United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EPA-450/3-80-001a
February 1980
Air
Benzene Emissions
from Maleic
Anhydride Industry —
Background Information
for Proposed Standards
Draft
EIS
-------
EPA-450/3-80-001a
Benzene Emissions from
Maleic Anhydride Industry
Background Information
for Proposed Standards
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
February 1980
-------
This report has been reviewed by the Emission Standards and Engi-
neering Division of the Office of Air Quality Planning and Stand-
ards, U.S. Environmental Protection Agency, 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 available through the Library Services
Office (MD-35), U.S. Environmental Protection Agency, Research
Triangle Park, N.C. 27711, or from National Technical Information
Services, 5285 Port Royal Road, Springfield, Va. 22161.
PUBLICATION NO. EPA-450/3-80-001a
11
-------
Background Information
and Draft
Environmental Impact Statement
for
Benzene Emissions from Maleic Anhydride Plants
Type of Action: Administrative
Prepared by:
Don R. Goodwill ^ (Date)
Director, Emission Standards and Engineering Division
Environmental Protection Agency
Research Triangle Park, N.C. 27711
Approved by:
_
DaVid G. Hawkins (Date)
Assistant Administrator, Air, Noise, and Radiation
Environmental Protection Agency
Washington, D.C. 20460
Draft Statement Submitted to EPA's
Office of Federal Activities for Review on
"//
(D
ate)
This document may be reviewed at:
Central Docket Section
Room 2903B, Waterside Mall
Environmental Protection Agency
401 M Street, S.W.
Washington, D.C. 20460
Additional copies may be obtained at:
Environmental Protection Agency Library (MD-35)
Research Triangle Park, N.C. 27711
National Technical Information Service
5285 Port Royal Road
Springfield, Va. 22161
m
-------
TABLE OF CONTENTS
INTRODUCTION I'1
SUMMARY S-l
1. THE MALEIC ANHYDRIDE INDUSTRY 1-1
1.1 General 1-1
1.2 Process Descriptions and Emissions 1-5
1.2.1 Benzene Oxidation Process 1-6
1.2.2 n-Butane Oxidation Process 1-11
1.2.3 Byproduct of Phthalic Anhydride Production 1-16
1.2.4 Foreign Process 1-16
1.3 Summary 1-17
1.4 References 1-17
2. EMISSION CONTROL TECHNIQUES 2-1
2.1 Alternative Emission Control Techniques 2-1
2.1.1 Carbon Adsorption 2-1
2.1.2 Thermal Incineration 2-3
2.1.3 Catalytic Incineration 2-4
2.1.4 n-Butane Process Conversion 2-4
2.2 Performance of Emission Control Techniques 2-5
2.2.1 Carbon Adsorption 2-5
2.2.2 Thermal Incineration 2-10
2.2.3 Catalytic Incineration 2-15
2.3 References 2-15
3. REGULATORY OPTIONS 3-1
4. ENVIRONMENTAL AND ENERGY IMPACT 4-1
4.1 Air Pollution Impact 4-1
4.1.1 Carbon Adsorption (99 Percent
Control of Benzene) 4-17
4.1.2 Carbon Adsorption (97 Percent
Control of Benzene) 4-19
4.1.3 Thermal Incineration (99 Percent
Control of Benzene) 4-20
4.1.4 Thermal Incineration (97 Percent
Control of Benzene) 4-20
4.1.5 n-Butane Process Conversion (100
Percent Control of Benzene) 4-21
-------
TABLE OF CONTENTS (continued)
4.2 Water Pollution Impact 4-22
4.3 Solid Waste Disposal Impact 4-23
4.4 Energy Impact 4-23
4.4.1 Carbon Adsorption (99 Percent
Control of Benzene) 4-24
4.4.2 Carbon Adsorption (97 Percent
Control of Benzene) 4-24
4.4.3 Thermal Incineration (99 Percent
Control of Benzene) 4-24
4.4.4 Thermal Incineration (97 Percent
Control of Benzene) 4-24
4.4.5 n-Butane Process Conversion 4-25
4.4.6 Summary 4-25
4.5 Other Environmental Impacts 4-28
4.6 Other Environmental Concerns 4-28
4.6.1 Irreversible and Irretrievable
Commitment of Resources 4-28
4.6.2 Safety Issues 4-28
4.7 References 4-28
5. ECONOMIC IMPACT 5-1
5.1 Summary of Economic Impacts of the Maleic Anhydride
Benzene NESHAP 5-1
5.1.1 Capital Budget Impacts 5-1
5.1.2 Shifts in Competitive Position 5-1
5.1.3 Price Impacts on Products that Use MA 5-1
5.1.4 Employment and Balance of Trade Impacts 5-2
5.1.5 Annualized Costs in Fifth-Year and Energy
Impacts 5-2
5.1.6 Impact of Requiring n-Butane (100 Percent
Control) at New MA Facilities 5-2
5.2 Industrial Economic Profile 5-2
5.2.1 Maleic Anhydride Supply and Capacity 5-2
5.2.1.1 General 5-2
5.2.1.2 The Individual MA-Producing Companies. . . . 5-4
5.2.2 MA Usage and Demand 5-7
5.2.3 Prices 5-9
5.2.3.1 Price of MA 5-9
5.2.3.2 Feedstock Costs 5-11
5.2.3.3 Transportation Costs 5-11
5.2.4 Briquettes vs. Molten MA 5-11
5.2.4.1 Imports of Maleic Anhydride 5-12
5.3 Cost Analysis of Alternative Emission Control Systems. . . . 5-15
5.3.1 Introduction 5-15
5.3.2 Summary of Technical Parameters Used as the Basis
in Cost Analysis 5-16
5.3.3 Control Costs for Maleic Anhydride Facilities .... 5-20
VI
-------
TABLE OF CONTENTS (continued)
5.3.3.1 Costs to Achieve the 97-Percent Regula-
tory Option 5-20
5.3.3.2 Costs to Achieve the 99-Percent Regula-
tory Option 5-20
5.3.3.3 Comparison of Control Levels 5-25
5.3.3.4 Monitoring 5-26
5.3.4 Cost-Effectiveness of the Alternative Emissions
Limits 5-26
5.3.5 Control Cost Comparison 5-33
5.4 Economic Impact of Regulatory Options 5-34
5.4.1 Introduction 5-34
5.4.2 Impact on Manufacturers 5-37
5.4.2.1 Capital Budget Requirements 5-37
5.4.2.2 Intraindustry Competition 5-41
5.4.2.3 Effect of Cost Pass-Through on Market Com-
petition Due to Benzene Emissions Control. . 5-41
5.4.2.3.1 Price Differentials in a Low-Demand
MA Market 5-46
5.4.2.3.2 Price Differentials and the Future
MA Market 5-50
5.4.2.3.3 Effect of Transportation on Intra-
industry Impacts 5-51
5.4.2.4 Effect of Imports 5-55
5.4.2.5 Summary of Impact on Manufacturers 5-56
5.4.3 Effect on Product Prices 5-56
5.4.3.1 Cost Pass-Through to the Final Customer. . 5-56
5.4.4 Employment and Balance of Trade 5-62
5.4.5 Fifth-Year Impacts 5-63
5.5 Economic Impacts of Using n-Butane as the Feedstock
at New Maleic Anhydride Plants 5-63
5.5.1 Introduction 5-63
5.5.2 Impact on Licensors 5-64
5.5.3 Impact on the Price and Availability of Feedstocks. . 5-64
5.5.4 Impact on the Economic Life of Existing Plants. . . . 5-66
6 5.6 References 5-66
Appendix A—Evolution of the BID A-l
Appendix B--Index to Environmental Impact Considerations B-l
Appendix C--Emission Source Test Data C-l
Appendix D—Emission Measurement and Continuous Monitoring D-l
Appendix E--Methodology for Estimating Mortality and Lifetime Risk
from Exposure to Benzene Emissions from Maleic
Anhydride Plants E-l
vn
-------
LIST OF TABLES
Summary
S-l Environmental and Economic Impacts of Regulatory
Alternatives S-2
Chapter 1.
1-1 Maleic Anhydride Usage and Growth 1-2
1-2 Maleic Anhydride Capacity 1-3
1-3 Benzene and Total VOC Emissions from Model
Uncontrolled Maleic Anhydride Plant 1-10
1-4 Waste Gas Composition—Product Recovery Absorber . . . 1-12
1-5 Benzene Emissions From Maleic Anhydride Plants .... 1-13
Chapter 2.
2-1 Technical Data—Carbon Adsorption System
(99 Percent Control) 2-9
2-2 Technical Data—Carbon Adsorption System
(85 Percent Control) 2-9
2-3 Technical Data—Carbon Adsorption System
(98 Percent Control at Design) 2-11
2-4 Design Criteria—Thermal Incineration
(99 Percent Control) 2-14
2-5 Technical Data—Thermal Incineration
(97 Percent Control 2-14
Chapter 4.
4-1 Maximum Annual Average Benzene Concentrations .... 4-2
4-2 Mean Annual Average Benzene Concentrations Produced
by Emissions from the Combined Sources 4-4
4-3 Mean Annual Benzene Concentrations Produced by
Emissions From Individual Sources 4-5
4-4 Maximum 1-hr, 3-hr, 8-hr, 24-hr, and Annual
Average Benzene Concentrations Produced by the
Combined Emissions from a Maleic Anhydride Plant
at Any Distance Downwind and at 0.1, 1.0, 10.0,
and 20.0 km 4-12
4-5 Identification of Model Source Numbers by Source
Name and Emission Control Alternative 4-14
4-6 Reduction in Benzene Emission for Selected
Control Levels 4-18
4-7 Total National Incremental Energy Requirement 4-26
4-8 n-Butane and Benzene Characteristics 4-27
Chapter 5.
5-1 Maleic Anhydride Capacity 5-5
5-2 Maleic Anhydride Usage and Growth 5-8
-------
LIST OF TABLES (continued)
5-3 Estimated Technical Parameters Used in Developing
Control System Costs 5-17
5-4 Cost Parameters 5-19
5-5a Existing Plant Costs for Achieving 97 Percent
Benzene Emission Reduction (Control Method:
Carbon Adsorption) 5-21
5-5b Existing Plant Costs for Achieving 97 Percent
Benzene Emission Reduction (Control Method:
Thermal Incineration With Primary Heat
Recovery) 5-22
5-6a Existing Plant Costs for Achieving 99 Percent
Benzene Emission Reduction (Control Method: Carbon
Adsorption) 5-23
5-6b Existing Plant Costs for Achieving 99 Percent
Benzene Emission Reduction (Control Method: Thermal
Incineration With Primary Heat Recovery) 5-24
5-7 Costs for Continuous Monitoring of Benzene Stack
Emissions 5-27
5-8 Cost Summary for Existing Maleic Anhydride
Plants 5-28
5-9 Estimated Total Investment Cost for Achieving
Benzene Emission Reduction 5-38
5-10 Ratio of MA Sales to Parent Company Sales 5-40
5-11 Comparison of Control Costs to Total Company Capital
Expenditures 5-42
5-12 Possible Cost Pass-Through Under Case Assumption
of 56 Percent Production Capacity 5-43
5-13 Possible Cost Pass-Through Under Case Assumption
of 100 Percent Production Capacity 5-44
5-14 Depiction of Possible Competitive Advantages Due to
Cost Pass-Through Under the 56-Percent Production
Capacity Assumption 5-47
5-15 Depiction of Possible Competitive Advantages Due to
Cost Pass-Through Under the 56-Percent Production
Capacity Assumption 5-48
5-16 Effect of Transportation Under a 97-Percent Control
Level 5-52
5-17 Effect of Transportation Under a 99-Percent Control
Level 5-53
5-18 Summary of Impact of 97 Percent Benzene Control
Level on MA Companies 5-56
5-19 Summary of Impact of 99 Percent Benzene Control
Level on MA Companies 5-57
5-20 Price Increases of Polyester Resins Due to Increased
MA Prices 5-59
5-21 Price Increases of Fumaric Acid Due to Increased
MA Prices 5-60
5-22 Price Increases of Malathion Due to Increased MA
Prices 5-61
IX
-------
LIST OF FIGURES
Chapter 1.0
1-1 Manufacturing Locations of Maleic Anhydride 1-4
1-2 Process Flow Diagram for Uncontrolled Model Plant . . . 1-8
Chapter 4.0
4-1 Layout of the Maleic Anhydride Plant Showing Source
Locations 4-13
Chapter 5.0
5-1 Manufacturing Locations of Maleic Anhydride 5-3
5-2 Captive and Merchant Sales of MA Companies 5-6
5-3 Price Fluctuations of Maleic Anhydride 5-10
5-4 Maleic Anhydride: A Comparison of Imports to U.S.
