of ore in
1950 to 4.5 kg in 1977.33 Currently, some of the copper deposits
under development contain an average of only 3.6 kg of copper per
megagram of ore with a cutoff grade of 1.8 kg.
Because of the low grade of copper ore, larger amounts of material
must be mined and processed to produce a given quantity of metal.
Moreover, the Bureau of Mines estimates that open pit mines, which
account for about 82 percent of domestic output, have average ratios
of overburden to ore of about 2.5 to I.33
9-17
-------
None of the annual reports or 10-K reports of the involved
companies presents pricing or cost schedules for their smelting
operations. Therefore it was necessary to go to indirect sources for
estimates. Fortunately, several researchers have attempted cost
estimates on an industry-wide basis, as shown in Figure 9-3 and Table
9-5.
The Bureau of Mines has provided one such estimate. The costs
were estimated for a 100,000-ton/yr smelter (copper content) built in
1973. Such costs were updated by indices recently provided by the
Bureau of Mines. The costs for October 1977 are 28.2
-------
240
230
220
O
O
n 210
in
en
X
UJ
O
UJ
O
200.
190
ISO
o:
UJ
a- 170
D.
O
O
Q
2
160
£ '50
UJ
CO 140
Ul
130
o
I2O
no
too
Copper
price
index
Mine and mill
capital cost index
till
1355 1966
1968
1970
'1972
1974
1976
1978
Figure 9-3. Comparison of copper price index and mine and mill
capital cost index.34
9-19
-------
TABLE 9-5. SMELTING COST ESTMATES'
Source
New plant
Bureau of Mines
Arthur D. Little
Green feed
Calcine
Existing plant
Arthur D. Little
Cleaver
Cost
-------
Arthur D. Little also estimated industry-wide production costs
for refined copper of $1.59/kg consisting of 94.84: for variable costs
and 63.9$ for fixed costs.39
Although the above two studies were performed on an industry-wide
basis, George Cleaver also estimated copper costs on a per-kilogram
basis for each major mine through the refining stage (Table 9-6). The
costs include the fixed and variable portion and include pollution
control by-product value and ore for 1976. The range in cost is from
$1.21 to $2.43/kg of refined copper with a median of $1.82.
In a more recent study, Leon Kovisars of MET Research estimates
the production cost for 1979, also through the refining stage. He
claims that two-thirds of the world copper produced in 1979 incurred
cash costs between $0.77 and $1.96/kg, averaging $1.37/kg. Costs in
North America (United States and Canada) are higher, with total cash
costs averaging $1.50/kg. The total cash cost ranges from $1.94/kg
to $1.06/kg for two-thirds of the North American producers. He also
estimates that the average variable cash cost in North America is
$1.01/kg and two-thirds of the production incurs cash cost between
$1.39/kg and $0.64/kg.41
9.1.7 U.S. Copper Resources
Various estimates of U.S. copper resources (identified deposits
that can be extracted profitably at a given price) show amounts ranging
from 6.18 to 99.1 teragrams (Tg).*
One U.S. Bureau of Mines report showed resources of 104.7 Tg of
copper in 1973 at a price of $1.65/kg.42 Meanwhile, the Bureau of
Mines Copper Commodity Profile, which reports resources on the basis
of corporate reports, currently uses a figure of 84.5 Tg. This figure
includes mines under operation and development as well as undeveloped
deposits. Of the 84 Tg, 60 are under operation and development and
are listed by company in Table 9-7 and updated to 91 Tg for 1979.
The lower figure of 60 Tg referenced above is from a 1973 Geolog-
ical Survey paper. All figures are subject to change based on price
changes in copper relative to general price changes.
*1 teragram =1.1 million short tons.
9-21
-------
TABLE 9-6. U.S. COPPER PRODUCTION BY MINE (1977), CENTS PER KILOGRAM
AND PRODUCTION CAPACITY40
Company
Kennecott
Amax-Anaconda
Kennecott
ASARCO
Hecla-El Paso
La. Land (Copper Range)3
Cyprus Mine
Atlantic Rich. (Anaconda)
Kennecott
Phelps Dodge
ASARCO
Atlantic Rich. (Anaconda)
Cities Service
Inspiration
Phelps Dodge
Cyprus Mines
Duval
Kennecott
Newmont
UV Industries
Phelps Dodge
Amax-Anaconda
Phelps Dodge
Mine
Nevada
Twin Buttes sulphide
Ray
San Xavier
Lakeshore
White Pine3
Bagdad0
Pi ma
Berkeley
Yerington
Inspiration
Christmas
Chi no
Metcalf
Mission
Silver Bell
Sacatoon
Victoria
Pinto Valley
Copperhill
Ox Hide
A jo
Tyrone
Johnson
Mineral Park
Esperanza
Battle Mountain
Sierrita
Utah
Magma
New Mexico
Morenci
Twin Buttes (leach)
Bisbee (leach)
Gg/yr
capacity
36
82
84
10
50
68
64
72
(27%) 456"
91
32
50
5
55
55
41
23
19
9
59
18
5
45
84
(34%) WL
5
18
18
14
91
205
164
23
(31%) 538
109
27
5
( 8%) 144
(100%) It739
-------
TABLE 9-7. COPPER RESOURCES OF U.S. COMPANIES34
(gigagrams)
Resources
Company
Anamax
ASARCO
ASARCO and Anamax (exclusive of the above)
Anaconda Company
Cities Service Company
Copper Range (Louisiana Land Exploration Co.)
Cyprus Mines Corporation (Standard Oil of Indiana)
Duval (Pennzoil)
Hecla Mining--El Paso Natural Gasa
Inspiration Consolidated Copper Co.
Kennecott Corporation
Magma (Newmont)
Phelps Dodge
Ranchers Exploration and Development Co.
UV Industries
Other
Total
1976
2,848
1,904
648
3,685
1,763
6,880
2,460
1,834
3,181
1,270
14,959
6,772
10,906
340
481
240
60,171
1979
4,329
3,146
894
-
2,659
4,694
1,674
1,555
-
2,466
19,438
6,929
13,594
289
492
-
91.165
aProperty now owned by Noranda Mines, Ltd.
9-23
-------
Copper resources that are as yet undiscovered are referred to as
hypothetical or speculative. The Bureau of Mines estimates 290 Tg of
such copper resources.
A discussion of the capability of these two sets of copper resources
to meet future demand is dependent upon several factors. The Bureau
of Mines estimates that copper demand will grow at an annual growth
rate of 3.6 percent to the year 2000 and that 31 percent of the demand
will be supplied by scrap. The likely primary copper demand over this
period would be 57 Tg compared with 84 Tg of resources.43 Therefore,
U.S. supply appears adequate to the year 2000. Beyond the year 2000,
demand is expected to strain supply sources. But, increased use of
old scrap and possible exploitation of sea nodules can supplement
onshore mining. In addition, microminiaturization, copper cladding,
and other conservation methods will be used more widely to extend the
supply of copper.
9.1.8 Smelter Capacity Growth
Smelter production in the United States reached its historical
peak of 1,582.1 Gg in 1973. Figure 9-4 shows that since reaching that
peak, production levels in the subsequent 8 years were considerably
below the 1973 level and showed no signs of returning to the historical
high. The lack of long-term growth in smelter production over recent
years serves as a historical indicator to suggest that on this basis
demand is not likely to grow over the next 6 or 7 years by an amount
that would require new capacity.
Neglecting the low production due to the 1980 strike and the 1982
recession, the previous 5 typical years averaged 1,380 Gg/yr. Current
capacity is 1,723 Gg. Possible shutdowns over the next 6 years amount
to 224 Gg. This potential lost capacity will be countered partially
by the announced expansion of 27 Gg at the Hurley smelter of Kennecott.
The net effect of these changes on capacity would be a drop to 1,526 Gg
by 1988. If production stays level, this represents a rise in utiliza-
tion to a 90-percent rate.
Should there be an unexpected upward pressure on production, two
corrective measures are available to ease a capacity shortage. First,
9-24
-------
Copper
Smelter
Production
(gigagrams)
1800-
1700-
1600-
1582.1
1500-
1400-
1300-
1423.9
1392.0
1335.4
1312.5
1301.7
1288.1
1200-
1100-
1000-
1973 1974 1975 1976~~ 1977 I978~
Figure 9-4. U.S. copper smelter production.44
1979
9-25
-------
many of the present smelters can be expanded 20 percent by increasing
the use of oxygen enrichment. This has the potential of adding approxi-
mately 200 Gg of capacity. Second, the presumed shutdown of the
McGill and Douglas smelters in 1988 may not occur. Also, the potential
supply provided by imports is assumed not to increase.
It is therefore apparent that current smelter capacity will be
adequate to meet demands over the next 6 to 7 years without the construc-
tion of facilities subject to a new source performance standard (NSPS).
Matters of utmost concern to the industry's smelter growth are
air pollution regulations affecting the location of new smelting
capacity. Any location for new sources will require stringent air
pollution controls. Smelter growth in the past 15 years in the copper
industry has occurred primarily by additions to existing facilities.
9.1.9 Trends in U.S. Productivity
The productivity movements from 1971 to 1977 are shown in Table
9-8, which was completed by the Bureau of Labor Statistics. The trend
has been a modest improvement. It took 60.2 employee hours to produce
a megagram of refined copper in 1977 and 67.6 hours in 1971. This
amounts to a 1.9-percent per annum gain. However, this figure is
below that for all U.S. industry.
Commodities Research Limited compiled a table (Table 9-9) of
output and productivity indices to examine the trend more carefully.
They feel it may be misleading to call this a trend because of the
setback to productivity in the middle of the period.45 Productivity
decreased in 1973 as a result of a drop in the mining and milling
sector. Productivity also fell in 1974 and 1975 as a result of decreases
in both the mining and milling and the smelting and refining sectors.
After the recovery in 1976, there was little net change in 1977 although
a gain occurred in smelting and refining productivity and a setback
occurred in mining and milling productivity. The gain in smelter
productivity occurred despite their suffering from the impact of
pollution controls that limited permissible throughput.
Table 9-8 shows that the number of employees and the total number
of hours worked increased significantly in 1973, Despite this being
9-26
-------
TABLE 9-8. PRODUCTIVITY IN THE COPPER INDUSTRY45
Mine-mill
Production workers
Year
Mine
Mill
Office
workers
Total
Smelter- ref inerv
Production
workers
~ • ^
Total
Grand
total
EMPLOYMENT (average number)
1971
1972
1973
1974
1975
1976
1977
EMPLOYEE
1971
1972
1973
1974
1975
1976
1977
16,800
16,500
22,000
23,400
21,500
18,700
17,600
HOURS
41,300
42,900
46,800
42,600
40,900
35,800
33,200
6,400
6,300
9,500
9,700
8,600
8,200
8,400
('000)
16,600
16,900
18,400
19,300
15,200
17,400
15,700
2,000
2,200
3,500
3,200
3,400
2,800
3,400
4,600
4,900
6,700
6,000
5,900
5,000
6,000
25,200
25,000
35,000
36,300
33,500
29,700
29,400
62,500
64,700
71,900
67,900
62,000
58,200
54,900
PRODUCTIVITY (employee hours per metric ton of
1971
1972
1973
1974
1975
1976
1977
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
45.3
42.8
46.1
46.8
48.3
40.0
40.2
12,700
14,200
14,300
14,300
12,900
11,400
10,100
28,133
30,939
31,529
31,008
27,100
24,600
22,741
16,000
17,600
17,800
18,100
16,500
14,600
13,100
35,443
38,347
39,245
39,248
34,663
31,583
29,496
40,200
42,600
52,800
54,400
50,000
44,300
42,500
97,943
103,047
111,145
107,148
96,663
89,783
84,396
copper production)
-
-
-
-
-
-
-
22.4
21.1
21.1
23.2
23.8
20.4
20.0
67.6
63.9
67.2
70.0
72.1
60.3
60.2
9-27
-------
TABLE 9-9. OUTPUT AND PRODUCTIVITY INDICES45
Year
1971
1972
1973
1974
1975
1976
1977
Mine
output
100.0
109.4
112.9
103.0
92.8
105.5
98.6
Mine-mill
100.0
105.3
98.1
96.6
93.2
116.7
111.9
Productivity
Smelter- refinery
100.0
105.4
105.4
96.1
93.6
108.9
110.3
Overall
100.0
105.4
100.5
96.4
93.3
110.7
110.9
Productivity indices are derived from Table 8-14.
