EPA-450/3-81-004
Control Techniques
for Sulfur Oxide
Emissions from
Stationary Sources
Second Edition
by
Emission Standards and Engineering Division
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air, Noise, and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
April 1981
-------
r
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 trade names or commercial products
is not intended to constitute endorsement or recommendation for use.
Copies of this report are for sale by the Superintendent of Documents,
U.S. Government Printing Office, Washington, D.C. 20402, and the National
Technical Information Services, 5285 Port Royal Road,.Springfield,
Virginia 22161.
ii
-------
CONTENTS
Figures
Tables
1. Introduction
2. The Sulfur Oxides:
Methods
Characterization and Sampling
2.1 Characterization of sulfur oxides
2.2 Sulfur oxides emission trends
2.3 Sampling and analysis methods
Considerations In Sulfur Oxides Control
3.1 Energy availability and usage trends
3.2 Determining emission reduction needs
3.3 Technical considerations
3.4 Environmental and energy impacts
3.5 The sulfur market
Combustion Processes
4.1 Nature and extent of sulfur oxide (SO )
emissions from combustion x
4.2 Control techniques
4.2.1 Fuel substitution
Coal
Oil
Natural gas
Source substitution
4.2.2 Fuel desulfurization
Coal cleaning
Synthetic fuels
4.2.3 Fuel gas desulfurization
Lime process
The limestone FGD process
Double alkali process
Nonregenerable sodium-based flue gas
desulfurization
Ammonia-based process
v
x
1-1
2.1-1
2.1-1
2:2-1
2.3-1
3.0-1
3.1-1
3.2-1
3.3-1
3.4-1
3.5-1
4.1-1
4.1-1
4.2-1
4.2-1
4.2-1
4.2-5
4.2-5
4.2-9
4.2-9
4.2-10
4.2-18
4.2-20
4.2-31
4.2-51
4.2-67
4.2-77
4.2-90
-------
r
CONTENTS (continued)
The Wellman-Lord process
The Citrate process
Magenesium oxide flue gas
desulfurization
Adsorption
Dry removal processes
4.2.4 Combined coal cleaning/FGD
Combined coal cleaning and FGD costs
4.2.5 Combustion process modifications
Conventional combustion systems
Fluidized bed combustion
Advanced combustion systems
5. Industrial Processes
5.1 Nonferrous primary smelters
5.2 Iron and steel production
5.3 Petroleum refineries
5.4 Natural gas industry
5.5 Sulfuric acid plants
5.6 Pulp mills
5.7 Coal mining waste disposal
5.8 Glass manufacture
5.9 Mineral products
5.10 Explosives manufacture
5.11 Petrochemicals
5.12 Incineration
Page
4.2-104
4.2-122
4.2-132
4.2-155
4.2-160
4.2-177
4.2-178
4.2-180
4.2-181
4.2-181
4.2-182
5.0-1
5.1-1
5.2-1
5.3-1
5.4-1
5.5-1
5.6-1
5.7-1
5.8-1
5.9-1
5.10-1
5.11-1
5,12-1
-------
FIGURES
Number Page
2.2-1 Nationwide Sources of Sulfur Oxide Emissions, 1977 2.2-4
2.2-2 Significant Areas of Sulfur Oxide Emission 2.2-6
2.2-3 Nationwide S0v Emission Trends 1970-1977 2.2-7
s\
2.2-4 Nationwide Trends in Annual Average Sulfur Dioxide
Concentrations From 1972 to 1977 at 1,233 Sampling
Sites 2.2-9
3.1-1 Cost per Kilojoule (British Thermal Unit) of Selected
Fuels and Purchased Electricity Consumed by All
Manufacturing Industries 1976, 1975, 1974, 1971,
and 1967 ' 3.1-8
3.2-1 Mandatory Class I Areas for PSD 3.2-7
4.1-1 Nationwide SO Emission Estimates 4.1-2
/\
4.2-1 World Share of Crude Oil Production, in 1976 4.2-7
4.2-2 World Daily Petroleum Demand, 1976 4.2-7
4.2-3 Proved Reserves of Liquid and Gaseous Hydrocarbons,
Year-End 1976 4.2-8
4.2-4 Typical Lime FGD System 4.2-34
4.2-5 Tray Absorber 4.2-36
4.2-6 Typical Mobile Bed Scrubber 4.2-38
4.2-7 Two Stage Venturi Scrubber . . 4.2-39
4.2-8 Horizontal Spray Scrubber 4.2-41
4.2-9 Typical Sludge Processing Circuit 4.2-42
4.2-10 Illustration of Capacity Penalty Concept 4.2-45
-------
FIGURES (continued)
Number
4.2-11
4.2-12
4.2-13
4.2-14
4.2-15
4.2-16
4.2-17
4.2-18
4.2-19
4.2-20
4.2-21
4.2-22
4.2-23
4.2-24
4.2-25
4.2-26
4.2-27
Energy Penalty for a Lime FGD System Utilizing Bypass
Heat at a Bituminous-Coal-Fired 500-MW Plant
Capacity Penalty'for a Lime FGD System at a Bituminous-
Coal-Fired 500-MW Plant
Capital Investment Excluding Cost of Sludge Pond and
Land for a Lime FGD System at a Bituminous-Coal-Fired
500-MW Plant
Capital Cost of Sludge Pond and Land for a Lime FGD
System at a Bituminous-Coal-Fired 500-MW Plant
Operation and Maintenance Cost Excluding Electricity
and Reheat for a Lime FGD System at a Bituminous-
Coal-Fired 500-MW Plant
Fixed Charges for a Lime FGD System at a Bituminous-
Coal-Fired 500-MW Plant
Diagram of a Typical Limestone FGD System
Actual and Projected Growth of Limestone and Other U S
FGD Capacity
Capital Cost of a Double Alkali FGD System on a Boiler
Firing 3.5 Percent Sulfur Coal
Annualized Costs of a Double Alkali FGD System on a
Boiler Firing 3.5 Percent Sulfur Coal
Basic Nonregenerable Sodium FGD System
Process Diagram of an FGD System, Nevada Power Co.,
Reid Gardner Station
FGD Capital Costs Versus Unit Size
FGD Annual Costs Versus Unit Size
Nonregenerable Ammonia-Based Process
Typical Wellman-Lord S02 Control System
Estimated Capital Cost for Wellman-Lord FGD Systems
Achieving 90 Percent S02 Removal Firing Either of
Two Eastern Coals
Page
4.2-47
4.2-48
4.2-54
4.2-55
4.2-56
4.2-57
4.2-61
4.2-63
4.2-78
4.2-79
4.2-81
4.2-83
4.2-87
4.2-88
4.2-94
4.2-105
4.2-119
-------
FIGURES (continued)
Number
4.2-28 Estimated Operating and Maintenance Costs for Wellman-
Lord FGD Systems Achieving 90 Percent S02 Removal
Firing Either of Two Eastern Coals
4.2-29 Typical Citrate S02 Control Process
4.2-30 Typical Citrate S02 Control System
4.2-31 Typical TCA Installation
4.2-32 Typical Ventri-Rod Scrubber With Mist Eliminators
4.2-33 MgO System With A Venturi Scrubber System
4.2-34 Rotary Calciner MgO Regeneration System
4.2-35 A Magnesium Oxide FGD System Using TCA Absorbers and
a Fluid-Bed Calciner
4.2-36 Capital Costs of MgO FGD Systems Given 90% S02
Removal Efficiency
4.2-37 Total Annual Operating Costs for Mgo FGD Systems Given
90% S02 Removal Efficiency
4.2-38 BF/FW Adsorption Process
4.2-39 Typical Spray Dryer Particulate Collection Flow Diagram
4.2-40 Nahcolite Dry Injection Flow Diagram
4.2-41 Once-Through S02 Reduction Versus Ca/S Molar Ratio
5.1-1 Section of Reverberatory Furnace
5.1-2 Peirce-Smith Copper Converter Operation
5.1-3 Total Capital Investment Costs of Double-Contact
Sulfuric Acid Plants - Dilute Feed Gas
5.1-4 Annual Operating Costs of Double-Contact Sulfuric
Acid Plants - Dilute Feed Gas
5.2-1 Simplified Flow Diagram of an Integrated Steelmaking
Facility Showing Only Major Sources of Sulfur Oxide
Emissions
4.2-120
4.2-125
4.2-126
4.2-137
4.2-138
4.2-142
4.2-143
4.2-144
4.2-152
4.2-153
4.2-157
4.2-166
4.2-168
4.2-183
5.1-4
5.1-7
5.1-27
5.1-28
5.2-2
vn
-------
FIGURES (continued)
5.2-3
5.2-4
5.3-1
5.3-2
5.3-3
5.4-1
5.4-2
5.4-3
5.5-1
5.5-2
5.5-3
5.5-4
5.5-5
5.5-6
5.5-7
5.6-1
5.6-2
Representation of the Overall Sulfur Balance at One
Steel Plant Producing 908 Gg (1 million tons) of
Ingots Annually
Flow Diagram of a Gray Iron Foundry
Limestone Scrubbing Process for Steel Mill Sinter
Plant Application
Major Processing Steps in a Typical Petroleum Refinery
Sulfur Recovery Unit
Process Layout of the Venturi Scrubbing System
Generalized Flow Diagram of Natural Gas Processing
Flow Diagram of the Amine Process for Gas Sweetening
Capital Cost of Claus (two-stage) Plus Incinerator
Flow Diagram of a Single-Absorption, Contact-Process
Plant that Produces Sulfuric Acid by Burning Sulfur
Flow Diagram for Ore-Roasting Contact Plant
Volumetric and Mass S02 Emissions From Contact Sulfuric
Acid Plants
Process Flow Diagram of a Double-Absorption, Contact-
Process Plant that Produces Sulfuric Acid by Burning
Sulfur
Flow Diagram for Ammonia Scrubbing of Sulfuric Acid
Plant Tail Gas
Capital Costs of S02 Control Systems for Domestic
Sulfuric Acid Plants
Operating Costs of S02 Control Systems for Domestic
Sulfuric Acid Plants
Typical Kraft Sulfate Pulping and Recovery Process
Showing Potential Emission Sources
Typical Magnesium-Based Chemical Pulping Recovery
Process Showing Potential Emission Sources
5,2-3
5.2-12
5.2-20
5.3-2
5.3-9
5.3-14
5.4-2
5.4-4
5.4-10
5.5-4
5.5-6
5.5-7
5.5-13
5.5-18
5.5-26
5.5-28
5.6-2
5.6-15
vm
-------
FIGURES (continued)
Number
5.8-1 Typical Flow Diagram for the Manufacture of Soda-Lime
Glass
5.8-2 End-Port Continuous Regenerative Furnace
5.8-3 Side-Port Continuous Regenerative Furnace
5.8-4 Typical Venturi Scrubber System
5.8-5 Dry Sorbent System
5.8-6 Nucleator
5.8-7 Reported and Estimated Installed Costs of Scrubber
Control Systems on Glass Furnaces
5.9-1 Process Flow Diagram for Lime Production
5.9-2 Process Flow Diagram for Cement Manufacturing
5.9-3 Process Flow Diagram for Clay and Brick Production
5.10-1 Generalized Explosive Manufacturing Process
5.10-2 Sulfur Acid Recycle System
5.11-1 Ortho-Xylene-Based Phthalic Anhydride Process
5.11-2 Ethylene From Ethane
5.11-3 Ethylene From Liquid Feeds
5.8-5
5.8-7
5.8-8
5.8-16
5.8-19
5.8-21
5.8-24
5.9-2
5.9-5
5.9-7
5.10-3
5.10-6
5.11-3
5.11-7
5.11-8
-------
TABLES
2.2-la
2.2-lb
2.2-2
2.2-3
2.3-1
2.3-2
3.1-2
3.2-1
3.2-2
3.4-1
3.4-2
3.5-1
3.5-2
3.5-3
Nationwide SOV Emission Trends, 1970-1977
(Tg/yr) x
Nationwide SO Emission Trends, 1970-1977
(106 short tons/year)
Nationwide SOX Emission Estimates, 1977
Summary of 1977 SO Emissions From Fuel Combustion
By Fuel Type
Automated Equivalent Methods for Ambient Sulfur
Dioxide Monitoring
Current Performance Specifications for Continuous
Monitoring System and Equipment
Percent of Total Energy Consumed for Selected
Individual Fuels and Purchased Electricity, by
Manufacturers
Air Quality Increments for the Prevention of
Significant Deterioration
Major Sources Subject to PSD Review
Comparison of Waste Liquor with Drinking Water
Criteria
Range of Concentrations of Chemical Constituents
in FGD Sludges From Lime Limestone, and Double-Alkali
Systems
Annual Capacities of All U.S. Producers of Virqin
Acid in 1979
Annual Capacities of All U.S. Producers of Smelter
Acid in 1979
Annual Capacities of Major U.S. Producers of Frasch
Sulfur in 1973
X
r ajjc
2.2-2
2.2-2
2.2-3
2.2-4
2.3-3
2.3-7
3.1-7
3.2-5
3.2-8
3.4-2
3.4-4
3.5-3
3.5-4
3.5-7
-------
TABLES (continued)
Number Page
3.5-4 Annual Capacities of the Five Largest U.S. Producers
of Recovered Sulfur in 1973 3.5-7
4.2-1 Demonstrated Coal Reserve Base by Rank and Potential
Method of Mining, January 1, 1976 4.2-2
4.2-2 Demonstrated Coal Reserve Base by Sulfur Content and
Potential Method of Mining, January 1, 1976 4.2-3
4.2-3 Estimates of Recoverable Reserves of Raw Coal
Characterized by S02 Emission Rate from Uncontrolled
Combustion 4.2-4
4.2-4 S02 Emissions from Burning Different Coals 4.2-6
4.2-5 Potential for Reducing Emissions by Physical
Desulfurization 4.2-13
4.2-6 Potential for Reducing Emissions by Chemical
Desulfurization 4.2-14
4.2-7 Summary of Physical Coal Cleaning Plant Costs 4.2-15
4.2-8 Major Coal Cleaning Process Data . 4.2-16
4.2-9 Coal Gasification Systems 4.2-19
4.2-10 Operating FGD Systems: June 1979 4.2-22
4.2-11 Categorical Results of the Reported and Adjusted
Capital and Annual Costs for Operational FGD
Systems 4.2-27
4.2-12 Sensitivity Analysis of a 500-MW.Lime FGD System
Capital Investment 4.2-52
4.2-13 Sensitivity Analysis of a 500-MW Lime FGD System
Annualized Costs 4.2-53
4.2-14 Capital and Annualized Costs of Operational Lime
FGD Systems 4.2-58
4.2-15 Costs of Limestone and Lime FGD Systems 4.2-66
4.2-16 Capital and Annualized Costs of Operational Limestone
FGD Systems 4.2-68
-------
TABLES (continued)
Number
4.2-17
4.2-18
4.2-19
4.2-20
4.2-21
4.2-22
4.2-23
4.2-24
4.2-25
4.2-26
4.2-27
4.2-28
4.2-29
4.2-30
4.2-31
4.2-32
Capital and Annualized Costs of Utility Double Alkali
FGD Systems
Industrial Sites Using Nonregenerable Sodium FGD
Technology
Costs of Nonregenerable Sodium-Based FGD Systems
in Industrial and Utility Applications
Water Pollution Impacts of a Nonregenerable Sodium
System
Summary of Estimated Capital and Operating Costs for
a Nonregenerable, Ammonia-Based Process with Ammonium
Sulfate Production
Summary of Estimated Capital and Operating Costs for
an Ammonia-Based FGD Facility on a New 500-MW Coal-
Fired Power Unit
Summary of Estimated Capital Cost of Flue Gas
Desulfurization Processes
Wellman-Lord Plant Installations in the United States
Wellman-Lord Plant Installations Overseas
Design Parameters for Wellman-Lord FGD Installations
at San Juan Station of Public Service Company of
New Mexico
FGD System Economics: Operational Systems
Energy Penalties Associated with Wellman-Lord SOo
Controls
Energy Penalties Associated with Wellman-Lord S02
Controls
Citrate FGD Process Units
Citrate FGD Process Capital Summary for a Coal-Fired
Power Plant, 2.5% Sulfur
Citrate FGD Process Annualized Operational Cost
Summary for a Coal-Fired Power Plant, 2.5% Sulfur
Paqe
4.2-77
4.2-85
4.2-89
4.2-91
4.2-99
4.2-100
4.2-102
4.2-111
4.2-112
4.2-116
4.2-121
4.2-123
4.2-124
4.2-129
4.2-130
4.2-131
xii
-------
TABLES (continued)
Number
4.2-33 Operating and Planned Magnesia Scrubbing Units on
U.S. Power Plants as of September 1978 - '
4.2-34 Operating Magnesia Scrubbing Units on Japanese Power
Plants as of 1978
4.2-35 Costs Associated with Three Magnesium Oxide Units
4.2-36 Costs of an MgO Absorption System at 90% Removal
Efficiency ; .
4.2-37 Capacity and Energy Penalties for Various FGD and
Fuel Types
4.2-38 Energy Requirements of the BF/FW System for a 500-MW
Plant .
4.2-39 Capital and Operating Costs of a 500-MW BF/FW System
4.2-40 Summary of Key Features of Dry FGD Systems
4.2-41 Summary of Key Features of Commercial Spray Drying
Systems
4.2-42 Estimated Costs of Two-Stage Dry Removal for 35 Years
4.2-43 Washability Data for High Sulfur Coals
5.1-1 Summary of Process Equipment at Copper Smelters in the
United States
5.1-2 Smelting and S02 Emission Control Practice
5.1-3 Summary of Capital and Annual Costs for Various FGD
Systems at Primary Copper Smelters
5.1-4 Estimated Capital and Annual Operating Costs of Options
for S02 Removal at Bunker Hill Smelter, Kellogg, Idaho
5.2-1 Sulfur Dioxide Emissions from Soaking Pits and Reheat
Furnaces
5.2-2 Capital Cost of Limestone Scrubbing System on an
Existing Sinter Plant Producing 6312 Mg/day
(5731 tons/day)
4.2-146
4.2-149
4.2-149
4.2-151
4.2-154
4.2-159
4.2-161
4.2-162
4.2-174
4.2-176
4.2-179
5.1-15
5.1-19
5.1-29
5.1-31
5.2-10
5.2-21
xm
-------
TABLES (continued)
5.2-4
5.2-5
5.3-1
5.3-2
5.3-3
5.3-4
5.3-5
5.3-6
5.4-1
5.5-1
5.5-2
5.5-3
5.5-4
5.5-5
5.5-6
5.5-7
Total Annual Operating Cost of Limestone Scrubbing
on an Existing Sinter Plant Producing 6312 Mq/day
(5731 tons/day) y
Estimated Capital and Operating Costs of Coke-Oven
Gas Desulfurization Systems
Net Energy Requirements for Coke-Oven Gas Desulfur-
ization Processes
Annual Costs for Claus Sulfur Recovery Plants with
Incineration
Costs for Claus Sulfur Recovery Plants with Wellman-
Lord Emission Control System (Oxidation)
Costs for Claus Sulfur Recovery Plants with Beavon
Emission Control System (Reduction)
Energy Impact of Emission Control Systems
Calculated S02 Emissions from Claus and SCOT Units
Potential Water Pollution Impact of Refinery Sulfur
Plant with Various Tail-Gas Treating Units
Typical SCOT Unit Costs
U.S. Sulfuric Acid Capacity
Uncontrolled Sulfur Dioxide Emissions from Single
Absorption Sulfuric Acid Plants
Status of Double Absorption Systems on Domestic Sulfuric
Acid Plants
Status of Ammonia Absorption Systems on Domestic Sulfuric
Acid Plants
Status of Wellman-Lord Systems on Domestic Sulfuric Acid
Plants
Status of Adsorption Systems on Domestic Sulfuric Acid
Plants
Status of Limestone Systems on Domestic Sulfuric Acid
Plants
Page
5.2-22
5.2-23
5.2-25
5.3-21
5.3-22
5.3-23
5.3-25
5.3-26
5.3-28
5.4-9
5.5-2
5.5-8
5.5-14
5.5-19
5.5-22
5.5-23
5.5-24
xiv
-------
TABLES (continued)
Number
5.6-1 Typical Emission Concentrations and Rates for SOX from
Kraft Pulp Mill Combustion Sources
5.6-2 Kraft Process Equipment Data
5.6-3 Typical S02 Emission Factors for Sulfite Pulp Mill
Sources
5.6-4 Sulfite Process Equipment Data
5.6-5 NSSC Process Equipment Data .
5.6-6 Criteria and Annual Operating Costs of a Sulfur Dioxide
Recovery/Control System on a Sodium-Based Sulfite
Recovery Boiler
5.7-1 Actively Burning Coal Mining Waste Piles in the
United States
5.7-2 Relative Costs of Demonstrated Methods of Extinguishing
Coal Mining Waste Fires
5.8-1 Glass Manufacturing Industry
5.8-2 Raw Materials Used in Manufacturing Soda-Lime Glass
5.8-3 Estimates of Annual SO Emissions from Glass Manufacture
/\
5.8-4 Emission Data from Site Testing
5.8-5 Summary of SO Emission Data Supplied by Glass
Manufacturers
5.8-6 Emission Factors for Glass Manufacturing Procedures
5.8-7 Demonstrated S02 Removal Efficiencies of Venturi
Scrubber Systems in the Glass Industry
5.8-8 Emission Data for Commercial Dry Sorbent Systems Using
Tesisorb Additives
5.8-9 Operating Parameters for Model Glass Plants
5.8-10 Performance and Cost Data for Emission Control Systems
at a Glass Furnace Producing 1.57 kg/s (150 tons/day)
5.9-1 Lime Kiln Modeling Parameters
5.6-5
5.6-6
5.6-9
5.6-10
5.6-12
5.6-18
5.7-1
5.7-6
5.8-2
5.8-3
5.8-9
5.8-10
5.8-11
5.8-13
5.8-17
5.8-20
5.8-22
5.8-25
5.9-3
xv
-------
TABLES (continued)
Number
5.9-2
5.10-1
5.11-1
5.12-1
5.12-2
SO Emission Factors for Brick Manufacturing Without
Controls
Emission Factors for Manufacture of TNT and NC
Uncontrolled Gas Emissions Vented from Phthalic
Anhydride Production Processes Using Ortho-Xylene
as Feed
Sulfur Oxide Emission Factors for Refuse Incinerators
S02 Emission from Controlled Sewage Sludge Incinerators
Page
5.9-8
5.10-8
5.11-5
5.12-3
5.12-5
xvi
-------
SECTION 1
INTRODUCTION
The first edition of "Control Techniques for Sulfur Oxide Air Pollu-
tants" was published in January 1969.1 The many changes and advances in
sulfur oxide emission control technology, as well as changes in fuel use
patterns since that time, have rendered many parts of the original document
obsolete. This second edition contains up-to-date information on available
control techniques, including cost, energy requirements, and environmental
impact of sulfur oxide control technology, as required in Section 108(b) (1)
of Public Law 91-604 of the Clean Air Act as amended August 7, 1977.
Sulfur oxides in the atmosphere are known to have many adverse effects
upon health and welfare, and reduction of emissions of this class of pollu-
tants is of prime importance to any effective air pollution abatement pro-
gram. Sulfur oxide pollutants originate from a variety of sources, and the
emissions vary widely in concentration and characteristics. Similarly, the
available control techniques vary in type, application, effectiveness, and
cost. The term "sulfur oxides" will be used when discussing emissions in
general, and "S02" will be used when referring to sulfur dioxide specifi-
cally. Sulfur oxides are emitted mainly as sulfur dioxide (S02) and to a
lesser extent as sulfur trioxide (S03), sulfuric acid mist (H2S04), and
sulfates. This document deals mainly with sulfur dioxide emissions control.
As in the first edition, the control techniques described herein repre-
sent a broad spectrum of information from many engineering and other tech-
nical fields. Many of the emission control devices, methods, and principles
have been developed and used over several years, and much experience has
been gained in their application. Discussions of methods still in various
stages of research and development are included to provide information about
the latest concepts under consideration, even though they are not yet avail-
able for general use.
1-1
-------
The many commercial, industrial, and municipal processes and activities
that generate sulfur oxide air pollutants are described individually. The
various techniques that can be applied to control emissions of sulfur oxides
from these sources are reviewed and compared. No attempt is made, however,
to review all possible combinations of control techniques that might bring
about more stringent control of each individual source.
The proper choice of a method, or combination of methods, to be applied
to any specific source depends on many factors in addition to the character-
istics of the emission stream. State or local governments may impose addi-
tional emission restraints; control regulations may also depend on relative
distance to a Class I Air Quality Control Region, whether the area is a non-
attainment area, and other considerations.
Some data are presented on quantities of sulfur oxides emitted to the
atmosphere; the effects of sulfur oxides on health and welfare are con-
sidered in a companion document, "Air Quality Criteria for Sulfur Oxides."2
Periodically the EPA publication "Compilation of Air Pollutant Emission
Factors" (AP-42) is referred to in this document. The user of AP-42 should
be aware that it is constantly being updated as new test data become avail-
able; therefore, the reader must be certain to utilize the most recent ver-
sion of AP-42.
The general 'organization of this report is as follows. Section 2
discusses the characterization and sampling methods for sulfur oxides.
Section 3 presents information on energy use patterns, determination of
emission reduction needs, technical considerations, environmental and energy
impacts of sulfur oxide control, and the sulfur products market. Sections 4
and 5 present information on the combustion process and industrial process
emissions of sulfur oxides respectively. These two chapters are presented
in the following sequence:
Process description and emissions
0 Control techniques
0 Control costs
Energy and environmental impacts.
1-2
-------
The costs discussed in this document are in mid-1979 dollars unless
specifically noted otherwise. Because of its wide applicability, the
Chemical Engineering plant cost index was used to update costs. A mid-1979
value of 236.5 was used. The Chemical Engineering index is given in every
issue of Chemical Engineering, a magazine published biweekly by McGraw-Hill,
Inc., of New York, New York.
This document was prepared primarily from information on sulfur oxiae
emissions found in the open literature and was supplemented by personal con-
tacts.
REFERENCES FOR SECTION 1
U.S. Department of Health, Education,
for Sulfur Oxide Air Pollutants.
January 1969.
and Welfare. Control Techniques
Public Health Service. AP-52.
U.S. Department of Health, Education, and Welfare.
Criteria for Sulfur Oxides. Public Health Service. AP-50.
(Document presently being revised.)
Air Quality
April 1970.
1-3
-------
-------
SECTION 2
THE SULFUR OXIDES: CHARACTERIZATION AND SAMPLING METHODS
This chapter addresses how SO are formed, the trends of sulfur oxide
s\
emissions, and both ambient and stack sampling equipment for measurement of
sulfur oxides.
2.1 CHARACTERIZATION OF SULFUR OXIDES
This section briefly defines and characterizes the sulfur oxides and
the reactions in which they are formed. Major sources and their contri-
butions to nationwide SO emission totals are discussed.
/\
Approximately 95 percent of the total pollution-related sulfur oxide
emissions is in the form of sulfur dioxide, most of the remainder being
sulfur trioxide, small amounts of sulfuric acid mist, and sulfates.1,2
2.1.1 Sulfur Dioxide
The primary source of S02 is combustion of fuels that contain sulfur.
During combustion, atmospheric oxygen combines with the sulfur in the
following reaction:
S + 02 •* S02
Sulfur dioxide is a colorless, nonflammable acidic gas. The odor threshold
for sulfur dioxide is 0.3 to 1.0 part per million (ppm) by volume; at con-
centrations above 3 ppm the odor is pungent.3
Formation of S02 occurs early in the primary flame at rates comparable
with the other combustion reactions. Formation will occur even in fuel-rich
flames; no practical combustion control techniques have been identified.2
The degree of photochemical or catalytic conversion of S02 into its deriva-
tives depends on atmospheric conditions such as humidity, catalytic sites,
concentrations of hydrocarbons, and particulate matter composition in an
effluent gas stream -or in the atmosphere. The reaction kinetics of S02 in
2.1-1
-------
the atmosphere are outlined in a literature review (through 1970) ' by
Bufalini.4
2.1.2 Sulfur Tri oxide
Sulfur tri oxide (S03) is formed from combustion sources directly or
from oxidation of atmospheric S02 as follows:
S0
1/20
Limitations in the chemical kinetic rates are such that only from 1 to 5
percent of the sulfur in the stack gases is observed in practice to be
present as S03.2 Sulfur trioxide is highly hygroscopic and normally com-
bines with water vapor in the stack to form sulfuric acid as a finely
divided aerosol:
S03 + H20 •* H2S04
Formation of S03 is found to occur only in air-rich mixtures and to be
governed by kinetic processes more amenable to combustion control.2
2-1-3 Sulfuric Acid Mist and Sul fates
Sulfuric acid and sul fates are formed by oxidation of sulfur dioxide by
several mechanisms, most involving reactive agents such as photochemical
smog, ammonia, catalytic metals, and fine particulates. Temperatures and
humidity also influence the reaction. Sulfuric acid can be found as a gas
phase component or a condensed liquid aerosol, and it can also be adsorbed
on carboneous particulate matter. In addition, free acid may react with
metal oxides formed in the combustion flame to yield sulfates.5 Sulfuric
acid and its reaction products are primary sul fate emissions, contrasted
with secondary sulfate derived from the transformation of S02 in the atmos-
phere.
Several factors influence the nature and extent of primary sulfate
emissions. These include fuel characteristics, boiler design and operation,
and emission controls. The extent of sulfate emissions can be affected by
boiler design parameters, including the number and type of burners, resi-
dence time and temperature distribution, and the amount of internal surface
area. Measurements have indicated that for a given fuel sulfur content, the
2.1-2
-------
total sulfate emissions from oil-fired sources are from 3 to 10 times
greater than from sources burning coal and contain a large fraction of free
sulfuric acid.5
Atmospheric sulfur dioxide may be oxidized to S03 and converted to
sulfuric acid aerosol, or it may form sulfite ions that are then oxidized to
sulfate. Subsequent to the oxidation, sulfuric acid or sulfate may interact
with other materials to form other sulfate compounds. Sulfate formation
rates are usually enhanced by increases in humidity. The mechanisms by
which sulfur dioxide is oxidized to sulfates are not well understood but are
important because they determine the formation rate and, to some extent, the
final form of sulfate.6 Most mechanisms involve reactive agents such as
photochemical smog, ammonia, catalytic metals, and fine particulates. These
agents can complicate the relationship between S02 and sulfates; for
example, reductions or increases in S02 concentrations may not result in
proportional reductions or increases in sulfate levels because of the other
agents that affect the formation reaction.7 Reduction in ambient sulfate
concentration will result ,in a concommitant reduction in particulate matter
concentration to the extent that sulfates are particulate matter.
2.1-3
-------
REFERENCES FOR SECTION 2.1
1.
2.
3.
4.
5.
6.
7.
U.S. Environmental Protection Agency. Position Paper on Regulation of
Atmospheric Sulfates. Research Triangle Park, N.C. EPA-450/2-75-007
September 1975. p. 13.
U.S. Environmental Protection Agency. Workshop Proceedings on Primary
Sulfate Emissions from Combustion Sources. Research Triangle Park
N.C. EPA-600/ 9-78-020b. Volume 2. August 1978. pp. 3, 14.
U.S. Environmental Protection Agency, Air Pollution Technical Informa-
tion Center. National Air Pollution Control Administration Air Quality
Criteria for Sulfur Oxides. PHS Publication No. AP-50. Research
Triangle Park, N.C. 1969. pp. 5-10.
Bufalini, M. Oxidation of Sulfur Dioxide in Polluted Atmospheres—A
Review. Environmental Science Technology. 5(8):685. August 1971.
Ref. 2, pp. 4-13.
Ref. 1, pp. 22-24.
Ref. 1, p. x.
2.1-4
-------
2.2 SULFUR OXIDES EMISSION TRENDS
Total SO emissions have been relatively stable over the first half of
/\
this decade. Ambient SO levels decreased in urban areas in the early
1970's, probably because of increasing use of fuels with lower sulfur
content and a general shift of sulfur-emitting sources away from urban
areas. After an initial improvement, the SO levels stabilized as the
y\
National Ambient Air Quality Standards were achieved in many areas. Since
reaching a low in 1975, however," SO emissions have risen, as seen in Tables
/\
2.2-1 (a and b);1 The increases are due to greater fuel use and to the
displacement of natural gas by oil and coal.
Sulfur oxides, primarily S02, are generated during combustion of any
sulfur-bearing fuel and also in many industrial processes that use sulfur-
bearing raw materials. Bituminous coal and residual fuel oil usually con-
tain 1 to 3 percent sulfur by weight. Ordinary combustion of fossil fuels
(at normal levels of excess air) forms S02 and S03 at a ratio of 30 to 1;
when power plants are operated with controlled reaction conditions, the
ratio of S02 to S03 is generally about 60 to I.2
This section presents a summary of the 1976/1977 estimated S0x emis-
sions by sector and fuel type. It also presents an estimate of SO emission
density for each county in the United States.
2.2.1 1977 SO Emissions
)\
Table 2.2-2 presents the estimated SO emissions in 1977. These esti-
mates are based on published data describing fuel use and industrial produc-
tion and on other data from the Environmental Protection Agency (EPA) de-
scribing emission factors and emissions. The transportation category in
Table 2.2-2 includes emissions from all mobile sources. Mobile sources
include aircraft, trains, shipping, and miscellaneous sources. Stationary
fuel combustion is defined as fuel used in nonmobile combustion equipment
such as boilers and stationary internal combustion engines. Emissions are
shown for electric utility power plants, industrial establishments, and
other fuel consumers (residential, commercial, governmental, and education-
al). Industrial processes include emissions from the operation of process
equipment by manufacturing industries. Solid waste includes emissions from
2.2-1
-------
TABLE 2.2-1 a. NATIONWIDE SOX .EMISSION TRENDS, 1970-19771
(Tg/yr)
Source category
Transportation
Stationary fuel combustion
Industrial processes
Solid waste
Miscellaneous
1970
0.7
22.6
6.3
0.1
0.1
29.8
1971
0.7
21.6
5.8
0.1
0.1
28.3
1972
0.7
22.0
6.7
0.1
0.1
29.6
1972
0.7
23.1
6.3
0.1
0.0
30.2
1972
0.7
22.1
5.6
0.0
0.0
28.4
1975
0.7
20.8
4.6
0.0
0.0
26.1
1976
0.8
21.9
4.5
0.0
0.0
27.2
1977
0.8
22.4
4.2
0.0
0.0
27.4
Note: A value of 0.0 indicates emission of less than 0.05 Tg/yr.
TABLE 2.2-lb. NATIONWIDE SOX EMISSION TRENDS, 1970-19771
(106 short tons/year)
Source category
Transportation
Stationary fuel combustion
Industrial processes
Solid waste
Miscellaneous
•
1970
0.8
24.9
6.9
0.1
0.1
32.8
1971
0.8
23.8
6.4
0.1
0.1
31.2
1972
0.8
24.3
7.4
0.1
0.1
32.0
1973
0.8
25.5
6.9
0.1
0.0
33.3
1974
0.8
24.4
6.2
0.0
0.0
30.7
1975
0.8
22.9
5.1
0.0
0.0
28.8
1976
0.9
24.1
5.0
0.0
0.0
30.0
1977
0.9
24.7
4.6
0.0
0.0
30.2
Note: A value of 0.0 indicates emission of less than 55,000 short tons/yr.
2.2-2
-------
TABLE 2.2-2. NATIONWIDE SOX EMISSION ESTIMATES, 1977
Transportation (total)
Highway vehicles
Nonhighway vehicles
Stationary fuel combustion (total)
Electric utilities
Industrial
Residential, commercial, and institutional
Industrial processes (total)
Chemicals
Petroleum refining
Metals
Mineral products
Oil and gas production and marketing
Industrial organic solvent use
Other processes
Solid waste
Miscellaneous
Forest wildfires
Agricultural burning
Coal refuse burning
Structural fires
Miscellaneous 'organic solvent use
Total
Tg/yr
0.8
0.4
0.4
22.4
17.6
3.2
1.6
4.2
0.2
0.8
2.4
o.e
0.1
0
0.1
0
0
0
0
0
0
0
27.4
106 tons/yr
0.88
0.44
0,44
24.69
19.40.
3.53
1.76
4.63
0.22
0.88
2.65
0.66
0.11
0
0.11
0
0
0
0
0
0
0
30.20
Zero indicates emissions of less than 0.05 Tg (55,000 tons) per year.
2.2-3
-------
the combustion of waste in municipal and other incinerators and from the
open burning of domestic and municipal refuse. Miscellaneous sources in-
clude emissions from the combustion of forest, agricultural, and coal
refuse, and from structural fires.3
Table 2.2-3 presents the 1977 S0x emissions from fuel combustion by
sector and fuel type.
TABLE 2.2-3.
SUMMARY OF 1977 S0v EMISSIONS FROM FUEL
COMBUSTION BY FUE^ TYPE4
Sector
Electric generation
Industrial
Commercial /Institutional
Residential
Percent by fuel type
Coal
68.5
7.6
0.5
4.3
Oil
7.7
6.1
3.7
1.0
Gas
<0.1
0.6
<0.1
<0.1
As shown in Figure 2.2-1, 22.4 Tg (24.7 million tons) or approximately
82 percent of the national total was produced by fuel combustion in station-
ary sources. This category covers all fuel use in stationary combustion
FUEL COMBUSTION
STATIONARY SOURCES
22.4 Tg (24.6 million tons)
TOTAL: 27.4 Tg/yr
(30.2 x 106 tons/yr)
OTHER 0.8 Tg
(0.9 million tons)
INDUSTRIAL PROCESSES
EXCLUDING FUEL
COMBUSTION
4.2 Tg
(4.6 Billion tons)
Figure 2.2-1. Nationwide sources of sulfur oxide emissions, 1977.
2.2-4
-------
equipment such as boilers and stationary internal combustion engines; it
encompasses electric utility power plants, industrial establishments, and
other fuel consumers (residential, commercial, governmental, and institu-
tional). Industrial processes generated an additional 4.2 Tg (4.6 million
tons). The remainder, 0.8 Tg (0.9 million ton), was emitted by miscella-
neous sources such as burning coal refuse banks, agricultural burning,
combustion of fuels for transportation, and disposal of solid wastes. It
should be noted that sulfur oxide emissions by source in any given locality
may differ markedly from those in Figure 2.2-1.
2.2.2 Emission Density of SO Emissions ,in the United States
This section describes the geographical variation in emission density
across the continental United States. Figure 2.2-2 is a map of the United
States with each county shaded according to its estimated S0x emission
density. These data represent the emission density at the midpoint of the
decade, the base year being 1975. The highest S0x emission densities are
found in the Northeast, where there is heavy usage of fossil fuels con-
taining sulfur compounds, and in several isolated counties in the West,
where many smelters are located. Approximately 26 percent of the total U.S.
population live in areas with SO emission densities exceeding 35 megagrams
/\
per square kilometer (100 tons per square mile). Over half of the popula-
tion live in areas with emission density greater than 3.5 megagrams per
square kilometer (10 tons per square mile); these areas represent 11 percent
of the land area of the continental United States.5
2.2.3 Trends in SO Emissions
. )\ '
Figure 2.2-3 represents the estimated S0x emissions from 1970 to 1977.
During this period the emissions decreased slightly. Emissions from elec-
tric utilities actually increased by 10 percent during this period. This
increase would have been substantially higher had it not been for the use of
fuels with lower sulfur content, because the amounts of coal and residual
oil burned during this period increased about 50 percent and 70 percent
respectively. Emissions of SO from industrial processes were significantly
~ )\
2.2-5
-------
to
to
<1J
<1>
x
o
S-
4-
o
to
CL)
to
c
o
£
c:
en
•p—
oo
CM
I
O)
zs
en
2.2-6
-------
Ill I I I
c
r^
01
t/i
-a
O)
CO
CM
10
V)
O)
X
o
00
O)
-o
_
O
•r—
(O
CO
O
1—
I I I I I
LO
CO
O
CO
Lf>
CM
O LO
CvJ i—
(suo^ ota^eui 9oi)
CM
a>
s-
3
CTl
2.2-7
-------
lower in 1977 than in 1970, mainly because of controls used by primary
nonferrous smelters and EPA regulations prescribing lower emissions from
sulfuric acid manufacturing plants.6
2.2.4 Nationwide Trends in SO Air Quality 1970-1977
X "
Sulfur dioxide levels in urban areas throughout the Nation have
gradually decreased since 1970. The 1972-77 trends show dramatic initial
improvement (see Figure 2.2-4) followed by fairly consistent continuing
improvement. In most urban areas, this is consistent with the switch in
emphasis from attainment of standards to maintenance of air quality; that
is, the initial effort was to reduce pollution to acceptable levels and
subsequent efforts have been to maintain air quality at these lower levels.
Sites providing data for these analyses were selected from EPA's
National Aerometric Data Bank. Sites for assessment of trends in the
1972-1977 time period were selected to ensure the historical completeness
and seasonal balance of data. For S02, 1233 sites recorded sufficient data
to qualify as trend sites.
Figure 2.2-4 illustrates nationwide trends in annual mean S02 levels
from 1972 through 1977. The graph shows that S02"levels' continued to im-
prove in the middle 1970's although the rate of improvement was much less
pronounced than earlier in the decade. From 1972 through 1977, the national
average S02 level dropped 17 percent, an annual improvement rate of 4 per-
cent per year. As would be expected, most sites showed improvement during
this period.
Short-term changes in S02 levels between 1976 and 1977 were mixed, with
no predominant trends. Most S02 monitors in urban areas reported levels
well below the annual standard. The remaining S02 problems are primarily
associated with specific point sources.7
2.2-8
-------
DO
o£ 48
3-
S 40
t— i
S
fe 32
UJ
o
0
"' 2*
o
»— 1
X
o
o 16
a:
u_
pj 8
0
-
-
_
"
'
it
S
*
I
I
72
H
h
>
<•
r
73
-Cl-^
••••^LH^
0-
>^-
lh
!>
*
r
n
c
>
>
*
n
»
x-
74 75 76
8
YEAR
90TH PERCENTILE
-
tt
0
X
]
-
—
77
•^ 75TH PERCENTILE
<
-<—
COMPOSITE AVERAGE
MEDIAN
25TH PERCENTILE
MOTH PERCENTILE
Figure 2.2-4. Nationwide trends in annual average sulfur dioxide
concentrations from 1972 to 1977 at 1,233 sampling sites.8
2.2-9
-------
REFERENCES FOR SECTION 2.2
3.
4.
5.
6.
7.
8.
U.S. Environmental Protection Agency, Office of Air Quality Planning
and Standards. National Air Quality, Monitoring, and Emissions Trend
Report, 1977. Research Triangle Park, N.C. EPA 450/2-78-052.
December 1978. pp. 5-5 to 5-12.
U.S. Government Printing Office, National Industrial Pollution Control
Council. Air Pollution by Sulfur Oxides. Washington, D.C. 1971.
Ref. 1, pp. 5-2, 5-3.
U.S. Environmental Protection Agency. Office of Air Quality Planning
and Standards. Research Triangle Park, N.C. OAQPS 1977 Data File.
Computer run date of March 27, 1979
Ref. 1, p. 6-3.
Ref. 1, p. 5-4.
Ref. 1, p. 3-7.
Ref. 1, p. 3-8.
2.2-10
-------
2.3 SAMPLING AND ANALYSIS METHODS7"
This section deals with manual and automated methods for sampling and
analysis of sulfur oxides in ambient air and at emission sources.
Ambient air monitoring is conducted for several reasons. Some state
regulations require that industries monitor the atmosphere for certain
pollutants specific to their source, and some states permit ambient air
monitoring in lieu of compliance with a specified emission limitation.
Ambient monitoring also provides information that is useful in determining
the normal levels of air quality, in following air quality trends, and in
providing guidance for emergency control procedures during air pollution
episodes.
Measurement of pollutant emissions in the stack of an industrial pro-
cess is referred to as stack gas sampling, source testing, or emissions
testing. Such measurements may be used to determine whether a source is
complying with emission regulations and to determine the efficiency of an
emission control device. Where a process design must be altered to correct
a pollution problem, source test data may provide a basis for determining
what changes will be most effective.
2.3.1 Ambient Monitoring
Ambient monitoring for oxides of sulfur, primarily S02, may be per-
formed by using either manual methods or continuous analyzers. Manual
methods involve operators collecting and analyzing samples at various steps
of the analysis; these methods yield values that are integrated over speci-
fied sampling periods. Even though many steps in the analysis may be auto-
mated, the procedure is still considered a manual method. In continuous S02
monitoring all sampling and analysis are performed without direct involve-
ment by an operator; the measurements are made continuously on a real-time
basis.
2.3.1.1 Manual Methods—
The EPA has designated the manual pararosaniline method as its refer-
ence method for analysis of S02 in ambient air.1 This colorimetric proced-
ure, developed by West and Gaeke in 1956,2 utilizes gas bubblers as the S02
collection system. Basically, the S02 is scrubbed into a potassium tetra-
2.3-1
-------
chloromercurate (TCM) solution to form a stable dichlorosulfitomercurate
complex, which then undergoes reaction with pararosaniline and formaldehyde
to form the colored complex of pararosaniline methylsulfonic acid. The
colored complex is measured colorimetrically to determine the S02 value.
These values are reported as an average number of micrograms of S02 per
cubic meter of air passing through the bubbler during the sampling period,
which is normally 24 hours.
2.3.1.2 Continuous Monitoring--
Automated methods for continuous ambient monitoring of sulfur oxides
must be EPA reference or equivalent methods. The definition and qualifica-
tions of reference and equivalent methods are given in Federal regulations
(40 CFR 53). Table 2.3-1 lists the analyzers that have been designated as
equivalent methods in accordance with 40 CFR 53. Each of these analyzers
must be used in strict accordance with the operation or instruction manual.*
The Lear Siegler Model SM 1000 S02 Ambient Monitor is based on second-
derivative spectroscopy. The narrow-band absorption of electromagnetic
radiation exhibited by S02 in the ultraviolet region produces an output
signal that is processed to give S02 concentration readings. A second-
derivative spectrometer processes the transmission-versus-wavelength func-
tion of an ordinary spectrometer to produce an output signal proportional to
the second derivative of this function. This technique enhances the sensi-
tivity of an ordinary spectrometer. The SM 1000 S02 Ambient Monitor has
been developed specifically to measure S02 gas concentrations in parts per
billion.
Both the Meloy Model SA185-2A Sulfur Dioxide Analyzer and the Monitor
Labs Model 8450 Sulfur Monitor use flame photometry, in which a hydrogen
flame excites the S02 molecules and causes the emission of light. A
narrow-band filter selects the light emitted at 394 nanometers (nm), which
A current list_of analyzers designated as reference or equivalent methods
Tor SOg is available from the U.S. Environmental Protection Agency, Envi-
ronmental Monitoring and Support Laboratory, Dept. E, MD-76, Research
Triangle Park, North Carolina 27711
2.3-2
-------
TABLE 2.3-1. AUTOMATED EQUIVALENT METHODS
FOR AMBIENT SULFUR DIOXIDE MONITORING3
Equivalent analyzer
Principle of operation
Lear Siegler Model SM 1000
S02 Ambient Monitor
Meloy Model SA 185-2A
Sulfur Dioxide Analyzer
Monitor Labs Model 8450
Sulfur Monitor
Thermo Electron Model 43
Pulsed Fluorescent S02 Analyzer
Beckman Model 953
Fluorescent Ambient S02 Analyzer
Philips PW 9700 S02
Analyzer
Philips PW 9755 S02
Analyzer
ASARCO Model 500
Sulfur Dioxide Monitor
Bendix Model 8303
Sulfur Analyzer
Meloy Model SA 285-E
Sulfur Dioxide Analyzer
Second derivative
Ultraviolet absorption
Flame photometry
Flame photometry
Fluorescence spectrometry
Fluorescence spectrometry
Coulometric
Coulometric
Conductimetric
Flame photometry
Flame photometry
Listing issued by Environmental Monitoring and Support Laboratory, EPA.
February 22, 1979
2.3-3
-------
Is then detected by a photomultiplier tube. The resultant photomultipl ier
signals are amplified and processed to give readings of S02 concentration.
The Thermo Electron Model 43 Pulsed Fluorescent S02 Analyzer and the
Beckman Model 953 Fluorescent Ambient S02 Analyzer use fluorescence spectro-
metry. Fluorescence is defined as the emission of light absorbed by a
molecule; the emitted light is of a different wavelength from that of the
absorbed light. In the fluorescence technique the S02 molecule is irradi-
ated with light at a given wavelength (usually in the near-ultraviolet
region), and the emitted light is measured at a longer wavelength. The
resultant fluoresence, directly proportional to the number of S02 molecules
present, is then measured by a photomultiplier tube. The photomultiplier
signals are amplified and processed to give readings of S02 concentrations.
The commercially available instruments contain either a continuous light
source (used in the Beckman model) or a pulsed ultraviolet light source
(used in the Thermo Electron Model).
The Philips PW 9755 and PW 9700 S02 Analyzers operate on the coulo-
metric principle. The analyzer measures the current generated in an elec-
trochemical reaction such as
2Br •* Br2 + 2e~.
Sulfur dioxide affects this reaction in the following manner:
S02 + 2H20 + Br2 -*• H2S04 + 2HBr.
The instrument measures the change of current flow due to the change in the
rate of Br2 generation caused by the presence of S02. The change in current
is processed to give readings of S02 concentration.
The ASARCO Model 500 Sulfur Dioxide Monitor senses the change in elec-
trical conductivity in water when a soluble substance such as S02 is dis-
solved in it. The change of conductivity is proportional to the concen-
tration of S02 added and is easily measured to give -S02 concentration
readings.
2.3.2 Source Monitoring
Source monitoring for oxides of sulfur may be performed by using manual
methods or automated continuous analyzers. Manual methods that may be used
2.3-4
-------
are EPA Method 6 and EPA Method 8. EPA Methods 6 and 8 are used as the
manual reference methods for determination of compliance with NSPS. Auto-
mated continuous analyzers are usually used as indicators of continued
complying operation. Two exceptions are the primary lead and electric
utility NSPS, where the instrumental method is required to generate compli-
ance data. Automated continuous analyzers fall into two main classes:
extractive and in-situ systems. Remote monitoring systems have been
developed, but they will not be discussed because they are mainly experi-
mental and still in the development stage.
2.3.2.1 Manual Methods—
The EPA Method 6 is the required reference method for determining emis-
sions of S02 from stationary sources (except sulfuric acid plants) subject
to New Source Performance Standards.3 Many states have adopted Method 6 for
all stationary sources. Both Methods 6 and 8 (discussed later) are valid
for any source emitting S02 and/or sulfuric acid mist.
In sampling for S02 using Method 6, a gas sample is taken at a single
sampling point located at the center of the stack or no closer to the wall
than 1 meter (3.28 feet). The sample must be extracted at a constant rate.
The collected gases are bubbled through solutions contained in
impingers, which trap sulfur oxides in a soluble and stable form for sub-
sequent analysis. As the gas goes through the sampling apparatus, the
sulfuric acid mist and S03 are removed in the first impinger containing 80
percent isopropyl alcohol, the S02 is removed by a chemical reaction with a
3 percent hydrogen peroxide solution contained in the two subsequent
impingers, and the sample gas volume is measured. The sulfuric acid mist
and S03 are discarded, and the collected material containing the S02 is
recovered for laboratory analysis. The concentration of S02 in the sample
is determined by titrating the sample with barium perch!orate in the
presence of the indicator thorin.
Method 6 requires a sampling time of 20 minutes per sample, and two
separate samples constitute a run. The method specifies three runs,
resulting.in six separate samples. The method can determine stack concen-
trations of 50 to 10,000 ppm of S02.
2.3-5
-------
The EPA Method 8 is the reference method for determining emissions of
sulfuric acid mist (including S03) and S02 from sulfuric acid plants.4 The
collection system utilizes an 80 percent isopropyl alcohol solution in the
first impinger and hydrogen peroxide in the two subsequent impingers. The
impingers are similar to those in Method 6, but in Method 8 the isopropyl
alcohol water solution that traps the sulfuric acid mist and S03 is also
titrated with barium perchlorate to determine the sulfuric acid and S03
concentrations.
2.3.2.2 Continuous Methods-
Continuous emissions analysis consists of extractive and in-situ moni-
toring systems. Extractive systems consist of removing a continuous sample
from the gas stream, conditioning the sample (removal of particulate and
excess water vapor), and transporting the sample stream to a remotely
located analyzer for analysis. Remote signifies only that the analyzer is
located outside the duct. Actual distance may be as little as 3m (10 ft) or
greater than 100 m (300 ft).
In-situ systems analyze the gas as it exists in the stack or duct,
generally by some advanced spectroscopic technique. The analyzers are
installed across a stack (cross-stack) or employ a sensor inserted into the
flue gas stream (in-stack). Both extractive and in-situ monitoring systems
used for source emissions measurements must meet EPA performance specifica-
tions at the time of installation and throughout their operation. A summary
of the current specifications is presented in Table 2.3-2.
Continuous methods using extractive systems—The extractive instruments
in use today for S0x monitoring utilize nondispersive infrared or ultra-
violet techniques, fluorescence, flame photometry, and electrochemical
methods.
Nondispersive infrared and ultraviolet techniques are based on light
absorption spectroscopy. Light is passed through a gas, .and the degree of
absorption of certain infrared or ultraviolet wavelengths is measured. For
S02, the sensitivity of nondispersive infrared (NDIR) instruments is usually
about 10 ppm. The sampling system used in tandem with an NDIR analyzer
2.3-6
-------
TABLE 2.3-2. CURRENT PERFORMANCE SPECIFICATIONS FOR
CONTINUOUS MONITORING SYSTEM AND EQUIPMENT5
Parameter
Specification
Accuracy
Calibration error9
Zero drift (2-hour)a
Zero drift (24-hour)a
Calibration drift (2-hour)a
Calibration drift (24-hour)a
Response time
Operational period
> 20 percent of the mean value of
the reference method test data
> 5 percent of each (50 percent,
90 percent) calibration gas
mixture value
2 percent of span
2 percent of span
2 percent of span
2.5 percent of span
15 minutes maximum
168 hours minimum
a Expressed as sum of absolute mean value plus 95 percent confidence
interval of a series of tests.
2.3-7
-------
contains filters that remove participates, sulfuric acid mist, and water
from the flue gas stream. Nondispersive ultraviolet (NDUV) analyzers,
sometimes called differential absorption analyzers, are more sensitive and
are not subject to the same interferences (e.g., H20, C02). They are,
however, heavy and bulky. Some manufacturers of NDUV monitors are DuPont,
CEA Instruments, Ester!ine-Angus, and Teledyne.
Fluorescence techniques are based 'on emission spectroscopy. Molecules
or atoms are energized by exposure to high-intensity ultraviolet radiation;
they emit light at specific wavelengths, and the amount of emitted light is
measured.
Ultraviolet fluorescence methods are used for analysis of S02 both in
ambient air and at stationary sources. Celesco Industries, Inc., manufac-
tures a fluorescence instrument with a continuously emitting UV light
source, whereas a unit made by Thermo Electron Corporation utilizes a pulsed
light source.
Flame photometry is another luminescence technique used to detect S02.
Instead of exciting the S02 molecules with an ultraviolet light source as in
fluorescence methods, flame photometry uses a hydrogen flame. The resulting
emitted light is measured and recorded as an S02 concentration. Flame
photometric analyzers are used primarily in ambient monitoring, but are
applied to stationary source monitoring by use of sample dilution systems.
Tracer, Melory Laboratories, Bendix Corp., Process Analyzers, and others
manufacture flame photometric analyzers.
Three classes of analyzers work on electrochemical principles: conduc-
timetric, coulometric, and polarographic (the electrochemical cell). The
conductimetric (electrical conductivity) principle was briefly discussed
earlier. Calibrated Instruments, Inc., (Mikrogas-MSK) produces a conducti-
metric analyzer that absorbs S02 in a hydrogen peroxide solution. Manufac-
turers of coulometric analyzers include Barton ITT and Beckman Instruments,
Inc. The coulometric analyzer can be operated unattended for extended
periods of time.
Polarographic analyzers measure the current induced by electrochemical
oxidation of S02 at a sensing electrode. Dynasciences, Beckman Instruments,
Inc., Theta Sensors, and Teledyne market polarographic analyzers.
2.3-8
-------
Automated methods using In-situ systems—In-situ, cross-stack, or in
stack monitors use electro-optical techniques based on infrared or ultra-
violet absorption. The four most common electro-optical principles used in
these instruments are 1) dispersive spectroscopy, 2) gas-filter correlation
spectroscopy, 3) dispersive correlation spectroscopy, and 4) second-deriva-
tive spectroscopy.
A dispersive absorption spectrometer can be set at any wavelength
within its range; in this it differs from a nondispersive instrument, which
looks at a broad spectral region and must be sensitized to detect each
particular gas by means of a detector cell. Environmental Data Corporation
currently markets an on-stack monitor utilizing this technique. Wilks
Scientific Corporation produces a series of portable dispersive infrared
analyzers that are applicable to stack gas monitoring and to analysis of
process streams and in-plant air.
A gas -filter correlation spectrometer discriminates between gases by
correlating the structure of the spectrum of the gas species to be measured
in the stack with the spectrum of the same species contained in a correla-
tion cell (gas cell). Correlation spectrometers have demonstrated potential
for stationary source monitoring. Environmental Research and Technology
manufactures an instrument that measures S02 by this technique.
Dispersive correlation spectroscopy combines the techniques of gas-
filtered correlation and dispersive in-situ monitors. Barringer Research,
Ltd., of Canada is actively developing correlation spectrometers and offers
units commercially for S02 source monitoring.
Second-derivative spectroscopy, discussed earlier, has only recently
been applied to the measurement of air pollutants. Lear-Siegler, Inc.,
markets a series of spectrometers based on this principle for S02 monitor-
ing.
Although in-situ monitors are relatively new, they are meeting perfor-
mance specifications and offer certain operational advantages for facilities
required to monitor source emissions continuously.
2.3-9
-------
REFERENCES FOR SECTION 2.3
1. U.S. Environmental Protection Agency. Reference Method for the Deter-
?nnrcD°cn °fA Su1^r Dl"oxide in the Atmosphere (Pararosanil ine Method).
40 CFR 50, Appendix A. July 1980.
2. West, P.W., and G.C. Gaeke. Fixation of Sulfur Dioxide as Disulfito-
(II) and Subsequent Colorimetric Estimation. Anal. Chem.
1956.
3. U.S. Environmental Protection Agency. Reference Method 6 - Determina-
tion of Sulfur Dioxide Emissions from Stationary Sources. 40 CFR 60
Appendix A. July 1980. '
4. U.S. Environmental Protection Agency. Reference Method 8 - Determina-
tion of Sulfuric Acid Mist and Sulfur Dioxide Emissions from Stationary
Sources. 40 CFR 60, Appendix A. July 1980. «i°nary
5* o;nVnEnV.1ronmental Protection Agency. Performance Specifications. 40
CFR 60, Appendix B. July 1980.
2.3-10
-------
SECTION 3
CONSIDERATIONS IN SULFUR OXIDES CONTROL
This section presents a general outline of items that should be evalu-
ated for sulfur oxide emission sources.
The first subsection presents information on energy availability and
usage trends in which energy policy is reviewed. The second subsection
discusses the determination of the emission reduction needs and includes
information on State Implementation Plans (SIP), New Source Performance
Standards (NSPS), Prevention of Significant Deterioration, and visibility
regulations, as well as an introduction to dispersion modeling; The third
subsection presents an outline of technical considerations in sulfur oxide
control. The fourth subsection discusses environmental and energy impacts
and technical considerations. The fifth subsection reviews the market for
sulfur and sulfur-related products that may be recovered and offered for
sale as products of some sulfur oxide emission reduction systems.
3.0-1
-------
-------
3.1 ENERGY AVAILABILITY AND USAGE TRENDS
The 1970's have brought soaring energy prices and the prospect of
future energy shortage. The oil embargo of 1973-74 pointed up an enormous
and growing energy demand, with which domestic fuel production has not kept
pace. As a result, alternative energy technologies are being sought to
reduce the dependence on energy from other sources. Short-term strategies
are being developed; their overall effect is intended to be a return to a
far greater dependence on fuels with proven domestic reserves, such as coal.
A move toward increasing use of coal brings special problems for in-
dustry. An increase in the use of coal for energy independence would neces-
sitate the installation of pollution control equipment, not only for clean-
ing the stack gas, but also for acceptable disposal of the tons of bottom
ash and fly ash generated in the firing of coal.
Combatting an energy shortage in an energy-intensive nation is a dif-
ficult problem. This section discusses some of the methods proposed to date
for coping with the situation.
3.1.1 U.S. Energy Policy
The availability and future usage of fuels in the United States is
being greatly influenced by a series of national acts aimed at stemming the
use of imported crude oil and natural gas. Reserves of petroleum have been
located in Alaska, but even they are insufficient to reduce the dependence
upon Middle East oil. As a result, several short-term plans have been
enacted to curb the rate of increase of petroleum fuel usage.
3111 The Energy Supply and Environmental Coordination Act (ESECA) of
1974 (PL 93-319) as Amended by the Energy Policy and Conserva-
tion Act (EPCA) (PL 94-163)—
As a direct consequence of the 1973 oil embargo, Congress passed ESECA
to institute a reduction of oil and natural gas usage by major combustors in
the United States. The purpose of the Act, as it relates to fuel usage, is
stated in a pamphlet issued by the Federal Energy Administration (FEA):
"ESECA is based upon expanding the use of U.S. reserves of coal, considering
the environmental consequences of that use, and recognizing that in some
cases, previous environmental requirements may unnecessarily preclude the
3.1-1
-------
use of coal."1 Knowing that the National Environmental Policy Act of 1969
would require extensive investigation regarding the effects of fuel changes
on air and water'quality, Congress required close coordination and coopera-
tion between the EPA and the FEA. The authority granted to the FEA under
ESECA consisted of the following:2
The FEA could—
Prohibit a power plant or major fuel burning installation (MFBI)
from burning petroleum or natural gas under certain defined condi-
tions,
Order power plants and MFBI's in the "early planning process" to
be constructed and designed for coal firing as their primary
energy source, and ^
Allocate coal supplies, if necessary, to implement their prohibi-
tion orders.
The EPA was included in the Act as a participant in the prohibition process
to ensure that the State Implementation Plan would be met assuming total
coal firing in a converted unit.
The authority of ESECA expired on June 30, 1977. During the 1974
through 1977 period, many plants were evaluated for their coal-firing
capabilities. Because of such factors as age, availability of fuel, size of
the unit, and environmental constraints, only a few plants were issued final
prohibition orders. A total of 58 coal conversion orders were issued as of
June 30, 1977, when the Act expired.3
In the Congress at that time were the rudiments of a new comprehensive
energy bill intended to supersede the ESECA authority and change the overall
context of the prohibition order process. The cabinet-level Department of
Energy was established in late 1977.
The Congress passed the National Energy Act on October 15, 1978. The
Act is composed of five parts:
The National Energy Conservation Policy Act of 1978,
The Power Plant and Industrial Fuel Use Act of 1978,
The Public Utilities Regulation Policy Act,
The Natural Gas Policy Act of 1978, and
0 The Energy Tax Act of 1978.
Although each of these acts has some indirect effect on the availability and
usage of fuels, it is the Power Plant and Industrial Fuel Use Act (FUA) that
3.1-2
-------
carries on the ESECA coal-firing strategies. Further prohibition of petro-
leum and natural gas usage under FDA will have the most direct impact on the
availability and utilization of the various fuel types in the industrial and
utility sectors.
3.1.1.2 The Fuel Use Act of 1978--
The Power Plant and Industrial Fuel Use Act of 1978, as quoted below,
has several provisions:4
Prohibition of New Oil and Gas-Fired Boilers
Prohibition against use of oil or natural gas in new electric utility
generation facilities or in new industrial boilers rated at 29.3 MW
(thermal) (with a fuel heat input rate of 100 million Btu's per hour or
greater), unless exemptions are granted by DOE.
Restrictions on Existing Coal Capable Large Boilers
DOE authority to require existing coal capable facilities, individually
or by categories, to use coal and to require noncoal capable units to
use coal-oil mixtures.
Restrictions on Users of Natural Gas for Boiler Fuel
Limitation of natural gas use by existing utility power plants to the
proportion of total fuel used during 1974-1976, and a requirement that
there be no switches from oil to gas. There is also a requirement that
natural gas use in such facilities cease by 1990 with certain excep-
tions. [If a power plant began operation on or after January 1, 1974,
the use of natural gas is limited to the average quantity of gas used
during the first 2 years of operation. Any new facility must be de-
signed to be coal-capable.]
Pollution Control Loan Program
An $800 million loan program to assist utilities to raise necessary
funds for pollution .control.
Supplemental Authority
Supplemental authority to prohibit use of natural gas in small boilers
for space heating and in decorative outdoor lighting and to allocate
coal in emergencies.
Other Provisions
Funding of several programs to reduce negative impacts from increased
coal production; energy impact assistance and railroad rehabilitation.
3.1-3
-------
A significant modification in the prohibition order process is that the
owner of an existing power plant or MFBI (or one that is proposed for con-
struction) must petition the DOE, stating the reasons for being exempted
from the oil prohibition. The petition is evaluated by the DOE's Economic
Regulatory Administration (ERA), which then issues a Fuels Decision Report.
This process requires considerable time and effort. It places the burden on
the petitioner rather than the United States Government. The DOE has issued
"interim final" rules for implementing the FUA. The final rules are to be
issued. A petitioner may be granted a temporary or permanent exemption from
coal firing for:
-
Thl'S exei»Pti°n
5°
Future use
combustion).
crude oil.
of innovative
be granted if the cost
technologies (such as f luidized-bed
Environmental constraints.
Public interest criteria.
Alternative fuels. A list of about 16 alternative fuels exists,
other than oil or natural gas cogeneration.
nS,!rJ!i«tUreS' KA fUB- Contai"nin9 at least 25 percent oil or
natural gas may be considered.
Site limitation.
USH-°f *ynVet1c fuel- Exemption may be granted for a
period pending development of the fuel.
0 Capital availability.
Product or process requirement.
utilitySunTtsk> rel1ab111ty» and intermediate-load exemptions for
3.1.2 Current Availability and Usage
The overall policy has not yet significantly altered the distribution
of fuels in use in the United States, however, as evidenced by Department of
Commerce data.
3.1-4
-------
The U.S. Department of Commerce, Bureau of the Census, compiles an
"Annual Survey of Manufacturers,"5 which presents information on the con-
sumption of fuels and electric energy. The survey issued in March 1978 is a
three-volume report presenting energy usage by all manufacturing industries
in 1976.6 The results of this study are summarized in Table 3.1-1. Both
consumption and cost of fuel have risen.
The relative distribution of fuel usage is presented in Table 3.1-2,
which shows the fuel trends for selected years since 1962.6 Although the
1977 survey is not yet available, it is anticipated that the use of coal
will have' increased slightly with decreasing oil usage. Table 3.1-2 shows
the decrease in natural gas usage; even with the recent release of gas
allotments to industry, it is expected that natural gas usage cannot in-
crease for a prolonged period.
Costs' of fuel, expressed in dollars per kilojoule (and dollars per
million Btu) have been tabulated and indexed by the Department of Commerce.
A composite cost chart is shown in Figure 3.1-1.7 The year 1971 is selected
to represent an index value of 100. The costs of all fuels have increased
sharply since 1971. The cost of natural gas increased nearly 40 percent
between 1974 and 1976. The relative cost of electricity as fuel has always
been high, since it is a secondary fuel.
In spite of the shortage of domestic crude oil, the Bureau of the
Census data show no indication of any radical curtailments due to the un-
availability of any fuel. No forced shutdowns of production or other criti-
cal actions of. any magnitude have been taken. It appears then, that the
energy supplies required by the industrial and utility sectors are available
under current conditions. It is generally believed, however,, that the
economic domination imposed by the Organization of Petroleum Exporting
Countries (OPEC) will tighten the reins on oil supplies and cause continual
increases, in energy cost. As a result, domestic users will be obliged to
switch to fuel types in reserve in the short term, and • we must undertake
concurrent research and development of future (long-term) technologies to
meet an increasing energy demand.
3.1-5.
-------
TABLE 3.1-1. QUANTITY AND COST OF PURCHASED FUELS AND ELECTRIC ENERGY
USED FOR HEAT AND POWER:
1976 AND SELECTED EARLIER YEARS BY MANUFACTURING INDUSTRIES^
Purchased fuels and
electric energy,
kHn x 101?
Btu x 10^
Total cost, $ x lOfi
Purchased fuels.
kWh x 1012
Btu x 1012
Total cost, $ x 106
Fuel oil,
Residual.
103 barrels
Cost, S x 106
Distillate,
106 m3
103 barrels
Cost, $ x 106
Bituminous coal, lig-
nite, and anthra-
cite,
Tg
103 short tons
Cost, S x 106
Coke and breeze
Tg
103 short tons
Cost, S x 106
Natural gas0
10s ro3
109 ft3
Cost, S x 10^
Other fuels*1
S x 106
Fuels not speci-
fied by kind
$ x 10°
Electric energy,
Purchased,
106 kWh
Cost, $ x 106
Generated
less sold
106 kWh
1958
2.415
8,247.8
5,067.0
2.162
7,384.9
2,836.2
b
b
b
26.44b
166, 301. Ob
522. 7 b
74.19
81,784.0
638.2
12.32
13,585.0
271.0
88.14
3,112.2
900.9
147.9
355.6
252,909.0
2,230.8
66,850.0
1962
2.873
9,810.5
6,184.1
2.559
8,739.2
3,360.7
23.99
' 150,885.0
432.4
1.111
44,730.0
190.9
81.14
89,438.0
639.5
16.10
17,747.0
304.8
122.0
4,308.1
1.-455.9
337.1
' 0.0
313,961.0
2.823.3
74,261.0
1967
3.458
11.810.3
7,691.7
3.031
10,351.7
3.974.9
17.96
112,958.9
298.7
10.44
65,653.9
236.9
68.13
75,100.0
551.7
12.30
13,562.5
248.9
150.3
5,306.9
1,749.1
220.2
669.4
27,465.1
3,716.8
78,355.8
1971
3.807
13.002.3
10,432.1
3.329
11,370.2
5,360.6
22.37
140,726.4
535.9
16.68
104,940.8
453.4
55.69
61,392.6
658.1
12.47
13,742.8
317.6
182.8
6,454.4
2,559.9
377.5
458.2
514,612.7
5,070.6
82,828.0
1974
3.956
13, 509. £
19, 433. *
3.345
11, 424. E
10,963.
27.20
171,095.3
1,964.9
18.10
113,823.6
1,350.4
43.37
47,806.8
1,083.1
13.80
15,215.2
744.6
186.0
6,566.4
4,360.1
702.6
778.1
611,094.5
8,449.4
80.932.3
•
1975s
3.527
i 12,044.4
23,237.2
2.936
10,026.7
12,951.1
3.697
12,625.3
27.5S&.9
3.062
10,458.9
i 15.5C5.1
I
28.21 | 36.51
177,452.4 j 229,614.6
2,149.0 2,718.2
16.68
104,894.2
1,369.7
40.48
44,623.3
1,310.3
14.92
93,836.1
1,292.5
43.38
47,817.1
1,341.6
11.94
13,156.8
880.8
164.4
5,804.8
5,653.1
849.5
" ' •'!
738.7
91,342.5
10,286.3
63,275.0
14.21
15,665.7
1,141.2
167.2
5,902.7
7,535.5
906.7
569.2
634,934.6
12,081.9
64,571.1
8 Revised.
For 1958. figures are combined for residual and distillate.
* For 1967 and earlier; Includes manufactured, still, blast-furnace, and coke-oven gas.
For 1971 and later; Includes gas (except natural).
3.1-6
-------
TABLE 3 1-2 PERCENT OF TOTAL ENERGY CONSUMED FOR SELECTED
INDIVIDUAL FUELS AND PURCHASED ELECTRICITY,
BY MANUFACTURERS6
Fuel or electricity item
___^_^^«*—
TOTAL
Residual fuel oil
Bituminous coal, lignite, and
anthracite
Coke and breeze
Distillate fuel oil
Natural gas
Purchased electricity
Other fuels
Data for fuels not specified by kind are distributed among detailed fuels,
Entries may not add to 100 because of independent rounding.
3.1-7
-------
INDEX
370 (350)
317 (300)
264 (250)
NATURAL GAS
DISTILLATE FUEL OIL
COKE AND BREEZE
RESIDUAL FUEL OIL
COAL
PURCHASED ELECTRICITY
211 (200)
158 (150)
106 (100)-
53 (50)
1967
1971
1974 1975 1976
Figure 3.1-1. Cost per kilojoule (British thermal unit) of
selected fuels and purchased electricity consumed by all
manufacturing industries 1976, 1975, 1974, 1971, and 19677
(1971 = 100)
3.1-8
-------
3.1.3 Future Energy Use
The major factors affecting the future growth of the boiler population
are the economic growth of the nation, technological advancements in energy
production and use, fuel use patterns, and energy and environmental regula-
tory trends.
Projections of fuel usage in the future are given in reports prepared
by the Edison Electric Institute (EEI) and the Energy Information Adminis-
tration (EIA). Results of the reports are included in the following subsec-
tion.
3.1.3.1 The EEI/EIA Projections--
The report issued by the EEI in 19768 analyzes the growth of energy use
as it relates to economic growth. The report does not attempt to factor in
the effects of forced fuel conversion on usage trends. Several scenarios
are given to represent interactions of nine separate elements: (1) popula-
tion, (2) agriculture, (3) growth of income and consumption, (4) mineral
demand and supply, (5) energy demand and supply, (6) conservation and envi-
ronment, (7) pricing policies, (8) capital requirements, and (9) relations
with the rest of the world. By varying some of these elements, EEI formu-
lates three scenarios of energy-demand growth: Case A--high economic
growth; Case B--moderate economic growth; Case C--low (,or no) economic
growth. Case A is predicted to result in a 4 percent annual increase in
U.S. energy demand by the industrial sector to the year 2000; Case B is
predicted to result in a 3 percent annual increase; and Case C is predicted
to result in only a 0.5 percent annual increase.
The EEI executive summary states: "To achieve greater energy indepen-
dence, it will be necessary to make a basic shift from oil and natural gas
to coal and nuclear fuels. Development of facilities to liquefy and gasify
coal will complement this shift by enlarging the areas of consumption to be
served by coal. Ultimately, oil and gas consumption will be limited by
market pressures to those uses where coal and nuclear power are not feas-
ible substitutes such as for petrochemical raw materials."9
3.1-9
-------
The Energy Information Administration (EIA) issues an annual report to
Congress. On the basis of its close contact with the DOE and the overall
energy picture with respect to the National Energy Plan, the EIA has specu-
lated on the future effects of the FUA in projecting total energy require-
ments by industry. A preliminary summary release of the annual report gives
the following forecasts:10
Coal Production is projected to grow from 695 million tons [0.63 Tg] in
1977 to 13 to 1.6 billion tons [1.2 to 1.5 Pg] per year by 1990 with
growth rate of western coal about four times' greater thar i eastern
om ^i1 deCll"ne Sl19htly over the mid-term (1990)
from 8.8 million barrels daily (MMBD) [1.4 million mVdavl to
between 5.9 and 8.3 MMBD [0.9 and 1.3 million mVday]^n 1990
consumPti°n will grow at an annual rate ranging
to 2 t I?01 2'8 PSrCent between 1977 and 1990> as compared
to 2.6 percent annually over the 1962 to 1977 period.
and 1ndustry will shift to coal and away from oil
DHces ri,a reSUJ- °f ^ .Natl'°nal Ener^ Act and r^ w°rld
ft'om 17 Potion of industrial energy consumption will
n electric ut mf ^ ™l t0 between 18 and 23 Percent in 199°-
oercent 7n iQQn ' COaL S Share w111 ^c^ase to between 52 and 58
percent in 1990, as compared to 45 percent in 1977.
i !°.ri> faster than the rate of
nan^i electriclty Prices increasing more slowly and
natural gas prices more rapidly than the prices of other fuels.
The further possible effects of NSPS for large coal-fired utility boilers
and possible NSPS for industrial and commercial boilers are not factored
into the projections.
3.1.4 Conclusions
Replacement of fuels of low sulfur content with coal will unavoidably
increase the generation of sulfur oxides. In the past, the sulfur oxide
control strategies could require the use of low sulfur fuels. With both the
current energy situation and recent energy legislation, the control of S02
emissions by switching to lower sulfur fuels, such as natural gas and oil,
is not a viable option.
3.1-10
-------
REFERENCES FOR SECTION 3.1
1. Federal Energy Administration. Implementing Coal Utilization ^Provi-
sions of Energy Supply and Environmental Coordination Act. Washington,
D.C. April 1976. p. 1.
2. Ref. 1, p. 2.
3. Coal Outlook. Washington, D.C., Observer Publishing, Company. May 1,
1978. pp. 4, 5.
4. U.S. Department of Energy, Office of Public Affairs. The National
Energy Act. Washington, D.C. November 1978. p. 6.
5. U.S. Department of Commerce, Bureau of the Census. Annual Survey of
Manufacturers. Washington, D.C. March 1978.
6., Ref. 5, p. 11.
7. Ref. 5, p. 6.
8. Edison Electric Institute. Economic Growth in the Future—The Growth
Debate in National and Global Perspective. New York. . 1976.
9. Ref. 8, p. 9.
10. U.S. Department of Energy. Department of Energy Information—Weekly
Announcements. Washington, D.C. 3(21):1. May 22, 1979.
3.1-11
-------
-------
3.2 DETERMINING EMISSION REDUCTION NEEDS
This section summarizes the various air programs that may require
control of sulfur oxides (S0v). It briefly discusses each program, its
/\
emission reduction requirements, and its current status. It also provides
information on the dispersion models available for estimation of the air
quality impacts of sulfur dioxide (S02) emission sources.
3.2.1 Emission Regulations
This discussion deals with the basic regulations that apply to S02
emissions. These regulations are the State Implementation Plans (SIP's) and
the New Source Performance Standards (NSPS). The SIP's are designed to
attain and maintain the National Ambient Air Quality Standards (NAAQS) and
to prevent the significant deterioration (PSD) of air quality. In addition,
a discussion is provided concerning the future regulations that will be.
promulgated regarding the SIP requirements to protect and enhance visi-
bility.
3.2.1.1 State Implementation Plans for the attainment and maintenance of
the NAAQS--
The Clean Air Act of 1970 gave the Environmental Protection Agency
(EPA) the responsibility and legal authority to control air pollution in the
United States. Among the many responsibilities given to EPA was the estab-
lishment of NAAQS for those pollutants "the emissions of which, in the
Administrator's judgement, cause or contribute to air pollution which may
reasonably be anticipated to endanger public health or welfare; and the
presence of which in the ambient air results from numerous or diverse mobile
or stationary sources."1 One of the pollutants included in this group was
sulfur oxides measured as sulfur dioxide.
The Clean Air Act of 1970 also mandated that these NAAQS must be
attained as expeditiously as practicable and that each state develop, adopt,
and submit to EPA for approval a plan that provided for the attainment,
maintenance, and enforcement of the NAAQS in every air quality control
region (AQCR), as designated by EPA under Section 107(C) of the Act.
In developing its plan, each state determined (on the basis of current
air quality levels) the degree of emission reduction required to attain or
maintain the NAAQS within all areas of the state. Additionally, the state
determined which air pollution sources must be controlled and to what
3.2-1
-------
extent, to accomplish the necessary emission reductions. The SIP then set
forth the necessary emission limitations and timetables for compliance to
ensure attainment and maintenance of the NAAQS.
Since the primary responsibility for development and enforcement of the
SIP fell to the states and each plan was tailored to an individual state,
the S02 regulations varied from state to state. They varied both in the
units of measure in which the limitation on sulfur or S02 was expressed and
in the equipment to which the regulations apply. In addition, some states
have a uniform regulation for all sources of either combustion or process
emissions, whereas other states have different emission limitations for
various sources according to the fuel used, type of material processed,
geographic location, size of the source, or type of source.2
In general, process regulations are of five types: (1) pounds of S02
emitted per hour; (2) pounds of total sulfur feed, expressed in pounds of
sulfur per hour or pounds of sulfur emitted per pound of input sulfur; (3)
pounds of S02 per ton of product; (4) flue gas concentrations of S02 (e.g.
ppm); and (5) ground-level concentrations of S02 (e.g., ppm).3
Sulfur dioxide emissions from fuel combustion are usually regulated
either by limiting the amount of sulfur or S02 emitted per unit heat input
(nanograms S02/joule, pounds S02/million Btu) or by limiting the sulfur
content of the fuel by weight. Sulfur dioxide emissions are also limited by
restricting the flue gas concentration of S02 in parts per million or grains
per cubic foot or by limiting the amount of S02 emitted per hour. A few
states specify a ground-level or ambient concentration of S02 that cannot be
exceeded. Additionally, a few states require a percent reduction of input
sulfur and application of "reasonable or best available control technology"
or "new proven technologies."
Some states or territories enforce their fuel combustion regulations on
a boiler basis, others on a stack basis, and still others on a total plant
basis. Depending upon the regulation, a source may be able to average its
emissions over all boilers (or stacks) rather than ensuring that each boiler
complies with the regulation.
Some states regulate specific fuel types. Other states have specific
S02 regulations for various geographic areas. In some areas (e.g., Ohio)
regulations have been promulgated to apply to specific plants. In a few
3.2-2
-------
states, the size of the source determines whether the source must comply
with an S02 emission -limitation and also determines the stringency of the
limit. In most cases, source size is defined by the heat input rate
measured in megawatts thermal (millions of Btu per hour). Other means of
defining source size include kilograms (pounds) of steam generated per hour
and megagrams (tons) of S02 emitted per hour. In some states, larger
sources are controlled more stringently than smaller sources. Over half of
the states have regulations incorporating more than one of the parameters
discussed above. In addition, about 35 percent of the states have separate
regulations for new sources.4
Finally, only a few states limit the emissions or the fuel quality as a
maximum value averaged over a given time period. Most states indicate only
that the emissions or sulfur content shall not exceed a maximum value. This
type of regulation implies that compliance is instantaneous.4
In most areas, sources have complied with the S02 emission limitations,
and considerable progress has been made in reducing the ambient levels of
S02. In some areas, however, the national ambient air quality standards for
S02 are still being violated. Because of this continuing nonattainment of
S02 as well as other NAAQS's, Congress passed the Clean Air Act Amendments
of 1977. These amendments required that the states evaluate the current air
quality levels and pursuant to section 107(d), designate as "nonattainment
areas" those areas in which levels of air pollution are above the national
standards. Once these nonattainment areas were designated, the states were
to develop a plan for attaining these standards as expeditiously as
practicable, but no later than December 31, 1982. The attainment plans were
scheduled for submission to EPA by January 1, 1979, and for EPA approval by
July 1, 1979. If a state did not have an approved plan by July 1, 1979,
certain limitations on funding and new source growth were to be invoked
until the state developed and EPA approved an adequate plan. A list of
nonattainment areas was published on March 3, 1978, with subsequent modifi-
cations for which states developed their attainment plans.5
In some of these areas, compliance with existing regulations will
provide attainment of the national standards. In other areas more stringent
emission limitations will be needed to attain the standard by December 31,
1982, as the Act requires. Since many of these state plans have not been
3.2-3
-------
approved, the types of new limits that must be met are unknown at this time.
As these plans are approved, the existing state regulations summarized above
are certain to be modified.
In addition to the control requirements affecting existing sources, the
Clean Air Act Amendments of 1977 require the states to develop a permit
program for new sources. As a minimum, this program must assure that all
new sources with actual S02 emissions equal to or greater than 91 Mg (100
tons) per year and which cause or contribute to a violation of the NAAQS in
an area designated as nonattainment must apply the .Lowest Achievable Emis-
sion Rate (LAER). The LAER represents the most stringent emission limita-
tion that is contained in any SIP for that particular class or category of
source or the most stringent emission limitation achieved in practice by
that particular class or category of source. In addition, no new source
permit can be issued unless the state determines that by the time the new
source is to commence operation either of the following will have occurred:
(1) the total allowable emissions from existing and new sources will be
sufficiently lower than the total emissions from existing sources under the
applicable implementation plan prior to the application of the new sources
so as to represent reasonable further progress toward attainment or (2)
emissions from the new source will not cause or contribute to emissions
levels that exceed those allowed under the SIP for new source growth. Also,
the owner or operator of the new source must demonstrate that all major
sources owned and operated by him within the state in which he wishes to
construct are subject to emission limitations and are in compliance or on a
schedule for compliance. The state must also be carrying out the provisions
of the plan.6
3.2.1.2 State Implementation Plans for Prevention of Significant Deterio-
ration—
In 1974, EPA issued regulations under the 1970 Clean Air Act for the
prevention of significant air quality deterioration. These regulations
establish a program for protecting areas with air quality better than that
specified in the NAAQS.
Under EPA's regulatory program, clean areas of the nation could be
designated under any of three "Classes." .For particulate matter and sulfur
oxides, specified numerical "increments" of net air pollution increase are
3.2-4
-------
permitted under each class up to a level considered to be "significant" for
that area. Class I increments permit only minor air quality deterioration;
Class II increments permit moderate deterioration; Class III increments
permit deterioration to the level of the secondary NAAQS.
The EPA initially designated all clean areas of the Nation as Class II.
States, Indian tribes, and officials having control over Federal lands
(Federal land managers) were given authority to redesignate their lands to
Class I or III status under specified procedures.
The initial area classification scheme was administered and enforced
through a program of preconstruction and premodification permits for 19
specified types of stationary air pollution sources. No such air pollution
source could begin construction or modification unless EPA (or a state) had
found that the source's emissions would not exceed the numerical "incre-
ments" for the applicable Class and that the source would use best available
control technology (BACT). The permit program applied to sources that had
not "commenced construction," as defined in the regulation, by June 1,
1975.7
The 1977 Amendments to the Clean Air Act essentially ratified,
extended, and generally made more stringent the PSD provisions promulgated
in 1974. Basically the new Amendments require classification of all areas
as Class I, II, or III. The air quality in each of these areas is allowed
to deteriorate only by a specific amount or increment. The increments for
each Class are presented in Table 3.2-1.8 Except for the Class I areas
listed in Figure 3.2-1 a"d the Northern Cheyenne Indian Reservation, the
entire country is designated as Class II.
TABLE 3.2-1. AIR QUALITY INCREMENTS FOR THE PREVENTION
OF SIGNIFICANT DETERIORATION8
(ug/m3)
S02 annual
24-hour
3-hour
TSP annual
24- hour
Class I
2
5
25
5
10
Class II
20
91
512
19
37
Clas.s III
40
182
700
37
75
3.2-5
-------
The 1977 Amendments designate the following areas as Class I areas and
prohibit redesignation into either of the other two classes: 1) interna-
tional parks, 2) national wilderness areas exceeding 20 million m2 (5000
acres), 3) national memorial parks exceeding 20 million m2 (5000 acres), and
4) national parks exceeding 24 million m3 (6000 acres). Figure 3.2-1 is a
map of the Class I areas.
The 1977 Amendments also required that each new or modified major
emitting facility or major stationary source obtain a preconstruction
permit. These amendments defined a major emitting facility as a stationary
source of air pollutants in any of 28 specific source categories (listed in
Table 3.2-2) that emit or have the potential to emit 91 Mg (100 tons) per
year or more of any pollutant regulated under the Clean Air Act and as any
other source (not specifically listed) that has the potential to emit 227 Mg
(250 tons) per year or more of any pollutant regulated under this act.
On June 19, 1978, the EPA published regulations to implement the 1977
Amendments. These regulations state that no major stationary source may be
constructed unless the following criteria are met: a permit is issued to
that source; the owner or operator of the source demonstrates that the-
emissions from the operation will not cause or contribute to air pollution
levels in excess of: maximum allowable increases (i.e., the increments for
TSP and S02 established under Section 163 of the Clean Air Act), NAAQS in
any region, or other applicable emission standards or standards of perform-
ance under the Clean Air Act; the proposed source is subject to the Best
Available Control Technology for each pollutant it emits, which is subject
to regulation under the Clean Air Act; and the owner or operator agrees to
conduct such monitoring that may be necessary to determine what effect
emissions of this proposed facility may have on air quality.
The regulations defined potential to emit as the capability at maximum
capacity to emit a pollutant in the absence of air pollution control equip-
ment. Annual potential will be based on the maximum rated capacity of the
source unless the source is subject to enforceable permit conditions that
limit the annual hours of operation. Enforceable permit conditions on the
type or amount of materials combusted or processed may be used in deter-
mining the potential emission rate of a source. The regulations also
3.2-6
-------
si SSf^---.-^
.' K9 f "^Ke
1.
2.
3.
4.
5.
i.
7.
8.
t.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
Olympic
Pasayton
North Cascades
Glacier Park
Alpine Lakes
Mount Ranier
Goat Rocks
Mt. Adacs
Mt. Hood
Eagle Cap
Mt. Jefferson
Mt. Washington
Three Sisters
Diamond Peak
Strawberry Mtn.
Crater Lake
Kal»iopsis
Mountain Lakes
Redwood
Gearhart Mtn.
Marble Mtn.
Lava Beds
South Warner
Thousand Lakes
Las sen
Caribou
Yolla Bolly Middle Eel
Desolation
Pt. Reyes
Mokeluane
Emigrant
Hoover
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
Yoseaite
Pinnacles
Ventana
Kaiser
Kings Canyon
Sequoia
Minarets
John Muir
San Rafael
DOM Land
Cucaaonga
San Jacinto
San Gabriel
San Gorgonio
Joshua Tree
Aqua Tibia
Selway-Bitterroot
Hell 's Canyon
Sawtooth
Craters of the Moon
Jarbridge
Cabinet Mtns.
Glacier
Mission Mtn.
Bob Marshall
Medicine Lake
Scapegoat
Gates of the Mountain
UL Bend
Anaconda Pintler
Red Rock Lake
North Absaroka
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.
93.
94.
95.
96.
Yellowstone
Washakie
Grand Teton
Fitzpa trick
Bridger
Capitol Reef
Bryce Canyon
Zion
Arches
Canyonlands
Grand Canyon
Sycaaore Canyon
Petrified forest
Pine Mt.
Manual
Sierra Ancha
Mt. Baldy
Superstition
Galiuro
Saguaro
Chiricahua
Mt. Zirkel
Flat Tops
Rawah
Rocky Mtn.
Eagles Nest
Maroon Bells Snowaass
West Elk
Black Canyon
la Garita
Great Sand Dunes
Weainuche
97.
98.
99.
100.
101.
102.
103.
104.
105.
106.
107.
108.
109.
110.
111.
112.
113.
114.
115.
116.
117.
118.
119.
120.
121.
122.
123.
124.
125.
126.
127.
Mesa Verde
Wheeler
San Pedro Parks
Pecos
Bandelier
Bosque del Apache
Salt Ck.
White Mtn.
Gila
Carlsbad Caverns
Guadalupe
Big Bend
Lostwood
Theodore Roosevelt
Badlands
Wind Cave
Witchita Mtns.
Voyageurs
Boundary Waters
Canoe Area
Isle Royale
Mingo
Hercules Glades
Upper Buffalo
Caney Creek
Seney
Maaaoth Cave
Great Saokey Mtns.
Joyce Kilaer-Slickrock
Sipsey
Cohutta
Okefenokee
128.
129.
130.
131.
132.
133.
134.
135.
136.
137.
138.
139.
140.
141.
142.
143.
144.
145.
146.
147.
148.
149.
150.
151.
152.
153.
154.
155.
156.
St. Marks
Chassahowitzka
Breton
Everglades
Wolf Island
Cape ROM in
Shining Rock
Linville Gorge
Swanquarter
Jaaws River Face
Shenandoah
Brigantine
Dolly Sods
Otter Creek
Lye Brook
Great Gulf
Presidential Range
Dry River
Acadia
Moosehorn
Bering Sea
Siaeonal
Mt. McKinley
Tuxedni
Haleakala
Hawaii Volcanoes
Rainbow Lake
Brodwell Bay
Roosevelt Caapobello
International Park
Virgin Islands
Figure 3.2-1. Mandatory Class I areas for PSD.
3.2-7
-------
TABLE 3.2-2. MAJOR SOURCES SUBJECT TO PSD REVIEW
Specific sources with potential emissions >91 Mg/yr (TOO tons/yr)
Fossil fuel-fired steam electric plants
[>73 MW thermal (>250 million Btu/h)]
Coal-cleaning plants (thermal dryers)
Kraft pulp mills
Portland cement plants
Primary zinc smelters
Iron and steel mill plants
Primary aluminum ore reduction plants
Primary copper smelters
Municipal incinerators [2.6 kg/s (>250 tons/day)]
Hydrofluoric acid plants
Sulfuric acid plants
Nitric acid plants
Petroleum refineries
Lime plants
Phosphate rock processing plants
Coke oven batteries
Sulfur recovery plants
Carbon black plants (furnace process)
Primary lead smelters
Fuel conversion plants
Sintering plants
Secondary metal production facilities
Chemical process plants
Fossil-fuel boil.ers [>73 MW thermal (>250 million Btu/h)]
Petroleum storage and transfer facilities
[capacity >47,700 m3 (>300,000 bbl)]
Taconite ore processing facilities
Glass-fiber processing plants
Charcoal production facilities
Any other source with potential emissions >227 Mg/yr (>250 tons/yr)
3.2-8
-------
provide that only certain sources would receive a full PSD review. The
detailed review would only be provided to those sources with allowable
emissions equal to or greater than 45 Mg (50 tons) per year, 455 kg (1000
pounds) per day, or 45 kg (100 pounds) per hour and to those sources that
would impact a Class I area or an area where the increment is known to be
violated. The detailed review includes an assessment to ensure that the
source has applied BACT and that the source will not violate any applicable
increment or NAAQS.8
Best available control technology is determined on a case-by-case basis
for each pollutant regulated under the Act and must represent an emission
limitation based on the maximum degree of reduction (taking into account
energy, environmental, and economic impacts, and other costs). In no event
shall application of BACT result in emissions that would exceed those
allowed under an applicable NSPS or National Emission Standards for
Hazardous Air Pollutants (NESHAPS). The NESHAPS are set for those pollu-
tants to which no ambient air quality standard is applicable and which, in
the judgment of the Administrator, cause or contribute to air pollution that
may reasonably be anticipated to result in an increase in mortality or an
increase in serious irreversible or incapacitating reversible illness.
The 1977 Amendments also provide protection for Class I areas in addi-
tion to the increments. The protection of "air quality related values" is a
large factor in determining whether a source may be granted a permit if it
would impact a Class I area. If the Federal land manager responsible for
the Class I area demonstrates that emissions from the proposed new source
would have an adverse impact on "the air quality related values" of the area
(even if the Class I increments would not be exceeded), a permit would not
be issued. Some special exemptions from this review process are given in
the regulations of June 19, 1978. Visibility is an example of an air
quality related value. It is likely that the potential adverse impact on
visibility will act as the primary triggering mechanism for a review of air
quality related values; other values, such as vegetative impacts and clim-
atological change, may also be identified.9
Many industrial and environmental groups petitioned the United States
Court of Appeals of the District of Columbia Circuit to review substantia-
tive portions of the June 1978 revised regulations. On June 18, 1979, in
3.2-9
-------
Ihe Alabama Power Company v. Costle (13 ERC 1225), the Court issued a deci-
sion that upheld some of the provisions of the June 1978 regulations and
overturned others. In its opinion, the Court summarized its ruling and
promised a supplemental comprehensive opinion at a later date. (The court
issued its final opinion on December 14, 1979.) On September 5, 1979,
following the court's mandate, the EPA proposed certain changes in the June
19, 1978, regulations to make the PSD requirements consistent with the June
1979 summary decision in Alabama Power. The proposed changes included:
Potential to emit
Fugitive emissions
Major modification
Preconstruction notice
Baseline definition
Ambient monitoring
De minimi's levels
Each of these changes under the proposed regulations is discussed below.
Potential emissions would be determined after application of emission
controls and would be calculated using maximum annual rated capacity, year-
round hours of operation, and any enforceable permit condition on the mate-
rial combusted or processed.
Fugitive emissions would be excluded from a source's annual potential
emissions unless these emissions are from the industrial source categories
listed below:
Coal-cleaning plants
Kraft pulp mills
Portland cement plants
Primary zinc smelters
Iron and steel mill plants
Primary aluminum ore reduction plants
Primary copper smelters
Municipal incinerators
Hydrofluoric acid plants
Sulfuric acid plants
Nitric acid plants
Petroleum refineries
Lime plants
Phosphate rock processing plants
Coke oven batteries
Sulfur recovery plants
3.2-10
-------
Carbon black plants
Primary lead smelters
Fuel conversion plants
Sintering plants
Secondary metal production plants
Chemical process plants
Fossil fuel-fired boilers
Petroleum storage and transfer units
Taconite ore processing plants
Glass fiber processing plants
Charcoal production plants
Fossil fuel-fired steam electric plants
Any other source category being regulated under Section 111 or 112 of the
Act at the time of the applicability determination would also be included in
the above list.
The current regulations subject a modified source to review if it is
one of the 28 source categories with emission increases above 91 Mg (100
tons) per year or if it is any other source with increases above 227 Mg (250
tons) per year; associated emission reductions were not allowed to exempt
the source from PSD review, but. reductions that offset the increase and
prevented a net increase were allowed to be used to avoid BACT review.
Under the proposed regulations, any modification to a major source would be
subject to PSD review if the modification would cause a net increase in the
source's potential to emit. The proposal also states that emission
increases offset entirely by contemporaneous emission reductions would not
be considered a modification. However, if a major stationary source
modifies its pollutant emissions so that the net increase in any pollutant
would be above the proposed de minimi's levels, it would be subject to PSD
review for all the pollutants it emits above the de minimi's levels as a
result of the modifications.
The September 5, 1979, proposal revised the definition of baseline and
established the baseline date as the time of the first completed permit
application, after August 7, 1977, within an Air Quality Control Region
designated as either attainment or unclassified.
More ambient monitoring would be required before and after construc-
tion, as a result of the proposal, for all pollutants regulated under the
Act—not just the criteria pollutants.
3.2-11
-------
The regulations also proposed to exempt, on a pollutant-specific basis,
major modifications and new sources from the BACT and air quality impact
assessment requirements, if the emissions of a specific pollutant are below
the de minimis level. The de minimi's air quality levels are to be used as
guidelines for exempting sources from PSD air quality analysis on a
pollutant-specific basis, if the emissions are above the proposed de minimis
emissions levels.10
The PSD program is currently being implemented for the most part by
EPA; however, the states are required to develop their own PSD programs
consistent with the requirements published by EPA.
3.2.1.3 State Implementation Plans for Visibility--
The Clean Air Act Amendments of 1977 under Section 169A require the EPA
Administrator to promulgate regulations setting forth guidelines for the
development of SIP's to remedy existing problems and prevent future visi-
bility impairment in those mandatory Class I Federal areas where visibility
is an important value. The current assessment of Class I areas where visi-
bility is an important value includes all mandatory Class I Federal areas
except two, Bradwell Bay (Florida) and Rainbow Lake (Wisconsin).
The SIP revisions will include an evaluation of existing major sources
and a requirement that those major sources which started operation after
August 6, 1962, and which cause or contribute to significant visibility
impairment in the mandatory Class I Federal areas install and operate Best
Available Retrofit Technology. These sources, which are listed in Table
3.2-2, have the potential to emit 227 Mg (250 tons) per year. The visi-
bility regulations were proposed on May 23, 1978 (45 FR 34762)
3.2.1.4 New Source Performance Standards (NSPS)--
The Clean Air Act of 1970 in Section 111 requires the EPA to develop
NSPS. The overriding purpose of Section 111 is to prevent the general
occurrence of new air pollution problems by requiring the installation of
best available controls during initial construction. Performance standards
for new sources are designed to allow industrial growth without undermining
air quality management goals. The NSPS are established at a national level
to provide uniformity and consistency to the requirements that a new source
must meet, regardless of location.
3.2-12
-------
The term "standard of performance" shall reflect the degree of emission
limitation and the percentage reduction achievable through the application
of the best technological system of continuous emission reduction which the
Administrator determines has been adequately demonstrated (taking into
consideration the cost of achieving such emission reduction, any nonair
quality health and environmental impact, and energy requirements). In
addition, the standards are to be established only for significant sources
(i.e. greater than 100 tons/yr potential emissions).11
The NSPS apply only to new sources. A new source is defined as any
stationary source on which construction is started after publication in the
Federal Register of the proposed NSPS for that source type. It is the
intent of the Clean Air Act that eventually NSPS will be promulgated for al1
significant emission-producing sources. Additionally,, any physical or
operational change to an existing facility that results in an increase in
the emission rate to the atmosphere of any pollutant to which a standard
applies shall be considered a modification within the meaning of Section 111
of the Act. Upon modification, an existing facility shall become an
affected facility for each pollutant to which a standard applies and for
which there is an increase in the emission rate to the atmosphere. .
In addition, when an existing facility is reconstructed, it becomes
subject to an applicable NSPS irrespective of any change in emission rate.
Reconstruction involves the replacement of components of an existing
facility to such an extent that (1) the fixed capital cost of the new
components exceeds 50 percent of the fixed capital cost that would be
required to construct a comparable entirely new facility and (2) it is
technologically and economically feasible to meet the applicable standards
set forth in 40 CFR Part 60.
Implementation and enforcement of NSPS may be delegated to the states;
however, until the state submits a satisfactory plan, EPA is required to
enforce the NSPS.
As of January 1, 1979, EPA had promulgated NSPS regulations covering 27
new source categories. Of these only six source categories include limits
for S02:
3.2-13
-------
1) Fossil-fuel-fired steam generators of capacity greater than 73 MW
thermal (250 x 106 Btu/hr)
2) Sulfuric acid plants
3) Petroleum refineries
4) Primary copper smelters
5) Primary zinc smelters
6) Primary lead smelters.
The Clean Air Act Amendments of 1977 reaffirm the NSPS process set
forth by the 1970 act and require that EPA significantly expand the coverage
of NSPS over the next 4 years. The EPA is directed to identify all cate-
gories of stationary sources that are not among the list of source cate-
gories regulated under NSPS. The Clean Air Act provides some guidance in
determining the priorities for promulgating standards for certain categories
of major stationary sources. The Administrator shall consider the fol-
lowing:
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 antici-
pated to endanger public health or welfare
The mobility and competitive nature of each such category of
sources and the consequent need for nationally applicable new
source standards of performance.
Given the priorities of the 1977 Act, EPA undertook a study12 to estab-
lish priorities for setting NSPS. The results of this study culminated in
the publishing of the final priority list for NSPS in the August 21, 1979,
Federal Register.
Additionally, the 1977 amendments provide some guidance on the desired
schedule for developing standards.13 By August 1980, NSPS must be estab-
lished for at least one-quarter of the categories listed in the August 21,
1979, Federal Register notice. By August 1981, standards must be promul-
gated for at least three-quarters of the listed categories, and for all
listed categories by August 1982.
3.2-14
-------
The 1977 amendments also introduce an important provision relating to
NSPS for new fossil-fuel-fired stationary sources, Section 111(a)(l)(A).
Formerly such sources were subject only to the customary emission limita-
tions expressed as nanograms per joule (pounds per million Btu). Many
fossil-fuel-fired stationary sources turned to firing of low-sulfur coal and
similar fuels from western mines in order to meet the emission limitations.
The 1977 amendments require that new and modified fossil-fuel-fired sta-
tionary sources achieve a percentage reduction in emissions in addition to
meeting emission limitations. By requiring a specified percentage reduc-
tion, this provision appears to eliminate the incentive for burning fuels of
naturally low sulfur content and requires these sources to adopt technolog-
ical controls.14
3.2.2 Dispersion and Dispersion Modeling
As stated earlier, an application for a preconstruction permit for a
new source or a major modification to an existing source must include an
assessment of the air quality impact of the emissions from the proposed
source. For a source that is to be located in an attainment area, it is
necessary to satisfy the requirements for the PSD. These requirements
include showing that the source will not cause violations of applicable
NAAQS nor will the maximum impact of the source exceed the remaining PSD
increment (Table 3.2-1). Further, it must be shown that the new source will
not significantly impact any area that has been designated as nonattainment.
Here a significant impact is construed to be the same as the maximum PSD
increments for a Class I area (i.e., for S02, 2 ug/m3 annually, 5 ug/m3 in
24 hours, and 25 ug/m3 in 3 hours, not to be exceeded more than once per
year at any receptor).
A new source of SO emissions that is to be constructed in a nonattain-
A
ment area must demonstrate that there will be a net air quality improvement
within the nonattainment area after the new source becomes operational. It
must also be shown that the new source will not cause violations of the
NAAQS, nor exceed the allowable PSD increments in any nearby attainment
area.
3.2-15
-------
The Federal regulation relative to the PSD (40 CFR 52.21) specifies
that "all estimates of ambient concentrations required under this section
shall be based upon applicable air quality models, data bases, and other
requirements specified in the Guideline on Air Quality Models."^ The
guideline includes a brief description of currently available dispersion
models and their applicability to individual and multiple sources This
guideline should be followed in any analysis pertinent to sources of SO
emissions. x
Congress recognized the state-of-the-art nature of currently available
dispersion models and the need to maintain consistency in the application of
these models. To this end Section 320 of the Clean Air Act, as amended
August 1977, requires that the Administration of the EPA conduct periodic
conferences on air quality models. The Act specifically provides that such
conferences provide participation by the National Academy of Sciences, state
and local air pollution control agencies, and'other appropriate agencies
such as the National Science Foundation, the National Oceanic and Atmos-
pheric Administration and others. These conferences will provide a basis
for the updating of the Guidelines on Air Quality Models.
Section 123 of the Clean Air Act, as amended in 1977, limits the credit
of the physical stack height for determining control requirements to that
height required to avoid excessive concentration caused by "atmospheric
downwash, eddies, and wakes." The EPA has proposed to define that height as
H + 1.5 L, where H is the height of a nearby building and L is the lesser
dimension of the height or width of that building, or the height that is
demonstrated as necessary through physical modeling or field studies.
3.2-16
-------
REFERENCES FOR SECTION 3.2
1. U.S. Congress. Clean Air Act, Section 108(a)(IXA)(B). 42 USC 1857 et
seq.
2. U.S. Environmental Protection Agency. State Implementation Plan Emis-
sion For Sulfur Oxides: Fuel Combustion. EPA-450/2-76-002. March
1976. p. 6.
3. Analysis of Final State Implementation Plans—Rules and Regulations.
APTD 1334. July 1972. pp. 9-11.
4. Ref. 2, pp. 7-10.
5. U.S. Environmental Protection Agency. States Attainment Status.
Office of the Federal Register. 43 FR 8961. Washington, D.C. March
3, 1978.
6. U.S. Congress. Clean Air Act. Section 173. 42 USC 1857 et seq.
7. U.S. Environmental Protection Agency. Prevention of Significant Air
Quality Deterioration. December 5, 1974. Code of Federal Regulations.
40 CFR 52.21.
8. U.S. Environmental Protection Agency. Prevention of Significant Air
Quality Deterioration. Office of the Federal Register. 43 FR 26380-
26384. June 19, 1978.
9. Goldsmith, B.J. , and J.R. Mahoney. Implications of the 1977 Clean Air
Act Amendments for Stationary Sources. Environmental Science and
Technology. February 1978. pp. 144-149.
10. U.S. Environmental Protection Agency. Prevention of Significant Air
Quality Deterioration. Office of the Federal Register. 43 FR 51924.
September 5, 1979.
11. McCutchen, G.D., and R.E. Jenkins. New Source Performance Standards.
Environmental Science and Technology. October 1972. pp. 884-888.
12. Argonne National Laboratory. Priorities for New Source Performance
Standards Under the Clean Air Act Amendments of 1977. (Draft)
February 28, 1978. pp. 59-64.
13. Ref. 12, pp. 1,2.
3.2-17
-------
14. Tnritt, T.H. , and R.M. Hall.
Publications, Inc. pp. 42-54.
Practical Environmental Law. Federal
15. Guideline on Air Quality Models. U.S. Environmental Protection Agency
Research Triangle Park, N.C. EPA-450/2-78-027. April, 1978 48 pp.'
3.2-18
-------
3.3 TECHNICAL CONSIDERATIONS
3.3.1 Introduction
When attempting to reduce S02 emissions from an existing source or to
prevent or minimize potential S02 emissions from a proposed facility, one
must be aware of local, state, and Federal standards limiting S02 emissions.
For example, new stationary source performance standards limiting S02 emis-
sions from power plants may be found in the Federal Register.1
Many processes are available for the removal of S02 from both process
and combustion sources. These processes involve direct conversion of S02
streams to sulfuric acid, sulfur, and liquid S02 for sale or disposal;
absorption of S02 in ammonia liquors, which can produce a salable ammonium
sulfate fertilizer product; absorption of S02 from lower concentration
streams with organic or inorganic reagents to produce a stronger, usable S02
stream for conversion to salable sulfur products; or the absorption of S02
by various alkaline reagents in nonregenerable systems. Many such systems
are used in the utility2 and industrial3 sectors.
3.3.2 Determination of the Required SO? Removal
The magnitude of S02 removal required can be obtained by comparing
existing S02 emission levels with S02 control regulations. For an existing
source, emission testing is performed to determine the degree of S02 control
required. For major modifications to existing installations, which ar^e
expected to yield increased sulfur oxide emissions, or for new plants,
background sampling is required to establish local environmental quality.