Production and Demand 5-13
5-5 U.S. Consumption of MA by Source in 1977 5-14
5-6 Cost Effectiveness of Alternative Control Systems--
Operating Factor 4,500 Hours With Monitoring 5-31
5-7 Cost Effectiveness of Alternative Control Systems--
Operating Factor 8,000 Hours With Monitoring 5-32
5-8 Installed Costs of Carbon Adsorbers 5-35
5-9 Installed Costs of Thermal Incinerators With Primary
Heat Recovery 5-36
-------
INTRODUCTION
This document describes the domestic maleic anhydride production
industry with emphasis on the benzene emissions to the atmosphere. The
first sections of the report discuss the characteristics of the industry,
the processes used to produce maleic anhydride, and the associated emissions.
Only ducted process vents are covered in this document. Fugitive emission
sources of benzene, storage and handling sources of benzene, and miscellaneous
secondary sources of benzene will be covered in other documents in preparation.
Applicable control techniques and alternative levels of control and their
environmental, energy, cost, and economic impacts are described.
1-1
-------
SUMMARY
REGULATORY ALTERNATIVES
Two regulatory alternatives were considered for best available technol-
ogy (considering environmental, energy, and economic impacts) (BAT) to
control benzene emissions from existing process vents of plants that use
benzene as a feedstock to manufacture maleic anhydride or maleic acid.
These alternatives are:
A 97-percent reduction from uncontrolled emission levels, based
on the best control that has been achieved at an existing maleic
plant using universally applicable equipment; and
A 99-percent reduction from uncontrolled levels, based on the
best control considered feasible using technology transfer.
While no particular control technology would be specifically required for
either alternative, carbon adsorption or thermal incineration could be used
in achieving either 97 or 99 percent control.
A third alternative, a 100-percent reduction in benzene emissions
based on the use of a substitute feedstock, such as n-butane, was con-
sidered as an alternative for BAT for new sources. This alternative was
not considered as an alternative for BAT for existing sources because of
uncertainties about the technical feasibility of converting each existing
source to n-butane and the impacts of such a conversion. However, few new
sources are expected to be built until the mid-1980's. Thus, a new source
could be designed to use n-butane and would therefore not encounter the
potential problems associated with conversion.
ENVIRONMENTAL IMPACTS—EXISTING SOURCES
Table S-l summarizes the environmental and economic impacts of the
regulatory alternatives considered. These alternatives would reduce nation-
wide benzene emissions from maleic anhydride plants from 6,400 Mg/yr to
1,030 Mg/yr for 97 percent control, to 540 Mg/yr for 99 percent control,
S-l
-------
TABLE S-l. ENVIRONMENTAL AND ECONOMIC IMPACTS OF REGULATORY ALTERNATIVES
Alternative
97% control
Impact on
benzene
+3
Impact
on HC
+ 3
Other aii-
impacts
-1
Water
impact
-1
Sol id waste
impact
-1
Enerpy
impact
-1
Space
impart
-1
Noise
impact
-1
Economic
impact
-2
i control
-1
-1
-1
-3
-1
-1
-3
Substitution
of feedstock,
such as n-butane
(new plants only)
-2
-1
-1
Delayed standard
-2
-2
CO
i
No standard
-3
-3
KEY + Beneficial impact
Adverse impact
0 No impact
1 Nrgliuible impact
2 Small impact
3 Moderate impact
A Large impact
5. Larger impact
-------
and to 0 Mg/yr for 100 percent control. These emissions include fugitive,
storage, and handling emissions, which are not covered by the proposed
standard.
The control systems likely to be used to meet a 97-percent or 99-percent
benzene emission reduction (incineration or carbon adsorption) would also
reduce emissions of other hydrocarbons, which might be toxic and which
contribute to photochemical oxidant formation and associated environmental
problems.
The reduction in nationwi.de benzene emissions would result in minimal,
adverse environmental impacts. These adverse impacts would include small
increases in nitrogen oxide because of high incineration temperatures and
sulfur oxide emissions into the air, if fuel oil were used as a supplemental
fuel for incinerators. There would be ^mall increases in solid wastes and
in benzene in wastewater. Compliance with the alternatives would increase
nationwide energy consumption by an estimated 50,000 barrels for 97 percent
control and 85,000 barrels of fuel oil equivalent per year for 99 percent
control by 1980.
ECONOMIC IMPACTS—EXISTING SOURCES
The capital investment required by the domestic maleic anhydride
industry to comply with the proposed standard would be about $6.6 million
for 97 percent control and $9.1 million for 99 percent control over the
2-year period from 1979 to 1981. The total annualized operating costs of
the industry would increase by about $2.5 million/yr for 97 percent control
and $4.5 million for 99 percent control by 1983. If continuous monitoring
were required, annualized costs would be increased by an additional $9,000
for each plant. The market price of maleic anhydride would increase an
estimated 1.2 percent for 97 percent control and 1.7 percent for 99 percent
control. In addition, one plant might close if the standard were based on
97 percent control and two plants might close if the standard were based on
99 percent control.
ENVIRONMENTAL IMPACTS-NEW SOURCES
The use of a substitute feedstock by new sources would eliminate
benzene emissions from process vents, storage and handling of benzene, and
fugitive sources. Information indicates there would be greater quantities
of volatile organic compounds (VOC's) from the uncontrolled n-butane process
S-3
-------
than from the uncontrolled benzene process. EPA is currently developing a
new source performance standard for air oxidation processes in the chemical
industry. This standard would cover new sources producing maleic anhydride
from n-butane.
ECONOMIC IMPACTS—NEW SOURCES
A plant using the n-butane process has been estimated to achieve as
much as a 7.3
-------
1. THE MALEIC ANHYDRIDE INDUSTRY
1.1 GENERAL
This chapter discusses maleic anhydride production, focusing primarily
on three basic processes used in the United States and the benzene emis-
sions related to these processes. The processes are: oxidation of
benzene, oxidation of n-butane, and recovery as a byproduct of phthalic
anhydride manufacture. Most of the current domestic capacity is based on
the oxidation of benzene. Foreign processes are also briefly described.
Table 1-1 shows the end uses of maleic anhydride and their expected
growth rates. The predominant end use is the production of unsaturated
polyester resins, which go into reinforced plastic applications such as
marine craft, building panels, automobiles, tanks, and pipes. The United
States maleic anhydride production capacity for 1977 was reported to be
o
229,000 Mg (5.0 x 10 lb) with only 56 percent of this capacity currently
1 2
used. ' At an estimated 11-percent annual growth in maleic anhydride
consumption, production would reach 95 percent of present capacity by 1982.
No shortage of benzene, the major raw material, is expected during this
period.
As of 1977, there were eight producers of maleic anhydride in the
United States with 10 plants. However, in 1979, Koppers announced the
mothball ing of its Pennsylvania facility. Table 1-2 lists the producers
and the processes being used, while Figure 1-1 shows the plant locations.
Approximately 82 percent of the 229,000-Mg/yr (5.0 x 108 Ib/yr) domestic
capacity is based on oxidation of benzene; 16 percent capacity comes from
oxidation of n-butane; and the remaining 2 percent results from phthalic
anhydride production, which yields maleic anhydride as a byproduct.
Because of anticipated increases in the price of benzene, work began in
1960 to develop a catalyst suitable for producing maleic anhydride from
1-1
-------
TABLE 1-1. MALEIC ANHYDRIDE USAGE AND GROWTH
End use
Unsaturated polyester resins
Agricultural chemicals
Lubricating additives
Fumaric acid
Copolymers
Maleic acid
Reactive plasticizers
Surface-active agents
Alkyd resins
Chlorendic anhydride and acid
Other
All MA products
1978
demand
Gg/yr 1978 demand
(in 1,000 as % of
Ib/yr) production
77.7 (171,200)
13.4 (29,600)
8.7 (19,200)
6.7 (14,800)
5.9 (13,000)
4.2 (9,200)
4.0 (8,800)
3.2 (7,000)
1.5 (3,400)
1.2 (2,600)
7.5 (16,600)
134.0 (295,400)
58
10
7
5
4
3
3
2
1
1
6
100
Average
annual
% growth
1978-83
13
10
12
5
9
10
8
8
4
13
5
11
SOURCES: Chemical Profile on Maleic Anhydride. Chemical Marketing
Reporter. February 18, 1978.
Blackford, J. C. Marketing Research Report on Maleic
Anhydride. In: Chemical Economics Handbook. Menlo Park,
Stanford Research Inst. July 1976.
1-2
-------
TABLE 1-2. MALEIC ANHYDRIDE CAPACITY
Capacity--1977
103 Mg (Ib)
Process
1. Amoco, Joliet, 111.
2. Ashland, Neal, W.Va.
3. Koppers, Bridgeville, Pa.
4. Koppers, Chicago, 111.
5. Monsanto,.St. Louis, Mo.
6. DENKA, Houston, Tex.
7. Reichhold, Elizabeth, N.J.
8. Reichhold, Morris, 111
9. Tenneco, Fords, N.J.
27 (59,600)
27 (59,600)
15 (33,000)
5 (11,000)
48 (105,800)
23 (50,800)
14 (30,800)
20 (44,000)
12 (26,400)
10. U.S. Steel, Neville Island, Pa. _38 (83,800)
TOTAL 229 (504,800)
Oxidation of n-
butane
Oxidation of
benzene
Oxidation of
benzene
Byproduct of
phthalic
anhydride
manufacture
(80%) oxidation
of benzene
(20%) oxidation
of n-butane
Oxidation of
benzene
Oxidation of
benzene
Oxidation of
benzene
Oxidation of
benzene
Oxidation of
benzene
SOURCES: Blackford, J. C. Marketing Research Report on Maleic Anhydride.
In: Chemical Economics Handbook. Menlo Park, Stanford Research
Inst. July 1976.
Letter from Hewett, P. S., Reichhold Chemicals, Inc., to
Patrick, D. R., Office of Air Quality Planning and Standards,
U.S. Environmental Protection Agency. March 27, 1978.
1-3
-------
j NORTH DAKOTA
i
i
L
.' SOUTH DAKOTA
|
I
i
L
\ Ml
(
\
I
i
i
i
MINNESOTA
NEBRASKA
f
I
1. Amoco, Joliet, III.
2. Ashland, Neal.W. Va.
3. Koppers, Bridgeville, Pa.
4. Koppers, Chicago, III.
5. Monsanto, St. Louis, Mo.
6. DENKA, Houston, Tex.
7. Reichhold, Elizabeth, N.J.
8. Riechhold, Morris, III.
9. Tenneco, Fords, N.J.
10. U.S. Steel, Neville Island, Pa.
Figure 1-1. Manufacturing locations of maleic anhydride.
SOURCE: Hydroscience. Emission Control Options for the Synthetic Organic Chemicals Manufacturing Industry: Maleic
Anhydride Product Report. March 1978.
1-4
-------
n-butane/butene (C,) streams available from naphtha cracking. This effort
was curtailed during the 1961-1967 period when the maleic anhydride market
was depressed and low-cost benzene was available. In 1967, however, demand
for maleic anhydride increased, and Kasei Mizuishima renewed work in Japan.
In 1974, DENKA, Chem Systems, BASF, Bayer, Alusuisse/UCB, and Mitsubishi
made announcements concerning the production of maleic anhydride from C^
feedstocks. Presently, Amoco and Monsanto are producing maleic anhydride
from an n-butane feedstock, and Halcon Catalyst Industries has completed a
plant to produce a catalyst to convert n-butane to maleic anhydride.