9-28
-------
the peak production year (the industry turned out 1,559 Gg of primary
refined copper), the rise in hours worked was more than the rise in
output. In 1974 the labor force was still rising, output declined,
and, despite a fall in hours worked, productivity dropped again. In
1976 the copper industry cut back a large amount of labor and reduced
hours worked while increasing output. Commodities Research argues
that the resulting sharp increase in productivity was probably due to
the closure of labor-intensive capacity.
9.1.10 U.S. Total Consumption of Copper
Total copper consumed in the United States over the last 10 years
has fluctuated considerably but shows overall a steady upward trend.
This conclusion is derived from data of copper consumption from refin-
eries and copper consumption from refineries plus scrap.
Table 9-10 shows each set of data for the year 1970 through 1980.
The 5-year averages in gigagrams for copper consumption from refineries
has declined (1970 through 1974 is 1,992.1 and 1975 through 1979 is
1,926.6.) However, copper consumption rose from 1975 to 1979 before
declining in 1980.
Five-year scrap consumption showed a decline from 892.3 Gg for
the 1970 to 1974 period to 838.0 Gg for the 1975 to 1980 period.
There are signs that the consumption of scrap has begun to increase
over the last 4 years.
The Bureau of Mines forecasts a long-range consumption growth
rate to the year 2000 of 3.0 percent per year. Prior to the 10-year
period analyzed above, copper consumption had increased steadily since
the early 1950's.
9.1.11 Demand by End Use. Refined copper and copper scrap are
further processed in a number of intermediate operations before the
copper is consumed in a final product. Refined copper usually consists
of one of the following shapes: cathodes, wire bars, ingots, ingot
bars, cakes, slabs, and billets. These shapes, along with the copper
scrap, then go to brass mills, wire mills, foundries, or powder plants
for subsequent processing. The copper is frequently alloyed and
transformed into other shapes such as sheet, tube, pipe, wire, powder,
9-29
-------
I
OJ
O
TABLE 9-10. U.S. COPPER CONSUMPTION46 4V 48
(gigagrams)
1970
Consumption of refined 1,864.1
copper
Consumption of scrap 813.5
Total consumed copper 2,677.6
Percent of total as scrap 30.4
1970-1974 and 1975-1979
averages for consumption
refined
1970-1974 and 1975-1979
averages for consumption
of scrap
1970-1974 and 1975-1979
averages for total
copper consumption
1971 1972 1973
1,837.3 2,034.2 2,225.7
861.5 948.2 958.6
2,698.8 2,982.4 3,184.3
31.9 31.8 30.0
1,992.1
892.3
2,884.4
1974 1975 1976 1977 1978
1,999.1 1,399.4 1,811.7 1,990.1 2,197.4
880.0 663.9 804.7 848.7 886.7
2,879.1 2,063.3 2,616.4 2,838.8 3,084.1
30.6 32.2 30.8 29.9 28.7
1,926.6
838.0
2,764.6
1979 1980
2,234.6 1,930.0
985.8 940.9
3,220.4 2,870.9
30.6 32.7
aWithout having to be refined again.
-------
and cast shapes. Ultimately, the copper is consumed in such shapes in
five market or end-use categories. The CDA uses the following
categories: building construction, transportation, consumer and
general products, industrial machinery and equipment, and electrical
and electronic products.
Table 9-11 shows the demand for copper in each of these five
markets over the 10-year period 1970 through 1979. The total figures
for these five markets will not equal the total consumption figures of
Section 9.1.9 since the United States is also a net importer of fully
fabricated copper products.
A look at the 5-year average demand shows that there has been an
increase in only two out of the five markets. The building industry
market sales showed the most substantial gain of 21.8 percent. This
is primarily due to the increase in residential sales. An increase of
10.5 percent also occurred in the electrical and electronic product
markets. The demand for electrical equipment has risen because of
increased emphasis on safety, comfort, recreation, and a pollution-free
environment. Automation, including the use of computers has also
boosted the use of copper.
Substitution of other materials has caused a sharp drop of 24 per-
cent in the consumer and general products markets. The 16-percent
decline in the industrial machinery and equipment market in the 1975
to 1979 period is largely due to the impact of the recession in 1975.
Since 1975, this market has returned to a prerecessionary level.
Declining car sales coupled with the decrease in copper content per
automobile resulted in a 10.4-percent decrease in sales in the transpor-
tation sector.
The Bureau of Mines has established that the most growth in
copper demand will occur in the electrical and electronic products
industries, consumer and general products, and building construction.
Copper will be in high demand for electric vehicles. General
Motors plans to produce an electric family car for mass marketing in
the mid-19801s. In addition, Chrysler is developing a four-seat
passenger car jointly with General Electric. A conventional internal
9-31
-------
TABLE 9-11. U.S. COPPER DEMAND BY MARKET END USES40 49
(gigagrams)
Market 1970
Building construction 556
Transportation 284
Consumer and general 586
products
Industrial machinery and 504
equipment
Electrical and electronic 693
products
Total 2,623
5-Year average demand
i
GO - Building construction
ro
Transportation
Consumer and general
products
Industrial machinery and
equipment
Electrical and electronic
1971 1972 1973 1974
624 693 762 602
318 379 418 356
575 652 694 607
493 568 606 530
710 748 868 764
2,720 3,040 3,348 2,859
647
420
623.2
540
601
1975 1976 1977 1978
515 633 893 926
260 379 409 438
512 572 382 442
349 415 461 481
500 610 667 746
2.136 2,609 2,812 3,033
789.1
376.9
473.0
453. S
657.3
1979
977
397
455
560
811
3,200
products
-------
combustion automobile contains from 6.8 to 20.4 kg of refined copper;
electric vehicles use much more. CDA estimates range from 45.4 to
90.7 kg, with an average nearer 45.5 kg.50
Another area for growth is the solar energy industry. Currently
this sector consumes approximately 4,500 Mg/yr of copper in the United
States. Representatives of the U.S. Solar Energy Industry estimate
that the consumption could climb to 31,000 Mg/yr by 1985.
In addition, the U.S. military demand for copper is expected to
increase. Increased military expenditures will have a significant
impact on copper demand because copper is an important element in
modern electronic weaponry. As seen in Table 9-12, during heavy
rearmament periods, the U.S. military demand for the metal has reached
18 percent of copper mill shipments. Although military demand is not
expected to return to the record high of 18 percent, analysts do
expect a large increase in military requirements for copper from the
low level in 1979 of less than 2 percent.51
The demand picture in the United States may also receive another
boost in the near future from the Federal government. The Government
is committed to acquiring, eventually, 1.1 Gg of copper for its currently
depleted strategic stockpile. The previous stockpile was largely
depleted in 1968; the final sale was in 1974 when copper prices had
soared. Further Congressional action is necessary to implement and
fund the purchase plan. Bills have been put forward to sell tin from
the government stockpile to fund the copper purchases. After authoriza-
tion by Congress the purchases may take 12 to 18 months to complete.
9.1.12 Copper Prices
Factors influencing the copper market and thus the price of
refined copper include production costs, long-run return on investment,
demand, scrap availability, imports, substitute materials, inventory
levels, the difference between metal exchange prices and the refined
price, and Federal Government actions (e.g., General Services Administra-
tion stockpiling and domestic price controls).
Among the many published copper price quotations, two key price
levels are: (1) those quoted by the primary domestic copper producers
9-33
-------
TABLE 9-12. U.S. SHIPMENTS OF COPPER-BASE MILL
AND FOUNDRY PRODUCTS-GROSS WEIGHT51
Year
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
Total
Gg
2,239
2,294
1,920
2,331
2,213
2,012
1,837
2,146
1,909
1,996
2,206
2,309
2,602
2,775
3,088
2,639
2,646
2,975
2,563
2,638
2,946
3,267
2,815
2,107
2,505
2,728
2,806
2,868
2,654
Gg
394
345
125
75
52
47
35
33
35
42
53
63
60
63
248
263
240
251
172
96
116
84
61
48
34
33
42
40
-
Military
Percent
17.6
15.0
6.5
3.2
2.4
2.4
2.0
1.5
1.8
2.1
2.4
2.7
2.3
2.3
8.4
10.0
9.2
8.5
6.7
3.6
3.9
2.6
2.2
2.3
1.4
1.2
1.5
1.4
-
9-34
-------
and (2) those on the London Metal Exchange and reported in Metals Week,
Metal Bulletin, and the Engineering and Mining Journal. The producers'
price listed most often is for refined copper wirebar f.o.b. refinery.
The London Metal Exchange (LME) price is also for copper sold as
wirebar. The LME generally is considered a marginal price reflective
of short-term supply-demand conditions while the producer price is
more long-term and stable and often lags the LME price movement.
Copper is also traded on the New York Commodity Exchange (Comex).
Arbitrage keeps the LME price and the Comex price close together (with
minor price differences due to different contract terms on the two
exchanges and a transportation differential).
A significant departure from the two-price system and toward one
world price occurred in mid-1978. At that time both Kennecott and
Anaconda abandoned the system of quoting a fixed price for an indeter-
minate period and began simply charging 2.5<|:/lb more than the closing
futures price for the current month on the New York Commodity Exchange.
In March 1982, Kennecott Corporation announced that, effective July 1,
1982, it would revert to quoting a fixed list price. Industry analysts
believe that Kennecott's use of a Comex-based price means that it
received less for copper than producers that used a firm list price.
Figure 9-5 is a plot of the quarterly movement of these two
prices for the years 1973 through the first quarter of 1981. Data
were obtained from monthly figures from two metals publications, which
are identified in the figure. The next graph, Figure 9-6 shows the
recent trends in more detail.
Three important points can be observed from the figure and from
the recent industry pricing policy of some producers: (1) that the
LME price has had wider swings than the producer price; (2) when both
prices are relatively high, the LME price has been considerably higher
than the producer price; during relatively low price periods, the
producer price has been moderately higher than the LME price; and
(3) marked change appears to be taking place away from a two-price
system and toward a one-price system with the difference between the
9-35
-------
oo
O)
X.
cc
111
a.
CO
H
2:
LU
O
05
u
o
E
a.
280-
270-
260-
250-
240-
230-
220-
210-
200 -
190-
180-
170-
160-
150-
140-
130-
120-
110-
100-
/'.
/
1973
o
London Metal Exchar«ge Prices
U.S. Producers' Prices
.„..! .„! i
i i f
i i i
i i i
! ! !
! ! !
i i i
1 I (
1
1974 1975 1976
1977
YEAR
1978
1979
1980
1981
1
Source: Metals Week
Source: Metal Bulletin
Figure 9-5. Quarterly price movements for copper wirebars (1973 to 1981).52
-------
d-300
280
260
240
220
200
180
160
CENTS/KG
US MERCHANT PRICE
US PRODUCER PRICE (DELIVERED)
LMECASHWIREBARS
AMJ.JA SOND J FMA
Figure 9-6. U.S. copper price
53
9-37
-------
LME and the U.S. producer price accounted for only by a transportation
differential. These earlier situations had occurred repeatedly over
the past 20 years.