The sampling, measurement, and analysis techniques practiced by the regula-
tory/control agency having jurisdiction over the emission source should be
used. An approximate method to determine S02 emissions from a combustion
source is to use the appropriate emission factor.4 For example, the amount
of uncontrolled S02 emissions from a new coal-fired boiler for a specific
heat input rate can be determined by knowing the sulfur content of the coal
to be fired and the heating value of the coal, and by assuming the frac-
tional conversion (such as 0.95) of sulfur in the coal to S02.5 Comparing
the peak uncontrolled value of S02 emissions with the applicable regulation
gives the required S02 removal. The necessary S02 removal for existing
sources can be determined similarly.
3.3-1
-------
3-3.3 Gas Stream Characteristics
The next step Is to check the characteristics of the gas stream to be
treated.
In addition to sulfur oxides, the chemical composition of volatilized
combustion products in the gas (such as N0x, chlorides, fly ash, etc.)
should be estimated in light of current and projected removal requirements.
The moisture content and temperature of inlet flue gas dictates the
adiabatic saturation temperature of the gas stream and determines the amount
of water that evaporates when the gas is cooled in either an S02 absorber or
a wet particulate scrubber. The inlet gas temperature also affects the
decision on whether the scrubber or absorber should be lined and what type
of liner should be used.
The volume of the flue gas and the desired gas velocity in the flue gas
desulfurization (FGD) train largely determine the dimensions of the particu-
late scrubber (if one is needed) and the S02 absorber. Furthermore, the
availability of space for such an installation can influence the design.
The erosive properties of the inlet flue gas and its corrosiveness
should be determined. The erosive characteristics can affect the equipment
upstream of the FGD system (such as a forced-draft fan) and can have a
significant effect on the downstream equipment that handles the gas stream.
Stress corrosion caused by the high chloride content of some gas streams is
crucial in selecting reliable materials of construction for the S02 absor-
ber.
The electrical characteristics of the fly ash particles in the gas,
such as the resistivity and the dielectric constant, play an important role
when an electrostatic precipitator (ESP) is used for particulate collection
prior to S02 removal.
3.3.4 S02 Removal Processes
As noted earlier, there are basically two choices in S02 removal pro-
cesses: either use the S02 stream directly in a process for conversion to a.
sulfur product for use, sale, or disposal (such as sulfuric acid, sulfur, or
liquid S02); or absorb the S02 in processes designed to produce an upgraded
S02-concentration stream or a product for disposal.
3.3-2
-------
The systems that produce a sulfur product directly may be typified by
sulfuric acid plants in use in the nonferrous smelter industry (see Section
5.1), sulfur-reduction units in use in the petroleum industry (see Section
5.3), and dimethyl aniline (DMA) scrubbing units also in use in the non-
ferrous smelter industry. Sulfur product processes also are used in con-
junction with weaker S02 stream removal processes where the absorbent is
regenerated. The number of units, which may be referred to as flue gas
desulfurization processes, that absorb S02 from process or combustion
streams, for ultimate upgrading or disposal is much greater than those that
directly produce a sulfur product.
There are four major categories of S02 removal (FGD) processes: non-
regenerable, regenerable, wet, and dry. Nonregenerable processes produce
either sludge or waste liquor that must be disposed of in an environmentally
acceptable manner. Regenerable processes involve absorbent regeneration and
production of elemental sulfur or a sulfur compound for sale or disposal;
the byproduct may be marketable S02, elemental sulfur, or sulfuric acid.6
Wet processes may require stack gas reheating to achieve the necessary plume
buoyancy and to avoid corrosion problems; particulate removal, if required,
is achieved prior to S02 removal. Dry processes generally do not require
stack gas reheating and often simultaneously remove particulate matter and
S02.
Wet lime/limestone FGD processes are most widely used in the utility
sector. These processes can have scaling and plugging problems, which have
been reduced as knowledge of chemical processes and as operating experience
have been obtained. The primary advantage of these processes is the low
reagent cost. System reliability and S02 removal capability can vary.
Details are available on the operation of existing systems and S02 removal
efficiencies achieved by these processes.2
Wet, sodium-based, nonregenerable FGD processes are most widely used in
the industrial sector. Operating histories have generally been good.
Sulfur dioxide removal efficiencies greater than 90 percent have been
achieved by many installations.3 The advantages of sodium-based processes
are that S02 absorption is achieved by a clear alkaline solution and the
reaction product is soluble. This eliminates scaling and plugging problems
3.3-3
-------
found in many calcium-based processes. The disadvantage is a significantly
higher cost for the reagent than for lime or limestone.
The selection of an FGD system is influenced by whether the application
is new or retrofit (an addition to the existing plant). In retrofit appli-
cations, space constraints and other site-specific factors are crucial to
the installation of an FGD system.
A nonregenerable process produces throwaway sludge or waste liquor. As
a result, the air pollution control problem can lead to both water pollution
control and solid waste disposal problems. A 500-MW power plant firing coal
with 3.5 percent sulfur and 14 percent ash [with a heat content of 27,924
kJ/kg (12,000 Btu/lb)] and employing a lime FGD system with 90 percent'ab-
sorbent utilization and 80 percent S02 removal can generate 104,000 Mg/yr
(115,000 tons/yr) of ash and 98,000 Mg/yr (108,000 tons/yr) of sulfur-based
sludge (dry basis).7 The disposal of these waste streams can require a
large lined pond and further treatment.8
A nonregenerable FGD system can operate in an open- or closed-loop mode
with respect to water usage. Current water pollution control legislation
requires closed-loop operation. In a closed-loop system, it is necessary to
control the makeup water flow at a level to avoid aqueous discharge and
insufficient water for proper operation. Maintaining the proper water
balance is critical in areas with limited availability of water.
In addition to S02 removal, the source may require a particulate col-
lection device such as an ESP, fabric filter (baghouse), or wet scrubber.
The choice of a particulate removal system is often determined by the physi-
cal and chemical characteristics of the fly ash and by whether particulate
control devices are currently used.9
3.3.5 Control Equipment Alternatives
3.3.5.1 Particulate Control Devices--
Electrostatic precipitators, baghouses, and wet particulate scrubbers
are used to remove particulate matter in order to meet emission standards
for a source. The particulates not removed ahead of the FGD system can be
removed in a properly designed absorber. The reagent for S02 absorption,
however, can become contaminated, which increases the system blowdown and
3.3-4
-------
operating cost. When an ESP is used for particulate removal and upgrading
of existing facilities is needed, the various alternatives include condi-
tioning the flue gas, reducing the temperature of the flue gas stream, and
installing additional plate area for particulate collection.
3.3.5.2 Presaturator--
Most FGD system suppliers prequench the gas before it enters the absor-
ber so that the absorber lining (if used) is not exposed to a high tempera-
ture gas stream and the area of the wet/dry interface is minimized. Quench-
ing may be done in a separate venturi stage in the inlet duct or inside the
main absorber (especially in a horizontal absorber).
3.3.5.3 Types of S02 Absorbers10'11--
Flue gas desulfurization system suppliers offer a wide variety of
absorbers: spray tower, packed tower, tray tower, or venturi; vertical or
horizontal design; single-stage or two-stage construction; and cocurrent,
countercurrent, or crossflow operation. The tradeoffs are simplicity,
easier maintenance, and possibly higher reliability, contrasted with a more
sophisticated absorber system with better S02 removal capability and absor-
bent utilization.
A spray absorber has no internals except spray nozzles. The mass
transfer capability is not as good in spray absorbers as in packed or tray
absorbers, but this can be offset by using a high liquid-to-gas ratio (L/G).
A crossflow (horizontal) absorber is a special type of spray absorber.
A packed absorber may have mobile or fixed packing; fixed packing can
be of various types. Packing improves mass transfer and allows the absorber
to be operated at a lower L/G ratio; however, packing -makes the absorber
more subject to solids deposition, scaling, and plugging.
Tray absorbers consist of one or more trays mounted transversely inside
the shell. The mass transfer capability is higher for tray absorbers than
for spray absorbers. The overall effect is a multiple .countercurrent con-
tactor for the gas and liquid streams.
A venturi scrubber-absorber can collect particles with high efficiency;
however, mass transfer capabilities are limited because of limited cocurrent
gas-liquid contact time.
3.3-5
-------
3.3.5.4 Materials of Construction--
The most important consideration in FGD system design is the selection
of materials of construction. Low pH, high chloride, and presence of ero-
sive solids complicate the selection of materials of construction. At high
chloride levels (>30,000 ppm) in the recirculation loop, the use of reason-
ably priced, unprotected material is questionable because of chloride stress
corrosion. Stainless steel can be used at average chloride levels (3000 to
5000 ppm). A number of coating materials have been tried, each having
positive and negative features; no liner can be used in all applications.
Site-specific considerations will dictate the type of liner that can be
used. Process conditions determine what materials of construction should be
chosen for other pieces of process equipment.12'13
3.3.6 Cost of Control
The cost of FGD systems is an area of considerable interest and sub-
stantial controversy. Many studies have been made to estimate capital and
annual costs.14'15'16
Capital costs consist of direct costs, indirect costs, contingency
costs, and other capital costs. Direct costs include the bought-out cost of
equipment, the cost of installation, and site development. Indirect costs
include interest during construction, contractor's fees and expenses, engi-
neering, legal expenses, taxes, insurance, allowance of startup, and spares.
Contingency costs include those costs resulting from unforeseen sources.
Other capital costs include the nondepreciable items of land and working
capital. The prefabricated equipment for a typical nonregenerable process
includes fans and motors, ductwork, reheaters (if required), S02 absorbers,
tanks and agitators, and pumps and motors. The cost of S02 absorbers is a
major capital expense that may vary from 50 to 85 percent of the total
installed cost of the equipment.
Annual costs consist of direct costs, fixed costs, and overhead costs.
Direct costs include the cost of raw materials, utilities, operating labor
and supervision, and maintenance and repairs. Fixed costs include deprecia-
tion, interim replacement, insurance, and taxes and interest on borrowed
capital. Overhead costs include plant and payroll expenses.
Various sources are available to estimate the capital and annual costs
of the S02 control system selected.14'15,16,1?
3.3-6
-------
REFERENCES FOR SECTION 3.3
1.
2.
3.
5.
6.
7.
8.
10.
New Stationary Sources Performance
Generating Units. Federal Register.
Standards; Electric Utility Steam
Part II. June 11, 1979.
Smith, M., et al. EPA Utility FGD Survey. December 1978 - January
1979. U.S. Environmental .Protection Agency. Washington, D.C. EPA-
600/7-79-022C. May 1979.
Tuttle, J., et al. EPA Industrial Boiler FGD Survey. First Quarter
1979. U.S. Environmental Protection Agency. Washington, D.C. EPA-
600/7-79-067b. April 1979.
4 US Environmental Protection Agency. Compilation of Air Pollutant
Emission Factors. 3rd ed. AP-42. Research Triangle Park, N.C. 1978.
Ponder T et al. Lime FGD Systems Data Book. U.S. Environmental
Protection Agency, Electric Power Research Institute, Research Triangle
Park, N.C. May 1979. pp. 2.2-6 to 2.2-10.
Bethea, R.M. Air Pollution Control Technology.
Reinhold Company. 1978. pp. 355-363.
New York, Van Nostrand
Leo, P.P., and J. Rossoff. Controlling S02 Emissions From
Steam-Electric Generators: Solid Waste Impact, Volume I.
ronmental Protection Agency. Research Triangle Park, N.C.
78-044a." March 1978. pp. 4, 5.
Coal-Fired
U.S. Envi-
EPA-600/7-
Rossoff J- , et al. Disposal of Byproducts From Nonregenerable Flue
Gas Desulfurization Systems: Final Report. U.S. Environmental Protec-
tion Agency. Research Triangle Park, N.C. EPA-600/7-79-046. February
1979. p. 4.
Szabo M F and R.W. Gerstle. Operation and Maintenance of Particu-
late Control' Devices on Coal-Fired Utility Boilers. U.S. Environmental
Protection Agency, Washington, D.C. EPA-600/2-77-129. July 1977.
pp. 2-4.
U S Environmental Protection Agency, Office of Air Quality, Planning
and Standards. Electric Utility Steam Generating Units, Background
Information for Proposed S02 Emission Standards. Research Triangle
Park, N.C. EPA-450/2-78-007a. July 1978.
3.3-7
-------
11.
12.
ar' F1+Ue GaS Desul^ization System Capabilities for
EsiHiS™^-™ »™«= »
Chemical
Study,
ity, Muscle
EPA-450/3-80-009a. March 1980.
14.
15.
-••—-— • V IT >— | I I U I I h« i^
Research Triangle Park, N.C.,
16' ^^r-£,jS^H?^"^-r «=
17.
Beach Calif
. - Inc- Richardson Rapid System.
ion Estimating Standards. 1979-80. Solana
3.3-8
-------
3.4 ENVIRONMENTAL AND ENERGY IMPACTS
This subsection discusses the environmental impacts of sulfur dioxide
removal and the attendant energy impacts. The streams involved are the
cleaned flue gas from coal combustion, the liquid portion of the scrub-
ber/absorber purge, and any sludge generated.
3.4.1 Air Quality
The primary air quality impact of a flue gas desulfurization (FGD)
system is the reduction of S02 in the exit gas stream. As discussed in
depth in Section 4.2, FGD systems have demonstrated the capability of re-
moving in excess of 90 percent of S02 emissions from combustion flue gas
streams. This reduction in S02 is also reflected in the reduction of,secon-
dary sulfates, which are formed by the oxidation of the S02.
As S02 emissions are reduced, the impact of sulfuric acid and sulfate
salts resulting from the chemical transformation of S02 in the atmosphere
(secondary sulfates) is reduced; however, the amount of primary sulfates
(sulfuric acid mist, sulfur trioxide aerosol, and sulfate salts) emitted
directly as emissions from combustion sources can increase.1'2 An air
pollution control system such as an FGD system changes the characteristics
of the gas stream. The gas becomes saturated with water vapor and is adia-
batically cooled. The sulfur oxides in the flue gas after the FGO system
can combine with water vapor to generate a fine aerosol of sulfurous and
sulfuric acid. The sulfurous and sulfuric acid mist emitted from an absor-
ber can produce secondary sulfates. It is generally necessary to reheat the
treated flue gas to increase the plume buoyancy; this improves the disper-
sion of mist in the air.
In some cases, the S02 absorber also emits fine particulate matter that
can be attributed either to poor removal efficiency of a particulate removal
device or to particulates generated in the absorber. These particulates are
solids present in the absorbent that are carried over -from the absorber.
The fine parti culates emitted to the air can increase the opacity of the
plume. Proper control of process conditions, improved mist elimination
design, and improved efficiency of particulate removal devices for fine
particulates can minimize the problem.
3.4-1
-------
3.4.2 Water Quality
An FGD system generates liquid waste streams mainly as a purge from
different points in the system (e.g., the quencher loop) and as supernatant
liquor from the effluent stream. These liquid waste streams can contain
significant amounts of total dissolved solids (IDS) and small, but signifi-
cant, amounts of trace elements. Table 3.4-1 compares the chemical con-
stituents of liquid waste streams from nonregenerable FGD systems with the
National Interim Primary Drinking Water Regulation (NIPDWR).3 The ratios of
constituent concentrations to water criteria are given for ranges of various
constituents of the samples analyzed.
TABLE 3.4-1. COMPARISON OF WASTE LIQUOR WITH
DRINKING WATER CRITERIA3
NIPDWR
drinking water
criteria, rag/liter
As 0.05
Cd 0.01
Cr 0.05
Pb 0.05
Mg 0.002
Se 0.01
F -x-2
TDS 500
pH (actual values)b
Range of
all samples3
<0.8 - 2.8
0.4 - 11
0.22-5
0.2 - 6.6
0.03 - 2.5
0.28 - 20
<0.5 - 5
6.6 - 48.5
6.7 - 12.2
Actual concentration/criteria concentration-
unit! ess number.
EPA-proposed secondary regulation is 6.5 to 8.5.
Table 3.4-1 shows that concentrations of all elements and the TDS and
pH exceed the drinking water criteria. Although trace elements are not
eliminated as a matter of concern by these data, there are indications that
in many cases the concentrations are quite low and that the primary item of
concern may generally be the concentration of dissolved solids and, in some
cases, pH.3
3.4-2
-------
3.4.3 Solid Waste
Under the provisions of the Resource Conservation and Recovery Act
(RCRA) of 1976, the Environmental Protection Agency (EPA) has been directed
to provide regulations and guidance to control the disposal of hazardous and
nonhazardous wastes including solid waste, which includes air pollution
control sludges. •
The sludge and ash generated by a nonregenerable FGD system can require
a large disposal area. It has been estimated that the amount of solid waste
(ash and sludge) produced by a 500-MW power plant firing a 3.5 percent
sulfur coal and having a limestone scrubbing system is 211,380 dry Mg/yr
(233,000 dry tons/yr) or 298,500 m3 (242 acre-ft) by volume.4
The solid wastes produced as a result of using regenerate processes
are approximately 50 percent of those from nonregenerable processes. The
wastes are primarily ash and are nearly independent of the regenerate
process.5
Table 3.4-2 presents the concentrations of various sludge constituents
from nonregenerable FGD systems.6 The trace element content in an FGD
sludge is a direct function of the combustion products of coal. Fly ash can
represent the major source of trace elements in sludge for all but the most
volatile elemental species (e.g., mercury and selenium) that are scrubbed
from flue gases.6 The FGD sludge may be treated chemically by several
processes and often can be used in landfill applications. It has been shown
that sludges chemically treated by commercially available processes can be
disposed of in an environmentally sound manner and that the disposal site
can be reclaimed as a structural landfill. The EPA considers permanent land
disposal of raw (unfixated) sludge to be environmentally unsound. Although
EPA has no regulatory authority to prevent raw sludge disposal, EPA antici-
pates states and local jurisdictions to require treatment of sludge.7
In regenerable processes, the S02 in the flue gas is absorbed and
subsequently most often released as S02 in the regeneration of absorbent.
The S02 may be processed further to form sulfuric acid or elemental sulfur.
3.4-3
-------
TABLE 3.4-2. RANGE OF CONCENTRATIONS OF CHEMICAL
CONSTITUENTS IN FGD SLUDGES FROM LIME,
LIMESTONE, AND DOUBLE-ALKALI SYSTEMS6
Scrubber
constituent
Aluminum
Arsenic
Beryllium
Cadmium
Calcium
Chromium
Copper
Lead
Magnesium
Mercury
Potassium
Selenium
Sodium
Zinc
Chloride
Fluoride
Sulfate
Sulfite
Chemical oxygen
demand
Total dissolved
solids
PH
Sludge concentration range
Liquor, mg/liter
(except pH)a
0.03 - 2.0
<0.004 -1.8
<0.002 - 0.18
0.004 - 0.11
180 - 2600
0.015 - 0.5
<0.002 - 0.56
0.01 - 0.52
4.0 - 2750
0.0004 - 0.07
5.9 - 100
<0.0006 - 2.7
10.0 - 29,000
0.01 - 0.59
420 - 33,000
0.6-58
600 - 35,000
0.9 - 3500
<1 - 390
2800 - 92,500
4.3 - 12.7
Solid, mg/kgb
0.6 - 52
0.05-6
0.08-4
105,000 - 268,000
10 - 250
8 - 76
0.23 - 21
0.001 - 5
2-17
48,000
45 - 430
9,000
35,000 - 473,000
1600 - 302,000
Liquor analyses were conducted on 13 samples from seven power
plants burning eastern or western coal and using lime, limestone
or double-alkali absorbents.
Solids analyses were conducted on six samples from six power plants
burning eastern or western coal and using lime, limestone, or
double-alkali scrubbing processes.
3.4-4
-------
3.4.4 Energy Impacts
The increasing demand for energy in the United States is projected to
be met in part by a significant increase in fossil fuel combustion. Energy
consumption associated with FGD systems can vary widely with the process and
vendor. Energy is required to run the recirculation and transfer pumps,
booster fans, and other process equipment. Different processes also require
varying degrees of energy use for absorbent makeup, absorbent regeneration
(if required), and/or sludge disposal. Additional energy is consumed by use
of fuel or steam to reheat flue gases and process steam in some of the
regenerate FGD systems. This energy consumption can be provided by addi-
tional power generation. If the source does not produce power or if maximum
power generation occurs, additional energy may have to be purchased; or the
power boiler may be derated. Additional S02 emissions would be generated
while producing the energy needed to operate the S02 control equipment.
3.4.4.1 Emission System Capacity Penalties--
The FGD systems cause losses in net generation by a power plant that
sometimes requires the addition of generation capacity. The additional
power-generating capacity required to compensate for the power used by the
emission control system is a capacity penalty. These penalties can be
expressed both as a percentage of the generating capacity of the unit con-
trolled and as an additional operating cost in mills/kWh.
3.4-5
-------
REFERENCES FOR SECTION 3.4
1. U.S. Environmental Protection Agency. Workshop Proceedings on Primary
Sulfate Emissions From Combustion Sources. Research Triangle Park
N.C. EPA-600/9-78-020a, b. August 1978. pp. iii, iv, 275 (from 020a)
and pp. 3, 4, 78, and 95.
2. Leavitt, C. , et al. Environmental Assessment of Coal- and Oil-firing
in a Controlled Industrial Boiler. U.S. Environmental Protection
Agency. Research Triangle Park, N.C. EPA-600/7-78-164a. August 1978.
p. 8.
3. Rossoff, J., et al. Disposal of Byproducts From Nonregenerable Flue
Gas Desulfurization Systems: Final Report. U.S. Environmental Pro-
tection Agency. Washington, D.C. EPA-600/7-79-046. February 1979
pp. 15-17.
4. Leo, P.P., and J. Rossoff. Controlling S02 Emissions from Coal-fired
Steam-Electric Generators: Solid Waste Impact, Volume I. U.S. Envi-
ronmental Protection Agency. Washington, D.C. EPA-600/7-78-044a
March 1978. pp. 14-19.
5. Ref. 4, pp. 23, 24.
6. Ref. 4, pp. 30, 31.
7. Ref. 3, pp. 3, 4.
3.4-6
-------
3.5 THE SULFUR MARKET
Air pollution control regulations require the limitation of sulfur
oxide emissions from combustion and industrial sources. The recovery of
useful sulfur products from these emissions could help conserve natural
resources and avoid the waste disposal problems often associated with non-
regenerable S02 removal systems. Many factors affect the marketability of
such products. Current data suggest that significant amounts of sulfuric
acid could be produced from the S02 in power plant flue gas and sold in
competitive markets.1
3.5.1 Types of Product
Sulfuric acid (H2S04) is relatively simple to manufacture and at
present is the most economically attractive product of S0x abatement.1
Although more complex to produce, elemental sulfur offers several marketing
advantages over sulfuric acid and could become economically competitive.2
The quality of elemental sulfur and sulfuric acid is important to their mar-
ketability.3 The economic feasibility of other possible S0x abatement
products, such as liquid S02, liquid S03, and oleum (S03 dissolved in
H2S04), has not been seriously studied.
3.5.1.1 Sulfuric Acid—
Sulfuric acid is the most widely used industrial process chemical in
the world.4 Its production has been used as an indicator of general eco-
nomic activity. High shipping costs and economies of production scale have
led to concentrations of sulfuric acid manufacturers in certain areas.4
The quality and concentration of the sulfuric acid influence its
marketability. Several items that can affect the resulting sulfuric acid
are S02 concentration, particulate loading, water content, and contaminants
in the incoming gas stream. Processes that yield a concentrated S02 stream
for sulfuric acid manufacture are preferable to those that do not. For the
protection of the vanadium pentoxide catalyst, particulate matter must be
removed from the gas stream; in many cases other than when sulfur is burned
to produce S02, the water is removed from the process gas stream. Another
concern is condensed moisture in the acid. Most commercial sulfuric acid
produced contains almost no dust and not more than 7 percent water.5 The
3.5-1
-------
most common grades of sulfuric acid produced are 93 percent (66° Baume) and
98 percent concentration.
3.5.1.2 Elemental Sulfur--
Elemental sulfur does not possess the hazardous, corrosive properties
of sulfuric acid. It can be stored and shipped more easily than sulfuric
acid and has a wider potential market in some areas. * The cost of the
reducing agent [such as methane (natural gas), hydrogen sulfide, etc.],
however, is often significant, and the reduction of S02 to elemental sulfur
requires a relatively complex plant. The Claus process is widely used to
produce elemental sulfur, especially in the petroleum and natural gas in-
dustries. There are two sulfur recovery units on utility FGD systems.
The color and purity of the sulfur product affect salability. A prod-
uct with a color different from or purity less than that of Frasch sulfur
would probably have to be priced lower. ? The Frasch process is the method
by which most of the virgin sulfur used in the United States is mined.
3.5.2 Industry Description
3.5.2.1 Sulfuric Acid—
The current annual production capacity of all manufacturers of virgin
acid in the United States is 37.92 Tg (41,796,000 tons).* In addition,
nonferrous smelters can produce 6.17 Tg (6,797,100 tons) per year of sul-
furic acid." Thus, the present total capacity is 44.08 Tg (48,593,100 tons)
per year.< Tables 3.5-1 and 3.5-2 list the annual capacities of individual
producers of virgin acid and smelter acid, respectively.
The yearly demand for sulfuric acid has increased since 1975 and is
expected to continue to do so through 1985. Past, present, and future
annual demands are summarized below:4
Annual demand,
Tg (10s tons')
Year
1975
1978
1979
1980
1983
1985
Sulfuric acid production increased at an annual rate of 3.4 percent
from 1968 through 1978.* This growth rate is expected to continue through
1983.4
29.628
36.299
37.533
38.811
42.906
45.871
(32.660)
(40.013)
(41.374)
(42.782)
(47.295)
(50.565)
3.5-2
-------
TABLE 3.5-1.
ANNUAL CAPACITIES OF ALL U.S,
OF VIRGIN ACID IN 19794
PRODUCERS
Producer
Agrico Chemical, South Pierce, Fla.; Dona Idsonvi lie, La.
Allied Chemical, Anacortes, Wash.; Baton Rouge, Geismar,
La; Buffalo, N.Y.; Chicago, 111; Claymont, Del.;
Elizabeth, N.J.; Fort Royal, Hopewell , Va.; Newell,
Pa.; Nitro, W. Va.; Pittsburg, Richmond, Calif.
American Cyanamid, Joliet, 111.; Linden, N.J.; New
Orleans, La; Savannah, Ga.
Beker Industries, Conda, Idaho; Hahnville, La.;
Marseilles, 111.
Borden, Chesapeake, Va.; Piney Point, Fla.; Streator, 111.
CF Industries, Bartow, Plant City, Fla.
Chevron, El Segundo, Calif.; Honolulu, Hawaii
Cities Service, Augusta, Ga. ; Lake Charles, La.;
Monmouth Junction, N.J.
Coulton Chemical, Oregon, Ohio
DuPont, LaPorte, Tex; Burnside, La.; Cleveland, North
Bend, Ohio; Deepwater, Gibbstown, Linden, N.J.; East
Chicago, Ind. ; Richmond, Va.; Wurtland, Ky.
Essex Chemical, Newark, N.J.
Farmland, Industries, Green Bay, Fla.
First Mississippi, Fort Madison, Iowa
Freeport Minerals, Port Sulphur, Uncle Sam, La.
W.R. Grace, Bartow, Fla.
IMC, New Wales, Fla.
Kerr-McGee Nuclear, Grants, N. Mex.
LJ&M La Place Company, Edison, N.J.
Mississippi Chemical, Pascagoula, Miss.
Mobil Chemical, Depue, 111.
Monsanto, Avon, Calif.; El Dorado, Ark.; Everett, Mass.;
Sauget, 111.
USI, DeSoto, Kan.; DeBuque, La.; Tuscola, 111.
NL Industries, Sayerville, N.J.
Northeast Chemical, Wilmington, N.C.
Occidental Chemical, White Springs, Fla.; Plainview,
Tex.; Lothorp, Calif.
01 in, Beaumont, Pasadena, Tex.; Curtis Bay, Md.; North
Little Rock, Ark.; Shreveport, La.
Ozark-Mahoning, Tulsa, Okla.
Phelps-Dodge, Jeffrey City, Riverton, Wyo.
Philipp Brothers, Nichols, Fla.
US Phosphoric Products, Tampa, Fla.
Rohm and Haas, Deer Park, Tex.
Royster Company, Mulberry, Fla.
Swift, Bartow, Fla.; Calumet City, 111.; Dothan, Ala.
J.R. Simplot, Pocatello, Idaho
Stauffer, Baton Rouge, La.; Baytown, Fort Worth,
Manchester, Pasadena, Tex.; Dominguez, Martinez,
Calif.; Hammond, Ind.; Le Moyne, Ala.
Texasgulf, Aurora, N.C.
Union Chemicals, Wilmington, Calif.
USS Agri-Chemicals, Bartow, Fort Meade, Fla.;
Wilmington, N.C.
Valley Nitrogen Producers, Helm, Calif.
All others
Total
Annual capacity.
Tg (tons)
2.177 (2,400,000)
2.297 (2,532,000)
1.039 (1,145,000)
1.429 (1,575,000)
0.581 (640,000)
3.112 (3,430,000)
0.120 (132,000)
0.286 (315,000)
0.163 (180,000)
2.218 (2,445,000)
0.163 (180,000)
1.138 (1,254,000)
0.501 (552,000)
2.083 (2,296,000)
0.756 (833,000)
1.823 (2,010,000)
0.127 (140,000)
0.091 (100,000)
0.907 (1,000,000)
0.363 (400,000)
0.529 (583,000)
0.317 (349,000)
0.544 (600,000)
0.109 (120,000)
1.981 (2,184,000)
1.173 (1,293,000)
0.091 (100,000)
0.100 (110,000)
0.408 (450,000)
1.814 (2,000,000)
0.640 (705,000)
0.345 (380,000)
0.243 (268,000)
0.581 (640,000)
3.454 (3,807,000)
1.880 (2,072,000)
0.145 (160,000)
0.807 (890,000)
0.544 (600,000)
0.840 (926,000)
37.919 (41,796,000)
3.5-3
-------
TABLE 3.5-2.
ANNUAL "PACITIE^OFALJ, U.S. PRODUCERS OF SMELTER
Annual capacity,
Tg (tons)
Amax Lead, Boss, Mo.
Amax Zinc Sauget, m.
Anaconda, Anaconda, Mont.
Asarco, Columbus, Ohio; Corpus Christi El
Tex.; East Helena, Mont.; Hayden, Ariz.-
r/3Stl •
Bunker Hill, Kellogg, Idaho
Cities Service, Copperhill, Tenn
Climax Molybdenum, Fort Madison
Inspiration Consolidated Copper'
Jersey Minere, Clarksville, Tenn
Kennecott, Hayden, Ariz.; Hurley
City, Utah y
Magna Copper, San Manuel, Ariz
National Zinc, Bartlesville, Okla
New Jersey Zinc, Palmerton, Pa.
W ?LDS-96' Aj°' Morenci> Ariz,; Hidalgo, N. Mex
St. Joe Minerals, Herculaneum, Mo.; Monaca Pa
Paso,
Tacoma,
Iowa; Langeloth, Pa,
Inspiration, Ariz.
N.Mex.; Salt Lake
6.164 (6,797,100)
0.054 (60,000)
0.113 (125,000)
0.210 (231,000)
0.670 (739,000)
0.226- (249,000)
1.143 (1,260,000)
0-179 (197,100)
0.397 (438,000)
0.118 (130,000)
0.925 (1,020,000)
0.397 (438,000)
0.082 (90,000)
0.123 (136,000)
1.159 (1,278,000)
0.368 (406,000)
3.5-4
-------
The price of sulfuric acid can vary enormously. Current published
prices for virgin acid are $58.16 to $61.46 per Mg ($52.75 to $55.75 per
ton) on the Gulf Coast, $61.36 to $64.66 per Mg ($55.65 to $58.65 per ton)
in the Midwest, $63.73 to $67.03 per Mg ($57.80 to $60.80 per ton) on the
West Coast, and $65.60 per Mg ($59.50 per ton) in the Northeast.4 Smelter
acid, however, can be purchased more economically from copper, lead, and
zinc mining operations, mainly in the intermountain region and on the Gulf
Coast. For smelter producers with little storage capacity on site, the
necessity of disposing of acid can override other factors. Smelter pro-
ducers state that the prices paid for smelter acid have been $15.44 to
$19.85 per Mg ($14 to $18 per ton) on the Gulf Coast and $6.61 to $19.85 per
Mg ($6 to $18 per ton) in the West.4 Transactions are even reported at as
low as $2.21 per Mg ($2 per ton).4
Sulfuric acid is used in many ways, although primarily for fertilizers.
These uses are listed below:4
Use
Fertilizers
Petroleum refining
Copper leaching
Titanium dioxide
Hydrofluoric acid
Alcohols
Explosives
Aluminum sulfate
Ammonium sulfate
Iron and steel pickling
Cellulosics
Uranium milling
Surface active agents
Other
Total
Percentage of
acid produced
60
5
5
3
2
2
2
2
2
2
1
1
1
12
100
The outlook for the sulfuric acid industry depends on many factors.
Producers of virgin acid are in a cost/price squeeze: they face increased
costs for Frasch and recovered sulfur (the main raw material in acid pro-
duction), but must lower prices to compete with producers of inexpensive
smelter acid. If recovered, all the S02 emissions from utilities and smel-
ters could supply over 40.8 Tg (45 million tons) of sulfuric acid each
year.4 The economic feasibility of recovering significant amounts of sul-
3.5-5
-------
furic acid from power plants is restricted at present, but is expected to
improve greatly by the end of the century.* Smelter acid will probably
continue to keep prices down. Some anticipate that most smelter acid will
be used for metallic ore leaching, whereas others anticipate that technolog-
ical developments in leaching will make more smelter acid available for sale
on the open market. Producers of virgin acid tend to claim that mining
operations produce acid only to comply with pollution control regulations
and warn that buyers of smelter acid are. subject to the vagaries of the
copper, lead, and zinc markets.4 Domestic producers of virgin acid also
face intense competition from Canadian Industries, Ltd., and as the Canadian
government places more stringent limits on S02 emissions from metallic ore
mining, smelter acid from Canada is likely to become a stronger factor in
the U.S. market.4
3.5.2.2 Elemental Sulfur-
Elemental sulfur is produced by mining via the Frasch process and by
recovery from sulfur-bearing gas streams (sour gas). In 1974, U.S. firms
produced 10.7 Tg (10.5 million long tons) of elemental sulfur.8 Sulfur
mined by the Frasch process accounted for 74 percent of the total, sulfur
recovered by the Claus process from sour gas amounted to roughly 20 percent
of the total, and sulfur obtained as byproduct sulfur and from pyrite and
other sources supplied the remaining 6 percent.8
Table 3.5-3 shows the annual capacities of the major U.S. producers of
Frasch sulfur in 1973. Two of the 13 Frasch sulfur mines in 1973 were on
anhydrite deposits, and 11 were on sulfur domes.* All the mines and sulfur
processing firms were in Texas and Louisiana. Freeport Sulfur Co. and Texas
Gulf Sulfur Co. together supplied about 60 percent of the total U.S. capa-
city to produce Frasch sulfur in 1973.9
Recovered sulfur in 1973 was produced by 132 plants, of which 64 re-
covered sulfur from refineries, 61 from natural gas sweetening operations, 4
from coke ovens, and 3 from other sources. 9 The plants were located in 23
states. 9 About 40 percent of them were in Texas, and most were relatively
small.9 Table 3.5-4 lists the annual capacities of the five largest U.S.
producers of recovered sulfur in 1973.
The price of sulfur has varied widely during recent decades because of
successive periods of shortage and oversupply.10 The market could absorb
3.5:6
-------
TABLE 3.5-3.
ANNUAL CAPACITIES OF MAJOR U.S. PRODUCERS
OF FRASCH SULFUR IN 197311
Producer
Arco Chemical, Fort Stockton
Duval, Culberson County, Tex
Freeport, Garden Island, La.
Freeport, Grand Ecaille, La.
Freeport, Grand Isle, La.
Freeport, Lake Pel to, La.
Jefferson Lake, Long Point,
Pan American Petroleum, High
Texas Gulf, Bullycamp, La.
Texas Gulf, Fannett, Tex.
Texas Gulf, Moss Bluff, Tex.
Texas Gulf, Newgulf, Tex.
Texas Gulf, Spindletop, Tex.
, Tex.
Tex.
Island, Tex.
Total
Annual
10 Tg
0.183
2.540
0.813
1.422
1.524
0.610
0.305
0.051
0.305
0.178
0.305
1.524
0.686
10.446 (
capacity,
(long tons)
(180,000)
(2,500,000)
(800,000)
(1,400,000)
(1,500,000)
(600,000)
(300,000)
(50,000)
(300,000)
(175,000)
(300,000)
(1,500,000)
(675,000)
10,280,000)
TABLE 3.5-4. ANNUAL CAPACITIES OF THE FIVE LARGEST U.S.
PRODUCERS OF RECOVERED SULFUR IN 197311
Producer
Exxon Company
Getty Oil Company
Shell Oil Company
Standard Oil Company of California
Standard Oil Company of Indiana
Total
Annual capacity,
Tg (long tons)
0.260 (256,000)
0.231 (227,000)
1.077 (1,060,000)
0.165 (162,000)
0.464 (457,000)
2.197 (2,162,000)
3.5-7
-------
large quantities of sulfur with little decline in price during shortages."
In times of oversupply, however, the production of large quantities of
sulfur from S0x abatement could force prices down.
The future of the sulfur market depends considerably on the supplies of
sulfur produced by mining and by recovery from sour gas. Mining reserves
appear to be declining and may be exhausted within 20 or 30 years.10 Re-
ports indicate that old reserves are being depleted faster than new reserves
are being found and that costs are increasing as less accessible reserves of
lower quality are mined." The amount of sulfur recovered from sour gas is
increasing in several countries, especially Canada.™ Such byproduct or
coproduct sulfur is recovered independently of market demand and thus tends
to depress prices.
3.5.3 Plant Location
Plant location is an important factor in the sale of sulfur products
from SOX abatement. The large quantities of such products that could come
from a power plant require making large sales to single customers, selling
through an established marketer of sulfur products, or incurring high
marketing costs. Phosphate fertilizer plants are the dominant consumers of
sulfur products in the United States and are large enough to provide the
individual points of high consumption for S0x abatement products. For
example, a fertilizer plant producing 1 Tg (1100 tons) per day of phosphate
fertilizer (i.e., a plant in the current upper size range) uses 2.80 Tg
(3080 tons) per day of sulfuric acid, the possible output from 3000 MW of
power generating capacity.12
If high transportation costs are to be avoided, the phosphate fer-
tilizer plant should be fairly close to the power plant. Most fertilizer
plants, however, are on the Gulf Coast, an area of relatively little power
generation." Much S0x is emitted in the Northeast, where little fertilizer
is produced." Fertilizer and power production are nearly balanced in the
East North Central region." Phosphate fertilizer is heavily consumed in
the upper Midwest, where about 90 percent of all coal with a sulfur content
of more than 3.5 percent is mined." Recovery economics tend to be less
favorable in the East because, in general, the coal mined there contains
3.5-8
-------
less sulfur.14 The low cost of barge shipment may make the long-distance
transportation of SO abatement products economically feasible from power
plants on or near navigable waters.14
3.5.4 Outlook for SO Abatement Products
The EPA has sponsored a study by the Tennessee Valley Authority (TVA)
to evaluate the market potential for sulfuric acid and elemental sulfur from
SO abatement. The TVA has developed a cost model to determine the least
/\
expensive method for power plants to comply with air pollution control regu-
lations. Three methods of compliance have been considered: (1) burning a
clean fuel, (2) scrubbing with a nonregenerable limestone FGD system, and
(3) scrubbing with an FGD system that includes the recovery of sulfuric acid
or elemental sulfur. The TVA has investigated the distribution of S0x
abatement products in competition with existing producers for power plants
where the production of SO abatement products appears economically feas-
)\
ible. These investigations indicate that significant amounts of sulfuric
acid could be recovered from power plant flue gas and sold in competitive
markets.15
In determining the least expensive strategy for compliance with air
quality regulations, the TVA has devoted much attention to the clean fuel
alternative, which is defined as the increased price that a power plant
would pay for low-sulfur fuel to meet applicable regulations concerning S02
emissions.16 Several price increases have been considered for several
increments of heat input.*
The TVA suggests that large new power plants with high load factors are
most likely to find recovery of sulfur products economically attractive.
The boilers of the best candidate plants are usually less than 10 years
old.17 The average size is about 600 MW, and the average capacity factor is
about 60 percent.17
This study uses the International System of Units (SI). Boiler capacities
are expressed in watts thermal, the SI units for power. The available
data about boilers are given in English units and have been converted to
SI values.
Equivalent to $0.50 in mid-1983 dollars.
3.5-9
-------
3.5.4.1 Sulfuric Acid—
The TVA projected that 94 power plants with 165 boilers will be out of
compliance with air pollution control regulations in 1983.18 Sulfuric acid
production in 1983 was considered for five of these plants at an additional
clean fuel cost of $1.30 (in mid-1979 dollars) per MW thermal ($0.38 per
million Btu/h ) and for 26 plants at an additional clean fuel cost of $1.81
(in mid-1979 dollars) per MW thermal ($0.53 per million Btu/h*). « Detailed
analysis was limited to the latter situation, in which the combined annual
capacity of the 26 power plants to produce sulfuric acid would exceed 4.17
Tg (4,600,000 tons)." At only seven plants, however, is a market potential
in 1983 anticipated. The sales from these seven plants would total 1.128 Tg
(1,243,000 tons) in 1983, if the additional clean fuel cost is $1.81 (in
mid-1979 dollars) per MW thermal ($0.53 per million Btu/h*).6
3.5.4.2 Elemental Sulfur—
The TVA data for 1978 and 1983 show no sales of elemental sulfur from
SOX abatement systems." Such FGD sulfur could, however, become competitive
in some places with sulfur from Port Sulphur, Louisiana, if the total costs
of producing FGD sulfur were reduced by relatively small amounts. The cost
of producing FGD sulfur at 16 power plants is expected to be relatively low
in 1983.19 Sulfur from one plant could become competitive with a reduction
in total FGD sulfur production costs of 3.1 percent." Sulfur from the
other 15 plants could become competitive with reductions ranging from 5.4 to
23.7 percent." The annual production of the 16 plants would amount to
0.241 Tg (265,723 short tons or 237,252 long tons).19
Even if there are no economic reasons for producing recovered elemental
sulfur, it may be desirable to produce elemental sulfur because of the ease
of storage and disposal of sulfur as opposed to the problems of disposing of
calcium-based sludge.
Equivalent to $0.70 in mid-1983 dollars.
3.5-10
-------
REFERENCES FOR SECTION 3.5
1. O'Brien, W.E., and W. L. Anders. Marketing Alternatives for FGD By-
products: An Update. (Presented at the Flue Gas Desulfurization
Symposium sponsored by the U.S. Environmental Protection Agency. Las
Vegas. March 5-8, 1979.) pp. 11-13.
2. Ref. 1, pp. 13-15.
3. Slack, A.V., and G.A. Hollinden. Sulfur Dioxide Removal from Waste
Gases. 2d ed. Noyes Data Corporation. Park Ridge, N.J. 1975. pp.
146-147.
4. Sulfuric Acid Producers Fighting to Stay Even As Costs Rise but Smel-
ters Add to Market Supply. Chemical Marketing . Reporter.
2J5(19):3,9,14-16. May 7, 1979.
5. Ref. 3, pp. 146-147.
6. Ref. 1, p. 13.
7. Ref. 3, p. 146.
8. Katari, V.S., and R.W. Gerstle. Sulfur, Sulfur Oxides and Sulfuric
Acid Industry. Prepared for the U.S. Environmental Protection Agency
under Contract No. 68-02-1321, Task No. 25, by PEDCo Environmental
Specialists, Inc., Cincinnati, Ohio. October 1975. p. 1.
9. Ref. 8, p. 2.
10. Ref. 3, p. 147.
11. Ref. 8, p. 3.
12. Ref. 3, p. 145.
13. Ref. 3, pp. 145, 146.
14. Ref. 3, p. 146.
15. Bucy, J.I., et al. Potential Abatement Production and Marketing of
Byproduct Sulfuric Acid in the U.S. EPA-600/7-78-070. April 1978. p.
ii.
16. Ref. 15, p. xxviii.
3.5-11
-------
17. Ref. 15, p. xxxi.
18. Ref. 1, p. ll.
19. Ref. 1, p. 15.
3.5-12
-------
SECTION 4
COMBUSTION PROCESSES
4.1 NATURE AND EXTENT OF SULFUR OXIDE (SOV) EMISSIONS FROM COMBUSTION
s\
Combustion of fuels at stationary sources generates by far the greatest
portion of total sulfur oxide (S0x) emissions nationwide, which was more
than 82 percent in 1977.* This section deals with the combustion processes
that make up this important category of sulfur oxide sources. A brief
overview is presented, followed by a detailed discussion of currently avail-
able technology for the control of S0x emissions from combustion processes,
as well as the technology that is anticipated or under development.
Stationary combustion sources consist of electric utilities and indus-
trial, residential, commercial, and institutional sources. Figure 4.1-12
shows the total estimated SO emissions nationwide and the portion due to
stationary sources in the period from 1970 through 1977. This figure shows
a slight decrease (9 percent) in total S0x emissions during the period.
Although the SO emissions from stationary fuel combustion sources also
/\
decreased slightly (1 percent) over the same period, these sources represent
an increasingly greater portion of the U.S. total. This increase can be
attributed to the electric utility industry, which is estimated to have been
responsible for more than 64 percent of the total S0x emissions in 1977 -as
compared with 53 percent in 1970.3
Coal burning by utilities contributes greatly to the increase. Given
equivalent sulfur contents, coal firing would release almost 1.5 times the
amount of SO released in oil firing to generate the same amount of heat.
J\
In 1977 coal-fired power plants represented 37 percent of all the electric
generating capacity in the United States4 and were responsible for almost 92
percent of the utility SO emissions.5 Oil firing accounted for the
s{
remainder.
4.1-1
-------
en
H-
X
S!
32
30
28
26
24
22
20
18
16
14
12
10
8
6
4
2
0
' i i i 1 r
0
0 0
o
-
* • • * •
•
-A A * * A *
0 TOTAL NAT 1 01
• STATIONARY
A ELECTRIC UT
• INDUSTRIAL
O RESIDENTIAL
AND INSTITU
•
0 * o ' o " '
— 1 — 1 1 _l 1 l_
1 _
O <
0 •
_
A -
-------
Projections are that the use of electricity will increase, although
possibly at a lower annual rate because of energy conservation measures and
a lower rate of population growth. In 1979, 39 percent of the electric
generating capacity in the United States was from coal; the percentage of
electric generating capacity from coal is expected to be 41 percent by
1990.6 According to the National Coal Association, over 250 coal-fired
plants are being planned or under construction for the 10-year period
1978-1987.7 Utility coal consumption will increase correspondingly from 433
teragrams (477.5 million tons) in 19778 to more than 635 teragrams (700
million tons) by 1985.9 Although not as dramatic as the increase in usage
by electric utilities, the industrial use of coal is also projected to
increase and to reach 96.2 teragrams (106 million tons).by 1985.9
4.1-3
-------
REFERENCES FOR SECTION 4.1
1. U.S. Environmental Protection Agency, Office of Air Quality Planning
and Standards. OAQPS Data File. Durham, N.C. September 12, 1979.
2. U.S. Environmental Protection Agency, Office of Air Quality Planning
and Standards. National Air Quality, Monitoring, and Emissions Trends
Report. Research Triangle Park, N.C. EPA-450/2-78-052. December
1978. pp. 5-5 to 5-12.
3. Ref. 2, pp. 5-5, 5-12.
4. U.S. Department of Energy, Energy Information Administration, Office of
Energy Data Interpretation, Division of Coal Power Statistics.
Inventory of Power Plants in the United States. Publication No. DOE/
EIA-0095. April 1979. pp. xx.
5. U.S. Environmental Protection Agency, Process Technology Branch, Indus-
trial Environmental Research Laboratory. Overview of Pollution from
Combustion of Fossil Fuels In Boilers of the United States. Research
Triangle Park, N.C. EPA Contract No. 68-02-2603, Task No. 19. p. 24.
6. Reference 4, p. xxii.
7. Lin, K., J. Dotter, and C. Holmes. Steam Electric Plant Factors 1978.
National Coal Association, Washington, D.C. 1978. pp. 124-128.
8. Reference 7, p. i.
9. U.S. Energy Information Administration. Annual Report to Congress,
Volume II, 1977 Projections of Energy Supply and Demand, and Their
Impacts. Washington, D.C. DOE/EIA-0036/2. April 1978. pp. 185-186.
4.1-4
-------
4.2 CONTROL TECHNIQUES
Various methods of reducing sulfur oxide (SOX) emissions from combus-
tion sources are either available or under development. These control
methods can be grouped into four major categories: 1) fuel substitution, 2)
fuel desulfurization, 3) flue gas desulfurization, and 4) combustion process
modifications. Economics and/or status of technological development are
major determinants in the selection of an SOX control technique.
4.2.1 Fuel Substitution
The most straightforward method of reducing SOX emissions is to burn
fuels that cause lower SO emissions than those now in use. This can
/\
involve either the switching of fuels in existing sources (i.e. low-sulfur
coal or oil for high-sulfur coal) or the substitution of energy sources
(i.e., hydropower for coal-fired power plants). Because of problems
associated with the availability of cleaner fuels in the long term and
because of energy legislation, the substitution of coal with oil or natural
gas may no longer be an acceptable sulfur dioxide (S02) control technique.
4.2.1.1 Coal-
Coal is perhaps the only domestic fuel with which the United States can
meet projected energy demands beyond the year 2000. The U.S. demonstrated
coal reserve is estimated to be about 397 petagrams (438 .billion tons).
(See Table 4.2-1.)1 At a minimum recoverability of 50 percent and the 1976
domestic consumption rate of 535 teragrams per year (590 million tons per
year),2 the domestic coal supply would last at least 370 years.
The United States has large reserves of low-sulfur (less than 1 per-
cent) coal predominantly in the western states. (See Table 4.2-2.)3 It is
estimated4 that 99.79 petagrams (110 billion tons) of recoverable coal could
6S02 . emission standard would
reduce the amount of compliance coal available,~as~ shown in Table 4.2-3.5
Initially, firing of low-sulfur western coal., appears to be an ideal
method of reducing SOX emissions; several considerations, however, limit the
widespread adoption of low-sulfur coal firing. Much of the low-sulfur coal
4.2-1
-------
^
2E
f^ ^Q
^,
m2
UJ «
to i—
CD >.
>• §*-»
UJ ^ c
CO 1-3 o
UJ I —
cc «
C£3 | *
§lj
o^10
UJ U. 0
§ a '
fe 0-,-
to 3; CQ
CD UJ
^ s: cn
UJ o.
0 _J
•a:
t— <
. 1—
i" t 'J
1 t
CVI **
. o
«3- a-
UJ Q
CO«=C
1 "
re
re
o
CU
jjf
E
•r-
— '
co
•^
Q
CO 3
"*~^
•r-
to
3
g:
B
E
•r—
CQ
CU
B
rr
re
.E
c
•=C
•o
0
-^ .
+-> re
cu cu
E i-
re
CD
E re
•r-
x-^-^^-.
to co cn
i — i — CM
en £?.5
og
^ ^.^
0 CO CO
cn cn
* *
r*< r^*
CT> en
LO r-~ CM
LO CM CO
^— ,
CM en i —
UD «3" ^~
«j- CM to
^^* *"^~"
r*^ r**^
^ ^
LO JD, LO
CO CO
to to
S- S-
(L) CU
^ ^
S '£L
t
^* CX"O
Q. Q. E
CO to O
CO to S-
IO CO S-
to to o
•i — -r- "O
3 4- 4-
0 O O i—
s- re cu
Ol+J +-> +J o
S- to co o re
cu re cu i— <+_
-o uj s s-
""""i /^
i — i — i — co
•* o «a- co
*^-*N ^-* ^
O CO CM LO
CM to cn co
co en CM en
i — CO
r~ CO ^d" ^"
CO CO CO
i— «* LO LO
cn cn oo co
o cn o o
CM CO CO
* ^* *•• ^
i— \ — CO
to to vo
*•— -*^— ^ r~~
" —
CD «d- <^- CM
CO CO CO
• • •
LO LO CO
LO LO LO
' —
™™
cn oo r-. cn
CO — -«* CM
• » — -CM
*^^ -
CO LO
CO
to
S- S_
CU CU
^ ^*
•i— -r— |
Q± 0±
•i— ff—
c~> r*t
O. Q.
•i — -r—
to to
to to cuja
•1- •!— O 1
to to re re
CO CO «t- 4J
•i— -I— t- o
CO
4—4— T3
O O i— E
re re
•M 4-> 4J S-
co to o CD
re cu i—
UJ 2
II
cn
• f«
-r^
E
3
O
S—
cu
E
cu
o.
cu
E
0
cu
to
3 •
CO f — "*
o to
OJ E
-O O
re •*-)
0 0
-f-3 f~*
to
r—
re E
3 O
cr -i-
CU i—
^5 -f^
E
LO
re o
to Cn
+•3 O-
E
CU LO
E «3-
O
Q- O
1 E
o re
JE
o
to
E <"
3 CU
CO —1
re JD
re
0 1
•r- S-
CT) CU "O
0 Q_ 0)
O O •!-
cu cn E
C3 ID
• E CU
to re .E
13 O
E E "~
re cu cu
CU CO
to s re
CU -M JD
E CU
•i- JD CU
«/> s-
tf- CU CU
O T- co
S- CU
3 re s-
re >
CU i —
s- >, re
3 4-> 0
CQ -i- o
CU 'I- I—
J= JD i—
4-> re re
C r
>, CU 0)
O O
-o o
cu cu cu
E ct: jz
•r- 4J
% .c_
•o cu o
u
to re cu
re •— s-
o. o
to E
•I- •!- J-
S- 0
O i —
cn re 4->
CU 0 E
+-> 0 CU
re cj
u cu s-
J= CU
•0 4J Q. .
cu cu
re o 4-> JD
o 4- re
•r- +J -r- S-
T3 E u_ CU
E 0) >
•r- O O
S- • 0
T3 CU co CU
E Q.-4-5 S-
re •!-
CD CO CU
•0 0 O JD
CU i — Q.
S- CU 0
3 to "O 4->
re E ,— -o
cu cu re cu
E to 3 •(->
cu -o re
to S- •!- E
-r-
•o cu T- -t->
3 S- -O to
i— E CU
o -o ••-
E E 10
>— i re s. •!-
o
>>4- t^
• • cu cu
CU > 4-) ^J
••-> S- E 03
O 3 CU +->
2: oo o to
.
to
cu
E
•I—
E:
4-
0
3
re
cu
3
CQ
cu
o
3
o
to
4.2-2
-------
o
O
a:
to
>- Q:
CD eC
to eg;
et --3
CO
-t->
s-
1 o
O O r-
00-^
CD
Q I—
UJ UJ
CD
Cu
I— «a:
to •— i
Q O
o_
. O
^
CO
^3-
LjJ
_l
CQ
«=E
^_
ro
o
1—
o
c
c~
— ^
^jj
i CO -
A
QJ
CD
c
ro s^
i. CO
1
i_ 1—
3 •
^~ '
3
to
^i
r—
V
-^23
O
J=
•i-i ro
QJ QJ
E -S~
ro
CD
c: TJ
•r- E
,^_
s: .
*—~~^~— ~^— -.
co ID i--
tD CO CTt
i — i— CO
r~. r~. «3-
O^ ^t" ^"
• • •
vo co cr>
*± CO V£D
r— r— CO
"S'^'cTr
CO i — CO
«^-^^— •'x^ *•
CTi O CO
ID r^ co
• • •
CO CO ID
CO i— CO
^— - .«— ^-^— ••
«=}• co co
^_^ ^_^
VD ^-
CO tTi i —
1 —
^__
QJ
., —
rv s_
QJ
•1- >
Q--I-
CLO:
•r—
CO -I— ~O
CO Q. C
•t- Q. 3
co •<— o
co co S-
•i- CO CD
21 -r- S_
co QJ
QJ CO T3
"O 4-* y 3
c
3 CJ- <»- •—
o o o ro
S- 4->
CD+-> +J O
S- co co l —
QJ ro QJ
•O UJ 3
c
^—~**—^*—~~*~~*
r— •— i— co
«a- o «a- co
^— f r— ^— Kd"
s.^% — **^*>
O CO CO ID
CO 1^5 CTl CO
• ' i ,• •
r***- *~~ r^** t*^-
CO CTi CO CTi
i — CO
|-~- ID i — i —
r- ID
ID ^3- CO r>»
CO ID CTi CO
• • • •
LO ^^ C*l ^D
^J-
*~^*~^,^~*.* — ^
CTl CO O CO
i— ^-^CO O~i
«^ CO «* I —
co r^~ i— co
» • • *
r~^ co co ^i-
i— •— CO
S — ^ — ^*- — .* — ^
CO *J^ ^3" CO
CO CO CTi
tc CTI ^~ r^.
CO ID CO CO
r~ co o «a-
CO CO CO
r^u-Tu:^
*^ -^ CJ~> f«^ {|~j
• CO
^_*~
ID CO ID ID
co co cn co
VD CTl CO CO
ID UD CO
1 —
CO
QJ QJ
•i- S-
S- O
o E
CD
QJ C S-
•*-> -r- O
ro
U r— -l->
ro c
~O O QJ
QJ U 0
4_) ^_
ro QJ QJ
O -C Q-
•|~~ rii ^
"O >-,
C H- 4->
•r- O 4- '•
•^ QJ
T^J «4-J [ i f "
EC JD
ro QJ ro
0 • S-
T3 S- CO QJ
QJ OJ -J-3 >
S- Q--r- O
3 CO O
CO O O QJ
ro O Q. s-
QJ i — QJ
E 130)
CO JO
co •*-> i —
QJ c ro o
T3 QJ 3 +-1
3 CO T3
r— QJ -r- -O
U S- > QJ
C CL-t- +•>
i— i QJ T3 ro
S- E E
•r— •! —
. -0 4J
CD E S- CO
E ro o QJ
•r- It-
"O •> CO
E >,4-> ••-
3 QJ E
O > QJ to
S- J_ <_> O)
<4- to QJ ro
0 0.4J
i — CO
QJ ro CD
CO O CTl "D
3-i- QJ
ro CDTD •*->
(_> O E -t-
QJ i — ro E
-00 =)
QJ O
«> c~
D -0 E 4J
JT E QJ
co ro QJ E
CO LO •*->
; i — QJ QJ QJ
ro E jo co'
\ 4-> -r- TO
i O 21 CO JO
S- S-
OJ QJ
*i— •?—
cy p^
•r— «r—
Q. Q-
CL CL
•i — *i —
co co QJ
CO CO (J
•r- -r- ro i —
co co 4— ro
CO CO S- 4->
•r- -r- 3 0
s: s: ui +->
14- <4- i — "O
• • o o ro c:
QJ •*-> ro
ro co co i — o
M- ro QJ
S- UJ 3
3
to
+-> QJ
<4- -i- QJ
O O S- >
+J ro i-
3 > QJ
•o ro co
-O QJ >, QJ
ro S- 4-> S-
3 •!-
4-» CQ •— i—
O -r- ro
E QJ JD O
j^ ro o
^-» -f ^ S—
ro QJ r—
£>>>•—
ja o ro
ro u s-
4-> T3 QJ QJ
ro QJ o: >
Q E 0
•T—
<4_ . QJ
. . QJ QJ JE
QJ T3 O 4->
-t-J ro
O co i — 4_
z ro D. o
co
QJ
3
ro
QJ
S-
3
CO
QJ
o
o
to
4.2-3
-------
8
ce
LU
-JO
co
S
O O
co,
UJ
azcx:
UJ I—
to 2:
UJ O
oc. o
uj r>
cos
-------
available in the East is coking coal used primarily by the steel industry.
Capacity for mining and production of eastern low-sulfur coal is also limit-
ed. Further, the vast reserves of low-sulfur western coal must be
transported over long distances for use by the majority of the coal-burning
plants, which are located in the East and Midwest. Facilities for transpor-
tation from western mines to eastern markets may be inadequate. Apart from
the transportation problem, firing of low-sulfur western coal in boilers not
originally designed for it is limited by the need for substantial boiler
modifications or boiler derating.6 Finally, because the heat content of
western coal is generally lower, more western coal is often needed to
generate the same amount of power as eastern coal.7 Table 4.2-47 gives data
on eastern and western coal firing.
4.2.1.2 Oil-
Residual and distillate oil accounted for approximately 21 percent8 of
the electricity generated by utilities in 1977 and almost 28 percent9 of the
energy consumed by industrial/commercial boilers in 1975. At one time,
national policies directed toward reducing SOX emissions prompted operators
to convert their boilers from coal to oil firing. In light of recent
developments in the international oil markets, this method of reducing emis-
sions is no longer practicable.
Figures 4.2-110 and 4.2-210 show world production of crude oil and
daily demand in 1976. The U.S. dependence upon foreign producers to meet
demand is obvious. In 1977 the U.S. portion of crude oil production dropped
to 14 percent11 of the total production worldwide.
In 1976 the United States imported about 40 percent of all its
petroleum. Current U.S. policies aimed at reducing dependence on foreign
oil virtually eliminate the substitution of oil as an SOX control technique.
4.2.1.3 Natural Gas--
Twenty-seven percent of all U.S. energy consumption in 1976 was pro-
vided by natural gas. Domestic production accounted for about 95 percent of
consumption.12 If sufficient quantities were available, substituting
natural gas for coal and oil could lead to large reductions of SOX emis-
sions; however, the proved reserves of natural gas peaked around 1970 and
4.2-5
-------
CO
o
LU
OS
UJ
u_
a
ca
E:
ez)
o;
u.
CO
CO
CO
CM
O
CO
•r-_Q
CU
CU
fO
fO
O
CU
CD
CU
O CQ
£= IO
S- i— O
CU "CD
4-> O CD
fO 4-> >, to
S- C ^vJ
CU (O CM
CU O.G
CO
-3
to 5= o
o •*•
•r— •
S- O
4-> O S-
O O >
CU r—
C3 i
1/1 CO •>
CO
CTl
LT3
»— co
to
CO
CM
co
o
CO
CO
10 O
CO CM
-Q O
CO
CU ;
UJ
_J
CO
CU
CU
CU
to o
to *—i—
U 3
CD +->
c co co
S- CO
cu «o
•MOO
tO CM CD
cu en
r— O
CU
CO
CU
r^. en
CM CTl
r— CO
CO
LO
fCj
JD
CDr—
o ca
«£>
s- i— o
CU "O
o o
I CO «
i CO
CO
CO
cr>
o
o
i— CO
cu
LO
ID
O)
CU
o
CO r—
i— CO
cu
o
CO
r— O5-Q
O 4->
cr co CQ
i-CO
cu «o
o o
CM O
cr>
CU
CO
O r—
CU
cu
ro
CU
CU
CO
cr>
CM
ID
tn
CO
IO tO
to
C7> CO
•i- T- 4->
JE J= CU
CU
IO tO E
O O
•I— 'r- tO
Q- Q. CU
>5 >) O
h- I— Q
o -a
cu
4.2-6
-------
(C
O)
-o
o>
o
O)
Q-
s-
o
CM
CM
O)
S-
3
3«3HdSIW3H
O)
-o
M-'-
o to
r---
d) en
i- •—
(O
x: c
01 •!-
-a •<
T o
O -I-
3 -(->
U
cu
4.2-7
-------
to
ai
i
s-
ro
0)
o
_a
s_
ns
o
o
S-
O
OJ
CO
(O
cn
-a
3
cr
co
OJ
>
a>
co
O)
s_
•a
o
Q.
co
i
cvi
•=*
O)
O r-
suouuq *lN31\/Ainb3 110
4.2-8
-------
have been declining since that time, as shown in Figure 4.2-3.13 Because of
gas curtailments and declining production, projections indicate a slightly
increasing demand in the residential and commercial sectors and a reduction
in industrial and utility use. This will probably result in higher SOX
emissions because of the substitution by higher sulfur fuels.14
4.2.1.4 Source Substitution--
Other sources of energy that could alleviate SOX emissions are nuclear
power, hydroelectric power, and other possible power sources. Growth of
nuclear power has fallen short of early projections. Although capacity is
expected to increase, public concern for safety and environmental effects
are constraining growth of this energy source. In any case the reduction of
SO emissions due to increased utilization of nuclear power will probably be
minimal, since it will be replacing the relatively clean but scarce fossil
fuels—oil and natural gas—rather than coal.
The total quantity of hydropower is also expected to grow; its per-
centage of contribution to the total energy supply, however, will decrease.
Thus, the increases in use of hydropower will lead to no substantial reduc-
tions in SO emissions.
The effects of other energy sources such as solar, geothermal, and
wind power on the total energy supply will be minima-1 in the mid-term
(1980's), as will be their contribution to reduction of SOX levels.
Extensive and rapid shifts in the current modes of power generation and
energy use would be difficult and would entail a great technological devel-
opment.
4.2.2 Fuel Desulfurization
Desulfurization of fuel prior to combustion, like fuel substitution, is
a principal near-term solution to the SOX emissions problem. Desulfuriza-
tion of oil and natural gas has been practiced for some time; current tech-
nology can reduce the maximum sulfur content of residual oil to 0.5 percent
and of natural gas to less than 0.1 percent. It is unlikely, however, that
substantial reductions of total SOX emissions can be achieved by oil or gas
desulfurization, given the anticipated limitations on usage of these fuels.
Cleaning or desulfurization of coal provides greater opportunities to reduce
4.2-9
-------
SOX emissions due to combustion. Processes yielding synthetic fuels, such
as liquefaction and gasification of coal, are being developed primarily
to provide substitutes for limited oil and natural gas resources; however
these processes also reduce S0x emissions from combustion and are therefore
considered here.
4.2.2.1 Coal Cleaning--
physical and chemical coal cleaning processes are being developed
specifically to provide cleaner fuels.
Physical coal cleaning-Sulfur occurs in coal primarily in one of two
nonelemental forms, inorganic and organic. Organic sulfur is chemically
bound to the coal substance. Inorganic sulfur occurs with iron as discrete
particles of pyrite in the coal. Inorganic sulfur is amenable to physical
coal cleaning, whereas the organic form is not.
Reducing the quantities of ash-forming impurities has been the primary
function of physical coal cleaning. When some of the impurities are
pyrites, the sulfur content of the coal is reduced simultaneously." Up to
90 percent of the pyritic sulfur in coal can be eliminated by physical
cleaning, yielding a total sulfur reduction of 10 to 40 percent.16
Physical coal cleaning processes generally involve crushing run-of-mine
coal to a point where some of the mineral and coal particles are then
separated by techniques usually based on differences in the densities or
surface properties of the particles. The coal is then often dried. The
maximum sulfur removal obtainable with most coals is about 40 percent."
Most existing physical coal cleaning techniques depend upon differences
in the density of coal and the impurities it contains." Hydraulic jigs
hydroclones, concentrating tables, dense-medium vessels, and classifiers can
separate ground coal from the more dense impurities.
In the use of hydraulic jigs, a pulsating fluid flow stratifies coal
particles from top to bottom. The less dense cleaned coal overflows at the
top. This is the most popular and the least expensive coal washer avail-
able; however, it may not give an accurate separation. It is used on coal
ranging in size from 6 to 200 mm (% to 8 in.).19
4.2-10
-------
In the use of hydroclones, the separating mechanism is located in the
ascending vortex. The particles of various densities or specific gravities
move upward in this current and are subjected to centrifugal forces that
effect separation. For maximum reduction of pyrites and maximum yield of
cleaned coal, supplemental processes are used such as screening and froth
flotation. Hydroclones are most often used to clean flotation-sized coal,
but they can be used for coal as coarse as 64 x 0 mm. (% x 0 in.).
When froth flotation is used, a coal slurry is mixed with a collector
to make a certain fraction of the mixture hydrophilic. Normally this is the
coal fraction although it may vary with ash physical properties and the
collector used. A frothier is added, and finely disseminated air bubbles are
introduced into the mixture. Air-bubble-adhering particles of cleaned coal
normally are floated to the top of the slurry and are removed by a skimming
device. The froth is then broken, and the coal concentrate is recovered.
Thus both density differences and the surface properties of the coal and ash
are used to achieve separation. This is normally used with coal in the 1.17
to 0.044 mm (14 to 325 mesh) range.
Another common method of coal cleaning is the use of concentrating
tables. A pulverized coal and water slurry is floated over a table, which
is shaken with a reciprocating motion. The lighter coal is separated to the
bottom of the table, while the heavier, larger particles containing most of
the undesirable impurities move to the sides. These tables are normally
used with coal in the 0.15 to 6.4 mm (100 mesh to \ in.) range.
For dense-medium vessels, a slurry of coal is prepared in a medium with
a specific gravity close to that at which the separation is to be made. Lab
tests on the coal determine at what density the desired amount of pyrites
can be removed. The lighter, purer coal that floats to the top is contin-
uously skimmed off. A major advantage of this system is that it allows a
sharp separation at any specific gravity within the range normally required.
Dense-medium vessels can handle coal in the 0.59 to 200 mm (28 mesh to 8
in.) range.
With pneumatic or air classification, coal and refuse particles are
stratified through the action of pulsating air. The denser layer containing
pyrites is collected in pockets or wells and is removed. The upper layer of
4.2-11
-------
cleaned coal travels over the denser refuse layer and is removed at the
opposite end of the equipment. Air classification can be used with coal up
to a particle size of 6.4 mm (% in.).
Other physical techniques that may be developed and become widespread
are oil agglomeration, high-gradient magnetic separation, two-stage flota-
tion, and use of fine-particle dense-medium cyclones.20 Although it is well
established, current technology cannot produce coal that can ensure
compliance with S0x emission regulations, and no major changes are expected
before 1990.21 Table 4.2-5 shows the potential for sulfur reduction by
different degrees of physical coal cleaning.22
Chemical coal cleaning-Chemical coal cleaning methods are under devel-
opment to increase the potential for sulfur removal over that offered by
physical coal cleaning. Some chemical processes have claimed to remove more
than 95 percent of the pyritic sulfur and about 70 percent of the organic
sulfur.23 Some 25 chemical coal cleaning methods are being actively devel-
oped, and others are in the conceptual development stage.24 Because many of
the processes are in the early development phases, it is estimated that
achieving commercial operation will require at least 5 to 10 years.
Economics are expected to limit the amount of sulfur removed to 95
percent of the pyritic sulfur and about 40 percent of the organic sulfur.23
Table 4.2-6 shows the potential for sulfur reduction by different degrees of
chemical coal cleaning.25
Economic and environmental impacts-The mid-1979 capital cost of phys-
ical coal cleaning of northern Appalachian coals has been estimated for a
454 Mg (500 tons) per hour, mine-mouth coal cleaning facility as being
between $11,500 and $57,500 per Mg ($10,400 to $52,100 per ton) per hour
capacity. The mean cost range is $19,200 to $22,900 per Mg ($17,400 to
$20,900 per ton) per hour capacity.
For this unit, the annual capital charges and operating and maintenance
costs in mid-1979 dollars yielded a cleaned coal processing cost of $5.46
per Mg ($4.95 per ton) of cleaned coal. Other more detailed costs are shown '
in Table 4.2-7.
The cost estimated for chemical coal cleaning is more difficult since
most of the systems are in the developmental or pilot stages and have not
been demonstrated. Economic information (shown in Table 4.2-8) on eight of
4.2-12
-------
(O
0
*""
1 —
^£
l^i
1— 1
a:
3
U_
_l
.
_j
CO
UJ
o
5:
1— 1
^*
IE
a.
>-
CO
CO
o
1— 1
CO
tO
t— 1
s:
1 1 1
•z.
1 — 1
o
"^
^^
UJ
cr.
D;
o
Ll_
«=C
H- 1
1 —
UJ
\-
o
D-
.f
ifS
CO
•*
UJ
CO
2
C 3
0 -I-"
C -i- CO
o ^^
in 3 O
in ^ r—
•r- £ '
E 0 OJ
oj u o
CO
QJ C
cr o ja
10 i—
QJ QJ
> > *~3
"~ c'
'c
0 •
•»- in
4J C
u o
3 •!-
•0 in
Qi in
u'i
O) O)
CD
IO OJ
V- 0
<: c
fr~
OJ
4->
S.
°""S
CD O
10 E
OJ 1-
^
3:
g
'
Ol
2
O
— ^
u
jr
CD
C
o
en
CO
•
— •
o
§
i
o
o
OJ
o
OJ
r^-
^.
C
CL
O
ro
0
U
TJ
c
(O
OJ
=3
LO
«J
0
CO
o
en
OJ
i
0
OJ
r—
CO
r
Ol
c
0
r-.
o
en
ro
o
c
ro
i
'
LO
in
CO
ro
00
OJ
ro
s-
1
OO
U QJ
c •«-
O +->
E u
CJ CL
^
«3-'
1
OJ
OJ°
C=
en
00
(
0
LO
en
LO
•^
en
c=
o
ro
LO
in
LO
•3-
O
OJ
1
^
O
OO
o
I —
ro
10
i
ro
"SJ
i
OJ
laj
O >•>
l_> Ol
o
si
I- 0
3 0)
O 4->
in ro
CO !_>
in
ro
i
en
_
— "
o
0
LO
O
OJ
CO
r—
1
CO
o
o
1
O
ro
0
i
O
in
LO
t
o
ro
O
en
o
en
ro
O) en
-4-> O
O t—
0- 1
in i
QJ 1
CO *J
l/l
C
o
en
aj
S-
ro
o
u
X
in
E
O
S-
M-
QJ
en
ro
s.
OJ
>
ro
ro
o
u
^
o
CO
CO
a:
c:
£
o
(O
OJ
OO
g
ro
4->
ra
o
i/i
CD
CO
l/l
C
o
D.
cu
S-
4->
l^
QJ
3
in "O
C -i-
0 i_
o. c
QJ S-
t. QJ
4-J
i — t/>
ro Q>
o 3.
u
T3
C C
s- ro
QJ
4J ••
»/] 4->
QJ l/l
3 QJ
3
•a TJ
c: -i-
ro 2_
» c:
ro s.
E QJ
"ro ro
r— QJ
c: ro
ro -t~
i_) ro
ro i —
r— ro
ro CL
a. a.
QL «i
c s-
S- QJ
QJ .£1
3 0
O C
l/l
O S-
S_ 4-
4-
01
i/i i —
• — ro
ro o
o o
S- 3
3 *4-
U- i —
3 t/l
Ul 1
5 en
.a u
•o
c
-o
QJ
M
si
QJ
>
3
a.
c:
o
4-J
(_)
ro
s-
u-
-5^
C '
tr-
0
VJZ/ OO
. — m
4-» « —
CO
4->
•Q ro
QJ
4-J ID
ro QJ
ro ro
CL S_
QJ ro
in CL
QJ
"O l/l
ro ~o
c: i —
QJ ro QJ
M >
.,- GJ a/
i/i rsi t—
O. ui C
C 0
-t-1 c_-«-
O in
-• — 4-> l/l
"5 —'i
C -C QJ
"~ c? s.
OJ -^01
~^ -C
r— co en
• — ro -C'
+j
& E ro
QJ
i — un i-
CO CTl • r-
n CD
o
0 ' 0 4-)
4-> 4-J
"O "O QJ
QJ QJ -D
.c: _c ro
i/i en S-
S- S- CL
O O 3
•o QJ
+->
c
QJ
U
S-
QJ
d
C
CTi
QJ
>
QJ
_c
U
ro
O
4->
eauired
i_
tn
QJ
3
O
C
_c
u
QJ
4-J
c
O
ro
ro
D.
QJ
in
fO
Ul
CL
•0
c
ro
Cn
c:
_f—
Ul
U
o ro
c: o
o E
4-> S-
C QJ
.O •*-
0 >,
LJ CL
<4-
o
4-J U ~&
C -^ C
••- S- fO
4-> T3
in QJ c i ~o
c E ro Cf< Q>
0 -P -r- C
••- > 4J en ro
cr» ro c QJ
OJ S- QJ S- r—
i- O 4J Q> U
• C 4- EZ
— >-, O 4- 3
ro _o u ••-
o -o >»
o TJ t- c:
QJ 3 C: ro
X > M- O £ •
•I- O *— -r- Ul
ui E 3 O-4- QJ
QJ l/l QJ O 3
QJ S_ S- •—
JC 4-> »C
4-> QJ . ^1 C >
^D ul O QJ
4- QJ fC 4-> .C
O C in QJ C C7
U S- 1- X)
Qy ro c: H- rs c
_c *i~ 4- ro
£ 4-> Ul • —
O QJ i — 33
t_ QJ 4-J ro en O
4- +-> T- O r—
•i— S~ LJ QJ
r— I S- >, J^ QJ
ro >> CLr— 4-> x:
ol CL ro •*-*
U; " — 3 •*•
u- ra -o ui c
QJ 1 O O •*- QJ QJ
O* U > 3 QJ
ro. QJ "- *— S
i- ! CT. 4- TD ro •*-»
Qji re O C > QJ
> 4-J ••- ^J
.ro' c QJ =
QJ CnM- i — QJ
QJ u ro o ro d
s: s- 4-> c c
4-j QJ c i — o ro
o. QJ ro ••— s-
o. u ?•- en
3 QJ S- 4-J Q» QJ
l_ 4-J CL
Oi -t-J
3 4~> ro
ro c: i- m
4-) O ui O QJ ' —
c QJ ro •— > •—
ai JD 4-J ro ro
••- ui ro **-
C ui QJ N QJ
QJ O. l/l **~ -C ui
c o QJ 3 ro
o s- i- 4- E O
U D'- LJ i — O (J
C 3 t-
LO 3 "- ui **- "O
•r- o' QJ 1*
. — C "O >i M
4-J O i— "-
•— -0 "~ r- 4-> t.
c 4-J ro c 3
ro ro o ro 4—
QJ .c ro ui ••— 3
*j en CL >^4- >
O »r- QJ .C ••- QJ
ar -c ui o. c -o
4.2-13
-------
C\J
LU
LU
oo
CO
CO
co
o
o
LU
or
LU
VO
I
CM
LU
•3- i— i—
OJ ft
«C C
4.2-14
-------
OO
00
o
o
-
Q_
a:
i
CSJ
00
M-
r— O
r— i- S-
O O> O>
0 D- Q-
^^/
10 c:
S- D) 0
re s: -P M-
r— O
i— S- £-
o cu cu
O Q. Q.
>;
3 CU
+-> >
CQ O
cu
£_
o>
i — D.
re .c
•r— CT
a. -s:
re in
CJ +->
10
o
(_)
^
5^ *~" "^
•r- -C \
re o> o
re ^-^ •
C_5
.-— s
3
03
C
O
r—
^2
,_
E
"*+*S
_»
re
o
u
re
cu
r—
U
re
o
u
5
re
S-
^
_c
c
o
3
Q.
23
O
-P
D.
I — i
in
in
01
O
o
r\
i QQ ^^- c\j o"^ r^^ co CM LO LO oo r*^ * * r^^ oo LO
co co ^d~ LO ^^ LO LO t o r^^ oo t c* r^^ co co c\j co
rHrH i— IrH i— 1 rH rHrH i-HrH rH rH rH i— 1 rH r- 1
OO OO OO OO OC3 CDCD OO OO
_v^/^_XV_^^_'^-'
CTiLO «^"f^ i — ICMi — ICD OOU3 UDCO CDOO LOOO
CO CO ^J" CO ^" CO ^j" ^d~ ^" CO LO *^" CO CO CO CO
^^ V«^ •^** *^^S \^ff N^rf* ^-^ *^-^
r — r~^ T— i CM UOCNI UDO oocn co^ CMOT corH
T — 1 CT> CTs UD ^~ CM OO CO CD f — < CT» LO CO O CM CTi
CNJt-H CMCM CNJCM CMCNJ COCN1 COCO CMCM COCM
v^_^ ^^/ \^s **^ v^^/ • N^^x ^^/ ^^
^^Or-l^CMrHCO
.
CTicrioocTicnoocncn
^^<, ^~*\ /^^N S* ~^» ' /*^N /^^N X**S ^"^N
i^ ^3 ^- C3 ^D CD vo CD LO <~> OO CD LO CD P*^- CD
r^O OCD COO UDO COCD OOCD LOCD f>CD
r*^- *JD LO co co c\j LO ^t* LO ^j~ ^t* CNJ LO ^~ LO co
^^ rHrH rHrH rH T-| rHrH C\J CM rHrH Cs] CSJ
*^S S^_X ^^-^ ^^^ ***** Vfc*^ "^-X
^^^^^^^^
^--j ^—^ pN^ ^j. {*-) ^_Q «^j- prj o*> t.Q ^.Q CD "^ CD **O ^"
C\J LO *^" rH rH t£) ^^D LO CO CO CO ^J~ C\] **D OS r*^-
coco t^r*1^ LOLO oocn COCD co*d~ coco r*-
v^> **^+s ***** *•— •* rH Si*-y s*-^' ^-^
^d~ CD r*"^ CD P"*» CD i — 1 CD CO CD ^" CD ^t" CD CD CD
^- CD CD CD CD CD LO CD ^O CD *Cj~ CD ^J~ CD rH CD
LO UD CO CD CO CD ^3" *+Q CSJ ^" LO t*D LO VO CO CO
^^ rH rHrHrHrHrH ^-^ ' — ' ^^,
1 1 I 1
O) r— Olr— O>r—
ID- 101 101 ID. mo. mo. in a. 1/5D-
O) E Cn+J O54J CDE CE CE CE CE
• r- •!— T- C T- C T- O O) T- O) O O) O O) O
•-Oin --3'r- -TS-r- --OU OlO O.O C3U C3U
4.2-15
-------
CO
o
o
Ol
LO
cn
oo
cn
cn
LO
CD
cn
CO
cn
en
CD
LO
O)
O
o •
LO
- c:
05 O
-w-
a>
£- O)
ID
CO
LO
CT)
00
CT>
O1
in
cr>
ro
o
o
O)
c
to
o
0)
re
in
CO<0
4J O
•r- t-H
0.4/5-
03
CJ
CO
CM
LLJ
I
03
OJ O
N 4->
C D)
03 S
LO
CM
oo
CO
en
CM
CM
CM
CO
CO
CM
cn
oo
CM
01
CM
CM
CM
CM
CM
CM
CM
LO
LO
CO
CO
00
oo
00
CD
ID
LO
ID
LO
cn
LO
OO
CO
LO
«3-
CO
CO
CD
CO
CO
CO
CO
CO
CO
a>
u
u j-
r— QJ
re
u
a>
O) r—
S O)
03
CQ
O)
N
re
03
CO
CO
CO
CJ5
u
re
a>
O)
a>
r—
I/) O
O t-H
re .a
fO
4->
03
T3
••- S-
re oj
•a
c
03
03
a.
re
re v-
re
oo
CM
o
c
O)
S-
Ol
Ol
Q£
(U
O)
CO
u
4.2-16
-------
the processes is compared with that for physically cleaned coal in the 1978
report.
In addition to the economic considerations of coal cleaning, the envi-
ronmental impact of these systems must be considered. Physical coal clean-
ing is normally achieved in mine-mouth facilities, and the process wastes
are normally disposed of on site. Some major areas of concern in, a physical
coal cleaning facility are as follows:
0 Operations causing major emissions of air pollutants are infre-
quent in physical coal cleaning. The largest air emissions are
fugitive dust from coal handling, transfers, and size reduction^in
the grinding step and both particulate matter and combustion
products from coal dryers.
0 Water-related problems have essentially two sources: additives
used in the physical coal cleaning cells that can be introduced to
the outside environment as system purges or when cells are dumped
and heavy metals and inorganic compounds that may be leached from
the coal. Heavy metals are often introduced into aquatic ecosys-
tems as byproducts of acid mine drainage. Heavy metals can be
highly toxic and bioaccumulative. Work in this area is ongoing.
.° The amount of coal solids that remains at a physical coal cleaning
site depends on the raw coal properties and the degree of coal
cleaning. Besides the problem of the sheer volume of solids to be
disposed of, there are two concerns: fugitive dust problems and
leaching of inorganic compounds and heavy metals from the solids.
A study of the possible stabilization of coal preparation wastes
is being performed.
0 One other area of concern is the possibility of fire in reject
coal piles. This is considered a mismanagement problem, rather
than a process problem.
The environmental problems associated with chemical coal cleaning are
less well identified because of the status of development of many of the
processes. In general, however, there are particulate matter problems
associated with coal handling, transfer, and size reduction, and there are
particulate and at times solvent release problems associated with drying
steps. The gas stream from some of the reactors can include carbon monoxide
and nitrous oxides, which may present problems.
Solvent handling and the handling of various alkalis can result in
fugitive losses and can present problems to operating personnel. In addi-
tion many of the water-related problems noted for physical coal cleaning may
also be applicable.
4.2-17
-------
The primary solid discharge stream associated with many of these
processes is gypsum; however, the tendency of the heavy metals to remain
with the solid stream may present leaching problems.
4.2.2.2 Synthetic Fuels-
Synthetic fuel processes involve conversion of coal to a liquid or gas
and the production of oil from oil shale. Substitution of these synthetic
fuels for coal can reduce S0x emissions because chemical reactions remove
sulfur during the conversion processes. Technology for production of syn-
thetic fuels is being developed to incorporate pollutant control technology.
Coal conversion—Synthetic gas or oil is produced by allowing coal to
react with hydrogen and oxygen in the presence of catalysts or solvents
while applying heat and pressure. In most of the processes sulfur is
released from the coal as hydrogen sulfide, which can then be scrubbed from
the gases and converted to elemental sulfur. Further technological develop-
ment and more favorable economics are needed before any of the myriad of
conversion processes can reduce S0x emissions significantly.28
Coal gasification processes are often categorized as low-, medium-, or
high-Btu systems, in accordance with the heating value of the product. Each
system includes coal pretreatment, gasification, and gas cleaning steps.
Low- and medium-Btu gas was produced from coal for many years before cheap
natural gas became available. If and when production again becomes eco-
nomical, such gas will probably be used first in process heaters and steam
boilers.29
Gasification in conjunction with combined-cycle power plants will
probably prove economically feasible for the utility industry, but technical
improvements in gas turbines and in the efficiency of coal gasifiers are
needed.30,** High-Btu synthetic gas will probably be too expensive for
widespread use in large combustion units but could eventually supplement
diminishing natural gas supplies for use in residential heating.32
Nearly 70 different coal gasification processes are under development
or have been operated commercially. Table 4.2-9 lists the coal gasification
processes that are considered most promising.33
4.2-18
-------
TABLE 4.2-9. COAL GASIFICATION SYSTEMS33
Commercially available
Commercially demonstrated
or under construction
Chapman
Foster-Wheeler/Stoi c
Koppers-Totzek
Lurgi
Wellman Incandescent
Wellman-Galusha
.Winkler
Woodall Duckham/Gas Integrale
BGC/Lurgi Slagging Gasifier
Bi-Gas ,
Coal ex
Pressurized Wellman-Galusha
Riley Morgan
Texaco
a Since publication of Reference 33, the Chapman gasifier has become
commercially available.
As with coal gasification, the technology for coal liquefaction has
been available for some time. During World War II, Germany produced trans-
port fuel and 90 percent of its aviation fuel from coal.34 Research and
further development of coal liquefaction in the United States was minimal
until it was realized that domestic oil supply is not infinite and that oil
may not always be available from foreign suppliers.35
Coal liquefaction is similar to gasification. The ratio of hydrogen to
carbon is increased in the progression from coal to synthetic oil to syn-
thetic gas.36,37 Coal liquefaction processes generate an acid gas stream
containing sulfur and other contaminants. As with coal gasification
processes, removal of hydrogen sulfide and recovery of sulfur may be
necessary.38
One solvent-refined coal method (SRC-I) is being investigated as a
method to convert coal to clean solid boiler fuel and liquid fuel products.
As in coal gasification and liquefaction, the SRC-I process increases the
hydrogen content of the coal, but to a lesser extent in order to reduce
processing costs.39 In tests, the process has reduced the ash content of
coal to 0.1 to 0.2 percent40 and has removed 71 to 93 percent of the total
sulfur content.41 Reduction of SOV emissions would therefore appear to be
s\
comparable with that achieved through chemical coal cleaning processes.
Major advantages of SRC-I are that most utilities are already equipped to
fire solid coal, the operation and maintenance of ash handling equipment is
4.2-19
-------
reduced, and pulverizer maintenance will probably be reduced because SRC-I
is easier to pulverize.39,42
Another solvent-refined coal method (SRC-II) produces only liquid
products. The resulting low-sulfur liquid fuel is being investigated as a
replacement for fuel oil and as a compliance fuel.
Shale oil-The energy potential of shale oil in the United States is
second only to that of our vast coal reserves. Although shale oil develop-
ment has occurred sporadically in the United States since the 1800's, no
shale oil is now produced commercially.« Recent emphasis on energy self-
sufficiency has spurred development of this resource." Several different
methods of retorting the shale (applying heat to release the oil) are under
investigation.« W1th respect to ^ emissionS) oi] from shgle ^^ ^
attractive because the sulfur content is relatively low, 0.5 to 0 7 percent
by weight.45
The major problems associated with shale oil recovery are the handling
and disposal of the vast quantities of solid waste associated with the oil
recovery process and the demand for water, which is required for the process
but which may not be available in areas where the shale oil is located.
4.2.3 Flue Gas Desulfurization46,4?
Flue gas desulfurization (FGD) has become a leading means of control-
ling S0x emissions in the United States; it is primarily used by the elec-
tric utility industry. The operating capacity of FGD systems has risen
sharply from 900 MW in the pilot plant era of the late 1960's to about
25,000 MW in 1979 and is projected to reach about 62,000 MW by 1986.4« It
is estimated that as of late 1979, the total operational FGD capacity on
utility boilers was 25,000 MW, roughly equivalent to 2.8 Tg (3 million tons)
of annual S02 removal.
Dry or wet FGD systems can be used to control SO emissions; wet sys-
tems, however, dominate the entire utility market in the United States.
Flue gas desulfurization processes are categorized as regenerable or
nonregenerable depending on whether sulfur compounds are separated from the
absorbent as a byproduct or disposed of as a waste. Nonregenerable proc-
esses produce a sludge that requires disposal in an environmentally sound
4.2-20
-------
manner. Regenerate processes have additional steps to produce byproducts
such as liquid S02, sulfuric acid, and elemental sulfur. The nonregenerable
group includes processes such as lime and limestone, sodium carbonate, and
double alkali FGD systems. The regenerable systems currently in operation
are typified by the magnesium oxide and the Wellman-Lord systems.
A listing of all the operable utility FGD systems as of June 1979 is
given in Table 4.2-10.47 This table shows the equivalent MW of the gas
treated, the FGD process, and whether the system is a new or a retrofit
installation. Utility systems account for the bulk of the boiler flue gas
treated in the United States.
During test periods, all the major wet FGD processes in current use
(lime, limestone, double alkali, Wellman-Lord, and magnesium oxide) have
demonstrated high S02 removal capabilities.48 It is important to note that
most test data concerning S02 removal efficiency are from a number of indi-
vidual tests run on various days and do not necessarily reflect long-term
performance. Because many problems have arisen regarding the use of con-
tinuous S02 monitors for inlet and outlet S02 concentrations, most data in
the literature are for short-term S02 removal tests, and sometimes at
different operating conditions. Such tests are not representative of long-
term averages. Averaging times are not normally reported even for short-
term tests. The U.S. EPA Method 6 is the test procedure normally used for
determination of S02 emissions from stationary sources.
In this FGD section the following terms are used: availability,
reliability, operability, and utilization. Definitions of these words are
as follows:
Availability Index
Reliability Index
Operability Index
Hours the FGD system is available for operation
(whether operated or not) divided by hours in
period, expressed as a percentage.
Hours the FGD system was operated divided by the
hours the FGD system was called upon to operate,
expressed as a percentage.
Hours the FGD system was operated divided by boiler
operating hours in period, expressed as a percent-
age. This parameter indicates the degree to which
the FGD system is actually used, relative to boiler
operating time.
4.2-21
-------
0)
ro
•P
co
oo en
en ID
en
co en
co ID
ro co
en
co
rH ID
If) CO
0)
in
ro
OJ
0)
0) 0)
O) 0)
CD
O> CD
CD
P P
O O
S- S-
•P -P
CD CD
re
01
CD
CD
co
>-
oo
ID
W
O)
U
0) 0)
c n
o o
-4->
w in
(U CD
(U 0)
E C
o o
0) CD
E E
CD
CD
CD CD
CD 0)
u
CD
-P 0)
U -P
CD in
CD
CD
O
Q) CD
CD
O
co
LLJ
en en
cn
rH
in
in ID
en en
en o
rH IT>
rH CO
o o
U
_Q
CM CO
CO
CM CO
CM
ID
0) fl)
I— 0) OJ
O) D)
re T- •!—
F» f^ o
ro e £
JQ O O
ro f— (—
03 ^ £
•r- »r-
in > >
3 en in
-D CD CD
EEC
300
i— c_5 O
J U)
re
•r- +J
3 C
cr a)
UJ U
in CD
•r- Q.
O 0)
00 jC
UJ P
0)
ai
4.2-22
-------
CM
m
CO
CM
CM
rH
S-
0) o>
E E E E E E E
.,- .,- -r- -r- -r- -r- •!-
CM oo-*t g tn o
CO O1 CT> 1 — «*}• CM CM
ID 00 Lf) rH ^1"
CO CO ^" rH rH ^ LO
•P
O) .C
•r- 0)
1 >r—
1
o^J -C
S- 0)
CU S- T-
3 cu -J
°3
r> o *>c5
i^
in O5 S-
•i- S- >» CU
r-3 +JCCCU 3 ,UU
Q.in ooool O- 01. c c
res- z=j=>, s-cueu
ecu m-P-PO io<»-s_s-
re+J re33 reM-33
•t- o) mrerere mcurere
•oo- c n: 3: — i c >-3 -J — i
c re re
w :*: . ^
LO CD r^* CT> 00 CO
C7> 00 CM ^" 00 ^
^
+J +J +J +J .+J
i^I 14- 14- it- q-
O O O O 3 O
S_ t. S- S- 5-
+J 4_1 4_> +J Z -P
CU CU CU CU CO
cc: ce cc. ce. ct.
t
^__
re
re
CU CU CO i — CO CU
E e E re E E
•1— 'I— •!— 3 •!— T-
_l —I —1 Q — 1 — 1
«ij- 00 O 00 CM CM
ID CO 00 *t t*~
rH r-H CM «*J"
00
I *3" ID l£> CO ID
O
•r-
S-
-P
U
Ol
tn n i
CU
•1 OO
•p
•i— in
1 — s- re c
!J3 > CU Qi
o -i- o> c c: c ,c ,-««*« o ^
OO) >CUCUCUi — "O
zjcu inccci — "d
4«> t_ »r— ro re re *i~ re
CCP 3C_3CJCJS:C1.
O) O
it: —i
•o
O)
-p
c
o
o
4.2-23
-------
cu
4->
re
in ID
to
oo oo
re
•P
co
cr> in
ID
CM f-
.
00
CU
HI 0)
p -p
•I— »i—
<<- M-
O O f
S- S- 0)
•P -P Z
CD O)
CD CU
CU CU
0)
o:
CU
-p -p
O O
I- 5-
CU CU
(D tD
CU
O
re
re
cu
cu
m
cu
O) CU CU
•p p p
re re re
c c c
o o o
re re re
u u u
E E
•--
O O O
CO 10 LO
re
cu
cu cu
c: c
re re
CU O)
c c
o o
p p
in in
cu cu
£ £
CU
cu cu
cu
c
D)
re
i. s-
o o
c c
re re
cu cu
o
co
LU
o o
i£> to
oo co
in in in
CM csj CM
in
o o
Ol
iH O
ID ID
CM
CM
CM CO
CM
CM
CM
CU
u
0)
re
csj
LU
re
cu
re
cu
a.
o
O D)
O C
t- O
a» >-
re c
•p o
o -p
cu
CO
c in o
o re re re
o_ CD cs CD
re T3 t3 -a
•O •!- -I- T-
re cu
-------
•o
cu
3
C
• r-
-P
O
u
I
CM
00
s_
•P
to
<*-
o
-p
cu
x^
^
cu
U5
in
CU
O
s-
D_
O
LU
*
O
CU
E
Q
C
.p
•r*
— ^
v^
£
re
Q.
o
o
T^ Ol Ol
f^* f^ f^'
X- *X- *Xk
r* in co
cu cu cu
z z z
•1—
p—
cu cu re
c c ^:
O O r-
-P -p re
in in
CU CU r-
£ e ro
•1- T- 3
_J _J 0
o «tf- m
«3- 00 ID
rH rH CM
CM •* rH
O CU
t- C_3 i —
CU LU
to S-
CU oS
U 5
•r- 0 W
i— o. re
.n is
3 in
a. T- re
0 C
re c re
C •!- -I-
•p™ i"™1 "C3 C
r t ^— *^
O HH »-t O
t. S-
re X: c c c on
o re s- o s-
>j o> T- eu .
JE c ^ i- ^r CD
4J -r- -P (0 -P •
3 jC 3 S 3 <
0 O O
vj to to
oo 01 r^-
f^* p^* r^
*v. \. V.
00 ID «*
? S 5
CU CU CU
z z z
0) CU O)
c c c
o o o
•p -p -p
W 1/1 V)
cu cu cu
EH E
•i— *r— *r—
_J _l —I
CM CM Ol
rH rH rH
rH CM rH
U
• r-
S-
4-> V)
U CU
cu •<-
1— -p
LU •!-
r—
•r- *r-
Q. -P
D. =3
•r-
tn >,
t/> *p
•f~> *r-
i^ 5 5 U
in o o
•r- S- S- -O -P
S s- s- •— in
o o cu cu
S- <«- .C
eu . . OI-P
^: o o c: 3
4J . . -r-O
3 Q£ a: s- to
0 0-
> to
CM CM r-^
r*** r** r^^-
^\ ^^ ^^
«* «* m
+J -p -p
•r— "r— *i —
<*-<*-<*-
O 0 0
s- s- s-
•p -p -p
cu cu cu
OS Q: D£
cu cu
c c
o o
•p -P
in in
CU 0) CU
EEC
•r- -r- 0
— 1 — 1 -P
^v ^^* wi
cu a> cu
E E E
•1— •!— 'f—
_1 1 _J
O O O
rH rH ID
ID
<: m
o o oo
rH rH
>*
•P
• r-
S-
o
x:
-P
3
s
cu
r— Jbtf?
r- CU
re cu
o
0) CU CU
d) cu cu in
to c c S
in 5 5 o
cu ro re T3
C JC -C ••-
c to to 3
cu
1—
r~- oo 01 oo
r**- r*** r*^ r^*
^s^ *"***. "^^ ^^
'sj- ID CM ID
CU CU CU' CU
z z z z
a> cu cu cu
c c c c
o o o o
4-^ -P -P
-------
Utilization Index Hours that the FGD system operated divided by total
hours in period.
In addition to the type of FGD system installed, much interest has been
focused on the cost of these systems and also the difference between new and
retrofit installations. In March 1978, an EPA-sponsored FGD system cost
study was conducted on each utility having at least one operable FGD system;
these costs were adjusted to a common base time period and were evaluated on
a common cost basis. The reported and adjusted costs of the utility FGD
systems are shown in Table 4.2-11;49 the costs shown as the adjusted costs
are for a June 1979 base. (The basis used to adjust the costs is avail-
able.50)
The retrofit FGD systems often require higher capital costs than
comparable new systems because the systems must be built within space limi-
tations imposed by existing plant facilities. The layout of existing plant
facilities governs the location of FGD system equipment. The space avail-
ability for scrubbing equipment has a major impact on the cost of the sys-
tem. The actual space required for the scrubbing equipment is not large;
however, the equipment must be located near the existing stack to minimize
extensive duct runs. The support facilities such as feed preparation and
sludge treatment can be located away from the stack and in a manner required
by the existing plant facilities without significant cost increments.
In general, the existing plants are built with compact layouts to
minimize the duct runs and without an allowance for additional equipment
items such as scrubbing units. In some cases the boiler stacks are located
on the roof to take advantage of the elevation. These plants may require
long duct runs or construction of a new stack.
For nonregenerable FGD's consideration must be given to land required
for disposal of the sludge. The disposal area needed will depend upon
factors such as FGD capacity, S02 removal rate, and capacity factor;
however, these areas could be hundreds of acres over the life span of the
FGD system. If onsite disposal of sludge is not possible, then it will have
to be pumped or trucked to a disposal site, which will increase the
operating costs of the systems.
4.2-26
-------
10
O LU
z f-
Q to
LU
I— a
at. CJ3
o u-
o_
UJ _l
a: <
•z.
LU O
U_ Qi
O LU
Q_
to O
_i cc:
zs o
tO LU
LU
C£. I/)
' QJ
O CMCMPOCSJOf-iCMOOOOO
O)
E <
**• i—I «*• r-t ^- iH «*
(Q . . . . . ID . O rH • CM O
DO) "d" O ^f O ^" • O ' ' ^~
CO) rHi-Hr-irHr-*r--r-4lOl.Oi-iCOO
C C t I I I I I. I 1 .1 I. J.
^ ro
•D
t/i
•q
S- jQ
o
CL
gj ' r~<
COCO^-CSJCMCSICM CO I—i
o^-f^rHif)^tHaitftiDair«-
01
10
ir> - o • CT)
in f~> LO «~> «r> . ^3 > po tA •
01 - - - -• - — -•
c
m
Categ
•o
01
•u
U
£T3
O
O. t.
Q.
X u
+j ro
£* (0 in
oaia>ai
*>«-i— lE
ai £ re •»- -r-
>,
r— O)
t- -4J
re
01 C
C O
i- XI
f- «-
CD o
j^u
ai-<-
E-D
f o
t-
0)
r— X
«t 0
.
at
c
O 1-
c
o
•o
t. Ol
O TJ
Q.
"S.
ID
*
V)
O
t)
U
-------
The market for recovered sulfur byproducts can determine the economic
feasibility of regenerate processes. A byproduct market in the vicinity of
the FGD installation reduces the operating costs by the amount received from
byproduct sales, whereas nonavailability of a market or location some
distance away from the FGD installation may add^'to the operating cost by the
amount required for byproduct disposal.
Nonregenerable FGD processes produce significant amounts of sludge that
must be disposed of in an environmentally sound manner. The quantity of
sludge that must be handled will vary greatly depending upon the sulfur
content of the fuel ffred, the absorbent utilized, the FGD system removal
efficiency, the degree of oxidation of the sludge, the degree of sludge de-
Watering achieved, the use of fixation agents, and other variables. In
addition, some FGD installations remove fly ash, which combines with the
sludge. Disposal areas must be available in the vicinity to minimize trans-
portation costs. An FGD system at a 1000-MW power plant, for example, may
generate from 190,000 to 470,000 Mg (210,000 to 515,000 tons) of sludge per
year on a dry basis, not including the fly ash, assuming a calcium-based
(lime or limestone) FGD system and the firing of coal in the 3.5 to 7.0
percent sulfur range.51 By 1998, the application of 90 percent S02 removal
to all new electric generating plants is estimated to result in the produc-
tion of 157 Tg (173 million tons) (dry basis) of waste per year.52 The
actual quantities of untreated wastes requiring disposal are at least double
that amount, assuming a solids content of 50 percent or less.52
Coal-fired power plants in the United States have been disposing of
coal ash by ponding or landfill ing for many years and have extended these
conventional ash disposal practices to the disposal of FGD sludge. Several
Utilities, however, chemically treat or fix the wastes before disposal.
Fixation improves the structural properties of the waste and tends to
decrease leaching.
At present, no Federal criteria specifically apply to FGD sludge
disposal, but the Resource Conservation and Recovery Act of 1976, which was
signed into law in October 1976, requires that the EPA establish regulations
or guidelines for disposal of wastes from air pollution control systems such
as FGD.53 Guidelines for FGD sludge and coal ash disposal have been
4.2-28
-------
prepared and are under review. The EPA has been directing efforts since
mid-1975 toward preparation of documents that can be used to set FGD waste
disposal guidelines.54
The EPA has indicated that disposal of raw sludge is unacceptable. In
September 1975, the EPA declared "permanent land disposal of raw (unfixated)
sludge to be environmentally unsound because it indefinitely degrades large
quantities of land."55 Eventually, however, disposal of raw FGD sludge may
be allowed if means of containing it are proved to be environmentally
acceptable. Chemical fixation or seepage elimination through the use of
impermeable liners are both possible methods by which FGD wastewaters can
meet current criteria for groundwater or drinking water quality.
Approximately 10 state regulatory agencies have considered FGD sludge
disposal. They have allowed ponding of untreated sludges, landfill ing of
fixed sludges, and discharge of excess water.55
Currently it appears that ponding with a liner or in an impermeable
basin will probably be satisfactory as long as land reclamation is not
necessary. Analyses show that FGD sludge can contain toxic trace
elements;56,57 therefore sludge leachate or overflow and runoff are possible
sources of ground water and surface water contamination.57 If sound struc-
tural properties are required, chemical fixation of the sludge, possibly in
conjunction with an impermeable basin, would provide sufficient strength and
minimize leaching to ground water. Chemical fixation also reduces disposal
volume.
Ocean dumping and mine disposal are also possible alternatives for
sludge disposal which have been actively investigated by the EPA.58 Mine
disposal also has a potential side benefit of preventing mine subsidence.
The subject of sludge handling and treatment is addressed in a number
of other recent reports, in addition to those referenced previously
here.ss-ee
The physical and chemical properties of FGD sludge affect the choice of
alternatives for handling and disposal and also any possible future land
use. Primary sludge constituents that affect the chemical and physical
properties are water, fly ash, calcium sulfate, and/or sulfite. Spent
slurry drained from an FGD system can contain as much as 85 to 95 percent
4.2-29
-------
water.67 The spent slurry is partially dewatered and thickened to a heavy
wet sludge before its ultimate disposal.
Generally, the sludge components of calcium-based FGD sludge consist of
calcium sulfite hemihydrate, calcium sulfate dihydrate, fly ash, and
unreacted absorbent. The relative amounts of each depend on many factors,
including the kind and amount of fuel burned, the efficiency of sulfur
dioxide and particulate removal, the purity of the absorbent, and the boiler
type and operating practices. Sulfate sludges are more easily dewatered
than sulfite sludges and thus result in smaller volumes to be handled.
Generally, the higher the water content in the sludge, the less desirable
are its physical characteristics. Sulfate sludges are less thixotropic than
sulfite sludges. Thixotropic materials will flow or deform upon agitation,
which affects structural properties and subsequent land utilization.68 On
the other hand, sulfates are more soluble and have high permeability, which
may create landfilling problems for sulfates. In some instances, to meet
regulatory requirements, a river may be required. If a leachate removal
system is installed, discharging of large quantities of generated leachate
will be required.
Direct landfilling of FGD sludge is possible by dewatering the sludge
to a high-solids cake. Although high-solids filter cake has been obtained,
very little data have been developed on filtration requirements. The energy
and filtering capacity requirements for producing a high-solids sludge in a
cake form may prove this alternative to be economically unattractive.
The permeability of FGD sludge is a measurement of the rate at which
water can pass through the material. Untreated scrubber sludge has a
permeability of 10"4 to Iff5 cm/s, which is approximately the same as fine
sand.69 As a comparison fixed sludge is reported to have a permeability of
10~5 to 10"7 cm/s.70
Elements can be leached from the sludge solids and carried into the
underlying soil as liquids move through the sludge. Leachate composition is
a function of the chemical composition of the sludge, the solubility of the
elements present, pH, and the age of the disposal site. It is important to '
determine the rate of pollutant migration and the chemical composition of
the seepage. The U.S. Army Corps of Engineers and others have been con-
ducting such research for the EPA for several years.71
4.2-30
-------
The quantities and physical characteristics of sludge warrant serious
consideration of land use and reclamation problems. If the sludge is not
treated, it may not remain sufficiently dry to support loads. Over a 20-yr
period, a 1000-MW power plant could require an area 3.5 to 4.5 km2, 3 m deep
(860 to 1100 acres, 10 ft deep) to dispose of lime FGD sludge on a dry
basis.54
Reclamation of disposal sites depends on the load-bearing capacity of
the waste. The thixotropic nature of a sulfite sludge could prevent
reclamation and pose a permanent hazard. Sludge that has been sufficiently
dewatered and is nonthixotropic could be reclaimed and revegetated to
produce an area adequate for recreation or building.54
Although FGD sludge could be used to produce gypsum for use in wall-
board or Portland cement, most utility power plants will dispose of the
sludge. Sludge utilization may not be economically attractive at the
present time.72
4.2.3.1 Lime Process—
The lime process is a wet, nonregenerable S02 absorption process, in
which an alkaline slurry formed from the lime is circulated through a
scrubber/absorber tower, where it reacts with S02 in the flue gas. Calcium
sulfite and sulfate formed by the reaction are then separated in settlers or
clarifiers and filters. The sludge produced by the system can be chemically
stabilized to produce an inert landfill material or can be stored in sludge
ponds equipped with adequate barriers to prevent contamination of surface or
ground waters.
Lime FGD systems have demonstrated the ability to remove in excess of
90 percent of the inlet S02 at a number of utility boiler installations in
individual, site-specific tests.73,74 Facilities at which high, removal
efficiencies have been obtained are briefly described as follows:75
1) The Mohave Station of the Southern California Edison Company
reported S02 removal efficiency of 98 percent with lime. The
tests were conducted intermittently over 1 year on low-sulfur
coal. The unit was a 170-MW equivalent, prototype scrubber. The
operation of the unit has been terminated.