In 1979 Monsanto announced plans to build a 45,000-Mg/yr (100 x 10 lb/
yr) maleic anhydride plant at Pensacola, Florida, using a proprietary
n-butane process. The plant is scheduled for completion in 1983. Later
in 1979, DENKA and Badger Company announced plans to build a demonstration
plant at DENKA's facility in Houston that would use n-butane as a feed-
stock. The projected growth rate for the n-butane oxidation process
through 1982 is 24.3 percent, compared to only a 9.1-percent growth rate
for the benzene oxidation process. This assumes the n-butane process
becomes fully commercialized. Substantive data regarding the economic
incentives for switching to n-butane oxidation were not available when this
document was prepared. Finally, no growth in the quantity of maleic anhy-
p
dride recovered during phthalic anhydride production is expected.
1.2 PROCESS DESCRIPTIONS AND EMISSIONS
The two major processes used to manufacture maleic anhydride in the
United States are benzene oxidation and n-butane oxidation, while a small
amount of maleic anhydride is recovered as a byproduct of phthalic anhydride
production. The only significant foreign process for maleic anhydride
production not used in the United States starts with a butene mixture
o
feedstock. While this process is currently used in France, no plans are
known to introduce this process domestically.
This chapter addresses only ducted process emissions (i.e., normal
vents from the process equipment) and does not include emissions from the
storage and handling of benzene, leaks from equipment, or secondary sources
of benzene (e.g., benzene evaporation from a wastewater stream). It is
important to recognize that no benzene emissions result from the recovery
process from phthalic anhydride production. No benzene emissions are
1-5
-------
anticipated from any process using n-butane or butenes, and none has been
9
found at lower detection limits of 1 ppmv.
1.2.1 Benzene Oxidation Process
Maleic anhydride is produced by the following vapor phase chemical
reaction:
0
9/2 09 ^ II 0 + 2H0 + 2C0
L.
BENZENE OXYGEN MALEIC WATER CARBON
ANHYDRIDE DIOXIDE
The process flow diagram with numbered streams shown in Figure 1-2 repre-
sents a typical process. This typical process is continuous, although some
plants operate dehydration and distillation batchwise. The emissions in
either case are judged to be equivalent.
A mixture of benzene and air enters a tubular reactor where the cata-
lytic oxidation of benzene is carried out at a temperature of 350° to
400° C (662° to 752° F). The catalyst contains approximately 70 percent
vanadium pentoxide supported on an inert carrier; most of these catalysts
also contain 25 to 30 percent molybdenum oxide. The reaction is highly
exothermic, releasing 24.4 MJ/kg (10,514 Btu/lb) of reacted benzene, with
the excess heat being used to generate steam. Maleic anhydride yields
range from 60 to 67 percent of theoretical.
The reactor feed mixture is provided with excess air to keep the
benzene concentration below its lower explosive limit of 1.5 volume percent.
The resultant large volume of reactor exhaust (Stream 3) directly influences
the size of the subsequent product recovery equipment. After reaction, the
stream passes through a cooler, partial condenser, and separator in which
40 percent of the maleic anhydride is condensed and separated as crude
product (Stream 4).10 The remaining product and other organics (Stream 5)
enter the product recovery absorber, where they contact water or aqueous
maleic acid. The liquid effluent from the absorber (Stream 6) is a 40
percent (by weight) aqueous solution of maleic acid. The absorber vent
1-6
-------
(Vent A) exhausts to the atmosphere or Is directed to an emission control
device.
The 40-percent maleic acid (Stream 6) is dehydrated by azeotropic
distillation with xylene. Any xylene retained in the crude maleic anhy-
dride (Stream 9) is removed in a xylene-stripping column, and the crude
maleic anhydride (Stream 10) from this column is then combined with the
crude maleic anhydride from the separator (Stream 4).
Color-forming impurities in the crude maleic anhydride are removed by
aging, which causes impurities to polymerize. After aging, the crude
maleic anhydride (Stream 11) is fed to a fractionation column that yields
purified molten maleic anhydride as the overhead product (Stream 12). The
fractionation column bottoms containing the color-forming impurities are
removed as liquid residue waste (Stream 13). This stream is either com-
bined with other effluent or is fed to a liquid incinerator. A small
percentage of the finished product is made into briquettes.
The vacuum lines from the dehydration column, xylene stripper, and
fractionation column are joined to the vacuum system (Stream 14). The
refining vacuum system vent (Vent B) can exhaust to the atmosphere, recycle
to the product recovery absorber (Stream 5), or be directed to an emission
control device. Water from the vacuum system can be recycled as makeup
water (Stream 7) or join the liquid residue waste (Stream 13).
Essentially, all process emissions exit through the product recovery
absorber (Vent A). These emissions include any unreacted benzene; this is
typically 3 to 10 percent of the total benzene feed. The only other
process emission source is the refining vacuum system vent (Vent B), which
may contain small amounts of maleic anhydride, xylene, and a slight amount
of benzene, because benzene could be absorbed in the liquid stream from the
product recovery absorber (Stream 6) or in the crude maleic anhydride from
the separator (Stream 4).
Analyses by producers have detected no benzene at the 10-ppm level in
the final maleic anhydride products produced by the benzene process. '*
Although the process just described and illustrated in Figure 1-2
typifies the benzene oxidation process, variations exist. In place of the
partial condensation system (cooler, partial condenser, and separator)
1-7
-------
i
00
AIR
01 I « —•
COMPRESSOR
(11—•
BENZENE
STORAGE
, STEAM
~<£H
VAPORIZER
STEAM
_ x
INTERCHANGED
mRTIAL
dt " "* ' r^v\ I mw i f I
REACTORIS) CONDENSER/|CRUDE MAVj PRODUC
-T VJ STORAGE \) RECOVEI
SPENT CATALYST ' ABSORB
I—M.
f==* XYLENE [-1- S
—,*A A L—J
DEHYDRATION
COLUMN
^&
tJ ^t
T I
XYLENE
STORAGE
XYLENE
STRIPPER
i
•»-
rnwuuv" i
RECOVERY T _
ABSORBER
MAKEUP
mMrvcur
" fit WATER
3>
® EXCESS
*" WATER
FRACTION AT ION
COLUMN
i
AGED ANHYDRIDE
STORAGE
^X
PRODUCT
*^
K
E
Y
A-
B —
C-
K -
Product recovery absorber vent
Vacuum system vent
Storage and handling emissions
Secondary emission potential
Figure 1-2. Process flow diagram for uncontrolled model plant.
-------
shown in Figure 1-2, a so-called switch condensing system can be incorpo-
rated. This system uses a series of condensers that are alternately cooled
to freeze maleic anhydride on the surface and heated to melt the maleic
anhydride for pumping to crude maleic anhydride storage. Switch condensing
can remove up to 60 percent of the maleic anhydride from the process com-
pared to 40 percent for the partial condensation system. Removal of this
additional maleic anhydride would allow the size of the product recovery
absorber to be reduced and would slightly reduce the maleic acid loss
through the product recovery absorber (Vent A).
Xylene is the only known azeotropic agent currently used for dehydra-
tion, although several other agents can be used, including isoamyl butyrate,
di-isobutyl ketone, anisole, and cumene. A vacuum evaporation system,
which replaces the dehydration column and xylene stripper, is used by at
least one plant to remove water and dehydrate the maleic acid to form
15
maleic anhydride. Because an azeotropic agent is not required in that
case, xylene is eliminated as a process emission.
The emission rates for the benzene oxidation process are based on a
model plant with a capacity of 22,700 Mg/yr (50 x 10 Ib/yr), assuming
8,000 hours of annual operation and 94.5 percent conversion of benzene.
Though not an actual plant, it is typical of most plants in capacity and
operating range. The model benzene oxidation process, shown in Figure 1-2,
best fits today's maleic anhydride manufacturing and engineering technology.
Single-process trains as shown are typical for the large plants except in
the reaction area where multiple reactors are common. The model process
uses partial condensation and azeotropic distillation with xylene. The
emission rates and sources for the benzene oxidation process are summarized
in Table 1-3 and represent compilation of data from several plants.
Emission rates depend on conversion. Conversion is the ratio of the
amounts of raw material reacted to the amount of raw material fed to the
reactor—in this case, benzene reacted to benzene fed. If conversion were
100 percent, there would be no benzene emissions from the process vent
because all of the benzene would be converted to something else. The exact
conversion rate is important. A maleic anhydride plant operating at a
95-percent benzene conversion would require a control system 80 percent
efficient in order to emit the same amount of benzene as an identical,
uncontrolled plant operating at 99 percent benzene conversion.
1-9
-------
TABLE 1-3. BENZENE AND TOTAL VOLATILE ORGANIC COMPOUND (VQC) EMISSIONS FROM
MODEL UNCONTROLLED MALEIC ANHYDRIDE PLANT9
Stream
designation
Source Figure 1-2
Product recovery A
absorber
Refining vacuum system B
kg (lb) of
benzene per Mg
(1,000 lb) of
MA produced
67.0 (67.0)
—
Total kg (lb)
of VOC per Mg
(1,000 lb) of
MA produced
86.0 (86.0)
0.1 (0.1)
Benzene
emitted
kg/hr
(Ib/hr)
190 (418)
—
Total VOC
emitted
kg/hr
(Ib/hr)
244 (537)
0.28 (0.62)
aEmission rates are annual averages based on 8,000 hr/yr of operation, 22,700-Mg/yr (50 x 10 Ib/yr)
capacity, 94.5% conversion.
SOURCES: Morse, P. L. Maleic Anhydride. In: Process Economic Program. Menlo Park, Stanford Research
Inst. November 1973.
Lawson, J. F. Trip Report for Visit to Reichhold Chemicals, Inc., Morris, Illinois,
July 28, 1977. Hydroscience, Inc. EPA Contract Number 68-02-2577.
V
Pervier, J. W., et al. Survey Reports on Atmospheric Emissions from the Petrochemical
Industry, Volume III. Houdry Division of Air Products, Inc. (Prepared for Office of
Air Quality Planning and Standards, U.S. Environmental Protection Agency. Research
Triangle Park, N.C.) EPA-450/3-73-005-C. April 1974.
-------
The highest possible benzene conversion is not necessarily optimum
because not all the benzene reacted forms maleic anhydride. In fact, at
high conversion rates, the yield (ratio of product formed to reactant
consumed) is often lower than at lower conversions. Thus, production often
increases with operation at a lower conversion; i.e., that which best
balances conversion and yield.
The composition of the waste stream from the product recovery absorber-
the location of the main process vent--of a model plant is shown in Table
1-4. All plants have this vent. The benzene-to-air ratio influences the
concentrations of emissions from this vent. Excess air must be fed to the
reactor to maintain the benzene concentration below its lower explosive
limit.
Some types of process upsets result in more benzene being released
because the product recovery absorber can only remove benzene up to its
solubility level in water or aqueous maleic acid. These upsets can result
in short-duration benzene and volatile organic compound (VOC) emissions of
three to five times normal. Process startup at an uncontrolled plant also
results in temporary benzene emissions three to five times the normal rate
because the benzene does not react completely until proper catalyst temper-
ature is reached and, again, absorption is only effective up to the limits
of solubility.
The refining vacuum system vent (Vent B, Figure 2) exhausts the non-
condensibles from the three vacuum columns used to dehydrate and frac-
tionate maleic anhydride. The VOC emissions--maleic acid, xylene, and
possibly benzene—are estimated to be relatively small, as indicated in
Table 1-3. Process upsets, startups, and shutdowns do not affect the VOC
emissions from this vent.
Table 1-5 summarizes the estimated benzene emissions from maleic
anhydride plants for uncontrolled and existing control conditions.
Descriptions of fugitive and storage benzene emissions are to be
covered in another document.
1.2.2 n-Butane Oxidation Process
Little information in the open literature is available on maleic anhy-
dride production by the oxidation of n-butane, as currently practiced in
the United States. Thus, it is particularly difficult to assess the effi-
1-11
-------
TABLE 1-4. WASTE GAS COMPOSITION—PRODUCT RECOVERY ABSORBER
Component Weight % kg/hr (Ib/hr)
(weighted average)
Oxygen
Nitrogen
Carbon dioxide
Carbon monoxide
Water
Benzene
Maleic acid
Formaldehyde
Formic acid
16.67
73.37
3.33
2.33
4.00
0.23
0.01
0.05
0.01
13,800 (30,360)
60,740 (133,628)
2,757 (6,065)
1,929 (4,244)
3,312 (7,286)
190 (418)
8 (18)
41 (90)
8 (18)
82,785 (182,127)
SOURCE: Lawson, J. F. Emission Control Options for the Synthetic Organic
Chemicals Manufacturing Industry Maleic Anhydride—Product
Report. Hydroscience, Inc. (Prepared for Office of Air Quality
Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, N.C.) EPA Contract Number 68-02-2577.