In theory, the U.S. producer price should be somewhat higher than
the LME price since ocean transport costs (6.6$ to IKt/kg) must be
incurred to get the refined copper to the United States. However,
this relationship appears to hold only during slack price periods.
When LME prices are high, the producers do not raise their prices as
much, which in theory appears contrary to profit maximization.
There are two principal reasons for this trend. First, U.S.
producers historically were concerned about the trend of substitution
by aluminum, particularly in wire and cable markets, and attempted to
maintain a U.S.-published producer quotation that was both reasonable
and relatively stable in order to meet competition from then lower
priced and stably priced aluminum. Second, government price controls
contributed substantially to holding U.S. producer prices down well
below world market levels, as quoted on the London Metal Exchange
during periods of peak demand. In the mid to late 1960s, price controls
came in the form of "jawboning" by the Johnson Administration. From
August 1971 until April 1974 formal price controls during the Nixon
Administration were responsible for the wide diversions between U.S.
producer prices and LME prices in the 1972-1974 period of strong
demand. Voluntary controls during the Carter years exercised a
restraining effect upon prices and the ability of prices to truly
reflect world market. During the 1970s there really were only two
short periods of good demand and strong prices, which would have been
reflected in profits and retained earnings as well as in investments
for the years of depressed prices: the 1972 to 1974 period and the
1978 to 1981 period, during both of which either formal or voluntary
price controls were in effect.
Considerable data exist to validate the point that the long-run
economic cost of producing copper is increasing. Commodities Research
Unit, Ltd.,54 has analyzed the cost of developing additional capacity
from the mine through the refining stage. Approximately 10 years ago
the capital costs per megagram of annual capacity for developing
9-38
-------
copper were $2,000 to $2,500; they have now risen to $7,200 to $7,700.
An Amax, Inc., official has stated that 1970 costs were $3,500/Mg of
annual capacity and $6,500 in 1976. In was estimated that costs would
rise to $7,500 in 1980. A Kennecott Corporation director has reported
similar cost-per-megagram figures and added that a price of $2.76/kg
to $3.30/kg for refined copper would be needed to support such capital
outlays. Likewise, an ASARCO official claims that a price of $3.30/kg
would be required for new mine developments.
The above costs are for conventional pyrometallurgical smelting.
The newer smelting processes such as Noranda and Mitsubishi offer some
capital cost savings at that stage due to lower pollution control
costs. The hydrometallurgical processes also require less capital.
However, the mining costs are the highest part of overall development
costs for which limited cost saving techniques exist. The mine develop-
ment costs in the United States have risen significantly, largely as a
result of the shift from higher to lower grades of available copper ores
and the sometimes remote locations that require infrastructure costs
for towns, roads, etc.
In 1979, the Bureau of Mines analyzed 73 domestic copper properties
to determine the quantity of copper available from each deposit and
the copper price required to provide each operation with 0 and 15 percent
rates of return. They based the study on a 1978 domestic copper
reserve base of 92 Tg of copper, of which 74 Tg are recoverable using
current technology.55
The Bureau estimates that a copper price of $4.56/kg would be
required if all properties, producing and nonproducing, were to at
least break even. This price increases to $8.40 for the properties to
receive at least a 15-percent rate of return. The average break-even
copper price for properties producing in 1978, $1.46/kg, was about
equivalent to the average selling price for the year. At this price,
analysts calculate that only 25 properties could either produce at
break-even or receive an operating profit. Of these properties, only
12 could receive at least a 15-percent rate of return.
Annual domestic copper production from 1969 to 1978 averaged
1,337,000 Mg. According to this study, to produce at this level and
9-39
-------
to receive at least a 15-percent rate of return, a copper price of
$1.81/kg is required, as seen in Figure 9-7. If the United States
were to produce the additional 248,000 Mg that were imported each year
over this period, a copper price of $1.94 would be necessary.56
The report concludes that increases in copper prices are required
for many domestic deposits to continue to produce. Many U.S. producers
cannot continue to operate at a copper price that has not kept pace
with inflated operating and capital costs. Assuming that copper
demand and other market conditions warrant it, the increases in copper
price that occurred in 1979 will help provide the additional revenue
necessary for many operations to continue producing and should encourage
companies to begin developing other properties.
In fact, the improvement in copper prices during late 1979 brought
back some idled U.S. mine capacity and, more importantly, began to
stimulate the consideration of some major new mine projects. More
recently, depressed copper prices have altered this situation. To
develop a list of possible new U.S. mines that could be brought onstream
in the next decade, Commodities Research surveyed the U.S. producers.
These results are shown in Table 9-13. They estimate that the total
tonnage from projects that stand a reasonable chance of coming onstream
before 1990 is about 792 Gg including 167 Gg from mines that had been
closed down over the last few years but that could be reactivated.
This would be an increase of 53 percent over the 1,523 Gg of capacity
that was in operation at the end of 1978.57
It has been suggested that long-term potential for higher prices
and the high cost of new capacity are significant reasons for the
increased purchases of copper properties by oil companies. The reason-
ing is that oil companies need places for recent heavy cash flows, and
diversification to other products is desirable. The oil companies
reportedly can wait for expected copper price increases to obtain
their return. Further, by purchasing existing facilities rather than
building new capacity, they avoid the escalating new capacity costs.
As shown below, U.S. oil (and gas) companies own or have major
interests in seven of the largest domestic copper producers:
9-40
-------
3.20 -
s r-i :—r~~ i
Potential coppeir that cou?d be? produced at, c, spscl.ic price based
on the 1978 capper reserve? base and 1978 costs; duration of
production not shown
15—percent rate of return
0 —percent roto of return
300 600 900 1.200 1,500 1,800 2.10O 2/JOO 2»7OO 2-.GOO
ANNUAL RECOVERABLE COPf-EH (Jhousaod wtrtric tons)
Figure 9-7. Annual recoverable copper available from domestic deposits
over a copper price range of $1.10 to $1.30/kg.56
9-41
-------
TABLE 9-13. U.S. COPPER MINE CAPACITY: CURRENT AND POTENTIAL57
(gigagrams/year)
Amax
T. Tolman
Anaconda
Berkeley
Yerrington
Carr Fork
Anamax
ASARCO
Mission
Silver Bell
San Xavier
Sacaton
Others
Troy Hills
Casa Grande
Cities Service
Pinto Valley
(leach)
Miami
Miami East
Copper Hill
Continental Materials/Union Miniere
Continental Oil
Copper Range
Cyprus
Bagdad
Johnson
Pi ma
Duval
Sierrita
Mineral Park
Esperanza
Eisenhower
Exxon
Crandon
Pinos Altos
End of 1978
91
112
39
21
10
20
2
65
7
18
54
65
5
91
18
9
Planned and
potential new
23
23
50
18
91
11
5
12
13
91
12
h
54°
h
14°
12C
27
(continued)
9-42
-------
TABLE 9-13 (continued)
Planned and
End of 1978 potential new
Inspiration
Inspiration
Christmas
Ox Hide
Kennecott
Ray
Chi no
Nevada (leach)
Utah
Noranda
Lakeshore
Newmont (Magna)
San Manuel
Superior
Vekol Hills
Phelps Dodge
A jo
Morenci
Tyrone
Metcalf
Safford
Ranchers Exploration
Bluebird
Old Reliable
UV Industries
Other
Total
54
81
59
204
118
39
45
112
86
53
4
21
20
1,523
k
7.
5
45
36
18
27
h
59D
13
29
18
31
54
H
4b
2D
792
Could resume operations for about 3 years before depletion of reserves.
Reactivation.
cMine capacity will be 23.6 Gg, but half will replace Mission (ASARCO).
9-43
-------
1. Amax--Approximately 20 percent owned by Standard Oil of
California
2. Anaconda--0wned by Atlantic Richfield Company
3. Cities Service—Also a primary copper producer
4. Copper Range—Owned by Louisiana Land and Exploration Company
5. Cyprus Pima Mining Company—Standard Oil Company
6. Duval—Owned by Pennzoil Company
7. Kennecott—Standard Oil of Ohio (British Petroleum).
These copper producers own or control a large portion of domestic
copper reserves and mine production and hold a considerable share of
U.S. refinery capacity. Their investment in the copper industry is
significant, and thus they must expect higher prices and substantial
profits in the future.
9.1.13 Substitutes
Substitutes for copper are readily available for most of copper's
end uses. Copper's most competitive substitute is aluminum. Other
competitive materials are stainless steel, zinc, and plastics. Aluminum,
because of its high electrical conductivity, is used extensively as a
copper substitute in high voltage electrical transmission wires.
According to the Bureau of Mines, some 4 percent of insulated power
cable and over 90 percent of bare conductor applicators are currently
provided by aluminum. Aluminum has not been used as extensively in
residential wire because of use problems and minimal savings.
Aluminum is also potentially interchangeable with copper in many
heat exchange applications. For example, automobile companies are
still experimenting with the use of aluminum versus copper in car
radiators. When copper prices are high or copper supply is limited,
cast iron and plastic are used in building construction as a copper
pipe substitute. A relatively new substitute for copper is glass,
which is used in fiber optics in the field of telecommunications.
Fear of long-run substitution for copper is one of the hypotheses
cited earlier to explain why the primary producer price of copper is
lower than the LME price during high demand periods.
9-44
-------
9.1.14 World Production and Consumption of Copper
The United States is the leading copper-consuming country. The
United States is also the leader in refined production but is third to
Africa and the communist block countries in mine production and second
to the communist block countries in smelter production. In 1978 the
United States produced 17.7 percent of the world's mine production of
copper, 17.4 percent of the smelter production, and 20.8 percent of
refinery production. The consumption of the world's refined copper by
the United States amounted to 29.9 percent. Tables 9-14 and 9-15 show
world production and consumption figures.
While the United States is essentially maintaining its consumption
and production levels, world consumption and production are increasing
quite rapidly. As a result, the U.S. share of world consumption and
production shows a relative decrease.
According to the Bureau of Mines, the United States has 18.4 per-
cent of the world's identified copper resources and 26 percent of the
other land-based copper resources.60
In 1979 there was a continuation of the large-scale decline in
free world copper inventories that began in 1978. Free world stocks
declined by 454 Gg in 1979 to a level of 726 Gg. This represents
about 1.25 months' consumption, which is a normal inventory. At their
peak in 1978, these stocks totaled 1,560 Gg. The drawdown occurred
because of steady growth in world consumption from the recessionary
lows of 1974-1976 and the stable level of world copper production.
Consumption of copper in the free world exceeded refined production in
both 1978 and 1979 by approximately 5 percent.61
Domestic demand for copper declined in 1980 by about 15 percent
due to reduced demand in the housing and automobile industries brought
about by periodic high interest rates. The supply of copper also
decreased by about 15 percent as a result of a lengthy industry-wide
strike.62 World copper inventories held by producers and commodity
exchange warehouses remained at 1979 levels in 1980, far below the
1978 peak of 1,560 Gg.63
9-45
-------
TABLE 9-14. UNITED STATES AND WORLD COMPARATIVE TRENDS IN
REFINED COPPER CONSUMPTION, 1963-197958
(gigagrams)
Years
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
Average annual compound
growth rate (%)
1963-1973
1964-1974
1967-1977
1969-1979
U.S.