2) Recent short-term tests at the Paddy's Run Station of Louisville
Gas and Electric have shown S02 removal efficiencies in excess of
4.2-31
-------
99 percent on 3 percent sulfur coal. This extremely high removal
efficiency is attributed to the addition of magnesium oxide to the
lime slurry.
3) Several tests were conducted at the 10-MW TVA Shawnee Pilot Plant,
where S02 removal efficiencies of 95 to 99 percent were reported
for lime-based systems.
4) At Bruce Mansfield Station of Pennsylvania Power, S02 removal
efficiency of 93.2 percent was reported for an FGD on Unit 1 for a
short-term period in September 1977. 76
Process chemistry— The following reactions take place in the absorber
during S02 absorption by an aqueous scrubbing liquor:
HSO
(1) S02(g) -» S02(aq)
(2) S02(aq) + H20 -> H2S03
(3) HSOs -» H+ + SOs
Lime in the slurry produces calcium through the following reactions:
(4) CaO + H20 -* Ca(OH)2(s)
(5) Ca(OH)2(s) -> Ca(OH)2 (aq)
(6) Ca(OH)2(aq) -»- Ca++ + 20H~
Sulfite ion generated (Reaction 3) combines with calcium generated
(Reaction 6) to yield the insoluble calcium sulfite hemihydrate:
(7) Ca++ + SOg + 1/2H20 -*' CaS03-l/2H20
In addition, sulfite ion may ultimately be converted to gypsum in the
following reactions:
(8)
(9) Ca
+ 1/202 -» 504
++
$0
4 2H20 -* CaS04-2H20(s)
Quantities of lime required by the process and sludge generated in the
process are calculated from Reactions 1 through 8. With assumptions of 95
percent lime purity and a 1.0 molar stoichiometric ratio, the lime require-
ment is 1.05 weight per unit weight of S02 removed.
4.2-32
-------
The sludge produced in the process consists of calcium sulfite hemi-
hydrate and gypsum, lime impurities, and any excess lime. The exact propor-
tions of calcium sulfite hemihydrate and gypsum are primarily functions of
the S02 content of the flue gas, the percent excess oxygen, and whether
forced oxidation is applied. An assumption of equal proportions gives a
sludge formation rate of 2.5 weight units per weight unit of S02 removed.
Because the FGD sludge is always generated wet, the weight of associated
water must be included.
System description—The equipment for a lime FGD system is generally
grouped under four major operations:
0 Scrubbing or absorption—includes S02 scrubbers, holding tanks,
and circulation pumps
0 Flue gas handling—includes inlet and outlet ductwork, dampers,
reheaters, and fan
0 Lime handling and slurry preparation--includes lime unloading and
storage equipment, and lime processing and slurry preparation
equipment
0 Sludge processing—includes clarifier and filters (if used) for
sludge dewatering, 'sludge pumps, and sludge handling equipment
A diagram of a typical lime FGD system is shown in Figure 4.2-4.
Individual systems may deviate from that shown, depending upon plant charac-
teristics and system manufacturer.
A forced draft fan (not shown in the figure) forces the flue gas via
ducting and dampers through the absorber, .in which the S02 is transferred
from the flue gas to the circulating slurry. The flue gas then passes
through a mist eliminator to a heat exchanger, where the flue gas is often
heated to about 80°C (175°F) before it is exhausted to the atmosphere.
Reheating the flue gas reduces the possibility that condensation of water
vapor, with its attendant acidification as sulfur oxides are absorbed, will
create a corrosive environment for the ducting, stack, and the fan, if an
induced draft fan were used. Reheat also increases the plume buoyancy. A
number of reheat plans are in use or have been suggested to reduce the
heat/energy requirement.77
4.2-33
-------
OJ
O
CM
4.2-34
-------
Slurry from the absorber goes to the hold tank, from which a fixed
amount of slurry is bled off and sent to the sludge circuit; an equal amount
of fresh lime slurry is added to the hold tank. Lime handling and slurry
preparation require lime unloading equipment, lime storage silos, conveyors,
a lime slaker, grit removal equipment, and slurry tanks.
The sludge must be dewatered to make it suitable for disposal. Water
recovered from the sludge dewatering circuit is returned to the scrubbing
circuit to maintain a closed loop system to prevent direct contamination of
fresh water supplies with effluent and to confine products of FGD to the
sludge. In most FGD installations in the United States, the sludge solids
are allowed to settle in a holding pond on the site. Flow to the holding
pond either comes directly from the recirculation tank or from the clarifier
(as partially dewatered sludge). Water from the sludge pond is returned to
the scrubbing circuit.
The availability of an FGD system is in most part dependent upon the
availability of the absorbers, whose basic function is to promote contact
between S02-laden gas and calcium in the circulating slurry. A residence
time of 1 to 4 seconds is necessary for effective S02-calciurn contact. The
two major parameters of an absorber-liquid to gas ratio (L/G) and the
typical pressure drop are functions of the absorber configuration and type,
the S02 removal required, and the inlet S02 content of the gas stream. The
energy requirement of an absorber is made up of two components: 1) the
energy needed to overcome the system pressure drop and 2) that needed to
circulate the lime slurry. To achieve maximum S02 removal with an optimum
amount of energy input, industry operators use various types of absorbers.
The major types are discussed in the following paragraphs. -,
Tray absorber78--A tray absorber promotes gas-slurry contact in a
vertical column with transversely mounted perforated trays. The S02-laden
gas enters at the bottom of the column and travels upward through the
perforations in the trays; the reagent slurry is fed at the top and flows
toward the bottom. Absorption of S02 is accomplished by countercurrent
contact between the gas and reagent slurry.
A schematic drawing of a tray scrubber is shown in Figure 4.2-5.
4.2-35
-------
MIST
ELIMINATOR
CLEAN GAS
IMPINGMENT
TRAYS
S02-
LADEN GAS
TO RECYCLE
TANK
Figure 4.2-5. Tray absorber.
4.2-36
-------
Packed scrubber79—A packed scrubber consists of an absorption tower
filled with packing material designed to provide a large surface area for
gas/liquid contact. The reagent slurry is fed at the top and travels down-
ward, wetting the packing surfaces; the gas travels upward from the bottom
through the packing material. The packed tower design offers a large area
for contact of reagent and S02-laden gas and provides the longest residence
time among all of the scrubber types.
These absorbers require careful control of reagent flow rates, which
must be high enough to prevent blowing and dry gas channeling, but not so
high as to cause flooding. Normal operation is 0.2 to 0.8 kPa/m (0.25 to
1.0 in. H20/ft) of packing, corresponding to 40 to 70 percent of the
flooding velocity. , ,
Mobile bed scrubber80—The mobile bed scrubber, shown in Figure 4.2-6,
extends the concept of a tray type scrubber. It consists of perforated
trays or grids filled with mobile elements such as plastic spheres and
mounted transversely in a vertical column. Flue gas introduced at the
bottom travels upward through the mobile packings; the reagent slurry, fed
at the top, flows downward. This countercurrent operation, coupled with the
action of the mobile spheres on the transverse trays, produces highly
turbulent zones in the scrubber. Each sphere is free to rotate, and the
constant movement results in self-cleaning of the spheres.
A typical overall pressure drop for a mobile bed scrubber is 1.5 to 2
kPa (6 to 8 in. of H20).
Venturi scrubber—In a venturi scrubber the S02-laden gas is introduced
at the top, passes through a converging section of the scrubber (the venturi
throat), and then exits the scrubber through a diverging section. The
venturi shape imparts high velocity to the passing gases at the throat. The
reagent slurry is also introduced at the throat, leading to turbulent mixing
of the gases and reagent slurry. This thorough mixing promotes a chemical
reaction between the S02 in the gas and the absorbent.
The annular orifice design, as shown in Figure 4.2-7, has the con-
verging section, throat, and diverging section. The gas impinges on either
a fixed or movable disc while liquid flows cocurrently down the walls of the
4.2-37
-------
TURBULENT-
CONTACT-
BED-STAGES
GAS INLETS
WASH
HEADERS ^
MIST
ELIMINATOR
GAS OUTLET
•*o ooooooooo
SPRAY X
HEADERS
WATER SPRAYS
7
0 °n° oSo^^^^O^O^
I W o
,0 0_ o
°0
>«4
! I I
l I
I I
I 1 I
WATER LEVEL
•~l
1
Figure 4.2-6. Typical mobile bed scrubber.
4.2-38
-------
SCRUBBING
SOLUTION
GAS OUT
GAS IN
TO REACTION
TANK
Figure 4.2-7. Two stage venturi scrubber.
4.2-39
-------
converging section. As the gas stream exits the throat, the gas stream
diverges, which accomplishes much of the liquid separation.
In the rod bank tower design, parallel rows of horizontal rods are
placed in the throat of the venturi, perpendicular to the gas flow The
reagent slurry is introduced just ahead of the rods and also on the walls of
the venturi section.
Spray tower-A spray tower scrubber can be vertical or horizontal The
reagent slurry is introduced in the scrubber in atomized droplets through
the spray nozzles at the top. The flow of gas and slurry is crosscurrent in
a horizontal design and countercurrent in a vertical design.
Figure 4.2-8 shows a modification of a conventional spray tower
scrubber supplied by M.W. Kellogg Company. Slurries of varying degrees of
richness can be introduced at the different stages in the tower. Often the
fresh slurry (recycle and makeup streams) is introduced at the rear of the
absorber (the last stage) where the S02 content of the gas stream is lowest
The slurry collected in the last stage is pumped forward to the next stage
In effect, the slurry "flows" countercurrent to the gas flow. The first
stage of the absorber has the highest S02 concentration gas stream and a
slurry that has had much of its active alkalinity exhausted.
Sludge d1sposa1"-8«-Process1ng of the sludge generated by an FGD
system may involve several steps. A stream is bled continuously from the
scrubber to the sludge circuit. Because this stream contains a large
proportion of water (90 percent is not uncommon), liquid-solid separation is
required. The FGD sludge is thixotropic.
The major constituents of lime FGD sludge and the typical percentages
are 73 percent CaS03.l/2H20, 11 percent Ca(OH)2, 11 percent CaS04.2H20, and
5 percent CaC03. The percentages vary from system to system.
A typical sludge processing circuit (Figure 4.2-9) involves solids
sedimentation, dewatering, fixation, and transportation of sludge for final
disposal. Clarifiers are generally used for sedimentation; the recovered
water is sent back to the scrubber circuit, and the partially dewatered
sludge is sent for further dewatering. When vacuum filters are used for
further dewatering, the dewatered cake contains about 60 percent solids
Any further dewatering leads to excessive energy consumption.
4.2-40
-------
>
Q-
rra
sz
o
N
'sZ
O
co
CXI
OJ
CT!
4.2-41
-------
CD
0)
O
0)
CO
fB
O
CT>
OJ
•CD
4.2-42
-------
After the vacuum filtration, the cake is transferred to a fixation
tank, where chemicals are added to the cake and mixed thoroughly. The
fixation process is a means of physically and chemically stabilizing the
sludge to reduce its pollution potential and facilitate handling. Two
companies, Dravo Corporation and IU Conversion Systems, Inc. (IUCS), have
systems in operation at utility FGD sites. The processes of both companies
use proprietary fixation agents to produce pozzolanic reactions.
In the Dravo process, a proprietary additive called Calcilox is added
to the sludge. CalciloxR is a hardening agent derived from blast furnace
slag. The fixed sludge is physically more stable, stronger, and less
permeable than untreated sludge.
The IUCS process utilizes pozzolanic (cementitious) reaction princi-
ples. The company markets a physico-chemical fixation system called
Poz-0-TecR, which it claims produces a sludge that is ecologically accept-
able. The Poz-0-TecR process incorporates the sludge into a chemically
stabilized matrix. The sludge is trapped and encapsulated within the hard
and relatively impermeable matrix. The process involves addition of lime,
dry fly ash, and other substances to the dewatered sludge.
Additional information on sludge handling treatment and disposal can be
found in the FGD Sludge Disposal Manual published by Electric Power Research
Institute (EPRI FP-977).
Energy and environmental impacts85-87—Operation of FGD equipment
requires significant amounts of energy. The FGD components consuming major
amounts of energy are the ID fan that overcomes pressure drop in the
absorber and ductwork, and the process pumps that maintain the flow of
fluids in the system. The pressure drop through the absorber depends on
such factors as the number of trays, type of packing, the height of the
unit, and L/G ratio. Energy is also used in reheating the gases as they
leave the absorber before they are discharged to the atmosphere.
Transportation of scrubber sludge to a distant disposal site also
consumes energy. This energy may be tapped from the electrical' energy
generated by the plant if the sludge is transported by pumps.
The energy requirement of the FGD system reduces the capacity of a
utility plant by the amount needed to operate the FGD system. For a new
4.2-43
-------
plant and FGD installation, the FGD energy requirement can be factored into
the design of the generation system, but this- will lead to higher initial
costs and higher annual amortization charges. At existing plants where an
FGD system is retrofitted, the energy must be tapped from that generated by
the plant for sale. Some installations may be required to compensate for
the reduced capacity by purchasing the additional energy from a power pool
at higher cost.
The loads for utility boilers vary during the day according to elec-
tricity demand. When a boiler is not operating at or near its rated
capacity, the energy requirement of an FGD can be met by generating addi-
tional energy. When a boiler is operating at or near its rated capacity,
the additional energy requirement must be purchased from a power pool or
other sources.
The concept of capacity penalty is illustrated in Figure 4.2-10, in
which a hypothetical load demand curve is assumed. The solid horizontal
line shows the maximum rated generating capacity* of the plant. The solid
curve indicates the load demand curve; the dotted curve indicates the total
plant load with its FGD operating. For the illustrated plant, the total
energy demand exceeds the generating capacity in the regions indicated as A
and B. The shaded areas indicate the amount of energy purchased from out-
side sources.
The energy requirement of a lime FGD system ranges from 3 to 5 percent
of the boiler capacity when 100 percent of the flue gas is treated and
reheated. This range excludes the energy required for sludge transporta-
tion. The energy required to overcome pressure drop is a function of
scrubber type and system configuration. The energy required for reheating
flue gases is a function of the degree of reheat and bypass, which in turn
depends upon the permissible exit gas temperatures in the stack and duct-
work. When only part of the flue gas is treated, the remainder is bypassed
directly to the stack. This gas stream is not subjected to temperature drop
of the FGD system and is hotter than the gas stream leaving the FGD system;
the higher temperature of the bypass stream reduces the amount of reheat
needed.
4.2-44
-------
(O
c
<1)
Q-
Q.
CO
O
c
o
rd
•t-J
i/l
I
CNJ
(U
cn
QN\/W3Q QV01
4.2-45
-------
The amount of bypass will depend upon the S02 removal requirements. In
general, to meet NSPS no bypass will be possible while firing high-sulfur
coal because of high S02 removal requirements. When low-sulfur or medium-
sulfur coal is fired, however, it may be possible to meet the NSPS by by-
passing a fixed quantity of flue gas. The amount of bypass will depend on
the applicable NSPS, the sulfur content of the fired coal, and the maximum
S02 removal potential of the FGD system.
The operation of an FGD imposes energy and capacity penalties on the
plant. Figures 4.2-11 and 4.2-12 show the relationships of FGD energy and
capacity penalties for a 500-MW plant firing bituminous coal. An energy
penalty is represented as the percentage of total generating capacity The
energy used by the FGD system does not depend on averaging time. A capacity
penalty represents an instantaneous derating in boiler capacity by the
amount required to operate the FGD system. The derating depends upon the
maximum power to be reserved for the system during sulfur peaks. This study
treats the capacity penalty as a percentage of total generating capacity.
Additional information on the energy and environmental impacts of FGD are
given in Section 3 of this report. The energy penalty curve is based on a
reheat requirement of 28°C (50°F). The sudden change in the slope of the
energy penalty curve at a gas flow rate of about 70 percent is due to the
need for reheat energy beyond this point. At gas flow rates below this
point, the bypass stream temperature provides the needed heat and eliminates
the reheat energy requirements.
The environmental impacts of lime FGD are positive with respect to
reduction of gaseous pollutant emissions and negative with respect to dis-
posal of sludge. The S02 emissions from the system depend on the sulfur
content of the boiler fuel. Sulfur dioxide emission rates as low as 50 ng/J
(0.12 lb/10* Btu) could be achieved with a low-sulfur fuel. The emission
rates could be as high as 800 ng/J (1.9 lb/106 Btu) with high-sulfur fuels.
A lime FGD system designed for S02 removal can also remove some partic-
ulate matter from the incoming gas. In general, a system designed for both
particulate and S02 removal will require a scrubbing circuit for particulate
removal in addition to the S02 removal equipment.
4.2-46
-------
10 20 30 40 50 60 70 80 90 100
0.5
10 20 30
40 50 60
GAS FLOW THROUGH FGD,%
80 90 TOO
Figure 4.2-11. Energy penalty for a lime.FGD system utilizing
bypass heat at a bituminous-coal-fired 500-MW plant.88
4.2-47
-------
10 20 30 40
60 70 80 90 100
10 20 30
40 50 60
GAS FLOW THROUGH FGD,
70 80 90 100
Figure 4.2-12 Capacity penalty for a lime FGD system at a
bituminous-coal-fired 500-MW niant 89
4.2-48
-------
If the stack gases are not reheated, localized emissions of acid mist
may occur. Some of the S02 and S03 remaining in the flue gas could condense
to form sulfurous and sulfuric acid in the stack under such conditions.
Proper choice of stack liners can reduce corrosion problems from this con-
densate.
Sludge disposal is the major potential environmental problem associated
with FGD systems. A lime FGD system generates approximately 3 to 5 weight
units of sludge (depending on the solids content of the sludge) for each
weight unit of S02 it removes. A 1000-MW plant burning 3.5 percent sulfur
coal and 14 percent ash, would.generate about 190,000 Mg (210,000 tons) of
FGD sludge per year at 90 percent S02 removal efficiency.90 This plant
would consume 2,152,000 Mg (2,373,000 tons) of coal per year and generate
201,000 Mg (222,000 tons) of ash requiring disposal.90
Operational status and current developments91-93—As of the third
quarter of 1979, lime FGD systems represented 24 percent of the total
committed capacity for utility FGD systems in the United States. The system
is available from various vendors, who are joining with other groups and
agencies in major efforts to improve such aspects of the process as chemical
control, sludge stabilization, and S02 removal efficiency.
The products of reaction have caused plugging and scaling of the
absorber in lime FGD systems. Scaling and plugging in lime systems,
however, have been relatively insignificant as compared with other calcium-
based FGD systems. Scrubber internals and mist eliminator surfaces are most
susceptible to scaling; plugging can occur in such components as nozzles and
tray passages.
The solutions have been improved for chemical control of the absorbent
loop; for example, seed crystals and scrubbers/absorbers with reduced
surface area have been used to control precipitation.
In efforts to improve S02 removal efficiency, some operators are using
reagent additives such as magnesium oxide (MgO) and adipic acid. Dravo
Corporation has patented an MgO-promoted lime reagent called Thiosorbic
lime and claims S02 removal efficiencies of 90 percent and greater by U.S.
EPA Method 6 tests.
4.2-49
-------
The EPA studied the effects of adipic acid at a pilot plant in North
Carolina and at the TVA/EPA Shawnee test facility near Paducah, Kentucky.
Adipic acid in concentrations of 700 to 1500 ppm in the scrubber liquor have
been reported to yield S02 removal efficiencies above 90 percent by U.S. EPA
Method 6 tests. The investigators report only minor'differences in the pro-
cessing of sludge and no scaling problems.
System costs^-^-Because of differences in system designs and in cost
estimating procedures, the costs reported by various FGD installations in
the early 1970's varied widely and followed no definite pattern. Most of
these systems were the first of a kind and incorporated various safety
factors. These early efforts, however, have led to standardization of
system design, and the range of reported costs has narrowed. The early
efforts have also generated enough data for development of cost models to
predict the costs of an FGD system of a given size.
The cost of an FGD system is often presented as a capital cost and an
annual cost. Capital costs indicate the total initial investment necessary
to install the system. The annualized cost indicates the cost of operating
and maintaining the system and also the annual expense of repaying the
initial investment; the annualized cost thus represents the total annual
revenue requirements of the FGD system.
Capital costs consist of direct and indirect costs incurred up to the
successful commissioning date of the facility. Direct costs include the
costs of equipment and the labor and material required for installing it and
interconnecting the system. Indirect costs are expenditures for the overall
facility that cannot be attributed to specific equipment; they include such
items as freight and spares.
The annualized costs consist of a capital cost component and an oper-
ating cost component. The capital cost component includes the charges for
using the capital investment: equipment depreciation, taxes, insurance, and
interest paid for capital investment. The operating cost component includes
the costs of raw materials, utilities, labor, maintenance and repair, and
overhead.
A sensitivity study of the individual cost items of a typical 500-MW
FGD system indicates that the S02 scrubbing module of the system constitutes
4.2-50
-------
a major portion of the capital costs. Table 4.2-1297 shows the results of
such a study. The original costs were estimated on the basis of 3.48 per-
cent sulfur in the coal and 90 percent S02 removal. A similar analysis of
annualized costs is shown in Table 4.2-13.97
Various publications predict the costs of FGD systems. Figures 4.2-13
through 4.2-16 present charts for predicting the value of various components
of a lime FGD98-101 system. The figures are intended for use with a combi-
nation of input parameters, the S02 removal requirements being a major one.
The costs for an FGD can be found by selecting the combination of S02
removal requirements and percent of flue gas treated. When the S02 removal
efficiency required to meet the regulations is lower than the maximum
achievable efficiency of the FGD system, partial scrubbing of flue gas is
possible. The percent of flue gas requiring S02 removal will depend upon
the sulfur content of coal, applicable . regulations, and maximum possible
efficiency of the selected FGD system. The cost basis of these charts is
mid-1978 dollars. Figure 4.2-13 shows the capital investment for system
equipment excluding the sludge pond and land costs, which are shown
separately in Figure 4.2-14. Figure 4.2-15 shows the operation and main-
tenance cost excluding energy costs. The fixed charges are shown in Figure
4.2-16. The costs presented in these figures do not include the costs
associated with energy or capacity penalties.
Table 4.2-14 shows the capital investment and annualized costs of the
operating lime FGD systems in the United States. These costs are escalated
to reflect mid-1979 dollars.
4.2.3.2 The Limestone FGD Process--
The limestone and the lime FGD processes are similar in many aspects.
In the limestone process, limestone slurry is used as the absorbent, as is
lime slurry in the lime process. The use of limestone, however, requires
different feed preparation equipment than is used in preparing lime slurries
and also necessitates other process differences; the limestone process, for
example, requires a higher liquid to gas (L/G) ratio because the absorbent
is less reactive than lime. The exact L/G required is a function of S02
removal required, the inlet S02 gas stream concentration, the absorbent
inlet pH, and other items. Even with such differences, the processes are so
4.2-51
-------
TABLE .4.2-12. SENSITIVITY ANALYSIS OF A
500-MW LIME F6D SYSTEM
CAPITAL INVESTMENT97
Lime preparation
Conveyors
Slakers and pumps
Storage silos
Storage tanks
Pumps and motors
S(L scrubbing
Absorbers
Fans and motors
Heat exchangers (reheaters)
Soot blowers
Valves and ducting
Hold tanks
Pumps and motors
Sludge disposal
Clarifiers
Chemical storage
Mobile equipment
Tanks and agitators
Pumps and motors
Total installed cost
Raw material inventory
Sludge pond
Total direct costs
Total indirect costs
Contingencies
Contractor fees
Land cost
Total capital investment
Percent of total direct cost
Individual
cost
1.4
0.4
3.0
0.7
0,1
39.9
5.1
7.8
3.0
4.1
2.6
7.5
2.5
0.1
0.2
0.2
0.6
Subtotal for
module
5.6
70.0
3.6
0.8
20.0
27.7
8.3
0.4
Total
;
79.2
100.0
38.3
174.7
4.2-52 '
-------
TABLE 4.2-13. SENSITIVITY ANALYSIS OF A
500-MW LIME FGD SYSTEM
ANNUALIZED COSTS^7
Percent of total annualized cost
Individual
cost
Subtotal
Total
Operation and maintenance
Lime
Sludge fixation chemicals
Water
Electricity
Reheat
Direct labor
Supervision
Maintenance labor
Maintenance supplies
Plant overheads
Payroll overheads
Sludge handling
Fixed charges
Total annualized costs
11.4
1.9
0.1
6.7
,1
.2
0.
10.
1.5
6.4
0.2
1.9
43.2
56.8
100.0
4.2-53
-------
130
Z-0 3.0 4.0 5.0 6.0 7.0 8.0
9.0
GAS FLOW THROUGH FGD =100%
0 8.0 9.0
S02 REMOVED, lb/10° Btu
(1 lb/106 Itu • 429.6 ns/J)
Figure 4.2-13. Capital investment excluding cost of sludge
pond and ]and for a lime FGD system at a bituminous-coal-fired
' ' 500-MW plant.98
4.2-54-
-------
10.0
1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0
8.0
^ 6.0
E 5.0
4.0
3.0
«*» 2.0
i.o
tt:GAS FLOW THROUGH FSD -100%
' "
=20%-
10.0
9.0
8.0
7.0
6.0
5.0
4.0
3.0
2.0
1.0
1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0. 9.0
S02 REMOVED, lb/106 Btu :>
(1 lb/106 Itu • 429.6 ng/J)
Figure 4.2-14. Capital cost of sludge pond and land for a lime
FGD system at a bituminous-coal-fired 500-MW plant. yy
4.2-55
-------
5.0
2.0 3.0 4.0 5.0 6.0 7.0 8.
GAS FLOW THROUGH FGD -1002
3.0 4.0 5.0
S02 REHOVED. lb/106 Btu
(1 Ik/106 ttu • 429.6
6.0 7.0 8.0 9.0
Figure 4.2-15. Operation and maintenance cost excluding electricity
and reheat for a lime FGD system at a
bituminous-coal-fired 500-MW pi ant.100
4.2-56
-------
1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0
5.0
GAS FLOW THROUGH FGD =100°-
SOZ REMOVED, lb/106 Btu
(1 Ik/106 Btu • 429.6 119/0)
Fiqure 4 2-16 Fixed charges for a lime FGD system at a
bituminous-coal-fired 500-MW pi ant JOT
4.2-57
-------
TABLE 4.2-14. CAPITAL AND ANNUALIZEDQ7
COSTS OF OPERATIONAL LIME FGD SYSTEMS
Station-unit(s)
Conesville 5
Elrama 1-4
Phillips 1-6
Hawthorn 3-4
Green River 1-3
Cane Run 4
Cane Run 5
Paddys Run 6
Col strip 1-2
Bruce Mansfield 1-2
System
capacity, MW
400
570
410
200
64
178
183
65
720
1650
Adjusted costs (mid- 1979)
Capital cost,
$/kW
81.88
147.10
162.60
100.96
89.74
93.21
78.06
88.47
89.40
119.00
Annual ized cost,
mills/kWh
8.58
9.03
9.91
5.03
6.06
6.68
6.43
7.53
4.70
10.04
Design S02
removal , %
89.5
83
83
70
80
85
85
90
60
92.1
4.2-58
-------
similar that it is possible to design a system that can use either lime or
limestone as the absorbent.
Process chemistry—The chemistry of the limestone process differs from
that of the lime process only in the way that the calcium ion becomes avail-
able for absorption of S02; calcium ion generation in the limestone process
takes place according to the following reactions:
CaC03 (aq)
•++
(1) CaC03(s)
(2) CaC03 (aq) •* Can
Other reactions are the same as those in the lime process.
The S02 in the flue gas stream contacts water forming the sulfite ion,
which reacts with the calcium ion to yield insoluble calcium sulfite hemi-
hydrate or gypsum, as shown in Reactions 7, 8, and 9 in the lime process
description.
The quantities of limestone required and the volume of sludge generated
may be calculated from the chemical reactions. With the assumption of 95
percent limestone purity and a 1.3 molar stoichiometric ratio, the require-
ment is 1.37 weight units of limestone per weight unit of S02 removed.
The sludge from a limestone process consists of calcium sulfite hemi-
hydrate, calcium carbonate, gypsum, limestone impurities, and any excess
unreacted limestone. The exact proportions of calcium sulfite hemihydrate
and gypsum depend on such factors as inlet S02 content, excess oxygen, and
absorbent pH. Because the FGD sludge is always an aqueous slurry, the
weight of the associated water must be included in the total sludge weight.
System description102--Most of the equipment in a limestone FGD system
is similar to that in a. lime-based system. The major difference is in feed
and slurry preparation. In a limestone system, the feed generally must
undergo size reduction before the slurry is prepared; although preground
rock of less than 200-mesh particle size can be purchased and used directly
for slurry preparation, this is rarely done because of the high cost.
Therefore, a limestone FGD system generally incorporates equipment for size
reduction.
4.2-59
-------
The raw limestone can be stored in open piles without protection from
weather. Because it can be stored in the open and because it is cheaper
than lime, large inventories can be maintained.
The S02 removal efficiency of the system ^s dependent on the limestone
stoichiometry up to a certain level; after this level is achieved, no
increase in S02 removal is obtained with an increase in the limestone
stoichiometry.
Unclassified byproduct limestone from quarries may be an ideal feed.
The unclassified crushed stone contains a high percentage of fines, and
although it requires special handling equipment, this limestone splits
readily along bedding planes, forming irregular, thin plates or blade-shaped
pieces that interlock and chip easily into fine particles.
Wet ball mills at the FGD installations grind the raw limestone feed to
the size suitable for slurry preparation. The crushed limestone slurry from
the ball mill is sent to a classifier, from which the larger-size particles
are fed back to the ball mill for regrinding.
The description of the lime FGD system in Section 4.2.3.1 is applicable
to the limestone system except that the feed preparation modules are
different because of the difference in the properties of the two feed mate-
rials. Other aspects of the system as well as the types of scrubber/
absorber are essentially similar.
Figure 4.2-17 is a diagram of a typical limestone FGD system. Indiv-
idual systems may deviate from that shown, depending upon plant character-
istics and the system supplier.
Because limestone is less reactive than lime, some of the process
parameters are different. The L/G ratio of a limestone system is higher,
and the residence time in process tanks is longer than that of the lime
system. Typical L/G ratios for limestone FGD systems using spray tower
absorbers often are in the 8 to 11 liters/m3 (60 to 80 gal/1000 acf) range.
Spray absorbers are the most common in limestone systems. The design of
scrubbing module equipment also requires a consideration of limestone reac-
tivity. 103
The configuration and design of a reaction tank affect the chemistry of
limestone dissolution. A plug-flow reaction tank designed to prevent the
4.2-60
-------
E
O)
CO
CO
Q
C£S
u_
QJ
O
CO
O)
fO
O
03
M-
O
to
en
Q
f^
CO
-------
backmixing of the reacting stream is reported to yield significant improve-
ments in limestone utilization and in S02 removal. 10« This plug-flow design
apparently drives the additive dissolution reaction further toward comple-
tion and makes more liquid-phase alkalinity available for reaction with
sorbed S02.104,105
In design of the scrubber/absorber for a limestone system, pH of the
circulating slurry must be considered. A low PH leads to better limestone
utilization but reduces S02 removal efficiency. The scrubber should be
designed for optimum S02 removal. Some proposed designs take advantage of
the flexibility afforded by operating multiple stages at different pH
levels. Borgwardt reports 97 percent limestone utilization with 81 percent
S02 removal in a two-stage, forced-oxidation, pilot system.106 A low-pH
liquor contacts the entering flue gas in an initial scrubbing stage, in
which part of the S02 is removed and the limestone dissolution is essential-
ly completed. High-pH liquor contacts the flue gas in the second stage,
where additional S02 removal takes place. Slowdown liquor from the second
stage is bled to the first stage. Makeup limestone slurry is added in the
second-stage loop:
Another design that may enhance limestone utilization is the Weir
scrubber. This is a multistage, crossflow scrubber, in which the lowest-PH
liquor contacts the entering flue gas and the highest-pH liquor contacts the
exiting flue gas. It is expected that this countercurrent flow arrangement
will optimize limestone utilization and S02 removal efficiency. The Weir
scrubber has not yet been tested in commercial limestone systems, but five
are currently being installed.107
Operational status and current developments-As nf the third quarter of
1979, limestone FGD represented 40 percent of the total committed capacity
for utility FGD systems in the United States.91 Figure 4.2-18 shows the
growth of limestone capacity in the United States from 1968, as estimated
through 1982; the figure includes the capacity growth trend for all the FGD
systems combined.108
Efforts to improve efficiency and other process areas of the lime and
limestone systems are carried on simultaneously, and usually a new develop-
ment in either process is applicable to both.
4.2-62
-------
CM
g
O
ra
Q.
rO
CD
O
CO
01
S-
OJ -.
CO
r^
CTi
CTI
:=> uj
H- i-
O£ OO
-
I— 00
OO
UJ
u_ z:
o O
>- •—i
puesnoqj) '
CM
1^
en
O
i^-
cr>
CO
in
CM
o
CM
CM
CM
LO
I—
LD
CM
LT>
in
CM
o
-a
OJ
E
O
cn
OJ
O
S-
cn
QJ
4->
0
a>
•i~5
O
s-
O-
T3
C
(O
-P
O
CO
I
CM
a>
CD
puBsnoqq.) 'AllOVdVO Q9J
4.2-63
-------
Current listings of operational limestone FGD systems are given in the
EPA Utility FGD Survey (published quarterly) and the EPA Industrial FGD
Survey (published semiannually). Operating problems and successes are well
documented in these publications.
Some additional operating histories for limestone FGD systems are as
follows:109
At the Mohave 170-MW facility of the Southern California Edison
U>. , S02 removal efficiencies in excess of 95 percent were
reported for a Turbulent Contact Absorber using limestone slurry
for a low-sulfur coal application.
At the Martin Lake Station of Texas Utilities Generating Company,
S02 removal efficiency of 99 percent with 90 to 92 percent lime-
stone utilization is reported for the FGD system on Unit 1. The
99 percent removal was achieved using wetted film contactors; the
efficiency dropped to 80 to 85 percent without using wetted
Reports on the
show 92-percent
scrubbing.
packed-bed module at the 115-MW Cholla facility
removal of sulfur dioxide using limestone slurry
has reported S02
During one run in
the TCA unit on a
The 10-MW test facility at the TVA Shawnee
removal efficiencies in excess of 90 percent.
February 1976, efficiency reached 96 percent in
high-sulfur coal application.
The initial operating performance of the Widows Creek No. 8 550-MW
scrubber facility indicates that S02 removal efficiencies sub-
stantially higher than the designed 75 percent value have been
measured. S02 removal efficiencies in the range of 85 to 94
percent were reported during the period of November 1977
February 1978.1J1
through
During the St. Clair No. 6 limestone scrubbing demonstration and
test program conducted by Detroit Edison and Peabody, S02 removal
values of 90 and 91 percent (90 percent design) were measured on
high (3.0 percent) and low (0.4 percent) sulfur coals
At the 167-MW
hdison Co., S02
Will County No. 1 facility of the Commonwealth
removal efficiencies as high as 86.8 percent were
- i-ia hJ9h-sulfur "a! (4.0 percent) burn program,
yielded inlet S02 loadings averaging 3573 ppm.
S02 removal efficiencies in excess of 90 percent have been
measured in tests conducted by Combustion Engineering and Kansas
rnnf ?"? 9 A at Lawrence ' Unit N°- 4, in 1977. The testing was
conducted on the recently installed rod scrubber and spray tower
4.2-64
-------
absorber system, which replaced the original limestone injection
and scrubbing system (1968 startup). Actual efficiencies in the
95.5 to 97.5 percent range were measured for low-sulfur (0.55
percent) coal.
Some additional information on several systems of particular interest
are as follows:
0 The Unit No. 1 steam generator of the La Cygne Power Station of
Kansas City Power and Light is a 820-MW (net) system that has_one
of the earliest limestone scrubber systems installed in the United
States (1973). The sulfur content of the coal ranges from 5 to 6
percent. Despite the problems at startup, the availability of the
system improved steadily.
removal efficiency.
with the seven
load, the removal
This system was designed for 76 percent S02
Actual S02 efficiency has been 80.18 percent
modules operating at 720 MW. Under maximum
efficiency averaged 76.2 percent.112,113
0 The No. 1 and No. 2 units of the Sherburne County Station of
Northern States Power Co. have a net capacity of 700 MW each and
burn 0.8 percent sulfur coal. Availabiliy for Unit No. 1 averaged
85 percent for the 4 months of operation after startup. During
one 12-month period, availability was in excess of 90 percent.
Unit No. 2 has shown even better startup performance, with opera-
bilities averaging about 95 percent for the first 4 months. For
the first 8 months of operation of No. 1 unit, the S02 removal
efficiency was 50 to 55 percent, which was sufficient to meet
local regulations.112,113
Based on the operating experience of limestone systems, there is
evidence from individual S02 removal test runs to show that limestone FGD
systems can operate at 90 percent S02 removal or greater and that they can
operate reliably (90 percent operability) with proper design and maintenance
on both low- and high-sulfur coals.114,115
System costs—The capital cost of a limestone FGD system is higher than
that of a comparable lime FGD system. This higher cost is attributable to
the additional limestone preparation equipment and to the larger sludge
processing system necessitated by the higher sludge generation rate of the
limestone process.
Table 4.2-15 lists the costs of limestone and lime FGD systems
installed on units burning Eastern coals with sulfur contents of 3.5 percent
and 7.0 percent.116 The FGD systems are designed to comply with a 516 ng/J
(1.2 Ib/million Btu) S02 emission limitation.
4.2-65
-------
CU
LO co LO
CD in r^-.
LO LO
co co
CU .
o
CU !
t-^ co
CO CO (
oo
co
CM CM
CO
c\j o
re
oo
LU
oo
oo
o>
OvJ r- r— C\J I—
C\J LO ID
Lf5 CO LO
1 CO ^~ (^ *^- CD
• en r-. co co LO
"^ co r^. vo i—•
CM O i— LO CO
o o o o o o
O O CD
C\J LO O
000
CM LO CD
oo
oo
ra U
re
CD QL
CO
r— CU
re +J
o to
cj >,
CU
re
UJ
cu
re
4.2-66
-------
Table 4.2-16 shows the capital investment and annualized costs of
several operational utility limestone FGD systems in the United States.96
These costs are escalated to reflect mid-1979 dollars.
To help put the costs given in these two tables into perspective, the
following items should be known:
0 The FGD systems in Table 4.2-15 include spare absorber modules and
equipment redundancy, not found in some older FGD systems, to help
assure greater reliability of operation.
0 Also the systems in Table 4.2-15 were designed to meet 90 percent
S02 removal.
t ° For coal of about 3.5 percent sulfur, the capital costs in both
tables are close.
For FGD systems designed to meet S02
than 90 percent, the cost is reduced.
removal requirements lower
The annualized costs reflect the higher capital
in Table 4.2-15. Capital costs are often a
annualized costs.
cost of the units
major portion of
Energy and environmental impacts—Because the limestone and lime FGD
processes are similar, the energy and environmental impacts are also
similar. Although there may be minor quantitative differences in the sludge
quantities and energy requirements, these are not significant. A secondary
energy impact of lime systems is the energy required in the lime calcining
operation, which is normally achieved outside of the utility. The energy
consumption for various lime processes ranges from 4.7 to 6.4 GJ/Mg (4 to
5.5 million Btu/ton) of lime.117
The discussion of energy and environmental impacts in Section 4.2.3.1
is generally applicable to the limestone system.
4.2.3.3 Double Alkali Process-
Several FGD processes may be classified as double or dual alkali sys-
tems. Basically, double alkali scrubbing is an indirect lime/limestone
process that removes S02 from exhaust gases, which avoids some of the
plugging and scaling associated with direct lime/limestone scrubbing. The
process normally involves absorption of S02 in a sodium solution in the
absorber followed by regeneration of the absorbent in a separate system
4.2-67
-------
0>
oo
00
oo
D)r-
•<- IB >>
in > o
O) O C
T3 E 01
N S- U
O •.-
00 t-
C\J
CTi
tn
00
00
ID
00
LlJ
z
o
oo
cn
u
O)
tB T-
oo
en
en
in
CM
CO
CO «tf- r-
CO i— i—
00
o
u
ce
LU
oo
j_
oo
o
LU
cu
u
(B
Q.
to
o
s- c
O 3 O
00
ID
<*• LO
ID r-
CO 00
r>- oo
CO
o
E +•>
Ol •!-
+-> u :
w to :
>> Q.
oo to
u
oo
CM
oo
CM
CM
Ol Ol
•)-> Q.
00
ID
01
cu
OJ
o;
CM
LU
_1
QQ
OO
IB
•P
OO
Ol
Ol
C
D)
O
IB
CM
O)
Ol
-C
oo
CM
cu
oo
Ol
Ol
o
4.2-68
-------
2NaOH + S02 -> Na2S03 + H20
Na2 C03 + S02 •* Na2S03 + C02 t
S02 + H20 5 2NaHS03
through reaction with a calcium-based alkaline slurry. Regenerated absorb-
ent is recycled to the absorption loop; the calcium sulfites and sulfates
are precipitated and discarded.
Process chemistry— The alkaline solution used to absorb S02 may be a
solution of either potassium, sodium, or ammonia compounds. In the United
States, the scrubbing liquor is generally a sodium salt solution.
The basic absorbent is formed by addition of soda ash (sodium carbon-
ate) or caustic (sodium hydroxide) to water. Several sodium species are
available for reaction with the S02, including sodium carbonate, sodium
hydroxide, and sodium sulfite. The primary products of reaction include
sodium sulfite and sodium bisulfite. Following are the main absorption
reactions:118-119
(1)
(2)
(3) Na2S03
(4) NaOH + S02 -> NaHS03
Also some of the S03 present in the exhaust stream may react with
sodium hydroxide to form sodium sulfate as follows:
(5) 2NaOH + S03 -»• Na2S04 + H20
In the absorber, and through the rest of the system to a lesser degree, some
sulfite is oxidized as follows:
(6) 2Na2S03 + 02 •* 2Na2S04
The sulfate species is inactive and is unavailable for further. S02 removal.
The extent of oxidation is a function of the oxygen and S02 content of the
flue gas, the temperature of the gas in the absorption vessel, and the
design of the absorber. As an example, the typical excess air level in
high-sulfur, coal-fired utility boilers leads to oxidation levels of 10 to
15 percent of the S02 removed.
After absorption of S02, spent absorbing liquor is bled to the regener-
ation system and reacted with either lime or limestone. The reactions vary,
depending upon which reactant is used. Regeneration with slaked lime
4.2-69
-------
(calcium hydroxide) takes place over several stages. The calcium hydroxide
reacts with sodium bisulfite, sodium sulfite, and sodium sulfate in the
following series of reactions:120-122
(7) 2NaHS03 -f Ca(OH)2 -> Na2S03 + CaS03- 1/2H20 4- + 3/2H20
(8) Na2S03 + Ca(OH)2 + 1/2H20 -* 2NaOH + CaS03- 1/2H20 4-
(9) Na2S04 + Ca(OH)2 + 2H20 •* 2NaOH + CaS04- 2H20 4-
The precipitated calcium species (sulfite and sulfate) are separated from
regenerated liquor.
Regeneration with limestone (calcium carbonate) involves the following
main reaction:123
(10) CaC03 + 2NaHS03 + 1/2H20
Na2S03 + CaS03- 1/2H20 4-
+ C02 t + H20
The bisulfite is the main reactive species, because the sulfite ion is not
sufficiently acidic to react with limestone.124
At present limestone regeneration is not in commercial use, although
several laboratory studies have been conducted. Therefore, the emphasis in
this section is on lime regeneration.
After removal of the calcium sulfite and sulfate species from the
regenerated scrubbing liquor, two additional steps may be required. The
first step, needed in all cases, involves the addition of makeup sodium to
replace the small amount of sodium lost in the waste solids. The sodium is
usually added as sodium carbonate or sodium hydroxide. The second step
involves reduction of the calcium ion concentration in the scrubbing liquor
to prevent scaling in the scrubber. This step is not always needed if the
calcium concentration in the return liquor is low. If excess calcium is
present, however, the solution must be softened by addition of sodium
carbonate as the makeup sodium. Calcium is removed by the following reac-
tion: 125 ,
++
+
(11) Na2C03 + Ca •* CaC03 4- + 2 Na
Sodium carbonate must sometimes be added in amounts greater than the
makeup requirement to soften the solution sufficiently.
4.2-70
-------
A residence time of about 1 to 4 seconds in an absorber is necessary
for S02 absorption. To achieve maximum S02 removal with an optimum amount
of energy input, various absorber types are used by the industry, the most
common of which are the tray tower or packed tower types. Typically the L/G
ratio is 1.3 to 2.0 liters/m3 (10 to 15 gal/1000 acf), the pressure drop is
1.5 to 3.0 kPa (6 to 12 in. H20), and the spent absorbent pH is 6.O.126
The sludge generated by the double alkali process is essentially the
same as that generated by lime scrubbing (Section 4.2.3.1) and is handled in
the same way. The presence of soluble sodium compounds in sludge from the
double alkali system may necessitate more care in disposal to prevent leach-
ing or runoff of sodium species.
Operational status and current development—Several vendors offer
variations of the double alkali system. In the United States, all double
alkali systems use a sodium salt as the absorbent and lime as the regen-
erant. The system offers high S02 removal capabilities and minimizes scal-
ing and plugging.
Several successful bench-scale, pilot plant and prototype double alkali
FGD systems have been tested on boiler flue gas applications in the United
States. The success of these programs resulted in commitments by three
separate electric utility companies to install full-scale double alkali FGD
systems on coal-fired boilers. Two of the new utility double alkali systems
began operation in early 1979, and the third came onstream in late 1979. In
addition, several systems are working on coal-fired industrial boilers, and
one pilot plant system and one prototype have been tested on utility units.
At the end of 1979, double alkali FGD systems comprised about 6 percent
of the utility FGD capacity in operation. Applications on utility and
industrial boilers are as follows:127
Company:
Plant:
Location:
Stream treated:
System size:
S02 inlet:
Startup date:
Southern Indiana Gas and Electric
A.B. Brown 1 (New Plant)
West Franklin, Indiana
Off-gas from coal-fired boilers
265-MW
2810 ppm (4.5 percent sulfur coal)
March 1979
4.2-71
-------
Company:
Plant:
Location:
Stream treated:
System size:
S02 inlet:
Startup date:
Company:
Plant:
Location:
Stream treated:
System size:
S02 inlet:
Startup date:
Company:
Plant:
Location:
Stream treated:
System size:
S02 inlet:
Startup date:
Company:
Plant:
Location:
Stream treated:
System size:
S02 inlet:
Startup date:
Company:
Plant:
Location:
Stream treated:
System size:
Fuel properties:
Startup date:
Company:
Plant:
Location:
Stream treated:
System size:
S02 inlet:
Startup date:
Louisville Gas and Electric
Cane Run 6 (Retrofit system)
Louisville, Kentucky
Off-gas from coal -fired boilers
288-MW (Net)
3471 ppm (4.8 percent sulfur coal)
April 1979
Central Illinois Public Service
Newton 1 (New plant)
Off-gas from coal -fired boilers
575-MW (Net)
— (4 percent sulfur coal)
November 1979
Caterpillar Tractor Company
Joliet Plant
Joliet, Illinois
Off-gas from coal -fired boilers
48.8 mVs (103,500 acfm) (18-MW)
2300 ppm (4 percent sulfur coal)
September 1974
Firestone Tire and Rubber Company
Pottstown Plant
Pottstown, Pennsylvania
Off-gas from oil-fired boiler
6.6 mVs (14,000 acfm)
1,000 ppm
January 1975
Caterpillar Tractor Company
Mossville Plant
Mossville, Illinois
Off-gas from 4 coal-fired boilers
113 ms/s (240,000 acfm) (57-MW)
Coal, 3.2 percent sulfur average
October 1975
General Motors, Inc.
Chevrolet Parma
Parma, Ohio
Off-gas from coal -fired boilers
124 m3/s (262,000 acfm) (32 MW)
Ma?ch°1974°
C0a1)
One prototype and one pilot plant double alkali system have operated
on utility coal-fired boilers:128
4.2-72
-------
Utility:
Unit:
Location:
Unit size:
Fuel properties:
Startup date:
Note:
Utility:
Unit:
Location:
Unit size:
Fuel properties:
Startup date:
Note: !
Utah Power and Light Company
Gadsby Station, Unit No.. 3
Gadsby, Utah
1.2 mVs (2500 acfm) (0.6 MW)
Coal, 0.4 percent sulfur average
1971
Terminated 1973
Gulf Power Company
Scholz,. Unit No. 1
Chattahoochee, Florida
35 m3/s (75,000 acfm) (20-MW)
Coal, 3 to 5 percent sulfur
February 1975 :
Terminated July 1976 ' • i
As a result of the success of pilot and prototype systems, three full-scale
double alkali systems are in operation on coal-fired utility boilers.
Utility operating results are sketchy because initial operations are still
in progress.
The GM Parma system has performed well with regard to S02 removal.
Results of a 1-week test in 1974 indicate S02 removal efficiencies in the
94- to 99-percent range, with relatively low inlet S02 levels (600 to 1200
ppm) and high excess air rates. A test was conducted by A.D. Little, Inc.,
and General Motors (GM) from August 19, 1974 to May 14, 1976. It consisted
of three 1-month intensive test periods and 18 months of lower-level tests.
Removal of S02 reflects the variations in operating modes employed by GM
during the period, but removal efficiencies were at 90 percent for the
viable operating modes. Operation during April and May 1976 was excellent
and A.D. Little, Inc., recommended continued operation in the mode used
during this period.129
The operability (hours the FGD system was operated per boiler operating
hours in a period expressed as a percentage) of the Parma system for the
1-year period from May 1976 through April 1977 was about 70 percent. The
system's best period of operation was May through August 1976, when opera-
bility averaged 94 percent.129 The GM Parma plant has several unique
characteristics that affect operability. Each boiler is equipped with its
own separate scrubbing module with no provisions for crossflow between
modules. The GM plant is a developmental system, and as such is subject to
4.2-73
-------
modifications. Many of the low operability periods were due to mechanical
outages or outages for modifications to accommodate and test new modes of
operation. Several different operating modes have been investigated, and
significant improvements have been obtained in both process and mechanical
performance. It is believed that in the latest operational mode the system
is capable of long-term reliability.
The Joliet system has achieved excellent S02 removal efficiencies of
between 85 and 95 percent under various operating conditions. Sulfur
dioxide inlet concentrations are high, about 2300 ppm. The system was
designed to attain an emission level of 860 ng/J (1.9 Ib S02/106 Btu) (75
percent S02 removal), but has consistently performed much better than
designed.13°
The operability of the FGD system has been improving steadily. Process
availability for the period October 24, 1975, through June 1976 has been 100
percent. Most problems at the Joliet plant are mechanical; the majority are
solved while still on stream or during scheduled shutdowns. As a con-
sequence, there have been few forced outages.130
The Firestone-Pottstown system has exhibited excellent S02 removal
efficiencies of 90 percent on high-sulfur oil, but no data are available for
its performance on coal. It has also achieved a very high availability 99
percent for the first 12 months of operation. Most downtime periods were
due to mechanical component failure or to maintenance, and not to unwanted
chemical changes or side reactions. No scaling problems have been experi-
enced.131
The Gadsby scrubbing system has performed well with respect to S02
removal. .Various modes of operation were tested using two types of ab-
sorbers. With the polysphere absorber, S02 removals of 90 percent were
achieved, giving outlet concentrations of 15 to 40 ppm S02. With the
venturi absorber, efficiencies ranged from 80 to 85 percent S02 removal.^
With the exception of the first 3-month operating period, during which
some gypsum scaling problems were encountered, dilute mode operations were
conducted for almost 2 years without any major problems. No operating
problems causing shutdown were experienced between October 1972 and August
4.2-74
-------
1974. For convenience, the system was shut down on weekends, but no drain
age of solution or cleaning of equipment took place during these shut-
downs.131
The Gulf Power Company, Scholz plant prototype system started up
February 3, 1975, and operated continuously through July 18, 1975, when it
was shut down for repairs and modifications.. The second period of operation
was from September 16, 1975, through January 2, 1976.
The system exhibited excellent S02 removal capabilities of 90 percent
and greater. Using the combined venturi/tray tower absorber configuration
at a venturi liquor pH above 5.2, outlet S02 concentrations below 50 ppm
were achieved, which corresponds to greater than 95-percent removal.
Raising the pH of the venturi liquor above 6.0 resulted in S02 removal
efficiencies greater than 98 percent.131,132
The Scholz plant was designed to demonstrate the viability of the
double alkali process technology for application on utility coal-fired
boilers. As such, this prototype plant had less spare equipment than would
be normal in full-scale applications. The operability of the system has
been steadily improved; during 4 months of operation, it. was 94 percent.132
The operability of double alkali FGD systems on coal-fired industrial
boilers and prototype utility installations has been improved. Most oper-
ability problems were due to design-related equipment shortcomings in these
prototype installations. It should be pointed out that most installations
did not have spare equipment; however, this is included in full-scale util-
ity systems.132
The vendors of double alkali systems have developed confidence in their
reliability: as evidenced by guarantees of 90-percent availability for the
first year of operation and 100 percent for the life of the plant (based on
a boiler operating rate of 70 percent for some of the new, full-scale util-
ity applications). The systems are all guaranteed to achieve 85 to 95
percent S02 removal efficiency on high-sulfur coal applications. No new
low-sulfur coal applications are planned, but similar guarantees would be
expected for such systems.133
Corrosion, erosion, and scaling problems have not been important
factors at double alkali FGD installations. Full-scale versions of these
systems are not expected to experience these problems either. The double
4.2-75
-------
a double alkali FGD
may be some potential
alkali system has demonstrated the ability to perform well under fluctuating
S02 inlet concentrations. At the Scholz plant, the design inlet S02 con-
centration was 1800 ppm. At inlet concentrations varying from 800 to 1700
ppm, removal efficiencies were above 90 percent.133
Energy and environmental impacts— Opprat.inn Of
system requires relatively little energy. There
adverse environmental effects, but none significantly more consequential
than from other types of FGD systems.
The electrical requirement is variable, usually in a range of 1 to 2
percent of the energy output of the boiler. "4 Tne reheating requirement is
usually about 2 percent of the energy output of the boiler for about 50°F of
reheat. 134 Tne concepts of energy and capacity penalties are discussed more
fully in Section 4.2.3.1.
Beneficial environmental impacts are the prime reason for operating the
scrubber. Sulfur dioxide removal efficiencies of 95 percent are achievable
with current technologies in both high-sulfur and low-sulfur coal applica-
tions. 134,135 A potential adverse effect .s em1ssion Qf sulfurous ^
sulfuric acid mist if the flue gas stream is not processed to remove the
entrained liquor. The principal adverse impacts, however, are due mainly to
the generation of sludge.
At a 1000-MW plant burning 3.5 percent sulfur coal with 14 percent ash,
90 percent S02 removal would generate about 210,000 Mg (232,000 tons) of dry
sludge per year.51 This plant would consume about 2,152,000 Mg (2,373,000
tons)^ of coal and generate about 201,000 Mg (222,000 tons) of ash 'per
year.51 The potential environmental effects of 'sludge are discussed in
detail in Section 4.2.3.1. Because the double alkali sludge contains
soluble sodium compounds, care must be taken to reduce leachage or runoff.
This can be achieved most often by constructing a plastic-' or clay-lined
sludge impoundment and managing the placement of waste and/or by chemically
treating to fix the waste.
System costs-Little cost information is available on the early double
alkali system applications, but some is reported for the large coal-fired
utility applications. Table 4.2-17 presents cost data on the three utility
4.2-76
-------
double alkali systems, adjusted to July 1, 1979 dollars.136 Two of the
systems were recently completed, and the third should be on line in, late
1979. .
TABLE 4.2-17. CAPITAL AND ANNUALIZED COSTS OF
UTILITY DOUBLE ALKALI FGD SYSTEMS136
Station, unit
Cane Run 6
A.B. Brown 1
Capacity,
MW
277
250
Adjusted costs (mid-1979)
Capital costs,
$/kW
72.7
61.4
Annuali zed costs,
mills/kWh
4.8
3,5
Generic cost data on double alkali FGD systems, have been developed in
support of the NSPS for utility boilers. These data are presented in
graphical form in Figures 4.2-19 and 4.2-20.137 The values are higher than
those presented in Table 4.2-17, but they represent costs for systems with
spare scrubbing trains and a high level of installed spare capacity.
4.2.3.4 Nonregenerable Sodium-Based Flue Gas Desulfurization—
Nonregenerable sodium-based FGD systems utilize a clear liquor absorb-
ent (soluble salts) that minimizes the plugging, scaling, and erosion that
are common in some calcium-based scrubbing systems. Solutions of sodium
hydroxide or sodium carbonate are currently being used to scrub SOX from
flue gases. :
Nonregenerable sodium-based scrubbing systems are well developed and
widely applied on industrial boilers. Whereas about 90 percent of utility
FGD systems treat flue gases with lime or limestone absorbent, sodium-based
scrubbing technology accounts for about the same percentage (93 operating
systems) in the industrial sector.138
Nonregenerable sodium-based scrubbing is based on the following
reactions (assuming the presence of an absorber recirculation loop):
(1) Na2S03 + S02 + H20 -> 2NaHS03
(2a) Na2C03 + 2NaHS03 -> 2Na2S03 + C02 + H20
(2b) NaOH + NaHS03 -» Na2S03 + H20
(3) 2Na2S03 + 02 -»• 2Na2S04
4.2-77
-------
400
300
o
o
o.
5
200
TOO
'—•516 ng/J (1.2 Ib/mm Btu)
90% S02 REMOVAL
200
_L
400 600
PLANT SIZE, MW
800
1000
Figure 4.2-19. Capital cost of a double alkali FGD system
on a boiler firing 3.5 percent sulfur coalJ37
4.2-78
-------
I 1
— 516 ng/J (1.2 Ib/mm Btu)
90% S02 REMOVAL
200
400 600
PLANT SIZE, MW
800
1000
Figure 4.2-20. Annualized costs of a double alkali FGD system
on a boiler firing 3.5 percent sulfur coal.'37
4.2-79
-------
Reaction 2a occurs in a sodium carbonate (Na2C03) scrubbing system, and
Reaction 2b is the primary reaction in sodium hydroxide (NaOH) scrubbing
systems. Reaction 3 is common to both types of systems. Reaction 1
accounts for the main S02 removal in both systems.
If fly ash is not collected concurrently with the S02 absorption, a
nonregenerable sodium scrubbing system produces only a liquid waste stream.
Typically, a bleed stream from a recirculation loop is discharged at a rate
equivalent in sulfur content to that at which S02 is being absorbed.139
Total dissolved solids in the waste stream are approximately 5 percent.140
The stream may or may not be processed to remove the sodium bisulfite
(product of Reaction 1) or to oxidize the sulfite to sulfate to reduce the
chemical oxygen demand of the waste stream. Small industrial units may be
able to discharge the waste stream directly into treatment equipment. If
treatment is necessary, vapor compression distillation, multistage flash
evaporation, or reverse osmosis may be used to reduce the amount of total
dissolved solids (TDS);140 however, this is not currently done to great
extent.
Figure 4.2-21 is a basic process flow diagram of a typical nonregener-
able sodium FGD system. The chemistry of sodium scrubbing dictates the
basic components of the system, but many variations occur.
Sodium carbonate, a solid, is often stored in onsite silos, whereas
sodium hydroxide, usually received in liquid form, is stored in onsite
tanks. Both absorbents are highly soluble and are used in aqueous solu-
tions.
Some coal-fired units utilize electrostatic precipitators or fabric
filters to remove fly ash from the flue gases before they enter the scrub-
bing system. In such cases, the S02 absorber can be a tray-type tower or
spray tower, which provides good scrubbing efficiency at low pressure drops.
Oil-fired units present a' similar situation. For simultaneous particulate
matter removal and S02 absorption, venturi scrubbers have been used. Fol-
lowing cleanup, the gas passes through mist eliminators and often is
reheated before being exhausted to the atmosphere.
Because the sodium-based liquor is highly reactive, it is possible to
use liquid-to-gas ratios (L/G) as low as 0.668 liter per actual cubic meter
4.2-80
-------
EXHAUST
NaC03 FAN
OR
NaCH
STORAGE
\/
WATER
LIQUOR
STORAGE
FLUE GAS
'RECIRCULATION
TANK
PUMP
BLEED
OFF
PUMP
PUMP
Figure 4.2-21. Basic nonregenerable sodium FGD system.
4.2-81
-------
of gas (5 gallons per 1000 actual cubic feet). This high reactivity also
permits rapid response of the system to loading changes. The liquor feed
rate is usually controlled by adjusting the PH at the absorber inlet or
outlet. Spent sodium sulfite solution is usually removed by bleeding the
recirculation loop.141
The only nonregenerable sodium-based FGD systems in operation at util-
ity plants are at the Reid Gardner Power Station of Nevada Power Company and
at Jim Bridger Unit 4 of Pacific Power & Light. A simplified process flow
diagram of the FGD systems on Units 1, 2, and 3 is shown in Figure
4.2-22.142 This figure shows several differences from,, the basic system
discussed earlier. The source of sodium at the Reid Gardner Station is
trona, a low-grade ore containing 60 percent sodium carbonate, 20 percent
sodium chloride, 10 percent insolubles, and 10 percent sulfates and inert
dissolved solids. From storage in onsite silos, the trona is conveyed to a
slurry tank, where the absorbent is dissolved. The solution is then pumped
to a clarifier, in which the insoluble impurities are allowed to settle.
The clarified sodium carbonate solution is injected into the venturi
recirculation loop.
Hot flue gas from the boiler first passes through mechanical collectors
for primary particulate removal (75 percent). After passing through a fan,
the flue gas is ducted into two streams, which enter twin throat venturi
scrubbers, where the gas is quenched. The scrubbed gas then enters the
droplet separator tower, where the finer droplets coalesce on the wall, and
the gas rises and bubbles through a single sieve tray flooded with clear
water from the ash pond. More S02 is absorbed on the sieve tray, after
which the gas passes through a horizontal radial-vane-type mist eliminator.
The gas is reheated before exhaust, both to enhance buoyancy and to prevent
condensation in the stack.
Slowdown from the venturi recycle tank may be mixed with the alkaline
clarifier underflow stream in the postneutralization tank. Spent liquor is
sent to the ash settling pond, and pond overflow is pumped to an evaporation
pond.143
Industrial applications of nonregenerable sodium-based scrubbing are
much more widespread than in the utility industry. In fact, some 90 percent
4.2-82
-------
CM
C
o
(O
4->
I/)
s-
cu
-o
S-
(O
CD
T3
OJ
O
O
O)
o
Q-
ea
-o
ro
>
CD
O)
to
Q
M-
O
to
CD
O
O
D_
CV1
CNJ
I
CM
OJ
s_
4.2-83
-------
of the industrial FGD applications use nonregenerable sodium-based scrub-
bing. At present, 93 nonregenerable sodium-based FGD systems are treating
flue gases from 158 industrial boilers. Capacities range from 18,900 to
386,350 NmVh (12,000 to 245,000 scfm, 6 MW to 123 MW), and systems are
being offered by at least 14 different vendors. Sodium hydroxide is used as
the reagent in 79 of the systems; the remainder use sodium carbonate.144
Table 4.2-18 lists the industrial applications of nonregenerable sodium-
based FGD technology.145
These applications in the industrial sector have been highly success-
ful. The users regularly report excellent operating system reliabilities.
The major advantage;of the sodium-based systems is that S02 absorption by
soluble sodium salts eliminates the scaling and plugging that occur in many
calcium-based FGD systems.1^ In addition to the lower maintenance costs
that result from scale-free operation, other attractive features of non-
regenerable sodium-based systems are relatively low power requirements and
high S02 collection efficiency.147
Although operating problems are relatively less frequent and less
severe in the sodium-based FGD systems than in calcium-based systems, oper-
ation is not flawless. Inadequate control of FGD system pH is of concern,
because effective pH control is essential to maximizing absorbent usage.'
Other mechanical problems, not peculiar to the nonregenerable sodium scrub-
bing systems, include failure of dampers and liners, pump problems, and
inefficiency of mist eliminators. At the coal-fired installations, oper-
ators face the problems usually associated with fly ash (abrasion, erosion,
and plugging).
Operational status and current developments—The three Nevada Power
nonregenerable, sodium-based FGD systems have been operated well over a
period of 5 years. Although the expected mechanical problems have occurred,
the availability and operability of the FGD systems on the three units have
exceeded 90 percent for extended periods.148,149
The 18 industrial locations, which use the 93 FGD systems, have similar
operating histories: problems have occurred, but the systems operate well
the majority of the time. As a case in point, the sodium-based system at
Alyeska Pipeline Service Company operated 1 year (second quarter 1978
through first quarter 1979) at 100 percent availability.150 Chevron U.S.A.,
4.2-84
-------
O--t-»
I- >)
* o
£ i
I 2
CMIO
O >
l/l O
C5
O
O
c >>
O) CT "
.11 =
8
•o
O
O
co
UJ
CO
ff.
CXL
CJ3
UJ
CO
CO
UJ
CO
•=c
I — I
o;
£4->
*J
E E 5
4J (J
Vl O
3 W> ,
.O Ul
E <
3
»— r- (O
I1
E
S
s
•a
>>
§
8 8
S
^o
O) *O
r— r— t
^ ^ .^ .^ Q
I— I— re 1—
O)
0 d
trt t/1
I- t-
S S
£ i
- -
S 2
r— r— C
•r- -r- O
.,_ 4_» tO
>» •— T.
4.2-85
-------
Inc., reports operability in excess of 95 percent since startup in July
1978.151
Cost—Although the reported costs of nonregenerable sodium FGD systems
vary considerably, some general conclusions can be drawn. Dickerman
reports152 that the capital costs of nonregenerable sodium FGD systems are
always lower than the costs of comparably sized double alkali, limestone, or
Wellman-Lord systems. Sodium-based systems are less complex and require
less equipment, largely because of the high solubility and reactivity of the
sodium absorbent. High solubility and the consequent absence of slurry
solids allow use of less exotic construction materials, and high reactivity
allows use of simpler or smaller absorber vessels.
These characteristics of the sodium-based system, however, do not lead
to comparable reductions in annual costs relative to other systems. The
sodium absorbent is more expensive than lime and limestone and wastewater
treatment may be required. Because of these cost items, the annual costs of
a nonregenerable sodium-based FGD system rise more sharply with increasing
unit size than do the annual costs of other FGD systems. Figures 4.2-23153
and 4.2-24154 show the capital and annual costs versus system size for
several industrial-scale FGD processes capable of removing 90 percent of the
S02 from flue gases generated by coal with 3.5 percent sulfur content.
Table 4.2-1915*-is7 presents costs of a nonregenerable sodium-based FGD
system as reported by several utility and industrial operators. All costs
are adjusted to mid-1979 dollars.
The PEDCo industrial cost data are derived from direct contact with the
industrial FGD system users. Little or no redundancy is employed. Also the
purged absorbent is often disposed of in municipal systems, deep wells, etc.
Redundancy and an involved waste treatment ponding system can double the
cost of the FGD system.
The utility cost numbers reflect the PEDCo utility FGD cost program
(with redundancy and waste treatment costs built in) near the low end of its
relevant range, based on the data used to prepare the program. The FMC
costs are the estimates of a system vendor, which include little redundancy
and waste treatment costs.
4.2-86 .
-------
J2 4000
'o
-o
f)
•o
c
as
o 3000
-M
r*
f—
o
_i
<
S 2000
1000
WELLMAN-
LORD
LIMESTONE
DOUBLE
ALKALI
SODIUM
NONREGENERABLE
3.5% SULFUR IN COAL
90% SULFUR REMOVAL
14.6 29.2 43.8 58.6 ,
(50) (100) (150) (200)
.FGD SIZE, MW(106 Btu/hr)
Figure 4.2-23. FGD capital costs versus unit size.153
4.2-87
-------
1400
s-
(O
o
TD
O
to
T3
C
-------
to
CJ3TO
LJ_ OO
•z.
O O
LJJ t— I
00 (—
-
DQ I
LU Z
CD
Q
00 Z
H- H- i
OO
o z
C_> I— I
cn
i—i
CM
CQ
«^
•p
o
re
•r—
Q.
ro
C_5
o>
cn
ro
o;
,— ^
3 /— v
-P 3
O CO -P
CDCO i—i nn
\ O i — CD
O i— -be- •
4^- . f—
CD
CO LO 0 CO
o co <=e- •
. CO O
o o -be-
CD
•be-
2 2
^^^ — ^
CO 1 —
CO CO
CO CM
co r^
3J 2
i— cn
CO CO
r*** LO
co cn
•fee- *be-
i i
CO 1 —
LO CO
CM l£>
i — LT)
•be- -be-
to tc
•r- m m
S- in TH >,
-P .H C -P
to o re *i—
3 0 -r- i—
•a o -a -i-
c: LU re -P
I— l Q- Q£ =D
J=. -C
3: 3:
J^. -^
^""^ *^s^
to t/)
r— r^
r— r—
•p— T—
E E
cn in
r— CD
» •
•3- LO
JZ -C
3: 3:
"*^^. ^*^
to (/>
r— r—
p— *• r—
•i — *r—
E E
r— «xt"
i — r—
• •
to r**-
i i
00 OO
P"** r—
. •
CO CO
2 2
^^, ^
1 — CM
cn r—
tri r-^
cn CM
•be- -be-
^
3; \
-^ *^"
^^^ r^^
CD
CM cn
i— CM
•be- "be-
i i
CM cn
CD <*•
t— *t
r~~ CM
^ <»
o> r*-
o w
C_J TH
*""•> C— 3
LU S
^\ f 1
-p
c
a>
ro
a>
s_
ro
ro
-a
c:
ro
u
c:
ro
•a
at
cu
Q)
Cn
a>
T3
O
Q.
3
cn
c:
a>
Q.
a>
T3
ai
-a
ro
to
O
U
•a
cu
-»->
s-
o
Q.
01
S-
cu
ro
4.2-89
-------
Energy and environmental impact—The waste streams from the sodium-
based nonregenerable systems consist of a solution of sodium sulfate and
sulfite/ sulfate only if oxidation is employed. Current users of nonre-
generable sodium alkali FGD systems dispose of the purge stream either by
discharging it to onsite or municipal liquid waste treatment facilities or,
in the West, by pumping it to large evaporative ponds, or injecting it in
deep wells. As long as these means of disposal are feasible, the innate
simplicity and high S02 removal capability of the system coupled with its
low energy requirement make this system attractive at many sites.
Table 4.2-20158 presents estimates of the liquid wastes from various
size boilers firing different coals at the same removal efficiencies.
4.2.3.5 Ammonia-Based Process
In the ammonia-based wet S02 absorption process, flue gas from a sta-
tionary source is pretreated to remove most of the particulate matter (if
necessary) by electrostatic precipitation, fabric filtration, or other means
and is then water-quenched to its adiabatic saturation temperature. The
conditioned, humid gas is brought into contact with an aqueous ammoniacal
solution, which rapidly absorbs the S02. There are two basic ammonia-based
processes, both with salable byproducts:, a nonregenerable process with
ammonium sulfate as byproduct and a regenerate process with elemental
sulfur or sulfuric acid as byproduct. A nonregenerable process without
oxidation would yield ammonium sulfite as byproduct.
Process chemistry—In a typical nonregenerable process with ammonium
hydroxide as the feed liquor and ammonium sulfate as the byproduct, the
following principal reactions occur:
S02 absorption:
(1) 2NH4OH + S02 -> (NH4)2S03 + H20
(2) (NH4)2S03 + S02 + H20 -> 2NH4HS03
These reactions take place in the absorber during
S02 absorption by ammoniacal solutions.
4.2-90
-------
co
in
co
>-
CO
Q
O
CO
LU
_J
CD
•
a:
z
o
o
CO
o
r~
S3
o -
S- •—
•P QJ
c >
O Q)
C_J r-
QJ
D.
^_>
-a
c
ro
01
N
•r—
in
s-
QJ
•r—
O
CQ
/*-\ t*~\ i — s
r^ CD *i-
CM CM CM
CM 00 i—
* * •
O CD CD
•3- m CM •=*
CO. CM i— CO
1 1 1 V^>
r~- co *tf- CM
O 1^- CM i —
o in in CD
CD co r^- m
/— -N
JC
^j
.!_}
CQ
to s-
o 00
^- ,_
r— r—
S^,^ Sta*-*
co r***-
> t
m «3-
m <^-
•=*• CD oo
CD m oo
CD m co
\^ \^^/ \^s
r~» CD oo
CM i— CD
CD r- •=}-
CM CM i—
o m CD
CD i — m
2
V^
33
-p
CQ
to
0 S-
i — Ol
NX
X 0
^_)
o in
m
r— &_
•^^ O)
T3
S ro
s: 01
s_
«*• Q.
«* CO
^^ /*~N x^*s
in I^Q r^-
OT r-^ tn
f— f^ r—
S^X N^^,/ S,^. /
00 tD ^1-
* * •
CO CO CD
r^ co in
co *d~ r^- CD
CM CO «3- CM
CD oo r*~ m
N*-^ V»^ S»^ v**-'
CM CD P-- CM
oo in CM o
^- r— CO O
CO CO CM CM
o m in co
CD oo r>» in
, V
^
-p
CO
y?
o ,
r^
i —
X ro
O
O U
0
CM -a
s^^ QJ
N
^C *r—
£: s-
QJ
CD >
00 <~^-
in a.
•P -P -P
c c c.
ru QJ QJ
(J U U
S- J- i-
QJ QJ QJ
Q. Q. D.
CM in co
in co i—
CO Tf CO
O 0 0
CO CO O
to to to
-z.-z.-z.
•p -p -p
c c c
QJ (!) QJ
u u o
S- i. i-
QJ QJ ru
a. Q. a.
0 CO CM
r^. t— r—
CO Tf CO
O O 0
CO CO O
IN IN CJ
to ro to
z z z
T3
•f—
r—
o
in
-a
r—
O
. in
in
• r—
~&
•
O)
>
CZ
QJ
CJ
- C
O
u
CO
o
I—
•
O)
>
*£
4.2-91
-------
Neutralization:
(3) NH4HS03 + NH4OH •* (NH4)2S03 + H20
In the neutralizer, ammonium bisulfite is converted to ammo-
nium sulfite
Oxidation:
(4) 2(NH4)2S03 + 02 -»• 2(NH4)2S04
In the oxidizer, the Oxidation of ammonium sulfite yields a
solution of ammonium sulfate as byproduct of the process.
A common parameter in the ammonia-based process is S/C. ratio,159,160
in which S is defined, as moles of S02 (as ammonium sulfite and bisulfite)
per 100 moles of water, and CA is defined as moles of NH3 (as ammonium
sulfite and bisulfite) per 100 moles of water.
The chemistry of the ammonia-based system is relatively simple. In the
pH range 4.2 to 7.0 the main dissolved species are HS03~, S03=, and NH+.
The formation of sulfate depends on the amount of S03 in the flue gas and
the degree of absorber oxidation. Other materials introduced with the gas
(ash, chlorides, etc.) may accumulate if the absorber circuit is a closed
loop. These materials do not appear to affect absorber efficiency.160
The S/CA ratio is related to the pH of the scrubbing liquor in the
absorber. The CA value could vary from 10 to 20, with a typical C. range of
10 to 14.I59_i6i curves of S02 partial pressures over (NH4)2S04-NH4HS03-H20
solutions versus S at constant CA show the difficulty of obtaining high
bisulfite values (high S/CA) unless the S02 partial pressure in the entering
gas is relatively high. Therefore, high-sulfur coal could improve economics
by increasing sulfur throughput and also could allow high bisulfite-sulfite
ratios for processes that are helped by a high S/C .16°
Mass transfer in S02 absorption by -an ammoniacal solution is controlled
mainly by the gas phase resistance. The chemical reactions involved are
rapid; hydration of S02 is the slowest but it is fast enough to be non-
limiting up to a concentration of 3 to 4 percent S02 in the gas. Liquid
film resistance is also quite low, except at low pH which would be
encountered in the first stage of a multistage absorber; the liquid-phase
4.2-92
-------
resistance becomes equal to the gas-phase resistance at an S/CA of 0.92 to
0.96. The transfer rate falls rapidly with increasing temperature. Thus,
cooling below the wet bulb temperature, although expensive, would yield
better S02 absorption by the absorbing solution.160
Assuming an S/CA ratio of 0.79 at pH 5.8, the ammonia requirement is
1.29 moles NH3/mole of S02 removed and the composition of the bleed-off
solution is about 71.4 mole percent NH4HS03, 26.0 mole percent (NH4)2S03,
and 2.6 mole percent (NH4)2S04. After neutralization and oxidation the
solution typically has 24 weight percent (NH4)2S04.
The chemistry involved in the regenerate process depends entirely on
the nature of the product. Detailed reliable data are available.162-165
System description—The nonregenerable ammonia-based processes have
been widely used to control S02 emissions from various industrial process
sources such as pulp and paper plants166 and sulfuric acid plants,167,168
and to a lesser degree from industrial boiler sources.169 No commercial
application of an ammonia-based FGD system in the U.S. utility industry has
been reported. The high energy requirement and unfavorable economics170 of
a regenerable, ammonia-based process have prohibited its use in that appli-
cation.
The three major operations171,172 of a nonregenerable ammonia-based FGD
system are as follows:
0 Flue gas pretreatment—particulate removal as required, and cool-
ing and humidification in a quencher.
0 S02 absorption—removal of S02 by reaction with ammoniacal liquor
in the absorber.
0 Neutralization and oxidation—the bleed-off solution _is neutral-
ized by pure ammonia in a neutralizer and later oxidized by air.
Figure 4.2-25 presents a typical process flow diagram for such a sys-
tem. Flue gas from a stationary source, depending on its particulate con-
centration, may pass through a particulate collection device before entering
a quencher (Q-l). Flue gas with low particulate content enters the
quencher through a forced draft (FD) fan (F-l). In the quencher the flue
gas is cooled to its adiabatic saturation temperature by recirculating
4.2-93
-------
at e£
O
O)
)
03
fd
O)
-Q
(O
O)
CT;
-------
water. As a means of controlling the formation of ammonia fumes (sometimes
called "blue haze") in the exit gas (a problem in most of the earlier
ammonia absorber applications),173 the gas must be cooled to temperatures
below its adiabatic saturation temperature. This is accomplished in two
cooling stages that precede the absorption section. Heat exchangers are
supplied for each stage to cool the recirculating water.
The cooled flue gas enters the absorber (A-l), which incorporates four
stages. The first three stages are required to maximize S02 absorption and
to minimize ammonia loss and fume formation. The final stage is mainly a
water wash, which removes entrained absorber liquor. The four absorption
stages have valve trays, which are independent of each other. Absorber
efficiency is increased by placing mobile plastic spheres on the first two
absorption stages. A pad-type (wire mesh) mist eliminator follows each
stage. Chevron-type mist eliminators are installed above the cooling and
absorption stages.
Makeup water is added to the fourth-stage recycle tank (RT-4). Recycle
tanks are arranged in such a way that liquor can flow successively from the
fourth stage to the lower stages. The proper concentration at each stage
can be maintained by monitoring the flow rates of bleed-off between stages
and the product bleed-off.
The product bleed-off liquor from the first stage passes through a
neutralizer (N-l), where the liquor is neutralized to ammonium sulfite.
Ammonia is fed to the neutral izer and the second absorption stage from a
storage tank (ST-1). The neutralized liquor is then oxidized to ammonium
sulfate by compressed air in the oxidizer (0-1). The 24 weight percent
solution of ammonium sulfate is stored in a tank (ST-2).
The water used in the heat exchangers represents a substantial recovery
of heat, which may be utilized within the plant for preheating the boiler
feed water or other process uses.171
Formation of ammonia fume is minimized by reheating'of the clean flue
gas, which then leaves the system through a stack.
The ammonium sulfate generated in the process is a widely used fertil-
izer, second only to ammonium nitrate as a world source of fertilizer
nitrogen.174 Oxidation of neutralized liquor is an old and fairly well-
4.2-95
-------
established practice, considered as the basic method of converting the
entire bleed-off liquor to sulfate. The dilute solution of ammonium sulfate
can be directly sold or concentrated and crystallized. Although oxidation
1S a relatively simple method of sulfite conversion, use of the byproduct is
sometimes a problem. Industries such as caprolactam and coke manufacturing
produce large quantities of byproduct ammonium sulfate.
Absorber design and types-The ammonia-based process differs from other
systems using alkalis in that both the cation and the anion are volatile-
therefore the absorber must be designed to recover both. In processes that
reqmre a high bisulfite content in the absorber effluent liquor, an enrich-
ing stage is needed in which the driving force of the S02 partial pressure
in the incoming gas produces a high bisulfite-sulfite ratio. In practice
four or more stages are preferable for maximizing the bisulfite content'
The composition of the solution circulated in each stage is controlled. «o
Several different absorber types have been used in ammonia-based
processes: a venturi/separator combination in an industrial boiler particu-
late and S02 removal system;"* bubble cap and tray tower absorbers ^ pulp
and paper plants;"* a two-stage, packed-bed tower with Brink mist elimi-
nators in sulfuric acid plants;^ and sieve tray impinger- type absorbers in
other miscellaneous plants. "s comparison of the performance data is
diff1Cult because of widely varying conditions. In an effort to obtain
comparative data, the EPA and the Tennessee Valley Authority (TVA) installed
a pilot plant at TVA's Colbert station in Tuscumbia, Alabama, "o Various
absorbers have been tested, including tray, plate, and mobile-bed types.
Each absorber has given good S02 removal, but some offer advantageous
operational features. A valve tray arrangement, widely used in pulp and
paper mills for S02 control, is a variation of the bubble-cap arrange-
ment. "V76 The valve trays are, in effect, perforated trays with variable
openings for gas flow. The perforations are covered with movable caps
which rise as the rate of gas flow increases. At low gas flows, with
correspondingly small openings, the tendency to weep is reduced; at high gas
flows the pressure drop remains low. Many varieties of valve tray absorbers
are available.
4.2-96
-------
Nonreqenerable processes177--For some years now, the ammonia process
has been used to remove S02 from flue gases emanating at pulp mills. In
particular, one pulp mill with scrubbers ranging from 5.2 to 7.6 m (17 to 25
ft) in diameter and from 18 to 27 m (60 to 90 ft) high successfully removed
both particulate and S02 from recovery boiler flue gases. The scrubbers
serving these installations are handling gases equivalent to coal-fired
boilers ranging from 80- to 100-MW capacity.177
Regenerate processes—The research effort on ammonia as an absorbent
for S02 has included studies of a wide variety of regenerable processes.178
One way of regenerating the absorber solution is by thermal stripping. In
this process the bleed-off solution is heated to evolve S02. The process
• apparently is not considered economical.178
Catalytic, Inc., and Institut Francais du Petrole (IFP) have developed
a regenerable process with elemental sulfur as the end product. The process
uses a staged absorber, a sulfate reducer, and a liquid Claus reactor. A
low-Btu gas produced from coal and rich in hydrogen and carbon monoxide is
utilized to reduce ammonium sulfate and sulfur trioxide, to produce hydrogen
sulfide for the Claus reaction, and to incinerate the tail gas. Major
problems associated with this process are the possible formation of a 'blue
plume1 and lack of commercial scale equipment for the sulfate reducer and
H2S generator.164
Another regenerable process is the ammonia-ammonium bisulfate (ABS)
process developed by TVA, which removes S02 from the flue gas by absorption
in a buffered aqueous solution of ammonium sulfite and bisulfite. The
solution is regenerated by acidulation to release the S02, and produce
ammonium sulfate solution. In an integrated ABS process, acidulation occurs
when ammonium bisulfate is generated by thermal decomposition of ammonium
sulfate. The liquor from the acidulator overflows into a stripper where the
remaining S02 is stripped with air. The stripped liquor is sent to an
evaporator-crystallizer to produce a slurry of ammonium sulfate crystals.
The crystals are fed to the ammonium sulfate decomposer.
Ammonia fumes in the exit gas—One of the most difficult problems in
the ammonia-based process is fume formation. The most complete evaluations
of fume formation have been done concurrently by Air Products, Inc./
Catalytic, Inc., and by TVA/EPA. The approaches are different, one being
4.2-97
-------
theoretical and the other based primarily on pilot plant operation. Regard-
less of the basis, these studies show that control of the partial pressures
of gaseous ammonia, S02, and water vapor in each absorption stage 'is
cntical to prevention of ammonia fumes.-3 The average particle size in
the fume has been determined to be about 0.25 Mm, with 10? particles per
cubic centimeter in the size range of 0.005 to 0.5 mm. Chemical and
petrographic analyses of the plume collected in an impaction sampler
indite that the major fraction of the ammonia-sulfur salt is ammonium sul-
fate The particulates probably were formed in the vapor phase as ammonium
suime and then oxidized to the sulfate form in the sampler. A portion of
the particulate was analyzed as ammonium chloride. "9
The TVA pilot tests showed that a water wash ahead of the absorber
materially reduced the chloride content of the entering gas. This also
would reduce plume formation.179
Another means of fume control is to allow the fume to form and simply
remove it in a candle-type mist eliminator. Some sulfuric acid plants use
such m,st eliminators in the absorbers to control fumes in the exit gas "7
Certain candle-type mist eliminators can eliminate liquid mist of sizes as
low as 0.1 Mm with over 99 percent efficiency.1*0
Costs-Various estimates of capital and operating costs of ammonia-
based processes differ for regenerable and nonregenerable processes
depending on the specific installation. i«-i.a Tne operating cost of ap ^
facmty is significantly affected by the S02 content in the flue gas
treated and the prices of both ammonia and the byproducts.
Table 4.2-21 summarizes the estimates of capital and operating costs
for an ammonia-based process that controls S02 emissions from a coal-fired
boiler and yields byproduct ammonium sulfate. The total installed cost
includes all material, labor, engineering, procurement services, supervision
for design and erection of the facility, contingencies, and contractor's
fees. These costs, however, do not include land or allowance for site
preparation.
Detailed cost estimates were made for a regenerable, ammonia-based
process with byproduct sulfuric acid and a nonregenerable ammonia-based
process w,th byproduct ammonium sulfate.»* Table 4.2-22 summarizes the
4.2-98
-------
TABLE 4 2-21. SUMMARY OF ESTIMATED CAPITAL AND OPERATING COSTS
FOR A NONREGENERABLE, AMMONIA-BASED PROCESS WITH
AMMONIUM SULFATE PRODUCTION
(minions of mid-1979 dollars)
Capital Cost
Fixed capital investment
Working capital
Total capital investment
Operating Cost
Raw material cost^NH3 at $115/ton
Conversion costs
Total direct cost
Total Indirect Cost0
Total Direct and Indirect Cost
Byproduct credit at $55/ton
General and administrative expenses
Total Annual Revenue Requirements
--
Plant size
100 MW
12.3
1.8
14.1
1.60
2.00
3.60
2.50
6.10
(2.54)
0.73
—
4.29
50 MW
8.1
1.2
9.3
0.80
1.40
2.20
1.80
4.00
(1.20)
0.54
'"—
3.34
••-
...
25 MW
6.2
0.9
7.1
0.40
1.10
1.50
1.40
2.90
(0.60)
0.45
__
2.75
. •
a Basis: coal 3.5 percent sulfur.content, 90 percent S02 removal 90
percent capacity factor, remaining plant life of 10 yr, U.S. GuIt
Coast plant location.
b Includes cost of utilities, labor and supervision, maintenance and
repairs, supplies, and laboratory operations.
c Includes cost of payroll, overhead, depreciation, taxes, and insurance.
4.2-99
-------
TABLE 4.2-22. SUMMARY OF ESTIMATED CAPITAL AND OPERATING COSTS
FOR AN AMMONIA-BASED FGD FACILITY ON A NEW 500-MW COAL-FIRED POWER UNIT3
(millions of mid-1979 dollars)
Capital Cost
Direct capital investment
Engineering design, field expense,
contractor fees, and contingency
Fixed capital investment
Land, working capital, and interest
during construction
Total Capital Investment
Operating Cost
Cost of raw material0
Conversion costs
Total direct costs
Total Indirect Costs6
Total Direct and Indirect Costs
Byproduct sales revenue^
Total Annual Revenue Requirements
Process type
Regenerate,
sulfuric acid
production
33.8
12.5
46.3
10.2
56.5
1.22
7.35
8.57
10.58
19.15
(4.86)
14.29
Nonregenerable,
ammonium sulfate
production
25.3
9.4
34.7
8.1
42.8
7.50
5.60
13.10
8.30
21.40
(8.40)
13.00
Basis: 3.5 percent sulfur in coal; 90 percent removal: power unit
on-stream time, 7000 h/yr; stack gas reheat to 80°C (175°F) by indirect
steam reheat; entrained water 0.5 percent by weight (wet basis); Midwest
b plant location; remaining life of power plant, 30 yr niawest
Includes cost of equipment and facilities for makeup handling and prepar-
ation; particulate removal; S02 absorption; reheat; flue gas handling-
services anUfaCtUre> handll"ng' and st°rage; utilities; and construction
d JnC]Ud.es COSt of catalyst for regenerate process; NH3 at $150/ton
and supplies1 °f °perating 1abor and supervision, utilities, maintenance,
* Rwn!ndeV°S? ^P^0,11 overhead, depreciation, taxes, and insurance.
Byproduct sale at $39/ton for 100 percent sulfuric acid and $57/ton for
ammonium sulfate.
4.2-100
-------
estimates of capital and operating costs for an FGD system on a new 500-MW
coal-fired power unit firing 3.5 percent sulfur fuel. Table 4.2-23 sum-
marizes cost estimates for ammonia-based processes and for several other
advanced FGD processes.184
Note that annual revenue requirement is reduced significantly by credit
for sale of the byproducts. Requirements for particulate collection ahead
of the absorber would increase the capital investment.
Energy and environmental impacts—The energy requirements of a non-
regenerable ammonia-based process are less than those of many other FGD
processes. The major components consuming energy are the FD fan required to
overcome the pressure drop in the absorber and, to a lesser extent, process
pumps. Thermal energy is also needed to reheat the gas leaving the absorber
before discharge to the atmosphere. The FD fan typically requires 60 to 70
percent of the total electrical energy consumed. The total energy consump-
tion of such a system operating on an industrial boiler is 1.0 to 1.5 per-
cent of the total heat input to the boiler. In a typical regenerate
process, the energy consumption could be as high as 3 to 5 percent of the
total heat input to the boiler. The energy requirements of different
process steps in a typical regenerable process with byproduct elemental
sulfur are 50 to 60 percent for ammonia scrubbing, 30 to 35 percent for
Claus gas preparation, and 10 to 15 percent for Claus reaction and ammonia
stripping.164
The cooling required to condition the gas for absorption of S02 yields
substantial heat recovery, which is normally lost in the flue gas or in any
hot quench and hot scrubbing system. The recovered heat may be utilized
within the plant for preheating boiler feed water or for other process use
in other plants.
Although the primary environmental consideration in a flue gas desul-
furization process is S02 removal efficiency, attention should be directed
to the overall environmental effects. Most of the particulate matter in
flue gas is generally removed before the gas enters the absorber. Pulp and
paper plants using nonregenerable ammonia-based FGD systems are reported to
have achieved S02 removal efficiencies as high as 95 percent with inlet
gases containing 5000 to 6000 ppm S02.166 Almost all particulate matter
4.2-101
-------
TABLE 4.2-23.'
SUMMARY OF ESTIMATED CAPITAL COST OF FLUE GAS
DESULFURIZATION PROCESSES?"4
(mid-1979 dollars)
Total capital
investment,
million $
Net unit revenue
requirement,
mills/kWh
Ammonia absorption—ammonium
bisulfate regeneration—
sulfuric acid production9
Ammonia absorption—scrubbing
liquors saturated with ammonium
su I fate—ammonium sulfate
production
Limestone slurry absorption
ponding of sludge
Magnesia slurry absorption
sulfuric acid production^
Sodium sulfite absorption-
sulfuric acid productionb
Regeneration process.
4.2-102
-------
larger than 7 micrometers is removed in the cooling stages; some of the
participates smaller than one micrometer are also removed.172 Chloride
removal occurs during flue gas cooling.
Pilot tests of regenerable ammonia-based processes have demonstrated
S02 removal efficiencies of more than 90 percent.164,185 One such process is
designed to remove up to 99 percent of the S02 with an 8 to 10 percent
increase in the cost of raw materials and utilities. Designers of another
regenerable process claim to remove S02 and N0x simultaneously in presence
of a catalyst.185
Accumulation of fly ash and other particulates in the absorber bottom
can be controlled or eliminated by a quencher loop purge stream.
The formation of ammonium salt fumes is an environmental consideration
unique to the ammonia-based process.
Operational status and development—Ground-breaking studies in ammonia
absorption date back to at least 1883, when a British patent was issued to
Ramsey.185 Exhaustive work later conducted by Johnstone and coworkers at
the University of Illinois produced a collection of fundamental data that is
the technical standard for current investigations.185
A pioneer ammonia-based S02 absorber began commercial service in 1936
at Consolidated Mining and Smelting Company. Ammonia absorbers are
successfully applied in sulfite paper processes, where unusually severe
operating requirements necessitate extremely reliable and versatile perform-
ance. With the sole exception of the 'blue plume,1 ammonia absorption has
been found totally acceptable. With a view toward adapting ammonia desul-
furization to power plant flue gas, TVA conducted an extensive pilot program
at its Colbert station.186
At present, ammonia-based FGD processes offered by many vendors are
widely used in the pulp and paper industry, sulfuric acid plants, and other
miscellaneous plants.166-168,172 All such commercial processes are of the
nonregenerable type.
Results of 3 years of operation and tests at a paper mill show that a
5000 ppm S02 concentration in the recovery boiler flue gases can normally be
reduced below the required level of 300 ppm S02 in the stack effluent. At
4.2-103
-------
the same time, ammonia concentrations in the effluent gas are relatively
low. Analyses of particulates (smaller than one micrometer) leaving the
system indicate that about 20 percent is in the form of ammonium salts, most
being sodium and potassium salts originating in the furnace. "2
Seven ammonia scrubbing systems have been operating on industrial
boilers, the first since October 1973 with no major problems. Great Western
Sugar retrofitted ammonia-based FGD systems on lignite-firing boilers (0 7
to 1.2 percent sulfur coal). The plants operate about 3 months per year in
the fall to process sugar beets. The FGD systems reportedly operate
well.187
Several ammonia scrubbing units operating since 1977 have easily
reduced the S02 content of sulfuric acid plant tail gas to the level
required to meet emission standards. "7 Tnese units have demonstrated the
operability and reliability of the ammonia scrubbing system, though large-
scale application of the system, as in the utility industry, is yet to be
shown.
4.2.3.6 The Wellman-Lord Process--
In this process an aqueous sulfite solution is used to absorb S02
Sodium bisulfite is formed as the S02 is absorbed from the gas stream; the
S02 is then released in a concentrated stream in the stripping step The
regenerated absorbent is returned to the absorber loop. The concentrated
S02 stream with water vapor enters a condenser, where most of the water is
removed. If necessary, the resulting S02 stream may be further dried in a
concentrated sulfuric acid drying tower. Sulfur values from the S02 stream
may be recovered as liquid S02, liquid S03) sulfuric acid, or elemental
sulfur. The product is determined by potential use, market demand, and cost
of transportation to the destination.
A typical Wellman-Lord system as applied to a combustion process is
shown in Figure 4.2-26.
Process chemistrv-ln sodium sulfite/bisulfite systems, it is desirable
that any fly ash or other particulate matter be removed before the absorp-
tion step to reduce the need for process purge and thereby reduce the need
4.2-104
-------
O
O
OO
rc
Q.
vo
C\J
CM
OO
-------
for makeup of fresh scrubbing solution. An electrostatic precipitator
(ESP), fabric filter, wet particulate scrubber, or other device may be used
for particulate removal. The gas stream normally is cooled to its adiabatic
saturation temperature in a wet scrubber or presaturator.
The basic process for absorption of S02 by aqueous scrubbing liquor is
given by the following reactions:
C1) S°2 (9.) * S°* (aq)
C2) S°2 (aq) + H2° -» H2S03 -> HS03- + H+
(3) HS03 -» H+ + S03=
Recovery processes are based on the chemistry of the sulfite/bisulfite
buffer system. After appropriate pretreatment, the flue gas containing S02
enters the absorber, where it is brought into contact countercurrently with
a sulfite solution. The sulfite absorbs and reacts chemically with the S02
forming the more soluble bisulfite product.189
Oxygen and S03 in the flue gas also react with the sodium sulfite
forming the unreactive sulfate/bisulfite. The presence of the unreactive
species m the system necessitates a purge from the absorber to maintain the
level of reactive sulfite and to reduce the possibility of scaling 19« »i
The principal chemical reactions in the S02 absorber are absorption'and
oxidation, discussed briefly as follows:192
S02 Absorption: Sulfur dioxide and sodium sulfite react to form
bisulfite.
(4) S02 + S03= + H20 -> 2HS03~
Oxidation: Some oxidation of sodium sulfite to sodium sulfate occurs.
(5) 2S03 + 02 -* 2S04=
In the sodium ion makeup reactions, sodium carbonate (soda ash) or
sodiun, hydroxide (caustic) reacts with sodium bisulfite 'to regenerate the
502 absorbent, sodium sulfite.
4.2-106
-------
(6) Na2C03 + 2NaHS03 -»• 2Na2S03 + H20 + C02t
(7) NaOH + NaHS03 -» Na2S03 + H20
If the product is to be a concentrated S02 stream, the bleed stream is
regenerated by use of single-effect evaporators, double-effect evaporators,
or either atmospheric or vacuum steam stripping. The basic chemical reac-
tion for regeneration of the alkali absorbent is then:
(8) 2NaHS03 * Na2S03 + H20 t + S02 t
The sodium sulfate formed by oxidation of sulfite or by absorption of
S03 must be purged at approximately the rate of formation. It can be dried
for sale or disposal, or it can be neutralized and discharged as an innoc-
uous effluent.
The concentrated S02 stream leaving the regeneration step can be used
to produce sulfuric acid, sulfur, liquid S02, or some combination of these,
depending on available markets. The chemistry of these regeneration pro-
cesses is as follows:
Sulfuric Acid: Sulfur dioxide reacts with oxygen in the presence of
vanadium pentoxide catalyst to form S03.
(9) 2S02 + 02 •* 2S03
The S03 reacts with water to form sulfuric acid.
(10) S03 + H20 •* H2S04
These are the reactions that occur in a typical sulfuric acid plant.
Sulfur: Methane (natural gas) reacts with S02 to form hydrogen
sulfide.
(11) 2CH4 + 3S02 -» 2C02 + 2H20 + 2H2S + S
The hydrogen sulfide reacts with sulfur dioxide to form water vapor and
sulfur.
(12) 2H2S + S02 -> 2H20 + 3S
The overall reaction is:
(13) CH4 + 2S02 -> C02 + 2H20 + 2S
4.2-107
-------
Other commercially available reductants that can be used in place of methane
(natural gas) are carbon monoxide, hydrogen, higher hydrocarbons up through
propane, and the products of coal gasification (low-, medium-, and high-Btu
coal gases).
Liquid S02: The S02 vapors contact silica gel to remove moisture The
vapors then are compressed and condensed; the resulting liquid is collected
and stored in pressurized tanks.
Liquid S03: The S02 reacts with oxygen in the presence of vanadium
pentoxide catalyst to form S03, which is then compressed, condensed, and
stored in pressurized tanks.
System descriptlon-A sulfite/bisulfite FGD system can be considered in
terms of the following general steps:
o
o
o
o
o
Flue gas pretreatment
S02 absorption
Absorbent regeneration
Sulfur product recovery
Purge treatment
Efficient flue gas pretreatment is important to these FGD systems
because, in reducing particulate contamination, the requirements for regen-
eration and purge are reduced. The flue gas to be treated is taken after
the electrostatic precipitator at a temperature of about 149°C (300°F) and
passed through a venturi or tray-type prescrubber,' in which it is cooled to
around 54°C (130°F) and humidified. A tray-type prescrubber satisfactorily
cools and humidifies the gas with low pressure drop, but removes less of the
fly ash and chlorides. Humidification of the flue gas in the prescrubber
prevents the evaporation of excessive amounts of water in the absorber
Scaling and plugging problems are virtually eliminated by use of the pre-
scrubber and clear scrubber solutions as well as by the solubility of the
absorption product, sodium bisulfite, which is more soluble than sodium
sulfite.
A well-designed prescrubber removes up to 99 percent of all chlorides
in the flue gas; this should help maintain a low level of chloride in the
scrubbing liquor and reduce the potential for stress corrosion. The absorp-
tion of chlorides and some S02 and S03 can cause the scrubber water to
become acidic. After neutralization with lime when necessary, the fly ash
4.2-108
-------
and other solids collected by the prescrubber are pumped to an ash disposal
pond as about a 5 percent slurry.
The S02 absorption step in the WeiIman-Lord process has been performed
primarily in a tray tower absorber. The humidified gas from the prescrubber,
is passed upward through the absorption tower, where it meets the counter-
current flow of the aqueous absorbent solution. The Wellman-Lord systems at
both NIPSCO and Public Service of New Mexico have shown individual S02
removal test results above 90 percent.193-195 Mist elimination can be
achieved with either chevron-type or polypropylene-mesh-type units. Reheat
of about 28°C (50°F) is often practiced to reduce condensation and the
resulting potential for corrosion downstream from the absorber.
When S02 is the product of the regeneration step, various sulfur prod-
ucts are possible. Production of elemental sulfur, which has been done at
two full-scale utility FGD systems, requires a reducing gas (such as methane
or natural gas, hydrogen sulfide, or carbon monoxide).196
Sulfuric acid, the most widely used commodity chemical in the world, is
another possible product. Although acid production consumes less energy
than the production of elemental sulfur, the availability'of a market in the
area must be considered.197 Since sulfuric acid is a byproduct, a definite
market is needed to maintain proper operation of the FGD system. The Public
Service of New Mexico, San Juan Station will produce sulfuric acid directly
when Unit 4 comes on stream.198
Other options are to produce liquid S02 or S03; however, the market for
these chemicals is relatively small. A decision to produce either would be
site specific.
In addition to the decision regarding the byproduct materials, disposal,
of the purge stream must be considered. Some of the sodium sulfite in the
absorbent solution will be oxidized to sodium sulfate. The degree of
sulfate generation is a function of the S03 content of the incoming gas
stream and the excess oxygen. Since the sulfate form has no S02- absorbing
value, it must be purged from the process. The sodium sulfate, present in
the decahydrate form (Glauber's salt), is continuously removed from the
regenerated absorbent solution by vacuum crystallization. Subsequently, it
is prepared for disposal or dried for storage and sale. The sodium sulfate
4.2-109
-------
is used in the pulp and paper industry and the fertilizer industry, two of
several potential markets.
Operational status and current developinents-Th* Wellman-Lord process
a proprietary process of Davy McKee, Inc. (formerly Davy Powergas), is in
wide use both in Japan and the United States. In Japan, the Wellman-Lord
process is used on 14 oil-fired boilers and 3 Claus sulfur plants; 3 of the
14 oil-fired boilers are electric generation units.™* In the United
States, the process is used on four Claus sulfur plants, two sulfuric acid
plants, and three coal-fired electric generation units.200 Un1ts are under
construction for two more coal-fired electric generation units and three
coke-fired boilers.^o Tables 4_2_24 and ^^ ^ ^ ^.^ rega
these units.
The first large installation of a Wellman-Lord unit at a Japanese
utHity plant came on line in late spring of 1973 at the Nagoya Station of
the Chubu Electric Power Company. The unit, constructed almost entirely of
stamless steel, has the capacity of treating 663,000 mVh (390,000 scfm) of
flue gas, containing 2100 ppm S02, from a 220-MW boiler that burns oil
containing 3 percent sulfur. It was designed and constructed by Mitsubishi
Kakoki Kaisha, a licensee of Davy McKee, Inc. The Nagoya station is a
peak-shaving power plant. The FGD system was contractually required to
handle stack gas fluctuations from 35 to 105 percent of design flow in 22
minutes, while maintaining outlet S02 emissions at less than 150 ppm. In
May 1973 the Japanese government measured the outlet S02 at 130 to 135 ppm,
which represents over 93 percent S02 removal.20'
The flue gas from the boiler is cleaned in an ESP and precooled to 58°C
(136°F) before entering the tray absorber. The L/G ratio is 0.7 liter/Nm*
(5.2 gal/1000 scf).2°2
When the unit was originally put on line, the operators inadvertently
switched from low-sulfur (0.7%) to high-sulfur (4%) fuel oil; this change
had mtle effect on the outlet S02 concentration, however, and because of
automated adjustments of the process equipment the concentration did not
exceed 150 ppm.203
The Nagoya unit has proved reliable and easy to control. The system is
h19hly automated; only two operators are needed for the entire operation,
4.2-110
-------
0
o
Csl
CO
LU
1—
H-
i— i
LU
i
r^
1 — 1
z
— - >
LU
31
H-
z
fr — |
^y
O
HH
1—
-P *~N
CO E
0) M-
t- u
+J t/)
» — '
OJ
E en
t— • n
0 E
> z
ro
0
c
>P_
Dl
'i
in
ro
,o>
OJ
LJ-
j
c
o
4-*
ro
u
o
•a
ro
ro
§•
o
o
jj "c 4-> *J c c c:
c 4-> 4-> 4-> 4->
Q; OJ QJ OJ QJ OJ OJ
r? 5* 5" u1 u u u
QJ cu a) cu a) cu aj
iS tS ee ce D: - a: oc
i i I i i i i
LO o o co" o o o
^- ro co f^ en co co
C\J r— •— CO »— •— p~
*J ^_3 *J
c c c
to m fo
'o. 'o. eu
-a -o-o
10 = ro -1 ro ro ro
•i- a. •!-••- a. CL o.
^ w 3 3 U) « «
i, -j i^_ L^ 333
,— - ro ' — i— (O ro o f-J o
<*- J*-
--^ i^* *4— *4— *~~
^ m *r- *r~ TO
* " - -z <->
0 ro >.**- **- M~
O •»" 4^ ^ *^ ir™ 5 5 £
t! £-^5 S--O-D-O
g 3 1 | » | | |
w .^ « f— .-f- *0 TO JO
.r_Qt/)^Ot^>t^(^
C
=>
-«->
c:
ro
a.
i_
ID
ZJ
fO
*J
c
QJ
e
O
o
o
CO
ID
«3-
4-J
rc
. CL
£_
a*
o
Q.
•D
0)
t-
"*~
, —
ra
0
u
^
in
•~
'v*%
c
ro
Q.
&
O
.^ s-
4-> (T3
Q
C
TO
CL
S-
3
v*_
-j
Ul
ro
^J
c:
OJ
E
01
UJ
O
O
0
§
CO
^^ ,
CO
CO
c
ro
a_
s-
s
o
CL
•o
QJ
S-
i in
, — t_
ro OJ
O • —
O "-
O
1 CSJ
O ^-*
O
r—
o
u
X
a>
s:
r~-
co
CM
c
ro
r—
CL
£-
O
CL
OJ
'C __^
r— in
ro s-
O QJ
U i—
^ O
1
O CM
0 --^
r—
O
(J
X
.£
c f
O OJ
U "4-
3 0
U
c ro
O CL
u ' E
ts in design o
ublic Service
Waterflow, N.
•i— O-
c
3
4->
ro '
CL :
T3 '
U
ro
u
£- 1
3
14—
CO
^-. !
O
§ I
o
£. i
CD
ro i
CM |
li
ii
'
1/1
£-
O1 i
•r-
O :
•o ;
t_
<*- '
\ :
1 !
o
oj :
OJ ;
i
t—
j
OJ
0
CJ
etty/Del aware
u
4.2-m
-------
'— in i—
•I- O •!-
CJ -I- -^- .r-
•1- Ol •!-
CO
CK
LU
ce
o
in
csi
esj
UJ
—i
CD
•— cr i—
•r- 3 .,-
g
o I
•— ra o i— r—
T- 1— CM •!- -f-
•I- ^ •,.
<_> •(- -r-
E -i- •,-
4.2-112
-------
including startup and shutdown. Automation is necessary because of the
widely fluctuating load and the shutdown of the boiler every weekend.
The first application of a Well man-Lord system to a coal-fired utility
boiler is the 115-MW Unit No. 11 at the D.H. Mitchell Station of Northern
Indiana Public Service Company (NIPSCo). Flue gas from the boiler passes
through an ESP for primary particle removal (about 98.5 percent effi-
ciency).204 The booster blower delivers the flue gas through a variable-
throat venturi prescrubber' to the absorber. The flue gas is cooled and
saturated in the prescrubber by water recirculated from the bottom of the
prescrubber and back to the venturi sprays. Fly ash captured by the scrub-
bing solution is purged continuously from the system to a fly ash pond. For
short periods, the scrubber can handle considerably greater than design fly
ash loading. Water from Lake Michigan is used to make up water lost through
the purge and evaporation processes.205
Absorption of the S02 from the prescrubbed flue gas takes place in a
three-stage absorber. Each stage consists of a valve tray and a collector
tray.206 A S0dium sulfite solution absorbs and chemically reacts with the
S02 to form sodium bisulfite. A mist eliminator removes entrained liquid
droplets from the gas leaving the absorber stack. The stack incorporates a
direct-fired, natural gas reheat system, to heat the cleaned gas to 82°C
(180°F), if necessary, to enhance dispersion of the steam plume.