March 1978.
1-12
-------
TABLE 1-5. BENZENE EMISSIONS FROM MALEIC ANHYDRIDE PLANTS'
Ashland, Neal
W. Va.
Koppers, Bridge-
ville, Pa.
Monsanto,
St. Louis, Mo.
DENKA,
Houston, Tex.
Reichhold,
Elizabeth, N.J.
Reichhold,
Morris, 111.
Tenneco
Fords, N.J.
U.S. Steel,
Neville Island,
Pa.
Totals
Capacity
(Mg/yr)
27,200
15,400
38,100
22,700
13,600
20,000
11,800
38,500
Total
uncon-
trolled
emis-
sions2
(Mg/yr)
1,849
1,047
2,589
1,543
924
1,359
802
2,616
12,729
Existing
control
device
efficien-
cy (%)
0
99
0
97
97
90
0
90
Emissions
with
existing
control 3
(Mg/yr)
1,849
26
2,589
68
44
154
802
295
5,824
Emission rates are based on a benzene conversion rate of 94.5 percent and
on: Lawson, J. F. Emission Control Options for the Synthetic Organic
Chemicals Manufacturing Industry Maleic Anhydride--Product Report. Hydro-
science, Inc. (Prepared for Office of Air Quality Planning and Standards,
U.S. Environmental Protection Agency, Research Triangle Park, N.C.) EPA
Contract Number 68-02-2577. March 1978.
2Total
uncontrolled
emissions
Emissions with
existing
control
[(2.6 kg/hr x 8,760 hr)
+ (190 kg/hr x 8,000 hr)]
x production capacity Mg/yr .
model plant capacity 22,700 Mg/yr
{(2.6 kg/hr x 8,760 hr)
+ [190 kg/hr x (1-control device efficiency)
x 8,000 hr]} x
production capacity Mg/yr
model plant capacity 22,700 Mg/yr
1-13
-------
ciency of this process relative to the conventional benzene oxidation
process. General information on n-butane oxidation is summarized below,
although it may not accurately represent the process as practiced by Mon-
santo and Amoco.
The overall chemical reaction of interest is:
0
II
CH,- CH9 - CH9 - CH, + 7 Q0 H-C-C\ + 4H00
•J ^ £ 3 "n f. ^, II \.n t-
£ ~"^^^ I I jS^J
^^^
II
0
n-BUTANE OXYGEN MALEIC WATER
ANHYDRIDE
The reaction involves both dehydrogenation and oxidation. The actual
catalysts used today have not been reported, although several possibilities
are listed in Hydrocarbon Processing. One is a mixture of iron and vanadium
pentoxide-phosphorous pentoxide on a silica-alumina support with unspeci-
fied proportions. The reaction proceeds at 500° C (932° F) at atmospheric
pressure and results in a yield of 14.2 mole percent. Another scheme uses
a cerium chloride, cobalt-molybdenum oxide catalyst supported on silica.
The cerium chloride dehydrogenates the n-butane to butene, after which the
cobalt-molybdenum oxide catalyst oxidizes the butene to maleic anhydride.
The reaction is carried out at 490° C (914° F), and a yield of 63 weight
percent is reported. A third catalyst system has been reported using
phosphorous and vanadium with yields of 55 to 60 percent. The actual
catalyst used by the two domestic producers (Monsanto and Amoco) of maleic
anhydride from n-butane has not been reported. Catalyst development re-
search is actively continuing, and the catalyst technology is a closely
held secret.
The DENKA plant was originally designed to use the cis- and trans-
isomers of 2-butene as a feedstock. This feedstock was used for 4 years
before being replaced by the more economical benzene process. Later, OENKA
experimented with n-butane for the life of one catalyst--6 to 9 months—and
18
encountered problems with catalyst stability.
In 1979, DENKA and the Badger Company announced plans to build and
operate a demonstration plant at DENKA's facility to use n-butane. The
1-14
-------
process uses a fluidized bed reactor. Badger reports high yields and low
utilities consumption, based on laboratory and model plant tests.
A simplified flow diagram from one producer indicates that the n-butane
oxidation process is similar to the benzene oxidation process shown previ-
ously in Figure 1-2. One obvious, major difference is in the raw material
storage facilities, where n-butane requires additional safety and property
19
protection safeguards. Because n-butane is in the gaseous state at
standard conditions, it will normally be stored as a liquid in pressurized
vessels, resulting in additional operating costs. Like the benzene process,
product recovery is by partial condensation, product absorption, dehydration,
20
and fractionation. Further indirect evidence supports the assumption
that the n-butane process is analogous to the benzene process. First,
Ashland Chemical, which started up a new benzene-based plant in 1976,
states it is able to convert to n-butane feed simply with a change in
21
catalyst. Second, it has been reported that the switch to n-butane feed
can be accomplished at a relatively small fraction of the cost of a new
21
plant. Third, the benzene oxidation process can be converted to n-butane
oxidation by changing the catalyst system, and conversion can be accomplished
21
for much less than the cost of a new plant. This implies that, for the
most part, the same unit operations are performed after the reaction module
and in equipment of similar design and capacity.
On the other hand, there is evidence that some process differences
exist between the benzene and the n-butane process. The n-butane process
is said to require a longer reactor residence time and, therefore, bigger
22
reactors. This means that a higher capital investment is required for
the n-butane process to achieve the same production output, although there
are few data upon which to estimate the increase in capital cost. One
source estimates the capital costs for an n-butane plant could be 10 to 20
23
percent higher than for a comparable, benzene-based plant. Also, produc-
tion capacity would decrease for an existing benzene-based plant, if it
were changed to use n-butane as a feedstock, unless the reactors were
replaced with larger equipment. An additional problem is that the carbon
steel reactors used in the oxidation of benzene may not withstand the
higher temperatures of n-butane oxidation. Furthermore, the composition
of the stream from the reactors is likely to differ in the n-butane process,
1-15
-------
op
even though many of the same compounds will be present. For this reason,
variations in the design and operation of the product recovery and refining
19
steps would be expected for a new plant based on n-butane. The impact of
this variation on the efficiency of a converted facility is not well known
but could be a significant problem. According to Chemical Week:
"One producer notes that the advisability of building new n-butane-
based capacity versus converting old benzene-based capacity to n-butane
is 'one of these questions not yet answered1."21
A switch in feeds for an existing facility would require a major investment
8 21
in new catalysts; ' for plants with an activated carbon adsorption unit
attached to the product recovery absorber, additional modification costs
19
would be incurred. Moreover, the overall yield in converted plants would
probably not be as high as in new plants designed specifically for n-butane;
this lowered efficiency might be serious enough to warrant other major
21 22
design changes. '
1.2.3 Byproduct of Phthalic Anhydride Production
Phthalic anhydride is manufactured from naphthalene and orthoxylene.
Maleic anhydride is recovered as a byproduct from the plant effluent. The
emissions associated with maleic anhydride recovery are believed to be
insignificant and are not being investigated at this time. There are no
benzene emissions from this recovery process. In addition, no growth in
the production of maleic anhydride by this process is expected.
1.2.4 Foreign Process
The only significant foreign process that differs from processes used
in the United States is a process using feedstocks of 65 to 80 percent
butenes with the remainder mostly n-butanes or isobutene (mixed C. oxida-
tion). The general process description is similar to that shown in Figure
o
1-2 for benzene oxidation. The exhaust from the main process vent contains
unreacted n-butane, butene, carbon monoxide, and various secondary products.
Except for the possible absence of benzene, the VOC emissions should be
about the same as for the benzene oxidation process.
The most significant process variation—the use of a fluidized cata-
lyst bed rather than a fixed bed—was developed for the mixed C^ process.
This variation provides good temperature control within the bed, allowing
optimum ratios of feed to air. In contrast, optimum feed-to-air ratios
1-16
-------
cannot be used with fixed bed systems because temperature cannot be precise-
ly controlled; therefore, excess air becomes necessary to stay below the
explosive range. The reduction of excess air in the fluidized bed feed
will reduce emissions from the product recovery absorber. However, this
advantage is offset by the lower product yields obtained with the fluidized
4
bed than with the fixed bed.
1.3 SUMMARY
This chapter has described the processes and associated emissions for
four routes to the production of maleic anhydride. The benzene oxidation
process is emphasized because it is the only one known to emit benzene.
Two process emission points have been discussed: the product recovery
absorber vent, and the refining vacuum system vent. Only the process
emissions are discussed; emissions of benzene from fugitive and storage
sources are to be addressed in another document that discusses these emis-
sions from other chemical industries and petroleum refineries.
The n-butane oxidation process is discussed to the extent that avail-
able data permitted; it may replace benzene oxidation in the future.
1.4 REFERENCES
1. Blackford, J. C. CEH Marketing Research Report on Maleic Acid. In:
Chemical Economics Handbook. Menlo Park, Stanford Research Inst.
July 1976.
2. Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency. Unpublished EPA Survey and Ranking Information.
Research Triangle Park, N.C. 1977.
3. Koppers Closing Maleic Plant, Firm Cites Outdated Technology. Chemical
Marketing Reporter. 215:3, 11. April 9, 1979.
4. Mitsubishi Chemical Details its C4-based Maleic Process. European
Chemical News. April 5, 1974.
5. Checkoff. Chemical and Engineering News. 57(40):13. October 1,
1979.
6. Neunreiter, R. L. Florida Site Selected by Monsanto for New Maleic
Anhydride Plant (press release). Monsanto Chemical Intermediates Co.
St. Louis, Mo. August 16, 1979.
7. Chementator. Chemical Engineering. 86(23):81. October 22, 1979.
8. Lenz, D., and M. De Boville. The Bayer Process for the Production of
Maleic Anhydride from Butenes. Rev Assoc Fr Tech Pet. 236(20-3):17.
1976.
1-17
-------
9. Telecon. Weber, Robert, Office of Air Quality Planning and Standards,
U.S. Environmental Protection Agency, with Pierle, Michael, Monsanto
Chemical Intermediates Co. January 18, 1979.
10. Lawson, J. F. Emission Control Options for the Synthetic Organic
Chemicals Manufacturing Industry Maleic Anhydride—Product Report.
Hydroscience, Inc. (Prepared for Office of Air Quality Planning and
Standards, U.S. Environmental Protection Agency, Research Triangle
Park, N.C.) EPA Contract Number 68-02-2577. March 1978.
11. Lewis, W. A., Jr., G. M. Rinaldi, and T. W. Hughes. Source Assessment:
Maleic Anhydride Manufacture. Monsanto Research Corp. EPA Contract
Number 68-02-1874.
12. Telecon. Warren, J. L., Research Triangle Inst., with Pierle, M. A.,
Monsanto Chemical Intermediates Co. November 16, 1978.
13. Telecon. Warren, J. L., Research Triangle Inst., with Hewett, S.,
Reichhold Chemicals, Inc. November 20, 1978.
14. Telecon. Warren, J. L., Research Triangle Inst., with Brennan, H.,
Amoco Chemicals Corp. November 17, 1978.
15. Lawson, J. F. Trip Report for Visit to Reichhold Chemicals, Inc.,
Morris, Illinois, July 28, 1977. Hydroscience, Inc. EPA Contract
Number 68-02-2577.
*
16. Pervier, J. W., et al. Survey Reports on Atmospheric Emissions from
the Petrochemical Industry, Volume III. Houdry Division of Air Pro-
ducts, Inc. (Prepared for Office of Air Quality Planning and Stand-
ards. U.S. Environmental Protection Agency. Research Triangle Park,
N.C.) EPA-450/3-73-005-C. April 1974.
17. Hatch, Lewis F., and Matar, Sami. Hydrocarbons to Petrochemicals...
Part 7 Petrochemicals from n-Paraffins. Hydrocarbon Processing.
56(11):355. November 1977.