1,590.0
1,690.0
1,845.6
2,157.8
1,797.5
1,701.4
1,944.3
1,854.3
1,830.5
2,028.6
2,218.6
1,956.4
1,396.0
1,783.0
1,986.0
2,181.0
2,218.0
3.39
1.47
1.00
1.33
World
5,519.3
5,995.4
6,193.2
6,444.8
6,194.8
6,523.3
7,148.0
7,283.4
7,309.9
7,944.5
8,791.6
8,325.4
7,460.0
8,509.0
9,006.0
7,289.0
7,412.0
4.77
3.34
3.81
0.36
U. S. as per-
cent of world
28.8
28.2
29.8
33.5
29.0
26.1
27.2
25.5
25.0
25.5
25.2
23.5
18.7
21.0
22.0
29.9
29.9
-
-
-
-
9-46
-------
TABLE 9-15. UNITED STATES AND WORLD COMPARATIVE TRENDS IN COPPER PRODUCTION: 1963-197959
(gigagrams)
Mine production of copper
(copper content)
Years
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
Average annual
compound growth
rate (%)
1963-1973
1964-1974
1968-1978
U.S.
1,100.6
1,131.1
1,226.3
1,296.5
865.5
1,092.8
1,401.2
1,560.0
1,380.9
1,510.3
1,558.5
1,445.7
1,282.2
1,457.4
1,364.8
1,352.0
1,450.0
3.54
2.48
2.15
World
4,624.3
4,798.6
4,962.7
5,215.9
5,057.6
5,456.5
5,951.2
6,387.3
6,473.9
7,071.5
7,591.4
7,885.6
6,968.2
7,452.8
7,688.7
7,633.4
-
5.08
5.09
3.41
U.S. as
percent
of world
23.8
23.6
24.7
24.9
17.1
20.0
23.5
24.4
21.3
21.4
20.5
18.3
18.4
19.6
17.7
17.7
-
-
-
-
Smelter production of copper
U.S.
1,176.3
1,214.2
1,300.9
1,330.3
782.3
1,148.9
1,438.3
1,489.0
1,360.8
1,533.5
1,582.1
1,424.2
1,357.5
1,438.3
1,346.6
1,343.0
1,396.0
3.01
1.61
1.57
World
4,634.8
4,851.4
5,024.4
5,167.0
4,891.0
5,507.8
5,972.9
6,309.5
6,380.0
7,003.2
7,445.5
7,933.6
7,535.4
8,026.3
8,187.8
7,708.7
-
4.85
4.77
3.42
U.S. as
percent
of world
25.4
25.0
25.9
25.7
16.0
20.9
24.1
23.6
21.3
21.9
21.2
18.4
18.0
17.9
16.4
17.4
-
_
-
-
Production of refined copper
U.S.
1,709.5
1,805.7
1,942.1
1,980.7
1,384.9
1,668.3
2,009.3
2,034.5
1,780.3
2,048.9
2,098.0
1,938.3
1,610.7
1,714.2
1,677.0
1,843.0
1,992.0
2.07
0.71
1.00
World
5,399.7
5,739.0
6,058.5
6,322.2
6,000.5
6,658.6
7,199.8
7,577.8
7,377.8
8,068.0
8,497.3
8,851.5
8,402.0
8,851.2
9,148.8
8,856.6
-
4.64
4.43
2.81
U.S. as
percent
of world
31.7
31.5
32.1
31.3
23.1
25.1
27.9
26.8
24.1
25.4
24.7
21.9
19.2
19.4
18.3
20.8
-
_
-
-
NOTE: 1 Gg = 1,000 Mg. One Mg (1,000 kg) equals 1.102311 short tons (907.185 kg = 2,000 Ib avoirdupois, where 1 Ib
avoirdupois equals 0.453592 kg or 453.5924 g).
-------
9.2 ECONOMIC IMPACT ASSESSMENT
9.2.1 Introduction
9.2.1.1 Introduction. This section presents the economic impact
analysis of possible revisions to the existing standards of performance
for primary copper smelters. As discussed in preceding sections, the
possible revisions include the deletion of the current exemption for
reverberatory smelting furnaces when the furnaces process high impurity
concentrates, and the establishment of emission standards for fugitive
emission sources. In addition, the analysis considers the effect of
the standards on future capacity additions on expansions at existing
smelters. The principal economic impacts analyzed are: the ability
of the smelters to increase copper prices in response to an increase
in costs caused by a revised standard, and the impact on profits if
part or all of the costs cannot be passed on in the form of price
increases.
Section 9.2.2 describes the methodology. Sections 9.2.3, 9.2.4,
and 9.2.5 describe the supply elasticities, the demand elasticities,
and vulnerability to imports, respectively. Section 9.2.6 presents
the calculations, and Section 9.2.7 presents the findings.
9.2.1.2 Summary. This analysis focuses on the control costs for
three groups of smelters that may become subject to the revised NSPS.
The three groups of smelters are new greenfield smelters processing
high impurity concentrates, new greenfield smelters processing clean
concentrates, and expansions to existing smelters.
The analysis indicates that in view of the competition presented
both by certain types of domestic smelters and by Japanese smelters,
selected greenfield and expansion scenarios are feasible, but not all
greenfield and expansion scenarios are feasible. Specifically, new
clean concentrate greenfield smelters are feasible, but new high
impurity concentrate greenfield smelters are not feasible. The high
baseline costs for the new high impurity concentrate greenfield smelter
suggests that it would not be built even in the absence of a revised
NSPS. In the case of clean concentrate expansions, scenarios Ib (or
9-48
-------
high impurity concentrate without the exemption), 7, 18, and 24 are
feasible among the scenarios that passed the screening process within
each of the five smelting configurations (I, II, III, IV, and V).
Scenarios 11, 15, and 23 are not feasible. For the flash furnace
scenarios, the revised NSPS would not require any additional control
costs. All of the five flash furnace conversion scenarios are feasible
(5, 6, 16, 17, and 22). The single smelting configuration for a high
impurity concentrate expansion, la, is feasible with the exemption.
9.2.2 Methodology of Impact Analysis
The purpose of this section is to provide an overview of the
methodology used in the analysis. Specifics of the analysis are
presented in subsequent sections.
Many of the companies that produce refined copper are integrated
producers; that is, they own the facilities to treat copper during
each of the four principal stages of processing, raining, milling,
smelting, and refining. Also, several of the producers are integrated
one additional step into the fabrication of refined copper. However,
not all companies in the copper industry are integrated producers.
There are companies that only mine and mill copper ore to produce
copper concentrate and then have the copper concentrate smelted and
refined on a custom basis (the smelter takes ownership of the copper)
or on a toll basis (the smelter charges a service fee and returns the
copper to the owner). The existence of both integrated and noninte-
grated producers introduces a complex economic element into this
analysis. That complex economic element manifests itself in the
choice of the appropriate profit center to focus on for an analysis
of this type that affects only one stage of the production process
(smelting) in a direct way but has indirect effects on the other
stages.
For accounting purposes, integrated producers frequently view the
smelter as a cost center, rather than a profit center. However, in an
economic sense, the smelter provides a distinct contribution to the
production process that ultimately allows a profit to be earned although
that profit may be realized for accounting purposes at another stage
such as the mine or refinery.
9-49
-------
Although this standard directly affects only the smelters, this
analysis focuses on the smelter plus refinery as a separate profit
center. The mine and mill are excluded. The choice of the smelter
plus refinery as a profit center is made to facilitate comparisons
between foreign smelting plus refining costs versus domestic costs.
This choice facilitates comparisons because when concentrate is shipped
to a foreign smelter, the concentrate is both smelted and refined
overseas. The transportation cost to ship concentrate overseas is a
significant cost. Although smelting and refining are separate produc-
tion processes, an economic relationship exists between the two.
Therefore, because financial information to apportion transportation
costs in a manner that properly reflects the economic relationship is
not published by foreign smelters, the combination of the smelter and
refinery eliminates the need for such an apportionment and permits the
use of total transportation costs. Finally, the choice of the smelter
plus refinery as a separate profit center presents a conservative
analysis in that the economic impacts on the smelter plus refinery
will be overestimated if a smelter plus refinery are actually part of
an operation that is integrated from mining through refining, but the
impacts will not be underestimated for a smelter plus refinery that is
not fully integrated from mining through refining. In other words,
profits from the mines might be available in some cases to help offset
the costs of controls at the smelter. However, this; analysis will
assume that this situation does not occur.
The choice of analyzing the smelter and refinery segments only has
a disadvantage. Most of the studies that have quantified price or
supply elasticities for copper have been for the total of mining
through refining and have not been segment-specific. Therefore, such
studies are useful only in qualitative ways in this analysis rather
than in direct quantitative ways.
9.2.2.1 Smelter Competition and Options. Mines have long-run
flexibility in deciding where they will send their copper concentrate
for smelting. Therefore copper smelters face competition from three
sources: other existing domestic smelters, new smelters that may be
built, and foreign smelters, especially Japanese.
9-50
-------
Japan is a major force among copper-producing countries in terms
of its volume of smelting, refining, and fabrication of copper.
However, Japan does not have copper ore deposits of any noteworthy
size. Therefore it must import concentrates in order to supply its
smelting, refining, and fabricating facilities. Japan seeks concentrates
from many countries, including the United States. Japan's ability to
be competitive with domestic smelters for U.S. concentrates is indicated
by the contractual arrangements it has established with Anamax and
Anaconda to purchase concentrates. Additionally, the Duval Corporation
is reported to have signed a contract to supply concentrates to Japanese
smelters. Also, the Japanese smelters have approached many other
copper mine owners in the United States. For example, Cyprus Corporation
is reported to have seriously considered shipping concentrates from
its Bagdad mine to Japan.
The cost of transporting concentrates across the Pacific Ocean is
significant. It is noteworthy that Japanese smelters can compete with
U.S. smelters in spite of these costs. One reason the Japanese can
compete is that their smelters are newer than U.S. smelters and, in
theory, are more cost competitive. Other factors that operate to the
advantage of Japanese smelters, including a tariff mechanism, are
described later.
The existence of competition for concentrates introduces what is
referred to in this analysis as a "trigger" price. The "trigger"
price is that price which triggers or provides an economic incentive
for the supplier of concentrate to change to another smelter and
refinery. If a given smelter charges a service fee in excess of
competing smelters, that smelter will lose business and eventually be
forced to cease operations. In the case of new smelters or expansions,
the new process facilities will not be built. Faced with an increase
in costs, a smelter could respond using one of three options, or any
combination of the three. First, the smelter could pass the costs
forward in the form of a price increase. Two important considerations
with respect to a price increase are: the prices of competitors in
9-51
-------
the copper business and the elasticity of demand for the end users of
copper. For example, even if all copper producers experience the same
increase in costs, at some point the end users of copper will consider
changing to a substitute. Second, the smelter could absorb the cost
increase by reducing its profit margins, and thereby reducing its
return on investment (ROI). If the smelter's profit margins are
reduced significantly it will cease operation. Third, the smelter
could pass the costs back to the mines by reducing the price it is
willing to pay for concentrate. An important consideration in setting
the service fee a smelter charges for custom or toll smelting is the
fact that the concentrate may be shipped elsewhere; e.g., to Japan.
Market conditions suggest that the option of passing costs back to the
mines does not seem feasible at this time, due to the existence of
excess smelting capacity.
9.2.2.2 Steps of the General Methodology. Existing smelter
capacity is presently underutilized. In terms of a supply function,
the lowest cost smelters would obtain concentrates first. Because in
the short run the existing underutilized smelters need to recover only
their variable costs to justify continued operation, they are the
lowest cost producers (both fixed and variable costs must be recovered
to justify continued operation in the long run). As a result, before
any new smelter capacity is added, existing smelter capacity must
return to a normal utilization rate. Once market forces have exhausted
the lowest cost capacity (existing smelters), then the choices of new
greenfield capacity, expanded existing capacity, or Japanese capacity
become relevant. Therefore, the following discussion of new capacity
is predicated on an increase in demand sufficient to require new
capacity.