The product solution collected on the bottom collector tray overflows
to the absorber surge tank, from which the solution is pumped through a
filter to ensure that no fly ash enters the evaporator system. A small
sidestream of the filtered solution is sent to the purge treatment unit to
remove the sodium sulfate. The purge treatment equipment consists of four
chilled-wall crystallizers; a slurry of sodium sulfate crystals forms in the
crystallizers and is then removed in a centrifuge. The resulting clear
solution is pumped to the evaporator for regeneration of the sodium sulfite.
The evaporation system consists of a forced-circulation vacuum evapo-
rator. The filtered solution is recirculated in the evaporator, where
low-pressure (345 kPa, 50 psig) steam is used to evaporate the water from
the sodium bisulfite solution. When enough water is removed, sodium sulfite
crystals form and precipitate. Sulfur dioxide is removed with the overhead
4.2-113
-------
vapors. The slurry formed by the sodium sulfite crystals is withdrawn con-
tinuously to a dump/dissolving tank, where condensate from the evaporator is
used to dissolve the crystals in the solution that is pumped back to the top
stage of the absorber.206
Water vapor is removed from the S02 in water-cooled condensers. The
S02 is compressed by a liquid ring compressor for introduction to an S02-
reduction facility designed and operated by Allied Chemical. The gas stream
is about 85 percent S02; the remainder is mostly water vapor.^oe Tne off_
gases from the reduction facility are burned in a tail gas incinerator and
are returned to the absorber inlet.
Sodium lost as sulfate in the purge treatment system is replenished by
addition of sodium carbonate to the absorber solution. Soda ash is brought
to the plant in trucks and transferred to the storage bin by a pneumatic
conveying system. It is metered to the slurry tanks by a bin activator and
belt feeder. The soda ash slurry, is pumped to the absorber feed tank by
parallel centrifugal pumps.
The operating history of the unit at Mitchell Station reflects both
operational problems with the boiler and normal operating difficulties. The
system went on line in integrated operation in November 1976. A boiler
mishap in early January 1977 required a 6-month outage for boiler repair
During the demonstration period from late August to mid-September 1977, the
unit met all process guarantees, including 91 percent S02 removal 207 (A
detailed description of actual operating performance is given in Reference
195"}20«SlnCe AUQUSt 19?8' the °Pe^ng history of the unit has been
good.^08
The largest Wellman-Lord installations are being constructed on four
coal-fired boilers at the San Juan station of Public Service Company of New
Mexico (PNM) in Waterflow, New Mexico. The retrofit installations on Units
1 and 2 began operation in April and September 1978, respectively 209 An
installation on the new Unit 3 boiler was completed in'December 1979 209
Unit 4 and its FGD system are scheduled for startup in January 1982 209
Table 4.2-26 provides further details on these units.
The wellman-Lord system for Units 1 and 2 was designed to remove 90
percent of the S02 from the flue gas when firing coal with sulfur content
4.2-114
-------
01
O"1
1-fN
1— O
^^ 2c
I— I LU
•y
0
CD U_
U_ O
o >-
at. z
0
ce:
O£ UJ
co to
u_
0
I/O i-t
:
ro
c
CM
_C
i
f^fff 1 1 ( '
n i CD
LU Q_ ,
^t U
fy O
£z
0
Z n
C3 h-
l/) 1 —
LU CO
O
§
10 ^
CXI
CXI ^3
•r- t***
« aT
u s-
OJ
HI eu
0 3
CD LL.
^
^*
o
r~ o
ro -
s s
CM
*
%w.
o"
r^ o
• o
ro
r-~ CM
ro en
f^
**^
/^*.
o
CM 0
• O
i— CM
*§
O
o y*^
r*» u.
. o
vc o
CM CO
re
*—*
i^
u
en
m
B
u-
re
ts
000
en CM m
<• -00
0
o o o
en CM in
"* r-~°°
C3
o o o
en CM in
<^* — GO
O
O O O
en CM in
'-
£
Q.
0.
C
O
"'""
4-*
re
t-
4->
C
CU
g...
U 3 3 D)
B S re
Wr- -r- t.
O C X CU
*>•££>
<->
01
"c
1-1
en ,-—•
ro <*^ ic
O en ro in
O 0 r- O
o o o o'
en ' — '
ro --^ ID
§cn ro in
0 — 0
o c> o o
\^s v-'
M /-^ en"
§?-o
0 0 0 C5
o •—
ro /^i— co
o r~ CM m
00—0
o o o o
—
en
c:
T3
Q
* r^ ^--
— 4-> t-
3 T 00 T3
U
•r-00 <0 t-
t£2«K^
CU
IE
I"H
0
o
0
in oo -
* o
o —
*»• in
CM •—
0
O
o
m oo -
. o
O i—
^- in
CM > —
I
*3- 0
O CM
o <*•
/•"s
o
§
«S- CM -
r^ ro
O «3-
CM ~~*
re
>r_
2
4->
^^
t_ C
of scrubbei
mVs
(acfm/traii
c>
ro < — in —
to en — CM
CM 1C CM VD
o en — o
ro i— in i —
ro en — CM
CM ID CM 1C
O Cr7r-'o
•"" ~~"~ C-
ro ro o ic
o ID ic en
CM VD ID O
o en o o
r"N ^-^
ro CM o in
CM CO CO —
CM ID 1C ' —
o en o o
^•x t-
f f CU
222 s-
•^^» 3 eu
33 S- XI
4J t) . r^
OQ 1.1 3
oT^'+J (J
4-> 3 U)
re o
U J= JC
*> ••->
f 5 -5
re
01
3:
i
m
CM
CM
i
CM
m
CM
CM
^
O
CM
in
CM
CM
^"T
^t
CM
^— V
in
CM
CM
i
CM
.C
c
o
4->
c.
"^.
s^
,3-
•o c
£.?
1- 0
r—
)
^
r—
*J"
'i
Ol
?
^
ct
^
c
re
>^
Jj
u
re
ex
re
u
eu
*
CU
in
£
eu
•c
o
o.
o
:
•r—
re
K-
4->
CU
i-
re
o.
re
c
0
•D
re
c:
re
CU
re
•^
r—
^
S_
eu
o
^:
u
re
LU
re
4.2-115
-------
ranging from 0.59 to 1.3 percent by weight with an average of 0.8 percent.
Prescrubbers with relatively high pressure drop were specified for the
system because of New Mexico's stringent regulation on fine particulate
emission and to provide backup for the ESP's. The purge from the pre-
scrubber is sent to the plant wastewater system, where it is treated and
then recycled.211
Scrubber-absorber operation does not affect the power plant operation
because the FGD system can be bypassed. During a normal unit startup,
scrubbers and absorbers are put on line after the ESP's are functioning'
At PNM's request, four scrubber-absorber modules were installed on each
power plant unit, each sized to handle one-third of the total gas flow.
Therefore, the plant has one complete spare scrubber-absorber module and can
establish a maintenance program whereby absorbers are rotated in and out of
service for routine and preventive maintenance. Each absorber is a five-
stage tray absorber.212
Each scrubber-absorber module is operated independently (except for
control data transmission to the control panel board), and is put on line
separately. Each unit has a reheat system to protect the stack from corro-
sive products resulting from condensation in the flue gas when three modules
are in operation simultaneously.213
Two 2,840,000-liter (750,000-gallon) tanks for absorber product and
feed solution provide surge in the system to prevent the chemical plant
operation from being affected by normal fluctuations in the scrubber-
absorber area. The ideal operating situation is to have the feed solution
tank full and the product tank level very low.
The chemical plant has two double-effect evaporators, which provide
steam conservation (the overhead vapors from the first effect are utilized
as the heat source in the second effect). For reliability, each evaporator
is connected to steam and off-gas compressor manifolds so that any one
evaporator can be taken out of service for maintenance without affecting the
operation of the three absorber units. The availability of the chemical
plant to operate depends on operation of the power generation unit. To
date, this has been the major operating problem with the FGD system.
4.2-116
-------
The purge treatment plant consists of three low-temperature crystal-
lizers where sodium sulfate is precipitated in the decahydrate form. The
crystal!izers are followed by a melt tank and evaporator. The evaporator is
similar to, but smaller than, the main evaporators. Water is driven off,
and the sulfate purge is centrifuged, then dried in a flash dryer. The
entire purge treatment plant was specified to obtain a concentration of 70
percent sodium sulfate and 30 percent sodium sulfite in the dried purge
salt. The actual sulfate content has been very high; purities have been
achieved in the area of 90 percent sodium sulfate. Residual moisture in the
dried purge salt has been less than 1.0 percent.212
Two identical Allied Chemical S02 reduction trains were installed as
part of Units 1 and 2 FGD systems. Each of the trains has a design .capacity
of more than 50 percent of the total FGD system capacity, based on the use
of low-grade coal (1.3% sulfur) in the boilers. Public Service Company of
New Mexico specified two S02 reduction trains with the objective of achiev-
ing an FGD system with essentially a 100 percent onstream reliability.
The system being designed for the Units 3 and 4 power plants is some-
what similar, with the following exceptions: the prescrubbers have a. lower
pressure drop for energy considerations. The five-stage tray absorbers also
function quite well for residual particulate removal; sulfate purge quality
is not degraded because the fly ash is filtered out of the solution.
In the interest of economy the absorbers are designed for four units in
operation per boiler when burning low-grade coal; therefore Units 3 and 4
will not have the one-module spare as on Units 1 and 2. The coal being used
at San Juan Station has rarely exceeded 0.95 percent sulfur for long periods
of time, however, so the plant operability and maintenance capability should
not be affected by lack of the spare module.212
A sulfuric acid plant, which will be installed while the FGD system for
Units 3 and 4 is being constructed, will handle the regenerated S02 streams
from all four units. The design capacity is based on low-grade coal [470
tons (426 Mg) per day of 100% sulfuric acid]. The sulfuric acid plant will
significantly reduce operating costs, reduce the demand for natural gas,
directly provide the product that the sulfur is used to produce without the
intermediate sulfur step, and deliver readily salable material for which
there is good demand.198
4.2-117
-------
Further current operational data on FGD systems Tn the United States
sites are available in the quarterly EPA Utility FGD Survey.
Control cost-A recent EPA publication gives estimates of the capital
costs and the operating and maintenance costs of FGD systems, including the
Wellman-Lord process, which achieves 90 percent S02 removal while firing
either of two high-sulfur eastern coals. *is These estimates are presented
in Figures 4.2-27 and 4.2-28.
The quarterly EPA Utility FGD Survey contains reported and adjusted
costs for utility FGD systems. The survey data for the Northern Indiana
Power unit and the two Public Service of New Mexico units are given in Table
4.2-27. The capital and operating and maintenance costs are given as
reported and as adjusted to July 1979 dollars by use of the Chemical Engi-
neering (C-E) plant cost index with the assumption of a July 1979 index of
236.5. The capital costs are in the $150/kW range.216
Energy and environmental impacts-This FGD process normally has only
one purge stream requiring treatment, the sodium sulfate bleed. A portion
of the sodium sulfite in the absorber is oxidized to the sulfate form, which
is not an absorbing species.
During the Fifth Symposium on Flue Gas Desulfurization sponsored by the
U.S. EPA in Las Vegas, Nevada, in March 1979, D.W. Ross, Director of Tech-
nical Services for Davy McKee, reported that the Wellman-Lord systems at
both D.H. Mitchell Station of NIPSCO and the San Juan Station of PNM have
experienced about 3 percent oxidation of the absorbing liquor to the sulfate
form. The sulfate (Glauber's salt) is crystallized from the purge stream
and is marketed. The actual sulfate content of the crystalline purge is
about 90 percent with residual moisture of less than 1 percent.
The positive side of this process is that there is a usable, salable
product produced (sulfur, sulfuric acid, liquid S02). There is no large
volume of possibly environmentally hazardous sludge that must be stored or
otherwise handled and disposed of. The sodium purge product (Glauber's
salt) may be salable. Regenerate sodium sulfite FGD systems are among
those having the least pollution resulting from S02 removal
4.2-118
-------
400 •
•09-
O
O
D-
<£
O
7.0% S COAL
3.5% S COAL
200
400 600 800 1000
BOILER CAPACITY, MW
Figure 4.2-27. Estimated capital cost for Wellman-Lord FGD systems 215
achievinq 90 oercent SOo removal firing either^ of two eastern coals.
4.2-119
-------
200
400 600 800 1000
BOILER CAPACITY, MW
Figure 4.2-28. Estimated operating and maintenance costs for
Wellman-Lord F6D systems achieving 90 percent S02 removal firing
either of two eastern coals.215
4.2-120
-------
oo
LU
00
OO
•=c
OID •
1— I-1
C£ OO I
LU I — |
O- OO :
°s
•• 0
00 LU
0 1—
I— i OO
O --3
z: Q
o «a:
1 1 1 f— i
"5T° •r^T"
2- *-3-.
LU
OO LU
>- 1—
OO OL
O
O D-
CD LU
1 i fV
•
OJ
-t-
o c
1— n
(.
•*•
1
(1
C
r
t
-> i
~
v -(
I) r-
*—
UJ
•
01
x
u_
*•* /•"•*
u t-
O) TO
U OJ
— >>
O *— '
TO
0
.
*^*x
^c s—
- J* TO
J_\ CU
!«••»• >>
J, -
J ° *%
a u
a. ro
0 **-
in
ZT3 •(->
-Cm
TO O
3 U
V r-
D TO —
D 4-> TO
- ••- 3
J O. C
C TO C
^ U TO
g
LJL.
.$".13
•t- TO -i-
=3 in =
en •—
ro 0
00 ^
r~- ID <• on
^- r-^ r- f^-
10 ^ 00 i—
s^'
ix) m
T3°° -D^ -0 -D T3 T3
ft* 52: 5 5 5 5
u in «- "> t £
030303
& -o' Q; j * Si
^ ^ Oi *31 CU *s-
cn i£? vc c? 01 r~ CM en en r-- CM en
uoS-i odcn r-^cn oocn r— en pocn
in r- ix> i— CM •— •* i— CM^- "^ C-
^ m
r~- ix>
""' . M-*
»••—?
o o
C u <->
- 2 . . 'i
o E ^
s * ^
<^;z ^ ».
TO Ol 0 O
TO~5 OJ eu
•f +> U <->
"c '£ '> ">
~ . oj TO oj ro
« 1,= .,=
(D C O W
5 S - - S - S
tea,— ^ato'- xitocM
£ S £
jj Ul
w O
o u
u
i — TO
TO £
£ a.
Q. TO
PO U
0 . e
C -I-* U) . •»-
• r- W +J (/)
o ui +•> "O
-o u O tn ,- ' u o -o
3 TO r— r-
r-. • ^J TO r— U
(j «r- 3 TO C
C Q. C 3 ••-
.,- TO C C
U • TO C. «•"»
,^^. v) TO in
m •»- in u •*-> Oi
rtj O QJ O U
U "D U S_ OJ O
O OJ ••- J- U
t_ T3 •— "D •*- O.
CL • ^ TO T3
-O — =) C C2
o • ~o o
.^- r— C . tf) 3 T3 TO 3-"D **-
U-*-4-> O »— 3-t-*' — ^3
"OC V1(_)(J' — OU< — "O
-§ "*" 4J S >»£ C -^ C t-
ZJ in v> +•> ••- c: *r- -^
•r-Ur— *f-uow>O s— '
r— 4^ TO **- U 3 U
O4->Q. CTO>-.UO>-)O
I_»i-TOCO3TOC:i-CTS-
4-> Q. U •*- — C =J -i- •»-> t- •*£
CTO 4->CC COJC
uaiu
^j A) "C QJ -*-1 m
TO t— T3 U Ol $— X TO QJ TO
i — «r- C C OJ •*-•!- T3 ' — E ^~
STj.r-'p- fc-TO1^ TO 3 QJ ^
U OJ O U <_>
+JTOTOOTO«tJTOt-4->»— 4-»
TOOO4->OOO>TOOJTO
ID JO O T3 «* «*- O»J= "- "->-^
4.2-121
-------
In addition to concern for the environmental impact of the FGD process
there is interest in the energy requirement of the system. In a recent EPA
publication, the power consumption (energy penalty) of the Wellman-Lord FGD
system is presented as a percentage of the generating capacity of the unit
that it treats. The energy penalty as a percentage of gross output is shown
in.Table 4.2-28.^ The energy penalty expressed in mills per kilowatt hour
is shown in Table 4.2-29.218
4.2.3.7 The Citrate Process-
Sulfur dioxide tends to be absorbed in aqueous solutions; however the
rate of absorption is pH-dependent. The more alkaline the aqueous solution
the greater the tendency of S02 to be absorbed. As S02 is absorbed'
however, the aqueous solution becomes more acidic as sulfurous acid is
formed. The absorption of S02 in aqueous solutions, therefore, is self-
limiting. By incorporating a buffering agent in the solution to inhibit the
PH drop during absorption, substantially higher S02 loadings can be
attained. This fact has led to the development of the citrate process.
Three citrate processes are or have been available: the Bureau of
Mines Process, the Peabody Process, and the Flakt Process. Figure 4 2-29
shows a typical citrate process that produces elemental sulfur as the end
product. Figure 4.2-30 shows a typical process that produces a concentrated
S02 stream for possible use in sulfuric acid production or liquid SO,
production.219
Process chemistry-22"-2^^ pr1n)ary absorbent 1n tMs process .g
water. The ability of the water to absorb S02 is enhanced by the addition
of citric acid (C6H807.H20) and either caustic (NaOH) or soda ash (Na2C03).
The absorption of S02 occurs in three steps.
1. The soluhilitv nf en •;„ ,.,,.•._„ _•--,••. . _.
The S02 is dissolved
The solubility of S02 in water is limited
and sets up the equilibrium:
O) S02 + H20 HS0
1ncreased by removing the hydrogen
4.2-122
-------
TABLE 4.2-28. ENERGY PENALTIES ASSOCIATED WITH
WELLMAN-LORD S02 CONTROLS217
Coal type,
% sulfur
Eastern, 3.5
Eastern, 7.0
Capacity,
MW
25
100
200
500
1000
25
100
200
500
1000
Energy penalty,
% of gross output
520-ng/J
(1.2-lb/106 Btu)
regulation
3.80
3.64
3.45
3.34
3.25
90% S02
removal
3.80
3.64
3.45
. 3.34
3.25
3.80
3.64
3.45
3.34
3.25
4.2-123
-------
TABLE 4.2-29. ENERGY PENALTIES ASSOCIATED WITH WELLMAN-LORD
CONTROLS
218
Coal type,
% sulfur
Eastern, 3.5
Eastern, 7.0
Western, 0.8
Anthracite
Lignite
Capacity,
MW
25
100
200
500
1000
25
100
200
500
1000
25
200
500
500
500
Energy penalty,
mills/kWh
520-ng/J
(1.2-lb/106 Btu)
regulation
1.05
1.01
0.95
0.92
0.90
90% S02
control
1.16
1.11
1.05
1.02
0.99
1.16 '
1.11
1.05
1.02
0.99
1.39
1.26
1.22
1.02
1.02
4.2-124
-------
BOILER
EXISTING
STACK BOOSTER
FAN
PREHEATER
SULFUR
SLURRY SULFUR
PRODUCT
CRYSTALLIZER
PRECONDITIONING AND
S02 ABSORPTION
SULFUR PRECIPITATION
AND RECOVERY •
Figure 4.2-29. Typical citrate S02 control process (producing
sulfur product).219
4.2-125
-------
CO
00
o
oo
O)
o
Oi
QJ
O
GO
O)
O
(O
o
CO
I
CM
O)
D-)
g
co
4.2-126
-------
(2) H+ + Cit" ± HCit"
2.
3.
(3) H + HCit j H2Cit
(4)
HoCit" -» H3Cit
Reportedly 90 percent of the S02 is
225
this manner.
for sodium ions
removed in
Sodium hydroxide or soda ash is added to make up
lost in the purge stream. The S02 absorption is a function of pH,
temperature, and the S02 concentration of the gas stream.
source also reports the citrate/sodium ratio can play
icant role. The solubility of S02 increases
The citrate requirement of a
and decreases with
about
each pH unit of
system increases with increasing temperature
increasing S02 concentration in the inlet gas stream.
One
a signif-
ten-fold for
Thiosulfate (S203 ) reportedly complexes absorbed 5
in inhibiting oxidation of sulfite to sulfate. The
follows:
02, which aids
reaction is as
(5) 4HS0
S0
23
3H20
Trithionate is the reaction product.
The remainder of the S02 may be removed by the reactions shown in
the Wellman-Lord process chemistry discussion. The regeneration
step is a function of the desired byproduct.
If elemental sulfur is desired, the reaction is as follows
(where hydrogen sulfide can be used as the reductant):
(6) HS03
2H2S -* 3S + 3H20
If S02 is desired, the S02-rich citrate solution is heated to
evolve S02 through the following reaction:
(7) HS03
H+ A H20 + S02t
The SO, may be recovered as liquid S02 or may be converted to
sulfuric acid in an acid plant; the reactions are shown in the
Wellman-Lord process chemistry.
System description—A citrate FGD system can be considered in terms of
the following general steps:
o
o
o
o
o
Flue gas pretreatment
S02 absorption
Absorbent regeneration
Sulfur product recovery
Purge treatment
4.2-127
-------
Basically, the equipment in a citrate process unit can be very similar
to that in a Wellman-Lord process system. The major differences are that
the absorber normally has a packed tower rather than a tray tower absorber
and that both the Bureau of Mines and Peabody processes normally use hydro-
gen sulfide generators to recover the absorbed S02 as elemental sulfur. The
Flakt process normally recovers the absorbed S02 as a concentrated S02 gas
stream for liquid S02 or sulfuric acid production.
As with the Wellman-Lord process, the sodium sulfate decahydrate
(Glauber's Salt) that is formed is purged. The rate of purge is a function
of the oxygen content of the incoming gas stream.
h °Perationa1 status and current dPvPlormPnt.-Two of the three processes
have been tested only in pilot facilities. Startup is in process of a 60-MW
Bureau of Mines process unit at St. Joe Minerals. Also, the Electric Power
Research Institute (EPRI) is currently installing a 1-MW pilot Flakt unit
and plans to install a 100-MW Flakt process unit. Table 4.2-30 gives more
details regarding several installations of these units.
The St. Joe unit was designed to be a 60-MW demonstration unit
Because the smelter associated with the boiler was shut down, the unit has
been operating at 20 to 30 MW; the power produced is fed into the local
power grid. The normal startup problems have occurred. Initial results
indicate S02 removal in the 80 to 90 percent range. 226
Control cost-The quarterly EPA Industrial FGD Survey reports that the
l? °f ^ BUreaU °f Ml'neS (BOM) CUrate Process Demonstration
MW) at St. Joe Zinc, Monaca, Pennsylvania, is $12.7 million in 1977
$2° /Iw'saT't " $14'8 "1l"°n " ^^ 19?9 termS' Thl'S e^tes to ab-t
$25Q/kW,^ which is greater than for Bother FGD systems; however, this
unit is a demonstration unit with a- great deal of redundancy built in
A 1978 publication shows the capital and operating and maintenance cost
estimates for the Bureau of Mines citrate process for 500- and 100-MW units
Tab \ *rC6nt SUlfUr C°al " fUe1'228 TheSe eStl'mates -e presented in
Tables 4.2-31 and 4.2-32.
Energy and environmental impact.s-These FGD/processes normally have
only one purge stream requiring treatment, the sodium sulfate bleed. A
4.2-128
-------
I,
4-> O
•r- 1/1
O l^
o. o
in
O
CMIO
O >
co o
a:
&s
c
0
CM4J
>' ! 8 £
J- 1 •£
S !l £
<^»
— '
CO
co
LU
O
O
Q.
o
U.
1 1 1
o:
rr-
i — i
.
CO
OJ
LU
CD
1 —
o
0
0
•>
•O
in
= in
3 -^
0-E
> z:
in
re
c
en
T
o
in
IO
en
•o
LJ-
Locatlon
3
l^.
3
in
re
^
LU
OI
Oi
1
ro
O!
in
^
i
o
o
o
CO
in
CO
01
o>
E
O.
o
CM
O
t.
Oi
a.
o.
.0
en
ID
z:
i.
3
<4_
3
IO
c
LU
o
01
A
in
O 0
i— >i
CO i —
• IO
o o
Q.
*•>
I
,~
CO
CM
i.
01
in
•o
re
01
u
3
in
CM
*re c
^- re 1/1
o * a)
/"» 'T3 4_>
T3 CJ QJ
l_ in
1 OJ C
*io O. IM
O 0
§ S co 1
.,_ to O) (U
. *— * •••—>. 4-> •*- »— 5
tsf t/> ro t/> 1C T3 C7>t/>
lp. gt CO t- C tS
i « ~- s 5 s. ~ 5 1
• ro xs ^~ oj o> re
r« U. CM i— i 14- in CO+J «»-.C
^)O I— O«-«^3 -UOJ
3° ,.-.3 •« Ji-U
= 2 ?: SS^ J^ = •**£
5S ^ o S ~ w c^
S. 1. 1— t. o wi- o) F—
C3 8»r-3O 0)1- 'CIO'
,SS -gS£-!S ££ ^5?
S fc «
s_
3
U-
3
in
i *
c
OJ
LU-
C
en
o ••-
en in
A OJ
CM
o
^_^
o
0
o
LO
o
CM
**
s-
s
a.
•o
a>
S- S-
•r- OJ
i-!- o
re x>
o
jsfc
"S o
^1
c c^ QJ tf
•r- in c re ^
i — J- • O
IBIO OJIO^O^E
c: r-- c Q. o co *
O Oi •«- Ol ^ "^
*— * *r r— SI * 3 , — •*" in
4^ 4_} 10 re .c >•>
.^ re • a> u OJ ^.4^ in
«£*-> orei-^-Im
f— OJ'Cl. "^c3i_l-D
Lu O. OJ O CO (o O O
*w^O CO • •£*•** LU *" Ll-
— i
o
Oi
j
"~* 1
^3" I
CM
•a- o . 1 j
C3 >, !
CM i— i;
' ">
O 0 "
O- !
-^ i
Jl
1
*~^
O
0
° 1
CO |
|
CO
CO
I
i.
" * ^ s-
Q. i ' — ;
•o 10 o
ai o J3
s- t. u
•r- OJ S-
l»- rZ. 3 OJ
i -^ US I
r— O 1 O :>
re JD o Q. :
0 0
o •—
IO
CO
CM >>
$ T -a
•f- IO 01
O
>> a ai o c
eui— c a) o
i — o>
re - r- in (J r~.
5» re o. Q) 01
•r- OJ "~~ • —
oj .0 . — - O- .a * — • OJ i
Oi E *-* 3 4-> in 13
in 5 re s- o *J ro a) E
a) in r— m co F— *>
C3U-4->Ol OJLl-->-C:
c i — — co " — •*-* — to ••-
OJ «r-
£ CO
4.2-129
-------
pnprn, cJIPE FGD PROCESS CAf>I™L SUMMARY
FOR A COAL-FIRED POWER PLANT, 2.5% SULFUR 229
(thousands of dollars except as shown)
Direct construction cost
Major equipment— by unit operation
I - Gas cooling and S02 absorption
II - Sulfur precipitation
III - Sulfur recovery
IV - Sodium sulfate removal
V - H2S generator
VI - Offsites and facilities
Subtotal
Foundations and concrete work
Structural steel
Buildings
Insulation
Instrumentation
Electrical
Piping
Painting
Miscellaneous
Subtotal
Total direct construction cost
Indirect cost and fee
Total capital cost
Cost in dollars/kW
Basis:
500-MW plant 1000-MW plant
14,593
1,050
1,844
350
1,647
162
1,436
997
116
773
955
959
5,962
154
657
29,137
2,141
3,448
466
3,165
268
38,625
2,768
1,911
127
1,357
1,694
1,717
10,935
293
1,074
85
New power plant application- Midwest location
Indirect cost includes engineering design, construction overhead
Julf ^^mrs. ^ admin1strative «Pense. '
4.2-130
-------
TABLE 4 2-32 CITRATE FGD PROCESS ANNUALIZED OPERATIONAL COST SUMMARY
' FOR A COAL-FIRED POWER PLANT, 2.5% SULFUR
(first year costs in thousands of dollars except as shown)
Direct operational expense
Chemical feedstocks
Utilities
Plant operations
Plant maintenance
Payrol 1 overhead
Subtotal direct
Indirect expense
Administrative and overhead
Insurance
Subtotal indirect
•stavl" im rn<;ts
• Total operational expense
Sulfur product credit
Cost (Net operational
Mills/kWh expense)
Interest
Dpnreci ati on
Total expense
Mills/kWh
500-MW plant
1,934
4,822
259
311
139
7,465
244
333
557
975
8,997
(1,360)
7,637
2.29
4,222
1,482
13,341
4.02
1000-MW plant
3,874
9,645
338
458
183
14,498
225
637
862
1 ,840
17,200
(2,721)
14,479
2.17
8,068
2,831
25,378
3.82
Basis: ,
95 percent plant availability; 80 percent average unit load.
Sulfur credit-- $40/ton.
Maintenance services from a fully staffed power plant maintenance
organization
(95 percent nonexempt bonds) SL depreciation-- 2b Years.
Utility financing method expressed in 1978 dollars.
4.2-131
-------
portion of the sodium sulfite in the absorber is oxidized to the sulfate
form, which is not an absorbing species.
The Bureau of Mines citrate process is reportedly capable of less than
2 percent oxidation because of the tendency of the citrate process to
suppress sulfate formation."! Flakt reports that the oxidation of the
suime absorber liquor at the Boliden smelter is between 0.1 and 0 5 per-
cent of the absorbed S02."o Flakt believes that oxidation .& ^.^ fay
the chemical reaction rate due to the complex binding ability of the citrate
ion. Peabody reports 0.4 to 1.0 percent oxidation from laboratory data
which they attribute to the effect of sodium thiosulfate to inhibit oxida-
tion to sulfate.224
The positive side of these citrate processes is that there is a usable
salable produced (sulfur, sulfuric acid, liquid S02). There is no large
volume of possibly environmentally hazardous sludge that must be stored or
otherwise handled and disposed of. The sodium purge product (Glauber's
salt) can be salable.
4.2.3.8 Magnesium Oxide Flue Gas Desulfurization—
In operation on boiler flue gases the magnesium oxide (MgO) slurry
scrubbing system has demonstrated S02 removal efficiencies above 90 per-
cent.^ Tnree full.scale utnity um.ts have been Qperated .n ^ Un.ted
States. Two units were designed for short-term demonstration on oil- and
coal-fired utility boilers. A third unit is operating, and four more are in
the planning stage.
The MgO process is a proven, regenerate FGD system. Magnesium sulfite
is formed in the absorption process and is then dried and calcined to
regenerate and recover 'the MgO for reuse. The same regenerating step
produces a gas stream containing 6 to 14 percent S02, which can be converted
to sulfuric acid, liquid S02, liquid S03, or elemental sulfur. The process
generates no waste streams.
Process chemistry-Fly ash and other particulates are removed from the
gas stream before it enters the absorption tower, either by an ESP or a wet
scrubber. Removing the particulates minimizes solids buildup and impurities
in the regeneration stream.
4.2-132
-------
High-efficiency removal of sulfur oxides from the gas stream requires
that the stream be precooled to 40° to 65°C (110° to 150°F) and that the
absorbent be highly reactive. Magnesium oxide is an excellent absorbent
when it is slaked with water to form magnesium hydroxide, Mg(OH)2 (slaking
is hydration, with supplemental heating when needed).
The main absorption reaction takes place between sulfur dioxide (S02)
and magnesium oxide (MgO) to form either magnesium sulfite hexahydrate
(MgS03-6H20) or magnesium sulfite trihydrate (MgS03-3H20). Some of the S02
may react with MgS03 in the presence of water to form magnesium bisulfite
[Mg(HS03)2], which then reacts immediately with excess MgO to yield addi-
tional MgS03 (hexahydrate or trihydrate). A portion of the MgS03 is
oxidized to magnesium sulfate (MgS04). Some of the sulfur trioxide (S03)
present in the flue gas is absorbed and reacts to form MgS04. The reactions
that occur are as follows:233-235
Slurry preparation:
(1) MgO + H20 -» Mg(OH)2
S02 absorption:
(2a) Mg(OH)2 + 5H20 + S02 -»• MgS03-6H20^
or
(2b) Mg(OH)2 + 2H20 + S02 •* MgS03-3H204.
(3a) S02 + MgS03-6H20 -> Mg(HS03)2 + 5H20
or
(3b) S02 + MgS03-3H20 -» Mg(HS03)2 + 2H20
(4a) Mg(HS03)2 + ,MgO + 11H20 -» 2MgS03-6H20^
or
(4b) Mg(HS03)2 + MgO + 5H20 -
(5) 2MgS03 + 02 -»• 2MgS04
S03 absorption:
(6) Mg(OH)2 + 6H20 + S03 -»• MgS04-7H20-i-
4.2-133
-------
The absorbent slurry Ts pumped into the upper area of the absorber. After
contacting the flue gas and absorbing S02 , the slurry normally falls into a
recirculation tank, in which the magnesium bisulfite formed in the absorber
is converted to magnesium sulfite by reaction with fresh magnesium hydroxide
from a makeup stream. The slurry in this tank contains primarily magnesium
sulfite, sulfate, and hydroxide.
A bleed stream of 5 to 15 percent of the recirculating flow is diverted
to the regeneration circuit. The quantity of slurry extracted is theo-
retically equivalent to the amount of S0x being removed in the scrubber
This bleed stream is either first routed to a clarifier/thickener for con-
centration and then to a continuous centrifuge system or is sent directly to
the centrifuge. The overflow liquid from the thickener and centrifuge is
either pumped back to the main recirculated slurry stream or is used to
slake the regenerated or fresh MgO. The underflow (wet cake) from the
centrifuge system contains crystals of hydrates of magnesium sulfite/sulfate
and some excess magnesium hydroxide.
The wet cake is then conveyed to a direct-contact or fluid-bed dryer
where free and chemically bound water are removed. Typical operating tem-
peratures of a dryer are from 175° to 235°C (350° to 450°F).235 Although
both rotary and fluid-bed dryers can be used, only the rotary kiln type has
been used in the three U.S. magnesium scrubbing units. 23e
The following chemical reactions occur in the dryer: 237 ,23«
(1) Mg(OH)2 -> MgO + H20 t
(2a) MgS03-6HO -> MgS03-3H20
3H20 t
or
(2b) MgS03-3H20
(3) MgS03-6H20
(4) MgS04-7H20
MgS03
MgS03
MgS04
3H20 t
6H20 f*
7H20 t
The anhydrous (water-free) crystals of magnesium oxide «1 percent)
magnesium sulfite, and magnesium sulfate are fed to a calciner (either a
rotary or fluid-bed type) to regenerate the MgO and to liberate the S02 in
the off-gas.239 The calciner temperatures can range from 670° to 1000°C
(1250° to 1800°F) 237>239j240 but
4.2-134
-------
850°C (1500°F) for a prolonged period because "dead burning" takes place at'
and above that temperature. "Dead burned" MgO is characterized by high bulk
density (450 to 700 kg/m3, 28 to 45 lb/ft3)241,242 and chemical unreactive-
ness, which render it ineffective for further use as an absorbent medium.
A reducing atmosphere increases the rate and effectiveness of 'the
calcining operation. In a fluid-bed calciner this can be achieved by
limiting the combustion air. The addition of coke in a rotary kiln calciner
also generates a reducing atmosphere. Using carbon in the range of 1.5 to
2.0 percent and keeping temperatures between 670° and 725°C (1250° and
1340°F) in the calcining operation will yield a product of low density and
high reactivity.242,243 The following reactions take place in the cal-
ciner:237,239,240,244
(1) MgS03 A MgO + S02 t
(2) 2C + 02 •* 2CO t
(3) MgS04 + CO -» MgO + C02 t + S02 t
The regenerated MgO is stored for later use in the flue gas scrubber slurry
system. The off-gas from the calciner is cooled, cleaned, and processed to
yield the desired product, usually sulfuric acid.
System description-The MgO process consists of two major parts: • the
S02 scrubbing system and the MgO regeneration system. Major components of
the S02 scrubbing system are the slurry preparation section, the absorber
section, and the bleed stream dewatering section. Major components of the
• MgO regeneration system are the regeneration section and the sulfur by-
product section.
Before entering the system, the flue gas must be free of particulates
and cooled to approximately 53°C (127°F). This can be accomplished in one
step by passing the flue gas through a venturi scrubber using circulating
water, from which a purge stream is sent to a wastewater treatment unit. A
two-step approach to preconditioning is to send the gas through an elec-
trostatic precipitator (ESP) and then through a cooling-humidifying chamber.
Two absorber designs currently have been considered for use in MgO
systems. One is the turbulent contact absorber (TCA), which was initially
adapted for use With MgO in the United States and has been used for S02
4.2-135
-------
removal at a Japanese smelter.*" The other absorber is the ventri-rod
scrubber, used in the United States by United Engineers and Contractors *"
Figures 4.2-31*- and 4.2-32*" are schematic representations of TCA and
ventri-rod scrubber installations.
In two demonstration units, venturi absorbers were utilized; however
the constructor (Chemico) believes that these venturi absorbers will not be
used on any other utility systems.
An improperly designed absorber is subject to corrosion and erosion
caused by the liquid flow, by action of the absorbent itself, and by
possible acidic conditions. To reduce costs, absorbers may be constructed
of mild carbon steel with an inner protective coating of fiberglass-
reinforced polyester (FRP), polyurethane, rubber, or flaked glass.*" The
uncoated sections (internals) of the absorber that have direct contact with
the flue gas and absorbent may be made of low-carbon stainless steel high-
nickel alloys, FRP, or other highly corrosion-resistant materials.
The liquid-to-gas (L/G) ratio in MgO systems is in the range of 4 0 to
5.3 liters/,^ (30 to 40 gal/10* ft'), about half of that used in calcium-
based "throwaway" processes.*" Other advantages are the high solubility of
magnesium sulfite, the controllability of slurry composition, the high
concentration of crystallization nuclei (for regeneration stream feed), and
a short slurry residence time.250
Size of an absorber module and the number of modules per system are
directly related to the turndown requirement (reduction of throughput)
system availability, and gas-liquid distribution. As the boiler load
fluctuates, the scrubbing rate should change to maintain optimum perform-
ance. One method of accomplishing turndown is to shut down scrubber modules
as the load decreases-the more modules in the system, the smoother the
transition. Scheduled cleaning and maintenance of absorber modules not in
use' reduce overall absorber downtime. Operation of multiple absorber
modules also permits the use of modules having smaller cross sections, which
may promote uniform gas-liquid distribution and improve efficiency. Capac-
ities of absorber modules normally range from 25 to 165 MW.251
The mist eliminator, located after the absorber, removes entrained
water droplets and slurry from the gas stream. Thus it reduces water loss
4.2-136
-------
FLUE GAS TO REHEATER
FLUE GAS FROM
ELECTROSTATIC
PRECIPITATOR
' MAKEUP WATER
MIST ELIMINATOR
SCRUBBER
SLURRY
RECYCLE
TO SCRUBBER
HOLD TANK
Figure 4.2-31. Typical TCA installation/
247
4.2-137
-------
VENTRI ROD
MODULES
TO SCRUBBER
HOLD TANK
FLUE GAS TO
REHEATER
FRESH H20 INLET
MIST ELIMINATORS
ENTRAPMENT
SEPARATOR
M_PJHER_LIQUOR
?P9?pP3 [Qftooqq)
>" , '**'"
£>PS9PP] [opopfjcg [opopop>
TO SCRUBBER
HOLD TANK
STACK •
GAS
INLET
SCRUBBER
SCRUBBING
SLURRY INLET
TO SCRUBBER
HOLD TANK
Figure 4.2-32. Typical ventri-rod scrubber with mist eliminators.
2if 8
4.2-138
-------
in the absorption process, reduces corrosion of downstream equipment, and
may reduce the need for reheat energy. Most mist eliminators are con-
structed of stainless steel or FRP, and are located either in the absorber
shell immediately above the absorption section or in a separate structure
just after the absorber. A mist eliminator is designed to remove the
entrained droplets by impaction as the gas is forced to change directions
while passing through chevron- or z-shaped sections.
The reheater immediately precedes the exhaust stack. The gas emerging
from the S02 scrubber often ranges from 48° to 54°C (120° to 130°F) and is
saturated with water vapor. Reheating of the gas is generally practiced to
prevent water condensation, reduce corrosion, and improve plume dispersion.
Reheating of the flue gas can be accomplished by (1) installing a gas or
low-sulfur oil burner (2) installing steam coils, (3) using a burner or
steam coil to heat ambient air and inject it into the flue gas stream, or
(4) bypassing untreated hot flue gas, which is mixed with the treated flue
gas.252
Usually about 10 percent of the total flow through the absorber
recirculation circuit is removed as a bleed stream that is routed to a
dewatering section and eventually to the MgO regenerating section. The
dewatering equipment is usually a train consisting of a clarifier/thickener,
centrifuge, and rotary dryer. Both thickeners and hydroclones (liquid
cyclones) can be used as first-stage dewatering units. Although such
first-stage treatment may not be required, it improves the centrifuge opera-
tion by providing a more consistent feed. The underflow from a typical
stainless steel, solid-bowl centrifuge is a wet cake, as discussed under the
Process Chemistry Section.253 The overflow from both the first-stage
dewatering and the centrifuge is usually recycled directly to the main
slurry recirculation loop. Another possibility is to use this water for
slaking the regenerated and makeup MgO before it is returned to the
absorber. The underflow of the centrifuge is usually transported by screw
conveyor to the dryer.
Most rotary dryers are of the counterflow, incline type. The feed to
be dried is introduced at the top and travels downward through the cylin-
drical, revolving dryer, where it meets the upward moving, hot combustion
4.2-139
-------
gases (175° to 230°C, 345° to 450°F). The free and chemically bound water
is removed, and the dried crystals of magnesium sulfite and sulfate are then
conveyed to storage hoppers. The dryer can be one of the most troublesome
areas for continuous operation; problems with dust may also occur.
The composition and temperature of the scrubbing solution are important
in determining which of the magnesium sulfite hydrates (hexahydrate or tri-
hydrate) is stable at equilibrium and will eventually be produced as the
solid product. For a given solution composition, the hexahydrate is the
stable hydrate at temperatures below the transition temperature. Above this
temperature, the trihydrate is the stable form. Keeping the temperature of
the absorbent medium below this transition temperature promotes maximum
growth of the hexahydrate. In pure solutions containing only water and
magnesTum sulfite, the transition temperature is 41°C (106°F)- however the
precipitation of hexahydrate crystals occurs at temperatures higher 'than
41 C (106°F) in some magnesium oxide scrubbing systems.254
The regeneration section consists mainly of a calciner, a particulate
collection device, and silos for storage of the regenerated MgO. The cal-
ciner can be similar to the rotary dryer or it can be of the fluid bed type.
Selection of an optimum temperature in a fluid bed regenerator repre-
sents a compromise. On the one hand high temperatures are needed to reduce
the magnesium sulfate to the sulfite form and to ensure the substantially
complete decomposition of the sulfate to MgO and S02. On the other hand
excess,ve temperatures will produce dead burned MgO, which is chemically
unreactive and thus ineffective for further S02 removal. As noted in the
discussion of process chemistry, the optimum temperature range for calciner
operatTon is from 670° to 725°C (1250° to 1340°F).237,239j24o Fluid bed
reactors are particularly well suited to precise temperature' control They
also allow for precise control of oxygen, which eliminates the necessity of
adding oxygen scavengers, such as carbon, to the bed to decompose the mag-
nesium sulfate.sss Essex Chemical at Newark, New Jersey, uses a fluid bed
calciner for regeneration of magnesium sulfite from Philadelphia Electric's
Eddystone Station Unit No. 1A.
The temperature range is the same as for a rotary calciner. Because
control of oxygen cannot be achieved in a rotary calciner, coke normally is
added to provide a reducing atmosphere; the total addition of carbon is
4.2-140
-------
about 1.5 to 2.0 percent of the feed.256 Figures 4.2-33 and 4.2-34 show
schematic layouts of a venturi scrubber and a rotary calciner (MgO absorp-
tion and regeneration system) such as were used at the first two full-scale,
demonstration units.257 Figure 4.2-35 shows a TCA and fluid bed calciner
MgO absorption/regeneration system.
The two earliest installations used a rotary kiln to regenerate the
MgO. High dust losses in the rotary kiln necessitate the use of a: hot
cyclone and venturi scrubber to recover the MgO. In a fluid bed reactor,
most of the MgO formed goes overhead with the S02 and combustion gases, and
separation equipment is required.258
From the calciner, the regenerated absorbent (MgO) is cooled and stored
in silos. Cooling can be accomplished by transporting it pneumatically. An
advantage of the MgO system is that the absorbency, potential afforded by MgO
in a given amount of storage space is roughly double that of a calcium-based
absorbent. This effective increase in storage capacity may reduce the
effect of an extended outage of the regenerating facilities and possibly
improve the availability of the total unit.
Several sulfur byproducts are possible with an MgO regeneration system,
the most prominant being sulfuric acid. Other possible products are liquid
S02, liquid S03, and elemental sulfur. Before the gas enters the sulfur
byproduct facility, the particulate MgO is removed and stored. The gas is
then cooled, usually in a venturi or spray tower. The S02 concentration of
the gas generally ranges from 8 to 10 percent and the temperature is about
38°C (100°F).259
The slurry preparation section links the regeneration circuit to the
absorption loop. The bulk (about 95 percent) of the MgO fed to the absorber
is regenerated material, the remainder being fresh feed.260 Heat is some-
times needed to increase the slaking rate of the regenerated MgO.261,262
Also, maintaining the pH of the main scrubbing slurry stream between 6.8 and
7.5 can increase its absorption capability.263 The pH is adjusted by con-
trolling the amount of regenerated slurry to be mixed with the main absorb-
ent stream. Mixing is done either in the recirculation tank, which is often
located beneath the absorber, or within the absorber.
4.2-141
-------
OJ
O)
o
cu
O)
oo
CSJ
4.2-142
-------
OL
O
O
cn
O _J
(_ Q.
> Q
3 <_)
*f
CSJ
00
Lj
5Cte
L_^nt
CD
UJ
C£
O
1
*— 1
00
•s.
_l
DL.
O
o.
4.2-143
-------
CT]
O
O
OJ
O)
cr>
OJ
O)
E
O)
to
O)
CD
•i—
Lu
4.2-144
-------
Development and current status—The feasibility of full-scale MgO
scrubbing systems has been demonstrated in Japanese and U.S. FGD systems.
Three retrofit units in the United States in the 95- to 150-MW range have
demonstrated 90 percent S02 removal during individual S02 removal test runs
on both coal- and oil-fired boilers. These and planned MgO units are listed
in Table 4.2-33.264,265 All the MgO systems installed to date have been
retrofits.
Scaling or plugging in the scrubber has not been a problem in magnesia
slurry scrubbing systems. Venturi or venturi-type scrubbers (single-stage,
double-stage, or ventri-rod types) have been used on all U.S. plants to-
date, and S02 removal efficiencies of over 90 percent during individual S02
removal test runs have been obtained.266
Some common problems to the U.S. units upon startup were exemplified at
Boston Edison. Most of the problems were related to materials handling and
resulted from characteristics of the solids formed in the scrubbing loop.
The production of small (10 to 15 urn) magnesium sulfite trihydrate crystals
rather than the larger hexahydrate crystals caused the centrifuge cake to
retain excessive amounts of unbound moisture. The operational chemistry of
the absorber is the controlling factor in crystal growth. The magnesium
sulfite trihydrate crystals are flat platelets, which form structures that
can trap unbound moisture. Extraction of this "trapped" water with a
centrifuge is difficult. The wet cake can readily cause buildup in the
rotary dryer or in solids handling equipment, with associated plugging prob-
lems.
The problems at Boston Edison were solved by several operating and
design modifications, which included changing the dryer to function as a
granulator and adding hammers to loosen materials adhering to the dryer
shell. The granulator discharge was screened and sent through lump breakers
to crush the oversize agglomerated granules of magnesium sulfite. The dryer
off-gas was sent to the S02 absorber to prevent high dust Tosses.
Other problems in MgO units have occurred in the calcining system.
Formation of the very fine trihydrate crystals in an oil-fired power plant
application resulted in dusting problems in the rotary calciner. Operators
of the Essex Chemical facility at Rumford, Rhode Island, eliminated the
dusting in the calciner by use of a cyclone followed by a venturi scrubber
to remove all the MgO fines from the gas. Leakage of air into the calciner
4.2-145
-------
•o
01
•o
01
i-
ro
t-
01
O.
o
4->
in
o>
ro ro
t- c
O) O
4-J Ol
in Q.
Ol O
CM 4J
P-- c
en 01
Ol Ol
o a.
t. o
o.
00
CO-Q
CQ «
_J |^J
c; oo
o t-^
to cr>
co uj
ui CQ
'
co uj
UJ
01
.a ..
~-m
Ol
ID
•o
Ol
c
O)
TD
Ol
4-1
U
01
••—>
o
o.
o.
s-
ro
_fll
O
Q.
LU
: u_
: o
_i co
o-ct
o to
co
: a.
i
cz
cc
CM O
UJ
CQ
«=c
4-1
i to
i
ij
W Ol
0 >
to o&«
U
•— C
tO T-
O
en
i=
01
Q.
Ol
to
c
Q.
J.
ro
to
in
en
4-1
C
Ol
^
H- O
o oo
Ol "~
4-1 01
4J C
Ol 3
. „
a
u.
D)
C
•r- O
i- OO
ai en
tn 01
c c
O 3
CJ -3
..
g
U.
O)
s- eo
01 en
•o r—
O 3
CM
in
csj
m
CM
-f 3
t *— »*-
f* OJ ^~
f|' 3 3
> LU UI
!;
e
t!
(I
i
1
1
tn
tn t-
01 o
0 -o
o c
a. >
•u
t*- *p-
o c.
.2 °
to LL.
^
g; r^
•t- Ol
51
CD
0
U
OJ
^c
C_)
0
in
(Z
o to
vn
•a o'
LU X
55
in in
ro
o
c_>
o
u
f
u
en
u
*^"
0 •
01 (-
LU 2
O
u a.
B T3
4-1 ro
CO
o
z
c
o
tn
t_
ai
-K:
u
o
to
ro
ro
-CM
o' .
m cj o
!.•" «
Ol 4-> O
•a o 4-1
ro ai in
f UJ '
ro
o
in
£-
01
QJ
c:
T3 -r-
01 0)
4-1 C
•r- LU
C
O
in
w
d
ro O
!.y
^- u
OJ 4->
•a u >i
m
-------
was a problem because the reduction of magnesium sulfate requires a reducing
atmosphere. Installation of new seals on the rotary calciner corrected the
problem.267
Other problems have been erosion and corrosion in the carbon steel
recirculating slurry piping. The use of rubber-lined pumps, valves, and
piping is considered a practical solution.261,262 The reelrculation pumps
also have withstood corrosion through the use of 316 stainless steel
impellers.268
Notwithstanding these problems, some positive observations are also
noted. The chemical and mechanical performance of the scrubber was better
than expected at Boston Edison. In the absorber, no internal plugging
occurred and the polyester lining was in good condition after 2 years of
intermittent operation.
The major problems encountered at Potomac Electric Power's Dickerson
Station were related to materials handling. The major equipment items
(scrubber vessel, thickeners, centrifuge, and dryer) all performed well.
Major problems were encountered, however, with construction materials for
handling systems. Carbon steel pipe and slurry pumps were found to be
inadequate against the corrosive/erosive properties of the slurry. Indi-
vidual pump suction and discharge lines were necessary since leaks in pipes
at or after a header could make installed spare pumps useless. The designs
of the centrifuge discharge hopper, weigh belt feeders, and the dry mag-
nesium sulfite bucket elevator were improved.269
Proper feeding of magnesium oxide to the slurry system became a prob-
lem, because of plugging in the mix tank and suction lines to the feed
pumps. Later, there were difficulties with the proper slaking of regener-
ated MgO. The final solution to both problems was the installation of a
steam-sparged, agitated premix tank/slaker to promote dissolution.269
Various rods, bellows, hangers, etc., corroded in the reaction vessel.
It was found that the wrong materials had been used,.particularly 304S.S.
instead of the specified 316S.S. The proper materials were used to replace
the corroded parts and this problem was solved. Corrosion and erosion were
severe in the recirculating piping of both first- and second-stage scrubber
slurry. Fiberglass reinforced polyester (FRP) and epoxy were used to make
repairs. For long-term commercial use, however, rubber-lined piping should
be specified.269
4.2-147
-------
When the calciner was used, it was discovered that magnesium sulfite
from Dickerson Station was predominantly hexahydrate, whereas that obtained
from Boston Edison's Mystic Station was mostly trihydrate. This necessi-
tated additional testing, to determine the proper operating parameters
(temperature and feed rate) for calciner operation.269
The bucket elevator conveying the dried magnesium sulfite to the
storage silo tended to overload and trip. The problem was traced to the
discharge chute from the centrifuge, where wet magnesium sulfite cake tended
to hang up and then break off in large chunks. Modifications to the dis-
charge of the centrifuge outlet hopper and installation of larger buckets
for the sulfite elevator helped overcome this.269
Problems affecting the centrifuge included buildup of material or wear
inside the centrifuge, and changes in the physical/chemical form of the
magnesium oxide. Use of washout ports and hardened steel surfacing helped
remedy this.269
The magnesia scrubbing process has been used on a commercial scale at
three locations in Japan, as summarized in Table 4.2-34. These units have
demonstrated S02 recovery of over 90 percent. In the Japanese operations
the large hexahydrate crystals were obtained at both the Onahama and Mitsui
installations. There were no problems in the filtration and drying steps,
as occurred at Boston Edison. The oxidation of sulfite to sulfate in the
absorber at Mitsui was only 7 to 10 percent, or half as much as was reported
at Boston Edison.
Control costs—The capital costs of the existing systems have varied
greatly because of the different types of equipment used and the quantities
of flue gas handled at each installation. One report lists the costs of the
three units in the United States. With all prices escalated to July 1979,
the capital costs, operating and maintenance costs (.both total and per
kilowatt-hour), and the FGD system capacities are shown in Table 4.2-35.
4.2-148
-------
CO
r-~-
o-i
ca u_
co o
Di OO
O eC
CO
CO
CO 3
C .O
m
^3 I-
._ •». .^
3
(O *r- *O
a. t-
3 ai
4J C
I- OJ "C
) C. E
O CL
u a.
••- ••-
tr o
3 .c
i
tn C
ss.
to to
00
•p- -i- *O .
4.2-'149
-------
TABLE 4.2-35. COSTS ASSOCIATED WITH THREE MAGNESIUM OXIDE UNITS271
Boston
Edison
(Mystic)
Potomac
Electric
(Dickerson)
Philadelphia
Electric
(Eddystone)
Capital costs,
$/kW
55
109
111
Operations and
maintenance cost,
Total annual,
mills/kWh
4.6C
6.3U
7.5C
FGD system
capacity,
MW
150
95
120
. Estimation.
Based on questionnaire.
Others have taken a more general approach to costing MgO systems. One
report lists a variety of costs for different conditions of input to the
absorption unites Table 4.2-36 is a reduced version of these costs,
showing 3.5 percent sulfur coal as the fuel and assuming 90 percent S02
removal efficiency. Figures 4.2-36 and 4.2-37 depict these costs and
similar ones for 7.0 percent sulfur coal at the same S02 removal efficiency.
The total annual operating costs include both operation and maintenance
costs and fixed charges.
Another source quotes $75.8 million as the capital costs of an MgO
system for a coal-fired boiler burning 3.5 percent sulfur coal; it is
assumed that regeneration will be done off site. This source also lists the
total annual cost as $10.4 million, which includes transport of the mag-
nesium sulfite to a regeneration site and fixed costs of 14.9 percent of
total investment. 272 (These CQsts ape a]so esca]ated to Ju]y ^ 19?g }
Capacity and energy penalties are also considered in comparing costs of
FGD systems. A capacity penalty is calculated by comparing the total
electrical requirements of the FGD system with the total original output of
electricity from the power plant, before the FGD was on-line. This number
is usually expressed as a percentage, as it is in Table 4.2-37.
4.2-150
-------
OO
>-
oo
Q_ CO
C£. >-
O C_)
oo ^z.
CQ LU
<: —i
o
O •-!
en Li-
o o
OO LU
1— OL
OO
o &s
t_) O
01
vo i—
CM
«*
CQ
oo
1—
o
o
JD
)
CO
•t— CD
Q-r—
ro
O -fcO-
1 CO ---.
CO 3 to
-(-> C r—
O C r-
r— CO -r-
E
3
•a -\
O> to
•r- t—
1 t tp—
E
-o
C OJ
ca o
C 3
Cn ro -ii
c: C ~-^
•i— a> to
CO C i—
S_ -I- -r-
CU CO E
0- E
O
rt
rcc 3
•r- ^1
O--t^-
. CO
O
S- +J
O) •!-
i — O
•r— CO
O CD-
CO CO
o
o
CM
CM
0
LD
•
f—
00
LT>
r-^
1 —
en
.
OO
r—
0
co
oo
CD
^J-
3
!>"
tn
CM
0
tn
en
CM
oo
o
oo
•
CM
CM
un
^j-
r—
LO
•
CO
O
LO
01
CM
00
3
^>~
O
O
1 —
o
en
oo
oo
«^j-
LO
en
•
<^-
i —
CM
OO
en
oo
•
LO
LO
en
UD
r—
CM
3
:>
O
O
CM
CD
P—
CM
r—
oo
LO
•
r—
r—
oo
en
IJD
r--
LO
•
^d"
CM
"=1"
CM
UD
i —
3
2:
CD
0
LO
0
co
CD
CD
O
1
LO
CM
O)
O
d
CD
S_
O)
a>
C£ .
£Z
6
s-
M-
-r-j
cu
CD-
CO
-a
-------
Ol
(J
o
CD
O
O
O
LO
o
CQ
CD
un
o
o
o
o
o
o
CM
J
o
CD
C\J
CD
O
tn
CM
03
>
O
I
CM
O
CO
O
CTi
Ol
>
•r-
cn
to
O)
4->
to
Q
O
O
cn
O
to
o
o
Q.
(O
CO
I
CM
-------
o
,
•4-J O
C/5 C
O O)
O •!—
O
CD-r-
4-5 OJ
(O
'(O
Q. >
O O
i— O>
its i-
3
E 00
SZ O
fO I/)
(O
4->
O
oo
i
CM
CL)
S-
CD
MM>I/SILLUI 'sisoo IVHNNV ivioi
4.2-153
-------
I/)
UJ
en en oi en en rororocMCM
CU
c
CU
ro ro co ro t
LU
ro ro ro ro ro
u
» o ro T—(<
p
(D
ro CM CM CM CM
CM CSJ CM CM
•a- CM o co o
a.
OJ
CU
O
O)
to
cc:
uu
r-. CD rot-ti
CM CM CM CM
ro CM CM CM CM
o
ro >—i ID r- o
GJ
Q.
CU
CU
) CM CM CM CM CO CM CM CM CM
-- J3 o o o moooo
CM O CD O CD CM O O O O
•—t CM LO O r-ICMinO
CO
I
CM
>£
CU
4.2-154
-------
Energy penalties are related closely to capacity penalties. Both are
calculated by finding the total electrical usage of the FGD system. The
energy penalties, however, also include any other type of energy that the
FGD system might need to operate. Other energy users could be a flue gas
reheater, a slaking tank, or any steam sparging unit. Table 4.2-37 also
lists energy penalties as percentages of the gross electrical generation.
The values listed are for purchasing the additional electrical power needed
rather than building to provide extra capacity.
4.2.3.9 Adsorption--
The phenomenon of adsorption can be used in collecting S02 from flue
gas; when the molecules of S02 are brought into contact with a solid adsorb-
ent, they adhere to the surface of the adsorbent material. Among several
materials that have been investigated for use as S02 adsorbents, activated
carbon has been investigated most intensively.
The Reinluft process, developed in Germany in the late 1950's, was the
first major carbon adsorption system. Several test units were operated
between 1959 and 1968.274 Bergbau Forschung later conducted pilot plant
studies at the Welheim power plant in West Germany,275 and together with
Foster Wheeler offer a dry adsorption system utilizing activated coke (char)
as the adsorbent.276 This system is known as the Bergbau Forschung/Foster
Wheeler (BF/FW) process. In the mid-1970's, demonstration plants were
installed at the Kellerman power plant in Lunen, West Germany, and at the
Scholz steam plant of the Gulf Power Company in Sneads, Florida. In Japan,
the Kansai Electric Power Company and Sumimoto Shipbuilding and Machinery
Company jointly developed a similar carbon adsorption system. Following
pilot plant tests in 1969, they constructed a demonstration unit at the
Sakai Port power station of Kansai Electric. Although this unit has been
operating since 1972, very little operating information is available.277,278
The discussion that follows therefore deals chiefly with the BF/FW process.
The demonstration plant in Lunen treats a portion of the flue gas from
a 350-MW boiler (150,000 Nm3/h, or 88,275 scfm--about a 35-MW equivalent).
The Scholz demonstration unit treated 40.4 ms/s (85,600 acfm) of flue gas,
roughly equivalent to 20 MW.279 The process includes regeneration of the
adsorbent for reuse; the regeneration step produces a concentrated stream of
S02, which is reduced to elemental sulfur. The Lunen plant utilized a
4.2-155
-------
modified Claus unit to reduce the S02 to elemental sulfur, and the Scholz
plant demonstrated Foster Wheeler's proprietary RESOX reduction process.
Pilot plant tests have shown that the BF/FW process can remove 97
percent of the S02 from flue gas streams and may also remove some parti cu-
late, N0x> and hydrocarbons. 28° At inlet loadings ranging from 0.49 to 3.20
g/Nm3 (0.2 to 1.3 gr/scf), particulate removal has been reported to range
from 93 to 96 percent. "i It is suspected that char abrasion contributes
particulate to the flue gas at the Scholz plant.282 At any rate> particu_
late removal in the adsorption section of the system could compromise S02
removal efficiency. Foster Wheeler reports that N0x removal efficiency for
the BF/FW process ranges from 40 to 60^3 percent. Tests at the Scholz
plant indicated that N0x removal ranges from 17 to 50 percent with an
average of about 20 percent.
Process description-The BF/FW system consists of adsorption, regenera-
tion, and reduction steps, occurring as follows:284
Adsorption (1) S02 + 1/2 02 -» S03
(2) S03 + H20 -> H2S04
(3) H2S04 -> S03 + H20
(4) 2S03 + C -* C02 + 2S02
Regeneration
Reduction
(5) S02 + C -> C02 + S
In the first two reactions, S02 collects on the surface of the adsorb-
ent, is oxidized, and is transported to the inner pores as sulfuric acid,
allowing more S02 to be adsorbed on the surface. Eventually the adsorption
ceases and the adsorbent must be regenerated. In the regeneration steps,
the adsorbent carbon is heated to about 650°C (1200°F) in an inert atmo-
sphere. The adsorption process is thereby reversed; S03 is released and is
reduced to S02. In the reduction phase the concentrated stream of S02 is
reduced to elemental sulfur in a RESOX unit. Elemental sulfur and ash are
the only byproducts of the system.285,286
Figure 4.2-38 is a simplified process flow diagram of the BF/FW sys-
tem.287 After passing through particulate collection equipment, boiler flue
gas enters the adsorber(s). The gas passes horizontally through a vertical
column of adsorbent in a crossflow, while the activated char (adsorbent)
4.2-156
-------
CO
CO
OJ
o
o
D.
Q.
S-
o
CO
T3
fO
oo
CO
I
OJ
OJ
CD
4.2-! 57
-------
moves downward in plug flow. Vibratory feeders control the rate of char
flow, which depends upon the quantity of S02 in the flue gas. Gas passing
through the lower half of the adsorber may be routed to a second-stage
adsorber unit. The cleaned flue gas passes through a fan at the adsorber
discharge before it is exhausted through the stack.288,289
The spent char, in the form of small pellets, is conveyed to the
regenerator, where the temperature of the char is raised to 650°C (1200°F)
by mixing it with hot sand at 815°C (1500°F). The sand and char are then
separated by a vibrating screen deck. The char is spray-cooled to about
120°C (250°F) and returned to the top of the adsorber by bucket elevators.
Makeup char is added to replace the char that is lost in the regenerator. A
fluidized-bed sand heater reheats the sand to 815°C (1500°F). Bleed-off of
the sand removes ash and small char particles.290-292
The regenerator off-gas (25 to 40 percent S02 by weight) enters the
RESOX process in counterflow to a mass flow of crushed anthracite coal. The
S02 is reduced, and molten elemental sulfur is recovered in an inclined
shell-and-tube condenser. The sulfur is stored in liquid form in an
insulated tank.293
The operating problems with the demonstration units at Scholz and LUnen
are primarily mechanical rather than chemical as in other types of FGD
systems.294>29s At Scholz, poor char distribution in the adsorber caused
imbalances in the bed level. This imbalance reduced S02 removal efficiency
and created hot spots in the adsorber that hampered cooling of the char.
Char consumption was about 5 times the expected level .295,296 In additl-on>
improper design of the char/sand separator and hot-sand bucket elevator
caused frequent outages,2" and the sulfur condensor of the RESOX system was
susceptible to plugging.295
^Operation of the process at LUnen has been more successful following
initial problems with fans, dampers, and bucket elevators.298,299
Energy and environmental impacts-Ait.hnngh the BF/FW process does not
include reheating of the exhaust gas, the energy requirements may be as high
as 10 percent of the power plant output.293 Tab1e 4.2-38 lists the energy
requirements for a BF/FW process that cleans the flue gas from a 500-MW
power plant firing 3.5 percent sulfur coal.300
4.2-158
-------
13
o
•n
| —
Z
^
Q.
s
I
0
o
LO
eC
Di
O
LL-
LU
I—
>-
3:
LU-
LU
03
LU
I—
LL.
Q
fe
1
1
I— (
cx
LU
Cti
C3
O£
LU
z
LU
00
CO
1
CM
•3-
LU
_J
03
> "2
py ^^
•r-
U
ll
4^
U
QJ
LU
^"N /^^N
Q) ^
."|f ^
U -P
re CQ
i-
•P 0
C i—
re >-/
^— '
^
i— \
re o
0 0
0
/^^
(/) /*^
— » f—
_j _i_
o ^.
•1— -p
E 03
3
•r- O
>^ >— '
: i — -C
re "^
o o
C_3 CIJ
1
1
O
j3
o
Ol
in
ui
in
OJ
U
O
Q.
<*• CO i—
^ 1
ID ' — O
^^
O
c^
*~
^^
\
00
ID
O
CM
•^^
LD
O
CM
CM
c
O
c •«-
O 4-> C
•r- re o
D. 0) -P
s- c u
o a) ^
(/) O) "D
•o o> -
-------
Since the BF/FW process is a, regenerate system, the environmental
effects are minimal in comparison with those generated by throwaway FGD
systems. Sulfur, dry fly ash, small amounts of char fines, sand, and
coollng water are the only effluents.^* Gases from the sulfur condenser
and sand heater can contain various sulfur compounds and may not be suitable
for release. These gases can be recycled to the adsorber. 3°2
Cost-The only applications of the BF/FW system to date are the demon-
stration units. Thus, the quantity of cost- information is extremely limited
and its accuracy is not known. Table 4.2-39 presents . capital, and operation
costs for a 500-MW unit.303_3os Data from ^ ^^ ^.^ .^.^ ^
operating costs may be higher than predicted because of the high rates of
char consumption in the regenerator and because mechanical maintenance costs
are approximately 50 percent greater than those incurred with other proto-
type systems tested.306
4.2.3.10 Dry Removal Processes—
The term "dry removal process" designates any FGD process from which a
dry product directly results. There are currently three major types of dry
FGD systems being developed today: spray drying, dry injection, and combus-
tion of fuel/limestone mixtures. Of these three systems, spray drying is
currently the only one being developed on a commercial scale 307 Table
4.2-40 summarizes the primary features of these three main types of dry FGD
systems.
Dry removal offers various advantages over wet scrubbing. Dry removal
systems do not require the sludge handling equipment that many wet scrubbers
need. Scaling and plugging, common problems at the wet/dry . interface in wet
scrubbers, are avoided in dry removal units because only a dry product
contacts the walls. Whereas wet systems often use special materials of con-
structs or coatings to prevent corrosion and erosion, the vessels and duct
work of dry systems can be made of low-carbon steel. Dry removal units
require less manpower to operate than wet scrubbers and can respond more
quipkly to fluctuations in S02 levels. Because dry systems operate with
restively low pressure drops through the absorption system and with smaller
volumes of spent absorbent, the operating expenses that wet systems incur
because of high pressure drops and greater volumes of spent absorbent can be
4.2-160
-------
TABLE 4.2-39. CAPITAL AND OPERATING COSTS OF A
500-MW BF/FW SYSTEM (MID-1979 BASIS)
Coal sulfur
content, %
or plant
0.9303
4.3303
3.530U
Adsorption
Regeneration
Reduction
Total
LUnen, West Germany305
49 MW
Sholz, 20 MW305
Low-sulfur coal
High-sulfur coal
Capital cost,
1000 $
16,410
32,825
3,016
616-862
1,724-1,847
$/kW
32.82
65.65
61.55
30.80-43.10
86.20-92.35
Operating cost,
$/yr
l,438,000a
4,673,000a
829,000
5,366,000
1,055,000
7,250,OOOb'C
l,219,000d
a Based upon 60 percent capacity factor, 1.05 x 107 J/kWh (10,000 Btu/kWh)
heat rate, and coal with a heating value of 2.78 x 10? J/Kg (12,000 Btu/lb).
b Based upon 60 percent capacity factor and 90 percent sulfur removal.
c Raw material and utility costs only.
d Includes fixed costs.
4.2-161
-------
TABLE 4.2-40. SUMMARY OF KEY FEATURES OF DRY FGD SYSTEMS
" Preeetl type
design features
tanget »f leegent
UUIUatten
laneet of SO, removal
Paniculate removal
$oel prtblemi or
a4»antagei
Spray drying/
paniculate collection
Employs • spray dryer equipped with atomizers
to spray sorbenl solution or slurry into Incomi
SO, laden flue gas. The spray dryer is coupled
with i btglnuse (or possibly ESP) to provide
collection of fly ash and entr.ined product
tolIds,
Sodium carbonate. liM. trona (NaCOj. KaHCOj.
ZH,0). ind limestone hive all been tested
Planned commercial systems will use sodium car-
bonate or I(K
80 to IOCS for sodium-based alkalis. 30 to
MX for UK on a "once-through" basis. 80 to
SOI for KM with partial recycle of product
10 ds. 20X or less for limestone. Reagent
utllilatlon is a itrong function of the outlet
Umperature of the gas; utilization Increases
•l the dryer outlet temperature of the gas
approaches it! adlabatic saturation temperature.
l!% '".I" ?»:1000 to 2000 pp.) 80 to 90X for
iodlum-based alkalis, 45 to (OX for UK on a
once-through" basil. 80 to 8SX for UK with
SJT'J'' recycle of product solids. Less than
30X for IlKStone.*
loth baghouse! and ESP'i have consistently
achieved S9«X removal of entrained product
lolldl and fly ash. laghouses have the
advantage of providing for additional
SO, reKival across the.fllUr cake tbat
collect! on the fabric surface. However '
•OK reports clala H Is possible to man
closely approach the adlabatic saturation
temperature of the gas with an ESP down-
itream of the spray dryer.
Spray drying results In a dry easy to handle
waste product. When sodium alkalis
are used, however, the products are quite water
•oluble and create disposal problems. Water and
energy requireKnts are less than for conven-
tional "wet" IlK/llKltone systems. High-
sulfur coal applications may be limited, but
are being Investigated further
Spray drying It currantly the only commercially
applied dry FGO technology; three utility
iylUu (400-500 IV each) arm being constructed
(startup In Wei. K. and §3), and two indus-
trial tystmms were scheduled to start up In
late 1979, Several other companies are
conducting extensive UE programs toward a
comKrclal systmm.
Dry injection/
paniculate collection
Pneumatic injection of dry alkali sorbent
Into a flue gii stream with subsequent par
ticulate collection Injection point
varies from imKdiately after the boiler
to Just upstream of the collection device
(Baghouse or ESP). A baghouse Is usually
employed as considerable S02 removal occur
across the filter cake collected in the baa
surface.
Sodium-based alkalis:, sodium carbonate.
sodium bicarbonate, trona, and nahcollte
(60-70X HaHCOj). UK and IlKStone
lave been investigated, but both require
600»°F flue gas for significant SO.
-emoval.
Baghouse systems) 40 to 60S for
nahcoltte, sodium bicarbonate at
lighest SO, removal conditions.
'OX or less for liKstone even at high
lue gas temperature. Utilization
ncreases at higher gas temperatures and
s a function of sorbent feeding Kthod.
0 to 90X for sodium-based alkali systems
depending on stolchioKtrlc ratio, flue gas
emperature, and Kthod of feeding. 90*
emovals have been achieved with nahcollte
t 290°F temperatures. 20 to 30X for 1 IK-
tone at high temperatures (600»°F).
ry injection systems with baghouses remove
9«X of entrained product solids and fly
sh. ESP's demonstrate 99»X removal also,
ut SO, removal is much lower than in bag-
reuses. Also the increased inlet-grain
oldIng will affect ESP sizing.
he dry product resulting Is very water
oluble, and teachability and stability
roblems are likely to occur in dts-
osing the waste solids. The use of
elatively inexpensive reagent (nahcollte)
nd minimal equtpKnt requireKnts make
ry Injection economically attractive,
* two major drawbacks are the availa-
'lity of nahcollte in amounts required
r comKrcial applications and the waste
sposal problem,
though dry Injection has be*n shown to be
clinically feasible. comKrcial application
at a standstill because of uncertain!
rbent availability.
Combustion of coal/
limestone fuel mix
The most promising technologies In
this area appear to be 1) combustion
of a coal/limestone pellet In a
spreader stoker boiler, and 2) com-
bustion of a coal/limestone fuel mix-
ture in a low-HO burner. The lower
adlabatic flmme temperature resulting
from the two-stage combustion schmme
employed In both technologies appears
to Increase the available IlKStone
reactivity.
Urn
stone (Pellet alto requires
type of binder).
Ca:S ratios of 7:1 have been used
In coal/llKstone pellets while
a 3:1 ratio was used for the low-
NO burner tests.
:oal/HK!tona pellets captured
75 to MX of the available sulfur
'n the fuel. Preliminary results
n telts with low NO burners In-
dicate that 80X retention Is
achievable.
ottHjstlon of the coal/1 iKstone
fuel mixture will result in increased
paniculate loading.
he additional costs of pre-
paring the coal/liKitone fuel
nd removing greater amounts of
sh are significantly lest than
onventlonal wet scrubbing systmm
osts However, these technologies
ave only been applied on imall-
cale industrial-type boiler system!
.onsfdermblo work remains to
develop the technologies for com-
Krcial scale applications, although
ndustrial comKrcial applications
ook promising. EPA l! currently
unding continued pilot plant
•sting on industrial boiler!, and
nre complete test work on low-HO
urneri has been proposed and i! *
nder review by the EPA.
•^amoral ana reagent utilizations may be lower or higher
>ftmflrature, Umperature drop over the ipray dryer, and
.
c. "oichioKtrlc ratio, flu. gas inlet
4.2-162
-------
reduced. It is estimated that dry removal units need only 25 to 50 percent
of the energy that wet scrubbers require.309 Finally, dry systems consume
much less water than wet systems and thus are particularly attractive in
western areas of the United States where water supplies are limited.
Several disadvantages have been put forth for dry scrubbing systems;
among them are higher priced absorbents-, • applicability • primarily to low
suTfur coal, and that there are no commercially proven systems. It is true
that the absorbents currently in use or planned employ dry lime scrubbing,
with the exception of one regenerate Aqueous Carbonate system, rather than
the less expensive limestone used successfully in many wet scrubbing sys-
tems.310,311 Additionally, other more expensive absorbents, such as sodium
carbonate, have been tested for use.. Most of the dry systems either in
design or under construction are for low sulfur (less than 1 percent sulfur)
coa-,.31^312 however, one industrial unit which is on stream reports 85
percent S02 removal when firing a 3 percent sulfur coal.313 Therefore, dry
systems may well be applicable to higher sulfur coals, but this will be
shown only by more operating experience and research. The first commercial
units, two identical systems, have begun operation,313,314 and the first
utility dry system is" scheduled to begin operation in 1980.315 More utility
systems are scheduled on stream thereafter.
Process descriptions—316Dry systems can be regenerate, such as the
Aqueous Carbonate Process, or nonregenerable, such as the lime systems.
.In spray dryer-based systems, the first of the major types of dry FGD
systems, flue gas at air preheater outlet temperatures [generally 135° to
204°C (275° to 400°F)] is contacted with a solution or slurry of alkaline
material in a vessel of 5 to 10 seconds residence time. The flue gas is
adiabatically humidified to within 28°C (50°F) of its saturation temperature
by the water evaporated from the solution or slurry. As the slurry or
solution is evaporated, liquid phase salts are precipitated and the remain-
ing solids are dried to generally less than one percent free moisture.
These solids, along with fly ash, are entrained in the flue gas and carried
out" of the dryer to a particulate collection device. Reaction between the
alkaline material and flue gas S02 proceeds both during and following the
drying process. The mechanisms of the S02 removal reactions are not well
4.2-163
-------
understood, so it has not been determined whether S02 removal occurs pre-
dominantly in the liquid phase, by absorption into the finely atomized
droplets being dried, or by reaction between gas phase S02 and the slightly
moist spray dried solids. The chemical reactions are identical to those
descnbed under corresponding wet scrubbing processes (e.g., lime, sodium
carbonate, etc.).
Sodium carbonate solutions and lime slurries are common sorbents A
sodium carbonate solution will generally achieve a higher level of S02
removal than a lime slurry at similar conditions of inlet and outlet flue
gas temperatures, S02 level, sorbent stoichiometry, etc. Lime, however, has
become the sorbent of choice in many circumstances because of the cost
advantage it enjoys over sodium carbonate and because the reaction products
are not as water soluble. Through the use of performance enhancing process
modifications, such as sorbent recycle and hot or warm gas bypass lime
sorbent has been demonstrated at the pilot scale to achieve high levels of
removal (85 percent and greater) at sorbent utilization near 100 percent.
Using a spray dryer for a flue gas contactor involves adiabatically
humidifying the flue gas to within some approach to saturation. With set
conditions for inlet flue gas temperature and humidity and for a specified
approach to saturation temperature, the amount of water which can be evapo-
rated into this flue gas is set by heat balance considerations. Liquid to
gas ratios are generally in the range of 0.03 to 0.04 liter/m^ (0 2 to 0 3
gal/1000 ft*). The sorbent stoichiometry is varied by raising or lowering
the concentration of a solution or weight percent solids of a slurry con-
taining this set amount of water. While holding other parameters such as
temperature constant, the obvious way to increase S02 removal is to increase
sorbent stoichiometry. However, as sorbent stoichiometry is increased to
raise the level of S02 removal, two limiting factors are approached:
ss oZfa\n°n decre«es' raisin9 s°rbent and disposal costs
oasis of S02 removed.
oronth
or on the
of the sorbent in the
percent of sorbent solids in a slurry.
There are at least two methods of circumventing these limitations One
method is to initiate sorbent recycle, either from solids which settle in
the spray dryer or from material collected in the particulate collection
4.2-164
-------
device. This has the advantage of increasing the sorbent utilization;
additionally, it can increase the opportunity for utilization of any alka-
linity in the fly ash.
The second method of avoiding the above limitations on S02 removal is
to operate the spray dryer at a lower outlet temperature; that is, a closer
approach to saturation. Operating the spray dryer outlet at a closer
approach to saturation has the effect of both increasing the residence time
of the liquid droplets and increasing the residual moisture level in the
dried solids. As the approach to saturation is narrowed, S02 removal rates
and sorbent utilization generally increase rapidly. Since the mechanisms
for S02 removal do not appear to be well understood, it is not obvious
whether it is the increase in liquid phase (droplet) resident time, the
increase in residual moisture in the solids, or both which account for the
increased removal.
Unfortunately the approach to saturation at.the spray dryer outlet is
set by either the requirement for a margin of safety to avoid condensation
in downstream equipment or restrictions on stack temperatures. The spray
dryer outlet can be operated at temperatures lower than these restrictions
would seem to allow if some warm or hot gas is bypassed around the spray
dryer and used to reheat the dryer outlet. Warm gas (downstream of the
boiler air heater) can be used at no energy penalty, but the amount of
untreated gas involved in reheating begins to limit overall S02 removal
efficiencies. Figure 4.2-39, a general flow diagram of a spray dryer based
system, illustrates these two "reheat" options.
The spray dryer design can be affected by the choice of particulate
collection device. Bag collectors have an inherent advantage in that
unreacted alkalinity in the collected waste on the bag surface can react
with remaining S02 in the flue gas. Some process developers have reported
S02 removal on bag surfaces on the order of 10 percent. A disadvantage of
using a bag collector is that since the fabric is somewhat sensitive to
wetting, a margin above saturation temperature [on the order of 14° to 19°C
(25° to 35°F)] must be maintained for bag protection. Electrostatic precip-
itator (ESP) collectors have not been demonstrated to achieve significant
S02 removal. However, some vendors claim that the ESP is less sensitive to
4.2-165
-------
LU
tO
O)
to
o
o
O)
o
O)
to
to
o
CT)
c\j
a:
LU
O)
CQ
4.2-166
-------
condensation and hence can be operated closer to saturation [less than a
14°C (25°F) approach] with the associated increase in spray dryer perform-
ance.
The choice between sorbent types, use of recycle, use of warm or hot
gas bypass, and types of particulate collection device tends to be site
specific. Vendor and customer preferences, system performance requirements,
and site specific economic factors tend to dictate the system design for
each individual application.
The second major type of dry FGD system, the dry injection process,
generally involves pneumatically introducing a dry, powdery alkaline mate-
rial into a flue gas stream with subsequent particulate collection. A
generalized flow diagram of this process is shown in Figure 4.2-40. The
injection point has been varied from the boiler furnace area all the way to
the flue gas entrance to an ESP or bag collector. Most dry injection
schemes use a sodium-based sorbent. Lime has been tested but has had little
success. Many dry injection programs have used nahcolite as a sorbent.
Nahcolite is a naturally occurring mineral, associated with western oil
shale reserves, and is about 70 percent sodium bicarbonate. Sodium bicar-
bonate appears to be more reactive than sodium carbonate because it loses
both two moles of C02 and one of water in reaction, while sodium carbonate
loses only one mole of C02 in reaction with S02. The following overall
reactions illustrate this point:
2 NaHC03 + S02 •* Na2S03 + 2C02 + H20
Na2C03.+ S02 -> Na2 + C02
Unfortunately, the availability of raw nahcolite in commercial quantities in
the near future is questionable due to the substantial investment necessary
before commercial scale mining can begin. Since the favorable economics of
dry injection are based to some extent on the use of inexpensive sorbents,
the use of commercially refined sodium bicarbonate is prohibitively expen-
sive. Recent research has been aimed at studying the use of raw trona ore,
which is currently mined in large quantities both in the Green River,
Wyoming area and the Owens Lake, California area. The mineral trona con-
tains one mole of sodium carbonate, one mole of sodium bicarbonate and two
waters of hydration (Na2C03-NaHC03-2H20). Trona has the potential for
4.2-167
-------
to
CD
to
O
03
CD
4.2-168
-------
providing a good compromise between reactivity, cost, and availability for
use in dry injection schemes.
An unresolved problem with this technology is disposal of the sodium-
based waste materials in an environmentally acceptable manner. Sodium waste
materials are highly soluble and can result in contamination of aqueous
streams. Disposal of sodium compounds is an area requiring further investi-
gation.
Both baghouse and ESP collection devices have been tested with dry
injection processes. However, the effect of the reaction between unspent
sorbent on collected bag surfaces and S02 remaining in the flue gas favors
the bag collector. Since a major portion of the S02 removal reaction
appears to take place on the bag surface, various methods of feeding have
been tested.
Three types of feeding are:
1)
2)
3)
After the bag is cleaned, sorbent is added to the bag surface only
as it is entrained with the flue gas from a continuous upstream
injection point. This method is called continuous feeding.
After bag cleaning, all sorbent is added to the bag as a precoat
before flue gas flow is resumed. This is considered batch feed
ing.
A compromise between types 1 and 2, after bag cleaning some
sorbent is added initially as a precoat and the remainder is added
continuously through the bag cycle at some upstream injection
point. This method of feeding is called "semi-batch feeding.
Also varied in dry injection programs are sorbent stoichiometry,
sorbent particle size, point and temperature of injection, baghouse air-to-
cloth ratio, and bag cleaning frequency.
The current research on the combustion of a coal/limestone fuel
mixture, the third major dry FGD system option, has taken two forms:
1)
Combustion of a coal/limestone pellet in an industrial spreader-
stoker boiler.
2) Combustion of a pulverized coal/1imestone mixture in
low-NO,
burner system.
Preliminary results of test work on both processes have indicated that
up to 80 percent of the available sulfur in the fuel can be retained by the
4.2-169
-------
limestone. The ratio of calcium to sulfur in the coal/limestone fuel
mixture is important in determining how much sulfur is retained
A spreader-stoker boiler has been used in testing the combustion and
sulfur retention characteristics of the coal/limestone pellet. A Ca:S mole
ratio of 7:1 has been used so far, but further work with a 3:1 Ca:S pellet
is planned. The emissions generated are dependent upon boiler-system
design, coal properties and combustion operating parameters. The inherent
staged combustion of the stoker-fired boiler (accomplished by supplying the
total combustion air as primary air through the grates and secondary air
through over-fire Jets above the bed) results in lower NCD emissions rela-
tive to conventional pulverized coal-fired boilers.
T *W°"Stafled combustl'°n Concept was employed by Babcock & Wilcox
to design an advanced low-NOx burner system. EPA has funded test work
to develop a concept of firing a coal/limestone fuel mixture in B&W low-NO
burners to reduce S02 emissions. Tests conducted on a 12.7 GJ (12 x 10*
Btu)/h scale by the Energy and Environmental Research Corporation (EERC)
with a Utah low sulfur coal have demonstrated 88 percent S02 removal with a
3:1 Ca:S mole ratio. This high S02 removal has been attributed to the lower
fame temperature found in the low-NOx burner which may help maintain lime-
stone reactivity. The EERC has reported that S02 removal increased substan-
tially when the reagent was passed through the pulverizer with the coal
Further research on a larger scale for both systems is needed to deter-
mine the effects of combustion of a coal/limestone fuel mixture on boiler
operation and maintenance. Collection of the increased ash loading and
be s'tudie'd °n °f the Pr°Pertl"eS "" diSP°Sal °f the WaSte Pr°dUCtS must also
Technological dovPlopm^nt-Many studies, which are discussed later
have been made of dry systems that use a fabric filter alone. The technol-
ogy of this process, however, is not advancing rapidly. No full-scale
applications of dry removal with only a fabric filter are currently planned-
owever, EPRI will support a 25-MW demonstration of this concept at the
Cameo Station in Colorado. Perhaps the most promising sorbent for this
process is nahcolite, which is found in tremendous reserves in the Piceance
Creek Basin of Colorado. The three companies that hold leases in the area
4.2-170
-------
are unwilling to mine nahcolite on a commercial scale until they receive
substantial commitments from customers.319
Several factors affect S02 removal by dry systems. For example,
removal efficiency generally increases as flue gas temperature and residence
time increase,320 but this depends to some extent upon the absorbent.
Removal efficiency also improves with a high stoichiometric ratio of sorbent
to S02. Sodium-based dry sorbents display significantly higher reactivity
than calcium- or magnesium-based sorbents.320
The S02 removal efficiency of fabric filter systems has improved since
early tests. Pilot-scale runs with only a sorbent-coated fabric filter at
the Edwardsport Station of Public Service of Indiana from 1967 to 1977
resulted in S02 removal efficiencies from 13 to 72 percent with alkali
utilizations of 22 to 93 percent; high removal efficiencies, however, were
possible only at unacceptably low alkali utilizations.321 The only sorbents
found consistently effective were Na2C03 and NaHC03. In mid-1974 tests at
the Nucla Station of Colorado-Ute Electric Association, the highest S02
removal efficiency with nahcolite as the sorbent was 70 percent with 56
percent alkali utilization when a 0.8 percent sulfur coal was burned.322 In
late 1974 tests at the Hoot Lake Station of Otter Tail Power Company, the
average S02 removal efficiency of 94 percent was observed immediately after
precoating.323 In late 1976 at the Leland Olds Station of Basin Electric,
the S02 removal efficiency with nahcolite was at first 83 percent with 77
percent alkali utilization and later 90 percent with 60 percent alkali
utilization.324 .
Two-stage dry systems can achieve high S02 removal efficiencies, as
shown by tests of the Aqueous Carbonate Process, a regenerable process that
Atomics International has developed, with a solution of sodium carbonate as
the sorbent. All the removal efficiencies observed during 26 laboratory
tests in May 1973 were 90 percent or more. During 40 similar tests in June
1973, the removal efficiencies ranged from 92 to 99 percent, and more recent
tests have yielded comparable results.325
Basin Electric anticipates that the S02 removal efficiencies of its
two-stage dry removal systems will be moderate to high. The spray dryer
using a lime and fabric filter system at Antelope Valley Unit 1 is designed
to operate at 62 percent S02 removal efficiency for lignite containing 0.68
4.2-171
-------
percent sulfur and 78 percent efficiency for lignite containing 1.22 percent
sulfur. 26 The spray dryer and Esp ^ Laram.e ^^^ un^ 3 ^ des_^ ^
operate at 85 percent S02 removal efficiency for 0.54 percent sulfur coal
and 90 percent efficiency for 0.81 percent sulfur coal.327 The 11me feed
rate of the Antelope Valley Unit 1 system will be essentially at a stoich-
iometric ratio of 1 because of the utilization of available alkalinity in
the fly ash. 326 The est1mated annual consumpt-on Qf ^ by ^ ^^
River Unit 3 system indicates that Basin Electric also expects the lime feed
rate for that system to be essentially at a stoichiometric ratio of 1 328
Moderate S02 removal efficiencies are also expected at two lime-based
two-stage systems at industrial facilities. The design removal efficiency
of the spray dryer and fabric filter at the Strathmore Paper Company in
Woronoco, Massachusetts, is 75 percent when coal with a sulfur content of
from 0.75 to 3.0 percent is fired.329 This is a retrofn installation Qn an
industrial boiler with a gas flow capacity equivalent to an 11-MW utility
boiler. * At the Celanese Fibers Company plant in Cumberland, Maryland
the S02 removal efficiency of the planned spray dryer and fabric filter wai
anticipated to be 85 percent for coal containing 1.0 to 2.0 percent sul-
fur; however, compliance tests showed 85 percent removal at 3.0 percent
sulfur coal.313
A 2-year test program of the Atomics International Aqueous Carbonate
Process, a regenerate dry FGD process, is scheduled to begin in 1982 at the
100-MW Huntley Station of Niagra Mohawk in Tonawanda, New York; this process
regenerates spent sodium carbonate and produces elemental sulfur with coal
as the reductant.331
Compilations of the status of dry sorbents and fabric filter filtration
for FGD have recently been published; these compilations address many of the
concerns, problems, and solutions for these systems.332^333
Costs-The costs of dry removal systems are very site-specific The
type of fuel and sorbent, the fuel sulfur content, the flue gas volume
treated, the desired control efficiency, the plant location, and other
variables greatly affect costs. Despite the wide variations such variables
can cause, sufficient information is available to allow rough cost esti-
mates.
4.2-172
-------
The key features of five dry systems including estimated capital and
operating costs are given in Table 4.2-41.334
The costs of two-stage systems have received much attention. Recent
studies indicate that the capital cost of a two-stage system depends greatly
on the flue gas flow rate. The mid-1979 capital cost of such a system is
estimated to be about $87/kW for a 440-MW utility plant and $65/kW for a
600-MW plant with about the same flue gas flow rate.335 Thus, the total
installed capital cost is roughly the same for the plants. In general, the
mid-1979 capital cost of a two-stage dry system designed to control both'S02
and particulate emissions ranges from $55/kW to $98/kW.335
The annual operating costs of a two-stage dry,system include the costs
of the reactant, operating labor, steam, electricity, water, chemical anal-
yses, maintenance, waste disposal, and items characteristic to a particular
system. With a two-stage system, a utility plant uses between 1.25 and 1.75
kg of lime per kg of S02 removed (1.25 to 1.75 Ib of lime per Ib of S02
removed).335 The cost of operating labor depends on the size of the system
and the degree of automation; two persons per shift should be adequate.335
A two-stage dry system may use steam in some climates for heat tracing of
feed tanks and lines. The electrical costs cover mainly the power to drive
the atomizers and the incremental power needed by the fans to overcome the
system pressure drop. The dry system uses about 1.14 liters (0.3 gal) of
water to treat 28.3 m3 (1000 ft3) per minute of flue gas. Because most of
the water can come from the ash pond, cooling tower blowdown, or other plant
or boiler waste streams, the cost of water may be small. Chemical analysis
requirements are also low because of the simplicity of the\ process.335
Maintenance costs can vary widely according to preventive maintenance sched-
ules and crews; as of mid-1979, however, annual maintenance is estimated to
cost about $1.09-for every installed kilowatt of capacity.,335 Waste dis-
posal costs are those incurred in disposing of the mixture of fly ash, dry
product, and unused reactant; this mixture contains less than 0.2 percent
water by weight and thus costs much less to transport to a disposal site
than sludge from a wet scrubber.335
One source estimates a total annual operating cost of roughly
$11,400,000 and an annualized cost of 3.27 mills/kWh for a two-stage dry
4.2-173
-------
co
to
C3
co
o
cc
co
LJJ
a:
UJ
u_
co
CM
UJ
_J
OQ
II
Is
•*"S.
UJ *
g;
gsf I
5s!5
-Is 5
'i . g
=11 s 5
33
O O
13
I. u
o o
MM
Si
8 S
.§88
I
;|-u"s:
o — */ v i
i -- s ^ t. s*— « i: *• -o i
£ J
>,
n) u L.
*»:
5 y.
£*Jl
? •= r T ° = •
J •• , 5 u> > o- E
- if * •_ * £
b J= ^< C wa^wi
LTWIVM >OUb
: ••- £ a. •- *- n c
i a M a u. v « ^
. a .rs •
>.«
at -a
V> V Oi
t
s
3§
5^ -SC.
= c
— vi 2 a
™ x -*
,£i
.
Kit
SS33
i — C O Q.
t- o c W
-
Si. 2
b - o c -p
u o w *J o £ «*
Jo *> «•«•>*>
o «• i. w •— Q
e: -i ja oi x-f
— C, « L. C
SI. |Mi-
•" -S Ifis'S
•sal =slt
8-^t Sgfct
i£S ssffe
— * w o v S *J
4.2-174
-------
removal system with lime sorbent at a 500-MW utility plant.335 (These
values have been adjusted from May 1978 to mid-1979.) The plant is assumed
to operate 7000 h/yr, to fire coal containing 2 percent sulfur and 10 per-
cent ash, and to achieve an S02 removal efficiency of 80 percent. About 60
percent of the annualized cost estimated for a new system covers direct
operating expenses, and 40 percent covers the capital charge on the total
investment, assumed to be about $76/kW as of mid-1979.335
Basin Electric Power Cooperative plans to use the two-stage dry removal
process at two lignite-fired units: the 440-MW Unit 1 of the Antelope
Valley Station near Beulah, North Dakota, and the 600-MW Unit 3 of the
Laramie River Station near Wheatland, Wyoming. Table 4.2-42 presents the
total costs of dry removal at these units for an estimated life of 35
years.336 The FGD system at Antelope Valley will consist of a spray dryer
with a rotary atomizer and a fabric filter. The system at Laramie River
will include a spray dryer with a "Y-jet" nozzle and an ESP. Lime will be
the sorbent for both systems; the estimated annual lime consumption rates
are 16,350 Mg (18,000 tons') at Antelope Valley and 19,000 Mg (20,920 tons)
at Laramie River.336 The annual power requirements will be approximately
5726 kW for Antelope Valley and 2451 kW for Laramie River. The estimated
manpower for each system is six operators and seven maintenance persons. At
Laramie River, the system will have to overcome a pressure drop of 1.6 kPa
(6.5 in. H20) between the air heater outlet and stack inlet. The S02
removal efficiencies are expected to range from 62 to 78 percent at Antelope
Valley and from 85 to 90 percent at Laramie River. According to Basin
Electric, the total savings of two-stage dry removal over wet limestone
scrubbing amount to a 36 percent cost reduction at Antelope Valley and 17
percent at Laramie River.336
Energy and environmental impacts—Dry removal requires less energy than
wet scrubbing because normally the temperature of the flue gas is not
greatly lowered and because saturation is not reached. 'Energy-consumptive
reheat is thus minimized or eliminated.337 According to Basin Electric, a
dry system needs only 25 to 50 percent of the energy required by a wet
system.309
4.2-175
-------
TABLE 4.2-42. ESTIMATED COSTS OF TWO-STAGE DRY REMOVAL
FOR 35 YEARS 6
(thousands of mid-1979 dollars)3
Capital investment
Lime
Electricity
Manpower
Replacement parts
Pressure drop
Total
Average annual cost
Antelope Valley
Unit 1
41,450
32,205
6,232
13,291
14,606
107,784
3,080
Laramie River
Unit 3
41,569
40,795
4,011
13,291
13,145
3,859
116,670
3,333
The Basin_Electric estimates were given for December 31 1981
Mnfth^1^6- ^r!.backdated to July 1, 1979, on the assump-
tion that the inflation rate for each item will be 7 5 per-
cent from mid-1979 to the end of 1981. '
4.2-176'
-------
Waste from a sodium-based system can pose environmental problems.
Because the portion of the waste that is soluble ranges from 50 to 60 per-
cent,338 substantial leaching may occur unless the disposal site is lined.
Waste from a calcium-based system is moderately cementitious and
impermeable. The solubility of such waste has been measured at 3 to 7
percent.338 Basin Electric plans landfill disposal of the calcium sulfate
and'sulfite waste from its two-stage dry removal systems. Although sig-
nificant leaching appears unlikely, other problems may occur, such as
weathering, erosion, fugitive dust, and structural instability.338 Disposal
procedures will be determined when the dry product becomes available.
Currently, the disposal of waste from a calcium-based system appears no more
dangerous or difficult than the disposal of fly ash.338 Disposal of the
used dry absorbent and fly ash at the mines where the absorbent is obtained
is being investigated.
4.2.4 Combined Coal Cleaning/FGD
The control of these S02 emissions can be partially or totally achieved
with current technology through coal cleaning, flue gas desulfurization
(FGD), or the combined use of cleaned coal with partial FGD. The subject of
coal cleaning is addressed in Section 4.2.2.1, and nine FGD processes are
discussed in Section 4.2.3. The combination of these technologies in the
achievement of S02 control will be discussed in this section.
Although coal washing cannot always eliminate the need for flue gas
scrubbing, the required S02 removal can be significantly reduced. For a 520
ng/J (1.2 lb/106 Btu) S02 emission limit with a 3.5 percent sulfur, 27,900
J/g (12,000 Btu/lb) coal being fired, the required S02 removal could be
reduced from 80 percent to 60 percent.339 Similarly for a 260 ng/J (0.6
lb/106 Btu) case with the same coal being fired, the required S02 removal
could be reduced from 90 to 81 percent. One impact of the reduced S02
removal requirement for an FGD system is that a partial bypass of the
untreated flue gas may be possible to provide the required reheat for the
cleaned flue gas. A second impact is to reduce the stringent, continuous
S02 removal requirement that may be required of an FGD unit. Another impact
is to improve the operation of the boiler because of the cleaner fuel being
4.2-177
-------
fired, which should also reduce the maintenance costs related to boiler
operation, as previously noted.
A major consideration that can affect FGD system design is that the use
of cleaned coal can significantly reduce the volume of sludge generated. In
a study based on the old NSPS for a utility boiler emitting 520 ng S02/J
Cl-2 Ib S02/106 Btu) and based on a 3.5 percent sulfur coal with a heating
value of 27,900 J/g (12,000 Btu/lb), a 500-MW generating unit, and 40 per-
cent sulfur removal by physical coal cleaning, the annual volume of sludge
generated is 149,000 m3 (121 acre-feet) for a lime-based FGD system and
155,000 m3 (126 acre-feet) for a limestone-based FGD system. Without coal
cleaning, the annual sludge volume generated is 286,000 m3 (232 acre-feet)
for a lime-based system and 299,000 m3 (242 acre-feet) for a limestone-based
system. This represents a 48 percent reduction in the sludge that must be
handled, treated, and stored or disposed of. If ash disposal is included,
the total solids that must be handled reflects a 44 percent reduction if
coal cleaning were used for lime and limestone FGD systems.340
4.2.4.1 Combined Coal Cleaning and FGD Costs-
Sulfur dioxide emission limitations when firing high-sulfur coal would
require additional S02 removal by an FGD system after coal cleaning. In one
study, several cases were examined to evaluate the economic benefits obtain-
able by the use of coal cleaning in combination with FGD versus FGD alone.
A single plant scenario is examined in which a single boiler is served by a
coal cleaning plant and a lime or limestone FGD system is installed to meet
the regulation level. In the first case, a 500-MW unit burning 3.5 percent
sulfur coal and required to meet a 520 ng S02/J (1.2 Ib S02/106 Btu) regu-
lation was considered. Considered in the second case were boilers of 25,
200, and 500 MW burning 7.0 percent sulfur coal and required to meet a 215
ng/J (0.5 lb/10* Btu) regulation level. Table 4.2-43 presents the washa-
bility data for the two coals.341
Case 1 involves 40 percent removal of sulfur by coal washing of a 3.5
percent sulfur coal. Conventional coal preparation can be applied to many
U.S. coals to achieve a 40 percent reduction in sulfur. In this situation,
the model coal selected is an Illinois coal with a raw coal sulfur content
of 3.48 percent. The U.S. Bureau of Mines washability data indicate that
4.2-178
-------
CO
o
CJ
CO
o
u_
13
CTiCQ
-P O)
re ^-s.
O)
're •>
^* ^>
Heating
recover
O> O)
•I- >
OJ O
3: <->
0)
re
o>
D)
in
re
CD co esj esj CD
esj LT> co en «tf-
esj CD r — en oo
ro CD en ^r*1 ^j*
i— i— i— CSl CO
CD CD O O CD
00000
r— r-. co <— •—
•3- CO CO CO CSJ
00000
O O O O CD
en o o CD o
CSJ CSl r— O CO
CO CO CO CO CSl
CO r— «^- CO O
CD CO CO CD CD
CM r»- en en o
r—
en LO ^" en o
t-^ CSJ CSJ LT> CD
CSl CD CO OO O
^j- CD cn
co i — i — i — re
o
i— X X X 0
co «*• CD 5
re
en r-- esi
i— CSl «3-
co «3- o
• — LT> en
CM CSl CM
O CD O
o o CD
o en r^
«3- co co
000
000
r-- •* CD
CM CSl CM
CO CO CO
CO r— <—
en •— co
tn co en
CSl «^- CO
CSJ *3" CO
LO r-»- oo
CO r— I—
r-^ X X
co «^-
CD
"""
CD
CD
O
O
CM
0
O
0
CO
CM
CD
0
O
CD
CD
O
"re
0
u
re
Q£
4.2-179
-------
cleaning at 1.8 specific gravity (s.g.) will reduce the sulfur content by
about 50 percent with a Btu yield of 93.4 percent; the data also indicate a
45 percent reduction in sulfur at 1.9 s.g. with a 96.3 percent Btu yield.
Assuming that the higher .cleaning gravity is used and that a grass roots
cleaning plant is built, the capital costs of cleaning should be in the
range of $8,650 to $26,000 per ton per hour of raw coal processed. For a
state of the art cleaning plant, operating 4000 hours/year and processing
approximately 1.45 Tg (1,600,000 tons) per year of raw coal, the capital
investment is estimated to be approximately $3,030,000 to $7,180,000. Since
the size of this cleaning plant is small, the cost is estimated on the high
side of the range at $6,700,000. Operating costs are estimated to be 2.47
to 3.72 mills/kWh. The additional coal required, because of heat value lost
with the coal cleaning refuse, is estimated to be about 90.7 Gg (100,000
tons) annually. At an assumed cost of $1.14/GJ ($1.20/106 Btu), the addi-
tional costs for coal would be $2,800,000 (0.98 mills/kWh).342
Case 2 was evaluated in exactly the same manner as Case 1 using wash-
ability data for the 7.0 percent sulfur coal. Costs do not differ appre-
ciably from those obtained for Case 1.
For a 520 ng S02/J (1.2 Ib S02/106 Btu) regulation case, combined coal
cleaning and lime or limestone FGD are more expensive than either lime or
limestone FGD alone. Capital costs are about 1.5 percent higher, while
annual costs are about 36 percent higher.*« Possible improved boiler
operation, reduced boiler maintenance, and reduced energy requirement by
utilizing flue gas bypass for cleaned flue gas reheat are not considered in
the annual costs.
It appears that the major benefit from the use of combined coal clean-
ing and FGD is in cases where FGD alone cannot attain the level of control
required.
One source states that, in some cases, under current state and Federal
standards, the S02 control costs of using FGD in combination with physical
coal cleaning may be lower than those for using FGD alone.343
4.2.5 Combustion Process Modifications
Any improvement in the efficiency of a combustion process that reduces
the fuel requirement will reduce both fuel costs and pollutant emissions.
4.2-180
-------
With conventional combustion systems, regular maintenance and proper opera-
tion will help ensure peak operating efficiency. Some advanced combustion
processes offer potential for further reduction in fuel requirements over
conventional systems and some also may reduce S0x emissions.
4.2.5.1 Conventional Combustion Systems--
Inefficient operation of the conventional combustion systems, whether
fired with coal, oil, or natural gas, can cause an increase in pollutant
emissions through increased fuel use or incomplete combustion. One cause of
inefficiency is an improper air/fuel ratio. Operation with insufficient air
(fuel-rich) can result in unburned fuel, whereas operation with too much, air
reduces efficiency by allowing the air to absorb heat unnecessarily and
carry it out the stack. Combustors should be checked periodically to ensure
the proper air/fuel ratio for optimum efficiency.
Incorporation of economizers and air preheaters is common in new
installations. An economizer recovers heat and raises the feedwater tem-
perature. An air preheater can also increase efficiency by reducing the
stack temperature and improving combustion conditions. Both types of equip-
ment can effect reduction of pollutant emissions by enhancing fuel utiliza-
tion. Retrofitting of such equipment at existing installations, however,
could be cost-prohibitive.
4.2.5.2 Fluidized Bed Combustion--
The fluidized bed process has been known since the 19th century and has
been used extensively in the petroleum industry. Its relatively recent
application to coal combustion offers several advantages over conventional
combustion methods.
Because heat release and heat transfer are higher in fluidized beds of
coal and air than in conventional furnaces, smaller (and perhaps less
costly) units can be used for a given generating capacity. Fluidized bed
combustion is being investigated at atmospheric and elevated pressures.
Operation above atmospheric pressures offers higher combustion efficiency,
greater unit capacity, and the potential for use of a combined-cycle system
for even higher thermodynamic efficiency.
4.2-181
-------
As an S0x emission control technique, fluidized bed combustion appears
to provide cleaner burning of high-sulfur coals. Limestone or dolomite
added to the fluidized bed absorbs sulfur released during combustion to form
sulfates. The temperature of the fluidized bed is high enough to calcine
limestone or dolomite to lime but low enough that the sulfate will not dis-
sociate. Research has shown that sulfur removal can be greater than 90
percent; the resultant waste is dry and therefore easier to handle than a
slurry or sludge.
Most of the development work on fluidized bed combustion has been with
small devices. It is not known whether data from this work can be directly
extrapolated to large, commercial-scale units. Operation of some larger
demonstration .units being planned will provide additional data on several
critical aspects of fluidized bed combustion.
Among the prime concerns in process development is the calcium/sulfur
(Ca/S) ratio, which is the ratio of the limestone/dolomite (sorbent)
required for each unit of sulfur removed. Ratios are reported to range from
2 to 4.5.344>345 High rates of S02 reduction can require large quantities
of sorbent and thus generate large quantities of waste. Methods being
developed to improve sorbent utilization include low bed velocities, regen-
eration of spent sorbent, and use of additives to the sorbent. Figure
4.2-41 shows the relationship of Ca/S ratio to sulfur reduction.a« Other
developments are concerned with recycling of elutriated fines to improve
efficiency, coal/sorbent feed systems, load response, bed dynamics and heat
transfer, and natural circulation.
4.2.5.3 Advanced Combustion Systems--
An extension of atmospheric fluidized bed combustion (AFBC) is pressur-
ized fluidized bed combustion (PFBC), in which combustion occurs at 6 to 16
times atmospheric pressure. Though not as highly developed as AFBC, PFBC
appears to offer advantages.
Combined-cycle power generation is one of the advantageous features of
PFBC. Utilizing the combustion gases in a gas turbine, as well as for steam
generation, offers the potential of attaining 40 percent system efficiency
versus 37 percent for AFBC.*" Pressur1zed FBC also prov1des potent1al for
smaller unit size and greater sorbent utilization. Development of gas
4.2-182
-------
ALL DATA ARE FOR UNITS FIRING
MEDIUM- TO HIGH-SULFUR EASTERN
COAL AND USING CRUSHED COAL AND
CRUSHED LIMESTONE AT A BED TEMP
ERATURE OF 843bC (1550CF)
• POPE, EVANS, AND ROBBINS
A NATIONAL COAL BOARD
• BABCOCK AND WILCOX
2 3
Ca/S MOLAR RATIO IN FEED
31+5
Figure 4.2-41. Once-through S02 reduction versus Ca/S molar ratio.