18. Letter from Babb, K. H., Research Triangle Inst., to Hinkson, R. E. ,
DENKA Chemical Corp. September 7, 1979.
19. Letter from Pierle, M. A., Monsanto Chemical Intermediates Co., to
Patrick, D. R., Office of Air Planning and Standards, U.S. Environ-
mental Protection Agency. March 22, 1978.
20. Lawson, J. F Trip Report for Visit to Amoco Chemicals Corp., Chicago,
Illinois, January 24, 1978. Hydroscience, Inc. EPA Contract Number
68-02-2577.
21. Maleic Makers Build on Hopes for Polyester. Chemical Week. 120(5):
37-38. February 2, 1977.
22. Maleic Builds New Bridges to Feedstocks. Chemical Week. 119(15):79.
October 13, 1976.
1-18
-------
23. Letter from Weber, Robert, Office of Air Quality Planning and Standards,
U.S. Environmental Protection Agency, to Gans, M., Scientific Design
Co., Inc. December 27, 1978.
24. Letter from Pierle, M. A., Monsanto Chemical Intermediates Co., to
Goodwin, D. R., Office of Air Quality Planning and Standards, U.S.
Environmental Protection Agency. February 22, 1978.
1-19
-------
1-20
-------
2. EMISSION CONTROL TECHNIQUES
2.1 ALTERNATIVE EMISSION CONTROL TECHNIQUES
This chapter addresses the control techniques that can be applied and
the emission reductions that can be achieved for each benzene source identi-
fied in Chapter 1.
The control techniques that can be applied to reduce benzene emissions
from maleic anhydride manufacturing facilities fall into two classes:
Add-on control devices such as adsorbers or incinerators, and
Use of another feedstock to eliminate benzene from the process.
For add-on control devices, it is assumed that the product recovery
absorber vent and the refining vacuum system vent can be controlled by the
same device. Only piping additions are required for controlling the refin-
ing vacuum system along with the product recovery absorber; no added utili-
ties, manpower, or other operating costs are involved. Emissions from the
refining vacuum system vent are included in all control system calcula-
tions. The control devices described here are applicable to the waste gas
streams defined in Chapter 1. Furthermore, add-on control devices are
expected to control the process vents during process upset conditions and
startup (as discussed in Chapter 1) with no increase in emissions.
Because of the low concentrations of benzene and other volatile organic
compounds (VOC) in the emissions from the main process vent (see Table
1-4), the use of flares is not generally believed to be an ^plicable
control technique. The high-volume, low-benzene concentration waste
stream would require large amounts of auxiliary fuel without the opportunity
2
for heat recovery.
2.1.1 Carbon Adsorption
Carbon adsorption systems have often been applied in other industries
to control waste gas streams carrying VOC for similar applications; e.g.,
with inlet organic concentrations ranging from 1 ppm to 40 volume percent
2-1
-------
and at air flow rates of 0.005 m3/sec (11 cfm) to over 100 m3/sec
3
(212,000 cfm). General discussions of the theory and application of
A C r
carbon adsorption can be found elsewhere. '
Adsorption onto activated carbon can be used to recover benzene from
gas streams from the product recovery absorber and refining vacuum system
vent. Two plants currently use a carbon adsorber system to control the
effluent from the product recovery absorber. To use carbon adsorption in
this application, the exhaust gas stream may be scrubbed with a caustic
solution to remove organic acids and water-soluble organics. Benzene is
probably the only VOC remaining in appreciable quantity after scrubbing.
The waste stream is conditioned by reducing the relative humidity, which
improves loading capacity.
Various levels of control can be achieved with carbon adsorption,
depending on the design and operation of the adsorber system. In general,
99 percent or greater reduction of hydrocarbon emissions, including benzene,
can be obtained. Factors influencing the efficiency of carbon adsorption
systems for benzene control in maleic anhydride plants include:
Relative humidity of the incoming waste gas stream;
Presence of other organic compounds that may interfere with
benzene adsorption (possibly forming polymeric materials on the
carbon beds) thereby decreasing capacity;
Temperature of the beds and the gas during adsorption;
Efficiency of the steam regeneration; and
8 9
Dryness of the bed when put back on line. '
Both a caustic scrubber and a heater should be included in a carbon adsorp-
tion system, the caustic scrubber to remove most of the other organics in
the stream and the preheater to lower the relative humidity of the water-
saturated stream. Generally, a system of two or more carbon beds is used.
The gas stream containing benzene passes through one or more beds in par-
allel, and the benzene is removed from the gas stream. At the same time,
another bed is regenerated with low-pressure steam. The steam and desorbed
benzene are condensed and decanted, after which the benzene returns to the
process. The aqueous layer can be combined with the other liquid waste
from the plant or recycled to the process. After regeneration, the carbon
bed, which is hot and saturated with water, is usually cooled and dried,
often by blowing ambient air through the bed. Bed size, number, and cycle
2-2
-------
times can be varied to achieve the desired removal efficiency. Finally,
because the system may be exposed to corrosive compounds, stainless steel
vessels are recommended. Specific systems that can achieve various removal
efficiencies are discussed in Section 2.2.1. A carbon adsorption system,
however, will not remove carbon monoxide, which is also present in the
waste gas from the product recovery absorber.
2.1.2 Thermal Incineration
Thermal incineration, also called direct-fire incineration, can be
used to control emissions from maleic anhydride manufacture. Three plants
in the United States use a thermal incinerator on the product recovery
absorber vent; one uses n-butane, while the other two use benzene as the
feedstock. ' General information on thermal incineration can be found
4 8
elsewhere. ' Because of the cost of the fuel required to operate a ther-
mal incinerator, heat recovery is generally used. The recovered heat can
be used either to preheat the feed to the incinerator or to generate steam.
Some of the factors that influence the efficiency of incineration are
temperature, degree of mixing, and residence time in the combustion chamber.
For maleic anhydride plants, a knockout demister tank is required to
prevent liquid droplets from reaching the burner area. Supplemental fuel
is required to maintain necessary combustion temperatures, and supplemental
combustion air may also be required if the incoming gas stream is not
preheated. Because the gas stream to the incinerator contains corrosive
materials, the equipment ahead of the combustion chamber must be made of
stainless steel. Specific incineration systems and their removal efficien-
cies are described in Section 2.2.2.
The waste gas to be incinerated is approximately 2.5 mole percent
fuel, which is below the combustible level (see Table 1-4). Thus, auto-
ignition can only occur if the lower flammability limit of the mixture is
2.5 mole percent or less. Based on LeChatelier's Rule and the modified
Burgess-Wheeler Law, the lower flammability levels of the waste gas mixture
are 6.4 mole percent at 487° C (908° F) and 3.2 mole percent at 871° C
(1,600° F). Consequently, autoignition of the waste gas does not occur
even at the design temperature of the incinerator.
Concern has been expressed that benzene emissions may leak through
recuperative heat exchangers and incinerators. Except for rotary exchangers,
2-3
-------
heat exchangers generally do not leak. However, emissions are routinely
prevented by maintaining the clean gas at a higher pressure than the side
contaminated with benzene emissions. This higher pressure causes leakage
into the feed stream to the incinerator rather than into the atmosphere.
Shell and tube exchangers are not expected to leak. If routine maintenance
is inadequate or operation is improper, leakage can occur because of rup-
tured tubes or warped flanges. However, this leakage is not caused by
1?
intrinsic equipment limitations.
2.1.3 Catalytic Incineration
Catalytic incineration can be used as an alternative to thermal inciner-
ation of a waste gas stream. The catalyst allows oxidation to occur more
rapidly at a lower temperature, thereby decreasing or eliminating the need
for supplemental fuel consumption. A catalytic incinerator is currently
used by one maleic anhydride producer to meet local hydrocarbon regulations.
The unit removes at least 85 percent of the total hydrocarbon content and
95 percent of the CO content of the product recovery absorber vent gas
without additional fuel.
Theoretically, a catalytic incinerator could be designed to have a
o
removal efficiency of 90 to 95 percent. In practice, however, no such
system is in operation on a maleic anhydride production facility because
cost may be prohibitive. Achieving high benzene removal efficiencies typi-
cally requires large catalyst volumes or temperatures close to those of
o
thermal incineration, so a catalytic converter is believed to be uneconomical.
In addition, depending on the nature of the catalyst used, fouling of
the catalyst may reduce its life. This may occur in maleic anhydride
plants because some of the components of the waste gas stream may polym-
o
erize. One catalyst producer believes this problem can be overcome by
using a monolithic support system with platinum catalysts and periodic
maintenance. Precious metal catalysts are not deactivated by water. The
deposition of high-boiling organics on the catalyst's surface merely masks
that surface. Periodic maintenance is often all that is required to alle-
14
viate this problem.
2.1.4 n-Butane Process Conversion
The n-butane oxidation process has the potential for zero benzene
emissions, permitting conversion from benzene feed to n-butane feed to be
2-4
-------
considered a control technique for reducing benzene emissions from the
manufacture of maleic anhydride. The process, emissions, and factors in-
fluencing conversion from benzene to n-butane are discussed in Chapter 1.
One company using n-butane has detected no benzene emissions from its
15
reactors at the lower detection limits of 1 ppmv. In addition, no benzene
has been detected at the 0.5-ppm level in the final product from the n-butane
process. Problems associated with conversion to n-butane were discussed
in Section 1.2.2.
2.2 PERFORMANCE OF EMISSION CONTROL TECHNIQUES
2.2.1 Carbon Adsorption
Based on engineering experience with similar applications for the
control of VOC, a carbon adsorption system can be designed and operated at
a sustained benzene removal efficiency of 99 percent. This efficiency
reflects optimal control of temperature, pressure, humidity, and the level
of other organics. Carbon adsorption systems operating at different levels
of control are described in this section. The first system, at 99 percent
control of benzene, is based on engineering design calculations and previous
engineering experience. The other systems described here are based on the
experience of existing maleic anhydride facilities, one system reported by
the company to achieve an average of only 85 percent control of benzene,
while the efficiency of the other has not been determined.
As discussed in Section 2.1.1, several key factors influence the
efficiency of the adsorption system, and optimum control of these factors
is responsible for 99 percent reduction of benzene emissions. The carbon
adsorption system uses a caustic scrubber. A heater is used before the
carbon beds to decrease the relative humidity.
The size of the activated carbon adsorbers is determined by the super-
ficial velocity through the bed and the carbon requirements, as determined
by the flow of adsorbable species and the carbon loading for each species.
The superficial velocity is important because it affects the pressure drop
through the bed. The usually acceptable range for superficial velocity is
25 to 50 cm/sec (10 to 20 in/sec), which gives a pressure drop of 25 to
65 cm H90/m (3 to 8 in H90/ft) bed. The air flow rate to the carbon adsorber
3 4
system is approximately 21 m /sec (4.5 x ion cfm). Assuming a superficial
velocity of 38 cm/sec (15 in/sec), the required cross-sectional area is
2-5
-------
2 2
about 55 m (590 ft ). Because adsorbers this large are usually horizontal
tanks, two tanks 3 m (10 ft) diameter x 9 m (30 ft) long operated in parallel
would give the required cross-sectional area.
To allow enough time for regeneration, additional adsorbers are neces-
sary. To ensure that 99 percent removal could be achieved for the system,
which is designed for comparison and costing purposes, two additional
adsorbers were chosen instead of one. Also, experience from one adsorber
system in this application suggests there is not sufficient time to com-
plete the regeneration cycle (including the cooling and drying cycle) and
thereby prevent breakthrough when only three beds are used.
The amount of carbon in the adsorber is determined by the desired
cycle time, flow rate of adsorbable species, and carbon loading. Although
shorter cycle times have been used, a 2-hour loading cycle was chosen
initially to allow a 1-hour steaming cycle and another hour to cool and dry
the carbon in preparation for adsorption service. The amount of carbon per
adsorber can be calculated from emission rate, loading, and bulk density:
kg (Ib) carbon _ 190 kg (418 1b) Bz x 2 hr
bed cycle hr cycle
x 1 kg (2.2 Ib) carbon x 1
0.06 kg (0713 Ib) Bz 2 beds
= 3>200 ^ <7>000
3 3
Volume of carbon = 3,200 kg (7,000 Ib) x 44g kffigsVs Ib)
= 7.1 m3 (250 ft3) .