In evaluating the economic feasibility of expanding capacity,
costs must be assembled for both new greenfield units and the several
expansion options. These costs are then compared against an estimate
of what it would cost for a Japanese smelter to buy concentrate in the
United States, smelt and refine it in Japan, and return it to the
United States for sale. Any options exceeding this cost level will
probably be rejected by a company as unsound.
9-52
-------
The next step is to calculate the maximum percent price increases
that will be required under each option. This yardstick serves as an
upper bound to show the maximum cost that must be passed through to
the consumer so that there will be no effect on a company's profit
position. Since copper prices tend to move only slowly, there is
little chance that costs can be passed through completely. The
opposite bound is the absorption by the company of all control costs
with the result that profit margins are lowered.
9.2.3 Price Elasticity of Supply
Considerable time and effort, by others, have gone into the
development of econometric models of the copper industry. These
models attempt to quantify the dynamic workings of the copper market.
Selected aspects of these models are summarized below in
the interest of thoroughness.
Table 9-16 displays the various price elasticities of supply
estimates produced by the better-known econometric copper models. The
Fisher, Cootner, and Baily (FCB) model considers U.S. mine production
as a function of the real price of copper (U.S. producer price deflated
by the U.S. wholesale price index) and U.S. mine production lagged
one period. The FCB model estimates the U.S. price elasticity of
supply to be 0.45 in the short run and 1.67 in the long run. This
implies that a 1-percent increase in the price of copper results in a
0.45-percent increase in the quantity of copper supplied in the short
run and a 1.67-percent increase in the long run.
Mi Resell has questioned the appropriateness of the FCB supply
model.67 First, he notes that the U.S. primary copper industry tends
to behave more as a price maker than as a price taker. Thus, one must
be skeptical of the behavioral underpinnings of the model. Second,
marginal or variable costs do not enter into the model's specification
at all. Third, the partial adjustment hypothesis (i.e., using supply
lagged one period to capture a partial adjustment toward a static
equilibrium) implied by the model is both an inadequate and an
unrealistic means of explaining changes in supply. Mikesell believes
9-53
-------
TABLE 9-16. PRICE ELASTICITIES OF SUPPLY ESTIMATES3
Time period Short run Long run Study conducted by
1950-66 0.45 1.67 Fisher, Cootner, and Baily
1950-67 0.34 0.85 Charles River Associates, Inc.C
1970 - 0.61 U.S. Bureau of Minesd
Elasticities evaluate at mean values.
Reference 64.
Reference 65.
Reference 66.
9-54
-------
the entire process of exploration and development over the last decade
or so cannot be explained simply by reference to past prices or the
lag structure assumed in the FCB model.
The Charles River Associates (CRA) model builds on and is a sub-
stantial improvement over the FCB model. It incorporates additional
explanatory variables, such as a capacity index and an index of factor
prices. As Table 9-16 shows, the CRA model estimates the short-run
supply elasticity of copper in the United States to be 0.34 and the
long-run supply elasticity to be 0.85.
The third estimate of the long-run supply elasticity of copper
noted in Table 9-16 is based on a 1973 Bureau of Mines study. Accord-
ing to this study (based on 1970 data), the total known U.S. domestic
copper reserves available at a price of $1.10/kg (and assuming a
12-percent return on investment) would be 76 million megagrams. At a
price of $4.40/kg (also assuming a 12-percent return on investment),
the available supply would be 164 million megagrams. The implied
price elasticity of supply is 0.61.
Given the difficulties with the FCB model, the range for the
long-run price elasticity of supply will be taken to be 0.61 to 0.85
for the purposes of this analysis. The price elasticity of supply
estimate is used here to provide added insight in a qualitative sense
rather than in a purely quantitative sense.
9.2.4 The Price Elasticity of Demand
Table 9-17 contains several price elasticity of demand estimates
for copper. As in the case of the price elasticity of supply estimates
above, the price elasticity of demand estimates are provided here to
add insight of a qualitative nature into the operation of the copper
market. Concerning short-run own-price elasticity (rather than cross-
price elasticity) estimates, the demand appears to be rather inelastic,
ranging from a low of -0.21 to a high of -0.47. In the long run, of
course, the demand is much more elastic, ranging from a low of -0.64
to a high of -2.88. For the purposes of this analysis, the relevant
(long-run) price elasticity of demand is taken to be -0.64, primarily
because this is the elasticity employed in the Midas II model.
9-55
-------
TABLE 9-17. PRICE AND INCOME ELASTICITIES OF DEMAND ESTIMATES1
Time
period
1950-73
1950-67
1950-66
Own price
elasticity
Short Long
run run
-0.47
-0.21
-0.21
-0.64
-2.88
-0.90
Cross-price
elasticity
(aluminum)
Short
run
0.61
0.46
0.24
Long
run
0.84
6.30
1.01
Income
(activity).
elasticity
Short
run
1.30
0.26
0.33
Long
run
1.78
3.56
1.40
Study
Arthur
conducted by
D. Little, Inc.c
Charles River Associates,
Inc.
Fisher
Baily6
, Cootner, and
Elasticities evaluated at mean values.
DThe income (activity) measures used in each model are as follows: The
ADL and CRA models used the FRB index of durable goods manufacturers; and
the FCB model used the FRB index of industrial production.
Reference 68.
Reference 69.
"Reference 70.
9-56
-------
Accordingly, a 1-percent increase in the price of copper is expected
to decrease the quantity of copper demanded by 0.64 percent, assuming
everything else remains unchanged.
Cross-price (aluminum) and income (activity) elasticity estimates
are also reported in Table 9-17. The demand for copper appears to be
somewhat insensitive to aluminum prices in the short run, but consider-
ably more sensitive in the long run. In fact, according to the CRA
model, a 1-percent decrease in the price of aluminum can result in as
much as a 6.3-percent decrease in the demand for copper. With respect
to the income (activity) elasticities, the results are mixed in the
short run, but are relatively elastic in the long run, ranging from a
low of 1.40 to a high of 3.56.
9.2.5 Analysis
This section presents the calculations of the control costs on
the common basis of cents per kilogram of copper to permit direct
comparison across greenfield smelters and expansions of various sizes.
The various total costs for greenfield smelters and expansions are
taken from Chapter 8. After the costs are converted to a common basis
the results are analyzed according to the methodology described earlier.
9.2.5.1 Costs of New Smelting Capacity. Table 9-18 shows the
costs in cents per kilogram of the control option for a new high
impurity greenfield smelter. The costs are 70.3C for process costs,
2.8$ for the revised NSPS, and 1.5
-------
TABLE 9-18. COST DATA FOR NEW HIGH
IMPURITY GREENFIELD.SMELTERS3
(MHR-RV-C)0
Cents/kilogram
(Process + S02) + NSPS = Total
70.3 + 2.8 = 73.1
Fugitive control costsc
(Collection, 103 $)
1. MHR-RV-CV
AC (1,401 + 234) -=- 110,000 = 1.5
-------
TABLE 9-19. COST DATA FOR NEW GREENFIELD SMELTER PROCESSING
CLEAN CONCENTRATES USING A FLASH FURNACE
Baseline
annual i zed
costs
(103 $)
NSPS
fugitive
costs
(AC)
(103 $)
Total
annual i zed
costs
(103 $)
Blister .
production
(Mg/yr)
Baseline
annual i zed
costs
(
-------
by-product revenues were included, it is not expected that the costs
for a new greenfield smelter processing high impurity concentrates
would decline to the level of a new greenfield smelter processing
clean concentrates.
The third smelter group of interest is the expansion situation at
an existing smelter processing either clean or high impurity concen-
trates. For expansions processing clean concentrates, there are 26
expansion scenarios that are divided into five, smelting configurations.
Within each of the five smelting configurations there is an additional
subdivision according to the percent expansion--20, 40, 50, 60, or
100 percent. Within the 26 expansion scenarios, most of the scenarios
are direct expansions. However, there are seven scenarios that represent
a conversion to a flash furnace--5, 6, 16, 17, 22, 24, and 25. Throughout
the remainder of the analysis both the lowest cost, direct expansion
scenarios and the five flash furnace scenarios are presented. Table 9-20
shows the lowest cost expansion scenarios within each percent expansion
(1, 7, 11, 15, 18, 23, and 26) plus the seven flash furnace conversion
scenarios. The above group is reduced to the lowest cost scenario
within each of the five smelting configurations (excluding the seven
flash furnace conversion scenarios). The 20-percent expansions have
lower costs than the 40-percent expansions (for IV the single available
scenario is a 40-percent expansion). For the five smelting configura-
tions, the lowest cost scenarios are Ib, 7, 18, 23, and 26. These
scenarios have costs of 32.0, 30.2, 14.9, 60.6, and 22.4
-------
TABLE 9-20. SMELTER COST DATA FOR EXPANSION SCENARIOS1
(t/kg)
Smelting
configuration
Expansion
scenario
Percent
expansion
Smelter
production
costs
NSPS fugitive
control costs
AC (capture +
collection)
Clean concentrates
I
(MHR-RV-CV)
II
(RV-CV)
III
(FBR- RV-CV)
IV (EF-CV)
Ib
5
6
7
11
15
16
17
18
22
23
24
25
20
50
100
20
50
40
50
100
20
60
40
50
100
32.0
18.9
22.9
30.2
42.4
67.1
22.0
25.2
14.9
20.5
60.6
17.7
18.8
5.6
0.0
0.0
4.8
17.4
1.3
0.0
0.0
2.1
0.0
3.1
0.0
0.0
V (FF-CV) 26 20 22.4 0.0
High impurity concentrates (with exemption)
I
(MHR-RV-CV)
la
20
32.0
0.0
Excluding sulfuric acid credits.
3Smelter production costs includes all process and control costs other
than those associated with a possible revision of the NSPS. See Tables
8-14 and 8-15.
9-61
-------
Table 9-20 also shows the costs for expansions at an existing
smelter that processes high impurity concentrates. The cost of scenario
Ib is 32.0$, without the revised NSPS control cost, and a revised NSPS
control cost of 5.6$, or a total cost of 37.6
-------
9.2.5.3 Costs of Japanese Smelters. The fifth, and last, smelter
group is also not subject to the revised NSPS, but must be considered
here because of its competition with the other smelter groups. The
fifth smelter group is composed of the Japanese smelters.
On December 17, 1980, representatives of the Anaconda Company and
a consortium of seven Japanese copper products, led by Nippon Mining,
signed a formal agreement providing for smelting and refining of
Anaconda Company concentrates by the seven Japanese firms.72 Under
the agreement, Anaconda will ship 390,000 Mg/yr of copper concentrates
to Japan over the first 2 full years, and the amount will be increased
to 500,000 Mg/yr for the remaining 5 years of the contract.73 This
figure accounts for approximately 80 percent of the company's annual
copper concentrate output. At an average of 26 percent copper, the
390,000 tons of concentrates equates to 101,400 Mg of copper and the
500,000 tons equates to 130,000 Mg of copper. Nippon Mining will
receive half of the concentrates, with the remainder divided between
Sumitomo Metal Mining, Mitsui Mining and Smelting, Mitsubishi Metal,
Dowa Mining, Furakawa, and Nittestu Mining.