4.2-183
-------
turbines suitable for coal-fired service is a primary impediment to commer-
cial application of PFBC technology, which is not expected before 1985-
1990.347-348
Magnetohydrodynamics (MHD) is a means of directly converting the energy
of a high-temperature, ionized gas stream into electricity by passing it
through a magnetic field. Extremely high gas temperatures and addition of a
"seed" material are necessary for satisfactory gas conduction/ionization.
Since the exhaust gases would be at a high temperature [1650°C (3000°F)], an
MHD could serve as a topping cycle for conventional steam power generation
with system efficiencies as high as 50 to 60 percent.349-35! This would,
however, require modification to the conventional system to make it compat-
ible with MHD.
The high system efficiencies achievable by MHD will reduce fuel use and
thus pollutant emissions. Researchers have found that a potassium seed
material will combine with the S02 to form potassium sulfate (K2S04), which
can be removed by conventional pollution control devices as it solidifies
upon gas cooling.3^ studies are under way to develop methods of regener-
ating the potassium seed material while collecting the sulfur. There are
indications that care must be taken to avoid extensive NO emissions from
MHD systems. x
The MHD is not expected to be available until the end of this century.
The United States and the USSR are participating in a cooperative research
program, but difficult technical problems still must be solved. Major
problems include combustor performance, slag/seed separation and recovery,
erosion, corrosion, and material requirements.349,353
Many other methods of increasing the efficiency of coal combustion have
been studied and proposed. Advanced power cycles promise to reduce the
amount of emissions per unit of useful power produced in addition to
reducing fuel consumption. Although most of these methods are not yet
economically feasible or developed far enough to permit widespread use,
combined cycle gas turbines are Currently being used to produce electrical
power on a commercial basis. The desire to make these systems compatible
with coal or synthetic coal-derived fuels complicates the problems. In
addition, new combustors are being developed to reduce the NO emissions,
4.2-184
-------
which are generally greater than for a similar sized unit burning distillate
fuel oil.
In addition to MHD, the primary advanced power cycle candidates for
commercial use are open-cycle gas turbine combined cycles, closed-cycle
power systems, and fuel cells.354 The combined cycles utilize steam turbine
and gas turbine technology together to increase efficiency. Use with coal
firing presents materials problems (erosion, stress, temperature) requiring
solutions. Methods must be found for cleaning the hot gases to maintain
efficiency and prevent pollution and corrosion.
Closed-cycle power systems also utilize gas turbines but with an inert
gas or liquid metal as the working fluid in lieu of combustion gases. High
efficiency can be attained by operation at high temperatures with a steam-
bottoming cycle. Again, materials problems preclude use of these systems
before 1990.355
Fuel cells involve the electrochemical generation of electricity by
combining hydrogen and oxygen or oxygen and a mixture of hydrogen and carbon
monoxide. Coal-derived synthetic gas will fuel second-generation develop-
ment units. Because combustion gases must be clean before entering the fuel
cell, techniques for hot gas cleanup are needed.356
4.2-185
-------
REFERENCES FOR SECTION 4.2
1. U.S. Energy Information Administration Annual Report to Congress
Energy suppiy'
2" vAi; E^rgy™f°rnmat.ion Administration Annual Report to Congress.
Tmn^rtc nnp/pT, n^^1"0"5 °f Energy SuWK and Demand and Their
Impacts. DOE/EIA-0036/2. April 1978. p. 179.
3. Ref. 1, p. 93.
4" =;?" E|™i™nmeDntal Protection Agency. Electric Utility Steam Gener-
B
5. Ref. 4, pp. 4-6.
6. Commission of Natural Resources, National Academy of Sciences, National
Academy of Engineering, National Research Council. Air Quality and
9io o23ry UrCS Emiss1on Control. Serial No. 94-4. March 1975 pp
rr '
nPvl
Fact
M!?! K'i r' nD°tter'. and C' Ho1mes- Steam Electric Plant Factors 1978.
National Coal Association, Washington, D.C. 1978. p. i.
T-w-.,Dev1tt- Overview of Pollution from Combustion
p ?,°llerS °f the United States- u-s- Environmental
Protection Agency. Washington, D.C. EPA-600/7-79-233. October 1979.
I"terior' Bureau of Mines. Minerals Yearbook
pp 985nd 986 Minerals, and Fuels. Washington, D.C. 1978.
12.
13.
rnv °f /T%r9y> ln&T-gy Info™at^n Administration, Office of
fcnergy Data and Interpretation. World Crude Oil Production Year 1977
Energy Data Reports. July 20, 1978. pp. 2, 3. auc^°n> *ear iy//.
Ref. 2, p. 155.
Ref. 1, p. 60.
4.2-186
-------
14. Ref. 2, pp. 157, 275.
15. U.S. Environmental Protection Agency. Engineering/Economic Analysis of
Coal Preparation with S02 Cleanup Processes. EPA-600/7-78-002.
January 1978. p. 2.
16 Kiloroe J.D. Coal Cleaning for Compliance with S02 Emissions Regula-
------ Coal Confer-
122.
17.
18.
tions. In: Fourth Symposium on Coal Utilization, NCA/BCR
ence and Expo IV, Louisville, Ky. October 18-20, 1977. p.
Leo, P.P., and J. Rossoff. Controlling S02 Emissions from Coal-Fired
Steam-Electric Generators: Solid Waste Impact. U.S. Environmental
Protection Agency. Washington, D.C. EPA-600/7-78-044a. March 1978.
p. 18.
Min, S., and T.D. Wheelock. A Comparison of Coal Beneficiation
Methods. In: Coal Desulfurization Chemical and Physical Methods,
Wheelock, T.D. (ed.). ACS Symposium Series, American Chemical Society,
Washington, D.C. 1977. p. 83.
19. U.S. Environmental Protection Agency. Division of Stationary Source
Enforcement. Inspection Manual for the Enforcement of New Source
Performance Standards: Coal Preparation Plants. Washington, D.C.
EPA-340/1-77/022. August 1977. p. 4-4.
20. U.S. Department of Energy, Assistant Secretary for Policy and Evalua-
tion, Office of Technical Program Evaluation. International Coal
Technology Summary Document. HCP/P-3885. December 1978. p. 73.
21. Ref. 20, p. 63.
22. Ref. 16, p. 131.
23. Kilgroe, J.D. Development Progress in Coal Cleaning for Desulfuriza-
tion. In: Energy/Environment II, Second National Conference on the
Interagency R & D Program. EPA-600/9-77-012. November 1977. p. 177.
24. Hall, E.H., et al. Physical Coal Cleaning for Utility Boiler S02
Emission Control. EPA-600/7-78-034. February 1978. p. 82.
25. Ref. 16, p. 132.
26. Holt, E.C., Jr. An Engineering/Economic Analysis of Coal Preparation
Plant Operation and Cost. United States Department of Energy, Solid
Fuels Mining and Preparation Division. Washington, D.C. EPA-600/
7-78-124. July 1978. p. 286.
27. Ref. 24, p. 96.
28. Ref. 24, pp. 30, 31.
4.2-187
-------
29. Corbett, W.E. Low-Btu Gasification—Environmental Assessment. In:
Symposium Proceedings: Environmental Aspects of Fuel Conversion Tech-
nology, III, September 1977, Hollywood, Fla. EPA-600/7-78-063 April
1978. p. 135.
30. Page, G.C., and P.W. Spaite. Low- and Medium-Btu Gasification Systems-
Technology Overview. EPA-600/7-78-061. March 1978. p. 12.
31. Balzhiser, R.E. R & D Status Report, Fossil Fuel and Advanced Systems
Division, Gasification-Combined-Cycle Power Plants. EPRI Journal
3(6):43. July/August 1978.
32. Ref. 24, p. 30.
33. Ref. 30, p. 23.
34. Ref. 20, p. 47.
35. Ref. 6, p. 377.
36. U.S. Environmental Protection Agency. Advanced Fossil Fuels and the
Environment. EPA-600/9-77-013. June 1977. p. 8.
37. Emerson, D.B. Liquefaction Environment Assessment. In: Symposium
Proceedings: Environmental Aspects of Fuel Conversion Technology, III
September 1977, Hollywood, Fla. EPA-600/7-78-063. April 1978 p
208. H'
38. Hossain, S.M., J.W. Mitchell, and A.B. Cherry. Control Technology
Development for Products/By-Products of Coal Conversion Systems. In:
Symposium Proceedings: Environmental Aspects of Fuel Conversion Tech-
nology, III, September 1977, Hollywood, Fla. EPA-600/7-78-063 April
1978. p. 392.
39. Balzhiser, R.E. R & D Status Report, Fossil Fuel and Advanced Systems
Division, Solvent-Refined Coal Technology. EPRI Journal 3(5)-37
June 1978. -
40. Koralek, C.S., and V.B. May. Flue Gas Sampling During the Combustion
of Solvent Refined Coal in a Utility Boiler. In: Symposium Pro-
ceeding: Environmental Aspects of Fuel Conversion Technology, III
September 1977, Hollywood, Fla. EPA-600/7-78-063. April 1978. p.
I O£ •
41. Ref. 39, p. 39.
42. McRanie, R.D. Burning Solvent Refined Coal. In: Preprints of Papers
Presented at Anaheim, Calif. American Chemical Society, Division of
Fuel Chemistry. 23(1):156. March 12-17, 1978.
43. Ref. 36, p. 16.
4.2-188
-------
44. Eckstein, L. EPA Program
February 1978. p. 1.
46.
Status Report: Oil Shale. EPA-600/7-78-020.
45. Ogden, G.E., and R.D. Ridley. A Marketing Prospectus for Shale Oil.
In: Preprints of Papers Presented at Miami Beach, Fla. 23(4).-47.
American Chemical Society, Division of Fuel Chemistry, September 10-15,
1978.
Laseke, B.A., and T.W. Devitt. Status of Flue Gas Desulfurization in
the United States. PEDCo Environmental, Inc., Cincinnati, Ohio.
(Presented at U.S. EPA Symposium on Flue Gas Desulfurization. Las
Vegas. March 5-8, 1979.) EPA-600/7-79-167a. July 1979. pp. 286-340.
47. Smith, M. , M. Melia, and T. Koger. EPA Utility FGD Survey:
1979. U.S. Environmental Protection Agency. Washington,
600/7-79-022e. August 1979. pp. 6-33.
53.
April-June
D.C. EPA-
for
48. Devitt, T. , et al. Flue Gas Desulfurization System Capabilities
Coal-Fired Steam Generators. U.S. Environmental Protection Agency.
Washington, D.C. EPA-600/7-78-032a. March 1978. Vol. I. pp. 8-22.
49. Smith, M. , et al. EPA Utility FGD Survey: December 1978-January 1979.
U.S. Environmental Protection Agency. Washington, D.C. EPA-600/
7-79-022C. May 1979. p. A-18.
50. Ref. 49, pp. A-l through A-18.
51. Leo, P.P., and J. Rossoff. Controlling S02 Emissions from Coal-Fired
Steam-Electric Generators: Solid Waste Impact. Vol. I: Executive
Summary. U.S. Environmental Protection Agency. Washington, D.C.
EPA-600/7-78-044a. March 1978. pp. 4-10.
52. Ref. 51, p. 13.
Smith, C.L. Sludge Disposal by Stabilization - Why? In: Proceedings
of the Second Pacific Chemical Engineering Congress, Volume I. New
York, N.Y., American Institute of Chemical Engineers, August 28-31,
1977. p. 358.
54. PEDCo Environmental.
Research Institute.
2.4-4.
Inc. Lime FGD Systems Data Book.
Palo Alto, Calif. EPRI FP-1030.
Electric Power
May 1979. p.
55.
56.
Duvel, W.A., Jr., et al. State-of-the-Art of Sludge Fixation. Elec-
tric Power Research Institute. Palo Alto, Calif. EPRI FP-671.
January 1978. pp. 1-2 and 1-3.
Rossoff, J. , R.C. Rossi, R.B. Fling, W.M. Graven, and P.P. Leo.
Disposal of Byproducts from Nonregenerable Flue Gas Desulfurization
Systems: Final Report. U.S. Environmental Protection Agency.
Washington, D.C. EPA-600/17-79-046. February. 1979. pp. 15-17.
4.2-189
-------
57. Ref. 55, pp. 2-7 and 2-24 to 2-28.
58. Ref. 51, pp. 33-41.
59. Duvel, W.A. , Jr. and R.A. Atwood.
trie Power Research Institute.
January 1979.
FGD Sludge Disposal Manual. Elec-
Palo Alto, Calif. EPRI FP-977.
60. Barrier, J.W., H.L. Faucett, and L.J. Henson. Economics of Disposal of
Lime/Limestone Scrubbing Wastes: Untreated and Chemically Treated
X? ™;/ U'S- Environmental Protection Agency. Washington, D.C.
EPA-600/7-78-023a. February 1978.
61. Leo, P.P., and J. Rossoff. Controlling S02 Emissions From Coal-Fired
Steam-Electric Generators: Solid Waste Impact Vol. II: Technical
Discussion. U.S. Environmental Protection Agency. Washington, D.C.
EPA-600/7-78-044b. March 1978. 221 p.
62. Ref. 56, 165 pp.
63. Leavitt, C. , et al. Environmental Assessment of Coal- and Oil-Firing
in a Controlled Industrial Boiler. U.S. Environmental Protection
Agency. Washington, D.C. EPA-600/7-78-164a,b,c. August 1978. 26 p
168 p. , and 328 p. at--,
64. Weaver, D.E., J. Schmidt, and P. Woodyard. Data Base for Standards/
Regulations Development for Land Disposal of Flue Gas Cleaning Sludges.
U.S. Environmental Protection Agency. Cincinnati, Ohio. EPA-600/
7-77-118. December 1977. 285 p.
65. Fling, R.B., et al. Disposal of Flue Gas Cleaning Wastes: EPA Shawnee
Field Evaluation - Second Annual Report. U.S. Environmental Protection
Agency. Washington, D.C. EPA-600/7-78-024. February 1978. 184 p.
66. Barrier, J.W., H.L. Fawcett, and L.J. Henson. Economics of Disposal of
Lime/ Limestone Scrubbing Wastes: Sludge/Flyash Blending and Gypsum
r™ !™', -, y-S- Environmental Protection Agency. Washington, D.C.
EPA-600/ 7-79-069. February 1979. 209 p.
67.
68.
69.
70.
Ref.
Ref.
Ref.
Ref.
55,
55,
55,
55,
P-
P-
P-
pp.
1-1.
2-16.
2-25
4-4, 5-3.
71' teo' rf:P-' and J- Rossoff. Control of Waste and Water Pollution from
Coal-Fired Power Plants: Second R & D Report. U.S. Environmental
Protection Agency. Washington, D.C. EPA-600/7-78-224. November 1978
pp. 37-168.
4.2-190
-------
72. Ref. 62, p. 6.
73.
U S Environmental Protection Agency. Electric Utility Steam Gener-
ating Units, Background Information for Proposed S02 Emission Stan-
dards. Research Triangle Park, N.C. EPA-450/ 2-78-007a. July 1978.
pp. 4-80, 4-81.
74. Ref. 73, pp. 4-1.4 to 4-43.
75. Ref. 48, pp. 10, 11.
76. Smith, M. , et al. EPA Utility FGD Survey:
1979. U.S. Environmental Protection Agency.
600/7-79-022C. May 1979. pp. 259-268.
77.
Choi, P.S.K., et al. Stack Gas Reheat for Wet
Systems. Electric Power Research Institute.
FP-361. February 1977. 73 pp.
December 1978 - January
Washington, D.C. EPA-
Flue Gas Desulfurization
Palo Alto, Calif. EPRI
78. Ref. 54, pp. 4.6-1, 4.6-3.
79. Ref. 54, pp. 4.6-3, 4.6-4.
80. Ref. 54, p. 4.6-4.
81 Bulger, L. , F.E. Diederich, and G.W. Vandervoort. Disposition of Power
Plant Wastes. (Presented at American Power Conference 36th Annual
Meeting. Chicago. April 29-May 1, 1974.) pp. 5, 6.
82 Evans R J Potential Solid Waste Generation and Disposal from Lime
and Limestone Desulfurization Processes. U.S. Bureau of Mines.
Washington, D.C. Bureau of Mines Information Circular 8633. 1974.
pp. 12, 13.
83 Elisson, W. , E.G. Kominek, and E.S. Robbins. Ultimate Disposal of
Wastes Collected in Wet Scrubbers. (Presented at the 66th Annual
Meeting of the Air Pollution Control Association. Chicago. June
24-28, 1973.) pp. 3-8.
84. Ref. 54, pp. 2.4-1 to 2.4-27.
85.
86.
Shah Y.M. , and D.H. Henz. Cost Analysis of Lime-Based Flue Gas
Desulfurization Systems for New 500-MW Utility Boilers. U.S. Environ
mental Protection Agency. Research Triangle Park, N.C. EPA-450/
5-79-003. January 1979.
Devitt T W. , et al. Assessment of Alternative Strategies for the
Attainment and Maintenance of National Ambient Air Quality Standards
for Sulfur Oxides. U.S. Environmental Protection Agency, Research
Triangle Park, N.C. EPA Contract No. 68-02-1375, Task Order No. 17.
January 1975. pp. 75-79.
4.2-191
-------
SmJ in h "••"?/•?; Zada' StatUS °f Flue Gas Desulfurtzatlon Sys-
nno r n i*nit-ed -tateS' (presented ^ the U.S. EPA Symposium on
Flue Gas Desulfurization. Atlanta. November 4-7, 1974. ) 73pp.
88. Ref. 85, p. 4-9.
89. Ref. 85, p. 4-8.
90. Ref. 51, p. 4.
91. Smith M and M. Melia. EPA Utility FGD Survey: July-September 1979
*>A-60?/
Lime6' '5'
The FGD Rea9ent Dilemma
94.
Proaram;N"AH? ^\ *?Bnt Results fr°m EPA' s Lime/Limestone Scrubbing
% ^ ! 5Crubber Additive. Bechtel National. San
. (P;esentedM at the U.S. EPA Symposium on Flue Gas
Las Vegas. March 5-8, 1978,) pp. 2-1, 2-2.
McGlamery, G.G. , et al. Detailed Estimates for Advanced Effluent
Desulfurizatfon Processes. U.S. Environmental Protection Agency
Washington, D.C. EPA-600/2-75-006. January 1975. 418 pp M9ency..
cists' fLorL
Aaencv
?978 pp
ea1r 1PaITtl"C,UlaDte. and Sulfur D1ox1de Emission Control
?3 "FlTd Boilers- u-s Environmental Protection
i to 4-Is1an ' N-C- EPA-450/3-78-007. February
96.
gmlth M and M. Melia EPA Utility FGD Survey: January-March 1980.
7-fln n?Qh M noJn Protectlon ^ency. Washington, D.C. EPA-600/
7-80-029b. May 1980. Appendix A and, p. 7-37'.
97.
98.
99.
TOO. Ref. 85
101. Ref. 85, p. 4-7
Ref. 94, pp. 83-89 and 162-167.
Ref. 85, p. 4-4.
Ref. 85, p. 4-5.
p. 4-6.
E?fe^'nfW*T" °/W; "ar9rove' and R-S. Merrill. A Summary of the
LimestonP wTP%LChemiCal Variables UP™ the Performance of Lime/
Limestone Wet Scrubbing Systems. Interim Report Electric Power
Research Institute. Palo Alto, Calif. EPRI FP-636. December 1977
4.2-192
-------
103. Ref. 102, p. 2-10.
104. Borgwardt, R.H. Limestone Scrubbing of
Progress Report 12. Monsanto Research Co.
S02
St.
at EPA Pilot Plant.
Louis, Mo. July 1973.
105. Borgwardt, R.H. Increasing Limestone Utilization in FGD Scrubbers.
U S. Environmental Protection Agency. Research Triangle Park, N.C.
(Presented at the 68th Annual AICHE Meeting. Los Angeles. November
1975.)
106. Borgwardt, R.H. IERL-RTP Scrubber Studies Related to Forced Oxidation.
In: Proceedings: Symposium on Flue Gas Desulfurization. U.S. Envi-
ronmental Protection Agency. Washington, D.C. May 1976. pp. 117-144.
107. Ref. 102, pp. 2-13, 2-14.
108. Smith, M. , M. Melia, and T. Koger. EPA Utility FGD Survey: October-
November 1978. U.S. Environmental Protection Agency. Washington, D.C.
EPA-600/7-79-022b. February 1979. pp. 90-100.
109. Devitt, T.W., et al. Flue Gas Desulfurization System Capabilities for
Coal-Fired Steam Generators. U.S. Environmental Protection Agency.
Washington, D.C. EPA-600/ 7-78-032b. March 1978. Vol. II. pp. 3-96,
3-98.
110. Ballard B. , and M. Richman. FGD System Operation at Martin Lake S.E.S.
American Power Conference. April 1979.
111. U.S. Environmental Protection Agency, Office of Air, Noise, and Radia-
tion, Office of Air Quality Planning and Standards. Electric Utility
Steam Generating Units - Flue Gas Desulfurization Capabilities as of
October 1978. Research Triangle Park, N.C. EPA-450/3-79-001. January
1979. p. 2-131.
112. Ref. 109, pp. 3-86, 3-89.
113. Ref. 73, pp. 4-91, 4-95.
114. Ref. 109, p. 3-98.
115. Ref. 73, p. 4-95.
116. Ref. 95, p. 4-16.
117. Energy Consumption In Manufacturing. Energy Policy Project of the Ford
Foundation. Cambridge, Ballinger Publishing Company. 1974. p. 406.
118.
Kaplan, N. An Overview of Double Alkali Processes for Flue Gas Desul-
furization. U.S. Environmental Protection Agency. Research Triangle
Park, N.C. (Presented at the U.S. EPA Symposium on Flue Gas Desulfuri-
zation. Atlanta. November 4-7, 1974.) pp. 453-454.
4.2-193
-------
Prnn^I ?/?'' TTT n ^ Report: Dual Alkali Test and Evaluation
Program Vol. Ill: Prototype Test Program-Plant Scholz. U.S Envi-
ronmental Protection Agency. Washington, D.C. EPA-600/7-77-050c. May
1-7 / / . p. 111""!.
*ntro5fuct1on to Double Alkali Flue Gas Desulfurization
n i ,"; Proceedln9s: Symposium on Flue Gas Desulfurization -
'h \ MarnCh ^^ ^' l' U'S' Environmental Protection
Washington, D.C. EPA-600/2-76-136a. May 1976. pp. 387-422.
Ref. 122, p. 942.
120.
np«,,ift,^?\- P D 9 Experience Wlth tne Zurn Double Alkali Flue Gas
Desu furizat on Process. In: Proceedings: Symposium on Flue Gas
Desulfurization - New Orleans, LA. March 1976; Vol. I. U.S Environ-
mental Protection Agency. Washington, D.C. EPA-600/2-76-136a. May
iy/o. pp. 503-514.
122. Kaplan, N. Summary of Utility Dual Alkali Systems. In: Proceedings:
Symposium on Flue Gas Desulfurization - Las Vegas, Nevada. March 1979
i/ni ii 11 v c7p«tif-«tA^vBim._.u.j-_i i-\ i .• . '
123.
124.
125.
126.
127.
128.
129.
130.
131.
132.
133.
134.
135.
136.
137.
La!)oratory Study of Limestone Regeneration
•, • * n . U.S. Environmental Protection Agency
Washington, D.C. EPA-600/7-77-074. July 1977. p. 2. Agency.
Ref. 109, p. 3-117.
Ref. 109, p. 3-186.
Ref. 47, pp. 10, 16, 27, 74-76, 107-108, and 134-135.
Ref. 73, p. 4-106.
Ref. 73, p. 4-107.
Ref. 73, p. 4-108.
Ref. 73, p. 4-109.
Ref. 73, p. 4- 110.
Ref. 73, p. 4-112.
Ref. 122, pp. 898, 899.
Ref. 48, p. 3-187.
R/f. 122, pp. 921, 923.
Ref. 95, pp. 4-16, 4-17.
4.2-194
-------
138. Tuttle, J-, and A. Patkar. The Status of Industrial Boiler FGD_Appli-
cations in the United States. PEDCo Environmental, Inc. Cincinnati,
Ohio (Presented at the U.S. EPA Symposium on Flue Gas Desulfunza-
tion'. Las Vegas. March 5-8, 1979.) EPA-600/7-79-167b. July 1979.
pp. 994, 995.
139 Legatski L.K., et al. Technical and Economic Feasibility of Sodium-
Based SO'2 Scrubbing Systems. (Presented at the U.S. EPA Symposium on
Flue Gas Desulfurization. Hollywood, Florida. November 1977.) U.S.
Environmental Protection Agency. Washington, D.C. EPA-600/7-78-058b.
March 1978. p. 983.
140. Dickerman, J.C. Flue Gas Desulfurization Applications to Industrial
Boilers. Radian Corporation. Durham, N.C. (Presented at the U.S. EPA
Symposium on Flue Gas Desulfurization. Las Vegas. March 5-8, 1979.)
EPA-600/7-79-167b. July 1979. p. 1144.
141. Ref. 73, p. 4-99.
142 Laseke, B.A. Electric Utility Steam Generating Units--Flue Gas Desul-
furization Capabilities as of October 1978. U.S. Environmental Pro-
tection Agency. Washington, D.C. EPA-450/3-79-001. January 1979. p.
2-181.
143. Gerstle, R.W., and G.A. Isaacs. Survey of Flue Gas Desulfurization
Systems, Reid Gardner Station, Nevada Power Co. U.S. Environmental
Protection Agency. Washington, D.C. EPA-650/2-75-057-J. October
1975. pp. 3-1 to 3-5.
144. Ref. 138, pp. 995, 996, 1000-1002.
145. Ref. 138, p. 1003.
146. Ref. 138, p. 996.
147. Ref. 140, p. 1143.
148. Ref. 49, pp. 187-224.
149. Ref. 47, pp. 84-89.
150. Tuttle, J. , et al. EPA Industrial Boiler FGD Survey: First Quarter
1979. U.S. Environmental Protection Agency. Washington, D.C. EPA-
600/7-79-067b. April 1979. p. 33.
151. Ref. 150, p. 56.
152. Ref. 140, p. 1147.
153. Ref. 140, p. 1157.
154. Ref. 140, p. 1158.
4.2-195
-------
155. Ref. 138, p. 1020.
156. Ref. 140, pp. 1157, 1158.
157. Ref. 139, pp. 993, 994.
158. Ref. 140, p. 1154.
159. Wil
160.
161. Ref. 160, p. 183.
162. Ref. 160, pp. 172-204.
163.
164.
165. Ref. 164, pp. 192-217.
166.
Desulfuriza-
InstHute- pai°
• TO. ' --......«,j, Report on SOo Control Svstpmc; fnr
lal Combustion and Process Sources. U.S. Environmental Protlc-
tion Agency. Research Triangle Park, N.C. December 1977Vol II.
167. Ref. 166, Vol. IV.
168. Ref. 166, Vol. VI.
169.
> iiJV ? ^' EPA IndustHa1 Boiler FGD Survey: Fourth Quarter
.^
170. Ref. 159, Abstract.
1?1' Katzen' Associates! C^*^^ RaPnael
172.
173. Ref. 160, pp. 175-178.
4.2-196
-------
174 Slack, A.V. Fertilizer Developments and Trends. Noyes Data Corpora-
tion.' N.J. 1968. pp. 146-149.
175. Ref. 160, pp. 184-187.
176. Ref. 166, pp. 2-59, 2-68.
177 Letter from Kaupisch, K.F., Raphael Katzen Associates, to Hartman,
JS., PEDCo Environmental, Inc. October 11, 1979. p. 2. Response to
request to comment on the ammonia FGD system section.
178. Ref. 160, pp. 179, 180, and 189-197.
179. Ref. 159, pp. 25-32.
180. Monsanto Enviro-Chem. Brink Fact Guide for the Elimination of Mists
and Soluble Solids. St. Louis, Mo. 1978.
181. Tennessee Valley Authority. Sulfur Oxide Removal from Power Plant
Stack Gas, Ammonia Scrubbing, Conceptual Design and Cost Study.
National Air Pollution Control Association. New York, N.Y. 2y^A.
1970.
182. Ref. 159, pp. 195-200.
183. Rust Engineering Co. Rayonier/Katzen Process for S02 Removal from Flue
Gas and Conversion to Ammonium Sulfate. Birmingham, Ala.
184. Ref. 159, p. xviii.
185 Ennis, C.E. S02 Removal with Ammonia: A Fresh Perspective. In:
Proceedings of the Second Pacific Chemical Engineering Congress,
Denver, Colo. American Institute of Chemical Engineers. New York,
N.Y. August 1977. Vol. I. pp. 345-351.
186. Tennessee Valley Authority. Pilot-Plant Study of an Ammonia
Absorption-Ammonia Bisulfite Regeneration Process. Topical Report
Phases I and II. U.S. Environmental Protection Agency. Washington,
D.C. EPA-650/2-74-049a. 1974.
187. Ref. 169, pp. 84-89.
188. Devitt, T. , et al. Flue Gas Desulfurization System Capabil Hies for
Coal-Fired Steam Generators, Vol. II. Technical Report. EPA-600/
7-78-032b. March 1978. p. 3-279.
189 Pedroso, R. An Update of the Wellman-Lord Flue Gas Desulfurization
Process. U.S. Environmental Protection Agency. Washington, U.t.
EPA-600/2-76-136a. May 1976. p. 720.
190. Ref. 189, p. 724.
4.2-197
-------
TPoWer Plant F1ue Gas Desulfurization by
tion • - In: Proceedings of the 12th Air Pollu-
t on and Industrial Hygiene Conference on Air Quality Management in the
siudleT ThTuni^r' ,C°T°Per' Hal ^ ' CeSfor Ene gj
Studies, The University of Texas at Austin. January 28-30, 1976. p
192. Ref. 188, pp. 3-261 to 3-267.
193. Ref. 142, p. 2-142.
194. Ref. 188, pp. 3-267 to 3-291.
Hodc;ni/-- Unk' Bating and Status Report-
Lord S02 Removal/Allied Chemical S02 Reduction Flue Gas Desul-
funzation Systems at Northern Indiana Public Service Comnanv and
' Pm on
196.
197.
198.
199.
200.
201 '
2°2'
Ref. 195, pp. 1, 3, 8, and 47.
Ref. 195, pp. 51-52.
Ref. 195, p. 51.
Ref. 195, p. 54.
Ref. 195, p. 53.
Cc°ntl'nU.1'n9 Pr°9ress for Wellman-Lord S02 Process. In-
1974 Vol IT Sy;P°,S1UFm °n Flue Gas Desulfurization - Atlanta, November
EPA-650/2-74-l9fih n Envikronn;^ta1 Protection Agency. Washington, D.C.
CCH bsu/^ /4-126b. December 1974. p. 750.
J'.i
600/-7fi
bOO/2-76-
.- . S02 Abatement for Stationary Sources in
Environmental Protection Agency. Washington, D.C. EPA-
January 1976. pp. 5-3, 5-8
the'
Power Plant
script p 2444. Arllngt°n'
204. Ref. 188, p. 3-276.
Public hearin9 and Conference on
fUr °Xlde Emissions Regulations by
October 18 to November 2, 1973, Tran-
" at" 'nWSCO'f n^*^ nStfUS °f the Wei Iman-Lord/Al lied FGD
llJ ^ \ Mitchell Generating Station. U.S. Environ-
pp 704-705. Y' WaSh1ngt°n' D'C- EPA-600/2-76-136b. May
206. Ref. 188, pp. 3-280, 3-281.
4.2-198
-------
207. Ref. 195, pp. 7-9.
208.
Smith, M. , M. Melia, and N. Gregory. EPA Utility _FGD Survey:
October-December 1979. U.S. Environmental Protection Agency.
Washington, D.C. EPA-600/7-80-029a. January 1980. pp. 250 to 257.
209. Ref. 208, pp. 27, 28, and 309-317.
210 Public Service Company of New Mexico. Information provided in response
to Edison Electric Institute Flue Gas Desulfurization questionnaire.
March 21, 1975. p. 2.
211. Ref. 195, p. 42.
212. Ref. 195, pp. 42-43.
213. Ref. 195, p. 46.
214. Ref. 195, pp. 42, 50.
215.
PEDCo Environmental, Inc. Particulate and Sulfur Dioxide^ Emission
Control Costs for Large Coal-Fired Utility Boilers. U.S. Environmental
Protection Agency. Washington, D.C. EPA-450/3-78-007. February 1978.
p. 4-17.
216. PEDCo Environmental, Inc. EPA Utility FGD Survey: October-November,
1978. U.S. Environmental Protection Agency. Washington, D.L.
600/7-79-022b. February 1979. pp. A-24, A-25.
217. Ref. 215, pp. 4-21, 4-22.
218. Ref. 215, pp. 4-25, 4-26.
219.
Madenburg, R.S.,
Construction and
(Presented at the
Vegas. March 5-8,
et al. Citrate Process Demonstration Plant--
Testing. Morrison-Knudson Co., Inc. Boise, Idaho.
U S. EPA Symposium on Flue Gas Desulfurization. Las
1979.) EPA-600/7-79-167b. July 1979.
220.
221,
Farrington, J.F., Jr., and S. Bengtsson. The Flakt-Boliden Process for
S09 Recovery Flakt, Inc. Old Greenwich, Connecticut. (Presented at
the 1979 Annual Meeting of the Metallurgical Society of the AIME. New
Orleans. February 1979.) 13pp. .
Korosy L et al. Sulfur Dioxide Absorption and Conversion to Sulfur
by the' Citrate Process. In: Sulfur Removal and Recovery from Indus-
trial Processes, Pfeiffer, J. B. (ed.). American Chemical Society.
Washington, D.C. April 4-5, 1974. pp. 194-208.
222 Madenberg, R.S., and R.A. Kurey. Citrate Process Demonstration
Plant—A Progress Report. U.S. Environmental Protection Agency.
Washington, D.C. EPA-600/7-78-058b. March 1978. pp. 711-717.
4.2-199
-------
223' LSSeDeSUWlVD^7D;ttnElkA1ncS; +3nd Wn'A- McK1nney- "irate Process for Flue
VJQO L/tr3iJiiiJ"i7^Tinn — — u v T -a +11 *- n«.— _._j. 11 <~ r- • •*"-
rt. U.S. Environmental Protection
2-76-136b. May 1976. pp. 845-847.
224.
225. Ref. 222, p. 712.
226. Ref. 208, pp. 31, 343, 344, and 345.
227.
228. Ref. 222, pp. 732-734.
229. Ref. 222, p. 733.
230. Ref. 222, p. 734.
231. Ref. 219, p. 28.
232. Ref. 188, p. 3-201.
233. Flue Gas
234.
235. Ref. 188, p. 3-196.
236.
Instr1al Boiler FGD Survey: Fourth
237. Ref. 188, p. 3-197.
238. Ref. 233, p. 5.
239. Ref. 233, p. 6.
240. Ref. 236, p. 315.
241. Ref. 188, p. 3-205.
242. Ref. 233, p. 14.
Ac1d P^uction Via Magnesia
Evaluat'°"
Reg.nerable Flue Gas Desul
Research
4.2-200
-------
243.
244.
245.
246.
247.
248.
249.
250.
251.
252.
253.
254.
255.
256.
257.
258.
259.
Koehler, G.R., and E.J. Dober. New England S02 Control Project Final
Results In: Proceedings: Symposium on Flue Gas Desulfurization -
Atlanta', November 1974; Vol. II. U.S. Environmental P;otectlonc^e?Qo'
Washington, B.C. EPA-650/ 2-74-126b. December 1974. pp. 690-692,
707.
Ref. 243, p. 688.
Matsuda, S. Trip Report-Nonferrous Smelters in Japan. U.S. Environ-
mental Protection Agency, Industrial Environmental Research Laboratory^
Cincinnati, Oh. EPA Contract No. 68-02-1321, Task No. 38. February
1977.
U S Environmental Protection Agency. NATO-CCMS Study-Phase 1.1-
Status Report on the Magnesium Oxide Flue Gas Desulfurization Process.
Washington, D.C. Prepared under Contract No. 68-01-4147, Task No. 7.
June 1978. p. 2-9.
Szabo M F. , and R.W. Gerstle. Operation and Maintenance of Particu-
late Control Devices on Coal-Fired Utility Boilers. U.S. Environmental
Protection Agency. Washington, D.C. EPA-600/2-77-129. July 1977. p.
3-71.
Anz B M et al. Design and Installation of a Prototype Magnesia
Scrubbing Installation. United' Engineers and Constructors, Inc. May
15, 1973. p. 15.
Ref. 188, p. 3-214.
Ref. 246, p. vii.
Ref. 246, p. 2-11.
Ref. 246, p. 2-12.
Ref. 248, p. 8.
Lowell P.S., F.B. Meserole, and T.B. Parsons. Final Report-
Precipitation Chemistry of Magnesium Sulfite Hydrates in Magnesium
Oxide Scrubbing. Radian Corporation. Austin, Tex. Contract No.
68-02-1319, Task Nos. 36 and 54. June 24, 1977.
Ref. 248, p. 9.
, Ref. 188, p. 3-209.
Ref. 233, pp. 4, 8.
Ref. 236, p. 315.
Ref. 236, p. 316.
4.2-201
-------
260. Ref. 233, p. 14.
261. Ref. 233, p. 13.
262. Ref. 188, p. 3-210.
263. Ref. 233, p. 6.
264. Ref. 236, p. 326!
265.
266.
267.
268.
269.
270.
271.
272.
273.
274.
Ref. 236, p. 325.
Ref. 236, p. 327.
Ref. 236, p. 328.
Ref. 188, pp. 3-219, 3-220.
275.
276.
277.
278.
- • —• —-^-MVIWII r
November 1978. p. 242.
Ref. 246, p. 5-2
Ref. 233, p. 17.
Ref. 215, p. 4-22.
Haug, N. , G. Oelert, and
Study Phase I—Survey -•*
Gas Desulfurization
lenges of Moderr
1979. pp. l-i,
Ref. 236, p. 82.
Washln9ton> D.C.
n. U.S. Environ
EPA-600/7-78-210.
Recovery fro. e
Chemical Society. Washington, D.C.
Ref. 274, pp. 1-2, 1-3.
Ref. 160, pp. 270, 271.
B
Apn 14-5 1974 p iso
4.2-202
-------
Strum J J et al. BF Dry Adsorption System. Part I: FW-BF Gulf
?ower' Demonstration Unit Interim Results. Part II: BF-STEAG Demon-
stration Unit Operational Experience and Performance ^.S Environ-
mental Protection Agency. Washington, D.C. EPA-600/2-76-136b. May
1976. pp. 885, 899.
280. Ref. 236, p. 63.
281. Ref. 236, p. 65.
282 Rush RE and.R.A. Edwards. Evaluation of Three 20-MW Prototype Flue
Gas Desulfurization Processes. Vol. 1. Electric Power Research Insti-
tute, Palo Alto, Calif. FP-713-SY. March 1978. p. 44.
283. Bischoff, W.F. FW-BF Dry Adsorption System for Flue Gas Cleanup U.S.
Environmental Protection Agency. Washington, D.C. EPA-650/2-73-038.
December 1973. p. 4.
284. Ref. 236, pp. 57, 60, and 61.
285. Ref. 276, pp. 181-185.
286. Ref. 236, pp. 57-60 and 68.
287. Ref. 236, p. 56.
288. Ref. 283, pp. 2, 3, and 6.
289. Ref. 236, pp. 55, 57.
290. Ref. 188, p. 3-331.
291. Ref. 236, pp. 60, 61.
292. Ref. 283, pp. 7, 8.
293. Ref. 188, p. 3-333.
294. Ref. 236, p. 94.
295. Ref. 282, p. 41.
296. Ref. 279, p. 888.
297. Ref. 279, p. 891.
298. Ref. 282, p. 43.
299. Ref. 279, pp. 902, 906.
300. Ref. 236, p. 70.
4.2-203
-------
301. Ref. 188, p. 3-334.
302. Ref. 236, pp. 62, 64.
303. Ref. 283, p. 12.
304. Ref. 236, p. 93.
305. Ref. 274, p. 5-2.
306.
n R"E;' ?nd R'A- Edwards. Evaluation of Three 20-MW Prototype Flue
tute Pa cTltn t10" P™cesses" D Vol. 2. Electric Power Researchlnstl-
pp 2-23 and 2-24 °rma' Publlcation No- EPRI-FP-713. March 1978.
307. Blythe, G.M. , J.C. Dickerson, and M.E. Kelly. Survey of Dry S02 Con-
Park NSremS>n^' Env1!;onmenta1 Protection Agency. Research Triangle
October'lS, 1979. p^4 prepared by Radian Corporation, Austin, Tx.)
308. Ref. 319, p. 5.
309. Janssen K E. , and R L. Eriksen. Basin Electric's Involvement with Dry
Flue Gas Desulfurization. Basin Electric Cooperative Bismarck N D
(Presented at the U.S. EPA Symposium on Flue Gas Desulfur zation'. Las
Vegas. March 5-8, 1979.) EPA-600/7-79-167b. July 1979. p. 633
310. Ref. 208, p. 395.
311.
-al> EPA Industri'al Boiler FGD Survey: First Quarter
312. Ref. 208, pp. 9, 10, 24, 25, and 35.
313.
p.^.
Newsletter' Northbrook, 111. Number
NeWSlette- Northbrook, 111. Number
314.
315. Ref. 208, p. 24.
316. Ref. 307, pp. 9 to 15.
317. Ref. 307, p. 11.
318. Ref. 307, p. 13.
319. j^McIlva^ine^Company. The Fabric Filter Manual. Northbrook, 111.
4.2-204
-------
320.
Shah N D , D.P. Teixeira, and R.C. Carr. In: Proceedings: Symposium
on Flue Gas Desulfurization - Hollywood, Fla. November 1977; Vol. II.
Application of Dry Sorbent Injection for S02 and Particulate Removal
U S. Environmental Protection Agency. Washington, D.C. EPA-bUU/
7-78-058b. March 1978. pp. 924-928.
FGD Systems
(Presented
Las Vegas.
for the Electric
at the U.S. EPA
March 5-8, 1979.)
321. Lutz, S.J., and C.J. Chatlynne. Dry
Utility Industry. TRW. Durham, N.C.
Symposium on Flue Gas Desulfurization.
EPA-600/7-79-167b. July 1979. p. 3.
322. Ref. 307, pp. 30, 32.
323 Estcourt, V.E., et. al. Tests of a Two-stage Combined Dry Scrubber/S02
Absorber Using Sodium or Calcium. Bechtel Power Corporation. San
Francisco Calif. (Presented at the 40th Annual Meeting of the Amer-
ican Power- Conference, Illinois Institute of Technology. Chicago.
April 26, 1978.) p. 6. . , '
324. Ref. 323, p. 8.
325 Gehri D.C., and R.D. Oldenkamp. Status and Economics of the Atomics
International Aqueous Carbonate Flue Gas Desulfurization Process U.S.
Environmental Protection Agency. Washington, D.C. EPA-600/ 2-76-136b.
May 1976. p. 801.
326. Ref. 309, p. 334.
327. Ref. 309, p. 6.
328, Ref. 309, Table 8.
329. PEDCo Environmental, Inc. EPA Industrial Boiler FGD Survey: First
Quarter 1979. U.S. Environmental Protection Agency. Washington, D.C.
EPA-600/ 7-79-067b. April 1979. pp. 138, 139.
330. Ref. 329, pp. 151, 152.
331. Ref. 321, pp. 5, 6.
332. Lutz, S.J., et al. Evaluation of Dry Sorbents and Fabric Filtration
for FGD. U.S. Environmental Protection Agency. Washington, D.C.
EPA-600/7-79-005. January 1979. 144 pp.
333. Ref. 307, pp. 66 to 108.
334. Ref. 307, p. 7.
335. Fockler, R.B., W.V. Botts, and J.H. Phelan. New Approach to Dry S02
Removal. Pollution Engineering. JO(5):46-48. May 1978.
336. Ref. 309, Tables 7 and 8.
4,2-205
-------
337. Ref. 321, p. 14.
338. Ref. 309, p. 7.
339. Leo, P.P., and J. Rossoff. Controlling S02 Emissions from Coal-Fired
Stream-Electric Generators: Solid Waste Impact. U.S. Environmental
Protection Agency. Washington, D.C. EPA-600/7-78-044a. March 1978.
p. 18.
340. Ref. 339, pp. 4, 6.
341. US. Environmental Protection Agency. Sulfur Reduction Potential of
U.S. Coals: A Revised Report of Investigations. EPA-600/2-76-091
pp. 71, 164.
342. PEDCo Environmental, Inc. Particulate and Sulfur Dioxide Emission
Control Costs for Large Coal -Fired Boilers. U.S. Environmental Protec-
tion Agency. Washington, D.C. EPA-450/3-78-007. February 1978. pp.
"" ~"
343. Hoffman, L S J Aresco, and C.C. Holt, Jr. Engineering/Economic
Analysis of Coal Preparation with S02 Cleanup Processes for Keeping
High Sulfur Coals in the Energy Market. Prepared by the Hoffman-
™?«?^ Corporation for the U.S. Bureau of Mines under Contract No.
J0155171. Published by U.S. Environmental Protection Agency,
Washington, D.C. EPA-600/7-78-002. January 1978. p. 66.
344. U.S. Environmental Protection Agency. Electric Utility Steam Gener-
ating Units: Background Information for Proposed S02 Emission
Standards. Washington, D.C. EPA-450/2-78-007a. July 1978. p. 4-44.
345. Walker, D.J., R.A. Mcllroy, and H.B. Lange. Fluidized Bed Combustion
Technology for Industrial Boilers of the Future— A Progress Report
Combustion. 50(8): 26-32. February 1979. p. 27.
346. Freedman^S.I. Fluidized-Bed Combustion. In: Energy/Environment III.
o ™ njf"vlr°nmental Protection Agency. Washington, D.C. EPA-600/
9-78-022. October 1978. p. 320.
347. U.S. Department of Energy. International Coal Technology Summary
Document. Washington, D.C. HCP/p-3885. December 1978. p. 19.
348. Ref. 344, pp. 4-57, 4-58.
349. Ref. 347, p. 62.
350. Commission on National Resources, National Academy of Sciences
National Academy of Engineering, and National Research Council. Air
QA3] YM an? Stationary Source Emission Control. Washington, D.C.
94-4. March 1975. p. 357, 358
4.2-206
-------
351
352.
Penny, M.M., S.V. Bourgeois, and W.C. Cain. Development Status and
Environment Hazards of Several Candidate Advanced Energy Systems. In.
Proceedings of the 12th Intersociety Energy Conversion Engineering
Conference, Washington, D.C., August 28-September 2, 1977. Vol. 1. p.
648.
Dicks JB., et al. A Description of the Direct Coal-Fired MHD
Facility at the University of Tennessee Space Institute. In: Pro-
ceedings of the 12th Intersociety Energy Conversion Engineering
Conference, Washington, D.C., August 28-September 2, 1977. Vol. L. p.
999.
353. Ref. 352, p. 995.
354. Ref. 347, p. 55.
355. Ref. 347, pp. 59, 60.
356. Ref. 347, pp. 60-62.
4.2-207
-------
-------
SECTION 5
INDUSTRIAL PROCESSES
In this chapter, the largest process emitters of sulfur oxides are
discussed. The method of presentation is a description of the process
including the source or sources of sulfur oxide, emissions from the process,
the applicable or currently practiced sulfur oxide control technique(s), the
capital and annual costs associated with the control system (presented in
mid-1979 dollars), and a discussion of the energy and environmental impacts
associated with the control system. The information reported here is a
condensation of information reported in the open literature.
5.0-1
-------
-------
5.1 NONFERROUS PRIMARY SMELTERS
5.1.1 Process Description and Emission Sources
This section examines the production of the five principal nonferrous
metals: copper, lead, zinc, aluminum, and molybdenum. With the exception
of aluminum, the ores of these metals contain sulfur as a principal con-
stituent. An integral part of the processing operations is to separate the
sulfur from the metals; this separation is usually accomplished by oxida-
tion, which results in sulfur dioxide emissions.
5.1.1.1. Copper Smelters--
The production of copper begins with the processing of various copper
ores, which contain a small percentage of copper sulfide, such as the fol-
lowing: *
0 chalcocite - Cu2S - 79.8 percent copper
0 bornite - Cu5FeS4 - 63.3 percent copper
0 tetrahedrite - Cu5Sb2S7 - 57.5 percent copper
0 chalcopyrite - CuFeS2 - 34.5 percent copper
The percentage of copper in most ores mined is rarely above 1 percent, and
therefore the ore is concentrated in almost all cases by flotation. Con-
centration ratios may range from 5 to 1 to 40 to I.1 Sulfide ores are
mixtures of varying proportions of copper and iron sulfides mixed with
acidic or basic gangue. The conversion of iron sulfides to iron oxides
occurs preferentially to the conversion of copper sulfides to copper oxides.
As practiced in domestic plants, smelting consists of either two or
three distinct pyrometallurgical processing steps. Only enough sulfur
should remain in the concentrates to insure that the copper will form a
copper sulfide matte in subsequent processing operations. Sulfur in excess
of this amount is eliminated by roasting or in the smelting furnace. The
following primary reactions occur during^roasting (the first two are thermal
decomposition):
FeS2 -> FeS + S
CuFeS2 -» CuFeS + S
4FeS + 702 -* 2Fe203 + 4S02
5.1-1
-------
The sulfur formed in these reactions oxidizes to generate S02.
Copper matte, which is formed in the smelting furnace, is a mixture of
molten iron oxide and copper sulfide. These mixtures are mixable over a
wide range of compositions, so that oxidation of the iron alone will not
separate iron from the concentrates; however, if silica is present, the iron
oxide combines with it to form a liquid iron silicate slag that is immis-
cible with the sulfide phase. The slag phase floats on top of the matte
layer. The primary smelting and refining reactions are as follows:1
Smelting:
FeS
0
FeS + S0
S0
Slag formation:
Converting:
Slag blow:
Copper or
finish blow:
FeS + 1-1/202 •* FeO + S02
CuFeS2 + 1-1/202 -*• FeO + Cu2S
3Fe203 + FeS -> 7FeO + S02
xFeO + ySi02 •* x(FeO) -(Si02)
2FeS + 302 + Si02 -> 2FeO-Si02 + 2S0
Cu2S + 02 ^ 2Cu + SO,
Side reaction: Cu + 1/202 -* CuO
Matte is processed in a copper converter to remove the remaining impurities
and form blister copper. About 1 to 2 percent of the sulfur entering a
smelter is lost in slags, 3 to 4 percent is released as fugitive emissions,
and the remainder is contained as S02 in the gases from roasters, smelting
furnaces, and converters.2 A total of about 1.8 Mg (2 tons) of S02 is
generated for each 0.9 Mg (1 ton) of copper produced.
Although the three steps of copper smelting have the same functions in
all smelters, there are large differences in the equipment used. Two types
of roasters, four types of smelting furnaces, and two types of copper con-
verters are used in this country.
Roasters-Four domestic smelters use multiple-hearth roasters.3 These
are tall, cylindrical devices through which concentrate is passed downward
against a rising stream of hot air and combustion gases. Multiple-hearth
roasters are used primarily by "custom" smelters that process a variety of
5.1-2
-------
concentrates from many sources.4 In these operations, the roaster burns off
only a fraction of the sulfur to adjust the composition of the concentrates
to allow successful treatment in the smelting furnaces. Some multiple-
hearth roasters are operated only intermittently with concentrates that
necessitate their use. Off-gas typically contains less than 2 percent S02.
With the exception of one installation, no roasters are currently equipped
with controls for S02 emissions. Control at one installation is accom-
plished by blending the weak gas stream from the multiple-hearth roaster
with a stronger gas stream for feed to an acid plant. Fugitive emissions
are also a problem with multiple-hearth roasters.5
In contrast, fluidized-bed roasters are engineered and operated with
the production of sulfuric acid equal in importance to the smelting of
copper. These units are fed with special high-sulfur concentrates of uni-
form composition and produce a continuous stream of off-gas that is con-
sistent in volume and contains 12 to 14 percent S02.6 Four fluidized-bed
roasters, each with an associated plant producing sulfuric acid, are oper-
ating in this country.7 Emissions of S02 from fluidized-bed roasters con-
sist only of the tail gas discharge of the sulfuric acid plants and minor
fugitive discharges from occasional leaks.8
Smelting furnaces—The four types of smelting furnaces used in domestic
primary copper production are the Outokumpu flash smelter, the Noranda
continuous smelter, the electric smelter, and the reverberatory furnace.
The first three types produce off-gas streams rich enough in S02 to be used
for manufacture of sulfuric acid; the five copper smelters operating these
furnaces use the gas for this purpose. The reverberatory furnaces operating
at 11 of the 16 U.S. primary copper smelters, however, generate a weak
off-gas stream that accounts for the greatest amounts of uncontrolled emis-
sions of S02 in the nonferrous metals industries.7
The reverberatory furnace (shown in Figure 5.1-1) is a large horizontal
chamber into which concentrate, flux, and various other materials are
charged.9 The furnace is heated by direct firing with natural gas, oil, or
pulverized coal. During the firing of the furnace, sulfur in the concen-
trate burns to S02, which mixes with fuel combustion gases and large quanti-
ties of infiltrated cooling air. At about 1000°C (1832°F), the charge melts
and undergoes the complex reactions to form copper matte, which is tapped
5.1-3
-------
O)
O
O)
oo
5.1-4
-------
from the furnace and sent to the converters for further treatment.10 The
molten slag that also forms is discarded.
A typical charge contains 65 percent concentrate9 with the balance
composed of fluxes and reverts. The charge is fed into the furnace in small
batches, resulting in wide variations in the S02 concentration and volume of
the furance off-gases. From 20 to 50 percent of the sulfur in the concen-
trate is oxidized to S02 in the roasting step and 10 to 40 percent in the
smelting step.11 Concentration of the S02 in the furnace off-gases ranges
from 0.5 to 3.5. percent, only occassionally exceeding 2.5 percent.12 Some
smelters have very large capacity furnaces, accepting up to 1800 Mg (2000
tons) per day.9 A smelter of this size will generate about 300 Mg (330
tons) of S02 per day in the smelting furnace and roasters (if used). Where
reverberatory furnaces are used alone or in combination with multiple-hearth
roasters, all of the S02 generated is vented to the atmosphere because weak
S02 stream control is not practiced in this country. One smelter using a
fluidized-bed roaster to produce calcine for reverberatory smelting,
however, reports a very low S02 emission rate from its furnace and an over-
all sulfur retention of 94 percent; this smelter uses only strong S02 stream
control with a double-contact acid plant.
The electric smelter is similar to the reverberatory furnace, differing
in that the heat is supplied by electricity rather than by combustion of
fuel. By elimination of combustion gases, the S02 concentration in the
off-gas is increased. The minimum and optimum S02 concentrations required
for a conventional metallurgical-type single-contact sulfuric acid plant are
4 and 7 percent, respectively.13 For a double-contact sulfuric acid plant
these values are 7 and 9 percent.13 Electric smelters are economical only
in areas where electric power is relatively cheap and plentiful.
Both electric and reverberatory furnaces generate fugitive S02 emis-
sions during tapping and charging operations. These amounts are small in
comparison with the direct process emissions, and are estimated as approxi-
mately 5 percent and 7 percent of the direct process emissions.14 Fugitive
emissions are also released from leaks in the furnaces and auxiliary equip-
ment.15 No methods have been developed to measure these emissions.
5.1-5
-------
The principal S02 emissions from the Outokumpu flash smelter and the
Noranda continuous smelter are from.the acid plant tail gases. There are,
however, fugitive S02 emissions at tapping, slagging, and feeding ports''
These emissions are estimated to contain 2.9 percent of the total sulfur
feed to the furnace.16 These units require a continuous and consistent feed
of specially prepared concentrates, which are introduced along with pulver-
ized flux. In the flash smelter, heat is added by preheating the inlet
combustion air; the exhaust gases from fuel combustion are not mixed into
the S02-rich gas stream. Flash smelter off-gas is a continuous stream
containing up to 13 percent S02." Tne Noranda continuous smelter burns
fuel with oxygen-enriched air, which is mixed into the mineral feed.18
Additional oxygen-enriched air is blown into the base of the Noranda unit,
where it reacts with copper matte, partially converting it into impure
blister copper. The Noranda unit was designed to function both as a
smelting furnace and a converter; however, in its single U.S. application, a
standard converter is also used to facilitate control of trace elements.
The remaining sulfur in the matte is removed in the converter.19 The
Noranda furnace creates an off-gas containing about 16 to 20 percent S02,20
but air infiltration around an exhaust hood reduces this concentration to 10
to 13 percent.21 This furnace does, however, convert more than 90 percent
of the input sulfur into a continuous-flow gas stream suitable for feed to
an acid plant. Fugitive emissions during smelting'are minimized because of
the tight seal between the reactor and its hood.22
Converters--A copper converter is a horizontal, cylindrical vessel
equipped with jets (tuyeres) through which air is blown. A batch of molten
copper matte is charged into the converter through an opening on the top.
The blowing of air through the matte oxidizes the remaining sulfur to S02
and converts the remaining iron into a slag, which is most often recycled to
the smelting furnace. The product is blister copper. A converter is
mounted on rollers and is rotated to pour out slag and blister copper
through the charging port.
The two types of converters used in U.S. copper smelters differ mainly
in the manner in which off-gas is discharged. In the Peirce-Smith con-
verter, off-gas is vented through the charging port. This type is used in
5.1-6
-------
all but one of the U.S. smelters.7 In the Hoboken converter, off-gas is
drawn through a mechanical syphon attachment on one end of the vessel.
Converters have lower capacities than smelting furnaces; usually two or
three converters are needed to process matte from,one furnace. During the
batch cycle, exhaust fans withdraw off-gas from each converter at a rate
matched to the blowing rate. From a single converter the volume of off-gas
is variable, and S02 concentration ranges from essentially 0 to 20 percent.
Smelter operators therefore schedule the cycles of a group of converters so
that the combined off-gas is relatively consistent. This combined stream
contains 3.5 to 7 percent S02,23 and all but three smelters use this stream
for sulfuric acid manufacture.7
Copper converters are the principal source of fugitive S02 emissions
into the atmosphere of the converter building. With Peirce-Smith convert-
ers, gases generated during blowing are pulled away from the converter
through a hood mounted above the charging port.24 Figure 5.1-2 shows that
the primary hood is isolated when the converter is rolled out for either
materials addition or skimming and pouring.25 The Hoboken converter, with
its attached syphon, does not lose suction when the converter is rotated but
does lose S02 because of thermal and pressure imbalances in the converter,
buildups in the syphon, and air currents above the charging port. These
emissions can be minimized by proper operating practices.26 As much as 3:5
percent of the S02 may be lost through fugitive emissions from a Peirce-
Smith converter.
CHARGING
BLOWING
SKIMMING
Figure 5.1-2. Peirce-Smith copper converter operation
25
5.1-7
-------
5.1.1.2 Lead Smelters-
Lead ore is mined in two regions of the United States. Deposits in
several western states usually include recoverable amounts of zinc and
copper; some of these ores are processed first in a zinc smelter, and the
sulfur-free residue is used as raw material in a lead smelter." Ore from
the rich deposits in Missouri contains no other recoverable metals. The raw
material to smelters that process Missouri ore is a concentrate containing
almost 90 percent galena (lead sulfide). Although actual emissions will
vary depending on the sulfur content of the raw material, total S02 emis-
sions from lead smelters is reported to be about 0.14 kg (0.3 Ib) of S02 for
each 0.45 kg (1 Ib) of lead produced.28
Lead smelting begins with a sintering step to remove almost all of the
sulfur by oxidation. The product from sintering is then processed in a
blast furnace to produce lead bullion, an impure metal. Sintering is the
only part of the lead smelting process that can emit large amounts of S02
Fugitive emissions of S02 also occur at the discharge end of the sinter
machine, where the discharged material is broken into pieces. Low con-
centrations of S02 may occur in blast furnace off-gas. Processes that
purify lead bullion emit no S02, although they do emit other pollutants.
The sinter machine, the principal source of sulfur oxides emissions
contains a horizontal metal belt that passes slowly through an enclosed
combustion chamber. The charge material, a pelletized mixture of lead
concentrate and flux, is spread on the belt and ignited with a gas flame.
Burning of the charge during its passage through the chamber creates S02
Once ignited, the charge needs no supplemental fuel because it contains
enough sulfide concentrate to furnish self-sustaining combustion."
Off-gas from a sinter machine contains about 2 percent S02, which is
too weak for use as feed to an acid plant. 30 With most sinter machines
however, the exit gases can be split into two streams. A strong gas stream
evolved at the front end of the belt contains about 6 percent S02 • four
smelters use this stream for acid manufactured A weak gas stream from the
back end of the belt contains about 0.5 percent S02.« Weak gas recycle is
the only S02 control technique on weak gas streams in the United States
Two lead smelters do not divide the gases into strong and weak streams and
5.1-8
-------
practice no control of S02 emissions. Weak gas recycle is required by New
Source Performance Standards.
The solids at the end of the belt are fused into a porous clinker
(sinter), which is discharged from the belt into mechanical equipment that
crushes and screens it to form particles of the proper size for use in the
blast furnace. These operations release S02 from incomplete sintering and
create fugitive S02 emissions. No emission factors are reported for this
source because techniques for accurate measurement have not been developed.
In the blast furnace, crushed sinter, mixed with coke and flux are
passed downward through a tal 1 .vertical rectangular column. Air is blown
into the bottom of the furnace to burn the coke and create a high-tempera-
ture reducing environment that forms molten slag and lead bullion. Gases
from a blast furnace, containing about 0.5 percent S02, are released without
control of S02 emissions.33
5.1.1.3 Zinc Smelters--
Zinc ore is another sulfide mineral, and the raw material for a zinc
smelter is a zinc concentrate made from zinc ores by processes that create
no S02 emissions. Zinc concentrate contains up to 62 percent zinc, 32
percent sulfur, and variable amounts of iron, lead, cadmium, and copper.34
Although a zinc smelter converts all the sulfur into S02, emissions are well
controlled in this industry.
Two different methods are used to produce zinc metal from concentrate.
The first step in both methods is to burn off almost all the sulfur by
roasting. About 95 percent of the sulfur is converted into S02.35 The
product of roasting, called calcine, is an impure zinc oxide, which usually
.contains less than 0.3 percent sulfide sulfur.36 Roasting is the only
significant source of S02 emissions in a zinc smelter. Although actual
emission will vary depending on the sulfur content of the raw materials, the
process releases about 0.25 kg (0.55 Ib) of S02 for each 0.45 kg (1 Ib) of
zinc produced.37
Calcine is processed either electfolytically or pyrometallurgical ly
into zinc metal. The electrolytic process is entirely chemical and creates
no S02 emissions.38 Pyrometallurgical production may create low concentra-
tions of S02 (0.1 to 2.4 percent) in off-gas from sintering machines.39
5.1-9
-------
All U.S. zinc smelters now operate modern roasting equipment and use
all the off-gas for manufacture of sulfuric acid. Two types of roasters are
in use. These are the flash and the fluidized-bed roasters, both of which
oxidize the sulfur while the concentrate particles are suspended in a moving
stream of hot gases. Because zinc roasters need no supplemental fuel, the
off-gas is not diluted with products from fuel combustion.40 Roasters
operate continuously and produce consistent volumes of an off-gas stream
that usually contains 10 to 13 percent S02.«i Because most zinc smelters
are located in heavily industrialized regions, sulfuric acid is often a
profitable byproduct.
5.1.1.4 Aluminum Smelters—
The processes of aluminum production are completely different from
those of other nonferrous metal industries. Bauxite, the ore of aluminum,
contains no sulfur and no other recoverable metals.42 It is processed by
chemical methods to alumina (aluminum oxide) in bauxite refineries. The
only S02 emissions at refineries result from the combustion of fuel by
calcining furnaces used to remove moisture from hydrated alumina. Calcined
alumina is the raw material for aluminum smelters.
A smelter contains rows of electrolytic cells (shallow pots lined with
carbon) in which carbon electrodes suspended above the pot serve as anodes.
The pots serve as cathodes. Cryolite, a double-fluoride salt of sodium and
aluminum (Na3AlF6), is used as an electrolyte and a solvent for alumina.43
Alumina is added to and dissolves in the molten cryolite bath. The cells
are operated between 950° and 1000°C (1742° and 1832°F) with heat that
results from the electrical resistance between the anode and the cathode.43
During the reduction process, the aluminum is deposited at the cathode
where, because of its heavier weight, it remains as a molten metal layer
beneath the cryolite. The byproduct oxygen migrates to and combines with
the consumable carbon anode to form carbon dioxide and carbon monoxide,
which continually evolve from the cell.44 Additionally, the gas stream
contains evolved hydrogen fluoride and fluoride-containing particulate
matter.
The reduction cells in use for aluminum production in the United States
are of two basic types, prebake and Soderberg. There are two types of
5.1-10
-------
Soderberg cells that are designated according to the manner of mounting the
stud in the carbon anode: vertical stud Soderberg (VSS) or horizontal stud
Soderberg (HSS).
Prebake cells are so named because the anodes are preformed and then
baked in a separate facility often referred to as an anode bake plant. The
anodes are then mounted in the cell and are consumed in the aluminum produc-
tion. The anode butts, which remain after the anode is consumed, are
recylced for use in the preparation of new anodes.
In the Soderberg process, continuously formed, consumable anodes are
used. The anode paste is baked by the heat generated in the reduction cell.
The primary source of sulfur oxide emissions in aluminum production is
the sulfur in the coke (normally petroleum coke) and the coal tar pitch
binder used to produce the anodes. In the prebake process, the combustion
fuel to bake the anodes may be a significant S02 emission source. Petroleum
coke usually contains 2.5 to 5 percent sulfur, but may vary from 1.5 to 7
percent, sulfur.45'46 Pitch normally contains about 0.5 percent sulfur.46
The sulfur content of the coke depends on the crude petroleum stock and the
tendency of the sulfur to concentrate in the still bottoms at the refinery
and thus in the coke. The trend appears to be toward coke with higher
sulfur content; however, marketing factors and price may affect this
trend.47 The production of aluminum by electrolytic reduction (the prebake
and Soderberg processes) consumes 0.5 to 0.6 kg carbon anode/kg aluminum
produced (0.5 to 0.6 Ib carbon anode/lb aluminum produced).48
As the coke is processed (during prebake) or consumed in the reduction
cell, sulfur oxides are released. The emissions include those from the
anode prebake operation (prebake), the "primary" emissions (which are
captured by the pot hood exhaust system), and the "secondary" emissions
(which escape the primary exhaust system and exit through the roof moni-
tors). The great majority of S02 emissions are collected by the pot hood
exhaust system.
One source reports uncontrolled S02 emissions from anode bake plants
range from 5 to 47 ppm, which is 0.7 to 2 kg S02/Mg aluminum produced (1.4
to 4 Ib S02/ton aluminum produced).49 Other data indicate that emissions
are in the range of 0.09 to 1.7 kg S02/Mg aluminum produced (0.18 to 3.4 Ib
S02/ton aluminum produced).49
5.1-11
-------
The total amount of S02 generated per unit of aluminum produced is
essentially the same for the prebake, VSS, and HSS cases. The "primary"
cell hooding configuration for collection of process fumes is affected by
the characteristics of the different cell types.50 There are two types of
prebake cells, center-worked prebake cells (CWPB) and side-worked prebake
cells (SWPB), as well as the two Soderberg processes, VSS and HSS, which are
in use by the domestic aluminum industry. Information from seven primary
aluminum plants indicates the following:51
Primary hood collection
Cell type efficiency, %
CWPB 65 to 98
(average 88)
SWPB 85
VSS 81
HSS 80 to 95
(average 90)
Primary collector exhaust rate,
NmVkg A1 (106 scf/ton AT)
128 to 158 (4.11 to 5.05)
[average 141 (4.51)]
107 (3.44)
21 (0.67)
158 to 245 (5.06 to 7.85)
[average 209 (6.68)]
This information indicates that the gas volume associated with the
production of a fixed amount of aluminum is in the range of 5 to 12 times
(average 8 times) greater for CWPB, SWPB, or HSS than for VSS. Consequently
the concentration of S02 in a volume of exhaust gas in the primary collector
system can be expected to be about 8 times greater for a vertical stud
Soderberg unit than for other units.
Reported data on uncontrolled "primary" exhaust system S02 emissions
are as follows:52
Unit
Prebake cell
Vertical stud
Soderberg cell
Source
A
B
C
A
B
C
S02 concen-
tration,
ppm
Total S02 emissions,
kg S09/Mg Al (Ib S09/ton A1)
Not reported •
Not reported 20.9 to 23.4 (41.7 to 46.8)
[average of 22.4 (44.8)]
Not reported 30 (60) [3% S in the coke]
80
200 to 300
200 (average)
5.1-12
Not reported
17.5 to 25 (35 to 50)
Not reported
-------
The trend in construction of new aluminum plants is toward prebake
systems. A major factor influencing this trend is the lower power require-
ment of the prebake cell compared with Soderberg cells.53 It is reported
that 9 of the 11 aluminum plants opened since 1960 are of the prebake type,
and 99 percent of the 324 Gg (357,000 tons) capacity added since 1973 has
been at prebake facilities.54 Of the industry's total annual capacity of
4.8 Tg (5.3 million tons) prebake units account for 68 percent, HSS plants
account for 20 percent, and VSS plants account for 12 percent.54
5.1.1.5 Molybdenum Smelters--
The raw material for a molybdenum smelter is an ore concentrate that
usually contains about 90 percent molybdenum disulfide. This is produced at
mines operated specifically for molybdenum production and also is a by-
product of copper mining and concentrating operations. The United States is
a principal world supplier of molybdenum, which is used mostly as an alloy-
ing additive to steel.55 Only a small fraction is processed to metallic
molybdenum.
The molybdenum smelting process is one of the simplest nonferrous metal
operations. Technical-grade molybdic oxide is made by roasting the concen-
trate to remove essentially all sulfur as S02, a process similar to the
roasting of zinc concentrates.56 Most molybdic oxide is sold to the steel
industry"^ this form. Some is converted into other products by processes
that do not generate S02 emissions.57
Multiple-hearth roasters are the most common type of equipment used in
this process.56 Supplemental fuel is necessary only during startup.
Because of the use of large amounts of infiltration air for gas cooling and
high excess air ratios in the roaster, the roaster off-gas contains only
about 1.3 percent S02 and cannot be used for sulfuric acid manufacture
without supplemental sulfur burners. By chemical balance, production of
4-17 kg (2.58 Ib) of molybdic oxide, equivalent to 1 kg (2.20 Ib) of ele-
mental molybdenum, creates at least 1.33 kg (2.95 Ib) of S02.
5.1.2 Control Techniques
In the nonferrous metals industry, controls are being applied to all
strong gas streams that contain about 3 percent S02 or more. Controls are
5.1-13
-------
rarely applied in this country to weak gas streams containing less than 3
percent S02. The techniques available for S02 control differ depending on
whether the stream is strong or weak.
In all the facilities that generate strong gas streams, this gas is
used as feed to a sulfuric acid plant. This control technique is used in
all zinc smelters, two molybdenum smelters, four lead smelters, and 14
copper smelters. Both single- and double-contact acid plants are in use;
Section 5.5 provides details of this control technique.
In Japan and Sweden weak gas streams are controlled by wet scrubbing.
One U.S. copper smelter attempted scrubbing a stream of marginal S02
concentration (about 2 percent) with an organic liquid, dimethylaniline
(DMA).
5.1.2.1 Description—
Copper industry-Copper smelters create strong S02 gas streams from the
operation of electric smelting "furnaces, fluidized-bed roasters, most con-
verters, and (in two modernized smelters) newer types of smelting furnaces.
The accepted industry practice for control of strong S02 gas streams is the
contact sulfuric acid plant or conversion to liquid S02. Table 5.1-1 sum-
marizes the process and S02 control equipment of the 16 U.S. copper
smelters. Except for the DMA absorption unit at the Phelps-Dodge Ajo
smelter, which did not operate properly and is now shut down, all S02 con-
trol equipment listed is applied to gas streams containing at least 3 per-
cent S02.
The DMA absorption system is a cyclic-regenerative process that incor-
porates an absorber with trays on which most of the incoming S02 is absorbed
in a countercurrent stream of DMA. . The residual S02 in the gases is
scrubbed with a weak sodium carbonate solution to give sodium sulfite or
sodium bisulfite. Liquid sulfur dioxide is recovered as a product, and its
absorbent is regenerated and recycled through the system.58 Typical S02
emissions from a DMA unit may be in the 2000- to 3000-ppm range.59 Cities
Service Company operates two DMA absorption systems rated at 36 and 50 Mg
(40 and 55 tons) of liquid S02 per day at Copperhill, Tennessee. The feed
stream for the DMA system is 7.6 percent S02, and is a mixture of off-gases
5.1-14
-------
to
LU
(—
r-
0
LU
1— t
Z
LU
I—
^^
1— H
to
LU
1—
LU
to
LU
Q-
Q.
O
0
^_
•a:
i—
LU
yr
0-
cx
LU
LU
O
o
a.
o
>_
01
I?
+J
c
.*-
3
CT
M a>
o
fc.
c
o
u
c
t icul ate
1 equipmei
fc. O
to fc-
'1
u
*m
H-
I£
tft
ton
O O)
ID
fc.
"c
IE
1
g
5
S
1
1
"o t- >
SS o
***' £ *"*
O "D
oce t=
ro ^-* «
4->
(J.
C
o
u
1
^
1-
wT*a
e o
O (J fc-
f— a> a»
u *n jo
in re "f~
to t- *J
co a. c
g
o
rH
CO
s
J_
at
4->
in
m
£
•a
T3
u_
!4J C
O O)
SI
ex
I
^^
CO
rH
^
TO
•o
tO tn
•oeo c
(0
fj at i m
tO «n •*- *-^
•- w o
">•!
u o o en
<0 H- U TC.
t.
5 t-
t, t. re 0 ^
^ •§^^ " t!
U t. 0 E fcj «
j§ E •" .^i
re a. a. mxx.
o- LU UJ a.
ii
00
gg
tO CM
ft)
(J
fc.
£•«,
II
11
at o
ai
ae co
in
"« -2
£«*- u
0 0>
4-* 4-1
fcSi
« >»a
m t— ex
K §5
O O O
CO CO CO
rH rM «-(
X x
>» ft |
% ^^ ^
>^o re re
40 "in r^ c "S c
4J C 4J O"^. O
C O C 4J ^4J
Q,O O-T"^ O O
o or^m
TJ IO *O CM CM CM
cn
c
c
fc. re
ai at
ja in in u
3 - -
fc. a- a. m
o co i/> re
CO LU LU U
X
S
CD O CD
CM CO CO
*"-'£!,;£,
r-l CO ^~
»H,ro r--
T-tCNJ
t-
in fc-
§| .
fc. in at
Hat 4->
i^t
•oSES
•»- at s o
3 DC •+- O
LU in en
O *t-
£*
*^. fc- JD
00 01 fc-
CM 4-> O)
CO tn >
^*- « a>
o fc-
m c
CM W
1
-o
u
CM ^- in
CM
CO
Hultlhearth
roasters
*
o
o •/>
tj re
ceo.
tn f—
•CLU
o"
O >
CD
CO
/•^ re
C ^j
ox:
u ^
+j ^j- re
4J C in
0—0
U Q--*->
O" T3 O
0 ^
o
chamber
CO
fc. D-
CO LU
CD
CD
CO
£*
"^
r-
CM
U
«o
fc.
**-
Reverberatory
converters
CO
^
3
u
u
01
fc.
CO
tn
CM
r*.
01
in
3
5
CT)
re
CO
Multlhearth
to
1
(A
o> t.
o_ re
4^
c o.
m -r-
u
re CL
s g
o o
CD
° °
O O
CD CD
CD CD
CD CD
co *r
co r^
co ^r
4 en
roasters
) Hulti hearth
roasters
Reverberatory
f— 1 CM
i
O
U
m
H-
•o
tn
o
4->
g
CM
V^ CM
O
X CO
re O) 3
a1 •-
5§ ^
«c S
in in
fc. •—
i l
re in 4^
u o> o
JO U
>> in jQ
fc. O_ fc. m
O.CO U *>-
to LU co Z
CM CM CD CM
^
O
CD
1
O
"""
1
o
furnaces
Converters
«*
o
o
S
^"'
tn
i_
|
c
0
*J
O
m
r^.
^™^
4J •§ ^*
re O) c
CD O
TJ CO C
j_
o
fc. 4^
"I fc fc i—
C1 "5 0"5 -r- t-
t- *— JD O.
i> a. fc. t_> fc-a. fc- vi
a) co a. o-co u •£
COLUtnCMCOLUCO£
I 1
o o
s a
* I
2 Multlhearth
roasters
Reverberatory
TH
-------
•Q "*- u t, +j <:
v,. nj a, .„ Q_
in a. c -a t- LU
-
O UJ 3 3
tO •— O> C 3 6-
•O Q.J5 O I- .O -4J
•v, (- L> *J 41) (-
D1"P T3 C CL
O Q* C O **- OJ
O ID t- (D O O t-
o
in ^
i-H O
tn i—i
in o 1- r-.
r-t t- i-H
rq T3
i
•*•* ID C\J ^-*
t— O > C
!! S
•a
i-H fc. r>. *—'CM v.^s^i-H v
I
in
LLJ
ca
i- ei a) *J
— \n t- i
X O LLJ c_)
CXI ^- *»x
u> O)
S-8
5.1-16
-------
£ O> O
£ J= U -C T) O) •*-> O
i !S-~a
""S- tl-g —
OS- U» « -«J T3 E
5 1 ID i^ t. , VD
- a. o>« ** *i «
i
.o m
o 'o.
O)« *^«
E E E
=> Z «*-
(0 O ID i/»
m •— •
R)
§4->
o
C c:
(O TO
a. a.
.o ai
^1 J3
- - *
UJ UJ CO LU
* u
O V)
•a
(U
o
u
£
S fc-C e
t. 13 3 <-
u at tf>
« o t-
C (O QJ
^-S 5^!
5m cn o o
m ret_> oe:
CX LU CO CO C^J *
LT)
UJ
CQ
-------
from electric furnaces and fluidized-bed roasters. The absorbers could be
operated to achieve a concentration as low as 1500 ppm S02 in the treated
exit gas.6' ASARCO, Inc., at Tacoma, Washington, uses DMA absorption to
recover S02 from smelter off-gases that have (or can be upgraded to have)
S02 concentrations in the range of 4 to 10 percent." The system is rated
at 180 Mg (200 tons) of liquid S02 per day.
Although no estimates are reported, a high percentage of the S02 gener-
ated by the industry is now being converted into sulfuric acid. Of the 16
domestic copper smelters of all types, 14 are equipped with one or more acid
plants.^ Total acid plant capacity is over 6.1 Tg (6.7 million tons) annu-
ally.61
Typical S02 emissions from a single-contact sulfuric acid plant oper-
ating on a strong S02 stream from a copper smelter may be in the 2700-ppm
range.6* Typical S02 emissions from a double-contact sulfuric acid plant
operating on a strong S02 stream from a copper smelter may be expected to be
in the 400- to 500-ppm range.63
Weak S02 streams are created from the operation of all reverberatory
furnaces, some converters, and most multiple-hearth roasters in the United
States. The acid plant tail gas also contains unreacted S02, and fugitive
ennssions from converters and reverberatory furnaces create a dilute S02
stream in the ventilation air of buildings. No control of S02 emissions
from weak streams is practiced in this country.
Use of the S02 from these weak S02 gas stream sources for acid manu-
facture is limited by the technical requirements of a conventional acid
plant. For efficient performance, the gas delivered to the plant must be
consistent in volume and S02 content and must consistently contain from 4 to
9 percent S02.2
In foreign smelters, control of weak stream S02 is practiced. Table
5.1-2 summarizes the process and control equipment used by 8 of the 14
copper smelters in Japan. This table indicates that weak-streams, primarily
from acid plant tail gases, are scrubbed with both lime and caustic solu-
tions in nonregenerable FGD equipment.6* m addition, the Onahama Smelting
and Refining Company of Japan used a regenerate magnesium oxide (MgO)
scrubbing system and still uses a nonregenerable lime scrubbing system to
treat exhaust gases from a reverberatory furnace containing 2 to 3 percent
5.1-18
-------
O-
_l
o
O
o
co
on
o
to
CJ
CNJ
c
CO
a
•a
5
i. -
3 U
i— 3
tn o.
u
Ul t-
re ai
•r- U
re u
*J in
re *>
en c
0,1
U 4J
re re
c a>
i. S.
3 4->
U_
in
i
1*- N
t- o
o >
in
in
a>
Smelting proc
- s-
s-^1-^
o, •»- en
a. u E
Q, re
0 0.0
c_> re c=>
u o
r-t
i-
01
0)
to
t-H
CT>
C
Q
"a.
•a
u
re
r—
lash furnace
-S converters
U_ Q-
O
Tl-
3
re
^.
t- 0
I!
1 *
r--
01 Ol
c: c
0 0
z z
c c
re re
0. 0.
-a -o
re re
r- r—
£3 Ul
CSJ
r-i ir;
1 1
C3 «»•
r-1
O
lash furnace,
enriched
-S converters
U_ Q-
0
in
r-4
0
a. •
a. o
z
en
c
U OJ
re E
*J to
•E
en
>> >>
(- (.
3 3
in in
<_> u
c c
re a
a. a.
T3T3
re re
££
i i
OO »!•
lash furnace
-S converters
U- Q.
0
§ •
0 0
a u
en
.. c
re ••-
.* c
&
OO
CTl
a>
c
o
z
c
re
a.
•a
re
t—
0
tH JO U
r-4
0)
CJ
re
c
u t-
01 3
luid-bed roast
everberatory 1
-S converters
u. ce o_
o
Lf)
r-l
•p- d
-C O
in
re J3 O)
E 3 *>
!c •n'at
in -f- E
O Z to
re
>i
t-
tn
O
re
u
c
re
"a.
•a
u
re
i —
to
o
in
o
c
-furnace conti
o
tn
in
I
U)
5?
CM
en
o
3!
o
in
n:
o
re
z
- O CM
s- +J-^
0) rH
-Q O 4->
"^^
CM
CM
a.
u
re
c
3
u
leverberatory 1
01
O
CM
re
re
J^ TD •
re c o
c re o
o
en Cn
C 1=
re+J c
re 'cu **-
.C E O)
re to oc
o
>>
!_
3
in
O
4J
c
re
"a.
T3
re
to
i
p-
'-S converters
t~-
c
o
•_£
3
O, O
c in
* 1
-p •»->
c c
re re
"a. "a.
•o TS
O U
re re
Q to
CM
o «t
N
en
:lash furnace 1
enriched
>-S converters
o
m
rH
.
en
•r-
e •-
01 c
m o
O Q-
C.Q-
§1Z
re
to
en
en
01
c
0
.p
re
Q.
•a
re
[
o
00
i re
:lash furnace
3-S converters
o
CO
o
o
*z
•r- O
.n o
c
•r"
d i—
C HI
E i/1
t—
en
en
c
O
'•£>
3
O
in
3:
o
re
z
+j
c:
re
O.
•a
re
V)
rt "
:lash furnace
P-S converters
C3
CM
rH
d
0
O
0 C
•i- C
3Z
re
o
*—
to
II
E
to
O)
u
re
01
• at
V) U
re re
en c
•r- Ol 3
a, u *t-
o. re in
c in re
u 3 en
3 o
M- 3 <-
c ai
II
to
I
in •*-> in
re c re
r— O O
t- u t-
r— 13 -a T3
f) a, cj o,
3 c c c
O •*— '•- T-
o -a-c^
11 c3
-------
SOj,.6* The MgO scrubbing system produced a strong S02 stream, which was
subsequently used to manufacture sulfuric acid. The lime scrubbing system
produces gypsum as a byproduct." The domestic copper industry has not yet
adopted control of weak S02 streams (i.e., <2 percent) by nonregenerable
liquid scrubbing.
Revision of production methods may be an economically valid method of
S02 control since adoption of newer equipment also usually leads to signif-
icant reductions in energy consumption and production costs. A principal
source of S02 is the reverberatory furnace, which has poor thermal effi-
ciency. Reverberatory furnaces consume about 90 percent of all energy used
in copper smelting4 and operate at less than 25 percent efficiency.66
It is possible to replace the reverberatory furnaces with other types
of smelting equipment and thereby to upgrade the concentration of S02 in the
furnace off-gases. Alternative processes are the flash smelter, the Noranda
furnace, and several other systems used abroad.6? All of these claim or
have demonstrated consistent production of off-gas suitable for acid plant
feed, and energy consumption from 50 to 70 percent of that used by the
reverberatory furnace. Oxygen enrichment will further lower' fuel require-
ment." None, however, is adaptable to all grades and types of concen-
trates, and most produce slag that contains too much copper to be discarded
without further treatment.
The Noranda and similar continuous furnace designs not only permit
simplified control of smelting furnace off-gas, but also minimize fugitive
emissions of S02 from copper converters. As used in the United States, the
Noranda furnace produces a high-grade matte (>60 percent copper), which
requires the use of a converter, although with much-reduced potential for
fugitive losses.69
The size of these advanced copper manufacturing systems almost always
precludes retrofit at an existing smelter. It usually is necessary to build
new buildings and relocate auxiliary equipment such as cranes. Because more
S02 is captured, the new acid plants usually must be added and markets for
the additional acid must be found. Auxiliary slag cleaning plants are also
needed. Two U.S. primary copper producers have elected to do this and have
achieved substantial reductions in S02 emissions.
5.1-20
-------
Replacement of Peirce-Smith converters with the Hoboken type is also
possible, although the extra space required complicates retrofitting.
Operation of the Hoboken converter calls for special auxiliary equipment and
different operating procedures. Properly operated and equipped with neces-
sary hoods, fans, and process controls, a Hoboken converter can prevent the
escape of more fugitive S02 emissions than a Peirce-Smith converter.70
Because fugitives are weak S02 emissions and off-gas is a strong S02 stream,
the Hoboken converter can permit more complete control of overall emissions.
The ultimate process change is abandonment^ of the copper smelter
through the adoption of hydrometallurgy.71 In hydrometallurgical systems,
chemicals leach sulfide copper ores to extract the copper content, which is
then electrochemically converted into metal. The sulfur in the ore is
changed into either elemental sulfur or the sulfate ion, neither of which is
an atmospheric pollutant. Emission of S02 is completely eliminated.
Interest in copper hydrometallurgy appears to be declining, however, because
these processes consume more energy than pyrometallurgical methods and also
present formidable problems of solid and liquid waste disposal.72
Lead industry—The principle source of S02 emissions in the lead indus-
try is the sinter machine in which the lead sulfide concentrate (PbS) is
converted into an oxide or sulfate form. During sintering approximately 85
percent of the concentrate sulfur is removed as S02.73 A hood is generally
installed over the sinter machine to capture the S02 emissions. The emis-
sions may be captured in separate effluent streams (one with strong S02
concentration from the feed end and one with weaker S02 concentration from
the discharge end) or in a combined stream; or only . the strong off-gas
stream from the feeder end may be captured. One control technique that
concentrates essentially all of the S02 emissions into a single stream
strong enough for sulfuric acid manufacture is the use of the updraft-type
sinter machine with a weak S02 gas stream recycle system. In this system,
air is passed twice through the sinter machine. Combustion air is drawn
from a hood near the sinter breaking equipment, thereby capturing fugitive
S02 emissions from this source. The air is first passed through the back
end of the machine, forming a weak S02 gas stream. The stream is captured
. by a high-temperature fan and recycled through ductwork to the front end of
the machine. The strong gas stream thus produced contains all of the S02
5.1-21
-------
generated and is strong enough for manufacture of sulfuric acid This
system is installed at one U.S. lead smelter and has been used at foreign
smelters for many years." It has been reported that the use of gas
recirculation will decrease the production capacity of an updraft sintering
machine compared with a similar machine without gas recirculation."
The alternative to this system is-liquid scrubbing, which is the tech-
mque applicable to acid plant tail gas and blast furnace off-gases The
Cominco ammonia absorption process has been used with very good recovery to
treat process off-gases having S02 concentrations as low as 0.5 percent In
tins process, the, S02-laden off-gases are treated by an ammonium sulfite
solut10n ln a countercurrent manner. The ammonium bisulfite solution
generated in the absorber is acidified with sulfuric acid to produce a
concentrated (25 percent) S02 gas stream. The dilute ammonium sulfate
stream 1S treated in an evaporator to produce a granular product, which is
used as fertilizer.™ Cominco has operated ammonium sulfite-bisulfite
scrubbing systems at their smelter located at Trail, B.C., Canada, since the
1930 s. The ammonia scrubbing systems serve Dwight-Lloyd lead sintering
machines and zinc roasters.™ The S02 concentration of inlet gases to the
Dw,ght-Lloyd lead sintering machine absorption system ranges from 0.3 to 2 5
percent; however, fumes cause major visible emission problems at this
installation.™ Some Japanese lead smelters scrub blast furnace off-gas and
weak streams from the sinter machine; lime and sodium compounds are in
use. Recently a study has been made to define alternatives applicable to
control S02 emissions from a lead smelter and zinc plant of the Bunker Hill
Company in Kellogg, Idaho." The study concludes that three FGD processes
(aluminum sulfate-DOWA, citrate, and ammonium sulfate) are technically
feasible to treat lead smelter sinter machine weak or strong stream off-
gases.78
Zinc industrv-The domestic zinc industry has demonstrated an accept-
able S02 control technique, and equipment that can give adequate control is
installed at all U.S. zinc smelters. The control technique employed is use
of the S02 for manufacture of sulfuric acid. The most profitable smelter
operations remove essentially all sulfur from ore concentrate by roasting
and use an energy-efficient type of roasting equipment. The waste gas
5.1-22
-------
stream thus produced is ideal for use in acid manufacture.79 Another option
is the manufacture of liquid S02 if a market for this material is available.
The largest emission of S02 from a zinc smelter should be from the tail
gas of the sulfuric acid plant. If the plant is properly designed and
operated, it should remove about 99 percent of the S02 fed to it and the
exit gases should contain no more than 0.05 percent S02 (as indicated in
Section 5.5).
Fugitive emissions of S02 may result from leaks in roasters and from
equipment that transports calcine. These emissions can be controlled by
proper maintenance and by recycling the polluted air to the roaster as part
of the combustion air.
Pyrometallurgical plants will probably continue to emit very small
amounts of S02 in sinter machine off-gas. These plants must avoid over-
roasting, which causes loss of zinc in subsequent processing steps.29 With
adequate process controls the off-gas should contain much less than 2 per-
cent of the total sulfur content. Sodium, lime, and zinc compounds are used
to scrub sinter machine off-gas at some Japanese zinc smelters80 but not in
this country.
Aluminum industry—The most significant air pollutants emitted by
aluminum smelters are fluorides and particulate matter. In general, S02
emissions have been considered relatively insignificant,81 but two control
techniques can be applied if control is required. With all process designs,
reducing the sulfur content of the anode coke will result in reduced sulfur
oxide generation. In addition, flue gas desulfurization (FGD) has been
demonstrated on VSS-type processes.
The sulfur content of petroleum coke is related to the quality of the
crude oil from which it is produced. Good quality feedstocks for anode coke
include thermal tar, cat craker slurry, decanted oil, and coal tar pitch.
Poor feedstocks include vacuum residuals and derivatives of high-sulfur
crudes.46 In 1979, it was projected that supplies of low-sulfur anode coke
derived largely from doemstic crude were expected to be available for the
next 5 to 10 years. Future supplies of low-sulfur anode coke will depend
primarily on the availability of low-sulfur foreign crude. Although low-
sulfur coke may be currently available, one petroleum coke manufacturer
5.1-23
-------
reported in 1978 that low-sulfur coke commanded a price four to five times
that of high-sulfur coke.47
Essentially all the S02 emissions from the Soderberg process and over
90 percent of the S02 emissions from the prebake process occur in the reduc-
tion eel 1.52 it has been estimated tnat pr.mary cell collection systems can
capture 80 to 95 percent of the cell gases.** The remainder escapes through
roof vents; or a portion may be collected by secondary controls. Therefore
it may be estimated that 70 to 95 percent of all the S02 emitted in aluminum
production exits from the process through the primary collection system
Older emission control systems in the aluminum industry consist of
water scrubbers designed for simultaneous removal of fluorides and particu-
late matter from the primary exhaust system. Although designed to capture
fluorides, these scrubbers also absorbed some S02, which was discharged with
the scrubber overflow as a wastewater constituent. One such wet scrubber
followed by a wet ESP at a VSS plant was reported to have an S02 control
efficiency of 70 percent.83
At newer aluminum plants, wet scrubbers are being replaced by dry
scrubbers, which use alumina as an adsorbent. Dry systems have the advan-
tage of adsorbing gaseous fluorides and mechanically capturing particulate
matter. These systems, however, are ineffective in controlling S02 because
any S02 captured is eventually released through the primary or secondary
exhaust systems.
Dry scrubbing systems allow the captured fluorides to be returned to
the cryolite bath without further processing. At most aluminum plants using
dry scrubbers, all of the alumina fed to the potline is first routed through
the scrubbing system. In addition to removing fluorides, the alumina also
adsorbs some S02; however, adsorbed S02 is immediately re-emitted when the
alumina reaches the reduction cell. Most of the S02 is once again removed
from the hooded cell and exhausted through the primary collection system
and the remainder escapes into the cell room.
Because of the lower gas flow rates in the primary collection system
and higher S02 concentrations, for the use of anode coke of a given sulfur
content, FGD is more feasible for VSS processes than HSS or prebake proc-
esses. Because of greater primary exhaust volumes and the resultant dilute
5.1-24
-------
S02 concentrations, HSS and prebake plants would require much larger FGD
capacities than a VSS plant.84 The VSS off-gases contain S02 concentrations
o o
6 to 10 times higher than those associated with other processes.
A sodium-based FGD system has been installed downstream of the dry
fluoride/particulate matter scrubber at each of the Martin Marietta Aluminum
Company's VSS plants in The Dalles, Oregon, and Goldendale, Washington. The
installation is in response to a determination of Best Available Control
Technology (BACT) as required by Prevention of Significant Deterioration
(PSD) regulations, and operating permits require 70 percent S02 removal to
achieve an emission limitation of 9.5 kg S02/Mg of aluminum produced (19 Ib
S02/ton of aluminum). The supplier of the control systems designed and
guaranteed the adsorbers for 90 percent removal (450 ppm S02 inlet, 45 ppm
S02 outlet). The S02 inlet concentration may vary from 150 ppm S02 to 450
ppm S02 with an average expected of 300 ppm S02. The systems are designed
to treat 94 mVs (200,000 cfm) of gas at 121°C (250°F). The cost was
approximately $2 million per system or $21,300/m3 per s ($10/scfm).85 The
application of FGD to HSS or prebake processes has not been demonstrated.
Sulfur oxides not collected by the primary system escape through the
cell room ventilator. Several aluminum smelters are now equipped with
secondary scrubbers to treat the ventilation air, but these are typically
water spray screens designed to capture fugitive fluoride emissions.86 The
application of FGD to these weak S02 gas streams has not been demonstrated.
Molybdenum industry—The roasting of molybdenum concentrate has been
examined less intensively than the processes of other nonferrous industries.
Molybdenum is made in much smaller quantity; the industry generates less
than 2 percent of the amount of S02 generated by the copper industry.
No reports describe potential process modifications to permit the use
of roaster off-gas for manufacture of sulfuric acid. Both the copper and
zinc industries have replaced multiple-hearth roasters with other equipment
that not only creates high-concentration S02 streams but also reduces energy
consumption. This may also have been accomplished with the aid of supple-
mental sulfur burners at the two molybdenum smelters that sell sulfuric
acid, although no published information is available.
5.1-25
-------
A recent study of one of the larger molybdenum roasters indicates the
possible use of a lime-based, nonregenerable S02 scrubber to control these
gases. Union Carbide and Dm/all employ byproduct colloidal lime slurry to
remove S02 from the off-gas of a small molybdenum roaster in Bishop,
California.*? A Hme slurry sulfur diox1de scrubbing system 1s glso ^.^
used to control off-gases from.two multihearth roasters processing molybde-
num copper ore at the Duval Sierrita Company processing plant near Tucson,
Arizona.** Unless the production process used in most molybdenum smelters
is modified, nonregenerable scrubbing may be the only feasible control
method.
5.1.2.2 Control Cost-
Cost curves have been generated for the capital and annual operating
costs of double-contact/double-absorption sulfuric acid plants.89 The
capital costs are presented in Figure 5.1-3, and annual operating costs are
presented in Figure 5.1-4. Details of the estimated control costs appli-
cable to electric smelting, flash smelting, and reverberatory smelting are
also available.90
A recent study sponsored by the EPA has calculated capital and annual
costs of installing wet scrubbers on weak S02 gas streams from a "typical"
reverberatory furnace and copper smelter converters.9* No systems of this
type are currently in use in the United States. The costs of two nonre-
generable scrubbing systems (lime and limestone) and two regenerate systems
(magnesium oxide and sodium citrate) were calculated for several streams,
two of which are
98,900 NmVh (58,200 scfm) at 1 percent S02
76,500 NnrVh (45,000 scfm) at 1.4 percent S02
A summary of the cost data adjusted to mid-1979 is presented in Table 5.1-3.
Data on regenerate systems are shown as two separate costs: one for the
scrubbing units and one for the S02 drying and liquefication equipment.
Detailed estimates of capital and annual costs for these FGD systems treat-
ing different off-gas flow rates are available.91
Actual costs would be somewhat higher than those shown, since actual
gas streams from the principal uncontrolled sources (the reverberatory
furnaces) result from cyclic operations and therefore vary widely both in
5.1-26
-------
10C
in
s-
(O
o
-O
O1
t/o
c/5
o
4.25% S02
5% SO
6% SO
8% SO
:
10'
Q-
-------
10C
1| " 1 1 I I I I I I
to
o
•o
CO
CO
c
CD
t—<
H-
Q.
o
I
10'
10C
9
(20)
14 '
(30)
1
1
19
(40)
24
(50)
47
(100)
SMELTER GAS FLOW RATE, Nm 3/s (1C3 scfm)
4.25% S0
S0
S0
S0
1
71
(150)
Figure 5.1-4. Annual operating costs of double-contact sulfuric
acid plants - dilute feed gas.89
94
(200)
5.1-28
-------
a:
o
u_
"3
CO OO
I— C*T
CO LU
O I—
CJ _I
UJ
_j s:
- •£-.
Of LU
•=C h-
21 CO
2: >•
^3 CO
CO
Q
CD
. U-
00
I CO
m i— <
cz
UJ >=C
_l >
CO
o
IN
O
CO
•w- **•
•. I— ^
in
tn
O
O
"re
13
c
tn
in
Q
C3
u_
0000
o o o o
«3- en oo o
in oo >4- to
o in f^ o
oo m oo oo
«^ oo oo •—
O CD O O
o o o o
to en CM oo
m r~- en ^
to to m CM
ro to o en
•si" OO «*• r-
o o o o
O CD O O
co ^" r*~ ' —
#. rt •» **
i — CM tD t£5
CM oo m «^
0 r- • i—^ ^
to co • en ^~
CD O O CD
O CD CD O
r~. co oo «^-
o^ oo r~» •—
CM i — crv m
to in oo r-
to oo en «*
Ol
•a
X
o
0) E
C 3
O *r- CJ
+J in +->
in o) ro
O) CD C S-
E E 05 +J
•r- -r- TO •<-
_j —i s: cj
CD CD
0 0
oo to
O CM
CD in
I—
CD CD
0 O
OO 00
in oo
OO i—
en in
0 0
0 0
CO •—
** **
in o
r- en
r—
0 0
CD O
o oo
00 00
^" P"**
0 OO
CO t>
+J *r~
in x
>> , o
in
E=
CM 3
o •<-
in
rH
°\ §
CL
re
•o
-------
volume and S02 content. The cost estimates shown in Table 5.1-3 are based
on an assumption of steady-state operation.
Lead and zinc industries—Most of the S02 controls used by lead and
zinc industries are directed toward sulfuric acid manufacture. The cost
curves presented earlier for sulfuric acid plants are applicable to S02
control of lead and zinc smelting operations.
The cost requirements of various S02 control alternatives for model
lead and zinc smelting facilities were also estimated.92'93 Estimated
capital and annual costs (adjusted to mid-1979) of a DMA scrubbing system
and dual-stage sulfuric acid plant with neutralization applicable to conven-
tional sintering machine operations are $42.7 million and $11.2 million.92
The capacity of the DMA scrubbing system is 85,000 Nm3/h (50,000 scfm) of
off-gas. The estimated capital and annual costs (adjusted to mid-1979) of a
DMA scrubbing system used in conjunction with a dual-stage sulfuric acid
plant and neutralization applicable to roaster/sinter zinc smelting opera-
tions are $50.4 million and $12.4 million.93 NO DMA scrubbing system is
currently in use at any domestic lead or zinc smelter. Details on cost
estimates for various control alternatives are available.92,93
A recent study has been made to determine the technical feasibility of
various alternatives that could be applied to control S02 emissions from the
lead smelter sinter machine, the lead smelter acid plant, and the zinc
smelter acid plant of Bunker Hill Company in Kellogg, Idaho.78
Table 5.1-4 provides the estimated capital and operating costs of six
different FGD systems [lime/limestone, double-alkali, citrate, ammonium
sulfate, zinc oxide, and aluminum sulfate (DOWA)] applicable to various
options.94 Cost information was not available for the Wellman-Lord process.
The costs were determined on a preliminary basis from information provided
by various FGD system suppliers and designers.9« Capital costs include
fabricated equipment, engineering, and installation. Annual operating costs
include both direct costs such as chemical reagents, operating labor, utili-
ties, etc., and indirect costs such as depreciation, taxes, and plant over-
head.
5.1.2.3 Environmental and Energy Impacts--
A general discussion is presented here about environmental and energy
impacts of S02 control techniques applicable to nonferrous primary smelters.
5.1-30
-------
_J
«=c
o
LU
Oi
cS1
CO
D:
o
1 1
CO
o
HH
fl
°S
U_ 0
o rr:
CO O
• i .
r— ' '
CO
O "
0 03
CD
O O
(— i — 1
g£
UJ •>- — -
o uj o
i i
«=C LU X
ZD ^.
•z. co-fa^
Q r-H
^*" "T*
tf— — 1—
_l UJ
F*™ ^~
I—* ^3
D- CO
tJ t rTT
0
UJ
g
i — i
CO
UJ
•
_
.
LO
LU
i
CD
<£.
1—
u
_
o
Q.
o
u
ro
o
.*-»
O.
0
0
CXJ
O
Q.
0
-0
rH
C
0
4->
Q-
O
OJ
u **- •*-> ~o
c m o 0-*
M- 3 M- 4->
PS* . — E «/> O in
Q. TO -r- =J
.c +j i- JC in X:
*J C +> U *H X
E ^ "* IE >> *
o ex m X ' — QJ
u in (U E
•*- TD TO in -*-> 15
V> U >> O> E 0
TO TO JO •*- >
O5 CT X
+J ^? C O r—
t— c tn o s- TO
"Jo i— -*-J Q. o
j— Q. TO in ro •*->
QJ
O CD
4— W •*->
ro TO ^~
O) QJ Q-
r— TJ
TO -C U
r- *J TO
C
O
Ov—
QJ C •*->
C *f~ LJ
••- J- C
.C 3 -D
E "D U "D **— 4->
O C TO I- r— C
t_ QJ E QJ ^ TO TO
M- -4-> TO E •—
CJ S- C CT. Q-
in Dl QJ *i— 1 "O
ro i- •+-* in **— C "D
OJ ro c M- ro ••-
1 J= ••- C7) 0 U
»4- U l/) C C TO
i*- in O QJ S
O ••— OJ S- C O QJ
TO QJ U 3
QJ _c M- s- ro -c *+-
3c +J o o E m o
tu
*~ Q^ W C
TO S- "O JT
01 TO ro u
t £" QJ TO
•*- U •— E
«*- tn
O ••— QJ t-
•O .C QJ
TO QJ C
^ *j O
CD CD O O O
CD O CD O CxJ rH
O O O O ^ 1
^•ijDCNJUDO •• . .-* , .f^
CT> LO^C\jCr\ rHCD OCQCD r-~
O CD
O O
*fr C\J
in ro •— **
I I O ^*
CD O CD O ^J"
O O O CD rH O
°°.°° ^ CSji CM
SKS3 S dd C^r^OrH
r—
ro -i-3
O C3 E CD
O CO t- l/l
o o o a.
LO ro i i
LO CO *^^ s~^
1 i CD O O
CD CD O CD t~- m
O O CD O CO CM ID
»^»^CDCD r— LO rHrH
f--COCSjro cncM .- ,1, °. J^ «~
^CNJCXJrH rHrH OO CDCSJOt^
CZ> CD
O CD
O CD *~*
. - O
O CD P^
LO ro CO
1 1 s_^ CT1
O O O O •
°°°° | r-LO ' rH rH
ScSSrH1 dd oSir-
s s
^ "o "o
O Su^ N~^
. — "m ai c_> c • . c
•*J E o> o o o
-
* o m - -4-> +•>
• C t- 0) -f- •!-
1- in in
a, — , — =o o
*J *J Q. i— O-
US-CUN o
§B o E U O- O TO O OJO C"J
Zinzm 6 M cs- in <_> O t_> 3=
i— ai O w .
u_ t— m O
a>
^
c
•r-
o
o
5.1-31
-------
u
•3-
in «s-
O *3-
m o
•*• CM
i-H f-H
oi rr>
Q.
O
m oo
tsj co
•a
O>
C
O
CM CM
f- O
oo oo
in t—i
CMt-i
CM
ID CM
•o
(U
C
C
O
u
in
LU
co
5?
a
CD
CU i—
E ra •—
:c£ §
•N. Q.C
eu CD c
E O «E
Q. C
ID C
*J
o
Q. cz
<—
.tf 5
a
a.
• • o
C
01
u
o
00
i— O
ro *J
> u
o ITJ
E x-
a>
f- 01
CZ Q. O!
ai o
£- °.
01 ID CM
d = I
c m
o cz
oo ro o
E a>
•!- U
CZ (.
•r— OJ
E Q.
01
CM
cz>
•a
ro
o
T3 •
ai
T) I— r-
ro V> Z3 Z3
C U u
VI O -^ -r-
+J •>- 4J +J
VI *J t. t.
o a. (D (o
CJ O Q. £L
5.1-32
-------
Environmental impact—For the primary nonferrous smelting industry, two
distinct potential sources of secondary pollution can be identified.94 The
first type results from the application of sulfuric acid plants to strong
sulfur dioxide smelter off-gases; if the sulfuric acid cannot be marketed,
it must be neutralized with the resulting production of a solid waste. The
other type is associated with application of scrubbing techniques to weak
gas streams, which create wastewater or waste solids. It is anticipated
that the production and possible disposal of solid wastes, elemental sulfur,
and liquid S02 will not produce secondary pollution if adequate safeguards
are taken.95
Neutralization of abatement-derived sulfuric acid may produce both land
pollution and water pollution if a marketable application for the neutra-
lized acid sludge (gypsum) cannot be found. Although no purposeful neutra-
lization is now practiced, some leaching of waste rock dumps is probably
intended at least in part to dispose of excess acid. Ponding has been a
technique favored by a number of industries for waste disposal of this type
of sludge. Procedures and techniques are available that will prevent the
water pollution problems of neutralization and subsequent ponding from
occurring:
0 Proper site selection and location of waste disposal ponds
0 The use of impermeable pond liners
0 Closed-loop operation of ponds to prevent pond liquor overflow
A "dry" sulfuric acid neutralization process shows promise for pro-
ducing minimal secondary pollution problems.95
The purge or spent scrubber solutions from the ammonia-based, sodium-
based, dimethyl aniline, and calcium-based scrubbing systems, if directly
discharged to a local water course, could produce water pollution. The
possible forms of water pollution include chemical oxygen demand, dissolved
solids, increased organic content, soluble salt content, and increased water
hardness. Methods to solve the problems' of water pollution are available
and have been demonstrated for both the ammonia- and sodium-based scrubbing
systems. Practical disposal methods for the dimethylani1ine purge are also
available. Closed-loop effluents or water treatment facilities will, in
5.1-33
-------
most situations, be required for the spent calcium-based scrubber solution;
solid waste pollution is a possible result of this scrubbing technique.9* '
Energy impacts96-Large amounts of energy are required in the execution
of most conventional metallurgical extraction processes. The production of
aluminum, for example, is one of the most energy intensive metallurgical
extraction processes and requires energy in the range of 162,900-209 400
kJ/kg (70,000-90,000 Btu/lb) of aluminum.96 The energy requirements a'5S0.
dated with the production of copper, lead, and zinc are considerably less
than this, although still significant.
It is possible to compare the energy requirements of new nonferrous
smelting technologies having S02 emission controls with the energy require-
ments of conventional domestic nonferrous smelting technology not having
emission controls." Because the energy impact depends on the type of
smelting used, a more accurate comparison would relate the overall produc-
tion of a unit of nonferrous metal with and without S02 controls.
In copper smelting, the energy effect attributable to S02 control' of
gas from conventional reverberatory furnace smelting would result in an
increase from 20,900-34,900 kJ/kg (9,000-15,000 Btu/lb) of copper to
23,300-37,200 kJ/kg (10,000-16,000 Btu/lb) of copper, with or without neu-
tralization of the sulfuric acid produced. This represents an energy
increase of 5 to 10 percent.96
With regard to zinc and lead smelters, complete S02 control would
result in an increase of 1 to 5 percent in the energy requirements asso-
ciated with the production of zinc and an increase of 5 to 10 percent in the
energy requirements associated with production of lead.96 These are
increases over the energy requirements for production of zinc and lead by
conventional domestic technology without emission controls.