The depth of the carbon required to give a 2-hour loading cycle is 25 cm
(10 in) in the adsorbers selected. Thus, in this case, the size of the
adsorber is dictated more by the total flow rate than by the carbon require-
ments. Because the incremental cost of putting more carbon in the adsorbers
is small, a carbon depth of 61 cm was used. Each adsorber would now hold
17 m3 (600 ft3) or 7,600 kg (16,800 Ib) of carbon. Using the same carbon
loading and benzene flow as before now gives an adsorption cycle of 4.8
hours, which should allow enough adsorption capacity to prevent premature
breakthrough while the companion bed is being regenerated.
2-6
-------
A loading of 0.06 kg benzene (Bz) per kilogram (0.06 Ib Bz/lb) of
carbon has been selected based on capacity data reported by Hoyt Manufactur-
ing (a carbon adsorber vendor). These capacity data are also similar to
Reichhold's experience with a benzene adsorber in this application. The
selected loading is conservative because other data in the literature show
benzene loadings as high as 0.25 kg Bz/kg (0.25 Ib Bz/lb) of carbon on
19
successive cycles for full-scale systems adsorbing benzene from town gas.
If higher loadings are achieved, the loading cycle would be extended,
allowing still more time for regeneration, cooling, drying, and lower
operating costs.
There is some indication that the relative humidity of the vent stream
can affect the loading capacity. Hydroscience has observed a 75- to 80-per-
cent drop in loading for some materials at a relative humidity of 80 per-
9 20
cent as compared to 20 percent relative humidity. ' Similar data for
activated carbon are reported by Rohm and Haas in their technical bulletin
21
on Ambersorb Carbonaceous Resins. It is important to note, however, that
changes were noted only in the loading capacity and not in the ability to
9
achieve a baseline outlet concentration of less than 5 ppmv.
To regenerate the carbon beds 20 kg steam/kg Bz (20 Ib steam/1b Bz)
are required. This amount should be more than adequate to complete the
desorption because steam requirements as low as 3 kg steam/kg Bz (3 Ib
22
steam/lb Bz) toluene have been quoted in the literature. Some flexibil-
ity is available when the duration of the steaming cycle is determined;
steaming can easily be completed in 1 hour or less if the condensing capacity
exists. For a new installation, it is best to size the condenser to handle
the steam flow equivalent for at least a 1-hour steam cycle as an additional
safety factor. The average steam flow (for calculating annualized costs)
of 1.1 kg/sec (2.4 Ib/sec) is determined from the average benzene flow and
the steam-to-benzene ratio.
An important part of the total operation is the cooling and drying
sequence. If the bed is not properly cooled and dried, both poor efficiency
and lower adsorption capacities will result. Note that this result con-
trasts with the previously discussed effect of relative humidity of the
inlet gas on loading capacity and efficiency. The full-scale experience
with benzene adsorbers at one facility where cooling and drying are not
practiced shows that there was an initial spike of benzene in the outlet
immediately after the hot bed was put back on line. In addition to this
2-7
-------
initial spike, the instantaneous removal was rarely better than 97 percent.
These data are consistent with the observations that poor efficiency and
reduced loading capacity occur when the adsorber is not properly cooled and
dried following a steam regeneration. In order to achieve greater than 99
percent removal, the bed must be cooled and dried. In addition, the exhaust
from the cooling and drying cycle must be recycled to a carbon bed that is
in the adsorption mode. This means that the adsorbers must be sized so
there is capacity to handle the additional benzene and air flow as well as
time to complete the cooling and drying cycle. The data from one facility
show that the cooling and drying period lasted for about the first 15 minutes
of waste gas incineration and that about 1 to 3 percent of the benzene in
the feed left the system during this period. If the worst case of 3
percent is used, the calculated adsorber duty would be about 197 kg/hr
(433 Ib/hr) instead of 190 kg/hr (418 Ib/hr). The calculated time to
breakthrough (at a loading of 0.06 kg Bz/kg carbon [0.06 kg Bz/kg]) would
only be reduced from 4.8 hours to 4.65 hours.
The effect of total flow rate on efficiency and pressure drop through
the on-line adsorber is more critical than the recycle of some adsorbed
benzene since outside air must be used for drying. The vent itself is not
suitable for drying (even though the relative humidity of the vent down-
stream of the scrubber is reduced from 100 percent relative humidity to
g
about 50 percent relative humidity by the preheater). For design purposes,
a cooling and drying cycle of I hour at a flow rate equivalent to one-fourth
to one-half of the flow rate during the adsorbing cycle was chosen. This
would raise the superficial velocity for the adsorbing carbon bed from 38
cm/sec (15 in/sec) to 49 to 57 cm/sec (19 to 22 in/sec) during the cooling
and drying cycle but would not create an excessive pressure drop. Because
two beds could be cooling at once, the blower has been sized to deliver
21 m /sec (4.5 x 10 cfm). A capital savings could be achieved if a blower
were sized at 11 m /sec (2.3 x 10 cfm). However, the additional capital
cost of the larger blower is considered justified to make the operation of
the adsorption system more flexible. Pertinent details of the system
designed to achieve 99 percent control of benzene are in Table 2-1, and
costs for such systems are presented for each plant in Chapter 5.
As indicated previously, two plants currently use carbon adsorption
systems to control benzene from product recovery absorber vents. One plant
2-8
-------
TABLE 2-1.
TECHNICAL DATA-CARBON ADSORPTION SYSTEM
(99 Percent Control)
Number of beds
Weight of carbon per bed
Loading cycle time
Number of stacks
Benzene emission rate per stack
Air flow rate from system (per stack)
7,620 kg (16,800 Ib)
2 hr
2
0.26 g/sec (5.8 x lo"4 Ib/sec)
18.0 m3/sec (3.8 x 104 cfm)
NOTE: References and basis explained in text.
TABLE 2-2.
TECHNICAL DATA—CARBON ADSORPTION SYSTEM
(85 Percent Control)
Number of beds
Loading cycle time
Regeneration cycle time
Number of stacks
Air flow rate from system4
3
2 hr
1 hr
1
20.3 m3/sec (4.3 x lo4 cfm)
At capacity.
SOURCE: Lawson, J. F. Trip Report for Visit to Reichhold Chemicals, Inc.,
Morris, Illinois, July 28, 1977. Hydroscience, Inc. EPA Contract
Number 68-02-2577.
2-9
-------
for which data are available reports that it has achieved a sustained
benzene removal efficiency of only 85 percent. Results of tests conducted
by the U.S. Environmental Protection Agency (EPA) showed a mean benzene
removal efficiency of 93 percent (with a 90-percent confidence interval of
23
89 to 97 percent). However, the plant was operating at only 40 percent
of full capacity when tested. Pertinent technical data on that system are
summarized in Table 2-2. This system does not use an organic-free gas
stream to cool and dry the beds after regeneration with steam. Immediately
after regeneration, the waste gas stream containing benzene is directed to
the hot bed. Consequently, until the bed cools benzene removal efficiency
is low, which partially accounts for the low overall benzene removal effi-
ciency.
The system at the second plant was designed for a benzene removal
efficiency of 98 percent, although no reliable data are currently available
upon which to estimate its performance. The system can be expected to show
better performance than the one discussed previously because the beds are
24
cooled and dried with air after steam desorption. Pertinent technical
data on this system are in Table 2-3.
2.2.2 Thermal Incineration
Based on engineering experience with similar applications for the
control of VOC's, it is expected that a thermal incinerator can be designed
and operated at a sustained benzene removal efficiency of 99 percent.
Limited information is available on direct-flame afterburners used on
maleic anhydride production facilities, but there are several cases in
which streams similar to the product recovery absorber and refining vacuum
system vent gas have been controlled at high efficiencies. In one case,
data reported for toluene indicate a destruction efficiency of 99.9 percent
at 766° C (1,411° F) and a residence time of 0.21 second for toluene.25
Another facility incinerates a toluene-xylene vapor at 760° C (1,400° F)
26
and reportedly achieves a destruction efficiency of 99.1 percent. A
third installation reportedly expects a destruction efficiency of greater
than 99 percent at 760° C (1,400° F) for an organic stream considered as
26
toluene. Finally, a review of several case studies indicates that combus-
tion efficiencies of less than 95 percent were achieved, except in one
case, at temperatures of 730° C (1,346° F) or lower. Conversely, efficien-
2-10
-------
TABLE 2-3. TECHNICAL DATA—CARBON ADSORPTION SYSTEM
(98 Percent Control at Design)
Number of beds 2
Loading cycle time 6 hr
Regeneration cycle time 1.5 hr
Cooling and drying cycle time 0.75 hr
Number of stacks 2
Waste gas flow rate3 11.3 m3/sec (2.4 x 104 cfm)
Cooling air flow rate 11.3 m3/sec (2.4 x 104 cfm)
Weight of carbon per bed 15,900 kg (35,000 Ib)
aAt capacity.
SOURCE: Weber, Robert. Trip Report for Visit to Reichhold Chemicals,
Inc., Elizabeth, N.J., July 15, 1978. U.S. Environmental Pro-
tection Agency.
2-11
-------
cies of 99 percent plus were achieved at temperatures of 760° C (1,400° F)
27
or greater.
Research data on a fume incinerator also indicate high removal effi-
ciencies for similar streams. With a toluene-contaminated gas, it was
found that approximately 99 percent destruction efficiency was achieved at
760° C (1,400° F) and a residence time of 0.33 second. As the temperature
increased to 816° C (1,501° F), the efficiency increased to 99.5 percent.
Recent laboratory studies on thermal incineration of benzene also show
28
high destruction efficiencies. With a detection limit of 2 ppmv, no
benzene was found at temperatures above 790° C (1,454° F) with residence
times as low as 0.08 second. Based on that research, a thermal oxidizer
with a residence time of 0.5 second would require a temperature of 750° C
(1,382° F) for 99.9 percent destruction of benzene.
One plant controls the emissions from the product recovery absorber by
routing the waste gas stream to a waste heat boiler. In tests conducted in
1977, the average benzene removal efficiency was calculated to be 99 percent
29
with benzene outlet concentrations ranging from 6.0 ppmv to 9.0 ppmv.
Although the operating temperature is not reported, it is probably near
2,000° F (1,090° C). In later tests in 1978, the benzene removal efficiency
was estimated to range between 95.8 percent and 98.5 percent, with benzene
outlet concentrations ranging from 8.5 ppmv to 13 ppmv. Further, it was
30
stated that "... the boiler now operates somewhat below 2,000° F."
During this set of tests, the inlet benzene concentrations were not measured.
A waste heat boiler, however, is a viable option only if the facility can
economically use the additional steam generated.
A temperature of 870° C (1,600° F) is specified to ensure complete
combustion of the waste gas. While it is possible that greater than 99
percent VOC removal can be obtained at lower temperatures, it cannot be
predicted dependably. The conditions of Table 2-4 are consistent with
5 31
various air pollution engineering manuals. ' While the manuals do not
provide data on combustion temperatures above 800° C (1,472° F), extrapola-
tion of the data presented combined with the similar incineration experi-
ence described above supports the projection of greater than 99 percent
removal at 870° C (1,600° F). The costs for systems of this type are
presented in Chapter 5 for each maleic anhydride plant.
2-12
-------
Based on both the data from the maleic anhydride facility with the
boiler and the data discussed above on similar incinerator systems, removal
efficiency of 99 percent for benzene at 870° C (1,600° F) is anticipated in
a we11-resigned and efficiently operated incinerator. Pertinent technical
data are summarized in Table 2-4.
A second plant uses a thermal incinerator on this stream operating at
32
760° C (1,400° F) with a residence time of 0.7 second. Total hydrocarbon
destruction efficiency has been reported to be at least 93 percent, although
the company has no data on benzene removal efficiency. Tests conducted by
33
EPA showed a mean benzene removal efficiency of 98.6 percent. Although
the plant was operating at about 70 percent of capacity when sampling was
conducted, the plant personnel did not think the lower production rate
would seriously affect the validity of the results when applied at full
capacity. Pertinent technical data are summarized in Table 2-5. This
incinerator is also used to generate steam, with a steam production rate of
34
about 7 kg/sec (15.4 lb/ sec) during the testing.