The Japanese copper producers have the option of toll-smelting
Anaconda's concentrates and returning them to the U.S. company in the
form of metal, or the concentrates may be purchased by the Japanese
firms if Anaconda is notified beforehand.73 It was estimated that
50,000 Mg of refined copper will be produced in Japan from Anaconda
Company concentrates in 1981 (1981 will not represent a full contract
year, hence only 50,000 Mg). Of this number, 9,000 Mg were to be
processed on a toll basis and returned to Anaconda.72 It should be
noted, however, that Anaconda has recently closed its Montana operations,
which supplied much of this concentrate.
Smelting and refining costs charged by Japanese producers vary
with what the market will bear. In August of 1978, Mitsui Mining and
Smelting contracted to process Philippines Marinduque concentrates for
31.9
-------
the remainder of 1979 and reached a level of 39.6$/kg for smelting and
refining in July 1980.74 A spot price is generally higher than what a
supplier can obtain for a long-term contract. More recently, the cost
of Japanese smelting was estimated at 23.5$ and refining at 17.6$/kg.
The Japanese willingness to outbid domestic smelters for U.S.
concentrates is reduced but is not offset by the higher transportation
costs to ship concentrate to Japan. The transportation costs of
concentrate overland by rail to a U.S. port are reported as 13.0
-------
capacity. The Japanese copper smelting industry is composed of 14
smelters with a total capacity of 1,235,000 Mg/yr. The U.S. copper
smelting industry is larger, with 15 smelters and a total capacity of
1,723,000 Mg/yr. The normal capacity utilization rate of the Japanese
copper smelter industry in recent years has been about 75 to 80 percent.
In the unlikely event that the Japanese operated at a 100-percent
capacity utilization rate for a sustained period of time, the extra
capacity between the historical average of 75 to 80 percent utilization
and 100 percent utilization results in 250,000 to 310,000 Mg/yr of
additional capacity. If announced Japanese capacity expansions of
72,000 Mg/yr are added, the maximum capacity to accept additional
concentrates would be 322,000 to 382,000 Mg/yr, using optimistic
assumptions and ignoring the effect of internal Japanese demand on
smelting capacity. The significance of the above is that in any
particular case the Japanese may be willing to outbid U.S. smelters
and attract concentrates from any U.S. mine; thus the trigger price is
a constraint operating on U.S. producers. However, the Japanese
cannot totally replace U.S. smelting capacity and, as a result, as
Japanese capacity utilization rates increase and approach maximum
capacity, the Japanese trigger price should increase and thereby
reduce pressure on U.S. producers.
9.2.5.3.1 Japanese copper industry. The Japanese copper industry
has increased production over the past 10 years in all segments except
mining and secondary smelting. The Japanese copper industry's installed
capacity is 1,235,000 Mg/yr.77 Capacity utilization rates achieved in
1979 and 1980 are 80.5 percent and 81.2 percent, respectively.77
The volume of Anaconda Company concentrates expected to be
processed in Japan is approximately equal to 10 percent of total
Japanese imports of copper concentrates.72 The 50,000 Mg of refined
copper processed from Anaconda concentrates in 1981 should raise the
Japanese capacity utilization rate to approximately 85 percent in
1981, assuming installed capacity remains constant. The utilization
rate will increase further in 1982 to about 89 percent, when the full
level of 390,000 Mg/yr of concentrates is sent by Anaconda. Nippon
9-65
-------
Mining, expected to receive about half of the concentrates each year,
is remodeling an idle 4,000-Mg/mo electrolytic refinery into a
5,000-Mg/mo facility. By 1983, when the remodeling is completed, the
Saganoseki smelter and refinery will operate at a rate of 15,000 Mg/mo.72
9.2.5.3.2 Tariff mechanism. One example of foreign government
assistance to the copper industry occurs in Japan. Japanese copper
producers operate under a system that permits the payment of a premium
for concentrates, which is then recovered through a premium for refined
copper due to a protected internal market supported by a high tariff.
Japan imposes high import duties on refined, unwrought copper while
allowing concentrates to be shipped into the country duty-free. Duty
on refined unwrought copper is currently (1981) 8.2 percent of the
value of the copper, including freight and insurance, as opposed to a
U.S. customs duty of 1.3 percent of the value of copper. The import
duties allow Japanese producers to sell their refined copper in Japan
at an artificially high price and still remain competitive with foreign
producers.
Specifically, copper concentrates and ore imported into Japan are
free of duty. Refined copper imported into Japan is subjected to a
tariff of 15,000 yen/Mg.78 Using a December 15, 1980, exchange rate
of $0.004633/yen, the tariff is $0.0849/kg. Refined copper may be
duty-free under the preferential tariff, subject to certain limitations.
As a result of the tariff situation, Japanese copper producers
can pay a premium to attract concentrates and can recover the premium
through a premium on the price of the refined copper used in Japan.
If the refined copper is returned to the customer outside of Japan,
the premium on the price of refined copper is not recovered because
world prices would prevail in this case, rather than the protected
internal Japanese producer price. As a result, the principal interest
of the Japanese copper producers is in producing copper for internal
consumption. Toll smelting in Japan is generally used as a means of
balancing inventories. The absence of a tariff on ore and concentrates
encourages companies to import ore into Japan. The presence of a
tariff on refined copper and the costs of holding metal in Japan
discourage companies from importing refined copper into Japan.
9-66
-------
The Japanese tariff on refined copper, combined with the cost of
holding the metal until users have a demand for it, provides an extra
margin for domestic copper producers. The Japanese producers can
charge what the market will bear for their copper and still remain
competitive with the importers. The loss incurred by Japanese producers
in charging toll customers low processing rates is covered by the
extra margin of profit realized by charging prices for domestic refined
copper at competitive import levels.
Robert H. Lesemann (industry expert, formerly with Metals Week,
now with Commodities Research Unit), in an affidavit for the Federal
Trade Commission, outlined the situation in September 1979:
It is generally true that operating costs of U.S. smelters
are the same as smelters in Japan, Korea, and Taiwan. The
competitive advantage is without doubt due to the subsidies
outlined above. Thus, while the terms of the Nippon-Amax
deal have not been revealed, the treatment charge is likely
well below the operating cost levels of U.S. smelters.79
9.2.5.3.3 Other Japanese advantages. The tariff mechanism
described above is one example of government assistance to the Japanese
copper industry. Another example is provided by the Japanese govern-
ment's approval for a brass rod production cartel. In an effort to
reduce stocks and boost profit margins for the ailing Japanese brass
rod industry, the government approved the formation of a temporary
cartel to cut production.80
Apart from government assistance, other reasons are cited for the
advantage of the Japanese copper industry over the U.S. copper industry.
Additional reasons include:
A high debt-to-equity ratio—a typical Japanese smelter
may have a debt-to-equity ratio that is 0.8 to O.9.81 82 83
Lower labor rates--Japanese hourly rates in the primary
metals industry were estimated to be about two-thirds
of the U.S. rate in 1978.84
By-product credits--the market for by-products, sulfuric
acid, and gypsum is better in Japan than in the United
States and reduces operating costs significantly.81
9-67
-------
9.2.5.4 Refining Costs. Preceding sections have described the
pollution control costs and process costs associated with the smelter
alone. This section presents the cost of refining copper. The costs
of smelting plus refining plus transportation form the overall cost of
the profit center. The addition of the cost of refining to the overall
cost of the profit center permits a more meaningful comparison of the
Japanese alternative and the U.S. alternative. The comparison is more
meaningful because concentrates sent to Japan are both smelted and
refined in Japan. Therefore, the U.S. alternative must be presented
on the basis of both smelting and refining.
As in the case of the smelters, the costs of the refineries vary
somewhat among the actual refineries. However, in keeping with a
model plant analysis, and because the focus of the analysis is on the
smelters rather than the refineries, a model refinery cost is used.
The model refinery cost used is 23.1
-------
en
<£>
LOSt
(«7kg
Refined
Copper)
Japanese
100
90
8U
70
60
bO
40
30
20
10_
0
83.0
5.9 trans
6.5
profit
17.6
refine
23.5
smelt
29.5
trans
to
Japan
62.3 1.3
-------
4.4
Tnm
FBRRMV
20%
II
17.7
«™ct
4.4
Tnm
SOX
24
4.4
Tnm
EF-CV
100%
404
!•«
tm*
13.1
RiKni
4.4
Tnm
20.5
bmll
23.1
Rrfln.
4.4
Tnm
«HR.RV<:V FBR.RV.CV
60% 00%
5 22
4U
22.0
23.1
lUKm
4.4
Tnm
RVCV
W
40.0
22.4
tatlt
21.1
tUKm
4.4
Tnm
20%
20
22.0
ta*tt
Rrtm
4.4
Tnm
MHR-RV«V
100%
1
4.4
Tnm
100%
17
SM
32.0
tm*
23.1
4.4
HI* MHR-
HV«V
m
la
•0.1
M
tan
3JO
tm*
U.1
Rlflm
4.4
Tnm
•U
HI or LI"
MHR-RV-CV
20%
1b
4J
Ntn
30.1
tmrft
211
iMm
4.4
Tnm
17.3
1.1 F«»
1«J
NtPt
a.4
fclMH
O.1
Rt«™
4.4
Tnm
nx
J.I
«.«
t~«
23.1
B.HO.
44
Tm
UFm.
•7.1
fc~ll
23.1
Rdtau
4.4
TnM
RVCV RVCV tF-CV RV*V
T r s* r
NOTES: Domestic refining - 23.1 cents/kg
4.4 cents/kg - Transport. Mine-Smeltw-Refinery
•High-impurity concentrate with exemption
• 'High- or tow-impurity concentrate without exemption
Figure 9-9. Costs for smelting and refining in Japan vs. costs at expanding smelters in the United States.
-------
9.2.6 Findings
9.2.6.1 Maximum Percent Price Increase. Insight into the economic
impact of the revised NSPS can be gained by examining the maximum
percent price increase that is necessary to pass all control costs
forward in the form of a price increase. A complete pass forward of
costs may not be possible in every case, and later this assumption is
relaxed. However, assuming a complete pass forward is possible in
every case introduces a common reference point, which then facilitates
comparison of various base cases and scenarios.
Table 9-21 shows such a percent price increase comparison for
selected cases. The cases selected are for high impurity greenfield
smelters, clean greenfield smelters, clean concentrate expansions, and
high impurity concentrate expansions. The cases selected are the
least-cost choices for each smelting configuration and for each percent
expansion scenario within the smelting configurations. The percent
price increases are calculated using a simplified method for ease of
presentation. For Table 9-21 the percent price increases are calculated
by dividing the costs associated with the revised NSPS by the appro-
priate baseline production costs (both expressed in cents per kilogram).
The results are shown in Table 9-21 for two circumstances—nonintegrated
and integrated producers. The first circumstance is the more conserva-
tive and presents the results on a nonintegrated basis for the smelter
plus refinery alone, excluding the mine and mill. Because many of the
producer are integrated from mining through refining, the nonintegrated
results are shown in comparison to the results for an integrated
producer if the price of refined copper is increased to pass the costs
forward. The refined copper price used is the average price for 1981,
187
-------
TABLE 9-21. MAXIMUM PERCENTAGE PRICE INCREASE3
Control
option or Nonintegrated Integrated
Smelting expansion Percent producer percent producer percent
configuration scenario expansion price increase price increase
High impurity concentrate greenfield smelter
NA 4.4 2.3
Clean concentrate greenfield smelter
I
II
III
IV
V
Clean
Ib
5
6
7
11
15
16
17
18
22
23
24
25
26
NA
2.1
0.7
concentrate expansion
20
50
100
20
50
40
50
100
20
60
40
50
100
20
High impurity concentrate
I
la
20
9.4
0.0
0.0
8.3
24.9
1.4
0,0
0.0
5.0
0.0
3,5
0.0
0.0
0.0
expansion (with exemption)
0.0
3.0
0.0
0.0
2.5
9.3
0.7
0.0
0.0
1.1
0.0
1.7
0.0
0.0
0.0
0.0
NA = Not applicable.
aData from Table 9-23.