Although double-absorption sulfuric acid plants require about 15 per-
cent more energy to operate than single-absorption sulfuric acid plants the
incremental impact in terms of the increase in the overall energy require-
ments associated with the production of a unit of copper, zinc, or lead is
only about 2 percent. The production of elemental sulfur requires from 4
to 6 times the energy required for the production of sulfuric acid per unit
weight of copper or lead, and from 8 to 12 times the energy required for the
production of sulfuric acid per unit weight of zinc.96
5.1-34
-------
Dimethylaniline scrubbing, for example, requires only about 121 kWh/Mg
(110 kWh/ton) of sulfur dioxide recovered when treating off-gases of 5
percent sulfur dioxide.96 When used to treat off-gas streams similar to
those that could be treated by sulfuric acid plants, DMA scrubbing requires
about the same amount of energy as a sulfuric acid plant. When DMA is used
to treat off-gases with low concentrations of sulfur dioxide, however, the
energy requirement per unit of sulfur dioxide recovered increases by as much
as an order of magnitude.
5.1-35
-------
1.
2.
3.
4.
5.
6.
7.
8.
9.
REFERENCES FOR SECTION 5.1
Pacific Environmental Services, Inc. Feasibility of Primary Copper
Smelter Weak S02 Stream Control. U.S. Environmental Protection Agency,
Industrial Environmental Research Laboratory. Cincinnati, Oh. EPA-
600/2-80-152. June 1980. pp. 16 to 18.
U.S. Environmental Protection Agency. Environmental Considerations of
Selected Energy Conserving Manufacturing Process Options: Vol. XIV.
Primary Copper Industry Report. EPA-600/7-76-034n. December 1976.
pp. 24-27.
U.S. Environmental Protection Agency, Office of Air Quality Planning
and Standards. Draft of Standards Support and Environmental Impact
Statement. Volume I: Proposed National Emission Standards for Arsenic
Emissions from Primary Copper Smelters. June 1978. p. 3-7.
Encyclopedia
Publishers,
147-150.
of Chemical Technology. Volume 6. New York,
a division of John Wiley and Sons, Inc.
Interscience
1967. pp.
Craig, A.B., et al. Present and Future Control of Fugitive Emissions
in the Primary Nonferrous Metal Industry. (Presented at EPA Symposium
on Control of Particulate Emissions in the Primary Nonferrous Metals
Industries. Monterey. March 1979.) pp. 4, 7.
U.S. Environmental Protection Agency. Background Information for New
Source Performance Standards: Primary Copper, Zinc, and Lead Smelters.
Volume I, Proposed Standards. EPA-450/2-74-002a. October 1974. p.
3-12.
Ref. 3, p. 4-2.
Ref. 5, p. 6.
Ref. 4, p. 148, 149.
10. U.S. Environmental Protection Agency, Industrial Environmental Research
Laboratory. Environmental Assessment: Primary Copper, Lead, and Zinc.
(Prepublication copy). Cincinnati. November 1, 1978. p. 65.
11. Ref. 6, p. 3-3.
12. Ref. 6, p. 3-5.
13. Ref. 1, p. 45.
5.1-36
-------
14 U.S. Environmental Protection Agency, Industrial Environmental Research
Laboratory. Control of Copper Smelting Fugitive Emissions. EPA-600/
2-80-079. May 1980. pp. 36, 39, 45.
15. Ref. 14, pp. 12, 24.
16. Ref. 14, p. 43.
17. Ref. 2, p. 44,
18 Bailey, J.B.W., et al. Oxygen Smelting in the Noranda Process.
(Presented at 104th AIME Annual Meeting, New York, February 16-20,
1975.) pp. 2,3.
19. Dayton, S. Utah Copper and the $280 Million Investment in Clean Air.
Engineering and Mining Journal. April 1979. p. 73-77.
20. Ref. 19, p. 78.
21. Ref. 2, p. 62.
22. Ref. 19, pp. 79-81.
23. Halley, J.H., and B.E. McNay. Current Smelting Systems and Their
Relation to Air Pollution. San Francisco, Arthur McKee and Company.
September 1970. p. 6.
24. Ref. 14, pp. 12-14, 17-18.
25. Ref. 6, p. 3-80.
26. Ref. 22, p. 25.
27. Ref. 10, p. 289.
28. U.S. Environmental Protection Agency. Compilation of Air Pollution
Emission Factors. 3d ed. (including Supplements 1-9). AP-42, 1977.
p. 7.6-4.
29. Ref. 6, p. 3-131.
30. Ref. 6, p. 3-175.
31. Ref. 10, pp. 8, 9.
. 32. Ref. 6, p. 3-179.
33. Ref. 10, p. 190.
34. Ref. 6, p. 3-125.
35. Ref. 6, p. 3-130.
5.1-37
-------
36. Ref. 10, p. 263.
37. Ref. 28, p. 7.7-1.
38. Ref. 10, p. 294.
39. PEDCo Environmental, Inc. Industrial Process Profiles for Environ-
mental Use: Primary Zinc Industry. Prepared for U.S. Environmental
Protection Agency. Cincinnati, Ohio. February 1980. p. 45.
40. Fejer, M.E., and D.H. Larson. Study of Industrial Uses of Energy
Relative to Environmental Effects. U.S. Environmental Protection,
Agency. Research Triangle Park, N.C. July 1974. p. XII-5.
41. Ref. 6, p. 3-136.
42. U.S. Department of the Interior, Bureau of Mines. Mineral Facts and
Problems. Bulletin 667. 1975 ed. p. 48.
43. U.S. Environmental Protection Agency, Office of Air, Noise, and Radia-
tion. Primary Aluminum Draft Guidelines for Control of Fluoride Emis-
sions from Existing Primary Aluminum Plants. Research Triangle Park,
N.C. EPA-450/2-78-049a. February 1979. pp. 4-1 to 4-6.
44. U.S. Environmental Protection Agency, Industrial Environmental Research
Laboratory. Environmental Assessment: Primary Aluminum. (Prepublica-
tion copy.) Cincinnati, Oh. November 1, 1978. p. 15.
45. Nelson, W. L. Petroleum Refinery Engineering. New York, McGraw-Hill
Book Company. 1969. p. 74.
46. PEDCo Environmental, Inc. Guidance for Lowest Achievable Emission
Rates from 18 Major Stationary Sources of Particulate, Nitrogen Oxides,
Sulfur Dioxide, or Volatile Organic Compounds. U.S. Environmental
Protection Agency. Research Triangle Park, N.C. EPA-450/3-79-024.
April 1979. pp. 3.7-8 to 3.7-10.
47. Ref. 46, p. 3.7-9.
48. Shreve, R.N. Chemical Process Industries. 4th ed. New York, McGraw-
Hill Book Company. 1977. p. 227.
49. Ref. 46, p. 3.7-5.
50. Ref. 43, pp. 6-2 to 6-23.
51. Ref. 43, p. 6-22.
52. Ref. 46, pp. 3.7-5, 3.7-6.
53. Ref. 46, p. 3.7-3.
5.1-38
-------
54. U.S. Department of the Interior, Bureau of Mines. Primary Aluminum
Plants, Worldwide. Part One. Washington, D.C. August 1977.
55. Ref. 42, pp. 701, 706.
56. U.S. Environmental Protection Agency, Industrial Environmental Research
Laboratory. Environmental Assessment of Primary Nonferrous Metals
Industry Except Copper, Lead, and Zinc. Cincinnati, Oh. EPA Contract
No. 68-02-1323. p. IX-8.
57. Ref. 42, pp. 701, 702.
58. Ref. 6, pp. 4-65 to 4-72.
59. Ref. 6, p. 4-70.
60. Ref. 6, p. 4-71.
61. Sulfuric Acid Producers Fighting to Stay Even As Costs Rise But Smelter
Add to Market Supply. Chemical Marketing Reporter, 2J_5(19). May 7,
1979. p. 9.
62. Ref. 6, p. 7-43.
63. Ref. 6, pp. 7-46 to 7-48.
64. Rosenbaum, J.B., et al. Sulfur Dioxide Emission Control in Japanese
Copper Smelters. Bureau of Mines, U.S. Department of the Interior.
Washington, D.C. 1976. Information Circular 8701. pp. 4, 5.
65. Onahama Smelting and Refining Company. Double Expansion of Onahama
Smelter and Refinery, Onahama Iwaki-City, Fukushima-Pref, Japan. June
1975. pp. 5, 6, 9-12.
66. Treilhard, D.G. Copper—State of the Art. Engineering/Mining Journal,
April 1973.
67. Ref. 10, pp. 54, 55.
68. Ref. 2, p. 6.
69. Ref. 5, p. 29.
70. Ref. 5, p. 42.
71. Ref. 6, p. 3-213.
72. Ref. 10, pp. 143 to 144.
73. Ref. 6, p. 3-175.
74. Ref. 10, p. 179.
5.1-39
-------
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.
93.
94.
Ref. 6, p. 3-185.
Ref. 6, pp. 4-75 to 4-80.
a^rt* Cfnv7°nmentan1 Protection Agency, Office of Air Quality Planning
n the japanese
hn' V;- F1.ueGas Desulfurization at Bunker Hill Company, Kellogg
Fction An*1 Enf°rment Investl"9ations Center. U. S. Environmental
Protection Agency. Denver, Col. EPA-330/2-79-011 . February 1979.
IQT-. o
ly/ 1 . p. 29.
Ref. 77, pp. 217, 225, and 228.
- clnter?or' Cont™1 of Sulfur Oxide Emissions in
inC Smeltln9- Bureau of Mines Information Circular
ncn Agency- Environmental Assessment of the
Domestic Primary Aluminum Industry (Pre Publication copy). industrial
Env ronmental Research Laboratory, Cincinnati, Ohio. September 1978.
p • I o *
Ref. 43, p. 1-14.
Ref. 46, p. 3.7-7.
Ref. 46, p. 3.7-12
C0pmmuni'"t1.on f™" J- Farrington, SF Air Control, Inc., May
Recorded in memo by J. Wunderle for PN 3310-N file.
Ref. 43, pp. 5-6 to 5-8.
fl™ DeFlle.ney' R;DV Rad1an Corporation, Austin, Texas, to
ment Re.Vrh'f'h r°nmen^al ?rotectl'on Agency, Industrial Environ-
mental Research Laboratory. Cincinnati, Oh. September 11, 1979.
Ref. 1, p. 138.
Ref. 1 , Appendix.
Ref. 6, pp. 6-29 to 6-41.
draft.) December
Ref. 6, pp. 6-114 to 6-123.
Ref. 6, pp. 6-75 to 6-85.
Ref. 78, pp. 11, 19 to 26.
5.1-40
-------
95. Ref 6, pp. 8-1 to 8-32.
96. Ref. 6, pp. 8-33 to 8-45.
5.1-41
-------
-------
5.2 IRON AND STEEL PRODUCTION
The primary processes at an integrated steel making facility are produc-
tion of coke, reduction of iron ore to pig iron in a blast furnace, and
refining of the pig iron into various grades of steel. These processes
involve a diversity of operations, as shown in the composite flow diagram,
Figure 5.2-1.
In the blast furnace, the combustion of coke provides the reducing
atmosphere that converts the iron ore to metallic iron or pig iron. The pig
iron is refined into steel by oxidizing the impurities and adjusting the
alloy content to specified levels. Refining is done in various types of
steelmaking furnaces: open-hearth, basic oxygen, or electric arc furnaces.
Sulfur dioxide emissions come primarily from the sintering process and
from combustion of coke-oven gas and fuel oil. Most of the sulfur from the
ore and coke combines chemically with other substances in slag from the
blast and steelmaking furnaces. As the slag is cooled, some of this sulfur
is released as hydrogen sulfide or S02. Figure 5.2-2 shows the overall
sulfur balance at a small integrated steel plant.1 The specific sulfur
balance will be different for each plant depending upon raw material
analyses, fuels used, and the specific processes employed. In general,
however, sintering, combustion of coke-oven gas, and heating processes are
the primary S02 sources, generating about 1.05 kg of S02 per megagram (2.1
Ib/ton) of ingot product. The S02 emissions and available control methods
are discussed in the following sections.
Although foundries and ferroalloy furnaces are sometimes located in
integrated steel plants, these sources are usually considered apart from the
iron and steel industry for purposes of industrial classification. They are
discussed in the following sections, however, because the processes are
similar. Foundries and ferroalloy furnaces are not major sources of S02,
and control techniques are not currently applied solely for S02 removal.
5.2.1 Process Descriptions and Emission Sources
5.2.1.1 Sintering—
Process description—The sintering process converts iron-bearing fines
into an agglomerated product that is suitable for charging to the blast
5.2-1
-------
to c
C. O
CJ (/)
o to
cu
CJ
s- x
cr- o
CJ
O O
E; )
(O CJ
s_ o
01 t-
r« 3
•r- O
O O
Q. CD
o
co
5.2-2
-------
EMISSIONS
22% 5% 4%
INPUTS
70%—>
COAL
|
12%-
10%-
8%'
ORES AND MISC.I
.1
H
FUEL OIL
•t
SCRAP
NOTE: WIDTH OF LINE
EQUALS Mg PER DAY
SCALE '
t
i—i CO
OOUJ
00
t
.
00
STEEL PLANT
CO
o
Z30
QCO
CD
«C
-J
OO
O
CO
ca<: . .
Mil
UJ
o
0
0
CJ
£0
•X.
-------
furnace. Iron-bearing fines consist of the dust collected from air pollu-
tion control devices, roll scale, and finely crushed iron ore. Particle
sizes of these materials range from less than 10 microns to 0.6 cm (0.25
in.), and they cannot be used as blast furnace feed without first agglomera-
ting. The fines are mixed with equally fine limestone or dolomite and coal
or coke. Mixtures with a high percentage of limestone or dolomite produce
what is called superfluxed sinter. Water is added to the mixture to provide
cohesiveness. The proportions of the constituents in the mixture can be
varied over a wide range; a typical mix for a superfluxed sinter is as
follows:2
Iron-bearing fines--70 percent
Limestone or dolomite—18 percent
Coal or coke—4 percent
Water--8 percent
The mixture is placed on the sinter strand (an endless moving grate
made from cast steel bars), and a burner hood located near the feed end of
the unit ignites the coal or coke. Combustion air is drawn through the bed
of material from the top by a fan. The combustion is self-supporting and
provides enough heat to cause surface melting, reaction of the constituents,
and agglomeration of the mix. The temperature at the combustion zone is
1300° to 1500°C (2400° to 2700°F).2
Typical heat input to the combustion furnace is about 174,000 kJ/Mg
(150,000 Btu/ton) of sinter.2 As a means of achieving uniform distribution
of combustion air, the underside of the sinter machine incorporates compart-
ments called windboxes. The process fan pulls the air through the bed into
the windboxes, and then into a duct to the air cleaning device. The temper-
ature of the exhaust gas in the duct is typically 100° to 150°C (215° to
300°F). Air requirements for sintering are about 3100 NmVMg (100,000
scf/ton) of sinter produced.2 For a plant producing 4540 Mg (5000 tons) per
day, the discharge gas volume would be about 9,900 NmVmin (350,000 scfm).
The sintered cake that falls from the end of the strand is crushed and
screened. The undersize portion is recycled to the feed mix, and the re- .
mainder is allowed to cool. In 1976 sinter production in the United States
was about 32.7 Tg (36 million tons).3
5.2-4
-------
Emission .sources--The only S02 emitted from the sintering process is in
the windbox exhaust gases. The amount depends on the sulfur content of the
raw materials and the amounts of fuel and limestone used. The sinter pro-
duct retains as much as 50 percent of the input sulfur.4'5 The concen-
tration of S02 in the exhaust gases ranges from 20 to 398 ppm and averages
about 110 ppm, which is equivalent to about 0.9 g/kg (1.8 Ib/ton) of
sinter.6 Based on the 1976 industry production of 32.7 Tg (36 million
tons),3 the annual uncontrolled S02 emissions from sintering are 29,400 Mg
(32,400 tons). For an uncontrolled S02 emission rate of 30 g/s (238 Ib/h),
the maximum ambient 3-hour average concentration of S02 at ground level is
21 ug/m3> and the maximum 24-hour concentration is 4 ug/m3.7 Height of
sinter plant stacks ranges from 46 to 61 m (150 to 200 ft).
5.2.1.2 Byproduct Coking Operations—
Process description-Coke is the solid residue from controlled pyroly-
sis of coal. Coke consists of carbon, 6 to 10 percent ash, 0.5 to 1.0
percent sulfur, and various trace elements.
In the byproduct coking process, which accounts for over 99 percent of
U.S. coke production, coal (usually containing less than 1 percent sulfur
and about 30 percent volatile matter) is' heated in the absence of air in
individual ovens. An individual oven is a narrow refractory channel, typi-
cally about 12.3 m (40 ft) long, 3 to 6 m (10 to 20 ft) high, and 0.5 m (20
in.) wide. A single oven holds 11 to 33 Mg (12 to 36 tons) of coal. Usu-
ally 40 to 70 ovens are arranged side by side to form a coke-oven battery.
Flues between adjacent ovens are heated to produce the high temperature
(1200° to 1370°C or 2200° to 2500°F) needed to drive the volatile matter
from the coal. Coal is usually charged through three or four ports in the
top of the oven from larry cars, which are wide-gauge vehicles fitted with
coal hoppers that traverse the entire length of the battery on rails.8 In
recent years, several other methods have been developed for charging pre-
heated coal. When the coal is preheated before coking, it can be charged to
the coke oven through a pipeline (with steam conveyance), through a Redler
conveyor and special charging machine, or with a specially designed larry
car.9
5.2-5
-------
When the oven is fully loaded, the charging ports are covered and
sealed for the remainder of the coking cycle, which lasts about 18 hours to
produce blast furnace coke and 24 to 36 hours to produce foundry coke. The
gases generated during the coking cycle are piped to a byproduct recovery
section adjacent to the coke-oven battery, where various byproducts (such as
tar, anhydrous ammonia or ammonium sulfate, and light oil) are recovered.
After the byproducts are removed from the gas, about 40 percent of the gas
is used as fuel to heat the coke ovens. The remainder is used as fuel
elsewhere in the plant or is flared.
At the end of the coking cycle, doors at the end of the oven are re-
moved, and the incandescent coke is pushed out of the ovens into a special
railroad car. The product is then transported to a quenching station, where
it is cooled by deluge water sprays.
Sulfur dioxide emissions—The coke-oven gases produced by the con-
trolled pyrolysis of coal contain reduced sulfur compounds in addition to
numerous hydrocarbons. About 25 to 30 percent of the sulfur in the coal is
discharged in gaseous form as a constituent of the coke-oven gas. Almost
all of this sulfur is present as hydrogen sulfide, with minor amounts of
mercaptans. The normal range of hydrogen sulfide concentration is 5.7 to 11
g/m3 (2.5 to 5.0 gr/scf).10 The average gas yield is 361 Nm3/Mg (11,500
ftVton) of coal.11 At a hydrogen sulfide content of 9 g/m3 (4.0 gr/scf),
the potential emissions, expressed as sulfur dioxide, are 6.1 kg/Mg (12.3
Ib/ton) of coal. In 1976, about 69 Tg (76 million tons) of coal was pro-
cessed in the steel industry. These figures result in a potential annual
uncontrolled S02 emission of 423,000 Mg (466,000 tons).
Use of the coke-oven gas to heat or underfire the coke ovens or as fuel
for other combustion operations results in S0x emissions unless the hydrogen
sulfide is removed. At integrated steel facilities, coke-oven gas provides
an average of 70 percent of the total underfiring energy and is the exclu-
sive fuel for most batteries. Among merchant or foundry coke plants, coke-
oven gas is used almost exclusively for oven heating. In 1976 the under-
firing of coke ovens consumed approximately 32 percent of the total coke-
oven gas produced.11 No data are available on the amount of gas flared, but
it is minor because most plants attempt to utilize the energy in coke-oven
gas (22,100 kJ/Nm3 or 550 Btu/ft3) for useful heating.
5.2-6
-------
Although some plants strip the hydrogen sulfide from the coke-oven gas
before burning the gas, this practice is not universal. It is generally
done only where the gas is used in operations metallurgically sensitive to
sulfur or where state regulations require It, as, in Pennsylvania and West
Virginia.10 The stripped hydrogen sulfide may be converted to sulfuric acid
or elemental sulfur.
Fugitive emissions occur during oven charging and during the coking
cycle. These charging and coking emissions are significant with respect to
particulate matter and various organic substances, but not with respect to
S02, since they total about 681 Mg (750 tons) of S02 per year based on the
coal usage cited earlier and the emission factors cited in Reference 12.
Sulfur dioxide emissions from coal preheating are insignificant because
the sulfur in the coal is not released at the low temperatures involved
(200° to 425°C or 400° to 800°F). Either coke-oven gas or natural gas is
used as fuel for preheating.13
As an example of the impact of plantwide coke-oven gas combustion on
ambient S02, the maximum contributions to average annual ambient S02 levels
are 13, 23, and 128 \ig/m3 from plants emitting 4.0, 9.4, and 36.5 Mg of S02
per day (4.4, 10.4, and 40.2 tons/day).10 The specific stack parameters
used to obtain these results are not given in the reference. The exhaust
stacks at coke-oven batteries are 61 to 77 m (200 to 250 ft) high. Exhaust
gas temperatures from coke-oven combustion stacks range from about 130° to
315°C (270° to 598°F)'.14 Flow rates vary from 850 to 1870 NmVmin (30,000
to 66,000 ft3/min) depending upon battery capacity, fuel used, and excess
air. The flow rates cited are for batteries processing from 37.5 to 52.5
Mg/h (41 to 58 tons/h) of coal with excess air ranging from 33 to 200 per-
cent.14 The coke-oven gas is also burned in plant boilers and heating fur-
naces, both of which have'stack heights of 61 to 77 m (200 to 250 ft). The
gas flow rates from plant boilers and heating furnaces vary widely depending
upon size and fuel used. Exhaust flow rates for soaking pits using gas or
oil are about 620 Nm3/Mg (20,000 scf/ton) of steel heated. The corres-
ponding rates for reheat furnaces are 1270 Nm3/Mg (41,000 scfm/ton).15 The
exhaust gas flows for a boiler with a capacity rating of 44 MW thermal (150
x 106 Btu/h) vary from about 815 NnrVmin (28,800 scfm) for oil or gas firing
to 1126 NnrVmin (39,800 scfm) for coal firing.16
5.2-7
-------
5.2.1.3 Heating Furnaces—
Process descrlption-The heating furnaces in steel plants are referred
to as soaking pits, reheat furnaces, heat-treating furnaces, or annealing
furnaces, depending upon their metallurgical function. Many of these fur-
naces are equipped to fire various types of fuel including oil, natural gas,
coke-oven gas, and blast furnace gas.
Soaking pits are box-shaped furnaces 6 to 12 m (20 to 40 ft) square and
5 m (15 ft) deep. Steel ingots are placed in the furnace and heated for 12
to 24 hours at about 1300°C (2400°F) to bring them to a uniform temperature
for subsequent rolling into slabs, blooms, or billets. Soaking pit furnaces
are arranged in groups of 4 to 16 furnaces or more and are fired with nat-
ural gas, low-sulfur oil, or coke-oven gas. The sulfur content of the fuel
is important metallurgical^ because certain grades of steel are sensitive
to absorption of sulfur. The presence of sulfur may affect the surface
quality of the product during subsequent rolling operations. The energy
requirement for the soaking pit operation is about 1.16 to 1.74 GJ/Mg (1.0
to 1.5 million Btu/ton).17"19
When steel is cast continuously, the ingot and soaking pit operations
are eliminated, because the steel is cast directly into the intermediate
shapes required for final rolling. Continuous casting has been adopted in
most newer plants and so-called minimi 11s because of the energy savings and
higher yield.
Reheat furnaces are of various sizes and types, depending on the prod-
uct being heated. They serve the purpose of reheating intermediate steel
shapes to a temperature high enough to permit further rolling or shaping
about 1300°C (2400°F). The energy used for reheating slabs is about 2.9 to
3.5 GJ/Mg (2.5 to 3.0 million Btu/ton). The fuels are natural gas, low-
sulfur oil, and coke-oven gas.20'22
Heat-treating furnaces, generally operating in the range of 425° to
870°C (800° to 1600°F), are used to impart strength and hardness to the
finished product. Heat treating can be a batch or continuous operation.
There are numerous variations in the design of heat-treating furnaces but
these differences are not significant with respect to emissions. Combustion
5.2-8
-------
products are normally vented into the building and escape through roof open-
ings. Heat-treating furnaces generally burn natural gas or residual oil.
An annealing furnace is a special type of heat-treating furnace used to
anneal (soften) steel that has been cold-rolled. The annealing furnaces are
normally indirect-fired to prevent formation of scale on the steel. A
cylindrical cover is placed over the charge, forming a chamber that is
filled with a reducing gas to keep products of combustion from contacting
the steel. Furnace temperatures for annealing steel range from about 600°
to 760°C (1100° to 1400°F). In the annealing of strip steel, the facility
consists of 10 to 50 batch furnaces. A continuous annealing furnace can
supplant many batch furnaces. It is a tall structure in which the steel
strip is looped several times as it travels through the furnace to achieve
long exposure time. Combustion products from annealing are usually vented
into the building. The energy requirement for anneal ing:is about 0.9 to 1.5
GJ/Mg (0.75 to 1.25 million Btu/ton) of steel.23 25
Sulfur dioxide emissions—The S02 emissions from heating operations are
a function of the fuel used. Table 5.2-1 shows the calculated S02 emissions
from soaking pits and reheat furnaces burning oil or coke-oven gas.15
Emissions of S02 from annealing furnaces~are minor because high-sulfur fuels
are seldom burned.
In 1976 the total energy consumed by the steel industry in heating and
annealing furnaces was estimated to be 6.1 x 10« GJ (5.77 x 1014 Btu).26
This consisted of 2.24 x 106 m3 (592 million gal) of fuel oil, 9.57 x 109 m3
(335 billion ft3) of natural gas, 8.14 x 109 m3 (285 billion ft3) of coke-
oven gas, and 3.49 x 109 m3 (122 billion ft3) of blast furnace gas. Calcu-
lations based on an average sulfur content of 1.0 percent in fuel oil and an
average hydrogen sulfide content of 4.5 g/m3 (2 gr/ft3) in coke-oven gas
(assuming that about half of the gas is desulfurized) yield the total uncon-
trolled S02 emissions from fuel combustion. The values are 112,500 Mg/yr
(124,000 tons/yr), or 0.95 kg/Mg (1.9 Ib/ton) of raw steel.
5.2.1.4 Foundries—
Process description—Foundries produce castings for automotive parts,
light and heavy machinery, pipe, and a wide range of miscellaneous products.
5.2-9
-------
CO
t— 4
Q.
CD
1—4
0
CO
£
Q:
co co
SS LU
O 0
CO 21
CO DC
t-t •=>
s: LU
LU
LU ea;
O LU
X LU
O O£
1— 1
Q Q
rr* ?
i^, «^
u.
— 1
CO
^
r—
OJ
LO
LU
t
CQ
to
0)
O
re
c
S-
£.
ta
QJ
QJ
a:
to
n.
CD
•r-
re
o
CO
CO
CM
CM
CO
„ .
LO
co
v3-
LO
r— •
c
o
•M
3
CD
O
*— ^
CD
E:
•-3
CD
>£
to ^: i—
QJ "•j
T3 CD )
*)- t— c
O QJ
CD CD
CD E O
E •!- S-
r.c^
•r- re ^:
E S- *3~
E O CD 4->
C f~^ ^_^
1- tO LO CIO
(O **— *• ^5 f*^
tO CD 4-> i —
£= CO \-^
O C E -Q -0
•r- gj -x^ , ,__
t/) > QJ^
tO O TD
•r- 1 T- WO
E o» M- s: CD
Q) .^ r— •• ^^^
O 3 CD CD
CMO tO J*£ -i^
0
CO
5.2-10
-------
The process involves melting scrap metal and/or pig iron (crude iron in the
form of blocks) and pouring the molten metal into prepared molds. The two
major categories of foundry product are "gray iron" and "steel." Both
consist mostly of elemental iron, but gray iron contains 2 to 4 percent
carbon whereas steel contains 1 percent or less. Gray iron contains various
amounts of other elements, generally less than 1 percent. Steel may also
contain alloying elements. Such terms as "malleable," "white," and "nodu-
lar" iron are .used to describe gray iron castings with specific properties.
Figure 5.2-3 illustrates the process flow in a typical gray iron
foundry. More than 75 percent of the U.S. installations use a cupola fur-
nace to melt the raw materials, but the use of electric furnaces is in-
creasing.
27.28 in 1973 electric arc furnaces accounted for 17 percent of
total gray iron production.28 Cupola capacities range from 1 to 45 Mg (1 to
50 tons) of melted metal per hour; over 60 percent operate in the range of
2.7 to 9.9 Mg (3 to 11 tons) per hour. Electric induction and reverberatory
furnaces are.also used in gray iron foundries.
The cupola is a refractory-lined, cylindrical furnace resembling a
small blast furnace. Raw materials consisting of iron scrap, pig iron,
fluxes, and coke are charged through a door in the top of the cupola.
Fluxes are limestone or similar minerals, which absorb or react with impuri-
ties after the charge has melted. The coke is burned by blowing air through
ports (tuyeres) near the bottom of the furnace. The air may be preheated as
high as 980°C (1800°F) to reduce coke consumption.29 At cupola blowing
rates of 142 to 425 NmVrnin (5,000 to 15,000 scfm) with a preheat tempera-
ture of 550°C (1,000°F), the melting rates are 8.2 to 22.7 Mg/h (9 to 25
tons/h).30 This corresponds to about 1090 NmVMg (35,000 ft3/ton); the
undiluted exhaust flow rates from cupolas average about 903 NmVMg (29,000
ft3/ton).31 The dilution air introduced from the combustion of carbon
monoxide as well as from the charging door is about equal in volume to the
undiluted gas flow.31 As the charge melts, the molten material flows to the
bottom of the furnace, from which molten iron is drained periodically or
continuously. Additional raw materials are added to keep the furnace full.
Operation of the cupola furnace is normally continuous, but it is operated
over a much shorter period than a blast furnace because the cupola must be
reconditioned about once a week.
5..2-11
-------
fO
O)
ro
CD
03
CO
CM
CO
CO
uj o
LU
to
5,2-12
-------
Where electric arc, electric induction, and reverberatory furnaces are
used in gray iron foundries, the charge consists mainly of iron scrap, pig
iron, and limestone. These furnaces are operated on a batch basis. The
reverberatory furnace is heated by firing gas or oil. The molten metal is
drained from the melting furnace at a temperature of about 1600°C (2900°F)
into a ladle and then into prepared molds. After solidification the cast-
ings are removed from the molds and cleaned by shot blasting.
Castings intended for certain uses may be heat-treated for several
hours at temperatures of 530° to 900°C (1000° to 1600°F). Heat-treating
furnaces, .fired by gas or oil, are referred to by many different names,
including "annealing," hardening," "car-bottom," and "traveling hearth."
Castings that have been annealed are often referred to as "malleable iron
castings."
The cores and molds are formed in the desired shape from sand and
binders and are cured either in a baking oven (core oven) at 150° to 260°C
(300° to 500°F) or at room temperature. Curing evaporates moisture and
hardens the sand mixture. Core ovens are fired with natural gas or oil.
In 1978 the total shipment of gray, ductile, and malleable iron
castings was 14.8 Tg (16.3 million tons). Gray iron castings accounted for
77 percent of the total.32 A 60 percent yield of good castings means total
production was about 24.7 Tg (27.2 million tons).
Steel foundries are similar to gray iron foundries; the main difference
is that electric furnaces and open hearth furnaces, rather than cupola fur-
naces, are used for melting. The raw materials consist of steel scrap, pig
iron, and fluxes. The open hearth furnaces are fired with gas or oil.
Sometimes supplemental oxygen is blown into the open hearth furnace and the
arc furnace to accelerate the process, a procedure called oxygen lancing.
Large steel foundries operate 24 hours a day and 7 days a week, whereas
smaller ones operate 8 hours a day. Capacities of foundries range from 5 to
216 Mg/day (5 to 240 tons/day).33 Total shipments of steel castings in 1978
were 1.7 Tg (1.9 million tons).32
Sulfur dioxide emissions-Sulfur dioxide emissions from cupolas depend
upon the sulfur content of the coke (about 1 percent), the quantity of coke
burned, and type of iron being produced. Over 70 percent of the input
sulfur is normally contained in the slag and the iron.34
5.2-13
-------
At a coke sulfur content of 0.8 percent, and a metal-to-coke ratio of 9
to 1, the S02 emission rate estimated by one source is 0.18 kg/Mg (0.36
Ib/ton) of metal charged, equivalent to about 57 ppm in the cupola off-
gas. 34 Another source reports S02 concentrations of 25 to 250 ppm in the
cupola off-gas.35 Based on the emission value of 0.18 kg/Mg (0.36 Ib/ton)
and the 1973 production of 20.5 Tg (22.6 million tons) of iron, the annual
uncontrolled S02 emissions from cupolas are 3700 Mg (4000 tons) Cupolas
are not generally considered a significant source of S02 emissions because
of the inherent metallurgical restriction on coke sulfur content, i.e., less
than 1 percent.34
The S02 emissions from electric arc furnaces producing gray iron aver-
age 0.12 kg/Mg (0.24 Ib/ton) of iron.36 The emissions are highly variable
however, depending on the sulfur content of the scrap and the grade of iron
being produced. Total annual uncontrolled emissions of S02, based on 1973
production of 4.61 Tg (5.1 million tons), are 545 Mg (600 tons).
The S02 emissions from open hearth furnaces are a function of the
sulfur content of the fuel and grade of steel being produced. The heat
required for 100 percent scrap charges is 1.16 to 1.74 GJ/Mg (1 0 to 1 5
million Btu/ton) of steel.37 If th1s neat is supplied by burn1ng Qf ^
sulfur oil, the S02 emissions may approach 1.0 kg/Mg (2 Ib/ton). Because
open hearths are not widely used and the sulfur content of oil is restricted
in most localities, S02 -emissions from open-hearth furnaces are not con-
trolled by add-on control devices.
5.2.1.5 Ferroalloy Production—
Process description-Ferroalloy is the generic term for solid solutions
of iron and one or more other elements. Ferroalloys are used in steelmaking
as both deoxidizing elements and as alloying constituents. The United
States is the world's largest producer and user of ferroalloys ** The
ferroalloy most widely used in steelmaking is ferromanganese, followed by
ferrosilicon, ferrochromium, and ferrophosphorus. Silicomanganese is also
considered as a ferroalloy product although it does not contain much iron.
The alloys are used in steel manufacturing for deoxidation and to impart
special properties to the product metal.
5.2-14
-------
Silicon metal and silicomanganese are not strictly ferroalloys because
they contain no iron. They are discussed here because they are manufactured
in the same way as ferroalloys, which are made in open-semi sealed, or sealed
electric submerged-arc furnaces. Ferromanganese is also made in blast
furnaces.
The raw materials for ferrosilicon production are iron ore or iron
scrap, a silicon ore such as silica sand, coke or coal, and limestone.
Production of silicon metal requires silica and coal or coke; production of
silicomanganese requires silicon ores, manganese ores, and coal or coke.
The raw materials are charged to the furnace, which is a refractory-
lined crucible, and carbon electrodes are immersed in the mix. The mix is
melted by resistance heating, i.e., resistance of the charge material to the
flow of current between the electrodes. Additional heat comes from chemical
reduction of the iron, manganese, and silicon oxides and from oxidation of
the coke or coal. The temperature near the electrode tips is estimated to
be 2200° to 2760°C (4000° to 5000°F).39 Impurities rise as a floating slag,
and the molten alloy and slag are drained periodically from tapholes near
the bottom of the furnace. Once started, the process is essentially con-
tinuous. Some 24 ferroalloy plants in the United States produce about 1.5
Tg (1.7 million tons) per year.40
Sulfur dioxide emissions-Measured S02 emissions from several ferro-
alloy furnaces have ranged from less than 1 ppm to 17 ppm. The S02 emis-
sions from furnaces equipped with control devices did not exceed 3.2 kg/h (7
lb/h).39 .
5.2.1.6 Blast Furnace Slag—
Process description-Sulfur is introduced into the blast furnace pri-
marily through the sulfur in the coke. Because of the reducing conditions
and slag composition, most of the sulfur is discharged in the slag. The
remainder is dissolved in the molten metal. The blast furnace produces 250
to 400 kg (500 to 800 Ib) of slag per megagram (ton) of metal. Sulfur
content of the slag is 1.2 to 2.0 percent. Slag is usually discharged from
the furnace every 3 to 6 hours; in some of the new, larger furnace opera-
tions, the slag is drained continuously. The slag may be drained into a pit
and sprayed with water for cooling or may be granulated by introducing water
5.2-15
-------
directly into the stream of molten slag on a rotary drum or by application
of a fine water spray. When the slag stream contacts water, some of the
sulfur is released as fugitive hydrogen sulfide. The quantity depends upon
many variables, including the initial sulfur content of the slag and the
method of spraying.41"43
Sulfur dioxide emissions--The amount of S02 released directly from slag
pouring is not known. Within the industry there are many variations in slag
volume, sulfur content, and pouring and cooling methods. Most of the sulfur
released apparently occurs as hydrogen sulfide. Given a slag production
estimate of 300 kg/Mg (600 Ib/ton) of hot metal, the 1976 production of 79
Tg/Mg (87 million tons) of hot metal,44 and a slag sulfur loss of 0.24
percent,4^ the potential annual hydrogen sulfide emissions would be 106 Gq
(118,000 tons).
5.2.2 Control Techniques
5.2.2.1 Description—
Sinterinq-Sulfur dioxide emissions from the sinter plant windbox
exhaust can be reduced by reducing the total sulfur input from the raw
materials and by increasing the quantity of limestone feed.* These methods
may not be feasible, however, because of constraints on the availability of
raw materials or metallurgical constraints on product quality. Increasing
limestone feed also increases the resistivity of the windbox dust, which
decreases particulate removal efficiency for plants using electrostatic
precipitators.45
Wet scrubbers used for particulate control can also remove as much as
99 percent of the S02 when supplemented by injection of caustics.46 One
scrubber removed about 60 percent of the S02 when using regular plant
water.4* NO wet scrubber systems are installed in the United States for the
express purpose of S02 control of sinter exhaust. The common element of all
S02 control systems discussed in the literature is absorption by use of a
separate liquid system or a wet scrubber. Four processes have been used in
Japan at plants having S02 concentrations of 400 to 800 ppm; ammonia is the
scrubbing reagent in one process, and lime-limestone in the other three.
Reported removal efficiencies range from 90 to 95 percent. In all cases,
5.2-16
-------
the S02 scrubbing systems follow particulate collection using an electro-
static precipitator. The flow rates at the 12 plants using these systems
range from 2,000 to 33,000 NnvVmin (71,000 to 1,200,000 ftVmin).46
Byproduct coking operations—Several processes are suitable for re-
moving hydrogen sulfide from coke-oven gases. These processes either re-
cover elemental sulfur or produce a concentrated stream of hydrogen sulfide,
which in turn can be converted into sulfuric acid or elemental sulfur. The
processes most commonly used to produce sulfur are the Stretford, Takahax,
Fumaks, and Giammarco-Vetrocoke processes.47'48 Those producing concen-
trated hydrogen sulfide are the Vacuum Carbonate, Sulfiban, Bravo/Still, and
Diamox processes.47'48 The Vacuum Carbonate and Sulfiban processes are the
most widely used processes in the United States.49
In the Vacuum Carbonate process, hydrogen sulfide is absorbed into a
3.0 to 3.5 percent solution of sodium carbonate. The hydrogen sulfide is
then stripped by steam from the absorbent in a reactivating tower. The
reactivation is performed under vacuum to reduce the quantity of steam
required. The hydrogen sulfide content of the coke-oven gas can be reduced
by approximately 93 to 98 percent using a relatively new two-stage process.
Conventional systems achieve about 90 percent removal.47'50
After the stream is condensed, the hydrogen sulfide is available for
further use. Of the eight plants using the Vacuum Carbonate system in the
United States in 1978, six had Claus sulfur recovery systems.49 The sulfur
recovery plants used in conjunction with the Vacuum Carbonate system typi-
cally operate at 95 to 97 percent efficiency. Thus overall hydrogen sulfide
removal efficiency using the two-stage Vacuum Carbonate system with sulfur
recovery is about 90 to 95 percent. '
The Sulfiban process can reduce hydrogen sulfide concentrations in
coke-oven gas to less than 0.23 g/m3 (0.10 gr/100 ft3) by use of a mono-
ethanolamine (MEA) scrubbing solution. This process also produces a concen-
trated hydrogen sulfide stream. Three plants planned in the United States
as of 1975 all used a Claus plant to recover sulfur from the H2S stream.49
In the Stretford process, the gas is scrubbed in packed towers by an
alkaline solution of anthraquinone disulfonic acid (ADA), sodium ammonium
vanadate, and buffering compounds. The hydrogen 'sulfide reacts with the
alkali to form sodium hydrosulfide (NaHS). In holding tanks at the bottom
5.2-17
-------
of the towers, the NaHS reacts with the vanadium to form free sulfur. The
vanadium, which is reduced in the reaction, is reoxidized by the ADA. In
subsequent steps the ADA is oxidized, and the sulfur recovered in the molten
state. Hydrogen sulfide removal efficiencies can exceed 99 percent.47»si
This process is currently being used at the Hamilton, Ontario, coke-oven
plant of Dominion Foundries and Steel, Limited, and at several plants in
Europe.52'53
The Takahax and Fumaks processes are similar in configuration to the
Stretford process, although the process chemistry differs significantly.
These processes are used in Japan with reported hydrogen sulfide removal
efficiencies of more than 99 percent."'« These processes all pose a
potentially serious water pollution problem if the thiosulfate and thio-
cyanate are not removed from the wastewater before it is discharged.
Coke-oven gas can come into direct contact with water in various by-
product operations, such as direct-contact coolers. Some hydrogen sulfide
is removed, but that which is dissolved in the water is released to the
atmosphere if the water is cooled in open cooling towers.
Heating furnaces-The only control for S02 from heating furnaces is the
choice of fuel. Natural gas and blast furnace gas are essentially free of
sulfur. Coke-oven gas that is burned for heating may be either desulfurized
or blended with natural gas to dilute the hydrogen sulfide content. Most
state regulations limit the sulfur content, of oil to 0.5 to 1.0 percent.
Foundries and ferroalloy produCt.inn-Thoy.0 are no controls for S02
emissions from the cupolas, electric arc furnaces, or open-hearth furnaces
now in use. Where wet scrubbers are- used for particulate control, addition
of caustic to the scrubber wastewater can remove significant amounts of S02
from the gas.54
Blast furnace slaq-A variety of operating practices and granulation
techniques have been studied to minimize emissions.«>** • Addition of oxi-
dizing agents to the quench water is effective in reducing hydrogen sulfide
emissions but may create higher S02 emissions. Reducing the sulfur content
of the initial slag to 1.0 to 1.5 percent is effective but is not always
feasible for metallurgical reasons.
5.2-18
-------
5.2.2.2 Control Cost--
All costs in this section are referenced to July 1979 dollars. Values
obtained from various references were adjusted to July 1979 by use of the
Chemical Engineering plant cost index for July 1978 of 220 and an inflation
rate of 7.5 percent for the period July 1978 to July 1979.
Sintering—The estimated cost of an installed limestone scrubbing
system on a large sinter plant producing 6312 Mg/day (6950 tons/day) with a
flow rate of 675,000 NmVh (394,000 scfm) is $11,900,000. Total annual
operating cost including plant and payroll overhead and capital charges is
$4,221,000 or $2.29/Mg ($2.08/ton).55 These costs include a venturi scrub-
ber to remove particulate matter, an S02 absorber, limestone preparation
equipment, and a water treatment system. Figure 5.2-4 illustrates the
components of the system. Tables 5.2-2 and 5.2-3 present the capital and
annual operating costs, respectively.
Several sinter plants have been retrofitted for particulate control.
Although no U.S. plants have required direct S02 control, retrofit of a
limestone scrubbing system would be possible where space is available; many
steel mills are very congested and might not be able to accommodate lime-
stone and sludge handling facilities.
Byproduct coking operations—Massey and Dunlap have presented costs for
various desulfurization 'processes and associated hydrogen cyanide pretreat-
ment.56 These costs, adjusted to July 1979 dollars, are summarized in Table
5.2-4.
5.2.2.3 Energy and Environmental Impact—
Sintering—The energy requirements for the limestone scrubbing system,
described earlier are 12.6 kWh of electricity and 48.6 kg of steam per
megagram of sinter (11.4 kWh and 97 Ib per ton of sinter).55 For a plant
producing 302,000 Mg (333,000 tons) per year, the energy, requirement for a
system incorporating an ESP and limestone scrubber is 19.8 kWh/Mg (18
kWh/ton).57 The limestone scrubbing system generates large quantities of
wastewater, which must be treated; treatment in turn produces a sludge that
must be disposed of. The limestone scrubbing system requires about 160
liters of water per megagram of sinter (38 gal/ton) and 10 kg of limestone
per megagram of sinter (19 Ib/ton).
5.2-19
-------
ro
o
Q.
Q.
fB
CO
CD
'5.2-20
-------
CD
"Z.
I-H
co
1 — 1
X
1 1 1
«^ ul
z. >>
CD ro
LU (/>
1— C
co o
CO
1
CD CO
^^ f^^
I-H LT>
CO - —
CO
g >^
o -a s-
oo *-^ ro
1 1 1 ^~ i —
z. o
o CM -a
i
CO CO O1
LU tO f~-
"Si CT>
I-H CD r—
_J Z 1
I-H "O
U- 0 -i-
0 13 E
I— 0^"
co a;
O D-
O
I —
LU
I
1 —
z.
• t—t
CM CO
1
C\J
IT>
LU
— 1
CO
^^
h—
t/i
TO C
CT) O
t-£ 4->
O> ro
n r—
" (J
03 S-
CJ O
ifl OJ
TO S-
1-3
~W
**
-o
c c
ro o
ro
~T CNJ r—
§ !§§§ § till I
°i roio-a-ini-nO'— •=! aioorooi aj S^S^ §
CP> ^— CO ^- f^ ^^ r~~ *~~ ^*^ r~~ ^~ ^^ ^V (i «i •«
" "* " »— " r— ro
OJ 1— i— r~
ooooooooo
OOC3OC3C3OC)O
r— O CO »— C7^ i — LD O C^
cocnooro^vooir^
0°r— 1^ — CSJ^-CSJ^-
*
i/:
in ^
4J 0
en <->
i— O QJ
to O m *~^
S QJ — — in c V
en ° rn 4J -o r— O 4-> JJ JJ
c T] c c .^QjroQ-U .T30. S-
"~ _ S -^ in r— in in OJ ex M-
?"9 ™ ° 4-> QJ r- 0 -1- t- roOJ -0
C-S ^ 13 c o -— ro O--o •>-._, c
•f— C 13 ^3 *-^ •- If_) 4->in*O CTJ •!- in -i- i« cni—
111 Iff g-l 8 S S 8 ^ Sig^sl | -.. 1 =8 1
ro 5-«,aiSS^-- "• | = ^^« „ "g «= •£ .£ » "> S ^ a,gg">, S
42 ST.? D.3 b"£ ro m S o O. +J O QJ>3OCT>
o U1QJ-04J H- > +J St-nSmmCE-r^rootaii-'-- -t-1 Qjs-i-rec
o in 4-> QJ !- tsjm inr-s-i- g.?tj-S^43£-o>S-airo^a o CQJ4JS-T-
4-> S^ii^Soo^ooS -r^ = c = o -wr- roa. -ao-o Qj "5, §• ^ c c
u o =L .__._> oc^i c3x >— ••— c m o o o
« £ iS^lX COlZ — UJ -5 mu-i CO T3 UJ 000
o
o
o
o
~o
ro
o
o i
o
IT)
o
0
^_
•—
4-J
in
o
ro
4J
a.
ro
ro
° i
i— 1
3:
cn
r>^
en
•o
'i
o
c
OJ
QJ
01
O
S-
o
en
in LO
4-> r—
1A •
O ^~
O
5.2-21
-------
X
UJ
CD!*-
z: >,
I-H
Z CO
p r^.
I—in
to —
UJ
E >>
»-« to
—ITS
U- O1
os:
I—CM
CO r—
CD CO
CJ U3
CD CD
ofa
UJ CD
a. cc:
CD a.
2:0.
^D.
UJ
CO
UJ
00
II CO S-
0> U
\\ Vt 01
TO I.
II
11
||
TO
1! ~°
c c
TO O
1 ** '"-
II in *J
II • u
1 *•
(U O
II *"
II *"*
II
if
II
if
1
j
!
m
3
Materials and
CO ^-*
O VO
i— o
X *""
— - X
O VO
oo m *vo
r— *o O
•*— * cn r-» i—
o «a- — *
n o -— x
o o vo
•— "0 0 0
x 2 g - «
• X VO
O *}• r—
P^ VO
•— OJ
CO ^^
O VO
•— o
X "~
CM O~
• CD t^
en vo •
• — • co r^ o
o cn • — i —
ro o —
CD CD vo X
- «TS°.r.
x t— tn CM
**x •
*sr cn ro
CO O CM
r^v cn
r— CNJ
. „
IO |
c: s- ^,
o cu .—
4J Q. *~* ro
^3 cn 0)
to e; * »
S*-o 3 tn
« ro
OJ S- tn ^ »i-
S- O 1- r^ g-
•r- J3 13 - a|
3 fl3 O "O *>.^ -•
CTr— J= Q> t. -bO-
» *
rtl «3 3 (T3 4-> (/)
2 C E 0- 2 -.- 4->
co 4-» S_ r— Q)
— i O to n. LU. u
5
o o o o o o o
O O O O O CD O
to UD vo «3~ r^^ ir> vo
o co o% oo un m •—
co cr^ in o CM ro
r- r— r— in in
§O CO O CO O CO
o o o o o o
^— m c\j c^ in vo vo
n ro co o^ co f^» §—
CM cj> i" co vo m
c
o
tn
•!— . .
O t. ****
.a QJ *
"3 tn tn
•— >) OJ 4J
+-> o (— tn
4-> t. >, (j
OJ S- ro OJ U C i— O)
E ai o -t~> a; ••- 03 i_
•r- Q. +J ro ^— m C ••—
-JOco3Luz:«a: -o
c
o o a
o o o
o ^- ^>
m ^~ cfi
cn co •—
m CM
o o o
o o o
O r— ^«
^=r *r cn
r^. co i —
P-. n
**
r
o •*->
VJ v^ 3 m
cn o ro ~o i-
i- fO O
^ «/* rjj J3
CM in i- r—
4-> (U OJ
fO 4J 4-> f— > CTJ
c ra o c
in (D tn *r—
OJ E T3 4J O* +-J
s- m a> o -f- t_
ro QJ -C U -4J O)
-c: > i- in ro Q.
(J C OJ 4-> ^- t. O
•r- > u
r- o (U -r- tn H-
ro •— i. I- ••— o
+j ra •*->••- (U C
•i- •<-> C "O 4J -r- &«
Q- O ro C (O E O
ro -M r- -r- £ -0 ,—
00. cC
0
CNJ
CSJ
ro
o
o
o
CM
CM
*
•>!
-4-J
tn
o
tJ
cn
c.
4-)
ro
O)
CL
O
ro
C
C
ra
(O
O
o
o
o
tn i
O CO <
o r-. t
cn a)
c "~ c I
co u_ x_ .
• in s_ Q.-I- i.
4-> r- >, o •—
c. C. a.1—
TO *—'.C CT) Q. TO <
"a-tj o cn * -i-
o o r— c cn
s- «3- o o T- •
4-> CT1 C 4J O
C f^ E O TO
tn o a; 4-» 14- -r-
4-> S- TO C
it- 4-> u -i- tn
O 4-> tn O 4-»
TO i— ^s in •
• U -
•r- TO cn
••- i— c: c:
cn tn c: CL o •»-
c: « n 4->
•r- Cn 4-> T3 TO
c: i. m a) s-
•r— _^ *><>
-i- i'
Q. I
TO CO
U
(1)
i— tn
TO TO
4-> o
o
5.2-22
-------
00
CD
UJ
o
( t 1
0
^-•^
u_
VI
oo oo
OO UJ
O 1—
o oo
CD OO
"2Z.
1— H ^^
H- 0
^ p
UJ <£
O- M
O i— i
o;
Q Z3
2: u_
_1 OO
£
-M
t/>
C
_a
•f—
U—
j5
Ol
^_^
^".*°.
^ r—
c.
ro n
ro 01
'
^- nT
CO' CO
U3^
co r--
„
LO CO
CO O
"*~^
,. —
LO cn
<-o cn
^— .
oco u
0 B w
•o -oco
S-
U1 3
0 t-
(_> r—
l/J
C i-
o o
•r- <«-
4->
l/l •!-
O "O
Q. QJ
E S-
0 0
o
Ol O
-o c
U -C
) i ,f—
>> 3
ro OJ
4-) -o
ro ••-
^ 5
4-> C
C O)
a> cn
Bo
s-
ro t3
o; >,
S- -C
1°
c,C
•o o
•I— LO
tT3 S-
&01
CfTO
O E
y ^^
-a r—
to
i^
O) C
> o
o ••-
0 +->
rij ro
S. C
i(- 0
r— C
3 O
tn u
10 4->
3 (U
ro r—
O -r-
' c~ !>•.
O) O t-
-0 O)
3T3 >
i — QJ o
U 10 O
C ro
-------
Byproduct coklnq-The energy requirements for the three desulfurization
processes described earlier are presented in Table 5.2-5. The major envi-
ronmental lmpact of the wet-oxidative processes (such as the Stretford pro-
ami • ^ P9eneratl"°n °f wastew^er containing cyanides, sulfides, and
ammonia For a plant producing 1.7 million Nm3/day (60 million ft-/day) of
gas the effluent rate is 32,500 liters/day (8600 gal/day).si Tne Vacuum
Carbon.* Su1fiban processes ^^ ^^^^ ^ quantni.er
nighly contaminated wastewater.
5.2-24
-------
CO
CD
•yr
LU
O
1
LU
O
^v_!
C^LO
Lu CO
1 1 1
CO CO
1— CO
12: LU
1 1 1 (_3
s: c
t— i
o-o
LU l— i
rv | —
^^
>— f^
CD •— i
LU ro
Z Lu
LU —I
~^
1— co
LU LU
'z. cj
LO
CM
*
Ln
LU
QQ
I—
(O
ea
•o
en 3
$- •»->
CU 03
C
E
C T3
3 "O
CT CU
LU S-
•i —
3
cr
Ol
s-
^
A
>j
O
c:
OJ
•r—
O
•r—
lf_
Ij —
LU
t«
LO l^» C\J CT» CT»
VO CO OO CO CTi CTi
LO LO «* CO r— *sj-
i— 00 t— CT> O l~~
O i— E: c s- s-
3 3 4- *»- +•>
> 2> CO CO CO CO
•
>^
o
El
CD
•^
o
41
^—
QJ
CO
g
ro
a;
in
^-~^.
<^
^^
^^
3
4-*
CO
f-,
O
1
en
•••^
2
^D
LO
LO
CM
O)
E
)
(/>
^^
(O
^
O)
ro
O)
^_
*->
•4->
El
OJ
3
r—
M-
"4-
O)
S-
o
4->
(H
B
ro
a>
4J
OJ
5-
Q.
•4-^
3
O
^j
•^*
"^
•°
! *
C
E
ro
cu
4-)
^
c;
cu
3
n—
M-
^—
cu
_^
•r—
^£
O
5.2-25
-------
REFERENCES FOR SECTION 5.2
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
Jablin, R. Environmental Control at Alan Wood: Technical Problems
Regulations^ and New Processes. (Presented at the 79th General Meeting
of the American Iron and Steel Institute, New York. May 26, 1971.)
PEDCo Environmental, Inc. Best Systems of Emission Reduction for
nl~ ! I A" ^e, Ir°n and Stee1 Industry. Background information
Sn PXn9 i°£n T , ;, Environmental Protection Agency under Contract
No. 68-02-1321, Task No. 10. Cincinnati, Ohio. 1977 pp 3-12 3-13
American Iron
Washington, D.C.
and Steel Institute.
1976. p. 70.
Annual Statistical Report.
Katan V.S and R.W. Gerstle. Industrial Process Profiles for Envi-
ronmental Use: Chapter 24. The Iron and Steel Industry, Parsons, T
(ed.) PEDCo Environmental, Inc., Cincinnati, Ohio; Radian Corporation
Austin, Texas. EPA-600/2-77-023x. February 1977. p. 30. IdLlon>
« 5*S'' !u a,Vc Desulfurization of Steel Mill Sinter Plant Gases.
Prepared for the U.S. Environmental Protection Agency. Radian Corpora-
tion, Austin, Tex. NTIS PB-261 922. October 1976. p. 19.
Standards and Engi-
Environmental Impact
of Emission Reduction
U.S. Environmental Protection Agency, Emissions
neering Division. Draft Standards Support and
Statement, An Investigation of the Best Systems
for Sinter Plants in the Iron and Steel Industr
Park, N.C. May 1977. p. 3-20.
Ref. 6, p. 7-8.
^^:^5_n_Vir,0n,me!?tal Protection Agency. Draft of Standards Support and
Statement. Volume I: Proposed National Emission
^
Massey M.J., and R.W. Dunlap. Economics and Alternatives for Sulfur
Removal From Coke Oven Gas. (Presented at the 67th Annual Meeting of
the Air Pollution Control Association. Denver. June 9-13 1974 ) p
5.2-26
-------
11. Ref. 3, pp. 67, 73.
12 US Environmental Protection Agency. Compilation of Air Pollutant
Emission Factors. 2d ed. AP-42. Research Triangle Park, N.C.
February 1976. p. 7.2-2.
13. Ref. 9, p. 48.
14 Midwest Research Institute. Study of Coke Oven Battery Stack Emission
Control Technology, Final Report. Volume I, Collection and Analyses of
Existing Emissions Data. Kansas City, Mo. Prepared for Emission
Standards and Engineering Division of the U.S. Environmental Protection
Agency, Research Triangle Park, N.C. EPA Contract No. 68-02-2609, Task
No. 5. pp. 91-101.
15. U.S. Environmental Protection Agency. Development of Air Pollution
Control Cost Functions for the Integrated Iron and Steel Industry.
EPA-450/1-80-001. July 1979. pp. 2-28 to 2-32.
16 US Environmental Protection Agency. The Population and Character-
istics of Industrial/ Commercial Boilers. EPA-600/7-79-178a. May
1979. pp. 93-102.
17. Gilbert, K. L. Soaking Pit Innovations—Allegheny Ludlum. Iron and
Steel Engineer. 48(7):33-38. July 1971.
18. Katofiasc, T.W. Launching of Ford's 48- x 96-in. Universal Slabbing
Mill. Iron and Steel Engineer. 48(6):49-55. June 1971.
19. Nemeth, E.L., and C.H. Wexler. Phoenix Steel's 160-in. Plate Mill.
Iron and Steel Engineer. 47(7):33-40. July 1970.
20. Easter, H.C. Operations at Inland's New 12-in. Bar Mill. Iron and
Steel Engineer. 49(6):41-56. June 1972.
21. Kinsey, C.J. Republic Steel Corp.'s New 134-in. Plate Mill at Gadsen.
Iron and Steel Engineer. 47(7):56-59. July 1970.
22. Wilthew, R.M., and R.M. Davidson. Youngstown's 84-in. Hot Strip Mill.
Iron and Steel Engineer. 49(5):53-63. May 1972.
23. Richard, N. L. Co.ld Rolled Sheet Expansion at Kaiser Steel. Iron and
Steel Engineer. 50(11):54-59. November 1973.
24. Baggley, G.W. Comparison of Direct-Fired and Radiant-Tube Annealing
Furnaces. Iron and Steel Engineer. 48(6):75-77. June 1971.
5.2-27
-------
25.
26.
27.
28.
29.
30.
31.
32.
33.
United States Steel Corporation.
Steel. 7th ed. New York. 1971,
Ref. 3, p. 73.
The Making, Shaping, and Treating of
p. 611.
Kearney and Company, Inc. Air Pollution Aspects of the Iron Foundry
Industry. NTIS PB-204 712. Chicago, 111. February 1971. rou™ry
U.S Environmental Protection Agency. Standards Support and Environ-
mental Impact Statement. An Investigation of the Best Systems of
Emission Reduction for Electric Arc Furnaces in the Gray Iron Foundry
Industry Draft. Research Triangle Park, N.C. November 1975. PP
o-o to 3-6. ^
American Society for Metals. Metals Handbook. 8th ed. Volume 5-
nnr9JS-?rc Castin9- Part B: Melting and Casting. Metals Park, Ohio.'
PP * OT-O^OT-D .
Ref. 29, p. 338.
PI^+V.- c*' . nal ReP°rt on Screening Study on Cupolas and
Electric Furnaces in Gray Iron Foundries. Prepared for the United
btates Environmental Protection Agency under Contract No. 68-01-0611
15 1975 w-2ttelle C°lumbUS Laborat°ries, Columbus, Ohio. August
Foundry Management and Technology. 107(7):74. September 1979.
Prnrocf1"9 ^TTC^D'D ^ Exhaust Gases from Combustion and Industrial
p£2S?S' * PS"224 861' PrePared for the U.S. Environmental
Protection Agency. Washington, D.C. October 2, 1971. p. VI-63.
34. Ref. 31.
36.
3?*
p. III-ll.
EJectrostatic Precipitator Manual. Northbrook,
p. 156.5.
Ref. 28, pp. C-17 to C-20.
1C.hem1.st7 of Steelmaking Committee, Iron and Steel Division
n91"V°Clety °f the Ameri'can Institute of Mechanical
Stel™ ,HC HPenrKearnth S.teelmakin9 with Supplement on Oxygen in
anri Jo?™?9' P • A^ncan Institute of Mining, Metallurgical,
and Petroleum Engineers, New York. 1964. p. 819.
al T, A-?rcKliyn'. En9ineerin9 and c^ Study of the Ferro-
alloy Industry. U.S. Environmental Protection Agency Research
Triangle Park, N.C. EPA-450/2-74-008. May 1974 p II-l Kesearch
5.2-28
-------
39. Ref. 38, p. VI-48.
40 The Ferroalloy Association. Statistical Year Book, 1977. 1612 K
Street, N.W., Washington, D.C. 20006. p. 2.
41 Jablin R. Expanding Blast Furnace Slag Without Air Pollution. Jour-
nal of the Air Pollution Control Association. 22(3):191-194. March
1972.
42 Rehmus, F.H., et al. Control of H2S Emissions During Slag Quenching.
Journal of the Air Pollution Control Association. 23(10):864-869.
October 1973.
43 Stoehr R.A.4 and J.P. Pezze. Effect of Oxidizing and Reducing Condi-
tions on the Reaction of Water with Sulfur Bearing Blast Furnace Slags.
Journal of the Air Pollution Control Association. 25(11):1119-1122.
November 1975.
44. Ref. 3, p. 59.
45. Ref. 6, pp. 4-5 to 4-8.
46.
47.
48.
49.
50.
51.
52.
53.
Ref. 6, pp. 4-15 to 4-22, 4-56.
Sheldrake, C.W. , and O.A. Homberg. Coke Oven Gas Desulfurization—
State of the Art. (Presented at the 85th General Meeting of the Ameri-
can Iron and Steel Institute. New York. May 25, 1977.) pp. 1-14.
Singelton, A.M., and G. Batterton. Coke Oven Gas Desulfurization Using
the Sulfiban Process. American Institute of Mining, Metallurgical, and
Petroleum Engineers (AIME), Iron Making Proceedings. 1975. p. 604.
Massey, M.J., and R.W. Dunlap. Assessment of Technologies for the
Desulfurization of Coke Oven Gas. American Institute of Mining, Metal-
lurgical, and Petroleum Engineers (AIME), Iron Making Proceedings.
1975. p. 594.
Kirk-Othmer Encyclopedia of Chemical Technology, Standen, A. (exec.
ed.) 2d ed. Volume 19. New York, Interscience Publishers. 1969. p.
383.
Economics and Alternatives for Sulfur
Journal of the Air Pollution Control
October 1975.
Massey, M.J., and R.V. Dunlap.
Removal From Coke Oven Gas.
Association. 25(10): 1019-1026.
Ref. 49, pp. 588, 594.
Ludberg, J.E. Removal of Hydrogen Sulfide from Coke Oven Gas by the
Stretford Process. (Presented at the 67th Annual Meeting of the Air
Pollution Control Association. Denver. June 9-13, 1974.) p. 3.
5.2-29
-------
54.
55.
56.
57.
Gray and Ductile Iron Founders' Society
Cleveland, Ohio. 1967. p. 50.
Ref. 5, pp. 77, 80.
Ref. 10, pp. 20-22, 27, 28, 35.
Ref. 6, pp. 7-24, 7-25.
Cupola Emission Control.
5.2-30
-------
5.3 PETROLEUM REFINERIES
Petroleum refineries convert crude oils, various intermediate petroleum
fractions, and light gases into useful products. These components are
refined by various physical, thermal, catalytic, and chemical processes into
liquified petroleum gas (LPG), gasoline, kerosene, aviation fuel, diesel
fuel, fuel oils, lubricating oils, waxes, tars, asphalts, coke, and petro-
chemical feedstocks. Most refinery products are not pure chemical compounds
but are mixtures of compounds. Some refineries also manufacture pure petro-
chemicals such as benzene, toluene, and cyclohexane.
Because each refinery is designed to process specific crude oils, no
refinery is typical. Most U.S. refineries, however, are designed to maxi-
mize production of gasoline. The following are basic operations in refining
of crude oil: 1) separation processes, which separate the crude oils to
isolate the desired products (e.g., distillation); 2) decomposition proc-
esses, which break large molecular chains into smaller ones by cracking
(e.g., catalytic cracking, coking); 3) formation processes, which build the
products by chemical reaction (e.g., reforming, alkylation, isomerization);
4) treating processes, which remove impurities or compounds that make the
products environmentally unacceptable or are detrimental to operation of the
refinery; 5) recovery operations (e.g., sulfur recovery, fuel gas recovery);
6) storage; and 7) auxiliary facilities. Figure 5.3-1 depicts a typical
petroleum refinery.1 Because of the complexity of the various processes and
the individuality of each refinery, intermediate storage may be needed for
certain fractions that will later undergo further processing.
As of January 1979 there were 303 operating petroleum refineries in the
United States with total refining capacity estimated to be 3.0 x 109 liters
[18 million barrels (bbl)] per calendar day.2 In some urban areas of the
United States there are several refineries with a combined crude processing
rate of over 1.6 x 108 liters per day (1 million bbl/day). Refinery pro-
cessing during 1977 resulted in sulfur dioxide (S02) emissions estimated at
800,000 Mg (880,000 tons), or approximately 2.9 percent of total S02 emis-
sions in the United States.3
In some areas considerable effort has been made to control S02 emis-
sions, and many modern refineries have, of necessity, integrated air pollu-
tion control into the plant operations.
5.3-1
-------
t vita iron was
„
31
SQN3 iWJJl 13MS
S3SS33QW MOW SON3 iffflll 133HS
SB
lit
SONS IHOn 133HS
sow IHDII anos
S3
3S
3KITOSW 3IUW1V3
U10SVS 31UTV1V3
sows iron anos
SOH3 1H311 HflOS
S0»3 JHOI1 B(I05
M"
SQN3 1H9I1 WWS
wnos
dg
53
Si
g»
r!
gsl
j
iinun OJHSIHI j ~
M*
I
t
O)
•r—
q-
OJ
E
OJ
O)
Q.
O
•t—
a.
co
a.
cu
4->
to
CD
00
CO
O
O
a.
S-
o
CO
*
LO
5.3-2
-------
Removal of sulfur from some refinery streams is a part of refining. It
would be desirable to remove all sulfur compounds before any processing of
the crude begins, but because this is impractical, sulfur is removed
throughout the refining process. There are several reasons, besides air
pollution control, for removing sulfur from intermediate fractions and
products of crude oil. Sulfur removal reduces corrosion, odor, occurrence
of breakdowns, catalyst poisoning, and gum formation and improves octane
rating, color, and lube oil life.4
5.3.1 Process Descriptions and Emissions Sources
Most oil refinery processing units are made up of at least five main
types of equipment: heaters, reactors, vessels, heat exchangers, and pumps.
The arrangement, type, and quantity of equipment are selected to fit the
specified function. .
5.3.1.1 Refinery Combustion Processes—
In many instances refinery S02 emissions come from organic sulfur
compounds in the fuel burned as energy sources for process heaters and
refinery boilers. Almost every major processing unit in an oil refinery
includes one or more process heaters, fired with such fuels as refinery gas,
natural gas, and heavy residual fuel oil. Sulfur dioxide concentrations
ranging from 700 to 1000 parts per million have been measured in flue gas
resulting from burning heavy residual fuel oil.5 The S02 concentrations in
the flue gas of boilers and heaters depend entirely on the sulfur content of
the fuel.
5.3.1.2 Coking Facilities--
Coking is a severe form of thermal cracking. The feedstock is a
residual that may resist cracking by other methods. Coking is an ultimate-
yield destructive distillation process, which produces gas, distillate, and
coke. Coking is done by two principal processes, the fluid and the delayed.
In fluid, coking the feed is sprayed into a reactor containing a flu-
idized bed of preheated recycled coke particles. The hydrocarbons in the
liquid feed crack and vaporize while the nonvolatile material is deposited
on the fluidized coke particles. As the size of the coke particles
increases, they sink to the bottom of the reactor and flow to a burner. In
5.3-3
-------
the burner, the particles are fluidized with air, partially burned, and
recycled into the reactor. A portion of the coke produced in the reactor is
withdrawn as product. The flue gas from the burner is discharged through
the stack after passing through a cyclone.
The vaporized products formed in the reactor are removed from the bed,
passing through cyclones that remove some of the coke to the bottom of the
scrubber. The heavier gases are condensed, forming a slurry with the coke
dust, which is recycled to the reactor. The remaining gases pass to the
fractional on zone of the scrubber. The heavy distillate is withdrawn to be
fed to the catalytic cracker, and the other gas and distillate products are
removed from the system.
The residual feedstock to the coker usually contains the highest weight
percent of sulfur of the products leaving the main distillation fractiona-
tion column. Thus the various coker products contain sulfur-bearing
compounds and must be treated further downstream. The major source of S02
emissions in the fluidized coking process is flue gas from the burner. The
quantity of emissions depends on the sulfur content of the fuel and the
coke.
In delayed coking the coker feed is fed to the bottom section of a
fractionator, where the lighter fractions are flashed off. The remaining
material is mixed with a recycle stream of heavy products and pumped to the
coking heater. The heated feed is passed to the coking drum, where cracking
occurs. As the process progresses, the cracked products are removed at the
top and coke forms along the inner surface of the drum. The major S02
emission point in the delayed coking process is from the heater, depending
on the type of fuel used. A liquid waste stream containing hydrogen sulfide
(H2S) is drawn from the overhead accumulator on the coker tower. This
stream is pumped to the sour water stripper for H2S removal. Flaring of
gases from blowdown of coke drums will also release S02. Some cokers used
closed systems or amine scrubbers for S02 control.
5.3.1.3 Cracking Processes—
Cracking of large hydrocarbon molecules into smaller ones is accom-
plished by the application of heat and/or catalysis. At the same time some
of the cracked molecules recombine to form larger molecules; the result is
5.3-4
-------
formation of a synthetic crude that can be separated into gaseous hydro-
carbons, gasoline, gas oil, and fuel. The two kinds of cracking are thermal
and catalytic.
Thermal cracking, at high temperature and pressure, is generally
applied to distillates heavier than gasoline. In addition to coking opera-
tions, visbreaking is an example of a thermal cracking process.
Visbreaking, a milder form of thermal cracking than coking, is used to
reduce the viscosity of some residual fractions so that they may be blended
into fuel oils. The reduced crude is preheated by heat exchange with vis-
broken fuel oil and fed to a furnace. Mild cracking in the furnace tubes
produces a mixture of residual oil, naphtha, and gas. The reaction products
are quenched and fractionated in a steam distillation tower. A sour water
waste.stream containing H2S is withdrawn feom the fractionator and sent to a
sour water stripper for processing.
Catalytic cracking uses high temperatures and chemical catalysts to
crack the molecules of gas oils into lighter gasoline material. Two well-
known catalytic devices are in use today, the fluid catalytic cracker (FCC)
and the thermofor catalytic cracker (TCC). The TCC is no longer generally
manufactured. In the FCC, preheated feedstocks are introduced into the
bottom of a riser with regenerated catalyst. Most cracking occurs in the
riser, and the catalyst is separated from the gaseous reaction products in
the reactor vessel. The reaction products flow on to cooling and separation
processes. Spent catalyst falls through a steam stripping section (to
remove volatile hydrocarbons) and into a regenerator. Controlled combustion
of the coke on the catalyst is carried out in the regenerator, and regener-
ated catalyst is returned to the riser to complete the catalyst cycle.
Regenerator flue gas is usually sent to a carbon monoxide (CO) boiler for
waste heat recovery before discharge to the atmosphere. The boiler recovers
heat from the oxidation of CO to carbon dioxide (C02) while reducing emis-
sions of CO. Newer FCC unit designs do not require CO boilers. This flue
gas is the major source of S02 emission from the catalytic cracker. Sulfur
dioxide concentrations in the FCC flue gas range from 150 to 3000 ppm,6 de-
pending on the amount of sulfur in the feedstock and on operating condi-
tions.
5.3-5
-------
5.3.1.4 Hydrocracking—
Hydrocrackers perform both cracking and hydrogenation and are used to
convert heavy feedstocks into lighter, more valuable products. Hydrocrack-
ing is done at high pressure and temperature, with a special catalyst and
hydrogen. The reaction section is usually divided into two stages. The
first stage is designed to remove sulfur and nitrogen compounds by hydro-
genating them to hydrogen sulfide and ammonia in a fixed-bed reactor The
second stage accomplishes the actual cracking of the feedstock. The stream
from the second stage reactor is fed to a fractional, where the desired
products are separated and recovered. The hydrocracker generates liquid
waste streams containing dissolved H2S from the separators and accumulator
The H2S in these streams is usually removed in a sour water stripper and
further processed in a sulfur recovery unit.
5.3.1.5 Alkylation and Spent Acid Regeneration-
Refinery alkylation is the chemical combination of two light hydro-
carbon molecules (an olefin and an isoparaffin) to form one molecule that is
in the gasoline boiling range and exhibits good octane characteristics The
feedstocks are catalytically reacted over either anhydrous hydrofluoric acid
or sulfuric acid to produce a high-octane component known as alkylate The
reactor effluent is separated into hydrocarbon and acid phases in a settler
The acid is returned to the reactor. From this point on, the alkylate is
treated differently in the two processes. In the hydrofluoric acid process
the alkylate and excess isoparaffin are sent to a stripper for separation'
An alkylate bottoms stream is charged to a fired heater to decompose any
organic fluorides that may have formed. The remaining alkylate from the
fired heater is the finished product. In the sulfuric acid process the
hydrocarbon liquid from the settler is washed with caustic and water before
being fractionated. The alkylate product is then separated from the
isoparaffin, which is returned to the feed stream.
Because the alkylation process is a closed system with no process vents
to the atmosphere, S02 emissions are negligible. Liquid wastes associated
with the water and caustic scrubbing in the sulfuric acid process are
generally treated in the refinery wastewater treatment plant. The hydro-
fluoric acid system produces a sludge waste from the bottoms of the acid
regenerator that is readily burned as fuel.
5.3-6
-------
Some refineries regenerate the spent sulfuric acid used in alkylation
and treating processes. The spent acid is burned with elemental sulfur
and/or hydrogen sulfide gas to form sulfur dioxide. The S02 and other
combustion products are passed through gas-cleaning and mist-removal equip-
ment, then through a drying tower and on to a sulfur trioxide converter.
The S03 is absorbed in a circulating stream of concentrated sulfuric acid,
which has a concentration of 98 to 99 percent. The nonabsorbed tail gases
pass overhead through mist-removal equipment to the exit gas stack. Air-
borne emissions from spent acid regeneration include S02 and acid mist.
Acid mist is formed when S03 combines with water vapor at a temperature
below the dewpoint of H2S04. Sulfur dioxide emissions, an inverse function
of the sulfur conversion efficiency, can be as high as 35 kg/Mg (70 Ib/ton)
of acid produced (95 percent conversion). Acid mist emissions range from
1.1 to 1.4 kg/Mg (2.2 to 2.7 Ib/ton) of acid produced.7 Sulfur dioxide
emissions can be controlled by increasing the plant conversion efficiency or
by adding a sodium sulfite-bisulfite scrubbing process.8 Acid mist emis-
sions are reduced by passing the exit gases through an electrostatic precip-
itator (ESP) or fiber mist eliminators.
5.3.1.6 Oil Desulfurization Processes--
Hydrotreating is used to remove sulfur from all types of petroleum
products, eliminate other impurities such as nitrogen and oxygen, decolorize
and stabilize products, and correct odor problems and many other product
deficiencies. This widely used process consists of bringing oil charge
stock and hydrogen into a fixed-bed, catalytic reactor at high temperature
and pressure. Hydrogen reacts with sulfur, nitrogen, oxygen, and olefinic
hydrocarbons to form removable hydrogen sulfide, ammonia, saturated hydro-
carbons, and water. The process gas is rich in hydrogen, hydrocarbons, and
hydrogen sulfide. The H2S can be extracted from the stream and converted to
elemental sulfur or sulfuric acid. The catalyst in a hydrodesulfurization
process is regenerated periodically to remove built-up coke. During regen-
eration, a steam-air mixture burns off the undesirable carbon buildup;
sulfur dioxide is released to the atmosphere during this process or is
contained in a closed cycle regeneration system.
5.3-7
-------
5.3.1.7 Sulfur Recovery Units—
Refinery sour gas streams are generally fed to a regenerative type of
hydrogen sulfide removal process. The concentrated acid gas is then sent to
the sulfur recovery unit (SRU). The Claus process (developed in about 1890)
is the most widely used method of producing sulfur from refinery hydrogen
sulfide. The modified Claus process (developed in about 1937) is based on
producing elemental sulfur by first converting one-third of the hydrogen
sulfide feed by precise combustion with air to achieve the following
reaction:
2H2S + 302 -»• 2S02 + 2H20.
The above products of combustion are then allowed to react thermally with
the remaining two-thirds of the hydrogen sulfide feed- in the presence of a
suitable catalyst to form sulfur vapor:
2H2S + S02 -> aS2 + bS6 + cS8 + 2H20.
The letters "a," "b,» and "c" represent the number of moles of the various
possible molecular forms of sulfur vapor. The sulfur is recovered by
cooling the gas to condense the sulfur. In addition to these reactions
some sulfur is produced directly by dissociation of hydrogen sulfide:
H2S -> H2 + 1/2S2.
This is a minor reaction, however, and does not contribute appreciably to
the overall sulfur recovery.
Figure 5.3-2 shows a typical Claus process employing both noncatalytic
(thermal) and catalytic reaction. Generally about 50 to 60 percent of the
feed sulfur is recovered in the sulfur condenser following the noncatalytic
or thermal reaction. Following the oxidation (combustion) reaction in the
thermal reactor, the hot gases are fed to a waste heat boiler, where steam
is generated. The cooled gases from the waste heat boiler are fed to the
first sulfur condenser, where the elemental sulfur made in the thermal
reactor is condensed. The uncondensed gases leave the first sulfur con-
denser and are heated prior to being fed to the first catalytic reactor.
The gases must be heated above the sulfur dewpoint to prevent sulfur con-
densation on the catalyst and to obtain the optimum sulfur recovery in the
5.3-8
-------
-(->
•I—
3
O)
O
O
S-
M-
3
co
CM
I
CO
s_
3
CD
5.3-9
-------
reactor. A fixed-bed catalytic reactor containing activated alumina (A1203)
catalyst is used to improve the sulfur recovery. The sulfur-laden gases are
then fed to another sulfur condenser for recovery of the molten sulfur.
This reheating, reaction, condensing cycle is repeated with each catalytic
reactor on the Glaus unit.
Following the sulfur condensation step, the remaining gases are fed to
an incinerator, in which all sulfur compounds in the tail gas are converted
to S02 by combustion before being discharged to the atmosphere through a
stack.
The overall sulfur recovery obtainable with a Claus plant is dependent
on the number of catalytic reactors, hydrogen sulfide concentration in the
feed, degree of carbon dioxide and hydrocarbon contamination in the feed
gas, and outlet temperature of the sulfur condensers. In the past, sulfur
recovery units were operated with only one catalytic reactor. This practice
is not commonly used today because with only one catalytic reactor the
overall sulfur recovery is limited to about 75 percent. A minimum of two
catalytic reactors is required to provide overall sulfur recovery of about
90 percent, and the maximum level of overall sulfur recovery with three
catalytic reactors is considered to be 97 percent.9 Although some Claus
units have been built with four catalytic reactors, this generally is not
done because of the reaction limitations.10
The Claus unit tail gas contains sulfur dioxide, hydrogen sulfide,
elemental sulfur, carbonyl sulfide, and carbon disulfide. Incineration of
the tail gas converts all of the sulfur to S02 before it is emitted to the
atmosphere through the stack. The concentration of S02 'in the stack gases
depends on the overall sulfur recovery in the Claus unit, the incineration
temperature, and the amount of excess combustion air. The S02 concentration
can range between 3,000 and 20,000 ppm, but normally averages 8000 to
12,000 ppm.9
5.3.1.8 Other Sulfur Oxide Emission Sources--
Separation of a mixture of light and heavy hydrocarbons into fractions
or intermediates of a specified range of boiling temperatures is usually
done by distillation and steam stripping. The crude charge is separated
into several petroleum fractions within the fractional. Several liquid
5.3-10
-------
side streams are withdrawn from the fractionator at different elevations
within the tower. The fractions are charged to the side-stream product
stripper, where lighter hydrocarbons are stripped and returned to the frac-
tionation tower. The stripping medium is usually steam or light petroleum
gas. In addition to the side-stream strippers, the crude distillation
column has a bottoms stripping zone, in which lighter hydrocarbons are
steam-stripped from the residual product. Almost every major processing
unit in the refinery includes a distillation section. Atmospheric distilla-
tion is a closed process with only fugitive air emissions. Sour water
containing sulfides is produced from the condensed stripping steam and is
sent to the sour water stripper for removal of hydrogen sulfide.
Catalytic reforming units are used to convert low-octane naphthas into
high-octane blending stocks to be used in production of gasoline. Reforming
is accomplished by rearranging the molecular structure of the feedstock.
Reformer feedstock in the presence of hydrogen reacts over a platinum-
rhenium catalyst. Hydrogen is produced and partly recycled to the reactor;
the excess is used in hydrogen treating units for sulfur removal and product
improvement. This process unit is a closed system. Although S02 emissions
can occur during catalyst regeneration, this happens rarely, and the emis-
sions are considered negligible. Some catalytic reforming units incorporate
a continuous catalyst regeneration system. Total emissions, including S02,
average from 0.006 to 0.06 g/liter (0.002 to 0.02 Ib/bbl)11 and are con-
sidered negligible.
Waste gas from a refinery is generally odorous and can be handled by
one or more flare systems. The sulfur content of the waste gas to each
flare system depends on its source, because it can come from one or more
refinery operating units. The combustible composition of waste gas and the
temperature in the combustion zone determine the effectiveness of the con-
trol system. Sulfur dioxide and other injurious substances in hydrocarbon
waste gases should be removed by some type of absorption-system before going
to a flare. Absorption systems for relief valves, however, may not be
practical because of the erratic nature and rates of release.
Vacuum distillation separates the atmospheric residue from the main
fractionator into a heavy residual oil and one or more heavy gas oil
5.3-11
-------
streams. Vacuum fractionators are maintained at approximately 100 mm
mercury (Hg) absolute pressure by either steam ejectors or mechanical vacuum
pumps. Noncondensable vapors removed by these systems must be discharged.
The vapor emissions, containing as much as 25 percent hydrogen sulfide by
volume, may be as high as 400 g/1000 liters (130 lb/1000 bbl) of vacuum unit
charge. In addition, aqueous wastes containing hydrogen sulfide result from
condensation of steam used for stripping during vacuum fractionation and for
maintaining fractionator vacuum by ejectors.
Asphalt from the crude refining unit can be made into roofing asphalt
by subjecting it to air blowing at elevated temperatures. Air is passed
through the charge in a steam-blanketed still at an approximate rate of
1.2 mVrain per Mg (40 ftVmin per ton) of charge until the desired hardness
is achieved. The overhead gas, rich in hydrocarbon vapors, is sent to a
knockout drum. Gas from the knockout drum flows to an incinerator. The
effluent stream contains negligible amounts of various sulfur compounds.
Treating is used in refinery processing to remove undesirable impuri-
ties such as sulfur, nitrogen, and oxygen to improve product quality.
Emissions from treating operations consist of sulfur dioxide, hydrocarbons,
and visible plumes. Emission levels depend on the methods used in handling
spent acid and acid sludges, as well as in recovering or disposing of hydro-
gen sulfide. As refineries continue to process increasing amounts of high-
sulfur crudes, chemical treating of final products is being replaced by
hydrogen desulfurization of many feedstocks.
5.3.2 Control Techniques
5.3.2.1 Description-
Four processing areas within a refinery generate the major sulfur
emissions:
1) Process heaters and boilers
2) The fluid catalytic cracking unit
3) The sulfur recovery unit
4) Flares burning H2S streams.
5.3-12
-------
Process heaters and refinery boilers are integral parts of almost every
processing unit in a refinery. Fuel requirements for heaters and boilers
range between 5 and 10 percent of the heating value of the crude that enters
the refinery.12 Sulfur dioxide emissions depend on the sulfur content of
the fuel. Concentrations of S02 emitted from heaters and boilers can be
reduced by burning low-sulfur fuel oil, low-sulfur process gas, or natural
gas. Flue gas desulfurization techniques are not currently applied to
process heaters or boilers, but are described in Section 4.2.
The removal of S02 from the regeneration gases of FCC and TCC units is
not widely practiced. Studies in the literature for reduction of S02 emis-
sions from FCC regeneration flue gas suggest several alternatives:13
1) Desulfurization of FCC feed.
2) Flue gas desulfurization techniques.
3) Specific catalysts for S02 emission control.14
Desulfurization of FCC feed is practiced on a limited basis and usually is
done to improve gasoline yield and reduce coke buildup on the catalyst.
A flue gas desulfurization system, developed and marketed by Exxon, is
currently used in four Exxon FCC units. This is an application of a jet-
ejector liquid scrubber for simultaneous removal of sulfur oxides and parti-
culates.15 The schematic for Exxon's jet ejector scrubbing system is shown
in Figure 5.3-3. An alkaline scrubbing medium, usually a sodium solution,
passes through a spray nozzle, which breaks the liquid stream into droplets.
The flue gas is drawn into the body of the scrubber by the draft-inducing
action of the liquid spray. The gas is intermixed with the scrubbing
liquid, and both enter the venturi section of the scrubber. The intense
turbulence in the venturi section causes the liquid droplets to strike and
capture the particulates .in the gas stream, and the S02 in the flue gas is
absorbed in the liquid.
The mixture of gas and liquid droplets is sent to a separator, where
the clean gas is separated from the contacted liquid and discharged to the
atmosphere. The purge stream from the scrubber will contain suspended
solids (catalyst) and soluble sulfite and sulfate salts. It is necessary to
treat the scrubber bleed before it can be discharged. The water treatment
should include removal of insoluble salts, reduction of chemical oxygen
demand, and reduction of soluble salts.
5.3-13
-------
FLUE GAS
AIR
BLOWER
:AUSTIC
STORAGE
TANK
POLYMER
ADDITION
SYSTEM
SLURRY
pH MONITOR
CLARIFIER
SULFITE
OXIDATION
REACTOR
0
WATER EFFLUENT
DISCHARGE
*• TO
PONDING
FOR
TREATMENT
AIR COMPRESSOR
Figure 5.3-3. Process layout of the venturi scrubbing system.
5.3-14
-------
At locations where the FCC.regenerator is operated at hot regeneration
or where sufficient pressure is available, the jet-ejector Venturis can be
replaced by high-energy venturi scrubbers.
Pilot plant performance data show efficiencies of 95 to 99 percent for
removal of sulfur oxides from an inlet gas stream with 200 to 500 ppm S02
and 85 to 95 percent for removal of particulates.15 Recent sulfur dioxide
sampling tests conducted by the Texas Air Control Board (TACB) on tU *-..-
FCC units at the Exxon Baytown Refinery indicate a scrubber efficiency of
94 percent for S02 removal from an inlet gas stream with 900. to 1040 ppm
SO,
16
Special cracking catalyst composition can also reduce regenerator flue
gas sulfur oxide emissions by adsorbing S02 onto the catalyst surface.
Basically, S02 is oxidized to S03 in the regenerator, absorbed on the
catalyst surface as the sulfate (S04), and then carried into the reactor
where it is reduced in the cracking reaction to H2S; the H2S then exits the
system with the cracked products. Special catalysts used to achieve S0x
reduction in both pilot unit and commercial trials have shown flue gas S02
reductions of up to 88 percent.17
Tail gas from the sulfur recovery unit is a major sulfur source in a
refinery and usually requires S02 control. Emissions can be as high as
20,000 ppm, but normally average 10,000 ppm. Air quality control restric-
tions normally require the use of a Claus plant tail-gas treating unit for
further reduction of S02 emissions. Operational tail-gas control systems
include the IFP-1500 process, the Sulfreen process, the Beavon process, the
Stretford process, the Shell Claus off-gas treating (SCOT) system, and the
Wellman-Lord process.
IFP-1500 process—The IFP-1500 process converts mixed hydrogen sulfide/
sulfur dioxide streams to sulfur and water by a liquid-phase Claus reaction
using a proprietary catalyst. The process is primarily used to clean Claus
unit tail gas. The technology is an extension of the Claus reduction pro-
cess, but is carried out in the liquid phase. In this process the tail gas,
at Claus unit exit pressure, is injected into the bottom of a packed tower,
where the packing provides necessary surface area for gas-liquid contact. A
low-vapor-pressure polyethylene glycol solvent containing a proprietary
carboxylic acid salt catalyst in solution circulates counter-currently to the
gas.
5.3-15
-------
The catalyst forms a complex with H2S and S02> which in turn reacts
with more of the gases to regenerate the catalyst and form elemental sulfur
The reaction is exothermic, and the heat released is removed by injecting
and vaporizing steam condensate. Temperature is maintained at about 120° to
132°C (250° to 270°F), high enough to keep the sulfur molten but not high
enough to cause much loss of sulfur or glycol overhead. The sulfur accum-
ulates in the boot of the tower and is drawn off continuously through a seal
leg. Overhead gases from the IFP-1500 unit are incinerated.
The vendor claims that the IFP-1500 process is insensitive to changes
in gas flow rates. It has been shown to operate at flows as low as
30 percent of design without adverse effect. Another stated advantage is
maintenance-free operation for about 24 continuous months, after which the
unit is shut down to wash away spent catalyst that deposits on the packing
material. A water wash is all that is required. Outlet S02 concentration
is 1000 to 2000 ppm.18
Sulfreen process-The Sulfreen process reduces the sulfur content in
Claus plant tail gas by further promoting the Claus reaction on a catalytic
surface in a gas/solid batch reactor. Claus tail gas is first scrubbed with
liquid to wash out entrained sulfur liquid and sulfur vapor. The tail gas
is then introduced to a battery of reactors, where the 'Claus reactions are
carried out at lower temperatures (127° to 149°C, 260° to 300°F) than those
utilized in the sulfur plant.
A regeneration gas, essentially nitrogen, periodically desorbs the
sulfur-laden catalyst beds. Nitrogen is heated and cycles through the
catalyst bed at approximately 300°C (570°F) until all water and CO, are
driven off.
The process reduces entrained sulfur, because the catalyst acts as an
absorbent for liquid sulfur. Tail gas passing through the catalyst bed
however, retains several hundred parts per million of sulfur vapor in
equilibrium with liquid sulfur. The H2S and S02 are reduced by 80 to 85
percent to levels of about 1800 ppm H2S and 900 ppm S02. "As with the
IFP-1500 process, the levels of H2S and S02 are highly dependent upon '
maintaining the 2:1 ratio of H2S to S02 in the Claus tail gas. Carbonyl
sulf!de and carbon sulfide are not affected by the Sulfreen process
5.3-16
-------
Beavon process—The Beavon process is a reduction- type tail-gas treat-
ment system. As a first step, all sulfur compounds in the Claus tail gas
are converted to hydrogen sulfide. This process takes place in a fixed-bed
reactor, in which a cobalt-molybdenum catalyst enhances the reaction.
Before the reactor, however, a fuel gas stream is combusted in an in-line
burner and mixed with the Claus tail gases to provide a reducing atmosphere.
Hydrogenation and hydrolysis reactions reduce all sulfur (as carbonyl
sulfide, carbon disulfide, sulfur, and S02) to hydrogen sulfide.
The following reactions are involved in this step:
8H
8HS
S02
COS
CS2
CS2
4H2
COS +
+ 3H
+ H2
+ H2
+ 2H
+ CS
4H2 -»
2
0
0
2
2
•* H2S
-"•H2S
-> H2S
S -> CH
-»• 2H2
H20 +
4
S
H2
2H20
C02
COS
+ 4S
+ CH4
S + C
The stream from this reaction step, rich in hydrogen sulfide, is dewatered
by direct contact cooling in a quench tower. The sour water must be cleaned
in a sour water stripper facility. The hydrogenated tail gas is fed to a
Stretford unit for sulfur recovery.
Stretford process— Although the Stretford process may be capable of
solely replacing Claus or amine sulfur removal processes, it has -gained more
recognition as part of the Beavon tail-gas cleanup process.
Sour fuel gas or Claus tail gas is fed to the hydrogen sulfide absorber
column that contains "Stretford solution." This is a solution of sodium
vanadate (NaVOg) and sodium carbonate (Na2C03). the hydrogen sulfide is
absorbed by the sodium carbonate in the reaction:
H2S + Na2C03 •*' NaHS + NaHC03.
It is precipitated by reaction with the NaVO^:
2NaV03 + NaHS + NaHC03 •* S + Na2V205 + Na2C03 + H220.
The sodium vanadate is regenerated by use of an acid called ADA (antra-
quinone disulfonic acid) by the reaction:
5.3-17
-------
Na2V205 + ADA (oxidized) -» 2NaV03 + ADA (reduced).
The ADA is regenerated to an oxidized state by bubbling air through the
solution. After the absorption step, the Stretford solution is held in a
tank for a period of time to promote sulfur precipitation. As a result of
the air bubbling step, the sulfur is floated to the surface of the solution
and the froth overflows into a settling tank. Sulfur is recovered from this
tank as a sludge, which is filtered to form sulfur cakes; the cakes are
dewatered in an autoclave separator. The resultant liquid sulfur is removed
for storage and/or sale. Overall sulfur recovery of the Claus process plus
the Beavon-Stretford process is over 99.9 percent. Typical sulfur emis-
sions are less than 50 ppm.19
SCOT process-The SCOT system is a reduction-type S02 control system
The f,rst commercial-scale SCOT units in the United States began operations
in early 1973.
In the initial phase of treatment the Claus tail gas is subjected to
reasons that convert all of the sulfur-containing compounds to hydrogen
sulfide. This initial phase is identical to that of the Beavon process for
controlling Claus emissions. An in-line burner heats the tail gas and a
reducing atmosphere is maintained. A cobalt-molybdenum (on alumina)
catalyst in the fixed-bed reactor enhances conversion of the materials to
hydrogen sulfide by the following reactions:
S + H2 -»• H2S
S02 + 3H2 -> H2S + 2H20
COS + H20 ^.H2S + C02
CS2 + 2H20 -> 2H2S + C02.
The gas is cooled to near ambient temperature by direct contact with
water (or air) in a packed quenching tower. Sour water condenses from the
stream of H2S-rich gas and is withdrawn. The sulfur in this sour water must
be removed before the water can be disposed of. An onsite sour water
stripper may be used for this purpose; or a stripper may be installed as an
integral part of the SCOT system.
5.3-18
-------
The cooled gas stream is piped to a tray tower absorber, where di-
isopropanolamine absorbs the hydrogen sulfide and carbon dioxide from the
stream. The amine, laden with hydrogen sulfide, is sent to a regenerator,
which is normally a conventional steam stripping column. The regenerator
off-gas (hydrogen sulfide and carbon dioxide) is recycled to the Claus plant
as feed material. Absorber off-gas containing less than 300 ppm hydrogen
sulfide is incinerated before release through the stack.
Wellman-Lord process—In the Wellman-Lord process, the S02-rich gas is
stripped of S02 in a countercurrent absorber containing a sodium sulfite
solution. The spent solution, rich in bisulfite, is discharged to a surge
tank and then pumped to a proprietary evaporator/crystal!izer in the regen-
eration section. Low-pressure steam is used to heat the evaporator and
to drive off S02 and water vapor. The sodium sulfite precipitates as a
dense slurry of crystals.
The gas stream leaving the evaporator is subjected to partial condensa-
tion to remove most of the water vapor before the product S02 is discharged
from the process. The final product S02 can be delivered at whatever
quality is required for further processing. It is suitable for conversion
to high-grade sulfuric acid or elemental sulfur.
The condensate is mixed with the sulfite slurry stream withdrawn from
the evaporator and is used for redissolving the slurry. The sulfite-lean
solution is them pumped to a surge tank and fed back to the absorber.
The process is based on a sodium sulfite/bisulfite cycle. The reac-
tions in the process can be abbreviated for simplicity as follows:
Absorption S02 + Na2S03 + H20 -» 2NaHS03
Regeneration 2NaHS03 -» Na2S03-v + S02t + H20t.
Apart from the two major reactions above, sodium sulfate (Na2S04),
which is nonregenerable, is formed in the absorber as a result of solution
contact with oxygen and sulfur trioxide. The sodium sulfate so formed is
controlled by maintaining a continuous purge from the system. A makeup of
caustic is required to replace that lost in the purge stream. The absorber
off-gas normally contains less than 250 ppm S02 and is vented to the
atmosphere.20
5.3-19
-------
5.3.2.2 Control Cost-
Marketing brochures published by Exxon have compared the cost of a flue
gas scrubbing unit with the cost of a feed hydrodesulfurization (HDS) unit/
electrostatic precipitator for a fluid catalytic cracking unit with a
capacity of 13.3 x 106 liters (80,000 bbl) per day at a Gulf Coast
refinery.21 Independent estimates adjusted to mid-1979 dollars indicate
that the base case of a feed HDS unit/ESP would require $7,000,000 for
capital investment and $800,000 for annual operating expenses.22"25 Exxon
estimated that the capital investment for the jet ejector flue gas scrubbing
system, including a heating oil HDS unit, was two-thirds the capital invest-
ment required for the base case of a feed HDS unit/ESP and that the annual
operating cost of the flue gas scrubbing system was one-half the annual
operating expense of the base case. The investment and operating costs
would be lower if a high-energy venturi were used in place of the jet
ejector venturi. The operating costs did not include increased gasoline
yield credits resulting from the improvement of feed quality gained through
HDS. Actual capital and operating costs of the Exxon units are considered
proprietary information and were not divulged.
Because of legislation requiring the reduction of S02 emissions from
sulfur recovery units, most refineries must provide tail-gas treating units..
The cost of recovering this incremental sulfur is high and varies from 50 to
over 100 percent of the cost of a new Claus unit.2^ Tables 5.3-1 to 5.3-3
show the typical capital and operating costs of tail-gas treating units.
These estimates are not firm because costs vary with location, process
differences that affect energy requirements, need for chemicals and
supplies, and regenerability of the catalyst.
5.3.2.3 Energy and Environmental Impact27—
Refinery sulfur plants are major point sources of sulfur dioxide emis-
sions within petroleum refineries. These emissions result from treatment of
the gases produced by various sulfur removal processes within the refinery.
A typical uncontrolled sulfur plant (two- or three-stage Claus plant)
recovers about 95 percent of the sulfur in the incoming gas stream. Thus,
emissions of sulfur dioxide are usually in the range of 8,000 to 10,000 ppm.
5.3-20
-------
00
o
I—I
00 rH
CO J2 C
.c: re re E i— c
\ _i .c c 4-> 4-
OO3: in CD
r- -^ M- co o c
. Jj3 ->x O > CD C
r- o m c rH c
•W- "3" CD CO *r- *--^ i —
•P CD ~~-
4J 4-> O M- <*- O C
re re -fee- o • tr
+J CD CO t/=
i)_ ^-, m .p o S^ -w-
.,- i co re CD co 4-
^ Q 05 rH 4-> n
in -~^ c s- \ 4-> re
•i- CO CD TO 4-
t- S- S- f <=> 4J -r
co o c u
2 +-> t. re co ^
oo • 4-> 4-> E :
D.CO>OI— ere*
C S- Q.CO O) -r- COr-
s:QooLuu_s:tnt/
0 0
o o
rH rH
oo m
in rH
en to
rH
0 O
l-H 0
00 «=)-
oo cn
CM en
rH rH
^ '
^
O O
0 0
tD CO
oo r---
CM UD
- — ' rH
U
4-
C
a
t
4-
ir
a
-3
> c
> »r
15 <*-
: c
>
&<
Is-
) r-
5
>• 4-> 4-
in n
> o
3 o o
a
* 01 c
C i
3 T- n
L) 4-> j:
re t
j i-
O) r-
_ Q. r
304
— *r
- 4-> £
3 CO r
0 ^ C.
0 0
0 0
O O
CO 00
CO tD •
i-H CO
tf) CO
CD O
01 o
m o
CD CO
r- Is--
rH
j—J"
<=e- •««-
o o
0 0
r- CD
00 -=f
CO 00
rH CD
"
)
)
J
i
>
3 ' 4->
in
1 O
L) 0
35
o re
; 3 4->
J C EZ
c: co
- re E
a 4->
j i — in
re co
a. 4-> >
DOC
.5 H- HH
01
CD
X
re
^
i-
m
CO
E
3
in
t/)
^^
4J
re
in
S-
re
CO
in
rH
i.
o
If—
i-
0
•4-^
U
ro
4—
>,
5-
QJ
>
O
O
CD
S-
^~
re
^-)
•r-
CL
re
o
.
^C '
°3
0
T3
C :
re
**
CO
o
c
ro
&.
3
in
c
•r—
in
"O
c
3
O
O-
CD
^*
CM
CM
II
C
O
•p
05
C
o
— 1
• •
CO
-p
o
re
5.3-21
-------
I/)
CO
I i § 1 1 1 I 1 1
o o o
000
o
*->
§§§§§§§
in -sf c\j co
3:
I— o)
0- -»-<•—
OS X i—
uu o o
LU !
DC
en
CO r—
CO 0)
co o;
co z
=) o
o
-
O *J •— U)
-1-1001
•r- 3 .^
O 'i- t— t
5.3-22
-------
to
K— 1
s:
UJ
z
o
- H- O i —
O UJ i —
CJ ce o
a: s""' u
-„, ^: ?J
n h- en
— i >- "~
3 to cu
tO d
_J 3
t/5 O 0
s~'l
,-X ^-N
OOOOOOOOOO
oooooooooo
OOinOOrHrHrOCMO
in «sf in ro **• 00 CM ^-m
ro un rH ro in UD r**.
rH ^ rH
W
/— s *^^.
OOCDOCDC3O OCD
OCDOOOCDO OCD
OOCOromcMO ^•^f
m"^'fx~rHCOIl*x-rH CMP^-
ro in CM ^h ro co
/""s
^s C -Q
in 0 J3 C
TO Zi > i— r— 0
_j .c ro ro ro 4->
O ^ i— (I) O5 O
^; CJ CO O D
&-Q '"^ N— ' *«— ' O O C
.COin O rH O
^ . 4-i > > r— rH O **<.
f^4J^4- C C fl3\OO
TO4^. I-H i— i u o . in
^j r** o -r- co ro w
•r- Qjrooroin -4->
s-t-s-io -a co -u •*-
ajo**-O'O) c -i- *c
ro t3 *& c a> ro-r-^jc W03UN
O U ' +-> 4J r- ED
Q. cu > u r— c ro+^ro**-
C ^ O QJ 4J *^~ "^* **"" Of '""
ro'r-3r—3IO (O.O-P3
0 O
o o
CM cn
t£) CO
o CM
r — CM
i— 1 rH
0 0
o o
00 O
CM O
U3 cn
ro
0 0
o o
O CO
m ?— i
in CM
r-H CO
u
>
c
&-S
i —
.4-) +-
tn ro
o
u m
QJ
o; D
c: s-
-•- ro
+-> jz
ro u
s_
Q) 'r-
o. ro
O +J
•p a
OJ ft
o o o
o o o
ro o o
r-- ro to
*j- in i£>
ro ro
o o o
o o o
CO O O
CM CM CM
in ro ic
^* rH rH
rH rH
0 0 O
O CD CD
CO O 0
IJ3 O"l "Ct"
•^ tJi 0"J
•Q
C
U
c
1 — 1
in
IS)
QJ
l/J 3
o ro
u • —
•j (_)
ro
ZJ +->
c c
c; a>
'ro o
_ -P >
o c
r- HH
O
O
O
cn
CM
CM
r-T
o
o
o
«cr
CD
CM
CM
4fr
o
0
o
ro
00
rH
i
ji
cu
c
cu
cu
>
c
£-
CU
Q.
o
0
in
ro
ro-
^.
01
>
o
u
a>
S-
t.
13
t+_
3
in
8
0
t-H
in
cu
E
i/>
i/l
£1
15 years at 10% interest plus 4% of investment for property taxe:
S-
o
l*-
s-
o
4-}
u
ro
"-S
S-
CU -D
> C
o ro
u
cu -
i- 01
u
i — C
ro ro
+-> i-
a. IA
ro c
U •!-
u
, "Air Pollution Control Technology and Costs in Seven Selected
3-73-010, page 629.
£ ^^
S- m
i^_ ^*
i
C 0.
CU ' ' '
W) O
3 Z
•f-J
•o -u
ro t-
o
t- Q.
O CU
ro
i. -
c in
•i- ro
t> cu
c. i.
XI
X
cu
•o
c
r— 1
D>
C
LU
E
cu
x:
o
cn
r^
•^
ID
CU
5.3-23
-------
The tail-gas scrubbing systems used to reduce emissions from the Claus
plant employ oxidation or reduction processes and result in residual emis-
sions of S02 (oxidation) or H2S (reduction). The reduction system general-
ly, however, leads to the overall lowest level of emissions.
Energy impact—Although not generally recognized, petroleum refineries
consume a significant amount of energy in processing crude oil into various
petroleum products such as petrochemical feedstocks, gasolines, fuel oils,
etc. The energy requirements of a typical moderate- or high-conversion
refinery, for example, usually represent about 10 percent of the crude oil
throughout. Thus, the energy consumption of a nominal 15,900-mVday
(100,000-bbl/day) refinery is equivalent to about 1590 mVday
(10,000-bbl/day) of fuel oil, or some 250,000 kWh/h.
The energy requirements of refinery sulfur plants are quite small in
comparison. A 102-Mg/day (100-long-ton/day) Claus sulfur plant, for
example, typically consumes less than 1000 kWh/h of energy, or less than 0.5
percent of the energy consumed within the petroleum refinery itself. Con-
sequently, the use of Claus sulfur plants to control emissions of sulfur
dioxide or hydrogen sulfide at petroleum refineries does not significantly
increase the energy requirements associated with petroleum refining.
The energy impact associated with each of the alternative emission
control systems is summarized in Table 5.3-4. Tail-gas treating units have
a slight energy penalty or a moderate energy benefit, depending on whether
an oxidation or reduction tail-gas scrubbing emission control system is
employed and on whether tail-gas reheat is required to increase plume
bouyancy. The moderate energy benefit associated with the reduction tail
gas scrubbing system arises because of reduced tail-gas incineration
requirements.
This energy impact will vary from refinery to refinery, depending on
whether an oxidation or a reduction tail-gas scrubbing system is employed.
As shown in Table 5.3-5, use of an oxidation tail-gas scrubbing system
without tail-gas reheat increases the overall energy consumption of a Claus
sulfur plant by about 17 percent. Use of a reduction tail-gas scrubbing
system without tail-gas reheat, however, reduces the overall energy con-
sumption by about 50 percent.
5.3-24
-------
to
to
to
O
u
o
»-H
to
to
«sc
CL.
o
Q£
LU
I
CO
un
CO
m
Ol Oi
C 4->
•r- in
.O >>
JS m
3
^ ^
U O
(A U
10 5
WO
1 O)
r- fc.
"m >»
H- ^
U
O) 0)
c *•>
•r- tf)
£3 W*
D
I— C
u c
in •»-
in re
TO "O
1 X
r— O
re >,
I— -O
£ Q
4-1 01
•— £
t.
3 4J
o re
f Ol
*J X?
•t— Ol
3: 1-
»J 01
5 "S
3 4J
o re
x: ai
•i- Oi
.a
m
re
u
01
in
re
CO
re
4->
u
re
t
o>
Ol
c
LU
•a oi
*""\ *••% ^«*
**"."*: So "!
r-t in fH r*
>*^ ^^ o ^w ^^
^.