Generally, an incinerator can operate at least 95 percent of the time
12
with proper attention to maintenance and operating parameters. Even if
repairs are required, they can often be made quickly or postponed until the
next plant shutdown.
As discussed previously, the process waste gases are destroyed in an
onsite boiler in one facility; however, other maleic anhydride plants have
high gas volumes that exceed the total air-firing needs of those plants'
steam boilers. (As discussed earlier in this chapter and Chapter 1, these
high values are required to maintain the percent by volume of benzene in
the feedstock to 1.5 or less. This leads to large volumes of waste gases
with low benzene concentrations.) Consequently, most plants cannot handle
the waste gases in onsite waste heat boilers because the existing equipment
o
is not designed for such high-volume, low-concentration gases.
In addition to controlling VOC emissions, thermal incineration will
reduce CO concentrations. At a temperature of 609° C (1,128° F), thermal
35
incineration converts 90 percent of the CO to CO^. For example, at the
higher temperatures of a plant with 99 percent benzene removal, 870° C
(1,600° F), more than 99 percent of the CO is expected to be oxidized.
2-13
-------
TABLE 2-4. DESIGN CRITERIA—THERMAL INCINERATION
(99 Percent Control)
Residence time 0.5 sec
Temperature 870° C (1,600° F)
3
Natural gas (supplemental 0.435 m /sec (922 cfm)
fuel)
Supplemental combustion 1.24 kg/sec (2.7 Ib/sec)
air
NOTE: References and basis explained in text.
TABLE 2-5. TECHNICAL DATA—THERMAL INCINERATION
(97 Percent Control)
Residence time 0.7 sec
Temperature 760° C (1,400° F)
o
Natural gas (supplemental fuel) 0.5 m /sec (1,066 cfm)
Supplemental combustion aira 6.5 kg/sec (14.3 Ib/sec)
aDENKA's incinerator is also used to produce steam for various processes.
The air used beyond theoretical combustion requirements may be needed
to produce enough steam for these processes.
SOURCE: Weber, Robert. Trip Report for Visit to DENKA Chemical Corporatioi
Houston, Texas, April 20, 1978. Office of Air Quality Planning
and Standards, U.S. Environmental Protection Agency.
2-14
-------
2.2.3 Catalytic Incineration
One maleic anhydride producer reports VOC removal of at least 85
percent and carbon monoxide removal of greater than 95 percent, with a
13
catalytic incinerator on the absorber offgas. In tests run on a pilot
o
catalytic incinerator with a nominal capacity of 1.4 m /sec (3,000 scfm) of
maleic anhydride exhaust gases, benzene control was about 98.7 percent, and
carbon monoxide control was about 99.3 percent. These tests were based
on approximately 300 hours of operating time.
2.3 REFERENCES
1. U.S. Environmental Protection Agency. Control Techniques for Volatile
Organic Emissions from Stationary Sources. NTIS. Springfield, Va.
EPA-450/2-78-022. May 1978.
2. Letter from Pierle, M. A., Monsanto Chemical Intermediates Co., to
Goodwin, D. R., Office of Air Quality Planning Standards, U.S. Envir-
onmental Protection Agency. September 19, 1978.
3. Wagner, N. J. Introduction to Vapor Phase Adsorption Using Granular
Activated Carbon. Calgon Activated Carbon Division, Calgon Corp.
November, 1977.
4. Hughes, T. W., et al. Source Assessment: Prioritization of Air Pollu-
tion from Industrial Surface Coating Operations. U.S. Environmental
Protection Agency. Research Triangle Park, N.C. EPA-650/2-75-019a.
1975.
5. MSA Research Corp. Hydrocarbon Pollutants Systems Study, Vol. L.,
Stationary Sources, Effects, and Control. U.S. Environmental Protec-
tion Agency. Research Triangle Park, N.C. Publication Number APTID-
1499. October 1972.
6. U.S. Environmental Protection Agency. Air Pollution Engineering
Manual, Stcond Edition, Danielson, J. A. (ed.). Research Triangle
Park, N.C. Publication Number AP-40. May 1973.
7. Lawson, J. F. Emission Control Options for the Synthetic Organic
Chemicals Manufacturing Industry Maleic Anhydride—Product Report.
Hydroscience, Inc. (Prepared for Office of Air Quality Planning and
Standards, U.S. Environmental Protection Agency. Research Triangle
Park, N.C.) EPA Contract Number 68-02-2577. March 1978.
8. Lewis, W. A., Jr., et. al. Source Assessment: Maleic Anhydride
Manufacture. Monsanto Research Corp. EPA Contract Number 68-02-1874.
August 1978.
9. Letter from Parmale, C. S., Hydroscience, Inc., to Evans, L. B., U.S.
Environmental Protection Agency. May 4, 1978.
2-15
-------
10. Twaddle, W., Charles Olson, and Karen L. Kramer. Heat Recovery Incin-
eration of Organic Emissions Saves Amoco $970,000/yr in Fuel Costs.
Chemical Processing. January 1978.
11. Report from Felder, R. M. , North Carolina State University, to Warren,
J. L., Research Triangle Inst. January 2, 1979.
12. Telecon. Weber, Robert, Office of Air Quality Planning and Standards,
U.S. Environmental Protection Agency, with Kirkland, John, Hirt Combus-
tion Engineers. November 28, 1978.
13. Mackay, J. S. Testimony by United States Steel Corporation on a
Proposed Standard by the U.S. Environmental Protection Agency on
Benzene Emissions From Maleic Anhydride Plants. (Presented to the
National Air Pollution Control Techniques Advisory Committee. Alexan-
dria. August 22-23, 1978.)
14. Letter from Madden, G. I., E. I. DuPont de Nemours & Co., to Weber,
R. C., Office of Air Quality Planning and Standards, U.S. Environ-
mental Protection Agency. August 29, 1978.
15. Telecon. Weber, Robert, Office of Air Quality Planning and Standards,
U.S. Environmental Protection Agency, with Pierle, Michael, Monsanto
Chemical Intermediates Co. January 18, 1979.
16. Telecon. Warren, J. L., Research Triangle Institute, with Brennan,
H., Amoco Chemical Corp. November 16, 1978.
17. Lawson, J. F. Trip Report for Visit to Reichhold Chemicals, Inc.,
Morris, Illinois, July 28, 1977. Hydroscience, Inc. EPA Contract
Number 68-02-2577.
18. Letter from Parmale, C. S., Hydroscience, Inc., to Henry, A. L.,
Reichhold Chemicals, Inc. May 3, 1978.
19. Smisck, M., and Cerney, S. Active Carbon. Elsevier Publishing Co.,
1970. p. 202.
20. Letter from White, R. E., Hydroscience, Inc., to Patrick, D. R.,
Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency. April 13, 1978.
21. Rohm and Haas. Ambersorb Carbonaceous Resins (technical bulletin).
August 1977.
22. Ray Solv. Inc. Product Brochure. Linden, N. J. October 1976.
23. Clayton Environmental Consultants, Inc., Emission Testing at a Maleic
Anhydride Manufacturing Plant—Reichhold Chemical, Inc., Morris, Illinois.
EPA Contract Number 68-02-2817. April 1978.
2-16
-------
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
Weber, R. C. Trip Report for Visit to Reichhold Chemicals, Inc.,
Elizabeth, New Jersey, July 5, 1978. Office of Air Quality Planning
and Standards, U.S. Environmental Protection Agency.
Rolke, R. W., et al. Afterburner Systems Study, Shell Development Co.
Office of Air Programs, U.S. Environmental Protection Agency.
EPA-R2-72-062. August 1972.
Industrial Gas Cleaning Institute. Study of Heat Recovery Systems for
Afterburners. (Prepared for Office of Air Quality Planning and Stand-
ards, U.S. Environmental Protection Agency.) EPA Contract Number
68-02-1473. August 1977.
Memo from Seeman, W. R., Hydroscience, Inc., to White, R. E., Hydro-
science, Inc. May 4, 1978.
Lee, K., H. J. Jahnes, and D. C. Macauley. Thermal Oxidation Kinetics
of Selected Organic Compounds. Union Carbide Corp. (Presented at the
71st Annual Meeting of the Air Pollution Control Association. Houston.
June 25-30, 1978.)
Letter from Lawrence, A. W. , Koppers Co., Inc., to Goodwin, D. R.,
Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency. February 28, 1978.
Letter from Lawrence, A. W., Koppers Co., Inc., to Goodwin, D. R.,
Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency. January 17, 1979.
U.S. Department of Health, Education, and Welfare.
Engineering Manual. 1967.
Air Pollution
Pruessner, R. D., ar. Broz, L. D. Hydrocarbon Emission Reduction
Systems. Chemical Eny neering Progress. 73(8):69-73. August 1977.
Midwest Research Inst. Stationary Source Testing of a Maleic Anhydride
Plant at the DENKA Chemical Corporation, Houston, Texas. EPA Contract
Number 68-02-2814.
Weber, Robert. Trip Report for Visit to DENKA Chemical Corp., Hous'^n,
Texas, April 20, 1978. Office of Air Quality Planning and Standards,
U.S. Environmental Protection Agency.
Ann Arbor Science Publishers, Inc. Pollution Engineering Practice
Handbook, Cheremisinoff, P. N., and R. A. Young (ed.). 1975. p. 262-264.
Letter from Lawrence, A. W., Koppers Co., Inc., to Goodwin, D. R.,
Office'of Air Quality Planning and Standards, U.S. Environmental
Protection Agency. October 31, 1978.
2-17
-------
3. REGULATORY OPTIONS
This chapter describes three regulatory options for two emission
points in maleic anhydride production facilities: the product recovery
absorber vent, and the refining vacuum system vent. The same control
device can be used to reduce emissions from both sources.
For existing maleic anhydride production facilities, two control
levels are considered viable alternatives for a standard based on best
available technology (considering environmental, energy, and economic
impacts). These options, and examples of applicable control techniques,
are:
Benzene control efficiency of 97 percent, which can be
ach. ad by carbon adsorption or thermal incineration; and
Benzene control efficiency of 99 percent, which can be
achieved by either thermal incineration or carbon adsorption.
The control devices; i.e., incinerators and carbon adsorbers, can be
designed and operated at various levels of control efficiency. The benzene
control efficiency of 97 percent was chosen to represent universally appli-
cable control already demonstrated on a maleic anhydride facility. Although
tests conducted at this facility indicated a 98-percent removal efficiency,
they were conducted at a production rate of oniy 70 percent of capacity.
Therefore, because it is uncertain whether the same control efficiency
would have been obtained if the plant were operating at 100 percent of
capacity, 97 percent benzene removal was selected as the basis for one
regulatory option. The 99-percent removal efficiency is based on tech-
nology transfer, as described in Chapter 2.
As discussed in Chapter 1, uncertainty continues regarding the viabil-
ity of converting existing benzene-based plants to facilities using n-butane
at- the feedstock. Little information is available on what would be required
to convert each existing benzene-based plant to an n-butane-based plant, or
what the consequences of such a conversion would be. Based on the limited
3-1
-------
information available, it appears that considerable effort continues to be
directed towards developing n-butane technology, particularly the catalyst.
Only the existing Amoco plant was originally designed to use n-butane as a
feedstock. Problems associated with converting existing plants to n-butane
include:
Potentially significant reductions in maleic anhydride production
when current n-butane catalyst technology is combined with equip-
ment designed for benzene as the feedstock, and
Unsatisfactory operation resulting from equipment changes needed
in the refining system.
Because of uncertainties concerning the technical feasibility of
converting each existing source to n-butane and the impacts of such conver-
sion, this approach is not considered a viable regulatory option for exist-
ing sources based on BAT.
The use of n-butane as a feedstock, however, is considered a viable
regulatory option for new sources. Because the industry was operating at
only 56 percent of capacity in 1978, few new sources are expected to be
built until the mid-1980's. This allows time for continued development of
the n-butane process. Furthermore, a new plant could be designed to use
n-butane and would therefore not encounter the potential problems associated
with conversion. In fact, one company has recently announced the construc-
tion of a new 45,400-Mg/yr (50,000 tons/yr) maleic anhydride plant based on
their proprietary n-butane technology, which is scheduled for completion in
early 1983.