Nonintegrated producer includes smelter and refinery, but excludes mine
and mill.
Integrated producer includes mine, mill, smelter, and refinery.
9-72
-------
9.2.6.2 Maximum Percent Profit Reduction. Apart from the
calculation of maximum percent price increase, additional insight into
the economic impact of possible revisions to the NSPS can be gained by
making the opposite assumption from maximum percent price increase;
that is, zero percent price increase, or complete cost absorption.
The assumption of complete control costs absorption provides a measure
of the reduction in profits if the control costs are absorbed completely.
Table 9-22 is similar in format to Table 9-21, but Table 9-22
shows the results of complete control cost absorption. Again, the
results are shown for both a nonintegrated (smelter plus refinery
alone) and an integrated producer (mining through refining). A
10-percent profit margin is used and the margin is reduced accordingly,
depending on the magnitude of the costs absorbed. Table 9-22 shows
that the maximum profit reduction ranges from zero profit reduction to
100 percent profit reduction.
Generally, the high impurity concentrate cases, both greenfield
and expansions, have maximum percent profit reductions that are consid-
erably higher than the profit reductions for the clean concentrate
cases. This situation assumes that the high impurity reverberatory
furnace exemption is deleted. The range for high impurity concentrate
cases is 23.0 to 93.3 percent, while the range for clean concentrate
is zero to 100.0 percent. In addition to the large relative difference
in profit reductions between high impurity concentrate cases and clean
concentrate cases, the actual level of profit reductions for high
impurity concentrate cases is also large.
9.2.6.3 Results. Table 9-23 summarizes the costs. The new
greenfield smelter processing high impurity concentrates is not feasible
and, due to its high baseline costs, would not be feasible even in the
absence of a revised NSPS. A new greenfield smelter processing clean
concentrates is feasible. For the clean concentrate expansions,
scenarios Ib (or high impurity concentrate without the exemption), 7,
18, and 26 are feasible. Scenarios 11, 15, and 23 are not feasible.
For the flash furnace conversion scenarios, the revised NSPS would not
require any additional control costs. All of the seven flash furnace
9-73
-------
TABLE 9-22. MAXIMUM PERCENT PROFIT REDUCTION3
Control
option or Nonintegrated Integrated
Smelting expansion Percent producer percent, producer percent
configuration scenario expansion profit reduction profit reduction
High impurity concentrate greenfield smelter
NA Exceeds trigger price
Clean concentrate greenfield smelter
I
II
III
IV
V
Clean
Ib
5
6
7
11
15
16
17
18
22
23
24
25
26
NA
21.3
7.0
concentrate expansion
20
50
100
20
50
40
50
100
20
60
40
50
100
20
High impurity concentrate
I
la
20
93.3
0.0
0.0
82.8
Exceeds trigger
Exceeds trigger
0.0
0.0
47.2
0.0
Exceeds trigger
0.0
0.0
0.0
expansion (with exemption)
0.0
29.9
0.0
0.0
25.7
price
price
0.0
0.0
11.2
0.0
price
0.0
0.0
0.0
0.0
NA = Not applicable.
aData from Table 9-23.
Nonintegrated producer includes smelter and refinery, but excludes mine
and mill.
clntegrated producer includes mine, mill, smelter, and refinery.
9-74
-------
TABLE 9-23. SUMMARY OF SELECTED CASES
(t/kg)
I
^g
CJl
Smelting
Smelter configu-
category ration
Greenfield
High impurity concentrate
Clean concentrate
Expansion
Clean concentrate I
II
III
IV
High impurity concentrate8 V
I
Japanese
Scen-
ario
Ib
5
6
7
11
15
16
17
18
22
23
24
25
26
la
Base-
1 i ne S02
70.3 2.8
33.5
32.0
18.9
22.9
30.2
42.4 16.3
67.1
22.0
25.2
14.9
20.5
60.6
17.7
18.8
22.4
32.0
Smelting
23.5
Fug.
+ 1.5 =
+ 1.3 =
+ 5.6 =
+ 0.0 =
+ 0.0 =
+ 4.8 =
+ 1.1 =
+ 1.3 =
+ 0.0 =
+ 0.0 =
2.1 =
+ 0.0 =
+ 3.1 =
+ 0.0 =
+ 0.0 =
+ 0.0 =
+ 0.0 =
Cost with
controls
74.6
34.8
37.6
18.9
22.9
35.0
59.8
68.4
22.0
25.2
17.0
20.5
60.6
17.7
18.8
22.4
32.0
Refining
cost
23.1
23.1
23.1
23.1
23.1
23.1
23.1
23.1
23.1
23.1
23.1
23.1
23.1
23.1
23.1
23.1
23.1
17.6
Transpor-
tation
4.4
4.4
4.4
4.4
4.4
4.4
4.4
4.4
4.4
4.4
4.4
4.4
4.4
4.4
4.4
4.4
4.4
35.4
Total
cost
102.1
62.3
65.1
46.4
50.4
62.5
87.3
95.9
49.5
52.7
44.5
48.0
91.2
45.2
46.2
49.9
59.5
83. Ob
With exemption.
76.5 plus 6.5 for profit.
-------
conversion scenarios are feasible (5, 6, 16, 17, 22, 23, and 24). The high
impurity concentrate is not shown because small changes in the relevant
financial and cost parameters would not cause a change in the conclusions.
Even if the costs for a new greenfield smelter processing high
impurity concentrates were more favorable, there are two additional
issues that complicate an analysis of the decision to build a new
greenfield smelter processing high impurity concentrates: (1) Do the
Japanese have the capability and the willingness to accept more high
impurity concentrates; and (2) what would be the source of supply of
high impurity concentrates for a new greenfield smelter processing
such concentrates. With respect to the first issue, the indicators
are mixed. One indicator that suggests additional Japanese capability
to accept high impurity concentrates is the two expansions mentioned
earlier. Two indicators that suggest a lack of willingness for the
Japanese to accept additional high impurity concentrates are: the
forecast of moderate to low growth in the Japanese economy for the
next few years, and the fact that the Japanese exchanged high impurity
concentrates for clean concentrates with the Taiwanese.87 88
With respect to the second issue, the source of high impurity
concentrates, the answer is complex. An examination of this question
several years ago concluded that the supply of such concentrates was
tight and likely to remain so for some time.89 A comprehensive exami-
nation of the two issues above would require an intensive effort that
focused specifically on these issues; such an effort is outside the
scope of this analysis.
9.3 SOCIOECONOMIC IMPACT ASSESSMENT
9.3.1 Executive Order 12291
The purpose of Section 9.3.1 is to address those tests of macro-
economic impact presented in Executive Order 12291, and, more generally,
to assess any other significant macroeconomic impacts that may result
from the revised NSPS. Executive Order 12291 stipulates as "major
rules" those that are projected to have any of the following impacts:
9-76
-------
An annual effect on the economy of $100 million or more.
A major increase in costs or prices for consumers;
individual industries; Federal, State, or local govern-
ment agencies; or geographic regions.
Significant adverse effects on competition, employment,
investment, productivity, innovation, or the ability of
U.S.-based enterprises to compete with foreign-based
enterprises in domestic or export markets.
The primary copper smelter industry is currently operating at a
low capacity utilization rate in response to the depressed general
economy. Starting from a low capacity utilization rate and given
moderate to low growth prospects for the demand for copper over the
next 5 years, the need for new sources appears minimal. If new sources
are built in the next 5 years they are more likely to be built to
replace existing smelters that are closed rather than to add capacity
to meet an increase in demand. The number of existing smelters that
may close by 1988 is estimated to range as high as six.90 The closure
of six smelters would have a severe impact on the industry. This
occurrence would seem to be unlikely because as one or several smelters
close, the lost capacity should enable the remaining smelters either
to charge more for their services or to operate at a higher capacity
utilization rate, and thus quite possibly to remain open.
A revised NSPS would only apply to new sources (new greenfield
smelters or expansion at existing smelters), not to existing sources.
As such, a revised NSPS will not cause closures. However, a revised
NSPS may have an impact if it prevents new sources from opening.
If a new greenfield smelter processing high impurity concentrates
is built and becomes subject to the revised NSPS, the control costs
would add a maximum of $13,150,000 for S02 controls and $1,635,000 for
fugitive controls, or a total of $14,785,000, as shown in Table 9-18.
This total of $14,758,000 represents the highest cost option, I-B. If
a' new greenfield smelter processing clean concentrates is built and
becomes subject to the revised NSPS, the costs for fugitive controls
would add a maximum of $1,401,000 as shown earlier. If a revised NSPS
prevented a new source from opening, the lost blister production would
be 110,000 Mg/yr in the case of a greenfield smelter.
9-77
-------
TABLE 9-24. NUMBER OF EMPLOYEES AT COMPANIES THAT OWN PRIMARY
COPPER SMELTERS
Company
ASARCO, Inc.
Cities Service Company
Copper Range Company
Inspiration Consolidation Copper
Company
Kennecott Corporation0
Newmont Mining Corporation
Phelps Dodge Corporation
Employees
12,700
18,900
3,049
2,180
35,000
12,400
15,220
Source9
Reference 91
Reference 92
Reference 93
Reference 94
Reference 95
Reference 96
Reference 97
Reference 91.
Copper Range Company is a wholly owned subsidiary of the Louisiana Land
and Exploration Company. Figures are for Louisiana Land and Exploration.
Prior to merger with Sohio on March 12, 1981.
9-78
-------
9.3.2 Regulatory Flexibility
The Regulatory Flexibility Act (RFA) of 1980 requires that differ-
ential impacts of Federal regulations upon small business be identified
and analyzed. The RFA stipulates that an analysis is required if a
substantial number of small businesses will experience significant
impacts. Both measures must be met, substantial numbers of small
businesses and significant impacts, to require an analysis. If either
measure is not met then no analysis is required. The EPA definition
of a substantial number of small businesses in an industry is 20 percent.
The EPA definition of significant impact involves three tests, as
follows: (1) prices for small entities rise 5 percent or more, assuming
costs are not passed onto consumers; or (2) annualized investment
costs for pollution control are greater than 20 percent of total
capital spending; or (3) costs as a percentage of sales for small
entities are 10 percent greater than costs as a percentage of sales
for large entities.
The Small Business Administration (SBA) definition of a small
business for SIC Code 3331, Primary smelting and refining of copper,
is 1,000 employees. Table 9-24 shows recent employment levels for
each of the seven companies that own primary copper smelters. All
seven have more than 1,000 employees. Therefore, none of the seven
companies meets the SBA definition of a small business and thus no
regulatory flexibility analysis is required.
9.4 REFERENCES
1. United States International Trade Commission. Unalloyed Unwrought
Copper. August 1978. p. A-12.
2. Atlantic Richfield Co. Form 10-K. December 31, 1980. p. 16.
3. ASARCO, Inc. Form 10-K. December 31, 1980. p. A2.
4. Cities Service Co. Annual Report 1980. p. 41.
5. The Louisiana Land Exploration Co. Form 10-K. December 31, 1980.
p. 16.
6. Inspiration Consolidated Copper Company. Annual Report 1980.
p. 2.
9-79
-------
7. Kennecott Corp. Form 10-K. December 31, 1980. p. 4.
8. Newmont Mining Corp. Form 10-K. December 31, 1980. p. 3.
9. Phelps Dodge Corp. Form 10-K. December 31, 1980. p. 2, 4.
10. Copper Development Association, Inc. Annual Data 1980, Copper
Supply and Consumption, p. 4.
