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

-------