LO r^ n co cf*
O fH CM O
•0^ ^
r*. *— m
fH 2 r-'
\^ o ^-^ *^^
tf> CO CO Cft
O CM O
01
. . CO ^^
oo in f-t in
o
CM r— CM O
CM <-*
S O Ci S
"" O
*** CO CO ID
CM ' CM O
CM *-* CO
CO
CO t— t f^*
^^ O v.^ ^-^
•e- 1-1 co o
CM CM
<-< 6 -c
N^ f in re E
\ ~«. ai re
o re 2 •* *n *j
•f- E C Ji in
4J C O H- Ol
tai •>- • . L. oi
f 4J X 4J 3 «-
4J re 4-1 — in 3
U U ••- I- U OL fc.
- O U 4-1 I CX
>, — S- C U >s £ 1
oi oi a. H- oi oi oi *
U 3 •—(-••- 0
g U. LO » Z -J
LU LU
O
rH
rH
O
m
o
en
o
CM
r«t
0
a>
.c
c 2
o
«J 4->
a.—
§T>
1>
V) t-
c u
&
ul
^^ Ol
(- in
ai
C 4J
ai 3
o
— .c
re 4J
is*
m
m
r**
o
S
»— t
o
CO
m
i
1
o
un
cr\
t
(
^ : '
Cn i;
CO 1
1
^s^ i
x:
4-1
•5 —
Ol "D
t- Ol
° U
^,
i- re
C 4->
Oi m
r— .C
re 4->
4^ *^-
re
ai
O
u
Ol
t-
o
•o
3
cr
a>
i.
O
II
J^
• o
-C 31 O
c
re
— a.
Q.
•D
01
3
CO
O
,— (
s^^
r—
$-
QJ
a i
1
rn
Ol
CM
o
E
, .
a/
i:
O
f-H
^^
UD
rH
O
.
E
re
ai
in
01
i-
Q
re
•o
c
o
1
O)
c
o
o
0
t— (
I/)
3
re
c
0
i »
5
in
c:
o
u
o>
fc-
in
01
in
re
en
c
o
+j
u
3
•o
Ol
i-
Ol
u
•o
0
a.
re
o>
re
- c
re
a.
3
1^.
3
in
i*-
o
01
•o
"3
0
u
4J
re
4-1
E
re
Ol
in
Ol
•o
3
*-~
Q.
O
in
in
0
o
o
•o
c
re
ro
a.
3:
00
ro
a.
o
r^-i t- Ol
a. c-
- i a.
"oi o- s
3 •— O
*J T3
O Ol
C U
3
CM
O
r— •— *r- in
_ Ol
C O t-
1-1 O O.
a> **- o>
re
•D
Ol
u
3
O
5.3-25
-------
Tail-gas treating units will reduce national energy consumption by some
54 million kWh/yr, or about 14,300 m3 (90,000 bbl) of fuel oil per year,
assuming that refinery sulfur plants accounting for half of the capacity
subject to compliance with New Source Performance Standards (NSPS) install
oxidation tail gas scrubbing systems without tail-gas reheat and that plants
accounting for the other half install reduction tail-gas scrubbing systems
without tail-gas reheat.
Ambient air quality impact—Tail-aas scrubbing systems reduce the
maximum ambient air concentrations of S02 to essentially zero for all
practical purposes, assuming an oxidation tail-gas scrubbing system is
employed. If a reduction tail-gas scrubbing system is employed, emissions
of S02 are eliminated. Use of a reduction system, however, leads to emis-
sions of H2S, carbonyl sulfide, and carbon disulfide and thus low ambient
air concentrations of these pollutants. Table 5.3-5 shows the calculated
sulfur emissions in S02 equivalent from a sulfur recovery unit and a reduc-
tion tail-gas treating unit.
TABLE 5.3-5. CALCULATED S02 EMISSIONS FROM CLAUS AND SCOT UNITS
Sulfur recovery unit capacity,
Mg/day (tons/day)
Claus unit emissions (no tail-gas
treating unit and 95% sulfur
recovery efficiency), g/s (Ib/h)
SCOT unit emissions (sulfur
recovery efficiency increased
to 99.5%), g/s (Ib/h)
9 (10)
12.5
(98.9)
1.2
(9.3)
45 (50)
61.9
(490.9)
6.1
(48.5)
Water pollution impact—Petroleum refinery Claus sulfur plants generate
a small wastewater stream. This stream results from condensation of water
vapor contained in the H2S gases as they flow from the am.ine scrubbing units
to the Claus sulfur plant. The volume of water involved is less than
0.25 liter (0.07 gal) per minute and normally contains 1500 to 2000 ppm H2S
and up to 1000 ppm ammonia. The refinery's wastewater treatment facilities
can easily handle this stream.
5.3-26
-------
The potential water pollution impact of the tail-gas treating units is
negligible. Table 5.3-6 summarizes the characteristics and flow rates of
the various wastewater streams discharged by these systems. As this table
shows, although the volume of wastewater discharged by some of these emis-
sion control processes is larger than that discharged by the sulfur recovery
unit, it is less than 50 liters (13 gal) per minute.
Generally, the wastewater streams generated by the various tail-gas
scrubbing processes consist of a sour water condensate and a purge stream
containing either organic or inorganic salts. The amount and composition of
these wastewater streams varies depending on the particular tail-gas
scrubbing process used. The sour water condensate is produced by cooling of
the gases prior to the scrubbing tower, and the purge stream is necessary in
most cases to prevent a buildup of impurities in the scrubbing solutions.
All the wastewater streams generated by the emission control systems can be
treated without difficulty in the refinery's wastewater treatment facility.
Because these waste streams are so small, they will have a minor impact on
the ability of petroleum refineries to meet water quality effluent regula-
tions.
Solid waste impact—There is essentially no potential solid waste
impact associated with tail-gas treating units.
The Claus process itself requires periodic replacement of the reaction
catalysts; the frequency of replacement depends upon the impurities present
in the acid gas feed. Usually the catalyst, made of bauxite or alumina, is
regenerated annually until a substantial loss of activity occurs, normally
in 2 to 5 years. Emission control systems will not affect the rate or
quantity of catalyst replacement in the Claus plant.
As for the emission control systems themselves, the oxidation tail-gas
scrubbing systems do not generate any solid waste. The reduction tail-gas
scrubbing systems, however, do require periodic replacement of the reduction
catalysts about every 2 years. These catalysts usually have significant
salvage value; because they are composed primarily of col bait-molybdenum,
they are normally returned to a vendor for reprocessing. Hence, even the
reduction tail-gas scrubbing systems generate essentially no solid waste.
Other environmental impacts—No environmental impacts other than those
discussed above are likely to arise from tail-gas treating units for
5.3-27
-------
to
D£
UJN
2TCO
t-< to
U. I—
UJ I—I
o: as
O CJ
O
Q_ UJ
E o:
o
uj o
I—
o
Q.
I
co
•
in
UJ
i
CO
i u> 1
1 {_> 1
iracteristil
J2
JC I
° 1
'
T3
CU
P
ro 1
•g
Ul
•1- 1
•0
CO
CL
CL E
oS '
O
o o
CM 0
1 O
0 rH
CO
in
"^- E
0) r-
tn
re
tn i.
01
CM E -U
>.= 5
z -o
u> en
CM
c "E
^i'i
re re
Ol O)
ro o
in ^
c c
E 'E
CM
CL
C
•r—
E
re
Ol
ro >,
in CM
tn s-
f£™
CO
3:
1
o
•£•
E
a
CM
T- t
C
5
tn
ro
n in s-
:r cu
Z E +J
3 re
01 -i- S
U T3
re o >c
S- Ul 00
•£•
i
re
Ol
CO
c
E
i^x"
CL O)
CL u
re
0 (-
in *-J
c
^i
»—
re
Ol
CM
c
'i
in o
c\j .
oS
V
tn en
i. s-
01 o>
o in
CM I-H
tn *~- f—
t- ro
Ol O O)
•u o
•f- O O
.— -o
CD *(-
o CM r-~
CM <> *->
in --^
s- in
O.
ro
in c
c o
O) •!-
-D 4J
C 3
O •—
U O
in
i-
re
C
O
u
t-
o>
o
u
s-
Ol
1
1 cu
1 Ul
1 ro
1 <°
1 CO
I_
3
O
CO
•o
i.
cu
CL
in
c
o
01
c
o
S-
O 3
CO <*-
Ul
CM ro
So
Ul
cu
>p.
.fj
re
cz
s
ro
o
Emission contr
f -a
•-•§
3 S-
O U
CO CO
Ul
tn
CD
U
O
i.
CL
Ol
C
JD
JQ
3 T3
i- i-
Oxidation/sc
Wellman-Lo
t- re
3 +J
o re
CO O
o
o
in
rH
Q-
LL.
»— i
2 -CJ
0 U
CO CO
tn
Ul
01
Ll
O
CL
01
c
.0
JD
3
Reduction/scr
Beavon
3
0
CO
I—
o
o
CO
5.3-28
-------
refinery sulfur plants. Furthermore, other than those resources initially
required to construct the emission control system (most of which could
probably be salvaged in one way or another), there does not appear to be any
irreversible or irretrievable commitment of resources associated with these
systems. There is even no overall increase in the energy requirements
associated .with refinery sulfur plants, because the emission control systems
result in a net reduction in energy consumption.
5.3-29
-------
REFERENCES FOR SECTION 5.3
1. U.S. Environmental Protection Agency, Office of Air and Waste Manage-
ment, Office of Air Quality Planning and Standards. Compilation of Air
Pollutant Emission Factors. 3d ed. Research Triangle Park, N.C
AP-42. August 1977. Figure 9.1-1.
2. PEDCo Environmental, Inc. National Petroleum Refinery Inventory -
Phase 1. EPA Contract No. 68-01-4147. 1979. p. 6.
3. U.S. Environmental Protection Agency, Office of Air Quality Planning
and Standards. National Air Quality, Monitoring and Emissions Trend
Report, 1977. EPA 450/2-78-052. Research Triangle Park, N.C.
December 1978. p. 5-8.
Petroleum Proc-
4. Sittig, M., and G.H. Unzelman. Sulfur in Gasoline.
essing. V[(8):75-95. August 1956.
5. Danielson, J.A. Air Pollution Engineering Manual. U.S. Public Health
Service. PHS-Pub-999-AP-i40. 1967. p. 539.
6. Monsanto Research Corporation. Refinery Catalytic Cracker Regenerator
SO Control Process Survey. Office of Research and Development.
Washington, D.C. EPA-650/2-74-082. 1974. pp. 22-24.
7. Ref. 1, p. 5.17-7.
8. Ref. 1, p. 5.17-4.
9. Letter from Parnell, D., Ford, Bacon, & Davis to Spuhler, F.J., Texas
Air Control Board. June 25, 1975. Claus Sulfur Plant Costs and
Overall Sulfur Recovery.
10. Pfeiffer, J.B. Sulfur Removal and Recovery From Industrial Processes.
American Chemical Society, Advances in Chemistry Series, No. 139
Washington, D.C. 1975. pp. 12, 75-76.
11. Dickerman, J.C., T.D. Raye, and J.D. Colley. The•Petroleum Refinery
Industry. Prepared for Dr. I.A. Jefcoat, U.S. Environmental Protection
Agency, Control Systems Laboratories. Research Triangle Park, N.C.
EPA Contract No. 5-02-5609B. 1975. pp. 41-42.
12. NPRA '74 Panel Views Processes. Hydrocarbon Processing.
54(3):130-131. March 1975.
13. Ref. 6, pp. 22-24, 32, 40-46, 70-73.
5.3-30
-------
14 Magee J.S., R.E. Ritter, and L. Rheaume. A Look at FCC Catalyst
Advances. Hydrocarbon Processing. 58:123-130. September 1979.
15 AD Little, Inc. Screening Study to Determine Need for SC>x and Hydro-
carbon NSPS for FCC Regenerators. Research Triangle Park, N.C. EPA-
650/2-74-082. August 1976. pp. 35-43.
16 Texas Air Control Board. Sulfur Dioxide Sampling and Continuous
Monitoring at Exxon. Baytown Refinery. January 11, 1979. p. 1-4.
17. Vasalos, I.A., et al. Oil and Gas Journal. 75(26): 142. June 27,
1977.
18 U S. Environmental Protection Agency. Summary Report on S02 Control
Systems for Industrial Combustion and Process Sources. Volume III.
Claus Processes. Research Triangle Park, N.C. EPA Contract No.
68-02-2603. Task No. 4. December 1977. pp. 3-85.
19 US Environmental Protection Agency. Standards Support and Environ-
mental Impact Statement. Volume 1: Proposed Standards of Performance
for Petroleum Refinery Sulfur Recovery Plants. Research Triangle Park,
N.C. EPA-450/2-76-016-3. September 1976. pp. 4.15, 4.29.
20. Ref. 19, pp. 4.23, 4.27.
21. Exxon Research and Engineering Company. Fluid Catalytic Cracking Unit
Flue Gas Scrubbing. Florham Park, N.J. March 1979, p. 7.
22 PEDCo Environmental, Inc. Analysis of S02 Emission Control Alterna-
tives for the Cabras Power Plant, Guam Power Authority. Prepared for
Region IX of the U.S. Environmental Protection Agency under Contract
No. 68-02-1321, Task No. 19. May 13, 1975. p. 5-7.
23 Industrial Gas Cleaning Institute. Electrostatic Precipitator Costs
for Large Coal-Fired Steam Generators. Prepared for the U.S. Environ-
mental Protection Agency under Contract No. 68-02-1473, Task No. 17.
February 1977. pp. 3-3, 3-5.
24 Industrial Gas Cleaning Institute. Particulate Emission Control Costs
for Intermediate Size Boilers. Prepared for the U.S.' Environmental
Protection Agency under Contract No. 68-02-1473, Task No. 18. February
1977. pp. 3-3, 3-8.
25. World-Wide HPI Construction .Boxscore. Hydrocarbon Processing. Section
2. 58(6):9. June 1979. p. 9.
26. Ref. 19, pp. 8-3 to 8-34.
27. Ref. 19, pp. 7.1 to 7.23.
28. Ref. 19, p. 8-6.
5.3-31
-------
29. Ref. 19, p. 8-8.
30. Ref. 19, p. 8-9.
31. Ref. 19, p. 7.20.
32. Ref. 19, p. 7.12.
5.3-32
-------
5.4 NATURAL GAS INDUSTRY
Natural gas often contains hydrocarbon condensates (such as natural
gasoline, butane, and propane) and water. These condensates are usually
removed at a field separator located near the well site (see Figure 5.4-1).1
Natural gas from some reservoir formations contains such acid gases as
gaseous sulfur compounds and carbon dioxide (C02). Approximately 95 percent
of U.S. natural gas production is free of sulfur compounds and is referred
to as sweet.2 Natural gas containing sulfur compounds is referred to as
sour. Sour natural gas contains hydrogen sulfide (H2S) in widely varying
concentrations, together with trace amounts of organic sulfur compounds such
as mercaptans (RSH), carbonyl sulfide (COS), and carbon disulfide (CS2).
Hydrogen sulfide rarely constitutes less than 95 percent of the total sulfur
content.3 To meet pipeline gas specifications of 6 mg H2S per standard
cubic meter (0.25 gr/100 std ft3) and heat content of. 37 MJ/std m3 (1000
Btu/std ft3), and to obtain fuel gas of low sulfur content for plant use,
the processors "sweeten" the sour natural gas; i.e., they remove the "acid
gases."
Sour natural gas is processed in 17 states ranging from Michigan and
Ohio to California and from North Dakota to Texas and Florida. Capacities
of the processing plants range from less than 27 x 106 std ms/yr (1 x
109 std ftVyr) to 4 x 109 std nrVyr (148 x 109 std ft3/yr) of sour gas
processed. Quantities of sulfur available in the sour gas range from less
than 1000 Mg/yr (1000 long tons/yr) to 155,000 Mg/yr (153,000 long
tons/yr).4 There is a general trend to larger plants as the production of
groups of wells is consolidated for processing in individual plants. The
size of a plant, however, is dictated by the amount of well production in an
area the plant can feasibly serve. Therefore, new plants are likely to
cover a considerable size range.
Emissions from gas treating plants may include unrecovered sulfur
compounds (H2S, COS, and CS2), S02 (from oxidation of sulfur compounds), and
CO.
5.4.1 Major Natural Gas Desulfurization Processes
Gas sweetening processes can be grouped into four major categories:
1) amine and amine-type processes; 2) carbonate and other chemical proc-
5.4-1
-------
to
ai
o
o
a.
re
01
3
•t-5
re
03
ns
•r-
-a
T3
ai
to
I
•vf
(L)
en
5.4-2
-------
esses; 3) physical absorption processes, and 4) solid bed sweetening proc-
esses. This section will only discuss'the amine processes since they are
used for over 95 percent of all gas sweetening in the United States.5
Sulfur .in acid gases not present in sufficient quantities to be
recovered may be vented, flared, or incinerated. When there are sufficient
quantities of sulfur, acid gas from the sweetening processes goes to a
sulfur recovery system, which in most cases consists of a Claus-type plant.
The system may include equipment for treating the Claus tail gas.
In the amine process (also known as the Girbotol process and invented
by Girdler) various amine solutions are used as absorbents for H2S. Amine
processes were developed to remove high concentrations of H2S and C02 in
large volumes of gas. Pressures may be as low as 240 kPa (35 psig). The
alkanolamines are the most generally accepted and widely used of the many
available solvents for removal of H2S and C02. The three alkanolamines
generally used in gas sweetening are monoethanolamine (MEA), diethanolamine
(DEA), and triethanolamine (TEA). Of these three, MEA is usually preferred.
The basic amine process is summarized in the following reaction:
2 RNH2 + H2S -» (RNH3)2 S
where: R = mono-, di-, or triethanol
N = nitrogen
H = hydrogen
S = sulfur
The basic amine process system is illustrated in Figure 5.4-2.
The hydrocarbon gas (sour gas) enters the bottom of the absorber. The
lean amine solution (RNH2) contacts the gas countercurrently in a tray or
packed tower and absorbs the H2S and some of the other acid impurities
contained in the gas. The lean amine solution at normal temperatures of 21°
to 49°C (70° to 120°F) forms a compound with H2S. The desulfurized gas
leaves the top of the absorber, while the rich amine solution is sent to the
regenerator column. In the regenerator, the volatile H2S and C02 are
5.4-3
-------
ACID GAS
PURIFIED
GAS
LEAN AMINE
SOLUTION
RICH AMINE SOLUTION
STEAM
\JLREBOILER
HEAT EXCHANGER
Figure 5.4-2. Flow diagram of the amine process for gas sweetening.
5.4-4
-------
separated from the relatively nonvolatile amine by steam stripping. The
regenerated (lean) amine solution is cooled and sent to the amine storage
tank for eventual return to the absorber. The acid gas stream from the
regenerator is cooled and then sent to the sulfur plant.
Small quantities of H2S are flared. When present in recoverable
quantities, sulfur is usually removed from the acid gas by means of a Claus
type sulfur recovery system whereby exothermic oxidation reactions convert
H2S to elemental sulfur. The basic Claus reactions are
H2S + 1/2 02 -» H20 + S
H2S + 3/2 02 •* S02 + H20
2 H2S + S02 -» 3S + 2 H20
Four major variations of the Claus system are as follows:
where acid qas contains 50 to
1)
2)
3)
4)
Straight-through Claus—used where acid gas
100 percent H2S.
Split-flow Claus—used where acid gas contains 15 to 50 percent
H2S.
Claus—used where acid gas contains 2 to
Direct-oxidation
15 percent H2S.
Sulfur-recycle Claus—also used where acid gas contains 2 to
15 percent H2S.
The straight-through Claus system is the one generally found in
petroleum refineries. The acid gas concentration is high enough for combus-
tion to occur when the acid gas is mixed with the quantity of air optimum
for oxidizing the H2S to elemental sulfur. A complete discussion of this
Claus system is contained in Section 5.3.1.7.
The split-flow Claus process is the primary system used in sour gas
processing plants because of the lower H2S concentration. One-third of the
acid gas flow and the process air are introduced to the combustion chamber
where they react. The remaining acid gas stream is injected into the com-
bustion chamber bypassing the burners and mixed with the combustion gases
for the Claus reaction to occur. Typically, the process gases leave the
combustion chamber and enter first a waste heat boiler and then a condenser
5.4-5
-------
followed by a reheater before the catalytic reactor. After passing through
the reactor where H2S and S02 react to form sulfur and water, the gases
again flow to a sulfur condenser. The gas must be reheated and sent through
additional catalytic stages to increase the sulfur yield and overall plant
efficiency.
A variation of the split-flow Claus-type process with process gas and
combustion air preheat is used when the H2S concentration is not high enough
to support combustion. The acid gas and the process air are preheated by
heat exchange with process gas from the combustion chamber or in direct-
fired heaters. Adequate acid gas and air preheat is essential to maintain-
ing a stable gas flame in the reactor furnace. Again, approximately one-
third of the acid gas flow and the process air are introduced to the combus-
tion chamber, and the remaining acid gas stream is injected into the chamber
bypassing the burners. Because the heat of reaction in the formation of S02
is about four-fifths of the total heat of reaction in the conversion of H2S
to sulfur, a waste heat boiler is used to produce process steam. A
reactor-condenser train with process gas reheat prior to each reactor
follows the waste heat boiler.
There is no Claus combustion chamber in the direct-oxidation or cata-
lytic process. The acid gas is heated, mixed with air and S02, and passed
directly to a catalytic reactor. The S02 needed for catalytic conversion is
generated in a sulfur combustion chamber with the addition of the necessary
combustion air. Two or three reactors are commonly used in this process.
In the sulfur-recycle process, product sulfur is recycled to the com-
bustion chamber. The quantity of sulfur burned is limited to maintain the
H2S to S02 ratio of 2:1. The S02 formed and the H2S in the acid gas feed
undergo combustion according to the Claus reaction. A reactor-condenser
train similar to those previously described follows the combustion chamber.
The direct-oxidation and sulfur recycle Claus processes are not
generally used. It is possible to use a split-flow Claus unit to process
acid gas streams with very low H2S concentrations (2 to 15 percent) if
supplemental fuel gas is added to insure combustion of the H2S.6,7 It may
not be possible, however, to produce a salable product in this case because
5.4-6
-------
of color and/or carbon content added by the supplemental fuel combustion.
Acid gases obtained from natural-gas sweetening processes vary greatly
in H2S content and may contain impurities such as C02, hydrocarbons, and
water vapor. (Ammonia is normally only found in refinery acid gas.) These
feed gas impurities create costly and troublesome problems affecting design,
maintenance, and operation of a Claus sulfur plant. The major problem is
reduced sulfur conversion because of the effect of dilution by inerts.
Excessive C02 in the feed increases sulfur emissions to the atmosphere
because of the formation of COS and CS2, which exits with the tail gas.8
The C02 acts as the main inert, instead of requiring large quantities of
additional combustion air. Excessive hydrocarbons in the feed also increase
sulfur emissions because of the formation of COS and CS2. Additional com-
bustion air is required for oxidation of the hydrocarbons forming C02 and
water vapor. There is also a corresponding increase in waste heat boiler
duty due primarily to the additional heat release from combustion of the
hydrocarbons. Water vapor acts as a true inert, as well as being a product
of the Claus reaction. Therefore, both the equilibrium of the Claus reac-
tion and the effective partial pressures of the reactants are affected.8
Factors also influencing sulfur plant efficiency are changes in the ratio of
H2S to C02 and the acid gas rate, which may occur because of changes in the
production of the gas field.8
If the acid gas stream contains excessive amounts of these impurities,
several alternatives should be studied to enrich the stream with respect to
H2S. Several absorption processes that are selective for H2S in the
presence of C02 are available. Charcoal adsorption units can be considered
for removing hydrocarbons from acid gases prior to sending the gas to the
Claus unit. Water vapor can be condensed and separated from the acid gas
stream.8
The water formed in the Claus reaction is emitted as vapor and is not
an emission problem. There is some sour water generated by overflow from
the absorber that may contain both sulfur compounds and spent amine. The
volume is small, but some form of treatment prior to disposal may be appro-
priate. This stream can be corrosive to metal insufficiently protected.
The tail gas from the sulfur recovery unit is a major sulfur source in
a natural-gas sweetening plant and usually requires S02 control. Air
5.4-7
-------
quality control restrictions normally require the use of a Claus-tail gas
treating unit for further reduction of S02 emissions.9
5.4.2 Control Techniques
Operational tail-gas control systems include the IFP-1500 process, the
SCOT system, the Sulfreen process, the Beavon process, the Stretford
process, and the Wellman-Lord process. These tail-gas control systems have
been described in detail in Section 5.3.2.1 of this document.
5.4.2.1 Control Cost-
Sour gas processing operations include a natural gas treater (usually
an amine absorption process) and a sulfur recovery system. Table 5.4-1
shows the capital and operating costs for two sizes of amine units. Cost
estimates for amine units will vary with location, gas pressure, H2S con-
tent, C02 content, and nitrogen content.
The hydrogen sulfide content in the natural and acid gas stream is the
predominant cost parameter of sulfur recovery plants. The volume of C02 in
the source stream is also significant in costing sulfur recovery plants. An
acid gas stream with 50 percent C02 will require equipment that is twice as
large as a 100 percent H2S acid gas stream to maintain a constant flow rate
for the same sulfur capacity. Efficiency of the catalyst decreases with
dilution, and larger beds are required.
Other investment costs developed for the natural-gas processing
industry show the effect of low H2S concentrations on sulfur recovery plant
capital costs. As can be seen from Figure 5.4-3, the cost of a 100-Mg/day
(100-long-ton/day) two-stage sulfur recovery plant more than doubles when
processing an acid gas containing 15 percent H2S versus an acid gas contain-
ing 90 percent HaS.11
Legislation requiring the reduction of S02 emissions from sulfur
recovery units may necessitate the addition of tail-gas treating units. The
cost of recovering this incremental sulfur is high and is approximately
90 percent of the cost of a new Claus unit.10 Capital and operating costs
for tail-gas treating units are discussed in Section 5.3.2.2. The costs are
based on a typical refinery acid gas stream in excess of 70 percent H2S.
5.4-8
-------
TABLE 5.4-1. TYPICAL AMINE UNITS COSTS
(Mid-1979 dollars)
Investment:
(1) Plant cost
(2) Working capital, 15% of (1)
(3) Total capital cost
Operating costs:
(4) Capital recovery cost,
18% of (3)
(5) Taxes and insurance,
3% of (1)
(6) Total fixed costs, (4) + (5)
(7) Operating labor
(8) Maintenance, 4% of (1)
(9) Supplies
(10) Utilities
(11) Chemicals
(12) Total direct costs, (7) + (8) +
(9) + (10) + (11)
(13) Total operating cost, (6) + (12
Plant capacity
0.57 M nrVday
20 x 10R ftVday)
$2,160,000
324,000
2,484,000
$ 447,000
65,000
512,000
108,000
86,000
15,000
646,000
124,000
979,000
1,491 ,000
2.27 M mVday
(80 x 10G ftVday)
$8,650,000
1 ,300,000
9,950,000
$1 ,790,000
260,000
2,050,000
108,000
346,000
60,000
2,856,000
497,000
3,597,000
5,647,000
Reference 10 used to estimate total capital cost.
on engineering judgement.
Other costs are based
5.4-9
-------
o
o
1*
1 1 1 IT 1 1 1
-
-
mn*m
~~ to to to
~ -=C — i
— 0 CJ 0
«c «a: «=c
*•* i— i (— i
to to to
CM CM CM
*« &«! 3S«
10 O 0
•— to en
O O EJ
1 1 1 1 1 1 1 1
O O OOOOOO O
o o ocr>oor~ 10 LO 1 — 1
0[t] III 1
O OQ
O GD
O GD
O GE
~_
< oa
<] oa ~
-
-
1 1 1
O O OCDCOI^tD Ln «3" ;r; CM i
<"> CM !— OOCD O O O O CD o
0
0
p
,0
(tons/day)
o >,
• — ea
•o
en
O "
LO | —
o
-
CD o;
• 1 1 1
o
CJ
LU
OH.
01
• u_
0 _|
oo
CD
S-
o
•»->
(O
i.
OJ
CD
ra
-(->
to
I
o
to
Z3
03
o
to
o
o
ra
-)->
•r—
CL
CO
O)
6Z6L
OL x $
-------
The effects on the cost of tail-gas treating units for acid gases with
reduced hydrogen sulfide concentrations are similar to those discussed above
for sulfur plants.
5.4.2.2 Energy and Environmental Impact--
The energy requirements and environmental impact for tail-gas treating
units are discussed in Section 5.3.2.3. The costs are based on a typical
refinery acid gas stream. Energy requirements will be increased for tail-
gas treating units when the H2S concentration of the acid gas is low because
of larger processing equipment needed to handle the increased gas flow rate.
5.4-11
-------
REFERENCES FOR SECTION 5.4
1. U.S. Environmental Protection Agency. Compilation of Air Pollutant
Emission Factors. 3d ed. Research Triangle Park, N.C. AP-42. August
1977. Figure 9.2-1.
2. U.S. Environmental Protection Agency. Standards Support and Environ-
mental Impact Statement, An Investigation of the Best Systems of Emis-
sion Reduction for Sulfur Compounds From Crude Oil and Natural Gas
Field Processing Plants. (Draft). April 1977. p. 8-8.
3. U.S Environmental Protection Agency. Atmospheric Emissions Survey of
the Sour Gas Processing Industry. National Air Data Branch. Research
Triangle Park, N.C. EPA-450/375-076. October 1975. p. 36.
4. U.S. Environmental Protection Agency. Sulfur Compound Emissions of the
MPpr°D?Dm Product1on Industry. Office of Research and Development,
CDA Icn/A Control Systems Laboratory. Research Triangle Park, N.C.'
EPA-650/2-75-030. December 1974. pp. 5-3 to 5-17.
5.
6.
7.
8.
Ref. 1, p. 9.2-1.
Grancher P. Advances in Claus Technology, Part 1: Studies in Reac-
tion Mechanics. Hydrocarbon Processing. 57:155-160. July 1978.
Grancher P Advances in Claus Technology, Part 2: Improvements in
*7 oK Icl e S and °Peratl'n9 Methods. Hydrocarbon Processing.
57:257-262. September 1978.
Goar, E.G. Impure Feeds Cause Claus Plant Problems.
Processing. 53:129-132. July 1974.
Hydrocarbon
9. U.S. Environmental Protection Agency. Compilation of BACT/LAER Deter-
minations. Research Triangle Park, N.C. EPA-450/2-79-003 1979
Section 5.5. '
10. World-Wide HPI Construction Boxscore
2. 58(10):1-62. October 1979.
11. Ref. 2, p. 8-40.
Hydrocarbon Processing. Section
5.4-12
-------
5.5 SULFURIC ACID PLANTS
Sulfuric acid, the most important mineral acid, is the most widely used
industrial chemical. The chief uses of sulfuric acid are in production of
fertilizer, manufacture of chemicals, oil refining, pigment production, iron
and steel processing, synthetic fiber production, and.metallurgical opera-
tions. Sulfuric acid production in 1975 was 29.3 Tg (32.3 million tons)
from a total of about 150 plants.1 Florida has the greatest number of acid
plants (20), followed by New Jersey (9); Virginia (8); and Louisiana, North
Carolina, and California (6 each). The basic processes used for the produc-
tion of sulfuric acid are the contact process and the chamber process. The
latter accounts for only 0.3 percent of the total production, however, and
no new chamber process plants are being built. Further, the older chamber
plants are being phased out. The entire discussion in this section is
focused on the contact process.
Sulfuric acid is produced by burning sulfur or sulfur-bearing materials
to form S02. Sources of S02 include 1) elemental sulfur, 2) spent acid, 3)
smelter off-gas, 4) pyrites, and 5) waste gas from fossil-fuel-fired boil-
ers. Table 5.5-1 lists the capacity and the percentage of total capacity of
the major sulfuric acid manufacturing sources in the United States for the
years 1976 to 1980. The average operating rate, which determines the actual
production, reached a peak of almost 90 percent of capacity during 1973-
1974. In 1975, the average operating rate dropped to only 67 percent of
capacity.
5.5.1 Process Descriptions and Emission Sources
Contact sulfuric acid plants are classified as hot gas (sulfur burning)
or cold gas (metallurgical and spent acid) systems. Plants operating on
elemental sulfur receive hot S02 gas directly from the sulfur burner and
waste heat recovery system. When S02 gas from a metallurgical operation or
other byproduct source (such as spent acid or iron pyrites) is used, it is
received cold from the wet scrubber-cooler and purification systems.
A basic variation of the contact process is the double absorption
technique, also known as double catalysis. Because use of this design is
largely based on the need to meet air pollution control regulations, it is
discussed under Section 5.5.2, Control Techniques. ---•
5.5-1
-------
>- O
h- 10
t-4 CM
=
O
O
Q
t-H
O
£ g
— J O
=3 4J
to
Ol O
£**
01 *
2 £ 2:
4J OJ
C 4J
HI O
CD ,
CQ O>
-a: j-
(— i—i
m
c -u
cu
u
cu
ev. o
01 o
u J->
t-
0} <<-
O- O
83 s
SS
o
0
OOJD ^ CM
•— cs enio
00 CO CO
55 '*• «*>
O
0
..
O CD
o o'
ro CM
C7> O
01
c
01
u
o
n t.
OJ
CD QJ
Q- i
•/> CO
5.5-2
-------
5.5.1.1 Sulfur Burning Plants2—
In sulfur burning contact plants, shown schematically in Figure 5.5-1,
sulfur is melted and filtered to remove traces of ash. The molten sulfur is
atomized and burned with dry combustion air in the following reaction:
'(1)
°2
(g)
s°2
, , AH = -0.297 kJ/kg-mol
(9.) at 25°C (77°F)
The excess heat of combustion of sulfur is utilized in a waste heat boiler
to generate steam for melting the sulfur and in other process areas. Nor-
mally the amount of steam produced, by weight, is greater than the amount of
100 percent sulfuric acid produced.2 The hot gas (7 to 12 percent S02, 9 to
13 percent 02) is filtered and passed through a catalytic converter (plati-
num mass units or units containing beds of palletized vanadium pentoxide) to
oxidize S02 to S03 by the following reaction:
S02
V2 02
S03
AH = -0.098 kJ/kg-mol
Kg) T "' U2(g) ^ OU3(g)
The exothermic, reversible oxidation of S02 involves an inherent conflict
between the high equilibrium conversions at lower temperatures and the
favorable reaction rates at higher temperatures. Plant operators attempt to
optimize the process by first passing the combustion gas over a part of the
catalyst at about 420°C (788°F) where the reaction rate is high until about
70 to 75 percent of the S02 is converted, with a' consequent rise in tempera-
ture to around 600°C (ni2°F), where equilibrium is approached. The gas is
then cooled in a heat exchanger to about 430°C (806°F) and passes over two
or three more catalyst stages with intermediate cooling. The conversion of
S02 to S03 is about 97 to 98 percent.
The gas leaving the converter is cooled in an economizer, with addi-
tional cooling sometimes obtained by air-cooled heat exchangers. The S03 is
absorbed in a stream of strong (98 to 99.5 percent) acid. The S03 reacts
with the water in the acid to form additional sulfuric. acid. Dilute sul-
furic acid or water is in turn added to the recirculating acid to maintain
the desired concentration. If oleum (fuming sulfuric acid) is produced, the
economizer exit gases are passed through an oleum tower, which is fed with
the strong acid from the absorption tower.
5.5-3
-------
in
in
0)
o
O
CL
rt3 S-
•*-> =3
C <4-
O i—
O 3
co
n
E CD
-»-> E
C. S-
i- 3
O -Q
CO
-Q >^
(t3 XI
CD ID
CD O
E (O
CO O
•r—
ro S-
a r—
£ to
(O
t- CO
co CD
res o
•r- Z3
-o -a
4-5
ro
. 4J
I
LO
•
Lf)
OI
=5
CD
5.5-4
-------
5.5.1.2 Cold Gas Plants2--
Cold gas sulfuric acid plants use raw materials other than elemental
sulfur. The typical raw materials are sulfide ores, spent or sludge acid,
or waste gas from metallurgical operations. The cold gas processes require
extensive gas pretreatment involving dust removal, cooling, and scrubbing
for further removal of particulate matter and heavy metals, mist, and mois-
ture. When gas streams contain less than 10 percent S02, the size of the
equipment required to produce a given quantity of acid is relatively greater
because higher gas volumes are required. The capital and operating costs
for a cold gas plant are therefore much higher than those for a correspond-
ing sulfur-burning acid plant.
The process configuration of a.cold gas plant depends upon the source
of the S02. Figure 5.5-2 shows the overall flow diagram for a cold gas
sulfuric acid process applicable to ore-roasting plants and spent acid
regeneration plants.2
5.5.1.3 Sulfur Dioxide Emissions-- ,
Normal operations—Figure 5.5-3 shows the relationship between volu-
metric and mass emissions of S02 and conversion efficiency at various con-
centrations of S02 at the converter inlet.4 Although it is necessary to
provide adequate residence time for the reacting gases in the catalyst mass,
the conversion efficiency of the contact process is more directly affected
by the effectiveness of .interstage cooling. The conversion is also inverse-
ly related to the S02 concentration of incoming gases. The converter unit
consists of three, four, or five fixed beds of catalyst with interstage
cooling to maintain the optimum gas reaction temperature-conversion profile.
Plants built before 1960 generally had only three conversion stages and
operated with conversion efficiencies of about 95 to 96 percent.5 Plants
built since 1960 have four or more converter stages and overall conversion
efficiencies between 96 and 98 percent.5 Typical S02 emissions from various
types of single absorption plants without S02 tail gas recovery systems are
shown in Table 5.5-2.6
Acid mist is also emitted from sulfuric acid plants. The quantity of
acid mist formed depends on the strength of acid produced, the type of
5.5-5
-------
fe &
OQu Z-—
— «LIJQ:
J O
) Q£
§
CD
f— CNJ in r^ e\j
CD
.
—11— uj or uj >
3
-------
0.5
(1)
SULFUR CONVERSION EFFICIENCY. % of feedstock sulfur
99.7 99 98 97 96 95
1 1.5 2 2.5 3 4 5
(2) (3) (4) (5)(6) (8) (10)
10 15 20 25 30 40 50
(20) (30) (40)(50)(60)(80)(100)
S02 EMISSIONS, kg/Mg (Ib per ton) of 100X H2S04 produced
Figure 5.5-3. Volumetric and mass S02 emissions from
contact sulfuric acid plants.4
5.5-7
-------
TABLE 5.5-2. UNCONTROLLED SULFUR DIOXIDE EMISSIONS FROM SINGLE
ABSORPTION SULFURIC ACID PLANTS6
S02 in converter feed,
% by volume
S02 emissions from three
stage converter,
kg/Mg (Ib/ton) 100% acid
ppm by volume
S02 emissions from four-
stage converter
kg/Mg (Ib/ton) 100% acid
ppm by volume
Feedstock
Sulfur
7.5-8.8
28-35 (56-70)
3000-5000
13-28 (26-56)
1500-4000
' ' Acid sludge
6-8
NA
NA
15-56 (30-112)
1500-4000
NA - Not available.
5.5-8
-------
sulfur feedstock, and the absorption efficiency. The acid mist is control-
led by high-efficiency vertical tube mist eliminators. Details on acid mist
control technology are published.7
Sulfur dioxide emissions may exceed the normal emission rate during
startup or abnormal operations. The frequency and duration of the abnormal
emissions depend on the plant design, type of control systems, and the
nature of the startup or operational problem.
Startup—The amount of S02 emitted during startup depends on the time
needed to bring all of the converter stages to the proper operating tempera-
tures. Time required to achieve a stable operation of the plant depends on
the length of the shutdown and the condition of the catalyst bed. When a
plant has been charged with new catalyst, a startup will require 1 to 2 days
of slowly increasing production rates until full production is reached.
Plants with catalyst exposed to moisture can be in full production in 3 to 4
days. During startup, the emissions may be five times the normal rate for
the first few hours if the final stage is not at the proper temperature
before S02 is introduced into the converter. If the preceding stages are
sufficiently heated to obtain nearly full conversion at reduced rates,
reaction heat then continues the heating process until ignition is obtained
in the final bed.
Upon completion of the preheating sequence, sulfur or sulfur-containing
feedstock is burned at a low rate with excess air to produce a weak S02
stream, which is fed to the converter. Adjustments are made to stabilize
all operations, bringing all temperatures to normal conditions and gradually
increasing the feed rate and inlet S02 concentration as the temperature of
the first bed decreases and that of the last bed increases to ignition.
These adjustments must be carefully coordinated to prevent loss of stability
and resultant excessive S02 emissions.
Abnormal operation—During routine operation several .types of abnormal
conditions can cause excessive emissions. Elemental-sulfur-burning plants
have the fewest problems because they operate with a relatively constant
concentration of S02 to the converter. Plants using spent acid or some of
the various metallurgical off-gases are more prone to operating problems,
the most common of which are listed below:
5.5-9
-------
1. Sudden change in concentration of S02 to the converter: occurs in
plants using spent acid or metallurgical off-gases as feed.
2. Oxygen starvation; occurs in plants using roaster gas as feed.
3. Equipment failure or power failure.
Emissions during these conditions usually range from 50 to 100 percent
higher than normal levels. Usually, the operations are stabilized within a
few hours.
5.5.2 Control Techniques
Technology for control of S02 emissions from sulfuric acid plants is
well established. The double absorption process is operating successfully
at over 200 plants throughout the world, including 40 plants in the United
States. In addition, several desulfurization processes are applicable to
tail gases from a sulfuric acid plant. These processes, which could be
applied to all classes of contact acid plants when operated with a high-
efficiency Brink-type mist eliminator in the final absorbing tower, provide
simultaneous control of S02, S03, and acid mist. The two processes with
maximum potential for controlling S02 emissions from acid plants are sodium
sulfite (Wellman-Lord) and ammonia scrubbing. These processes and. others
that may also be applicable to sulfuric acid plants are discussed in Section
4.2.3. The discussion here is limited to a brief review of the processes
currently in operation on acid plants in the United States.
The size of an emission control system for tail gas is decided by the
flow rate of the exhaust gas to be handled. The exhaust gas flow rate at an
acid plant is a near linear function of the daily production rate. The
relationship between exhaust gas flow rate and daily production rate is
shown by the following equation:8
Exhaust gas flow, NrnVs = 0.0357 x daily production rate
in megagrams/day - 1.31
(Exhaust gas flow, 1000 scfm = 0.074 x daily production
in tons/day - 3.0)
This relationship applies to sulfur-burning contact acid plants with S02
concentrations at the converter inlet of 10 percent. For cold gas plants,
which have much lower S02 concentrations, the exhaust gas flow will be cor-
respondingly higher for a given production rate.
5.5-10
-------
5.5.2.1 Description—
Double absorption—The sulfur combustion portions of the single and
double absorption plants are similar. Combustion air is dried in a tower
with 93 to 98 percent sulfuric acid before being introduced in the sulfur
furnace. Furnaces normally operate with gas strengths of 9 to 12 percent
S02.
Using a double absorption process, a plant can convert 99.7 to 99.8
percent of the S02 produced to S03. The primary difference between the
single and double absorption processes is the addition of a primary S03
absorber for gas leaving the third catalyst bed.9 Some processes use this
absorber after the second bed. In the primary absorption tower the concen-
tration of S03 in the gas is reduced to approximately 100 ppm by contact
with 98.5 percent sulfuric acid. The gas stream is cooled before the inter-
stage absorber and is reheated before it goes to the next catalyst bed. The
type and arrangement of heat exchangers varies, but this cooling and reheat-
ing operation is included in all designs.
Approximately 97 percent of the S02 remaining in the gas stream is
coverted to S03 in the fourth catalyst bed. The lower partial pressure of
S03 drives the reaction to a higher overall conversion rate than is possible
in a single absorption plant
The gases leaving the fourth catalyst bed are cooled in a second heat
exchanger. The cooled gas passes to a secondary absorption tower containing
98.5 percent sulfuric acid. The gases leaving the secondary absorption
tower will contain about 100 to 300 ppm S02.
With the exception of the primary absorber and arrangement of the heat
exchangers, all major designs of dual absorption sulfuric plants use similar
equipment configurations. Design variations are found in converters, heat
exchangers, and absorbers.
The double absorption process has proved to be the S02 control system
of choice for the sulfuric acid industry since the promulgation of NSPS. Of
the 32 new units built since the promulgation of NSPS, 28 use the double
absorption process for S02 control.10
5.5-11
-------
Figure 5.5-4 is a flow diagram of the double absorption process;"
Table 5.5-3 summarizes the status of double absorption processes in the
United States as of December 1977.12'13
Ammonia scrubbinq-This process is described in detail in Section
4.2.3.5; only the salient features of its application to sulfuric acid
plants are discussed here. The tail gas from an acid plant is at about 85°C
(185°F) and contains only traces of moisture. The adiabatic saturation
temperature is about 32°C (90°F). At this temperature, the vapor pressure
of ammonia and S02 are low, and thus the formation of the "blue haze"
characteristically associated with ammonia scrubbing is reduced.
The tail gas, containing 2000 to 3000 ppm of S02 and 1.5 g/Nm3
(0.6 gr/scf) of S03 and acid mist, is contacted with ammonium bisulfite/
sulfite solution in a two-stage absorber. The lower stage is operated at
high bisulfite/sulfite ratio and lower pH (5.5) to minimize ammonia consump-
tion; the upper stage is operated at low bisulfite/sulfite ratio and higher
PH (6.5) to maximize S02 removal. Within bounds, each stage operates in-
dependently. The ammonium sulfite/bisulfite bleed solution from the
absorber is fed to the acidulation tank, where H2S04 converts the
bisulfite/sulfate to ammonium sulfate, evolving equimolar amounts of S02.
The acidulated liquor is further stripped of dissolved S02 in a packed
stripper by air. The stripper off-gas is recycled to the acid plant drying
tower along with the acidulation tank off-gas. Ammonium sulfate liquor in
concentrations of more than 40 percent is withdrawn from the stripper
bottoms. The blue haze, which consists of fine particles of ammonium
bisulfite/sulfite, is removed from the stack gases by high-efficiency
candle-type mist eliminators.
Figure 5.5-5 is a flow diagram of an ammonia scrubbing process-14 Table
5.5-4 summarizes the status of ammonia scrubbing systems at acid plants in
the United States as of December 1977." The table shows that f-ve 1nstal_
lations can meet NSPS, four cannot, and no emission data 'were available on
the other four.
Sodium sulfite TWellman-Lord^ ^..hMn-c^.K w1th sodfum
is described in Section 4.2.3.6. The basic difference in the desulfuriza-
5.5-12
-------
HW1S ;
HV31S
03iV3HH3dnS
s
s
N,
V
v
X
^
x^
^
Ul UjT
UJ Ct
h- UJ
ce __i
UJ O
L_L_*i
10
(1)
O
O
Q.
u
O S-
O 3
E 3
O IO
«r-
-1-5 CD
O- C
S- •!-
O C
co S-
-Q 3
fO J3
_Q
3 T3
O T-
•a o
(O
(O
o
S- =S
C7> to
to
•I- I/)
•a CD
o
I-i
i— O
M- i-
o.
CO
to 4->
a> res
o -c
o -u
o.
i
tn
-------
to
CL.
O
£ E
i—i O -O O
o
O
CO
LLJ
E
s
E1 ^
ID O
•u S
CO
>-
CO
14-02:
a:
o
CO
ca
•r- E 10 -I-
CO
I
i— O CM O
X CM X r—
'g
j^~ <* o
x co x .—
CO
CO
co
co
to
o o • E E f-
*-> • U J= >, . x: -t-
I-ro row toi— Ur—
r- C « f- 0) ••- TO
LU
_l
CO
I- IO r— (O
CT1
-------
•o
OJ
o
o
CO
I
in
LO
CQ
XI
re
re •
"o in
c
in v
in^'i
Ol
O CM
S- 0
+-> in
c
0
re
•o
c
re
O_
in
£
o.
*;
E
E
C7> C U
21 0-0
^£ _Q fO
~
1
o.
s_
01
Ol
in
o.
3 5.
+» >>
S_ ~^
re o
4-> E
•i- S_
O) OJ
ce c
1 -3£
T3 O
O> O
0) 4->
u_ in
*. re • — .
4-> ~^ re
•t- O1T3
us: ^
re in
CL X C
re 0
z^
c
o
4->
re
u
o
>i
1C
c
«r
z
z
in
re
Ol
3 re o
_J O D-
CM
s-
0)
1 *—
S~ OJ in
•r- E re
«X in 01
O
O- — •
10 O
1C
X 1C
,-"~^
»
re
"O .
C
0 •
O 4->
re c
c o
«r z:
o
re
"*2
s
re
c
Ol
t-
• •• 3
O -r-
•r- TO
in n-
0 3
U O.
s_ s_
o o
x>
s-
S. Ol
QJ XI
4-> 1-
tn o
o in
0X1
co ro
r- CM
CM ^I
i — CM
O
C
CM
01
in
3
O
1
C
^
1 —
ro
z
S-
3
O
ir> o
p— CM
r--
X i —
•— ~"*
re o
•o x:
c re
o -o
in
0)
•jj"
H
S- in
11
CO >—
s!- o
o
l_) 4->
U
~ 3
l/i ^3
OJ
•O tt> i-
C •— Q)
re 4J X
U 3 O
O 4->
*+-
o c in
•i- in
en ro
ceo.
•r- O l-
Ol"- OJ
01 in 4-1
re o c
I/O S- "-
1C CM
ro r~-
o
1C
in
re
-~- Ol
•f- i-
01 >> OJ
>- > S
3 re o
in
~-
CM
a:
i.
3
4-
3
O O
Ol O
o
X t—
CM
I •
i~ re
re • —
CQ U.
1
in
3
"&
C
*_4
in
. QJ
Lu t-
• t-
0 4->
CM «- CM «»
CO 1C CO VC
i— ' ro r— m
§O
c
ro ro
0 0
+- *J
C C
(O (O
tn tn
1 1
in in
CO CM
s_ s-
3 3
l^- *^-
"5 3
o"o
X t—
CM
^
>,
•*—
S "
4J . C .
L- (D 1C *O
«3 f— *— r—
CO U- Q_ U-
m o co co «d- cc
o re o 3 ro o
s; o. z: _i o D-
i-^ in •—
t^ r- r— el
CM CM Ol
z. ce a: <£
D^ ^
C C "
aj a> ^
^ . 'i
O re in s- «x in 01 m
r-~ CM 0
co—-~ T- — • o ic^~
o o in o »^* *—- o ^ — ro c
CMCD i — C3 rod cno i — CD
n -o i~~ >— 0 m
XCMCXr— X«3- Xr— X^,
•— •— •— •" •"
*. 1
in u
0) C *
, — i — 3 • ro
r— i- *"3 *~3 '
I 5 x:z B
"t (_) 4-> C7>
QJ . O * (O *
o. c: 0> g c u tn
S c •** • C o m vi
o m-—
£ i_ re re
«_ m O U
o " «->"-
O
O
-.5.5-15
-------
ID C\J i— CXJ
§
Dl >> O)
•r- t. -I-
Q) -t- t—
i- OJ Ul r—
O)
u
i— 10
x in x •— x.—
CO O i—.
E -r^
•-KC
CO
LO
*>•§
.
03
a
+»•- -MT3 O
1- rt3 CD -i-
00 J. i— J.
O--= 1- O CJ
-------
-O
0)
3
O
O
00
I
in
CQ
et
re
(C '
•O in
e Q
•*-> in
in in
01
O CM
t- O
c
o
re
ro
c:
re
in
V
.0
£
-5-
in
c
HI
o
O) CXI
: o -o
^ •*-> -r-
^ .Q re
s-'
|
Q.
U
O>
c
Ol
in
Ol
o
£fc
•«-> >>
re o
*•> E
C 0 Z
o ^-*
•<-> a? S
O) **— ' d
ce c:
•0 o
Q) O
CD 4->
u- in
- re f*
>>T3 >>
.<-> ~-. re
•r- OIT3
re3^
0. X C
re o
CJ . *>
o *• ~*
c
0
.1
o
£?
re
is
O
U
^
z
z
in
re
~^ O)
•»> &-
o> >, oi
3 re o
_J Q Q-
C TJ
Q. U
to re
0
CM in
X CM
CM
£
i.
a>
u_
in
o •
0 Z
o
o
01 re
it- U
1- -r-
3 E
O Ol
+-> f
1/5 O
tf
Z
z
in
c
0
in
re
o.
ce
t-
14-
uo
CM <— • •
in in
x in
— ""'
1C
U-
^
Q
t-
re
eo
E .
O) CX
0 0
u
*->
•i- re
i35-"
•— CM
en oo
o •—
i i
m o
oo r-
o^J
*~*
z
o
c
re
in
c.
o
z
^
r—
CM
z
u
l*-
5
ro
OO «^^
ro in
i— CM
IT)
CM
cj
Z
re
t-
o
t.
3
*
Ol
tn
re •
x u
(— "— '
0
CM CO
^- CM
i i <:
VD CM Z
t— ro
C3 0
*~*
Z Z
0 0
c. c
re re
in in
c c
0 0
^ *E
10 to
r- r^
r~ en
z ce
i. i-
i*— i«-
3 3
to irt
in to
•— O i— O
10 10
X •— X —
— ^ — ""
0
re ••-
iZ o
*^ s^
o o>
re o
CO CJ
o>
u
re
t.
C3
• 0
ce o
£
2 «S PO
01
fc: TJ
O) CD
O. <->
O
o
o
o>
z:
~^
CM
s?
Ol
o
Q.
O)
!-
E
01
lo
£
CL
re
O>
U
Ol
o>
•o *>
•i- 1C
u u
co re •!—
TD
4J H- C
00-^
01 J<
11 ^ i
« Ol i—
Z -* CO
5.5-17
-------
5.5-18
-------
C-l
•H
t/5
|
CL.
O
i — i
O
to
o
1— I
I—
C/5
O
O
o
CO
CO
co
a.
o;
o
CO
CO
>
IB 0
•<-> E
00
£ o z
o • — •
+•* o£ 5
ai — 33
i -a*
•a o
01 O
0) *J
u_ in
. IB'—
4J ~~. IB
•r~ en"o
IB in
0. X C
IB O
O • +J
O— '
z
o
) •
IB
O
o
_J
Si
c
IB
a
•r- +J O
IB U
C.— •—
O •!-
gal
J3 x:
s- s. *>
O O'i-
o in x
LD CM
r- LO
I— CO
^""*
o
0
CO
cu
3
O
c
I-H
s
""*•
a:
i.
3
oo
CO
LO'—
«* O
X LO
-"".
IB
O-
•t
^ .
Z
s
jc a.
•a o
0)
*^ U
«*••-
i
in ^J CU
0.
Fusion of packing
corrosion of scru
ber exterior; exc
sive pressure dro
CO LO
Ol CO
O r-^
O
LO
L.
0)
0)
1—
g
cs?
°^
i.
3
14-
3
IO
CM. — .
X CO
-
CU
r
"^
c
IB IB
3r _i
§
in •
3 W
IS
•t
&. in
OJ 0)
CO •(->
in
01
Stack corrosion;
sagging of demist
LO O
CVJ^
1
Vt
IB
cn
i.
o cS
LO
r—
r-
0:1
s-
oo
r—
f^o'
LO
X CO
•~"~'
"
Q
S- IB
IB »—
CO U-
1
in
3
•o
C
LO
• Ol
1 1 .^
• fc.
VC CM
co r>-
O
o
in
IB
>> cu
££.
LO
LO
CU
i.
3
o'o'
cno
X r—
•~~"
•
Q
•M .
(- IB
IB i—
CO U.
Demister housing
had corrosion
CO f-
Z CM LO
Z §
LO
S
0 C
•i- IB
§in
c:
5 i
cC S
r-
g^
z a:
s-
3
oo
CO
o. — .
•X. *f O
Z LO
X "3"
—
>> c
4-> 0
•l~ 4-*
O cn
4-> -r- It-
IB IB • — r—
ai d 3 CJ
1
S- •
S- 0
•i- «e •
.— CL
•— C S-
o o o
t_> ja o
cu
3 4J
in u
i*- S *J
o +•> «
e""E
o cn
•i- C -O
in -i- -r-
O T3 >
«_> S- J3
«J- CO
cn co
T— CO
o
LO
CO
\>1
O
JD
IB
Ol
CL.
r-
O
"
=
*»—
3
oo
LO
LO O
O
x r~-
CM
£
o
U IB
Z tH
onserv
u
Blower corrosion;
instrument over-
sizing
CO'LO'
CO CO
«s- cn
o
CM
co
in
IB
cn
>, (jj
IB O
a cu
LO
CO
i-
3
3
LO
CO
coo
x cn
CM
S
o
S- IB
IB i—
CO U.
in
O>
ir
IB <->
i— in
E-i
IB C
U. i—i
co co
VO CO
f Ol
o
CM
co
LO
LD
i.
3
«]
3
to
cno"
cn o
X i —
CM
~
o
i. IB
-------
o^-
mo i— o
t— CO
U
to
•
10
LU
_J
CO
0> 0)
J3 « -^ 10
E > O O
-------
tion of a flue gas and of an acid plant tail gas is in the adiabatic satura-
tion temperature, which is much lower for the tail gas. The lower moisture
content of the saturated tail gas (4 percent) further enhances the mass
transfer by allowing the tail gas absorber to operate at about 32°C (90°F),
and fewer mass transfer stages are required for the same S02 removal effi-
ciency.
- Table 5.5-5 summarizes the status of sodium sulfite (Wellman-Lord)
scrubbing systems on acid plants in the United States as of December 1977.12
Adsorption process—In adsorption systems, the tail gas is first passed
through a mist eliminator and then into an adsorbent bed, which effectively
removes the S02, S03, and acid mist. As the bed approaches saturation, the
tail gas is passed to another bed, and the saturated bed is regenerated with
a stream of hot, dry air. The effluent purge stream, rich in S02, is fed
back into the acid plant. The adsorption/regeneration cycle operates con-
tinuously and automatically.
Even though any of the commercial adsorbents could be a suitable mate-
rial, only synthetic molecular sieves or zeolites have been used in commer-
cial applications of the adsorption process. As shown in Table 5.5-6, the
process has not achieved satisfactory operations in any of the acid plants
in the United States.12 •
Limestone scrubbing—Limestone scrubbing is described in detail in
Section 4.2.3.2. Table 5.5-7 summarizes the status of limestone scrubbing
systems on domestic sulfuric acid plants as of December 1977.
Hydrogen peroxide scrubbing15—One U.S. chemical company uses hydrogen
peroxide (H202) scrubbing to control S02 emissions at two sulfuric acid
plants. In this process, S02 in the gas stream is reacted with hydrogen
peroxide to produce sulfuric acid. Dilute sulfuric acid (typically <50%)
containing a small amount of H202 (<0.1%) is circulated over polypropylene
packing in a scrubbing tower made of fiber-reinforced -plastic. A rapid,
high-yield reaction takes place in the recirculating acid medium, and the
acid produced becomes part of the plant's product through blending with
high-strength acid in either the drying or the absorbing towers.
5.5-21
-------
o
co
o
s
oo
cc:
UJ
tr>
in
LO
UJ
5.5-22
-------
CM
"oo
t—
_i
CL.
O
l — i
0
"t
O
1 — 1
fV
Li.
_j
13
OO
O
t '1
r—
00
UJ
y
O
0
•z.
o
1 t 1
1—
00
•^
o
1 1
r—
D;
o
oo
0
^t
u-
o
ro
oo
—i
£5
oo
*
i
LO
•
LO
UJ
1
CO
1^^
r^
X3
5
fO "
C
i°
+j i/>
(U
o
*•> Irt
c
o
o
IO
re
•o
c
ro
Cu
l/>
1
XI
0
0.
l/>
4->
C
i
u
Ol CO
z o-o
^t XI IO
1
0-
i-
01
c
O)
•r-
l/>
ai
o
|^
L. --^
4J §
1/1
+,
•T o z
• o ^-r
•^2;
Ol ^_^ Q>
IX C
T3
>1
^4o >^
4-* ^* ro
t- CT>T3
ro in
O. X C
ro 0
0 • -M
O— '
z
c
o
ro
(j
O
«J
£
0
1 f— r— O 1 O •-
1 0. 0*0 EQJM-l)- 'f.'r
4.J Q ft) O XI *r- ^ CU >>XI
CC4^ >
>> 5i/)>in1oE+J ri'sa'ota)'^'S"Q
^ 4-> ro O '^^ c S c .c cu x: c
ro ^~ c r— t- +J cy o in x: 10 D.4J 4J *c o
oa> roio-ocuine OIC-UE -CST^R
•r- > >> U C 4-> >> 0) EO inO)lO*J4J*i
E T- C-o ••- ro r- -in u •>- *> •- c > r- fc- a. C
O4->roa>+-> ro ro l--4->x:COpO.a>>-ra
0 ro gm S-S gi'o'g. •§ | f-i -g § J U g
LtJ -4-> O QJ T- t/) *r- +J i- SZ M— .O ^ (J •*- +J L> (D £
g §
0 0
o ">
T3
C -i- C
O XI ••-
•E fe =o
^ CJ
01 • en
•D T3 T3 T3
-1-3 -i-3
L) r~ O ^~
•S in
I* ^
>>-«= o-—
t/> O 3 •— '
in •—
c ro O
0 u
*e ^
*35 Ol •
ox: x: o
t3 cj mo
^~
o
CM
^
*J
C
a>
o
i-
01
o.
o
o
*i
0
€
C>M
o
* •
r-. cu o>
cr, r— -^:
i— XI
ro u
^ ,•-
01 -i- -o
Xi ro '^~
§> o
ro :ro
ai •*-> >*-
o o o
"o ii S
in ^ C7)
«t z J<:
re Xi o
5.5-23
-------
) O) .» 4-> •- I
- 'f~ E ^~
J C S •— -T>
•f— (J & •
: ^ *O T3 ^ Oi
) i— Q) QJ 3
IB 3 CLf— .>
O
^—^
a;
to
CJ)
oo
LU
00
LU
•g
(O
Sin,—
—~ O O
ir> in f— u>
CM r--
in
•
10
CO
IOr- +J i—
CO 1- T3
5.5-24
-------
Because the reaction is
S02 + H202 -> H2S04,
the "make" acid from the scrubber is usable as "drip" acid in lieu of dilu-
tion water. There is no byproduct and no purge stream to dispose of.
Process experience at the two plants has shown that the process is
stable and easy to control. The effects of acid plant upsets are moderated
by the scrubbing facility. Both plants have been in compliance with local
standards covering emissions from existing sources. Although the process
does not generate acid mist per se, mist entering the tail gas scrubber
picks up the dilute recirculating acid; this increases the size of the
droplets and the visibility of the mist. Thus, a high-efficiency mist
eliminator must be used for opacity control.
No published data are available regarding the costs or energy and
environmental impacts of this process. An approximate capital investment of
$2.5 million, however, is reported for a 360-Mg/day (400-ton/day) plant.
The electricity required for circulating scrubber reagent is about 75 kW at
a circulation rate of 0.127 ms/s (2000 gal/min). The hydrogen peroxide
consumption is reported to be 0.5 kg/kg (0.5 Ib/lb) of S02 removed.15
5.5.2.2 Control Costs--
Following a recent study conducted by the Tennessee Valley Authority
(TVA), extensive cost data were published on retrofit emission controls
applied to acid plants burning elemental sulfur.16 The report identifies
the sources of the economic data and also lists the assumptions involved.
In the data on ammonia scrubbing, no credit is taken for the byproduct
ammonium sulfate, and the equipment for crystallization of ammonium sulfate
is not included in the capital costs. Estimates of operating costs of the
Wellman-Lord process include neither credit for possible sales of byproduct
sodium sulfate nor costs of disposal if the byproduct is not marketable.
Because of the specialized adsorbent in the Purasiv-S process, the estimate
includes a service contract for adsorbent renewal rather than costs of
outright purchase of new adsorbents.
Capital costs—The data presented in Figure 5.5-6, show that on four
retrofit systems the ammonia scrubbing process is the least expensive, the
5.5-25
-------
4.0 -
to
o
•u
in
c
o
CO
to
o
o
3.0 -
£ 2.0
a.
S
S02 IS 10% AT CONVERTER
INLET, 3000 ppm IN EXHAUST
GAS FROM FIRST STAGE AND
300 ppm FROM RETROFIT
CONTROL SYSTEM.
, PS PURASIV-S
, DA DOUBLE ABSORPTION
, WL WELLMAN-LORD
, AS AMMONIA SCRUBBING
PLANT CAPACITY, Mg/day (tons/day)
Figure 5.5-6. Capital costs of S02 control systems
for domestic sulfuric acid plants.16
5.5-26
-------
Wellman-Lord system is more expensive, and the two most capital-intensive
systems are the double adsorption and Purasiv-S processes. These costs are
for a sulfur-burning contact acid plant with S02 concentration at the con-
verter inlet of 10 percent. For cold gas plants, which have lower S02
concentrations and higher gas volumes, the costs will be much higher.
Operating costs—The data in Figure 5.5-7 show considerable variation
in overall operating costs, depending on plant size. The data indicate that
the double absorption and Purasiv-S systems are cheaper to operate for
plants with capacities of 45 to 90 Mg/day (50 to 100 tons/day). For plants
producing 227 Mg/day (250 tons/day), the double absorption and ammonia
scrubbing are the least expensive to operate; and for plants producing 680
to 1360 Mg/day (750 and 1500 tons/day), the lowest operating costs are
associated with ammonia scrubbing.
Application of the ammonia scrubbing system would mainly depend on the
marketability of the generated ammonium sulfate as fertilizer. The costs
will also depend on such site-specific factors as location and availability
of space for retrofit.
5.5.2.3 Energy and Environmental Impacts—
Double absorption—Double absorption can reduce S02 emissions below
300 ppm and S03 and acid mist emissions below 0.05 g/Nm3 (0.022 gr/ft3).12
It produces no solid or liquid waste stream and thus has no secondary
pollution effects. In addition, the sulfur consumption is low; in compari-
son with a single absorption plant operating at 97 percent conversion
and producing 907 Mg/day (1000 ton/day) of 100 percent acid for 350 days/
year, a double absorption plant at 99.7 conversion and the same production
rate consumes nearly 2900 Mg (3200 tons) less sulfur per year.
A double absorption plant, however, consumes more energy and produces
less bonus steam than its single absorption counterpart because the gas from
the primary absorber must be reheated to conversion temperature. Further,
extra power is needed for the blower, more cooling water circulation is re-
quired, and the acid recirculation pumps for the added absorber consume
extra power. Together, these typically add the equivalent of 4 MW to the
power demand of a 907-Mg/day (1000-ton/day) acid plant.17
5.5-27
-------
t
in
to
5
-------
Adsorption process—The Purasiv-S system is completely equivalent to
the double absorption system in environmental impact. It could, in fact,
reduce the S02 emissions below 50 ppm; however, as shown in Table 5.5-6,
this process has not proved to be commercially successful.
The energy penalty, primarily due to the pressure drop in adsorber/re-
generator and the heating of ambient air used for regeneration of adsorbent,
would be equivalent to about 1 MW for a 907-Mg/day (1000-ton/day) plant.18
Ammonia scrubbing—Ammonia scrubbing can achieve S03 and acid mist
removal comparable to that by the double absorption process. It often
operates at 90 percent S02 removal efficiency. The process generates
ammonium sulfate (liquor or crystals) as a byproduct and may cause formation
of a blue haze from emission of very fine particles of ammonium sulfate and
sulfite.
Primary energy consumption is by the blower to overcome the pressure
drops in the absorber and the stripper, and by the recirculation pumps.
Based on data for combustion sources, the energy penalty for a 907-Mg/day
(1000-ton/day) plant is estimated to be equivalent to 0.6 MW.
Sodium sulfite scrubbing—The environmental impact of sodium sulfate
scrubbing is identical to that of ammonia scrubbing except that it does not
produce blue haze and the byproduct is sodium sulfate instead of ammonium
sulfate.
The energy penalty would be greater than for the ammonia scrubbing
process because the evaporators in the regeneration subsystem consume steam.
Again, based on data for combustion sources, the energy penalty for a
907-Mg/day (1000-ton/day) plant will be about.1.1 MW.
Limestone scrubbing—Although limestone scrubbing provides adequate
removal of the pollutants (S02, S03, and acid mist), it also generates a
waste product, calcium sulfite/sulfate sludge, that needs proper handling
and disposal. The sludge could be thickened, dewatered, stabilized, and
landfilled; however, operators of limestone FGD systems on combustion
sources currently tend to discharge unstabilized sludge to settling ponds.
The primary pieces of energy-consuming equipment are the booster
blower, recirculation pumps, ball mill (if any), thickener, filter, and
conveyors and pumps. The overall energy penalty for a 907-Mg/day
(1000-ton/day) plant is estimated to be 0.8 MW.
5.5-29
-------
REFERENCES FOR SECTION 5,5
Bucy, J.I., et al. Potential Abatement Production
Byproduct Sulfuric Acid in the U.S. EPA-600/7-78-070
17-18.
and Marketing of
April 1978. pp.
4.
5.
6.
7.
8.
A
Acid.
R.N. Chemical Process Industries. 3d ed. Sulfur and Sulfuric
New York, McGraw-Hill Book Company. 1967. Chapter 19. pp.
Calvin, E.L. , and F.D. Kodras. Inspection Manual for the Enforcement
of New Source Performance Standards as Applied to Contact Catalyst
Sulfuric Acid plants. Prepared for the U.S. Environmental Protection
Agency DSSE, Contract No. 68-02-1322 by Catalytic, Inc., Charlotte,
N.C. November 1976. Figure 1. p. 15.
Calvin, E.L., and F.D.
Startup, Shutdown and
Figure 7, p. 45.
Kodras. Sulfuric Acid Plant Emissions During
Malfunction. EPA-600/2-76-010. January 1976.
Donovan, J.R., and P.J. Stuber. The Technology and Economics of Inter-
pass Absorption Sulfuric Acid Plants. (Presented at the AIChE Annual
Meeting. Los Angeles. December 1-5, 1968).
Drabkin, M. , and K.J. Brooks. A Review of Standards of Performance for
New Stationary Sources—Sulfuric Acid Plants. Prepared for the U S
Environmental Protection Agency, Publication No. EPA-450/3-79-003, by
4-TI6 '°n °f the MITRE CorP°ratl"on> McLean, Va. January 1979.
U'5' Environmental Protection Agency, Office of Air Quality Planning
and Standards. Final Guideline Document: Control of Sulfuric Acid
Mist Emissions From Existing Sulfuric Acid Production Units. Research
Triangle Park, N.C. EPA 450/2-77-019. September 1977.
Engineering Science, Inc., Washington, D.C. Exhaust Gases from Combus-
tion and Industrial Processes. Prepared for the Division of Compli-
ance, Bureau of Stationary Source Pollution Control, Office
Programs, U.S. Environmental Protection Agency, Durham, N C
861. October 1971. pp. IV-71.
9. Ref. 4, pp. 16-17.
10. Ref. 6, pp. 4-24, 4-25.
of Air
PB-204-
5.5-30
-------
11. Ref. 4, Figure 2, p. 18.
12 Tuttle, J.D., et al. Summary Report on S02 Control System for Indus-
trial Combustion and Process Sources. Vol. IV: Sulfuric Acid Plants.
Prepared for the U.S. Environmental Protection Agency, IERL, under
Contract No. 68-02-2603, Task No. 4, by PEDCo Environmental, Inc.,
Cincinnati, Ohio. December 1977.
13. Ref. 6, Table 5-1. p. 5-2.
14 Friedman, L.J. Ammonia Scrubbing of Sulfuric Acid Plant Tail Gas.
(Presented at the Fertilizer Institute Meeting, New Orleans. January
1976.)
15. Letter and attachment from Kusko, J.D., E.I. du Pont de Nemours & Co.,
to Shah, Y.M., PEDCo Environmental, Inc. October 25, 1979.
16. Ref. 1, pp. 165, 166.
17. R.M. Parsons, Company. The Parsons Double Catalysts/Double Absorption
Sulfuric Acid Process. 1970.
18 Collins J.J., et al. The Purasiv-S Process for Removing S02 From
Sulfuric Acid Plant Tail Gas. (Presented at 66th Annual AIChE Meeting.
Philadelphia. November 15, 1973.)
5.5-31
-------
-------
5.6 PULP MILLS
Pulp and paper manufacturing, one of the 10 largest industries in the
United States, is conducted in two phases: pulping of wood, and production
of paper and related products from the pulp. In the pulping process, wood
is reduced to fiber, sometimes bleached, and dried. Most pulp mills use a
chemical cooking liquor to dissolve lignins and free the wood fibers, then
recover the chemicals by a combustion process. The pulp manufacturing
process generates gaseous and particulate emissions in quantities that
depend on the type of pulping operation, the type of recovery process, and
the effectiveness of control equipment.
5.6.1 Process Descriptions and Emission Sources
Three major pulping and recovery processes account for nearly 80 per-
cent of the pulp produced in this country: sulfate (kraft), sulfite, and
neutral sulfite semi chemical (NSSC). The other 20 percent is produced by
specialized processes.
The sulfate or kraft process has created the greatest air pollution
problem, mainly because of the large quantity of visible particulates and
highly odorous reduced sulfur compounds. Acid-based sulfite and neutral
sulfite pulping mills are important in this study because, more than other
types of mills, they emit S02. Many of these mills use highly efficient S02
absorption systems as part of the chemical recovery system.
5.6.1.1 Sulfate Process--
As with most modern pulping processes, kraft or sulfate pulping (Figure
5.6-1) begins in a digester, where wood chips from debarked logs are cooked
with a chemical solution. In the kraft process, the cooking solution, known
as white liquor, is sodium hydroxide (caustic soda) and sodium sulfide. The
caustic soda in the cooking liquor permits the pulping of nearly all wood
species. The liquor and wood chips are carefully cooked under controlled
conditions of temperature and pressure. During digestion the liquor dis-
solves the lignin in the wood and thus frees the cellulose fibers. When the
cooking is completed, the residual pressure within the digester is used to
force the contents to a blow tank. The sudden decrease in pressure and the
5.6-1
-------
O)
o
o
Q.
O
O
Ol
"O to
E 0)
re O
=5 O
Q.-I-
to
O) LO
+-> -i-
fO JH
H-. OJ
ZS i—
to tc
•r—
4J .4-)
fO
S-
O
CL
o
•I— •
CL
I
to
cu
5.6-2
-------
impact against the tank walls cause the soft wood chips to explode and
become a fiber pulp. Gases and steam from the blow tank are vented to a
condenser. The noncondensable gases, which are the sources of odors, are
either confined and treated or released to the atmosphere.
The pulp, containing spent liquor (called black liquor), is diluted in
the blow tank and then is pumped to washers. The pulp is washed counter-
currently to remove black liquor and waste wood products. The pulp may be
bleached and sent to subsequent operations for formation of paper products
or dried and sold as market pulp.
A typical kraft waste liquor recovery process begins with multiple-
effect evaporators, to which black liquor containing about 15 percent solids
is pumped from the pulp washers. The liquor is further concentrated in
direct-contact or forced-circulation concentrators to a concentration of 60
to 70 percent solids. Weak liquor from the pulp washers or strong black
liquor from the multiple-effect evaporators is pumped through an oxidation
tower to facilitate odor control and chemical recovery. Oxidation converts
sodium sulfide to innocuous salts to prevent the release of hydrogen sul-
1 * The concentrated black liquor then is sprayed into a recovery furnace,
where the organic content supports combustion. The inorganic compounds,
consisting of the cooking chemicals, fall to the bottom of the furnace where
they form a molten smelt that drains into a smelt-dissolving tank. The
dissolved smelt, containing mainly sodium sulfide, sodium sulfate, and
sodium carbonate, is called green liquor.
The green liquor is clarified, then mixed with slaked lime in a causti-
cizer to give products of sodium hydroxide, sodium sulfide, and calcium
carbonate. The calcium carbonate is removed in a clarifier; the resultant
clarified (white) liquor is recycled as cooking solution. The calcium
carbonate is pumped to a thickener, then discharged into a rotary kiln to
produce lime, which is sent back to the slaker.
Various reactions among the kraft mill cooking chemicals generate
characteristic gaseous emissions, including malodorous reduced sulfur com-
pounds. Such compounds are methyl mercaptan (CH3SH), hydrogen sulfide
(H2S), dimethyl sulfide (CH3SCH3), and dimethyl disulfide (CH3SSCH3). Kraft
5.6-3
-------
mills also generate oxides of sulfur, but to a much lesser degree. The
major source of S02 emissions is the recovery furnace in which the sulfur-
containing black liquor undergoes combustion. The furnace can emit some S03
under certain conditions. Table 5.6-1 shows typical S02 and S03 emissions
from kraft pulp mill combustion sources.
Concentrations of S02 from the recovery furnace depend on the fol-
lowing: 1) sulfidity of the cooking liquor, 2) manner in which liquor is
sprayed into the furnace, 3) ratio of primary to secondary combustion air,
and possibly 4) liquor firing temperature.2 Lesser quantities of S02 can, be
released from the lime kiln and smelt dissolving tank.
The major potential sources of particulate emissions are the same
sources that generate S02 (i.e., the recovery furnace, smelt dissolving
tank, and lime kiln).
For kraft process units, the process information on various pieces of
equipment is given in Table 5.6-2.
5.6.1.2 Sulfite Process—
Sulfite pulping is similar to kraft pulping except that a sulfurous
acid base solution is used to dissolve the lignin in the wood chips rather
than a caustic solution. A bisulfite of sodium, calcium, ammonia, or mag-
nesium is used to buffer the cooking solution.
After the wood chips have been cooked in a digester, the pressure in
the digester is reduced and the contents are charged to the blow tank, where
the chips are defibered. The pulp is screened for knots and foreign sub-
stances such as grit, then separated from the spent liquor in a series of
washers. The pulp is sent on for bleaching and finally the papermaking pro-
cess.
In the past, much of the spent sulfite cooking liquor was discharged to
sewers, but greater emphasis is now placed on burning the liquor, which
supports combustion when concentrated to approximately 55 percent solids
content, to reduce liquid effluent, recover chemicals, and generate steam.
In calcium-based systems (only relatively few are now operating)
chemical recovery is not economically practical. Some mills thicken the
liquor in evaporators and sell it for use in various products such as animal
5.6-4
-------
TABLE 5 6-1 TYPICAL EMISSION CONCENTRATIONS AND RATES FOR
'SO FROM KRAFT PULP MILL COMBUSTION SOURCES3
Emission source
Recovery furnace:
No auxiliary fuel
Auxiliary fuel added
Lime kiln exhaust
Smelt dissolving tank
Concentration, ppm by vol.
S02
0-1,200
0-1,500
0-200
0-100
S03
0-100
0-150
Emission rate, kg/Mg
S02
. 0-40
0-50
0-1.4
0-0.2
S03
0-4
0-6
5.6-5
-------
i— n
i— CM U>
LU
CO
CO
LU
O
Q-
U-
i—O r-
•—i— (M
in
LU
CO
t- ••- O) - o -
O) u ro o> u) oj
Q. 0) S- O) C7>
(0 0) nj
O) E -
s .>>
5.6-6
-------
feed and dispersants. For environmental reasons, one mill did install a
recovery system in mid-1973 as a standby for use when the thickened liquor
cannot be sold.5
When recovery is practical in magnesium-, sodium-, and ammonium-based
systems, the spent liquor is pumped through multiple-effect evaporators and
then is sprayed into a recovery furnace, similar to that in the kraft pro-
cess. When the magnesium-based liquor is burned, magnesium oxide (MgO)
particles and S02 are liberated and exit with the flue gas. The MgO is
collected in cyclones and later slaked to form a magnesium hydroxide
[Mg(OH)2] slurry. The hot gas stream containing S02 leaving the cyclones is
cooled, and the S02 is absorbed by the Mg(OH)2 in absorption towers to
produce the usable pulping acid. This system is described later in more
detail.
Recovery of chemicals to regenerate pulping acid in the sodium-based
system is somewhat more involved than the MgO system. Usually the sodium-
based liquor is burned in a reducing atmosphere. The flue gas contains S02,
and some sodium sulfate particles, which are collected in electrostatic
precipitators and returned to the furnace. Most of the inorganic dry solids
fall to the bottom of the furnace and form a smelt consisting of sodium
sulfide and sodium carbonate. When the market is good, some sodium-based
mills can sell the smelt to kraft mills for use in their process as "green
liquor." The S02 from the furnace is cooled and passed through an absorber
utilizing a sodium carbonate scrubbing solution. The resultant solution,
sodium sulfite/bisulfite, is returned as pulping acid. Sodium-based mills
that totally recover their furnace products utilize one of several modern
recovery processes (e.g., Stora, SCA-Billerud, Tampella). The Stora pro-
cess, which is commonly used, is described in the following paragraph.
The smelt from the recovery furnace is dissolved in water and clari-
fied. The resulting "green liquor" is carbonated with pure carbon dioxide.
The sodium sulfide is converted to H2S, which is stripped from the liquor by
carbon dioxide as the carrier gas. The H2S reacts in a Claus reactor with
S02 to form elemental sulfur. The carbon dioxide is recirculated to the
5.6-7
-------
process. In this way the green liquor is converted to sodium bicarbonate
(NaHC03), which is then reacted with sodium bisulfite to produce sodium
sulfite.6 Part of the sulfite solution absorbs S02 in an absorption tower
and is returned as sodium bisulfite. The sodium sulfite is returned to the
pulping process as pulping acid.
Ammonium-based mills can practice only partial recovery, by burning the
spent liquor, cooling the flue gas, and absorbing the S02 in the flue gas
with an ammonium hydroxide slurry to produce ammonium sulfite/bisulfite
pulping acid. The ammonium base is broken down to nitrogen, nitrogen ox-
ides, and water vapor during combustion and thus cannot be recovered.
Unlike the kraft pulping process, in which the cooking solution con-
tains sulfide/sulfate chemicals, the sulfite pulping cooking liquor contains
a sulfurous-acid base and sulfite/ bisulfite compounds. When heated, these
latter compounds give off considerable quantities of S02. Because sulfides
are not present, no organic reduced sulfur compounds are produced in the
sulfite process. Hydrogen sulfide emissions may occur if alkaline sulfite
liquor is burned in recovery furnaces under reducing conditions.7
The main sources of S02 emissions at sulfite mills are the digester
blow tanks, multiple-effect evaporators, and the chemical recovery system
(if recovery is practiced). The makeup acid preparation plant and the pulp
washers are minor sources of S02. Table 5.6-3 shows a range of typical
emission factors from controlled and uncontrolled sources of S02.
For sulfite process units, process information on various pieces of
equipment is given in Table 5.6-4.
5.6.1.3 Neutral Sulfite Semichemical Process—
The NSSC process differs from the kraft and sulfite pulping processes
in two major aspects. The cooking liquor is a neutral solution of sodium
sulfite and sodium bicarbonate or sodium carbonate, rather than being acidic
or basic. The sulfite ion reacts with the lignin .in the wood, and the
sodium bicarbonate acts as a buffer to maintain a neutral solution.8 More
importantly, only a portion of the lignin is dissolved in the digester
during cooking. The partially cooked wood chips then are subjected to
mechanical attrition to produce the pulp. This chemical/mechanical method
produces pulp yields of up to 80 percent, versus 42 to 58 percent yields
typical of complete chemical pulping processes.9
5.6-8
-------
TABLE 5.6-3. TYPICAL S02 EMISSION FACTORS FOR
SULFITE PULP MILL SOURCES30
Emission source
Blow pit:
Hot blow
Cold blow
Evaporators
Recovery process
Washers
Acid preparation
Emissions, kg/Mg (lb/ton)a
Uncontrolled
30-75 (60-150)
2-10 (4-20)
1-30 (2-60)
80-250 (160-500)
0.5-1 (1-2)
0.5-1 (1-2)
Control! edb
1-2.5 (1-5)
0.05-0.3 (0.1-0.6)
0.025-1 (0.05-2)
6-20(12-40)
Per mass, Mg (ton), of air dried pulp.
Alkaline scrubbing of gases.
5.6-9
-------
OJ -*!.
•i- C.
i— to
O) 4->
O) =3
+J TJ
in
OJ T3
D3 C
•r- to
01
CM
CM
tO
LU
to
LU
O
S- (J
c
O S-
U =3
QJ U TJ
to in o
S 05T3
o s:
• — v. s-
•*->o
CO
O) (LI
O5 -
co
U •!-
CO O) i—
V)
C
> CO
to
to
OJ
CO
>
CO
to
+J
to
cu
+->
CO
5.6-10
-------
The fate of spent liquor from the pulp washers depends on arrangements
at the specific pulp mills. Some mills recover the cooking chemicals in a
manner similar to that used at sulfite mills, i.e., by concentrating the
spent liquor and then burning it in a recovery boiler. Mills without NSSC
recovery systems treat and discharge the spent liquor; if a kraft operation
is nearby, the NSSC spent liquor can be mixed with that from the kraft
process and burned in the kraft recovery furnace. The recovered chemicals
are used entirely in the kraft system.
The digester and blow tank emit S02 during the digester relief and
blow. In batch-type digesters, however, the S02 pressures during neutral
sulfite cooking are considerably lower than those exerted in acid bisulfite
cooking; thus S02 emissions from NSSC sources are lower than those from
sulfurous-acid-based (sulfite) operations. The evaporators are another
source of S02, usually less than 1 kg S02 per Mg of pulp (2 Ib S02/ton of
pulp).12 At NSSC mills with a recovery furnace (sometimes a boiler, reac-
tor, or waste liquor incinerator), this unit is the main potential source of
S02 emissions.
For NSSC process units, process information on various pieces of equip-
ment is given in Table 5.6-5.
5.6.2 Control Techniques
Use of the various S02 control techniques described here depends on
site-specific factors such . as the type of pulping operation, quantities of
S02 emitted, the applicable environmental regulations (particularly con-
cerning air and water), problems associated with retrofitting control equip-
ment, and the economics of obtaining and operating the controls. Often the
end products of the SO controls are chemicals that can be reused in the
/\
pulping operation and serve as a credit to offset operating costs.
5.6.2.1 Description—
Sulfate process—Although it is the malodorous sulfur compounds and not
S02 that pose the greatest air pollution problem at sulfate mills, there is
generally always some release of S02 from such mills, particularly from the
recovery boilers. Control devices are not used specifically for S02 control
at sulfate mills, but reduction of the pollutant can take place in certain
5.6-11
-------
•o o
0) -f->
M U
•r- (0
-a > cu
t? u
CU TO
> c
O £-
U 3
CU t-
CM
o
o
oo
CO
00
CM
00
in
CM
CM
LT>
C\J
CO
co
10
CM
CM
o o
co +->
CSJ
CO
CO
LU
to
CO
LU
O
CJ
co
co
LT)
I
ID
co
oo
CO
oo
(X)
us r^.
00 CM
o o
o o
<3- m
in co
CM
l-H
CM
CM
O t-H
O CM
CM ^^
O O
O rH
CO
CTt
re
cu
re
u -o
re to cu
t. v^-r-
<*-«
re
cu
re
= o
4-> O
re i—
s- -re
)
re
-uo
re
a. s-
•^* Z3
U 4->
CU C
S- CU
Q. >
J- S-
cu cu
en
cu
u
re
co
cu
lo
re
re
>
re
re
4->
re
•o •
cu
4->
re
re
5.6-12
-------
process equipment. For example, the direct contact evaporator used at most
kraft mills to concentrate black liquor also acts to absorb about 75 percent
of the S02 emitted from the recovery boiler and nearly all of the S03 (con-
centrations shown in Table 5.6-1).14
Lime kilns and fluidized bed calciners, although sources of S02 at
kraft mills, are generally low emitters (given as about 34 ppm S02 in one
report)15 because once S02 is formed, most of it reacts with lime and forms
calcium sulfite and calcium sulfate. In kilns equipped with wet scrubbers
for control of particulate matter, this reaction is particularly effective.
In a limited series of tests of a fluidized bed calciner, it was not pos-
sible to measure the presence of S02 in the exhaust gases.16
Tests of lime kilns in lime manufacturing plants showed that S02
removal efficiency with dry control devices ranged from 82 to 93 percent.17
Normally dry control devices do not remove S02 but in this case the intimate
contact between lime particulate from the kiln and the S02 occurs in the
kiln, and also in the duct and control device. These kilns ranged in size
from about 18 to 155 Mg/h (20 to 171 tons/h). Tests on three kilns using
high-sulfur coal with wet scrubbers showed removal efficiencies of 98 to
99.5 percent.17 Outlet S02 concentrations for all cases except one ranged
from 5 to 45 ppm. One particularly large kiln (155 Mg/h or 171 tons/h)
firing high-sulfur coal had an outlet concentration of 199 ppm.17
Sulfite process—Emissions from the blow tank/pit and evaporator con-
taining S02 need not be vented to the atmosphere, but can be directed to
acid towers. The acid towers are used to produce makeup cooking acid,
usually by absorbing S02 (produced from sulfur burners) with a solution of
the hydroxide of the base used for making pulp (e.g., magnesium hydroxide).
Proper introduction of S02-laden gas streams from digester blows and evapor-
ator vents to the acid towers can adequately control S02 emissions from
these two sources . Information from a 127-Mg/day (140-ton/day) calcium-
based sulfite mill shows that emissions from their acid tower vent before
and after tie-in of the digester blow stack emissions and evaporator vent
emissions remained approximately the same at 0.54 Mg (0.6 ton) of S02 per
day.18 Similar findings are reported at another calcium-based sulfite mill
in which digester off-gases and the S02 stream from a spent liquor flash
tank are tied into the mill's acid tower.19
5.6-13
-------
Chemical recovery systems are operated at all types of sulfite pulping
mills, whether the process is magnesium, calcium, ammonia, or sodium based.
Many sulfite mills installed the recovery systems because environmental
regulations prohibit discharge of the spent liquor to water bodies (e.g.,
streams, rivers) and the mills cannot market the liquor. Recovery of the
cooking chemicals is important today, especially at magnesium-based sulfite
mills, because of the cost of makeup chemicals. Installation of a chemical
recovery system reduces water pollution problems but causes generation of
S02 from the recovery boiler, furnace, or incinerator. The purpose of the
recovery boiler, however, in addition to burning the organic portion of the
thickened liquor to generate steam for mill use, is to liberate the S02 (as
a gas) and inorganic solids (as particulates or as smelt) so that they can
be collected and reused in the pulping operation.
In a typical S02 control/recovery system at a magnesium-based opera-
tion, S02 and magnesium oxide from the recovery furnace (Figure 5.6-2) are
passed through cyclones to remove most of the magnesium oxide, which is
discharged to a slurrying tank. The hot gas stream containing S02 exits the
cyclones and is cooled in a scrubber and/or cooling tower. The cooled S02
stream then is sent through absorption towers (often a series of three), in
which the S02 is absorbed by a magnesium hydroxide slurry pumped from the
slurry tank. The absorption product is a slurry containing magnesium bisul-
fite and sulfurous acid; the slurry is returned to the pulping operation as
cooking acid for the digesters.
Although the recovery boiler generates the S02, it is actually the
absorption towers (which are not 100 percent efficient) that emit the S02
from the chemical recovery system; i.e., the absorption towers are an integ-
ral part of the recovery system and are not a control system per se. None-
theless, because the absorbers are reported to operate at high S02 removal
efficiencies (95 percent plus at many mills) and with high operability, they
are considered here as an important S02 control technique.20
It is important for chemical recovery and for meeting S02 regulations
that the absorption systems do operate reliably and efficiently. If further
S02 control is required, another absorption tower(s) with auxilliary equip-
ment such as fans could be installed to receive off-gas from the recovery
5.6-14
-------
in
to
a)
o
O)
>
O
o
O)
en
O. 1/1
i— QJ
0. S-
r— O
(O
U
•i- E
E O
O) O
C Q.
C7)
res cr>
to o
u -c
OJ
CU
3
U-
5.6-15
-------
absorption tower. Operators of one magnesium-based sulfite process in-
stalled an ammonia absorption system to reduce S02 emissions from the mag-
nesium hydroxide absorption towers. It was reported that the ammonia system
reduced the S02 concentration from approximately 2500 ppm to 390 ppm for a
control efficiency of 84.5 percent.21
Neutral sulfite semichemical process—At NSSC mills that thicken their
spent liquor and burn it in recovery furnaces, the S02 generated is gener-
ally removed or controlled in absorption towers with a sodium carbonate
solution as the absorbent. Product from the tower is mainly sodium bicar-
bonate, which is used in the digester cooking liquor. Smelt from the re-
covery furnace consists of sodium carbonate and sodium sulfide. The sulfide
is not usable in the NSSC cooking liquor. In the Mead process, the clari-
fied smelt solution is treated with a gas containing carbon dioxide, which
converts the sodium sulfide to sodium bicarbonate while expelling hydrogen
sulfide; the hydrogen sulfide is converted by pyrolysis in the presence of
excess air to S02, which together with the S02-containing flue gas from the
liquor recovery process is brought into contact with carbonated liquor in an
absorption (S02 control) system to produce fresh cooking liquor containing
sodium sulfite and sodium bicarbonate.22
The S02 from other NSSC process sources (e.g., evaporators) can be
routed to the chemical recovery absorption towers or to cooking liquor
(acid) makeup towers.
5.6.2.2 Control Cost-
Data on costs of the techniques for controlling S02 at pulp mills often
are not available, and it is difficult to formulate estimates that represent
a typical installation. Mill operators control S02 in various ways, even at
mills of similar size with similar process steps. One mill may purchase
special equipment for S02 reduction, while another mill converts unused
onsite equipment into usable S02 control components. Also, although data on
costs of chemical recovery at a mill might be available, they seldom show
the costs of specific S02 absorption equipment.
The following cost information from trade journals pertains to specific
mills. All costs are updated to July 1979 through the use of Chemical
Engineering cost indices.
5.6-16
-------
In January 1975 an emission control system to recover S0x, remove
particulates, and recover heat from a recovery boiler was retrofitted to ITT
Raynoier's sodium-based sulfite mill in Hoquiam, Washington. Total capital
cost in 1979 dollars is approximately $3.6 million ($2.5 million repor-
ted).23 Flow rate of the gas stream from the recovery boiler is about
118 mVs (250,000 acfm at 340°F). Effluent S02 from the control system has
been reduced to less than 1 ppm.24 The primary control system components
are a low-pressure-drop venturi scrubber (1.7 to 2.5 kPa, 7 to 10 in. H20)
and a crossflow scrubber nucleator.
The reported operating costs were based on the design conditions of
operation and the criteria shown in Table 5.6-6.
The emission control system is capable of producing a net annual gain
of $644,000 ($496,750 reported); the pulp mill, however, is unable to assim-
ilate the full amount of sodium sulfate and heat, and therefore shows a net
annual gain of $169,000 ($130,000 reported).24
In 1976 a calcium-based sulfite mill installed a system for receiving
condensed digester blow gases containing S02 (42 to 63 g/s, 4 to 6 tons/day
average), low-pressure digester gas, and pressure absorber relief gases and
sending these gases to the acid plant for S02 absorption by the weak cooking
acid.25 The system, installed at a capital cost of $1.43 million ($1.16
million reported), includes the following: a 320,000-liter (84,700-gal)
stainless steel vessel to receive the gases, sized to accommodate two blows
in rapid succession; 284-1iter/s (4500-gal/min) heat recovery and 63-liter/s
(1000-gal/min) cooler pumps for steam condensation and gas cooling; a
25-1iter/s (400-gal/min) heat exchanger feed pump and three heat exchangers
to cool the 82°C (180°F) condensate formed during a blow; an eductor and a
95-1iter/s (1500-gal/min) eductor feed pump to absorb the cooled S02 gases
from the digester blow into the weak cooking acid; an additional gas fan
between the weak- and strong-acid storage towers to relieve'Overloading of
the existing gas fans; instrumentation that allows operation of the system
with no additional manpower; and rehabilitation of an existing deep well for
use in cooling raw acid and acid-cooler shower water during summer months.25
With a few exceptions, this system has eliminated S02 discharge (formerly
5 Mg/day [5.5 tons/day]) to the atmosphere from digester blows.26
5.6-17
-------
TABLE 5.6-6. CRITERIA AND ANNUAL OPERATING COSTS OF A SULFUR DIOXIDE
RECOVERY/CONTROL SYSTEM ON A SODIUM-BASED SULFITE RECOVERY BOILER
Criteria used to determine
annual costs
Recovered Na2SOd
Recovered S02
Electrical cost
Steam cost for turbine
drive and evaporation
Recovered thermal value
Amortization
Maintenance
Operating manpower
Operating time
Mid-1979
estimate
$104/ton
$26/ton
$0.04/kWh
$1.95/10^ Btu
$1.30/10b Btu
1 yr recorded
data (less than
1% of capital
cost)
•$19,500/yr
8200 h/yr
Reported - October 1975
$80/ton
$20/ton
$0.03/kWh
$].50/lo5 Btu
$1.00/10° Btu
Straight line--10 yr
1 yr recorded data
(less than 1% of
capital cost)
$15,000/yr
8200 h/yr
Emission control system capability—financial recovery (700 tons per day)
Credits/costs
Recovered (credits)
NapSO* at average inlet
1 gr/sdcf--4217 tons/yr
SOp at average inlet 1000
ppm— 6841 tons/yr
Thermal
Design 1270 gal /mi n
55,000,000 Btu/h
Total credit
Costs
Electrical power 216 hp
1.6 x 106 kWh/yr
Steam for turbine fan
drive 1200 hp
Amortization
Maintenance
Operation (24 man-hours/wk
Evaporation of water from
additional salt formation
Total costs
$/yr
437,000
178,000
585,000
$1,200,000
62,000
82,000
207,000 .
21,000
12,000
171,000
$555,000
$/yr
337,000
137,000
451,000
$925,000
48,000
63,000
160,000
16,000
9,000
132,000
$428,000
"5. "6-18
-------
In 1973 a sodium-based sulfite mill (110 Mg/day, 120 tons/day) instal-
led an evaporation, burning, and smelt-handling system. The cost was just
over $3.3 million ($2 million reported).27 The primary function of the
system is to thicken spent liquor of about 8 percent solids content
to 60 percent solids, burn it in a Broby-type furnace, collect the smelt
from the furnace bottom, and make smelt flake for sale to kraft mills.27
Practical processing of the spent liquor in this manner enables the mill to
meet effluent standards. The significance of this system for S02 control,
however, is that an integral part of the system is a two-stage venturi
scrubber, the first stage of which is used for S02 scrubbing with a weak
sodium carbonate solution. The first-stage scrubber recovers S02 from three
sources—the furnace (in the flue gas), the digester (vent gases), and the
acid fortification tower (vent gases).27
The principal cost items are the vacuum evaporator, two-stage venturi
scrubber (with S02 scrubber), modified Broby furnace, and smelt flaker.
In 1973 a calcium-based sulfite pulp mill (191 Mg/day, 100 tons per
day) installed a waste sulfite liquor concentration and burning plant simi-
lar to that described above for the sodium-based mill, but using direct,
triple-effect evaporation. Capital cost was $3.9 million ($2.4 million
reported).5 Part of the system is an S02 scrubber placed above the direct-
contact evaporator, similar to a venturi scrubber. An alkaline solution
(sodium carbonate) absorbs S02 from the incinerator flue gases. The main
components of the system are three vacuum evaporators, an air contact evapo-
rator, a direct-contact evaporator, liquor concentrate tanks, a vertical
incinerator, and a cyclone to remove ash from incinerator flue gas.
The only available operating . costs are those given earlier in Table
5.6-6.
5.6.2.3 Energy and Environmental Impacts—
Figures on the energy impact of S02 removal systems used in the sul-
fate, sulfite, and neutral sulfite pulping operations are not available.
The S02 removal systems discussed, however, are not really distinctly con-
structed as air pollution control methods to reduce S02 emissions. The
systems are primarily part of the overall chemical recovery systems employed
at the mills; thus the energy penalties/credits are intertwined with the
5.6-19
-------
other recovery system components (e.g., evaporators, chemical recovery
furnace). With respect to the overall recovery system, the sources of
energy consumption are pumps, evaporators, recovery furnace, coolers, S02
absorption tower recirculation pumps, and fans. The recovery furnace,
however, generates steam during the burning of the concentrated waste
liquor, which at least partly offsets the energy penalties.
It should be noted that the chemical recovery systems used at pulp
mills are inherently water pollution control systems, which eliminate the
direct discharge of spent liquor. Many mills, particularly sulfite and
NSSC, had to install chemical recovery systems to meet water pollution
effluent standards.
The environmental impact with respect to air quality can be viewed
differently for pulp mills than for chemical processes installing S02 con-
trol systems to reduce air pollutants. Because the sulfite process has the
greatest potential for emitting S02, it will be discussed. When a recovery
system is installed at a sulfite mill, a new source of S02 emissions is
created-the S02 absorption towers, which are at the end of the recovery
system. The importance of the S02 towers, however, is- that at many sulfite
mills, they operate at high S02 removal efficiencies (95 percent plus) and
high operability; thus it is an important S02 control technique.
Table 5.6-3 shows typical S02 emissions (per megagram and ton of air-
dried pulp) from sulfite mill sources. The impact of an uncontrolled
recovery process would be great (80 to 250 kg/Mg, 160 to 500 Ib/ton), but
would not be run uncontrolled because the loss of S02 would mean a loss in
sulfite/bisulfite chemical recovery. Operabilities and reliabilities of
several sulfite mills reported in one study were 100 percent.™ The
absorption towers are normally overhauled during scheduled pulp mill shut-
downs. Mills usually have a certain capacity for storing liquor in the
event an S02 absorption system must temporarily shut down for repairs.
Controlled emissions from the recovery process as shown in Table 5 6-3
are 6 to 20 kg/Mg (12 to 40 Ib/ton) of air-dried pulp." The same report
showing S02 system operabilities and reliabilities, also shows the outlet
concentrations (from the absorbers) from six sulfite mills. The average
concentration from the towers is approximately 225 ppm with a range from 60
(calculated from reported S02 values) to 400 ppm.10
5.6-20
-------
The pulp mill recovery system may be best summarized as a "controlled
source." Although a source of S02 emissions, the absorption towers at
reported mills have been very efficient in removing S02. The scrubber
products are not disposed of in the environment, but are recycled to the
pulping process.
5.6-21
-------
REFERENCES FOR SECTION 5.6
1. U.S. Environmental Protection Agency. Environmental Pollution Control:
Pulp and Paper Industry. Part I, Air. Technology Transfer. EPA-625/
7-76-001. October 1976. p. 9-1.
2. Blue, J.D., and W.F. Lewrlyn. Operating Experience of a Recovery
System for Odor Control. TAPPI. 54(7):1143-1147. July 1971.
3. Ref. 1, pp. 1-5 to 1-7.
4. Engineering Science, Inc. Exhaust Gases from Combustion and Industrial
Processes. Springfield, Va. NTIS PB-204861. October 2, 1971. pp.
IX"1 to 6.
5. MacLeod, M. Novel Approach to Sulfite Liquor Disposal: Evaporate
Sell, or Burn. Pulp and Paper. 48(9):58-62. September 1974.
6. Kirk-Othmer Encyclopedia of Chemical Technology. 2nd Edition Vol
16. 1968. New York, John Wiley and Sons, Inc. p. 718.
7. Ref. 1, p. 14-1.
8. Ref. 6, p. 700.
9. Ref. 6, p. 680.
10. Ref. 1, p. 1-10.
11. Ref. 4, pp. IX-14 to 17.
12. Ref. 1, p. 14-10.
13. Ref. 4, pp. IX-9 to 24.
14. Ref. 1, p. 10-53.
15. Ref. 1, p. 11-6.
16. Ref. 1, p. 11-8.
17. U.S. Environmental Protection Agency. Standards Support and Environ-
mental Impact Statement, Volume 1: Proposed Standards of Performance
for Lime Manufacturing Plants. Research Triangle Park, N.C. EPA-
450/2-77-077a. April 1977. p. C-13.
5.6-22
-------
18 PEDCo Environmental, Inc. Summary Report of S02 Control Systems for
Industrial Combustion and Processes Sources. Vol II: Pulp and Paper
Processes. U.S. Environmental Protection Agency. Research Triangle
Park, N.C. December 1977. p. 2-50.
19. Ref. 18, p. 2-63.
20. Ref. 18, pp. 2-43 to 2-111.
21. Ref. 18, pp. 2-103, 2-104.
22. Ref. 6, p. 701.
23 Teller A J et al. Emission Control for Sulphite Recovery Boilers.
Pulp and Paper Canada. 78(2):T37-41. February 1977. p. 4.
24. Ref. 23, p. 5.
25 Fahrbach, J.C. Control of S02 Emissions in a Sulfite Manufacturing
Operation. American Can Company. Green Bay, Wisconsin. pp. B2Ub,
B206.
26. Ref. 25, p. B207.
27. Evans, J.C.W. Unique Process for Sulfite Spent Liquor Keeps Older Mill
Viable. Pulp and Paper. September 1975. pp. 63, 64.
28. Ref. 18, pp. 2-9 to 2-13.
5.6-23
-------
-------
5.7 COAL MINING WASTE DISPOSAL
5.7.1 Process Descriptions and Emission Sources
The mechanized extraction of coal from the mine and its preparation at
the plant by current mechanized mining methods generate large quantities of
waste to be disposed of. The principal problems associated with coal waste
disposal are pollution of air and water, nonproductive use of land, and loss
of aesthetic value. Landslides are another potential problem.1
Coal mining waste (including "coal refuse", "culm", and "gob") consists
primarily of a mixture of coal, rock, carbonaceous shales, and pyrites
(FeS 3.1 These rejected materials may amount to 20 percent or more of the
tonnage mined.1 Coal mining waste has no immediate use and is disposed of
as economically and conveniently as possible. The most common practice has
been simple open-end dumping of the mining waste in piles or banks. In
Appalachia, the waste is usually disposed in hillside dumps, valley fills,
or earthen dams. Dumptrucks, conveyor belts, mine cars, and aerial tramways
are used to move the waste material from the mine to the disposal site.1
Disposal sites for coal mining waste often become dumping grounds for
other discarded items such as grease-soaked rags, grease and oil containers,
trash and garbage from nearby homes, wood, and other combustible organic
matter. Depending on the age of a pile and the local mining and preparation
methods, the combustible content can vary from 10 to 60 percent.2
Until recently, coal mining waste piles commonly burned or smoldered
continuously, sometimes for years, releasing noxious gases and particulate
matter into the atmosphere. Fortunately there is growing evidence that the
number of actively burning piles is steadily diminishing, as shown in Table
5.7-1. Most of these piles are located in Virginia, West Virginia,
Pennsylvania, and Kentucky,3 and are believed to be confined largely to
older waste banks.4
TABLE 5.7-1.
ACTIVELY BURNING COAL MINING WASTE PILES
.IN THE UNITED STATES3,5,6
Year
1964
1968
1972
Number
495
292
206
5.7-1
-------
Spontaneous combustion is probably the major cause of coal mining waste
fires, although some are caused by lightening; grass, brush, or forest
fires; trash fires; and intentional ignition to create "red dog" residue for
road base or other foundation purposes.',e Spontaneous ignition is
influenced by the temperature, the coal rank, and the pyrite, moisture, and
oxygen content of the pile. These factors in turn depend on rainfall and
windspeed, and on the particle size distribution, void ratio, and surface
characteristics of the pile. If the waste pile is porous because of
improper layering and compaction, water and air can infiltrate deep within
the mass. Organic and pyritic materials become oxidized and release heat
The trapped heat builds up within the pile, leading eventually to auto-
ignition.
Air pollutants released from burning or smoldering coal mining waste
banks include sulfur oxides (SOX), hydrogen sulfide (H2S), sulfuric acid
(H2S04), nitrogen oxides (NCy, ammonia (NH4), carbon monoxide (CO), hydro-
carbons, and particulates. Sulfur dioxide is formed by thermal decomposi-
tion and oxidation of pyritic material (FeSv) in the refuse:
t "
FeS2 + 302 •» FeS04 + S02.
Sulfuric acid, produced through the reaction
2FeS2 + 2H20 + 702 -> 2FeS04 + 2H2S04
can react with pyritic material to form H2S:
H2S04 + FeS •* FeS04 + H2S.
Hydrogen sulfide is also likely to be formed by a number of other possible
reasons. Elemental sulfur, often observed on the surface of burning
Piles, is formed by the thermal breakdown of the pyrite FeS2, or by the
reaction of hydrogen sulfide and S02:
2H2S + S02 -*• 3S + 2H20.
Estimates of atmospheric S02 emissions from burning coal mining waste
piles have been variable. In 1968 it was estimated that 0.54 Tg (600 000
tons) was emitted annually.? In 1970 the EPA estimate was 0.128 Tg (141,000
tons) of SOX emitted to the atmosphere from 245 Tg (270 million tons) of
5.7-2
-------
burning coal mining wastes.8 More recent estimates place the annual
national S02 emission from such sources at 0.039 Tg (43,000 tons) per year.9
The latter estimate may be the most accurate because the number of actively
burning piles is believed to be decreasing and also because the estimate
assumes that only 21 percent of the available refuse mass is burning.
Air samples taken from communities near burning coal mining waste piles
were reported in 1971 to contain, on the average, more than 1 ppm S02, with
peak concentrations exceeding 4.5 ppm.5 Hydrogen sulfide concentrations
exceeding 0.4 ppm have also been measured.10
5.7.2 Control Techniques
The control of SO and other air emissions from coal mining waste piles
/\
involves prevention and extinguishment of fires.
Proper layering and compaction are the two principal techniques used in
newer coal mining waste banks to prevent fires.11 Because these methods
significantly reduce the permeation Of air and water into the waste mass,
spontaneous fires seldom occur.5 In addition to layering and compaction,
the following preventive measures are important.