11. Reference 10, p. 10.
12. Bureau of Mines. Copper Mineral Commodity Profiles. September
1979. p. 5.
13. AMAX. 1979 Annual Report, p. 12.
14. AMAX. 1980 Annual Report, p. 12.
15. ARCO. Form 10-K. 1980. p. 16.
16. Phelps Dodge Corp. Form 10-K. 1980. p. 5.
17. Newmont Mining Corp. Form 10-K. 1980. p. 3.
18. Reference 10, p. 14. )
19. Reference 10, p. 6.
20. Bureau of Mines, U.S. Department of the Interior. Bureau of
Mines Yearbook. 1975. p. 3.
21. Bureau of Mines. Bureau of Mines Minerals Yearbook. 1977
(Vol. I), p. 331.
22. Phelps Dodge Corp.. Form 10-K. December 31, 1976. p. 6.
23. Newmont Mining Corp. Form 10-K. December 31, 1976. p. 4.
24. Kennecott Copper Corp. Form 10-K. December 31, 1976. pp. 3-5.
25. American Bureau of Metal Statistics. New York, New York. Non-
Ferrous Metals Data Book. 1977.
26. Inspiration Consolidated Copper Co. Annual Report. 1976. p. 5.
27. EPA and Smelter Operators Square Off at Arizona Hearings on S02
Issue. Engineering and Mining Journal. 146:18. February 1976.
28. Cyprus Mines Corp. Form 10-K. December 31, 1976. p. 5-7.
29. ASARCO, Inc. Form 10-K. December 31, 1976. p. A3.
9-80
-------
30. Cities Service Co., Form 10-K. December 31, 1976. pp. 12-13.
31. Copper Range Co. Form 10-K. December 31, 1976. p. 3.
32. UV Industries, Inc. Form 10-K. December 31, 1976. p. 5.
33. Reference 12, p. 13.
34. Rosenkranz, R. D., R. L. Davidoff, and J. F. Lemons, Jr. Copper
Avail ability—Domestic: A Minerals Availability System Appraisal.
U.S. Bureau of Mines. 1979. p. 22.
35. U.S. Bureau of Mines. Cost of Producing Copper From Chalcopyrite
Concentrate as Related to S02 Emission Abatement. 1974. p. 12.
36. U.S. Environmental Protection Agency. Draft of Standards Support
and Environmental Impact Statement, Volume 1: Proposed National
Emission Standards for Arsenic Emissions From Primary Copper
Smelters. June 1978. p. 7-18.
37. Cleaver, George. Merrill, Lynch, Pierce, Fenner, and Smith, Inc.
June 1977. p. 9-10.
38. Weiss, Moshe. The Impact of Environmental Control Expenditures
on the U.S. Copper, Lead, and Zinc Mining and Smelting Industry.
National Economic Research Associates, Inc. January 1978.
Chart B-3.
39. Reference 36, p. 7-19.
40. Reference 37, p. 5.
41. Kovisars, Leons. Copper Production Costs vs. Required Prices.
Presented at SME-AIME Fall Meeting and Exhibit, Tucson, Arizona.
October 17-19, 1979. p. 2.
42. Reference 35, p. 10.
43. Schroeder, H. J. Bureau of Mines Commodity Report on Copper.
June 1977. pp. 16, 17.
44. Copper Development Association. Annual Data. 1981. p. 6.
45. Commodities Research Unit, Ltd. Trends in U.S. Productivity,
Copper Studies. New York. January 15, 1980. pp. 7,8.
46. U.S. Bureau of Mines. Mineral Industry Survey, Copper Survey,
Copper Industry Annual Supplement.
47. U.S. Department of Commerce, U.S. Industrial Outlook. Copper,
Quarterly Report. January 1981. p. 208.
9-81
-------
48. Reference 44, p. 23.
49. Reference 44, p. 31.
50. Commodities Research Unit, Ltd. Copper's Hope: Electric Vehicles,
Copper Studies. New York. March 30, 1979. p. 5.
51. Commodities Research Unit, Ltd. Copper in Military Uses, Copper
Studies. New York. February 15, 1980. p. 1.
52. Metal Statistics. 1981. New York, Fairchild Publications.
p. 71.
53. Commodities Research Unit, Ltd. CRU Metal Monitor Copper.
April 1980. p. 3.
54. Commodities Research Unit, Ltd. The Capital Cost Picture, Copper
Studies. New York. August 18, 1975.
55. Reference 34, p. 13.
56. Reference 34, p. 15.
57. Commodities Research Unit, Ltd. CS Survey: U.S. Expansion
Potential, Copper Studies. New York. March 30, 1979. p. 1.
58. Arthur D. Little, Inc. Economic Impact of Environmental Regula-
tions on the United States Copper Industry. U.S. Environmental
Protection Agency. January 1978. p. V-4.
59. Reference 58, p. V-8.
60. Reference 12, p. 6.
61. Lesemann, Robert. Record Demand and Speculation Drive Copper Over
$1.00 Mark. Engineering and Mining Journal. March 1980. p. 82.
62. Bureau of Mines. Mineral Commodity Summaries - 1981. January 1981.
p. 40.
63. Reference 3, p. 8.
64. Fisher, F. , P. Cootner, and M. Baily. An Economic Analysis of
the World Copper Industry. Bell Journal of Economics and Manage-
ment Science. 3:577. Autumn 1972.
65. Charles River Associates, Inc. Economic Analysis of the Copper
Industry. In: The World Copper Industry, Raymond F. Mi Resell.
March 1970. p. 177.
9-82
-------
66. U.S. Department of the Interior, Bureau of Mines. An Economic
Appraisal of the Supply of Copper from Primary Domestic Sources.
Bureau of Mines Information Circular No. 8598. Washington, DC.
1973. p. 36.
67. Mikesell, Raymond F. The World Copper Industry: Structure and
Economic Analysis. Baltimore, MD, John Hopkins University Press,
1979. pp. 175-176.
68. Reference 58, p. VII-28.
69. Reference 65, p. 280.
70. Reference 64, p. 587.
71. Virginia Polytechnic Institute and State University. A Disaggre-
gated Engineering Supply Model of the U.S. Copper Industry Operat-
ing in an Aggregated World Econometric Supply/Demand System.
U.S. Bureau of Mines Contract J0188158. July 1980. Appendix B,
p. B31.
72. Nippon Mining makes Agreement with Anaconda. Japan Metal Journal
(Tokyo). December 22, 1980. p. 1.
73. Anaconda Will Ship Copper Concentrates to Seven Japanese Copper
Producers. Metals Week. December 22, 1980. p. 2.
74. Japan Metal Journal (Tokyo), August 7, 1978, p. 1; April 16,
1979, p. 2; August 6, 1979, p. 1; July 21, 1980, p. 1.
75. Reiber, Michael. Smelter Emission Controls: The Impact On
Mining and the Market for Acid. Arizona Mining and Mineral
Resources Research Institute, Tucson, Arizona. Office of Surface
Mining. March 1982. pp. 3-59, 3-60, 5-14, 5-15.
76. Returning Shipments of Refined Copper to Anaconda will be Increased
Fairly. Japan Metal Journal (Tokyo). March 22, 1982. p. 1.
77. MITI's Revised Supply-Demand Prospects for FY 1980. Japan Metal
Journal (Tokyo). November 24, 1980. p. 1.
78. Copper Imports on Preferential Tariff. Japan Metal Journal
(Tokyo). December 8, 1980. p. 3.
79. Affidavit of Robert J. Lesemann, Commodities Research Unit/CRI
and former editor-in-chief of Metals Week, to the Federal Trade
Commission. September 27, 1979. FTC Docket Number 9089.
80. Brass Rod Production Cartel Starts. Japan Metal Journal (Tokyo).
July 6, 1981. p. 1.
9-83
-------
81. Smelter Pollution Abatement: How the Japanese Do It. Engineering
and Mining Journal. May 1981. p 72.
82. Reference 75, p. 5-10.
83. Custom Copper Concentrates. Engineering and Mining Journal. May
1982. p. 73.
84. Everest Consulting Associates, Inc., and CRU Consultants, Inc.
The International Competitiveness of the U.S.. Nonferrous Smelting
Industry and the Clean Air Act. Princeton, NJ. April 1982.
p. 9-9.
85. Reference 35, pp. 1, 14, 19, 22, 25, 30, 34, 37, 42, 52, 57, 62,
and 71.
86. Lewis, F. Milton, and Roshan B. Bhappu. Copper Production Costs
Update. Mountain States Research and Development, Tuscon, Arizona.
October 1979. p. 4.
87. Demand Prospects of Copper Electric Wires. Japan Metal Journal
(Tokyo). July 5, 1982. p. 3.
88. NM Agrees with TMM on Exchange of Concentrates. Japan Metal
Journal (Tokyo). October 19, 1981. p. 7.
89. CS Survey "Dirty Concentrates." Copper Studies, New York.
August 31, 1979. pp. 1-6.
90. Reference 82, pp. 1-11.
91. Reference 3, p. A7.
92. Cities Service Company. Form 10-K. 1980. p. 6.
93. Copper Range Company. Form 10-K. 1980. p. 22.
94. Inspiration Consolidated Copper Company. Form 10-K. 1980. p. 2.
95. Reference 7, p. 10.
96. Reference 8, p. 9.
97. Reference 9, p. 1.
9-84
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO. 2.
EPA-450/3-83-018a
4. TITLE AND SUBTITLE
Review of New Source Performance Standards for
Primary Copper Smelters
7. AUTHOR(S)
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
12. SPONSORING AGENCY NAME AND ADDRESS
Office of Air Quality Planning and Standards
Office of Air, Noise, and Radiation
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
3. RECIPIENT'S ACCESSION NO.
5. REPORT DATE
March 1984
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO.
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-02-3056
13. TYPE OF REPORT AND PERIOD COVERED
Draft
14. SPONSORING AGENCY CODE
EPA/200/04
13. SUPPLEMENTARY NOTES
Standards of performance for the control of emissions from primary copper smelters
were promulgated in 1976. Developments since promulgation necessitated that the
following be included in the periodic review of the standards: (1) reexamination
of the current exemption for reverberatory furnaces processing high-impurity materials,
(2) assessment of the feasibility of controlling particulate matter emissions from
reverberatory furnaces processing high-impurity materials, (3) revaluation of the
impact of the current standard on the ability of existing smelters to expand
production, and (4) assessment of the technical and economic feasibility of controlling
fugitive emissions at primary copper smelters. The results of the review indicated
that no changes should be made to the existing standard. This document contains
background information and environmental and economic assessments considered in
arriving at this conclusion.
This report is published in two volumes. Volume 1, EPA 450/3-83-018a, contains
Chapters 1 through 9. Volume 2, EPA 450/3-83-018b, contains the Appendixes.
17 KEY WORDS AND DOCUMENT ANALYSIS
1 DESCRIPTORS
Air pollution
Pollution control
Standards of performance
Primary copper smelters
Sulfur oxides
Particulate matter
8. D'STRIBUT.QN STATEMENT
Unlimited
b. IDENTIFIERS/OPEN ENDEDTERMS
Air Pollution Control
19 SECURITY CLASS ( Tins Report)
Unclassified
20 SECURITY CLASS (This page i
Unclassified
c. COSATI F-ield/Group
13B
21. NO. OF PAGES
579
22. PRICE
:Fi. Form 2220-1 (Rev. 4-77) PREVIOUS EDITION is OBSOLETE
-------
DATE DUE
L.ee^, Hoom 1670
60604
------- |