0 Proper selection and preparation of the disposal site: The ideal
site has flat terrain. In mountainous regions, valleys and hill
sides are used. The site should be close to an adequate supply of
noncombustible material to be sandwiched between refuse layers and
compacted or sealed around the sides to prevent air infiltration.
Cross-valley fills should be avoided.12 All vegetation is
removed.13,14
0 Optimum refuse bank design: Side dumps are less likely to ignite
than end dumps. Exposed surface area /is minimized. Terracing
reduces fire hazard by checking the draft along a slope. Slopes
of 50 percent or less are recommended.13
0 Removal of combustible organic materials: Dumping of grease rags,
domestic garbage, vegetation, and the like is prohibited. Carbo-
naceous content of the coal waste is reduced, if possible, by
improvement in coal preparation techniques.
0 Increased percentage of fines: A maximum of 15 percent fines
improves compaction and impermeability of the waste mass.
0 Water flow: Ground, surface, and runoff waters are avoided or
diverted to prevent heat generation from infiltration of the waste
mass, and also to prevent water pollution.
5.7-3
-------
°r,dl>t 1S Used to seal exP°sed edges of the pile
p"e "^ ^^ation, which lead to heat buildup
the possibility of
n^!amatl0n:- Ve9etatl'°n is planted to finish the pile both to
prevent erosion and to enhance aesthetic effects.
The following methods have been developed to control burning coal
raining waste piles.16
Blanketing: The top and sides are sealed with fly ash clay
tff
hein
!K thl'S blanket^9 or smothering genera iy ?s
^ts6"^"1^ ^^ 9nd c™k t0 b-°-
Sprayed °ver the entire
blanketed
h
l
This
the
isolate the hot area from the rest
is then quenched with water or is
coaflnt " Slury. »'•.<««*<• and pulverized limestone, fly ash
SSI tShelas°sranT?ni Jo"^ "" "^ 1nt° the ^"^ '"' *
i-: .ExPlosives are placed deep within a burning
in the had c?.uSC1ZOnv1.h0le5- The exP1osi^ creates fissures
in tne hard, crusty, clinker surface of the pile so watpr
penetrate and quench the. fire.
Accelerated combustion and quenching: Burnina refuse is l
ndr° erS (5° to 10° ^ though 'the air nto a
the t TK bu-rns. ""ibustible material and the water
ea is 9enerat1on of
waste,
5.7-4
-------
Further details are available regarding prevention and control of coal
mining waste fires.17,18,19
5.7.2.1 Control Costs--
Information on the cost of controlling S0y emissions from coal mining
/\
waste piles is limited. No cost information is available in the literature
concerning SO emission control through fire preventive design of coal waste
/\
piles. Table 5.7-2 presents cost information relative to the methods used
to extinguish coal mining waste fires. Hauling and repiling extinguished
material by tractor-scraper may be less expensive than using bulldpzers
because of the elimination of a material handling step.2ci Fire extin-
guishing techniques utilizing water appear to be considerably less expensive
than other techniques using grouting, foam blanketing, or explosive mate-
rials.
5.7.2.2 Energy and Environmental Impact—
As noted earlier, methods for extinguishing coal mining waste fires
with water are relatively inexpensive and are frequently used. The appli-
cation of large volumes of water to the burning mass results in runoff,
which often is contaminated with acids, metals, and other pollutants
requiring neutralization and sedimentati.on treatment. Obviously., methods
that do not involve using water, such as blanketing or sealing, do not lead
to water pollution problems. Fire preventive methods that prevent the
infiltration of water into the coal mining waste pile will help prevent pol-
lution of ground and surface runoff waters. Examples of such preventive
methods are layering, compaction, sealing, and revegetation.
5.7-5
-------
TABLE 5.7-2. RELATIVE COSTS OF DEMONSTRATED
METHODS OF EXTINGUISHING COAL MINING WASTE FIRES16,20,21
Method
Cost per cubic meter, $
July 1979
Blanketing/sealing
Polyurethane foam
Clay
Sand
Limestone
Explosives
Grouting
Lime slurry
Lime/limestone slurry followed
by limestone seal
Ponding
Water monitors and lagoons
Spraying
Quenching, bulldozers, and drag
lines
Water sprinkling, dozer digout
with repile in compact layers
Cooling and dilution
Quenching, hauling by tractor-
scraper
Hydraulic jets
Isolation
Trenching followed by quenching
or blanketing
Digging out and spreading
Accelerated combustion and quenching
4.37 (1971 or earlier)
2.86 (1968)
2.85 (prior to 1971)
1.46 (1969)
1.63 (1971 or earlier)
0.92-1.35 (prior to
1971)
1.48 (1971 or earlier)
0.86 (1971 or earlier)
0.60 (1968)
0.57 (1970)
0.98 (1968)
0.86 (1970)
5.7-6
-------
REFERENCES FOR SECTION 5.7
1. Coal gate, J.L. , D.J. Akers, and R.W. Frum. Gob Pile Stabilization,
Reclamation, and Utilization. PB-224-561. May 1973. pp. 5-7.
2. Flegal, R.C., and N.J. Gahr. A Summary of Demonstration Methods for
Extinguishing Culm-Bank Fires. U.S. Environmental Protection Agency,
Office of Air Programs. July 1973. p. 3.
3 Chalekode, P.K., and T.R. Blackwood. Coal Refuse Piles, Abandoned
Mines and Outcrops—State of the Art. EPA-600/2-78-004V. July 1978.
pp. 9, 10.
4. Wahler, W.A. Pollution Control Guidelines for Coal Refuse Piles and
Slurry Ponds. . EPA-600/7-78-222. November 1978. p. 19.
5 McNay L.M. Coal Refuse Fires, An Environmental Hazard. Information
Circular 8515. Department of the Interior, U.S. Bureau of Mines.
1971. pp. 1-26.
6. Ref. 1, p. 9.
7. Ref. 1, p. 13.
8. U.S. Environmental Protection Agency, Office of Air Quality Planning
and Standards. OAQPS Data File. Durham, N.C. July 1972.
9. Ref. 3, p. 15.
10. Sussman, V.H., and J.J. Mulhern. Air Pollution from Coal Refuse
Disposal Areas. Journal of the Air Pollution Control Association.
J4(7):279-284. 1964.
11. Ref. 4, pp. 8, 10.
12, Ref. 4, pp. 12, 54.
13. Ref. 1, pp. 35-43.
14. Ref. 4, pp. 60-61.
15. Ref. 4, p. 13.
5.7-7.
-------
16. Ref. 3, pp. 17-19.
17. Ref. 1, pp. 29-43.
18. Ref. 3, pp. 17-21.
19. Ref. 4, pp. 60-79.
20. Ref. 1, p. 33.
21. Ref. 2, pp. 13-46.
5.7-8
-------
5.8 GLASS MANUFACTURE
5.8.1 Process Descriptions and Emission Sources
5.8.1.1. Introduction—
The glass manufacturing industry is made up of several different seg-
ments, classified by the Standard Industrial Classification (SIC) System as
shown in Table 5.8-1. In early 1978, 129 primary glass producing companies
operated 338 plants, most located east of the Mississippi River.1 In 1976
the industry produced nearly 17 Tg (19 million tons) of glass, most of which
was soda-lime glass.2 The reported value of glass shipments that year was
over 6 billion dollars. Table 5.8-1 gives further details.
On a nationwide basis, the glass industry is estimated to be respon-
sible for 0.1 percent of total annual SO emissions.3 Historically, this
industry has maintained compliance with applicable S0x regulations through
process control techniques such as batch sulfur control and fuel selection;
add-on SO emission control devices have generally not been used.
/\
5.8.1.2 Process Description—
Temperatures ranging from 1500° to 1700°C (2732° to 3092°F) are used to
convert inorganic oxides of silicon (Si02), sodium (Na20), calcium (CaO),
and other elements to liquid mixtures that, after cooling, are homogeneous,
amorphous, multicomponent, and often transparent; such mixtures are known
generically as glass. More than 50 different glass compositions are avail-
able, but soda-lime glass reportedly accounts for 90 to 95 percent of all
the glass produced.2,4 • The finished composition of soda-lime glass is
typically 70 to 74 percent Si02, 10 to 13 percent CaO, and 13 to 16 percent
Na20.5 It is used to make flat glass (such as windows), container glass
(such as food, beverage, and drug containers), and pressed and blown glass
(tableware, tubing, and light bulbs).
Sand, soda ash, and limestone are the major raw materials used in
soda-lime glass manufacture. Cullet (ground and recycled glass) is used as
makeup material and may contribute 10 to 50 percent or more of the finished
product.6,7 Other ingredients are considered minor because they generally
constitute 5 percent or less of the material mixed in a batch. Table 5.8-2
gives further information about the raw materials used in manufacturing
soda-lime glass.8 11
5.8-1
-------
TABLE 5.8-1. GLASS MANUFACTURING INDUSTRY
Industry
segment
Flat glass, including
sheet, plate and
float, laminated,
and tempered auto-
mobile glass
Container glass,
including food,
beverage, and phar-
maceutical glass
Pressed and blown
glass, including table-
ware, television
tubes, light bulbs,
lamp enclosures, tub-
ing, and textile
fiberglass
Wool fiberglass
Total
SIC
code
3211
3221
3229
3296
Number
of plants
in 1978
32
117
165
24
338
Production
in 1976,
Tg (million tons)
2..56 (2.91)
11.80 (13.00)
1.73 (1.95)
0.896 (0.986)
16.976 (18.846)
Value of 1976
shipments,
millions
of dollars
645
3251
1598
817
6311
5.8-2
-------
TABLE 5.8-2. RAW MATERIALS USED IN MANUFACTURING SODA-LIME GLASS
8-11
Ingredient
Purpose
Amount of ingredient
mixed in batch, weight *
Major components
Glass sand
(>99% Si02)
Soda ash
(Na2C03)
Limestone or burnt lime from dolomite
(CaC03> MgC03)
Minor components
Feldspars
(Na20/K20-Al203-6Si02)
or nepheline syenite
(Na20-1.7Al203-4Si02)
Fining agents
Sulfates
(NapSOi, BaSO*, or
[NH4]2§04
Peroxides; nitrates; chlorates;
chloride salts; arsenic, cerium
and manganese oxides; and other
chemicals
Powdered coal
Coloring and decolorizing agents
(many metallic oxides, selenium)
Oxidizing agents
(KN03, NaN03)
Provides SiO,
Provides Na?0, the primary
agent for fluxing or
lowering of the melting
point
Provides CaO and MgO
Provide A1C>3 to lower
melting point and add
stability to the glass by
retarding devitrification;
other ingredients' are glass
forming oxides
Help in reduced sulfite
form to condition glass;
bubbles are removed from
the molten glass by gas
evolution
Help to condition glass
by removing bubbles
through gas evolution or
evaporation
Provides carbon for
reducing sulfates to
sulfite form
Add or remove color
from glass
Oxidize iron contamin-
ants to make them less
visible
60-70
17-20
7-15
0-6
0.5-5
5.8-3
-------
Typically, 540 kg (1200 Ib) of raw material is needed to produce 454 kg
(1000 Ib) of container glass, 90 percent of which is salable.4 The remain-
der is recycled as cullet. About 18 percent of the raw material fed to the
furnace is given off as gas, most of which is carbon dioxide (C02).« For
each kilogram (2.2 Ib) of raw material that is lost, over 2 m3 (70 ft3) of
off-gas is created at 1500°C (2732"F).« Finished glass occupies less than
half of the volume of the unmelted raw materials.12
The chemical reactions involved in soda-lime glass manufacture are as
follows:6
Na2C03 + xSi02 £ Na20-xSi02 + C02t
CaC03 + ySi02 £ CaO-ySi02 + C02t
C + Na2S04 + zSi02 £ Na20-zSi02 + S02t + COt
The last reaction may take place in two steps:
Na2S04 + C -* Na2S03 + COt
Na2S03 + zSi02 £ Na20-zSi02 + S02t
Carbon dioxide may also form according to the following reaction:
2Na2S04 + C •*• 2Na2S03 + C02t
Figure 5.8-1 is a typical flow diagram for the manufacture of soda-lime
glass; steps in the production of other types of glass resemble those for
soda-lime glass. The major and minor ingredients are premixed and stored in
hoppers until charged to the furnace, where temperatures generally ranging
from 1500° to 1700°C (2732° to 3092°F) convert the mixture into a molten
mass. The mass is held until it is of uniform consistency and until re-
fining (the process of removing gas bubbles, such as C02, H20, and S02) is
complete. The temperature is slowly lowered to about 1300°C (2372°F) to
increase the viscosity and to condition the melt for the next processing
steps forming.13
In the forming and finishing processes, the molten glass is extracted
from the furnace, shaped to the desired form, and annealed within a specific
temperature range. The final product is either inspected and shipped or
sent for further finishing, such as tempering or decorating. Rejected
product is crushed and recycled as cullet.
5.8-4
-------
GLASS SAND
SODA ASH
LIMESTONE
OR BURNT LIME
FROM DOLOMITE
SIDE-PORT,
CONTINUOUS, REGEN-
ERATIVE FURNACE
TEMPERATURE:
1500°C (2732°F)
SUBMERGED
THROAT IN
BRIDGE WALL
TEMPERATURE:
1300°C (2372°F)
TEMPERATURE: 8000-1100°C
(1472°-2012°F)
DEPENDING ON ARTICLE
AND PROCESS
FINISHING
OR ROLLING
MINOR COMPONENTS:
FELDSPARS OR NEPHELINE
SYENITE, SULFATES
AND OTHER FINING
AGENTS, POWDERED
COAL, COLORIZING AND
DECOLORIZING AGENTS,
AND OXIDIZING AGENTS
CRUSHED CULLET
OF SAME
COMPOSITION
AS THAT TO
BE MELTED
PACKING, WAREHOUSING,
• AND SHIPPING
CULLET
CRUSHING
, VISCOUS
Y PRESSING,
RAWING,
NG
ING
N AND
ESTING
Figure 5.8-1. Typical flow diagram for the manufacture of soda-lime glass
5.8-5
-------
Both batch and continuous furnaces are operated in the glass industry.
Day pots and day tanks are used if only a few tons or less of a specialty
glass, such as optical glass, art glass, or cast plate glass, is produced.15
Day pots are made of selected clay or platinum and are heated to about
1400°C (2552°F) in multiple-pot furnaces;16 the capacities of day pots range
from 9 kg (20 Ib) to 1800 kg (2 tons).17 Day tanks are refractory lined and
can hold somewhat more than day pots.
Most glass is produced in continuously operating regenerative or re-
cuperative furnaces. The capacities of these furnaces range from 0.9 to
1350 Mg (1 to 1500 tons).15,16 A regenerative furnace generally consists of
two chambers of refractory, called checkerwork. While combustion flue gases
heat the refractory in one checkerwork chamber, the other checkerwork cham-
ber preheats combustion air. Every 10 to 30 minutes the gas flow is re-
versed, combustion air is drawn through the chamber previously heated by
flue gases, and flue gases heat the refractory in the other chamber. The
firing pattern of a regenerative furnace can be end-port (Figure 5.8-2) or
side-port (Figure 5.8-3).18 These furnaces are operated continuously for 4
to 5 years at sustained temperatures up to 1600°C (2900°F).19 When operated
at lower temperatures, additional furnace' life can be expected. A recupera-
tive furnace consists of one continuously operating shell with tubular heat
exchangers instead of checkerwork heat exchangers to preheat combustion
air.20 Furnace combustion heat is recovered
and transferred to cooler
incoming air by passing the incoming air through pipes surrounded by the
outgoing combustion gases.
S0x emission rates for the production of
5.8.1.3 Emission Sources—
Table 5.8-3 presents annual
container glass, pressed and blown glass, and flat glass.21 Rates estimated
in the 1976 source assessment documents are less than the rates obtained by
site testing from June 1976 to September 1978 for the production of con-
tainer glass and flat glass. Site testing of container glass furnaces
primarily firing fuel oil showed that in most cases the emission rate for
SOX was higher than for any other criteria pollutant in the production of
container glass. Table 5.8-4 summarizes emission data from EPA site test-
ing,22 Table 5.8-5 presents S0x emission data provided by industry,23 and
5.8-6
-------
MOVABLE BAFFLE
INDUCED-DRAFT FAN\
PARTING WALL
SECONDARY CHECKERS
REFINER SIDE WALL
GLASS SURFACE IN REFINER
FOREHEARTH
THROAT
GLASS SURFACE IN MELTER-
COMBUSTION AIR BLOWER
MELTER SIDE WALL
DER
PRIMARY CHECKERS
CURTAIN WALL
RIDER ARCHES
1 o
.Figure 5.8-2. End-port continuous regenerative furnace.
5.8-7
-------
REFINER SIDE WALLx
MELTER SIDE WALLv THROATN
GLASS SURFACE IN MELTER-\ \ MELTER BOTTOM\
NATURAL-DRAFT
STACK
GLASS SURFACE IN REFINER
.FOREHEARTH
RIDER ARCHES
MOVABLE REFRACTORY BAFFLE
BURNER—1
-COMBUSTION AIR BLOWER
Figure 5.8-3. Side-port continuous regenerative furnace.
5.8-8
-------
TABLE 5.8-3. ESTIMATES OF ANNUAL SOX EMISSIONS FROM GLASS MANUFACTURE
[Mg/yr (tons/yr)] 21
Source of data
EPA source assessment
documents, 1976
EPA site testing,
1976 to 1978
Container glass3
72-111 (80-123)
200-342 (220-380)
Pressed and,
blown glass
51 (56)
0-31 (0-34)
Flat
glass
349 (380)
597 (630)
EPA Source Assessment Document data derived from 46 flint container
glass furnaces that were probably all gas-fired. Site testing data
derived from 8 container glass plants, 6 of which burned oil, and 2 gas.
EPA Source Assessment Document data derived from 5 pressed and blown
furnaces. Site testing data from 2 pressed and blown borosilicate,
1 pressed and blown lead, and 1 pressed and blown soda-lime furnace.
c EPA Source Assessment Document Data derived from 5 flat glass furnaces.
Site testing data from 1 soda-lime flat glass furnace.
5.8-9
-------
TABLE 5.8-4. EMISSION DATA FROM SITE TESTINQ22
Type of glass
produced
Container glass-
soda-lime
Container glass-
soda-lime
Container glass-
soda-lime
Container glass-
soda-lime
Container glass-
soda-lime
Container glass-
soda-lime
Pressed and blown-
borosilicate
Container glass-
soda-lime
Container glass-
soda-lime
Pressed and blown-
soda-lime
Pressed and blown-
borosilicate
Pressed and blown-
lead
Flat glass-
soda-lime
Process
rate,
kg/h
(Ib/h)
5,470
(12,058)
8,391
(18,499)
6,227
(13,728)
5,162
(11,380)
5,384
(11,871)
7,274
(16,039)
950
(2,094)
9,873
(21,767)
11,204
(24,701)
1,657
(3,653)
931
(2,053)
811
(1,789)
15,876
(35,000)
Fuel
(sulfur
content unknown)
No. 6 oil
No. 6 oil
No. 6 oil
No. 6 oil
No. 6 oil
No. 2 oil
No. 2 oil
Gas
Gas
Gas
Gas
Gas
Gas
so
kg?h
(Ib/h)
16.21
(35.7)
24.3
(53.6)
23.8
(52.3)
53.35
(117.6)
54.10
(119.3)
11.3
(25.0)
2.2
(4.8)
25.21
(55.57)
39.22
(86.46)
0.46
(1.013)
<0. 1
(<0. 1)
0.10
(0.21)
68.20
(150.3)
Kg SO/Mg
glafs
produced
(Ib/ton)
2.45
(5.95)
2.38
(5.79)
3.14
(7.62)
8.51
(20.67)
8.28
(20.12)
1.29
(3.12)
1.88
(4.57)
2.09
(5.11)
2.88
(7.00)
0.23
(0.55)
<0.04
0.09
(0.23)
3.54
(8.58)
5.8-10
-------
TABLE 5.8-5. SUMMARY OF SOX EMISSION DATA SUPPLIED
BY GLASS MANUFACTURERS23
Type of glass
Container
Amber
Flint
Soda-1 ime
Pressed and blown
Soda- lime
Soda- lime
Borosilicate
Fuel
Gas and boost
No. 2 oil and
boost
No. 2 oil
SO emitted,
kg/hr (Ib/hr)
a
a
10.48 (23.1)
52.94 (116.7)
5.49 (12.1)
S02 emitted.
kp/Mg (lb/ton)
5.1 (10.2)b
0.6 (1.2)b
a
a
a
Value cannot be calculated from available data.
Values are derived by difference, based upon batch input and retention of
sulfur. They assume a fuel with no sulfur content. The total S02 emis-
sion will be a function of fuel sulfur content.
5.8-11
-------
Table 5.8-6 gives emission factors for glass manufacturing procedures.24
It should be noted that emissions from container glass furnaces and
their pattern of fuel usage are disproportionately represented in the site
testing data in Table 5.8-4. Six out of eight (75 percent) of the container
glass furnaces tested burned fuel oil, but fuel usage statistics for the
container glass segment of the industry indicate that only one out of five
burns fuel oil.25
The major source of S0x emissions in the glass industry is the glass
melting operation. Forming and annealing operations are minor sources.
Furnace emissions appear to be attributable to both the manufacturing
process and the fuel burned, the latter being the predominant source (at
least in the pressed and blown segment of the industry).26 Fuel-derived SO
emissions are lower from natural-gas-fired furnaces than from oil-fired
furnaces, unless the oil has been desulfurized.27 Flue gases from furnaces
burning natural gas have been reported to contain 2 ppm SO or less.23
Roughly 600 ppm S0x can be expected in flue gas from a furnace burning fuel
oil containing 1 percent sulfur.28 The S0x emissions from glass manufac-
turing would be expected to increase with increased use of fuel oil.
Greater use of electric furnaces or electric boosting, however, may decrease
SO emissions.
/\
Process-derived S0x emissions come from sulfur compounds such as sodium
sulfate (salt cake) used to condition glass, as in the manufacture of soda-
lime glass and wool fiberglass. The greater the sulfur content of the raw
batch, the higher the SO emissions.27 About 40 percent of the sulfur added
/\
as sulfate is vaporized and exhausted as gaseous SO or condensed sodium
sulfate.27
The vaporization of sulfur compounds involves several reactions.
Sulfur dioxide is released when sodium sulfate chemically reacts with the
melt.29
Na2S04 + xSi02 + C •* Na20-xSi02 + COt + S02t
At the same time, sulfur trioxide may be produced from the thermal decompo-
sition of sodium sulfate.
Na2S0
Na20 + S0t
5.8-12
-------
TABLE 5.8-6. EMISSION FACTORS FOR GLASS MANUFACTURING PROCEDURES3'5'0
Raw material handling
Melting furnace
Container glass
Uncontrolled • ,
With low-energy scrubber
With venturi scrubber"
With fabric filter
With electrostatic
precipitator
Pressed and blown glass
Uncontrolled _
With low-energy scrubber
With venturi scrubber
With fabric filter
With electrostatic
precipitator
Flat glass
Uncontrolled f
With low-energy scrubber
With venturi scrubber*1
With fabric filter
With electrostatic
precipitator
S0x
Kg/Mg
0
1.7(1.0-2.4)
0.9
0.1
1.7
1.7
2.8(0.5-5.4)
1.3
0.1
2.8
2.8
1.5(1.1-1.9)
0.8
0.1
1.5
1.5
Ib/ton
0
3.4(2.0-4.8)
1.7
0.2
3.4
3.4
5.6(1.1-10.9)
2.7
0.3
5.6
5.6
3(2.2-3.8)
1.5
0.2
3.0
3.0
. Source: Reference 24 •
Emission factors are expressed as grams of SO per kilogram of glass
produced and as pounds of SO per ton of glassxproduced.
When literature references report ranges in emission rates, these ranges
. are shown in parentheses along with the average emission factor. Single
emission factors are averages of literature data for which no ranges are
reported.
Emission factors for raw materials handling are not separated into types of
glass produced because batch preparation is the same for all types.
Particulate emissions are negligible because almost all'plants utilize
some form of control (i.e., fabric filters, scrubber, or centrifugal
collectors).
Control efficiencies for the various devices are applied only to the
. average emission factor.
Approximately 52 percent efficient in reducing particulate and sulfur oxide
emissions.
** Approximately 95 percent efficient in reducing particulate and sulfur oxide
emissions.
5.8-13
-------
In the vapor state over the melt, sulfur dioxide may be oxidized to form
additional sulfur tri oxide.
S0
1/20
S0t
Vapors over the melt may recombine to form sodium sulfate.
Na20
2NaOH
H0
1000°C
NaOH
S03 1000°C> Na2S04t
As the temperature in the exhaust stack falls to about 200°C (400°F), the
rising vapors condense as fine, usually submicrometer particulates. Sodium
sulfate has been found to be the major component of parti cul ate emissions
from soda-lime glass manufacture; 75 percent of these parti cul ate emissions
are less than 1 micrometer in diameter.29
5.8.2 Control Techniques
5.8.2.1 Description —
Essentially all of the SOX emitted during glass manufacture is gener-
ated in the melting process. Glass furnace emissions can be reduced by
three means: process modifications, fuel changes, and add-on control equip-
ment.
Process modifications that may reduce SO emissions include altering
/\
the raw material charge to reduce the sulfur content or to increase the
fraction of recycled glass, changing the furnace controls or equipment, and
altering the pull rate. Process modifications that reduce the salt cake
content in the raw batch can significantly reduce SO emissions. For
example, one California flat-glass plant reportedly reduced furnace emis-
sions of S02 by 78 percent from 2.1 to 0.5 kg/Mg (5.0 to 1.1 Ib/ton) by
reducing the salt cake in the raw batch 60 percent (from 12 to 5 kg/Mg, 30
to 12 Ib/ton of sand).30 Similarly, another California flat-glass plant has
reportedly reduced its S02 emissions 75 percent (from 1.6 to less than 0.4
kg/Mg, 4 to less than 1 Ib/ton of batch constituents) by reducing the input
of salt cake. Glass quality was not compromised in either case.31 The salt
cake cannot be reduced below certain minimums without effecting glass
quality. The minimum salt cake required varies depending upon furnace type,
pull rate, glass type, and other variables.
5.8-14
-------
Fuel changes have also been shown to reduce SO emissions. These
include switching to natural gas or low-sulfur fuel oil, switching to all-
electric melting, and using electric boosting for melting. Electric melters
significantly reduce SO , NO , and particulate emissions because they elimi-
x\ /\
nate the combustion of fossil fuels. Electric melting also is reported to
minimize SO and other gaseous losses from the vaporization of raw materials
because the surface of the melt is insulated by a semisolid crust. Gases
discharged through the crust of the melt consist mainly of carbon dioxide
and water.33 Today, borosilicate, opal, and green,glass are produced with
electric furnaces.34 The capacities of such furnaces are about 100 to 110
Mg/day (110 to 120 tons/day). Electric melters have not been demonstrated
for larger operations, such as large container furnaces, the nominal
capacities of which are about 220 Mg/day (240 tons/day), and flat-glass
furnaces, which range from about 600 to 800 Mg/day (660 to 880 tons/day).35
Several emission control systems that are available to the glass indus-
try for particulate control are also capable of achieving various levels of
secondary SO control. For example, a venturi scrubber system of the type
/\
shown in Figure 5.8-4 can control SO emissions from commercial glass
X.
plants.36 As shown, the system includes a packed tower where part of the
sulfate particulates are removed from the hot furnace flue gases, a dual-t
hroat venturi scrubber where S02 and additional particulates are removed by
alkaline washing, and a cyclone for final particulate collection. A pilot
system of this type at a glass furnace in early 1973 achieved S02 control
efficiencies as high as 90 percent.37 Similar full-scale systems have
reduced S02 emissions 75 to 90 percent.38 Specific parameters for these
systems are given in Table 5.8-7. Currently, only the container glass
segment of the glass industry is reported to use scrubber systems for emis-
sion control.39
Electrostatic precipitators (ESP's) and fabric filters appear to be of
limited use for SO control. One company reports an S02 emission reduction
/s
of 15 percent and S03 emission reduction of 40 percent with an ESP, but such
effects have not been noted in other ESP tests.40
Injecting a sorbent such as alumina, limestone, or nepheline syenite
into a fabric filter system can effectively remove SO from furnace flue
5.8-15
-------
t— o
to «/) «C
*z zsS:
x CD in
UJ QC =>
0"-^
uj 1/5 x
m t_> o uj
10
oo
(1)
+J
to
>>
to
s_
S-
O
I/)
s_
c
0)
fO
,o
CL
I
OO
LO
O)
5.8-16
-------
O
UU
C3
a
UJ CO
I— S
-
•z. co
UJ UJ
o co
CO
i co
co
. i— i
If) OS.
m
** N*s,
re w *-
t- \ re
trt O)
i- i- N_,
OJ
1
u_
c
c
*0 E £
t- «l O.
£-sz
0 >>U
"•"X
PO
m
CO
06
en"
CNJ
O
o
0
m
3
>*-
V) O
m 01
C3 <*-
,
1
•o
t.
.C 0)
•;|
E U
S"
tf) «^-
4-»
*re >
u
3 -C
u. •*->
V)
V)
re
'01
j-
i
^
u>
S
U3
:-'
-•
^
in
t^
i
o
m
esj
v>
re
01
2
3
5
.c
'*
g
1
CU
U
3
u_
(/]
re
"CT
1
re
1
o>
"re
i_
>
•o
c
re
t.
1
u
s
s-
cu
ja
|
V)
u
3
1
re
I
4-*
c
o
3
O
U)
I
I
1
o
0
4->
•o
cu
re
3
O
recin
§
in
s
m
en m
A V
O NW
5*
m
in
CO
o
C3
cri
^
0
re
£
-C t
?!
ll
vi re
're •*-*
u
1 CU
si
u. o.
(A
re
*o>
i
1
o
t-
Ol
J3
3
l_
(A
S-
c
cu
>
0
m
en
CNJ
2
O
in
CM
CM
O
ift "O tn
>» c: CO)
« re •*- o c
*J t- 3 4-> •—
o cu •*-* 3 re
^ o cu'o^-
o. +* > v> re
t.
|
re
o
U tft
tn
t. re
i_
3
(J
C
5.8-17
-------
gases.41,42 The spent sorbent may be landfilled or possibly recycled. In
the fiberglass industry, solids produced by a sorbent system using hydrated
lime (Tesisorb X) are being recycled.42 In the container glass industry
solids produced by a similar dry sorbent system have not been successfully
recycled, but have been disposed of as landfill.
One patented system of dry removal (Figure 5.8-5) involves the combined
use of hydrated lime and nepheline syenite for acid gas neutralization and
fine particle agglomeration.43 In this system hot furnace flue gas is first
mixed with quench water, hydrated lime (Tesisorb X) for primary S02 removal,
and secondary air to cool the gas stream to a temperature range of 94° to
427°C (200° to 800°F). Next, nepheline syenite (Tesisorb A) is added to the
gas stream to capture residual S02 and submicrometer particulates. The gas
stream enters the fabric filter where the solid product is removed for
either recycling to the furnace or landfilling. Shake cycles reportedly
vary from 24 to 36 hours, and pressure drops range from 1.5 to 3.0 kPa (6 to
12 in. H20).42
Table 5.8-8 summarizes
several commercial glass furnaces.42'44 .__
f\
reduced 80 to 95 percent at a container glass furnace, 50 to 90 percent at a
fiberglass furnace, and 88 to 98 percent at a flat-glass furnace.
Mist eliminators apparently have no effect on SO gases. One sampling
/\ '
test indicated no decrease in S02 and S03 concentrations through the control
device.45
A nucleator or double-alkali system (Figure 5.8-6) is a wet method for
recovering submicrometer particulates and SO with relatively little energy
s\
input; the pressure drop is less than 3.7 kPa (15 in. H20).43 Formation of
a steam plume is avoided by use of a water recycle system with a cooling
tower. As Figure 5.8-6 shows, recyclable solid product is produced. Such a
system is reported in use on a fiberglass furnace and on a fluoride-opal
furnace; SO removal efficiencies of 95 to 99 percent -have been documen-
ted.44
Table 5.8-9 presents operating parameters for model glass plants and
emission control systems, including production rate, stack height, stack
diameter, stack gas exit velocity, and stack gas temperature.46 Calculated
SO emission data for dry sorbent systems at
s\
Total SO concentrations were
5.8-18
-------
FURNACE
FLUE GAS
QUENCH
WATER TESISORB X
1
MIXING AND
COOLING
SECONDARY
AIR
FABRIC
FILTER
TESISORB A
SOLID PRODUCT
FOR RECYCLE
TO FURNACE
CLEAN
EXHAUST
INDUCED-DRAFT
FAN
Figure 5.8-:5. Dry sorbent system 43
5.8-19
-------
CD
t—<
GO
to
LU
to
to
to
00
LU
CO
I
CO
•
to
CO
> u
O C
wt
n_
E
U
oi .c
01 re -
o>.e ai
C Ul r-
re u
cei- >,
0 U
Ul
^•x
****)
ui re re u
>i Q.T3 Ul
co re c v-/
u re
4->
Ul
i- +>
o c TT
ai at
01 J3 Ul
Q. t- 3
^°
^3
*»- :
? -s.
0 0
Ul Ul
•r- Ul •!-
Ol •— O)
t— Q.I—
01
C
re
c
o
CJ
o
01
o
in
in
CM
o
0
CO
o
in
C>J
i
CO
o
CO
o
CM
CM
<**
i
in
CO
"
o
a.
a>
t.
o
•a
ai
o
a.
ai
i %
o
u
o
Ol O
• o
O CM
1 1
r- o
• o
o in
"~"
<:
•o
o
(/I
Ul
01
1—
4J
re
o
u_
CO
en
o
CO
1
in
in
in
CM
ro
01
t.
o
a.
£
o
z
•o
ai
o
a.
ai
. *
o
z
u
CD
Ol O
• O
O CM
1 1
l»- 0
• o
o in
^
«* X
•e -e
o o
Ul Ul
•1— Ul
-------
FURNACE
FLUE GAS
QUENCH
TURBULENCE
PROMOTER
NUCLEATOR
RECYCLE CIRCUIT
FOR DOUBLE-ALKALI
SOLUTION
, CLEAN
EXHAUST
INDUCED-DRAFT
FAN
tt
-i COOLING
jf_ TOWER
CALCIUM
"OXIDE
REACTOR
PUMP
SEPARATOR
SOLIDS FOR POTENTIAL
REUSE IN FURNACE
Figure 5.8-6. Nucleator.^3
5.8-21
-------
vo
^
O>
u.
o.
t.
At
L
•S
re
o.
u
t.
Q.
t.
Q)
(-
U
t-
£
TJ
Q*
O
e
o
u
TJ
V
£
TJ
CU
O
4-1
O
u
c
QJ
£
1
U
TJ
V
O
I-
4-1
C
o
(J
1
£
o
u
TJ
QJ
£
4J
§
s
u
re
c
**-
«*
£
<*_
u
u
re
£
*•
Q*
U
i
3
H-
oi
u
re
t-
«*-
SJ
U
ra
c
t-
3
*
t-
3
i
CM
CO
rH
CO
rH
CD
CD
CM
VO
rH
CO
rH
O
8
vo
rH
CO
rH
O
in
CM
**«.
CM
CM
CD
Lfl
CM
CM
CM
O
tn
CM
^
CD
U1
CM
r-.
CM
CM
re
TJ
tA
3
>»
to
TJ
•
E
§
u
1
0*
CO
3
in
CO
cn
in
CM
o
cn
^.
CM
CD
ID
CO
3
O
ID
CO
CO
rH
in
CO
cn
^
o
CO
rH
CM
J?
C
*r
if
0)
re
4J
CO
CM
cn
CO
rH
UD
in
rH
r*.
rH
CM
^•^
cn
CO
rH
m
^
cn
CM
CM
tn
•a-
rH
CO
tD*
r-.
CD
CM
^
in
NM"
cn
CM
CM
<»-
«
3
1
TJ
1
CO
0*
rH
cn
CD
CO
s
cn
0
^>
rH
rH
cn
CD
CO
rH
cn
o
CO
s
cn
^
CD
CO
rH
cn
/-S
(A
^s.
«*-
^^^
VI
^
|
"g
*>
2
4->
I
tn
CO
in
CM
^-^
o
o
tD
in
o
rH
CM
o"
S
*~^
CO
CO
in
CM
§
cn
rH
VD
£
O
O
CM
ID
O
CO
rH
CM
O
g
in
(D
CO
o
4-*
CD
0
£
•o
1-
rH
Cn
£
•o
c
o
+•>
o
in
!
in
*r
Electrostatic
u
4->
re
4-»
Electro*
1
u.
u
4J
3
Electros
u
1
Ll_
precipltator-
t-
o
4->
»O
'a.
u
£
0.
o>
4->
«£
1
t-
o
-+J
re
a.
u
O)
i_
a.
1
«*-
controlled
Uncontrolled
•o
CU
£
4-»
C
0
u
•g
control
I
£
4->
i
TJ
CJ
£
4-*
C
O
u
controlled
Uncontrolled
furnace
furnace
o>
u
c
(-
3
(«_
0)
re
c
t-
3
i»_
O)
U
i
3
«*-
OJ
U
3
H-
furnace
furnace
o
s
l£>
in
ro
ID
o
g
V--
ID
in
m
"
<*-
*— '
E
g
D)
I
J£
5
V)
^-
a
Cft
ro
«a-
^*%
^-
s
en
m
^
r^.
rH
^N
tD
r»«.
^*
in
^l-
rH
s
\_^
in
«a-
rH
*O
r*.
s
rH
ID
cn
CO
CNJ
O
rH
*— \
t£>
CO
ro
CO
o
rH
*-^
4J
l*~
C
fc.
Q>
4->
•5
^
«
0
rn
*r
rH
cr>
§
S
o!
o
m
s
en
*-^
0
•»—•
^r
rH
cn
O
ro
S
cn
O
m
s
cn
o
ro
s
cn
o
m
s
cn
&
^j
-x.
M
1
*k**
M
.&
U
O
V
>
>*•>
8
^
2
»
i
8
CO
s
*— .
o
8
c»f
cn
CNJ
CO
eg
si
rH
I
f*>
ro
in
in
CO
rH
O
O
O
m
s
in
»-i
o
o
o
CM
s— '
O
rH
S
O
o
r»»
rH
>S
cn
«^«s
O
O
o
S
in
in
r^
S
m
m
r^
u
re
Vt
i
3
2
8
C
^
3
tn
m
0
*»•
§
m
o
«*
i
I
o
cn
m
m
*r
***
in
<=
o
§
1
o
cn
m
m
o
^~
o
CO
^~
in
o
^~
\— '
o
CO
**•
:k temperature, °K (°F)
5
V)
5.8-22
-------
exhaust gas velocities are based on an assumed 10.5 percent oxygen concen-
tration in the furnace exhaust, which is equivalent to 100 percent excess
oxygen. Stack diameters are calculated to maintain a stack gas exit velo-
city of 9 m/s (30 ft/s).
5.8.2.2 Control Cost-
Information on the costs of SO pollution: control in the glass industry
is limited. Figure 5.8-7 presents reported and estimated capital costs for
the installation of scrubber control systems on glass furnaces.47 The
reported costs are for scrubbers designed primarily for control of particu-
late matter. Values are in January 1978 dollars. Although the data do not
permit a reliable comparison between the estimated and reported costs, the
location of the points suggests relatively good agreement.
Table 5.8-10 summarizes other reported cost data and performance param-
eters for control systems applied to glass manufacturing furnaces.48
Although these systems are designed primarily for particulate control, they
also achieve various levels of secondary SO control, as noted in the table.
/\
All costs have been adjusted to July 1979 dollars by using the Chemical
Engineering Cost Index.
Data on the cost of electric melting appear contradictory. One report
indicates that electric melting is several times more costly than conven-
tional pollution control devices for reducing air emissions.48 Another
source presents investment and operating costs for various melting process
and pollution control systems and suggests that the overall cost of electric
melting compares well with the cost of firing natural gas or oil.49
5.8.2.3 Energy and Environmental Impact—
In 1971 a total of 264 PJ (250 trillion Btu) was consumed by the glass
industry.50 This energy came from the following' sources: 1.6 percent from
coal, 4.5 percent from fuel oil, 5.2 percent from electricity, and 88.7
percent from natural gas.51 Coal is not a furnace fuel, but can be used as
a batch ingredient or as a plant boiler fuel. , Data for 1976 show a trend
toward increased use of fuel oil and electricity. Fuel oil provided 14
percent of the energy consumed by the glass industry in 1976, electricity
provided 11 percent, and natural gas supplied 74 percent.52 Although,
5.8-23
-------
D
D
in
uj uj i— a
=3 rs 2: uj
_J _l CD UJ h-
> i—i i— o >-
Q cl Cu Cd
Q Q ^> UJ UJ I—
LU uj i c; o; c>o
^ I— o i— :s
c/o «=c 2: tr> a
•-1 s: t-i a: LU z
—J I—i L'J —i i—i
£9 I— I— I— —I
^ 00 O -
Q- UJ ^ 3 >• CQ
D
o
o
o
«t
o
o
o
o
o
o
o
CD
o
8
o
o
o
ID
o
C3
CD
CM
ID
o
in
o
LO
O
o
o
o
O)
3
(J
O
CO
CO
•o
a;
i— O
c
S-
-a co
co
T3 >,
a: c
o
o
i
oo
CD
s_
CD
X
LO
x
CM
P 8Z61
«
1SOO IVlIcfVD HZ
5.8-24
-------
TABLE 5.8-10. PERFORMANCE AND COST DATA FOR EMISSION CONTROL SYSTEMS
AT A GLASS FURNACE PRODUCING 1.57 kg/s (150 tons/day)^0
SOX removal , >>
Opacity, %
Sol Ids production,
kg/h (Ib/h)
Installed capacity, watt
(horsepower)
InstalleduCapUal
cost, $ b
Cost of emission control
$/Mg ($/ton)b
Utilities'1
Amortization6
Maintenance
Chemicals"
Additional h
operating labor
Solids reuse or
disposal i
Total additional
cost over fossil
fuel
Credit for recover.
ble chemicalsJ
Net cost
Dry forbent
tysteiis
25-95 (all sorbents)
90-96 (Tesisorb A+X)
0
59-60 (131-132)
s 44,742
520,000°
0.20 (0.18)
1.06 (0.96)
0.37 (0.34)
0.20 (0.18)
0.20 (0.18)
0.08 (0.07)
2.00-2.11
(1.81-1.91)
i-
0.09-0.22
(0.08-0.20)
1.89-1.91
(1.71-1.73)
Venturi
scrubber
system
90-98a
<10
38 (84)
186,425
(250)
715,000
0.81 (0.74)
1.46 (1.33)
0.80 (0.73)
0.33 (0.30)
0.40 (0.36)
0.04 (0.04)
3.84 (3.50)
0.22 (0.20)
3.62 (3.30)
Nucleator
95-99
<10
37 (82)
119,312
(160)
715,000
0.62 (0.56)
1.46 (1.33)
0.89 (0.81)
0.15 (0.14)
0.40 (0.36)
0.04 (0.04)
3.56 (3.24)
0.22 (0.20)
3.34 (3.04)
This value may De mgn; bu ernciencies reponea eisewnere 101 »cni.ui i
scrubber systems range froffi 75 to 90 percent.
b All costs adjusted to July 1979 based on 237 as Chemical Engineering Cost
Index value; 100 is the value for 1957-59.
c In multiple furnace installations.
d Utilities: fossil fuel, Jl.30/109 J ($1.37/106 Btu)
electricity, $0.03/kwh;
steam, $1.72/Mg ($0.78/1000 Ib).
e Amortization: 10-year straight line.
f Maintenance: dry sorbent, 3%
wet systems, 6%.
9 Chemicals: NaOH, $143/Mg ($130/ton);
Ca(OH)2, $36/Mg ($33/ton);
Tesisorb A, $72/Mg ($65/ton);
Tesisorb X, $40/Mg ($36/ton).
h Labor: $19,500/man-year
1 Solids reuse or disposal: $8/Mg ($7/ton).
3 Recovery value: $29/Mg ($26/ton).
5.8-25
-------
theoretically, coal could be used as an energy source, its use appears
impractical because of problems associated with poor glass quality, refrac-
tory damage, and diminished furnace life.53
Natural gas has continued to be the preferred fuel because it, burns
clean, does not affect glass characteristics, and allows for longer furnace
life than other fuels. Nevertheless, because of rising prices and alloca-
tion problems, some natural-gas-fired units have been replaced by oil-fired
ones. Consumption of electricity is also expected to continue to increase,
especially in the flat glass segment.54 Preliminary studies on preheating
the melt indicate that a 50 percent process energy savings is possible, as
well as reductions in pollutant emissions.55
The energy required to manufacture glass varies considerably according
to the glass type. Glass manufacturers indicate that fossil-fuel-fired
melting operations require 4.2 to 7.4 GJ (4 to 7 million Btu) to produce
0.907 Mg (1 ton) of container glass, 6.3 to 12.7 GJ (6 to 12 million Btu) to
produce 0.907 Mg (1 ton) of flat glass, and 6.3 to 42 GJ (6 to 40 million
Btu) to produce 0.907 Mg (1 ton) of pressed and blown glass.56
Estimates of energy costs for air pollution control range from 0.1 to
15 percent of the energy consumed by the industry.57'58 These energy costs
are indirectly related to S0x control because S0x emissions are only con-
trolled by the industry to the extent that S0x is removed with particulates
or to the extent that low-sulfur fuels are used.
The environmental effects of a dry sorbent system are minimal because
it creates no wastewater discharge and because the solids produced may be
recycled or safely landfilled.59 A venturi scrubber system typically incor-
porates a recycle system with an alkaline injection to control pH and a
bleed stream of about 0.13 liter/s (2 gal/min) on a furnace producing 2.1
kg/s (200 tons/day) of soda-lime glasses.60 The concentration of sodium
sulfate in this wastewater stream is 3 to 4 percent. Although discharges of
this nature from individual venturi systems may have significant local
environmental impact and require treatment, the total impact on water pollu-
tion from the glass industry is considered negligible because of the limited
number of glass facilities that are expected to utilize this system.61
5.8-26
-------
REFERENCES FOR SECTION 5.8
1. Spinosa, E.D., D.T. Hooie, and R.B. Bennett. Summary Report on Emis-
sions from the Glass Manufacturing Industry. EPA-600/2-79-101. April
1979. pp. 9, 12, and 17.
2. U.S. Environmental Protection Agency. Glass Manufacturing Plants.
Draft Environmental Impact Statement. Background Information:
Proposed Standards of Performance. EPA-450/3-79-005A. June 1979. pp.
3-1 through 3-3.
3. U.S. Environmental Protection Agency. OAQPS Data File. Durham, N.C.
September 12, 1979.
4. Arthur D. Little, Inc. Environmental Considerations of Selected Energy
Conserving Manufacturing Process Options. Vol. 11: Glass Industry
Report. EPA-600/7-76-034k. December 1976. p. 17.
5. Shreve, R.N., and J.A. Brink, Jr. Chemical Process Industries. 4th ed.
New York, McGraw-Hill Book Co. 1977. pp. 181,182.
6. Ref. 5, p. 184.
7. Kirk-Othmer Encyclopedia of Chemical Technology. 2nd ed. Vol. 10. New
York, John Wiley and Sons, Inc. 1969. p. 551.
8. Ref. 5, pp. 180-184.
9. Ref. 2, p. 3-15.
10. Ref. 7, pp. 550-557.
11. Ref. 4, p. 18.
12. Ref. 1, p. 19.
13. Ref. 2, p. 3-6.
14. Ref. 7, p. 549.
15. Ref. 5, p. 185.
16. Ref. 7, p. 552.
17. Ref. 2, pp. 3-7, 3-8.
5.8-27
-------
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
Compilation of Air Pollutant Emission Factors. 2nd ed
December 1977. p. 8.13-3.
AP-42.
Ketels, P. A., J.D. Nesbitt, and R.D. Oberle. Survey of Emissions
Control and Combustion Equipment Data in Industrial Process Heating
EPA-600/7-76-022. October 1976. p. 39.
Ref. 2, p. 3-8.
Ref. 1, p. 31.
Ref. 1, p. 28.
Ref. 1, p. 30.
Ref. 18, pp. 8.13-4, 8.13-5.
U.S. Department of Commerce, Industry and Trade Administration. Annual
Survey of Manufacturers. 1976. pp. 20-21.
Schorr, J.R. , et al. Source Assessment: Pressed and Blown Glass
Manufacturing Plants. EPA-600/2-77-005. January 1977. pp. 42-43.
Ref. 19, p. 47.
Reed, R.J. Combustion Pollution in the Glass Industry. The Glass
Industry. 54(4): 24-26, 36. April 1973.
Ref. 2, p. 3-17.
Ref. 2, p. 3-16.
Letter from Friesen, R.A. , Chief, Industrial Project Evaluation and
Control Strategy Development Branch, California Air Resources Board to
£™,n™' D-R-' Dl"rector Emissions Standards and Engineering Division,
EPA/RTP. August 8, 1979. p. 2. Docket No. OAQPS-79-2-IV-D-23.
Letter from Welebir, D.S. , Air Pollution Program Director, San Joaquin
^CaLx alth District> Stockton, California, to Central Docket Section
70 !P^, nWa.Shingt°n' D"C- August ]> 1979' PP- L 2- Docket No.
— /y— c.~ IV— D— 6.
Ref. 2, p. 4-6.
Ref. 19, p. 49.
Ref. 1, p. 5.
FMC Corporation, Glen Ellyn, Illinois 60137. Environmental Equipment
Division. Glass Furnace Emission Control System. 1973. p. 8.
37. Ref. 36, p. 7.
5.8-28
-------
38. Ref. 2, pp. 4-16 to 4-19.
39. Ref. 2, p. 4-30.
40. Ref. 2, p. 4-31.
41. U.S. Environmental Protection Agency. Capsule Report: Control of
Acidic Air Pollutants by Coated Baghouses. EPA 625/2-79-020. January
1979.
42. Teller Environmental Systems, Inc. Promotional Material. Tesi Dry
Systems. A: Data Summary, Commercial Installations, Tesisorb Addi-
tives. March 16, 1975.
43. Teller, A.J. Control of Glass Furnace Emissions. The Glass Industry.
57(2):18. February 1976.
44. Ref. 43, pp. 19, 22.
45. Ref. 2, p. 4-28.
46. Ref. 2, pp. 6-9 to 6-13.
47. Ref. 2, p. 8-61.
48. Ref. 43, p. 22.
49. Ref. 4, p. 64. .
50. Ref. 19, p. 44.
51. Ref. 19, p. 53.
52. Ref. 2, p. 3-7.
53. Hanks, G.F. A Trial on 100 Percent Coal Firing. The Glass Industry.
58(4):10. April 1977.
54. Ref. 19, p. 54.
55. Darvin, C.H. Pollution Control in the Glass Industry. The Glass
Industry. July 1979. p. 17. :
56. Ref. 19, p. 44.
57. Ref. 19, p. 55.
58. Ref. 2, p. 7-26.
59. Ref. 2, p. 4-27.
60. Ref. 2, pp. 4-14 to 4-16.
61. Ref. 2, p. 7-19.
5.8-29
-------
-------
5.9 MINERAL PRODUCTS
The three major mineral products industries that generate S02 emissions
are the port!and cement, lime, and clay and brick manufacturing industries.
The primary source of S02 emissions in each of these industries is com-
bustion of fuel used in kilns and drying operations. The fuels are natural
gas, oil, coal, and wood.
5.9.1 Process Descriptions and Emission Sources
5.9.1.1 Lime Production--
The manufacture of lime (quicklime) involves the calcining of limestone
(CaCO or CaC03 •• MgC03) to release carbon dioxide and form quicklime (CaO or
CaO • MgO). Calcitic limestone is approximately 95 percent calcium car-
bonate. Dolomitic limestone contains 30 to 40 percent magnesium carbonate,
and the remainder is calcium carbonate. Quicklime may be further processed
to yield dolomitic pressure-hydrated lime, high-calcium lime, or dolomitic
hydrated lime, depending on the limestone used.1
As illustrated in Figure 5.9-1,2 lime production operations include
limestone quarrying, crushing and sizing, calcining, hydrating, milling,
sizing by screening and air separation, storage, packaging, and shipping.
Calcining in the U.S. plants is performed either in stationary vertical
kilns of various designs or in horizontal rotary kilns. The oldest type
of continuous kiln is the vertical or shaft kiln, which is most efficient
in terms of fuel economy but is limited in capacity per individual unit.
Horizontal rotary kilns, are used to produce slightly more than 90 percent of
the total lime production in the United States.3 Even though fuel economy
is lower and capital investment is greater for rotary kilns, the trend is
toward use of these types because of their high capacity per unit.
Rotary kilns and vertical kilns may be fired with natural gas, fuel
oil, or coal (pulverized coal for rotary kilns). Because of the uncertainty
of natural gas availability and rising energy costs, the trend has been
toward coal firing; coal is now used to produce more than 60 percent of the
commercial lime manufactured in this country.4
The following discussion concerns sulfur dioxide (S02) emissions from
rotary lime kilns (the only source of S02 emissions at the plant), because
5.9-1
-------
HOPPER CAR
DUMP TRUCK
FROM QUARRY
HIGH-CALCIUM
AND DOLOMJTIC
HYDRATED LIME
RAIL
SHIP BARGE
/
Figure 5.9-1. Process flow diagram for lime production.2
5.9-2
-------
virtually all of the new kilns installed in the past 10 years have been
rotary kilns. Because of limited availability and high costs of oil and
natural gas, most installations will be rotary lime kilns designed to burn
coal.
Sulfur is present in most limestone and in all fuels used in the in-
dustry, except natural gas. The sulfur in the limestone does not normally
contribute a substantial portion of the total S02 emissions from a rotary
kiln. The major source of S02 is the sulfur,in the fuel.
During fuel combustion, most of the sulfur in the fuel is converted to
S02. Some of the S02 reacts with the lime product or with the lime dust,
and some is emitted with the kiln off-gas. The amount of S02 that reacts
with the lime product or lime dust depends on the chemical composition of
the stone, the temperature in the kiln, the amount of excess oxygen in the
kiln, and the amount and particle size of the lime dust inside the kiln.5
Table 5.9-1 presents parameters required to perform dispersion calculations
on two model lime kiln plants. When coal or oil with a sulfur content of
less than 1.0 percent is fired, only about 10 percent of the sulfur in the
fuel is vented to the atmosphere as S02. When fuels with higher sulfur
content are used, the S02 removal efficiency of the lime may be reduced to
about 50 percent.6'7
TABLE 5.9-1. LIME KILN MODELING PARAMETERS
Type of
collector
Dry
Wet
Stack gas
temperature
188°C (370°F)
66°C (150°F)
Stack
diameter
0.94 m (3 ft)
0.88 m (2.9 ft)
Stack gas
velocity
5.79 m/s (19 ft/s)
5.99 m/s (19.7 ft/s)
5.9.1.2 Portland Cement Manufacture--
The cement industry includes all establishments engaged in the manu-
facture of hydraulic cement (generic name: portland cement) and of masonry,
natural, or pozzolana cements. This description is limited to the produc-
tion of portland cement because it accounts for 95 percent of the total
cement manufactured in the United States.
5.9-3
-------
Cement is produced by heating to the point of fusion a finely ground
combination of limestone, cement rock, marl, or oyster shells with shale,
clay, sand, iron ore, or aluminum. The fused product, called cement clin-
ker, is ground to a fine powder and shipped in bags or by bulk carrier.
Figure 5.9-2 illustrates the process flow of a typical cement plant.
Initially, the raw materials (limestone, cement, rock, marl, or oyster
shells) are combined with shale, clay, sand, iron ore, or aluminum and with
other trace materials and ground to the desired gradation. Either the dry
process or the wet process is used. In the dry process, heat for drying is
provided by direct dryer firing or hot kiln exhaust gases. The finished
finely ground raw material is then conveyed to the blending operation and
later fed to the kiln. In the wet process, individual raw material slurries
may be blended after grinding. The finished slurry used as a kiln feed may
be 30 to 40 percent water or it may be dewatered to approximately 20 percent
water and fed as a filter cake. The mixed materials are heated in a rotary
kiln and transformed into clinker at approximately 1595°C (2889°F).8 The
clinker is discharged from the kiln, cooled, ground to the desired fineness,
and combined with gypsum to control the setting time of the concrete. The
finished cement product is then stored for later packaging and shipment.9
Emissions also include the products of combustion of the fuel used in
the rotary kilns and drying operations; these emissions are typically NO
and small amounts of S02.9 x
The limited available data on S02 emissions from uncontrolled kilns
using both wet and dry processes indicate a factor of 5.1 kg/Mg (10.2
Ib/ton) attributable to the mineral content of the raw materials and factors
of 2. IS and 3.4S with combustion of oil and of coal (where S is the percent
sulfur content of the fuel in percent)." Emissions of S02 attributable to
firing of gas are negligible. These factors take into account the reactions
with alkaline dusts when no fabric filters are used. With fabric filters,
about 50 percent more S02 is removed in reactions with the alkaline filter
cake." The 5.1-kg/Mg value accounts for part of the available sulfur
remaining in the product because of its alkaline nature and affinity for
S02. Total emissions from the kiln are the sum of the mineral content
factor and the fuel sulfur factor.
5.9-4
-------
u
03
OJ
O)
o
O
q-
(O
•I—
~o
o
to
-------
5.9.1.3 Clay and Brick Production--
The clay and brick production industry consists of those companies
involved in mining and beneficiating of clay minerals and subsequent pro-
cessing of these clay minerals to make bricks. The clay minerals vary
widely in chemical composition and physical properties, but are basically
natural, earthy, fine-grained, hydrated aluminum silicates, which are
plastic when wet, rigid when dry, and vitreous when fired. The U.S. Bureau
of Mines classifies the clay industry into six categories: ball clay,
bentonite, fire clay, fuller's earth, kaolin, and common clay plus shale."
Figure 5.9-3 is a process flow sheet for the clay and brick manufac-
turing industry. The clay ore is excavated from open pit mines, then
crushed in jaw or gyratory crushers or hammer mills (primary crushing);
secondary crushing is done if needed. The material is then screened, dried
from 20 to 30 percent moisture to 1 to 15 percent moisture content in rotary
dryers, and optionally dry-ground in ball, rod, or roller mills. Crushed
common clay or beneficiated clay is cut, formed, and molded in the extruding
process; it is then dried and finally fired in continuous tunnel kilns."
Batch-type periodic kilns are also sometimes used to fire clay and brick,
but they are not widely used especially in modern plants. Periodic kilns
require two to three times more fuel as compared with a tunnel kiln.14
The kiln may be fired with natural gas, oil, coal, or wood. The cur-
rent trend is toward coal and wood. Waste heat from the cooling section of
the kiln is generally used to dry the bricks before they enter .the kiln.
Makeup heat for the dryers, if necessary, is supplied by combustion of the
same fuels used for firing the kiln.15
Table 5.9-2 presents emission factors for S02 emitted from the com-
bustion of natural gas, fuel oil, and coal in the firing operation of
periodic and tunnel kilns without control systems.i*-« Based on these
erosion factors, from a 90.7-Mg/day (100-tons/day) kiln firing 6.8 Mg/day
(7.5 tons/day) coal, gases are discharged to the atmosphere at an estimated
rate of 7.08 ms/s (15,000 ftVmin) and a temperature of 132°C (270°F); these
gases contain 36 kg/h (80 Ib/h) S02.« It is not known what portion'of the
S02 reacts with the clay of the bricks or what portion of the sulfur in the
clay is emitted as S02; however, it is likely that these amounts are mini-
mal.
5.9-6
-------
O
3
•o
O
s-
Q.
o
.0
•o
c
eO
O
S-
o
li-
re
s-
cn
to
to
O)
O
O
s_
O-
ro
en
LO
o>
S-
rs
CD
5.9-7
-------
TABLE 5.9-2.
BRICK MANUFACTURING WITHOUT CONTROLS16-18
SO EMISSION FACTORS FOR
s\
Tunnel kilns
Gas-fired
Oil-fired
Coal-fired
Periodic kilns
Gas-fired
Oil-fired
Coal-fired
a Negligible.
kg SO /Mg brick
/\
2. OS1
3.6S
a
2.95S
6. OS
(Ib S0x/ton brick)
(a)
(4.OS)
(7.2S)
(a)
(5.9S)
(12.OS)
"S" denotes the percent of sulfur in the fuel.
5.9.2 Control Techniques
5.9.2.1 Description—
The primary means of reducing S02 emissions from the lime, cement, and
clay and brick manufacturing industries are the utilization of low-sulfur
fuels and/or the application of some type of emission control device. Clean
fuels, as a means of attaining desired S02 emissions, may not be available
as a viable option in the future because of increasing use of coal and
reduced supplies of cleaner fuels. The sulfur content of coal is typically
higher than that of natural gas and oil, which results in increased S02
emissions. The S02 emissions from the mineral products industries are 0.22
percent (600,000 Mg/yr, 660,000 tons/yr) of the nationwide S02 emissions.19
Lime—The fabric filter and electrostatic precipitator (ESP), which are
the control devices used predominantly for control of particulate matter
from lime kilns, also provide S02 control by increasing the time available
for the reaction of S02 with the lime. Wet scrubbers are generally more
efficient for S02 control than ESP's or fabric filters. In tests of a
rotary lime kiln firing 2.96 percent sulfur coal, S02 emissions were
measured at the inlet and outlet of a wet scrubber with a pressure drop of
5.5 kPa (22 in. H20); emissions were 3.8 kg of S02 per megagram of limestone
(7.6 Ib/ton) uncontrolled, and the estimated control efficiency (from reac-
tion with lime in the kiln) was about 50 percent, based on the sulfur con-
tent of the coal. The wet scrubber alone provided a control efficiency of
5.9-8
-------
96 percent, yielding an overall control efficiency of 98 percent, again
based on the sulfur content of the coal.6 A similar test on a lime kiln
firing 2.97 percent sulfur coal and controlled by a fabric filter indicated
an overall S02 control efficiency of 82 percent. Another similar'test on a
lime kiln burning 1.05 percent sulfur fuel oil and controlled by an ESP
indicated an overall S02 removal efficiency of 88 percent.6
Cement—Most of the S02 emissions are inherently controlled in the
process of cement manufacturing because about 75 percent of the raw feed is
converted to calcium oxide, which reacts with S02. In addition, the
presence of sodium and potassium compounds in the raw material aids in the
direct absorption of S02 into the product. The variable chemistry and
operating conditions in U.S. cement plants affect the amount of S02 removal
and, in some cases, the quality of the product. Sulfur dioxide removal of
this type is 75 percent in plants for which data are available. Sulfur
dioxide also is removed by this same mechanism by fabric filters, in which
the S02-laden gas contacts the collected cement dust. The degree of control
by S02 absorption depends upon the alkali and sulfur contents of the raw
materials and fuel.20 Limited information is available concerning specific
S02 control systems for these sources.21
Brick--No specific information is available on S02 control systems for
the dryers and kilns of brick and clay production industries.21
5.9.2.2 Control Cost--
The cost of burning low-sulfur fuels would be determined by the dif-
ference between the cost of the standard fuel burned and that of the low-
sulfur fuel, which usually is higher.
Little information is available on the cost of S02 scrubbers applicable
to these industries. Investment capital costs of wet scrubbers for lime
kilns range from $311,000 for a scrubber with a pressure drop of 2.24 kPa (9
in. H20) on a 114-Mg/day (125-ton/day) kiln to $793,000 for a scrubber with
a pressure drop of 5.5 kPa (22 in. H20) on a 454-Mg/day (500-ton/day) kiln;
total annualized costs of these units are $137,000 and $537,000, respec-
tively.22 These costs are for scrubbers designed primarily for particulate
control. The operating costs of scrubbers on lime kilns are more than twice
the costs of dry control systems such as ESP's and fabric filters.23
5.9-9
-------
5.9.2.3 Energy and Environmental Impacts--
Section 3.2.1.3 discusses the energy and environmental impacts from S02
control by fuel switching. Scrubbers typically use considerably more energy
than dry collectors, which are the control devices used predominantly in
these industries. In the lime industry, a scrubber with a pressure drop of
5.5 kPa (22 in. H20) needs six times more energy to operate than a baghouse,
which is equivalent to a total plant energy increase of 4 percent. If this
energy is produced in a coal-fired power plant, an additional 21 Mg (23
tons) of S02 per year would be produced, assuming that the power plant
conforms to the standard of performance of 520 ng/J (1.2 Ib of S02/million
Btu) heat input.24
The temperature of the stack gas following a scrubber is lower than
that following a fabric filter or ESP. Consequently, the dispersion charac-
teristics of the emissions are not as favorable. Because the scrubber gases
at lower temperatures are less buoyant, the maximum predicted concentration
of S02 in the ambient air is only slightly lower from the scrubber. If the
possibility of aerodynamic downwash is neglected, no difference is predicted
in the maximum ambient concentrations resulting from scrubbers and dry
collectors.25
Because they produce a sludge that necessitates solid waste management
and water pollution control, scrubbers might pose important problems where
effluent guidelines require zero discharge.
5.9-10
-------
REFERENCES FOR SECTION 5.9
1.
2.
3.
8.
10.
U.S. Environmental Protection Agency. Technical Guidance for Control
of Industrial Process Fugitive Particulate Emissions. Research Tri-
angle Park, N.C. EPA-450/3-77-010. March 1977. p. 2-297.
Ref. 1, p. 2-300.
U.S. Environmental Protection Agency, Office of Air Quality Planning
and Standards. Compilation of Air Pollutant Emissions Factors. 2d ed.
Supplements 1-8. AP-42. Research Triangle Park, N.C. July 1979.. p.
8.15-1.
4. Gutschick, K. Lime Outlook. National
D.C. November 29, 1978. p. 2.
Lime Association. Washington,
Schwartzkopf, F. Lime Burning Technology—A
Operators. Kennedy Van Saun. 1974.
Manual for Lime Plant
6 US Environmental Protection Agency. Standards Support and Environ-
mental Impact Statement. Volume 1: Proposed Standards of Performance
for Lime Manufacturing Plants. EPA-450/2-77-007a. Research Triangle
Park, N.C. April- 1977. p. C-13.
7. Ref. 3, p. 8.15-4.
Barrett, K.W. A Review of Standards of Performance for New Stationary
Sources—Portland Cement Industry. U.S. Environmental Protection
Agency. Research Triangle Park, N.C. Contract No. 68-02-2526. April
1979. pp. 4-6 to 4-8.
U.S. Department of
metallic Minerals.
February 23, 1979.
the Interior, Bureau of Mines,
Mineral Industry Surveys — Cement
p. 3.
Division of Non-
in December 1978.
Ketels, P.A., J.D. Nesbitt, and R.D. Oberle. A Survey of Emissions
Control and Combustion Equipment data in Industrial Process Heating.
U.S. Environmental Protection Agency, Industrial Environmental Research
Laboratory, Office of Energy, Minerals, and Industry. EPA-600/
7-76-022. October 1976. p. 69.
11. Ref. 3, p. 8.6-3.
5.9-11
-------
12.
13.
14.
U.S. Environmental Protection Agency, Industrial Environmental Research
Laboratory Industrial Process Profiles for Environmental Use--The
Clay Industry. Chapter 19. EPA-600/2-77-023s. February 1977. p. 1.
Ref. 12, pp. 26-30.
U.S. Environmental Protection Agency, Research Triangle Institute. A
Screening Study to Develop Background Information to Determine the
Significance of Brick and Tile Manufacturing. Research Triangle Park,
N.C. Contract No. 68-07-0607, Task No. 4. December 1972. pp. 2-7,
£ O *
Ref. 12, p. 28.
Ref. 3, p. 8.3-4.
Resources Research Inc. Air Pollutant Emission Factors. Final
SnK™' M repare5 f°!n Nationa1 A1r Pollution Control Administration,
Durham, N.C. under Contract No. CPA-22-69-119. Reston, Va. April
Norton, F.H. Refractories. 3d ed. New York, McGraw-Hill Book Co.
15.
16.
17.
18.
19' and' 4«nn!£Hinent?1 Proiec^"?n A9ency, Office of Air Quality Planning
and Standards, Research Triangle Park, N.C. 1977 OAQPS Data File
Computer Run Date of March 27, 1979.
20. Ref. 10, p. 72.
21* n;-w;-c-,-EnV1>?Tn^a- 1Protection Agency, Industrial Pollution Control
IndHUS*rial Environmental Research Laboratory. Multimedia
* Environmental Research Needs of the Cement Industry.
No LPum 1Clncinnatn1' Oh- Contract No. 68-03-2586, Work Directive
No. 2586-WD1. January 1979. p. 153.
22. Ref. 6, p. 7-27.
23. Ref. 6, p. 8-8.
24. Ref. 3, p. 1.1-3.
25. Ref. 5, p. 8-7.
5.9-12
-------
5.10 EXPLOSIVES MANUFACTURE
An explosive is a material that decomposes rapidly and spontaneously
under the influence of thermal or mechanical shock, with the evolution of
large quantities of heat and gas. Primary characteristics by which explo-
sives are classified are brisance (shattering power) and sensitivity to
explosion initiation. Other properties, such as heat and production of
toxic gas, may be important in the selection of explosives for specific uses
(e.g., underground mining operations).
Explosives are classified as either high or low explosives, and as
primary or secondary. Primary high explosives are not only very powerful
but also are very sensitive to thermal or mechanical shock. Because of
these properties they are used only in small quantities as initiating ex-
plosives or detonators to set off larger quantities of other explosives.
Secondary high explosives are less sensitive to mechanical or thermal
shock, but explode with great violence when set off by an initiating explo-
sive. Examples are ammonium nitrate mixtures, nitroglycerine (NG), and
2,4,6-trinitrotoluene (TNT).
The low explosives undergo relatively slow autocombustion when set off,
and evolve large volumes of gas in a definite and controllable manner.
Nitrocellulose (NC) is a common example. Black powder was once a principal
low explosive, but industrial use has not been reported since 1971.
The explosives manufacturing industry consists of a commercial and a
military sector. Ammonium nitrate mixtures are the explosives most widely
produced and used by the commercial sector because ammonium nitrate is
inexpensive and readily available. Ammonium nitrate is used in over 90
percent of all commercially manufactured explosives.
Other commercially produced explosives include NG (in bulk and in
dynamite forms), nitrostarch, RDX, and PETN. Apparent annual consumption of
commercial explosives and blasting agents in the United States in 1977
increased 11.4 percent above the 1976 level to 1.68 Tg (3.7 billion Ib).1
The military sector of the explosives industry produces large quanti-
ties of NG (^0.68 Gg, 1.5 million Ib), NC (M3.6 Gg, 30 million Ib), and TNT
(not produced in the private sector).2 The current major U.S. military
explosive, Compound B, is a blend of 40 percent TNT and 60 percent RDX. The
5.10-1
-------
current peacetime production of military explosives is very slow, and active
production sites are few.
Most explosives currently produced are nitrogen-based organic com-
pounds. From a process viewpoint, the nitrogen-based explosives of major
concern with respect to SO emissions are TNT, NC, and NG.
P\
5.10.1 Process Description and SO Emission Sources
S\ " ' '
Manufacture of TNT, NC, and NG generally follows the flow scheme shown
in Figure 5.10-1.3 Concentrated acids are reacted with an organic material
in a nitration step. One of the acids is a nitrating acid (HN03), and the
other is a reaction catalyst, sulfuric acid (H2S04). The explosive product
is separated from the acid phase, washed, purified, and dried. The acid
phase is recovered, reconcentrated, and recycled. Acid fume recovery may be
economically advantageous in larger processes.
Emission sources can be classified as either vents from nitration
vessels, washing units, and acid storage tanks, or tail gases from absorp-
tion towers on acid production, recovery, or concentration units. In addi-
tion to nitrogen oxides, major pollutants from this industry are S02 and
sulfuric acid mist.
5.10.1.1 TNT—
TNT may be prepared by either a batch, three-stage nitration process or
a continuous process known as the Canadian Industries Limited process. In
either case, toluene and nitric acid are the raw materials, and oleum
(fuming sulfuric acid) is used as a reaction catalyst. The overall reaction
may be expressed as:
TOLUENE
H2S04
3HN0
NITRIC
ACID
CHc
+ 3H20
TNT WATER
In the batch process a special mixture of nitric acid and oleum is fed
to each of the three reactors containing the toluene-nitrobody mixture to
5.10-2
-------
o
UJ
co
L.
I 8
•S O LU
o- z: ;>
OO U1
o
2 s!
p
o o
1—1 <_J
<_> UJ
O >•
LU O
—I <_>
to
01
o
o
Q.
CD
c
O
ro
3
C
0>
to
O
i —
Q.
X
O)
-o
(U
tsl
(O
(U
C
OJ
o
o
OJ
O
ce
Q-
5.10-3
-------
produce first mono-, then di-, and finally trinitrotoluene or TNT. Sulfuric
acid is recovered from the spent acids in the primary nitrator by first
separating it from the nitric acid and then boiling it in a drum concen-
trator to remove excess water.
The product from the third nitration vessel is sent to a wash house,
where it is water-washed to remove excess acids. The purification waste-
water, known as "red water," is either concentrated and sold to the paper
industry, incinerated, or discharged directly as a liquid waste stream.
In the continuous process, the nitration of toluene is accomplished by
moving the toluene-nitrobody mixture countercurrent to the oleum-nitric acid
mixture. The spent acid mixture from the TNT nitrators is heated to boil
off the nitric acid, leaving in the tower bottoms dilute sulfuric acid,
which is routed to the sulfuric acid recovery unit. The crude: TNT is puri-
fied in a similar manner as the batch process. Details on the TNT purifi-
cation process are available.4
Although they are not the major emissions, S0x and sulfuric acid mist
are emitted in the manufacture of TNT. The principal sources of SO emis-
sions are the vents from 1) nitration vessel acid fume collection and re-
covery systems, 2) sulfuric acid concentrators, and 3) water, sellite, and
post-sellite purification washes. Of the three washes, the post-sellite
wash is expected to produce the largest SO emission. Emissions of SO from
• x x
the incineration of red water or the onsite production of oleum used in TNT
production can also be considered significant. (Emissions from oleum or
sulfuric acid manufacture are discussed in Section 5.5.)
5.10.1.2 Nitrocellulose (NC)—
Nitrocellulose can be prepared by either a batch-type "mechanical
dipper" process, or a continuous process developed by Hercules, Inc. Both
methods follow the same nitration and washing steps.5,e The overall reac-
tion may be expressed as:
C6H702(OH)3
Cellulose
3HNO£
Nitric
acid
Ho SO
4 C6H702(N03)3 + 3H20
Nitrocellulose Water
5.10-4
-------
When nitration is complete, the reaction mixtures are centrifuged to remove
most of the spent acid, which is fortified and reused or disposed of. The
centrifuged NC undergoes a series of water washings and boiling treatments
for purification.7
Principal sources of SO emissions from batch nitrocellulose manu-
facture are the vents from the reaction pots and centrifuges, spent acid
concentrators, and boiling tubs used for purification.
5.10.1.3 Nitroglycerine (NG)--
Nitroglycerine is the principal explosive component in dynamite. The
older batch method of production is gradually being replaced by the con-
tinuous Biazzi process.8,9 In either case, glycerine is nitrated in the
presence of sulfuric acid according to the following formula:
3HN03 + C3 H803
Nitric Glycerine
acid
'4 C3H5N309 + 3H20
Nitroglycerine Water
Sulfur dioxide emissions from the absorber vent, the reactor, and washwater
system are expected to be small.
5.10.2 Control Techniques
Liquid scrubbers and acid mist eliminators are the control devices
reported to be used to reduce emissions of SO and sulfuric acid from pro-
/\
duction of TNT, NC, and NG.10 The emissions being controlled are primarily
those arising from the onsite recovery and regeneration of sulfuric acid and
oleum. A flow scheme for sulfuric acid recycling is given in Figure
5.10-2.11 Recovery systems such as acid fume recovery and spent acid re-
covery are usually found in the larger explosives manufacturing opera-
tions.12 Both large and small producers reportedly discharge emissions from
the purification and drying of TNT, NC, and NG directly to the atmosphere
without treatment.
Producers of TNT at the Volunteer Army Ammunitions Plant (VAAP) in
Chattanooga, Tennessee, used packed column water scrubbers to recover sul-
furic acid from captured nitration process fumes and spent acid recovery
exhausts.13 Recovered sulfuric acid (68 percent) is converted first to S03,
5.10-5
-------
t3-98% H,SO.
Z 4-
RECYCLE H,SO,
£• 4.
SULFUR
AIR
COOLER-HUMIDIFIER
•WASTEWATER
T
DRYING TOWER
PURIFIED GAS
S02 OXIDIZER
(VANADIUM PENTOXIDE CATALYST)
25%
OLEUM
98%
•40X OLEUM
TAIL GAS -».
SOX.H2S04
MIST. NOX AIR
c.w.
•25X OLEUM
Na2C03
SOLUTION
SELLITE-*-
Figure 5.10-2. Sulfur acid recycle system.
11
5.10-6
-------
then to oleum in a three-step regeneration process involving a rotary kiln,
a vanadium catalyst, and a sulfuric acid/oleum scrubbing system. Sulfuric
acid mist in the exhaust gas from the scrubber is removed by a Brinks mist
eliminator. Residual S02 and S03 are scrubbed again, this time with a
sodium carbonate (Na2C03) solution, to produce a sellite solution of
sodium sulfite (Na2S03) and sodium bisulfite (NaHS03), which is recycled for
use in TNT purification.
The final stack gas emission from the sulfuric acid recovery and
regeneration system at VAAP is reported to contain 240 to 275 ppm S02, plus
residual sulfuric acid mist. VAAP has indicated that the SO emission
control system, as described above, is 95 percent efficient for both acid
mist elimination and S02 removal.14 VAAP officials plan to add a second
mist eliminator behind the carbonate scrubber to remove the remaining acid
mist.
Fluctuations in processing variables such as temperature and pressure
changes within a system, breathing losses, process upsets or spills, and
efficiency of absorbers towers can influence the efficiency of SO and acid
s\
mist controls. Table 5.10-1 presents 1975 data on control of S02 and acid
mist emissions from the manufacture of TNT and NC.15 More recent data are
not available.
Information on costs, energy consumption, and environmental impacts of
S0x control devices in the explosives industry is not available in the open
literature. Section 5.5 gives additional information on sulfuric acid
processing, emissions, and control techniques.
5.10-7
-------
o
2=
O
•a:
co
C£
C3
O
£
CO
CO
LU
O
LO
LU
_J
CQ
E
(tj O
to "v^
•r; X)
K «*™*^ r™"
T3 O
•r- CO
O O,
O ^
•r- O
1- 0
^d-s1
*3 "ro
CO ^
LO
co"
CO
LO
O
0
o
LO
•
CT}
OJ
•4-)
to
3
ro
JE
X
QJ
QJ
-(->
OJ
tvo
NT: continuous process
t—
^~^
CO
.
o
LO
o
0
OJ
•
o
•*~^
OJ
OJ
o
1
LO
OJ
o •
0
OJ
^~
o
Red water incinerator
Z (nitrocellulose)
z:
^.
OJ
CO
0
**J '
,_!
i
o
.
o
Nitration reactors
CO
o
CO
o
LO
CO
1
.
0
CO
vo
1
OJ
o
^j-
co
i.
0
-M
Sulfuric acid concentra
c—
•r—
4 —
O "
E
O) O)
CD >
£ 'CD
OJ
> to
re •!-
OJ OJ
1— r^
(O
>
•
~O QJ
QJ E
•*-> 0
S-
§.£»
O) E
S- 0
E QJ
01 S_
O) O)
-^
1 ^»
OJ
>
re •
-E to
to to
O) O)
tO -E
to •»->
O OJ
0 S-
t- re
CL Q.
E-C
C H"
0
to to
QJ •
E Cn+J
O E to
S- (O O)
"+- S- 4->
to .E OJ
E •*-> 0
0 -r- i-
•r- S 3
to O
to •* to
E to QJ
QJ S- E
•r- O
•^ ^ >•>
E i—
Considerable variations
values reported is show
represents data from on
rO
3
4->
M-
M-
Ol
to
to
O)
OJ
-o
i.
O
CD
E
re
s-
-o
E
Ol
E
fO
O)
-r-
O
OJ
O)
-o
O
S-
o
q-
.QJ
cr>
E
re
q-
o
T3
0)
O)
to
ZD
5.10-8
-------
REFERENCES FOR SECTION 5.10
1. Bureau of Mines. U.S. Department of the Interior. Mineral Industry
Surveys. Apparent Consumption of Industrial Explosives and Blasting
Agents in the United States, 1977. Washington, D.C. May 12, 1978. p.
1.
2. Radian Corporation. Screening Study to Determine the Need for New
Source Performance Standards in the Explosives Manufacturing Industry.
July 1976. pp. 210, 215.
3. Ref. 2, p. 25.
4. Ref. 2, p. 35.
5. Ref. 2, p. 29.
6. U.S. Environmental Protection Agency. Compilation of Air Pollutant
Emission Factors. Supplement No. 5. AP-42. December 1975. p. 5.6-3.
7. American Defense Preparedness Association. State-of-the-Art: Military
Explosives and Propellents Production Industry. Vol. I: The Military
Explosives and Propellents Industry. EPA 600/2-76-213a. October 1976.
p. 69.
8. Ref. 7, p. 72.
9. Ref. 2, p. 28.
10. Ref. 2, pp. 77-80, 191-238.
11. Ref. 2, p. 50.
12. Ref. 2, pp. 47, 52.
13. Ref. 2, pp. 77-80.
14. Ref. 2, p. 212.
15. Ref. 6, pp. 5.6-4, 5.6-5.
5.10-9
-------
-------
5.11 PETROCHEMICALS
This section discusses sulfur oxide emissions from phthalic anhydride
and ethylene production.
5.11.1 Process Description and Emission Sources
5.11.1.1 Phthalic Anhydride—
Phthalic anhydride is produced from either ortho-xylene or naphthalene
by vapor phase oxidation in the presence of a vanadium oxide catalyst. In
1975, 10 phthalic anhydride plants were operating in the continental United
States and 1 plant was operating in Puerto Rico. Nominal production capa-
city of the 11 plants was 495,000 Mg (544,000 tons) per year. Of this
capacity, 70 percent (8 plants) used ortho-xylene as the feed stream, and
the remainder relied on naphthalene. Plants representing another 208,000 Mg
(229,000 tons) per year were not in operation. Of this capacity, all but
45,400 Mg (50,000 tons) was based on the use of naphthalene.1 New plants
are more likely to use the ortho-xylene process because the ortho-xylene
feed is less expensive and has a higher yield ratio.
The two processes are similar except for the reactors used and the
catalyst handling and recovery facilities. The naphthalene process uses
fluidized-bed reactors that require catalyst recovery equipment to separate
the entrained catalyst from the reactor effluent and return it to the fluid-
ized bed. The ortho-xylene process uses fixed-bed tubular reactors. This
process requires that 65 percent of the vanadium catalyst be in the tetra-
valent oxidation state, which is usually achieved by adding sulfur dioxide
to the process airstream. A new ortho-xylene process has been developed by
Rhone-Progil of France, and the process does not require catalyst regenera-
tion with S02. Only one plant, however, is known to use this process.2
The catalyst regeneration step is not required in naphthalene-based
plants;3 consequently, so long as natural gas is used as fuel and the napha-
lene is sulfur free, these plants are not a source of S02 emissions.
The oxidation of ortho-xylene is carried out at 340° to 385°C (650° to
725°F).4 The overall reaction is:
5.11-1
-------
30,
catalyst
0
II
C
/
\
C
II
0
3HoO
o-xylene oxygen
phthalic ' water
anhydride
If uncontrolled, the reaction can proceed further than desired, and the
phthalic anhydride product can be oxidized to maleic anhydride, or even to
carbon dioxide and water.
In the ortho-xylene process, the feed is about 95 to 96 percent ortho-
xylene, with the balance consisting of meta- and para-xylenes. The meta-
and para-xylenes do not react to form acetic anhydride, but are oxidized to
form carbon dioxide, carbon monoxide, and water. The ortho-xylene reaction
produces benzoic acid and maleic anhydride as solid byproducts, which are
emitted in significant amounts as particulates. Both of these are more
volatile than phthalic anhydride.
Figure 5.11-1 is a flow diagram of the ortho-xylene process. At the
start of the process,- filtered air is compressed to between 69 and 97 kPa
(10 to 14 psig) and is passed through a preheater. Liquid ortho-xylene is
vaporized and mixed with the preheated air. The mixture of air and ortho-
xylene then enters the reactor along with sulfur dioxide. The result is an
exothermic reaction. Molten salt is circulated around the reactor tubes to
draw off the heat, which is used to generate low-pressure steam.
The effluent from the reactor contains product phthalic anhydride,
nitrogen, excess oxygen, carbon dioxide, carbon monoxide, water, and small
amounts of maleic anhydride and benzoic acid. This effluent is passed to
switch condensers to extract the solids. The effluent gas stream from the
switch condenser, which contains S02, is usually scrubbed with water or
incinerated before release to the atmosphere.5
The switch condensers are alternately cooled and heated by successive
heat transfer oil streams. The time of the cycle is automatically con-
trolled. Crude phthalic anhydride that has collected during the cold phase
of the condenser cycle is melted from the condenser-tube fins during the
hot-oil circulation period.
5.11-2
-------
o
I—I
I—
o
CO
in
-------
Crude phthalic anhydride is then sent to the pretreatment section to be
heated. Crude phthalic anhydride is dehydrated to form the pure anhydride.
In this process, the associated water, maleic anhydride, and benzoic acid
are partially evaporated. A liquid stream of anhydride is then sent to a
vacuum distillation section where 99.8 percent pure phthalic anhydride is
recovered. It can either be stored as a liquid or be solidified, converted
to flakes, and bagged. The distillation residue consists of phthalic anhy-
dride dissolved in nonvolatile organic compounds.
The sulfur dioxide used to maintain the vanadium catalyst in the tetra-
valent state is emitted in the gas vented from the process. This gas stream
also contains significant amounts of carbon monoxide and organic particu-
lates that must be controlled. A material balance, shown in Table 5.11-1,
has been calculated for a plant with a yearly production capacity of 59 Gg
(130 million Ib). The values shown are for a discharge gas stream from a
system with fresh catalyst and without pollution control equipment. If
fresh catalyst is used, the sulfur dioxide content of the discharge gas
stream is normally 135 ppm (by weight); with aged catalyst, sulfur dioxide
levels can be two to three times higher (the normal upper limit is 400
ppm).6
5.11.1.2 Ethyl ene—
Ethylene is used to make polyethylene, ethylene dichloride, ethylene
glycol, and other important products. Total production for 1979 is expected
to be 13.2 Tg (29 billion Ib), and the market is projected to grow at a rate
of 6 to 7 percent per year.7
The choice of raw materials for ethylene production includes gases
(such as ethane, propane, and butane) and petroleum refinery liquids (such
as naphtha and gas oil). Although the production of ethylene from ethane is
still the most important process, the use of liquid feeds is becoming more
popular because they are becoming more available and allow the coproduction
of gasoline and aromatics.
In ethylene plants using gaseous raw materials, natural gas is the
primary source of the feed. A nearly pure ethane fraction is obtained when
natural gas is processed in an amine unit to remove hydrogen sulfide and to
separate the various components. A process flow diagram for the subsequent
5.11-4
-------
TABLE 5.11-1. UNCONTROLLED GAS EMISSIONS VENTED FROM PHTHALIC
ANHYRIDE PRODUCTION PROCESSES USING ORTHO-XYLENE AS FEED3'0
Component
Sulfur oxides
Carbon monoxide
Carbon dioxide
Nitrogen
Oxygen
Phthalic anhydride
Maleic anhydride
Benzoic acid
Water
Total
Emissions,
kg/h
34
1,094
3,777
184,960
46,240
167
315
20
6,955
243,563
(Ib/h)
(75)
(2,411)
(8,326)
(407,760)
(101,940)
(368)
(694)
(45)
(15,333)
(536,952)
Ratio of emi
to product, by
0.0047
0.1507
0.5204
25.4850
6.3713
0.0230
0.0434
0.0028
0.9583
ssions
weight
33.5596
Source: Reference 6.
Plant capacity: 59 Gg (1.3 x 108 lb)/yr.
5.11-5
-------
production of ethylene from ethane is shown in Figure 5.11-2. Propane and
butane are converted to ethylene in a similar process. The only source of
sulfur emissions from the production of ethylene is from sulfur impurities
in the gas feed.
In the process shown, the gas stream is preheated before it is passed
into the cracking furnace, where thermal decomposition takes place. The
process typically takes place at a temperature of 857°C (1575°F) and a
reactor residence time of less than 1 second.7 The gas from the cracker is
cooled in a fractionator by water scrubbing. A heavy gasoline is condensed
in the fractionator along with some of the water.
Next, the gas stream is compressed and washed with aqueous sodium
hydroxide to remove carbon dioxide and other impurities. Acetylene and
other unsaturated hydrocarbons in the stream are then further converted to
ethylene or saturated hydrocarbons. The next step involves drying and
cooling of the stream to about -156°C (-250°F) and compression to about 3800
kPa (550 psi).8 Hydrogen gas, which does not condense under these condi-
tions, is removed.
The hydrocarbons are then separated by distillation into ethylene,
ethane, and heavier hydrocarbons. Ethylene is the primary product; uncon-
verted ethane is recycled; heavier hydrocarbons up to C4 may either be
recovered as product or used as fuel; and the C5 hydrocarbon fraction is
used in blending gasoline.
In plants producing ethylene from a liquid feed, hydrodesulfurization
is required of gas oils and heavier fractions (which are particularly high
in sulfur) prior to the introduction of the feed into the ethylene plant.
Removal of sulfur from the feed is necessary to prevent the production of a
cracked fuel oil with too high a sulfur content. The sulfur content of the
raw feed is the only source of sulfur dioxide emissions in plants using a
liquid feed.
A process flow diagram for the production of ethylene from liquid feeds
is shown in Figure 5.11-3. The capital investment per kilogram (pound) of
ethylene produced from liquid feeds is larger than that for ethane because
the yield of ethylene from liquid feeds is less. It takes only 0.567 kg
(1.25 Ib) of ethane versus 1.42 kg (3.13 Ib) of naphtha to produce 0.454 kg
5.11-6
-------
O)
CO
_c
+->
-------
i
fcci
>-==
.Is
O)
OJ
cr
-------
(1.0 Ib) of ethylene.9 The sale and disposition of byproducts is therefore
important to offset the greater capital cost of processing liquid feeds.
The temperature and cracking furnace residence time used for processing
liquids are similar to those used for gases. Somewhat higher steam pres-
sures and temperatures are required to avoid fouling the heat exchanger
surfaces with coke and pitch, and the coke deposits on the cracker must be
burned off periodically. The combustion of this coke is a potential source
of sulfur oxide emissions. .
As in ethane processing, the feed stream is first preheated and then
cracked. Rather than water, however, recirculated fuel oil is used for
quenching. Fuel oil not required for quenching is recovered and sold. From
the quench tower, heavy gasoline and sour water are removed.
The gas stream is compressed and washed with aqueous sodium hydroxide
to remove impurities that may include sulfur compounds. The,gases are then
dried, cooled, and compressed. Hydrogen gas, which does not condense with
the other gases, is,removed for use elsewhere in the process. There are no
atmospheric emissions of sulfur dioxide. An amine unit may be required to
remove H2S from the byproduct hydrogen.
The Cj to C4 hydrocarbons are separated by distillation. Ethylene,
propylene, propane, and C4 hydrocarbons are used as products; ethane is
recycled to the process; and methyl acetylene and other unsaturated hydro-
carbons are converted to propylene and propane. The C5 and higher hydro-
carbons are used in blending gasoline.
Potential sources of sulfur emissions are the sour water removed with
the heavy gasoline and the caustic solution used in washing the process gas
stream prior to refrigeration. Sulfur compounds are removed from the sour
water in a stripper. The sulfur-laden gases may be sent to a Claus unit for
sulfur recovery or flared.
5.11.2 Control Techniques
5.11.2.1 Phthalic Anhydride--
The gas stream vented from ortho-xylene-based processes is the only
source of sulfur dioxide emissions. All phthalic anhydride plants use
pollution control devices on this stream to reduce emissions of organic
species.10 The control systems in use include water scrubber-incinerator
combinations and incinerators.11
5.11-9
-------
The water scrubber-incinerator removes organic particulate matter with
a wet scrubber, and the liquid purge stream is subsequently- incinerated.
The carbon monoxide and sulfur dioxide vented from the scrubber are not
controlled.
Direct incineration of the gas stream vented from the process controls
both carbon monoxide and organic compounds, but again S02 emissions are
uncontrolled.
Several methods are available for S02 control, but each has its disad-
vantages. If water-phase S02 controls are used, previous incineration will
cause the following problems.
The additional fuel (assuming natural gas) and air needed to
ensure incineration dilutes the sulfur dioxide in the off-aas
stream, which makes S02 scrubbing inefficient.
The increase in the temperature of the gas stream from the incin-
erator increases the water requirements of the S02 scrubber.
Wet alkali scrubbing of the gas stream before incineration is not
acceptable. In aqueous alkaline solution, phthalic and maleic anhydrides
react as readily as sulfur dioxide and would be converted to salts. The
sodium, potassium, and calcium salts of phthalic and maleic acid are water
soluble and would appear in waste streams.
Because S02 emissions are greatest late in the life of the catalyst,
one S02 control strategy is to regenerate or replace the catalyst more
frequently. Catalyst replacement may be attractive if more frequent shut-
downs and higher catalyst costs are economically feasible.
Rhone-Progil in France has developed a different process for producing
phthalic anhydride from ortho-xylene. A commercial plant in France using
this process has been in operation since 1971. The catalyst used does not
require regeneration by S02. It has not been established, however, whether
this catalyst is compatible with existing domestic plants.2 Researchers at
the French plant suggest that it may be possible to develop a catalyst
that can be used in existing U.S. ortho-xylene plants and that does not
require the addition of sulfur dioxide or other sulfur compounds.2 Research
in developing, new catalysts and improving present catalysts is being con-
ducted in the United States, mostly by private industry. Because the work
5.11-10
-------
is financed by the industries themselves, progress reports and research
results are not disseminated. Consequently, it is not possible to determine
the amount of research being conducted or the status of developments.
5.11,2.2 Ethylene--
Sulfur dioxide emissions within an ethylene plant are primarily con-
trolled by removing the sulfur compounds from the feed streams. When gases
are used as the feed, the hydrogen sulfide in the natural gas stream is
absorbed in an amine solution and then processed into elemental sulfur or
sulfuric acid. A more detailed discussion is presented in Section 5.4,
Natural Gas Industry. If gas oils are the feed source, it is necessary to
remove the sulfur through hydrodesulfurization. A hydrogen sulfide stream
from the desulfurization process is then converted to elemental sulfur. A
more detailed discussion of this process is presented in Section 5.3,
Petroleum .Refineries.
In ethylene plant feeds containing significant amounts of sulfur,
potential emission sources are the sour water coproduced with the heavy
gasoline, and the sulfur dioxide produced by burning coke .deposits from heat
exchanger surfaces during maintenance periods.
Sulfur is disposed of in several ways. Sulfur stripped from the sour
water in the form of hydrogen sulfide is either converted to elemental
sulfur in a Claus plant or incinerated to form sulfur dioxide, which is
released to the atmosphere. Sulfur dioxide emissions from the burning of
coke deposits are also released to the atmosphere.
Specific plant operating procedures can be followed to reduce the
buildup of coke on the furnace tubes and thus minimize S02 emissions .from
decoking operations. Control costs and environmental impacts of removing
sulfur from natural gas and heavy gas oils are discussed in other sections
of this report.
5.11-11
-------
REFERENCES FOR SECTION 5.11
1.
2.
3.
4.
5.
6.
7.
T»W> HlJ9hes- Source Assessment: Phthalic Anhydride
° ^oration. EPA-600/2^032?
8.
9.
10.
11.
Ref. 1, p. no.
Ref. 1, p. 13.
Ref. 1, pp. 106-109.
Ref. 4, pp. 5, 6, 14.
Stinson S.C. Ethylene Technology Moves to Liquid Feeds
Engineering News. 57(22):32. May 1979. L|Mum reeas.
Ref. 7, p. 33.
Ref. 7, pp. 33, 34.
Ref. 1, p. 106.
Ref. 4, p. i.
5.11-12
-------
5.12 INCINERATION
5.12.1 Process Descriptions and Emission Sources
Municipal and offsite incinerators are large, centrally located com-
bustion systems that are designed to handle a variety of wastes. Municipal
incinerators burn a mixture of residential, commercial, and industrial solid
wastes. Private offsite incinerators may be designed to handle liquid,
solid, and sludge wastes, especially those generated by industry.1
Municipal incinerator capacities range from 45 to 90 Mg (50 to 100
tons) per day for smaller units to more than 900 Mg (1000 tons) per day for
larger installations.2 Most municipal incinerators operate continuously,
and use little or no auxiliary fuel. Details of design and operation are
available from several sources.3-5
In contrast to the large municipal incinerators, smaller sized inciner-
ators that handle industrial, commercial, domestic, institutional, or patho-
1-ogical .wastes are usually located near the point of the waste generation.
For this reason they are often referred to as onsite incinerators. Nine out
of ten are multichambered, operate up to 8 hours a day, have intermediate-
sized capacities of 450 kg/h (1000 Ib/h) or less, and utilize auxiliary fuel
(especially natural gas).6-8
Incineration is alsro an effective technique for sewage sludge disposal.
Sludge incineration systems usually include a sludge pretreatment stage to
thicken and dewater the incoming sludge.9 Auxiliary fuel may be required
during startup or when the sludge cannot support combustion as a result of
moisture content.9,10
The major emission point from an incinerator is the furnace stack, and
the only pollutant typically controlled is particulate matter. Sulfur oxide
emissions occur at relatively low levels compared with other pollu-
tants.11'13 Studies have shown that sulfur dioxide (S02) emissions appear
to be more dependent upon the sulfur content of the incoming waste than upon
operating conditions associated with the burning process.14-16
On the average, municipal refuse contains about 0.1 percent sulfur.17
Sulfur compounds, usually in the form of sulfates and sulfides, are present
chiefly in paper, food waste, garden waste, and rubber. During incinera-
tion, some of these compounds are converted to S02 and possibly some S03.18
5.12-1
-------
Based on the assumptions of the average sulfur value for refuse composition
and total conversion to S02, it was estimated in 1970 that the S02 emission
factor was 2 kg/Mg (4 Ib/ton) refuse fired.19 Another estimate placed the
value between 0.5 and 0.9 kg/Mg (1 and 2 Ib/ton) refuse fired.17 These
values agree reasonably well with S02 emission rates measured at 13 munici-
pal incinerators; the average value reported was 1.2 kg S02/Mg (2.3 Ib/ton)
refuse, with a standard deviation of 0.8 kg/Mg (1.8 Ib/ton).19 In a similar
study from 1968 to 1969, four New York City incinerators were monitored for
S02 emissions. Reported values ranged from 0.6 to 2.9 kg/Mg (1.3 to 5.8
Ib/ton) refuse, with an overall average of 1.5 kg/Mg (3.0 Ib/ton).20
Another municipal incinerator study recently found S02 emissions
ranging from 17 to 120 parts per million (ppm) with an average concentration
of 66 ppm over a 3-day test period.21 These values are compared with the
ppm S02 values obtained in the earlier New York study mentioned above and
with data from several other studies.22
Sulfur oxide emissions from refuse incinerators represent only a frac-
tion of a percent of total national sulfur oxide emissions. It was esti-
mated that total S02 emissions from municipal incinerators in 1977 were 6.2
Gg (7300 tons) compared with 17.2 Gg (19,000 tons) for onsite incinera-
tors.23 These values may have been on the conservative side because they
were based on an emission factor of 0.75 kg of S02 per Mg (1.5 Ib/ton)
refuse. Table 5.12-1 lists sulfur oxide emission factors for various types
of refuse incinerators.24
The average sulfur content of sewage sludges is about 1 to 2 per-
cent.9,25 Much of this sulfur is in the form of sulfates or other stable
compounds and is not converted to S02 during incineration. Emissions
measured at several incinerators were less than 750 g of S02 per Mg (1.5 Ib
of S02 per ton) of sludge burned.25 Average S02 emission concentrations are
under 14 ppm.26
5.12.2 Control Techniques
Nearly all incinerators are equipped with some type of particulate
control equipment; i.e., afterburners, settling chambers, water sprays/
scrubbers, electrostatic precipitators.4'27 Primary sulfur oxide emission
5.12-2
-------
TABLE 5.12-1. SULFUR OXIDE EMISSION FACTORS FOR
REFUSE INCINERATORS 2k
Incinerator type
Sulfur oxides as S02>
kg/Mg (Ib/ton)
Municipal
Multiple chamber, uncontrolled
With settling chamber and water spray
system
Industri al/commerci al
Multiple chamber
Single chamber
Domestic
Flue-fed, single chamber
Flue-fed with afterburners and draft
controls
Single chamber with or without
primary burner
Pathological
Open burning/trench
Wood
Rubber tires
Municipal refuse
Auto body
1.25 (2.5)
1.25 (2.5)
1.25 (2.5)
1.25 (2.5)
0.25 (0.5)
0.25 (0.5)
0.25 (0.5)
Negligible
0.05 (0.1)
N.A.a
1.25 (2.5)
NA
NA - Not available.
5.12-3
-------
control is not practiced, and gaseous S0x emissions from incineration
apparently are unaffected by the particulate control systems with one excep-
tion. There is some evidence that wet scrubbers used for particulate con-
trol from sludge incineration may also remove about 20 percent of the S02.28
Higher S02 removal efficiencies have apparently been realized in one sludge
incineration system utilizing a fluidized-bed incinerator with a venturi
scrubber having a 4.5 kPa (18 in. water) pressure drop.29 See Table 5.12-2
for details. The use of venturi scrubbers on municipal incinerators is
expected to decrease because the device has not generally been successful in
meeting the New Source Performance Standards (NSPS) for particulate
matter.^ Because venturi scrubbers have successfully met emission stan-
dards for sludge incinerators, however, their use in this type of incinera-
tion process is expected to continue.31
5.12.3 Control Costs
Cost figures for S0x emission control are not available because it is
not practiced today in incineration.
5.12.4 Energy and Environmental Impact
Impacts are not available because control of sulfur oxides is not
typically practiced.
5.12-4
-------
ID
CM
CO
o
O£
LU
z
1— 1
C_3
)«4
LU
a
ID
CO
LU
CD
i— re
E
O-
a.
•P
c
cu
3
<4_
'[
CU
\s
U
re
in
E
O
J-
M-
(/5
c
o
• ^
in
in
'i
CM
cr
cu
D)
t3
jj
>
^^
CO
c.
3
"**
Csl
c:
Z5
rH
C
3
Di
re
• )n
u
re
u_
/"-s
O
r~- .
CD
LO
CO
o
CO
.
o
\J
CO
.
o
V
CO
.
0
V
CO
CD
V
^^ t i ^.^ ,-^
«^- co ^ Csl
cn csi CM r-^
CO CO CO CO
CD co co cn
OV .CD CD 00
. . • •
rH rH rH rH
CO CM 00 f*-
tr> • • o
<^- co
Csl rH rH CM
CO «si-
W LTJ
rH rH Csl
1 1 rH
CO CM •
Csl
CO «J"
rH rH
CD
•
«cj~
LD rH 00 1^-
r~~ i o
ID «3-
Csl r- 1 CM
•dj"
rH
r^* cn CM
m
LT> rH
CM rH rH rH
1 1 1 0
cn cn cn •
CM CM CM
Csi rH 00
CC CO C_5 Q LU
cn
-P
re
cu
D-
o
u
re
a. •
re Q.
u o
CD CU
rH f-
*• in
>, in
•P CU
•r- i-
U Q.
re
Q. i-
ro cu
u -P
re
in 3t
T3
O C
in -i-
£S
"O '^-s
^™s re
x: Q-
X2
LD
O '
o
\ j-
cn cu
J^ XJ
XJ
o 3
O f-
LO U
in
TD T-
CU i-
^Q ^3
I -p
-a c.
cu cu
M >
3 in
*— CO
-------
REFERENCES FOR SECTION 5.12
3.
4.
7,
8.
9.
10.
n.
12.
Gardner, R. , et al. Source Category Survey for Industrial Inciner-
ators. U.S. Environmental Protection Agency, Emission Standards and
Engineering Division, Office of Air Quality Planning and Standards.
Research Triangle Park, N.C. EPA Contract No. 68-02-3064. Draft
Report 79-S02. August 1979. p. 4-37.
Devitt, T.W., R.W. Gerstle, and N.J. Kulujian. Field Surveillance and
Enforcement Guide: Combustion and Incineration Sources. U.S. Envi-
ronmental Protection Agency Office of Air and Water Programs, Office of
Air Quality Planning and Standards. Research Triangle Park, N.C
APTD-1449. June 1973. p. 3-14.
Ref. 2, pp. 3-1 to 3-24.
Engineering Science, Inc. Exhaust Gases from Combustion and Industrial
Processes. U.S. Environmental Protection Agency. NTIS Publication No
PB-204861. October 2, 1971. pp. III-l to III-ll.
Hefland, R.M. A Review of Standards of Performance for New Stationary
Sources—Incinerators. Mitre Corporation Technical Report MTR-7983
EPA Contract No. 68-02-2526. March 1979. pp. 4-1 to 4-17.
Brinkerhoff, R.J. Inventory of Intermediate-size Incinerators.
Pollution Engineering. 5(ll):33-38. November 1973.
Ref. 1, pp. 4-16 to 4-18, 4-23.
Colonna, R.A., C. McLaren, and E. Sano. Decision-Makers Guide in Solid
Waste Management. EPA-SW-500. 1976. p. 85.
U.S. Environmental Protection Agency. Compilation of Air Pollution
Emission Factors. 3d ed. AP-42. 1977. p. 2.5-1.
Devitt, T.W., and N.J. Kulujian. Inspection Manual for the Enforcement
of New Source Performance Standards: Sewage Sludge Incinerators.
Prepared for the U.S. Environmental Protection Agency, Washington, D C
under Contract No. 68-02-1073. January 1975. p. 3-4.
Ref. 2, pp. 3-23 and 3-24.
Jahnke, J.A., et al. A Research Study of Gaseous Emissions from a
Municipal Incinerator. J. Air Pollution Control Association. 27(8)
pp. 751-753. —
5.12-6 •
-------
13.
14.
15.
16.
17.
18.
19.
22.
23.
24.
25.
26.
Ref. 9, pp. 2.1-2, 2.5-2.
Ref. 12, pp. 751, 753.
Ref. 9, p. 2.1-3.
Carotti, A.A., and R.A. Smith. Gaseous Emissions from Municipal Incin-
erators. Prepared for the U.S. Environmental Protection Agency under
Contract Nos. PH-86-67-62 and PH-86-68-121. 1974. p. 38.
27.
Ref. 2, p. 3-24.
U.S. Environmental Protection Agency, Office of Air Programs.
of Air Pollution from Municipal Incinerators. (Rough Draft.)
N.C. August 1971.
Control
Durham,
A.D. Little, Inc. Systems Study of Air Pollution from Municipal Incin-
eration. U.S. Environmental Protection Agency. 1970. pp. V-49 to
V-54.
20. Ref. 16, p. 47.
21. Ref. 12, pp. 751, 752.
Ref. 12, p, 752.
U.S. Environmental Protection Agency, Office
and Standards. Research Triangle Park, N.C.
Computer run date of March 27, 1979.
of Air Quality Planning
OAPQS 1977 Data File.
Ref. 9, p. 2.1-2.
U.S. Environmental Protection Agency.
Sludge Incineration. January 1972.
Task Force Report on Sewage
U.S. Environmental Protection Agency, Office of Air and Water Programs,
Office of Air Quality Planning and Standards. Background Information
for Proposed New Source Performance Standards: Asphalt Concrete
Plants, Petroleum Refineries, Storage Vessels, Secondary Lead Smelters
and Refineries, Brass or. Bronze Ingot Production Plants, Iron and Steel
Plants, Sewage Treatment Plants. Volume 2, Appendix: Summaries of
Test Data. APTD-1352b. Research Triangle Park, N.C. June 1973. p.
60.
U.S. Environmental Protection Agency, Office of Air and Water Programs,
Office of Air Quality Planning and Standards. Background Information
for Proposed New Source Performance Standards: Asphalt Concrete
Plants, Petroleum Refineries, Storage Vessels, Secondary Lead Smelters
and Refineries, Brass or Bronze Ingot Production Plants, Iron and Steel
Plants, Sewage Treatment Plants. Volume 1, Main Text. APTD 1352a.
Research Triangle Park, N.C. June 1973. p. 57.
_
5.12-7
-------
28. Ref. 9, p. 2.5-2.
29. Ref. 26, p. 60.
30. Ref. 5, pp. 1-1, 6-1.
31. Axetell, K. , T.W. Devitt, and N.J. Kulujian. Inspection Manual for the
Enforcement of New Source Performance Standards: Municipal Incin-
erators. Prepared for the U.S. Environmental Protection Agency,
Division of Stationary Source Enforcement under Contract No. 68-02-
1073. Washington, D.C. January 1975. p. 3-6.
5.12-8
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
REPORT NO.
EPA-450/3-81-004
3. RECIPIENT'S ACCESSION NO.
TITLE AND SUBTITLE
Control Techniques for Sulfur Oxide Emissions
From Stationary Sources - Second Edition
5. REPORT DATE
6. PERFORMING ORGANIZATION CODE
AUTHOR
-------
------- |