In summary, only two of the three options outlined above were considered
viable regulatory options that could serve as the basis for a standard for
existing sources based on BAT. These two regulatory options are designated:
Option 1, 97 percent benzene control, and Option 2, 99 percent benzene
control. For new sources, a regulatory option of no detectable benzene
emissions (100 percent control) is considered a viable option as the basis
for BAT. One hundred percent control was also considered as an option
beyond BAT for existing sources. The environmental and energy impacts of
these alternatives for existing and new sources are presented in Chapter 4
and in the economic impact section of Chapter 5.
3-2
-------
4. ENVIRONMENTAL AND ENERGY IMPACT
The environmental and energy impacts of each regulatory option pre-
sented in Chapter 3 are discussed in this chapter. Both beneficial and
adverse environmental impacts are assessed. The two regulatory options
considered for existing plants are:
A removal efficiency for benzene of 97 percent based on carbon
adsorption or thermal incineration, and
A removal efficiency for benzene of 99 percent based on carbon
adsorption or thermal incineration.
Catalytic incineration is not assessed as a control technique at this
time because control efficiencies in the 97-percent to 99-percent range
have not been demonstrated on a commercial scale. However, if these effi-
ciencies could be achieved, catalytic incineration would be a desirable
control technique because it would:
Operate with essentially no fuel consumption,
Produce no secondary emissions,
Provide for changing catalyst if required by further regulations,
Oxidize CO to C09, and
1
Probably have lower capital costs.
The additional regulatory option for new plants of elimination of
benzene by substituting n-butane, or similar material, as the feedstock is
also assessed. Both primary and secondary impacts of these regulatory
options are discussed in the next sections.
4.1 AIR POLLUTION IMPACT
This section addresses both the positive and negative effects on air
pollution expected to result from the application of the regulatory options.
The emission rates from a model plant as described in Chapter 1 are present-
ed in Table 4-1 for the following cases:
4-1
-------
TABLE 4-1. EMISSION RATES AND MAXIMUM ANNUAL AVERAGE BENZENE CONCENTRATIONS
a,b
Model plant control
alternatives
Emission rate
in g/sec
(lb/sec)c
Combined
maximum
ground level
concentration
in ug/m3 (ppmv)
Contribution of the
process vents to the
combined maximum
concentration of all
sources yg/m3 (ppmv)
A. Uncontrolled model
plant
B. 99 percent control by
incineration
C. 99 percent control by
carbon adsorption
D. 97 percent control by
carbon adsorption
96.0 (2.12 x 10"1)
0.96 (2.12 x 10"3)
60.7 (1.9 x 10"2) 56.4 (1.8 x 10~2)
20.7 (6.5 x 10"3) 1.5 x 10"2 (4.7 x 10"6)
0.48 (1.06 x 10"3) (Stack A) 20.8 (6.5 x 10"3) 1.4 x 10"1 (4.4 x 10* )
0.48 (1.06 x 10 •*) (Stack B)
2.88 (6.35 x 10"3)
20.7 (6.5 x 10:3) 5.8 x 10"2 (1.8 x 10"5)
Storage, handling, and fugitive emissions are assumed to be uncontrolled in all cases.
Based on a benzene conversion rate of 90 percent.
c Only refers to emission rate from the product recovery absorber and vacuum system vent or to
the control device for these emission points.
SOURCE: H. E. Cramer Co., Inc. Dispersion Model Analysis of the Air Quality Impact of Benzene Emis-
sions From a Maleic Anhydride Plant for Four Emission Control Options. Salt Lake City, Utah.
(Prepared for Source-Receptor Analysis Branch, U.S. Environmental Protection Agency. Research
Triangle Park, N.C.) EPA Contract Number 68-02-2507. August 1978.
-------
A. Uncontrolled model plant,
B. 99 percent control by thermal incineration,
C. 99 percent control by carbon adsorption, and
D. 97 percent control by carbon adsorption.
Table 4-1 also includes the estimated maximum ambient ground level
concentrations of benzene in the vicinity of the model facilities, determined
2
by atmospheric dispersion modeling. This atmospheric dispersion model was
originally run for 97 and 99.5 percent control levels. Thus, the data
given for 99 percent control are revised from the original data in Refer-
ence 2 to reflect a change in control level from 99.5 to 99 percent. This
effectively doubles both the amount of benzene emitted and the level of
ambient benzene concentrations. Table 4-2 shows the mean annual average
ground level benzene concentrations produced by the emissions from combined
fugitive, storage, and process sources based on the atmospheric dispersion
2
model. Table 4-3 summarizes the mean annual benzene concentrations produced
by emissions from storage, fugitive, and process sources; it is based on
2
the atmospheric dispersion model with the following assumptions:
Storage tanks A and B are controlled to the 90-percent level.
Fugitive sources are controlled to the 90-percent level.
Emissions from the product recovery scrubber, the incinerator
vent, Vents A and B of the adsorber system, and the adsorption
system vent are assumed to remain the same with or without fugi-
tive and storage controls.
Percentage of total emissions is the percentage of the total
emitted for that particular source, storage, fugitive, or process,
and option combination.
This table shows the differences in ambient concentrations with and without
fugitive and storage emissions control; fugitive and storage emissions
control may be considered in the future.
The Industrial Source Complex (ISC) Dispersion Model was used to
evaluate the air quality impact of benzene emissions from the model maleic
anhydride plant. The program was used to calculate the maximum annual
average ground level benzene concentration resulting from each model plant
control alternative, for both the combined emissions from all sources of
benzene in the plant as well as the concentration from the product recovery
absorber and the refining vacuum system vent (both uncontrolled and con-
trolled). This information is presented in Table 4-1. The program also
4-3
-------
TABLE 4-2. MEAN ANNUAL AVERAGE BENZENE CONCENTRATIONS PRODUCED BY EMISSIONS
roriM TUC rnMDTkicn cniiorcc™
FROM THE COMBINED SOURCES'
Model plant con-
trol alternative
A
B
C
0
Mean concentration (ug/m )
0.1 km
1.71X101
1.28X101
1.30X101
1.29X101
0.2 km
2.42X101
6.36
6.77
6.77
0.3 km
3.22X101
3.95
4.46
4.63
0.5 km
3.43X101
2.09
2.48
2.85
0.7 km
3.04X101
1.35
1.60
1.98
1.0 km
2.48X101
8.58X10"1
9.89X10"1
1.37
1.5 km
1.74X101
S.OSxlo"1
5.64X10"1
8.48X10"1
2.0 km
1.29X101
3.46X10"1
3.73X10"1
5.89X10"1
10.0 km
1.39
3.69xlO~2
3.13xlO"2
5.63xlO~2
20.0 km
5.03X10""1
1.11*10~2
l.lOxlO*2
2.02xlO~2
3 Concentration in ppmv = (0.000314) x concentration in (jg/m (at 25° C and at 1 atm).
SOURCE: H. E. Cramer Co., Inc. Dispersion Model Analysis of the Air Quality Impact of Benzene Emissions from a Maleic Anhydride Plant for Four
Emission Control Options. Salt Lake City, Utah. (Prepared for Source-Receptor Analysis Branch, U.S. Environmental Protection Agency.
Research Triangle Park, N.C.) August 1978.
i
.£»
-------
TABLE 4-3. MEAN ANNUAL BENZENE CONCENTRATIONS PRODUCED BY EMISSIONS FROM INDIVIDUAL SOURCES
0.1 km
No fugitive or
storage controls
Source/a 1 ternat i ve
Storage tank A
A
B
C
0
Storage tank B
A
B
C
D
Fugitive sources
A
B
c
n 0
Product recovery scrubber
A
Incinerator vent
B
Adsorber system vent A
C
Adsorber system vent B
C
Adsorption system vent
D
TOTALS
A
B
C
0
Concentration
pg/m3
5.
5.
5.
5.
1.
1.
1.
1.
5.
5.
5.
5.
4.
3.89 x
1.19 x
1.14 x
1.46 x
1.71 x
1.28 x
1.30 x
1.29 x
57
57
57
57
39
39
39
39
75
75
75
75
29
io-2
lO'1
10'1
lO'1
IO1
IO1
IO1
IO1
% of total
emissions
32.6%
43.5%
42.8%
43.2%
8.1%
10.9%
10.7%
10.8%
33.6%
44.9%
44.2%
44.6%
25.1%
0.3%
0.93%
0.89%
1.13%
Fugitive and .
storage controls '
Concentration
pg/m3
5.57
5.57
5.57
5.57
1.39
1.39
1.39
1.39
5.75
5.75
5.75
5.75
4
3.89
1.19
1.14
1.46
5
1
1
1
x 10" 1
xlO-1
xio'1
x lo"1
x 10'1
x 10'1
x 10"1
xio'1
x 10'1
x 10" 1
xlO'1
x 10" l
.29
x 10"2
xlO"1
x 10" l
x 10"1
.56
.31
.51
.42
% of total
emissions
10.0%
42.5%
36.9%
39.2%
2.5%
10.6%
9.2%
9.8%
10.3%
43.9%
38.1%
40.5%
77.2%
2.97%
7.9%
7.6%
10.3%
0.2
No fugitive or
storage controls
Concentration
pg/m3
3.09
3.09
3.09
3.09
5.50 x 10"1
5.50 x 10"1
5.50 x 10"1
5.50 x 10" l
2.60
2.60
2.60
2.60
1.79 x IO1
1.18 x 10'1
2.71 x 10'1
2.59 x lo"1
5.22 x 10"1
2.42 x IO1
6.36
6.77
6.77
% of total
emissions
12.8%
48.6%
45.6%
45.6%
2.3%
8.7%
8.1%
8.1%
10.7%
40.9%
38.4%
38.4%
74.0%
1.86%
4.0%
3.8%
7.7%
km
Fugitive and .
storage controls '
Concentration
pg/m3
3.09 >
3.09 x
3.09 x
3.09 x
5.50 x
5.50 *
5.50 x
5.50 x
2.60 x
2.60 x
2.60 *
2.60 '
1.79 *
1.18 x
2.71 x
2.59 x
5.22 x
1.85 x
7.42 x
1.19
1.15
lO'1
ID' 1
lO'1
lO'1
10"2
10" 2
io"2
ID'2
lO'1
lO'1
10- l
lO'1
IO1
10- x
ID' J
ID' l
ID' l
IO1
10- 1
% of total
emissions
1.7%
41.2%
26.0%
26.9%
0.3%
7.4%
4.6%
4.8%
1.4%
35.0%
21.8%
22.6%
96.8%
15.9%
22.8%
21.8%
45.4%
Notes are found at the end of Table 4-3.
(continued)
-------
TABLE 4-3. (continued)
0.3
No fugitive or
storage controls
Source/a 1 ternat i ve
Storage tank A
A
B
C
D
Storage tank B
A
B
C
D
Fugitive sources
A
B
C
•P» n
I
cn
Product recovery scrubber
A only
Incinerator vent
B only
Adsorber system vent A
C only
Adsorber system vent B
C only
Adsorption system vent
D only
TOTALS
A
B
C
D
Concentration
ug/m3
1.96
1.96
1.96
1.96
3.18 x lo"1
3.18 x io"1
3.18 x 10"1
3.18 x 10" X
1.47
1.47
1.47
1.47
2.85 x 1Q1
2.03 x io"1
3.6 x IO"1
3.55 * IO"1
8.82 x io" 1
3.22 x lO1
3.95
4.46
4.63
% of total
emissions
6.
49.
44.
42.
0.
8.
7.
6.
4.
37.
33.
31.
88.
5.
8.
7.
19.
1%
6%
0%
3%
1%
1%
1%
9%
6%
2%
0%
8%
5%
14%
07%
96%
1%
km
0.5 km
Fugitive and b
storage controls '
Concentration
pg/m3
1.96 x
1.96 x
1.96 x
1.96 x
3.18 x
3.18 x
3.18 x
3.18 x
1.47 x
1.47 x
1.47 x
1.47 x
2.85 x
2.03 x
3.60 x
3.55 x
8.82 x
2.89 x
5.78 x
1.09
1.26
10'1
10"1
10"1
10" l
10"2
10"2
10"2
ID'2
lO'1
1C'1
10"1
10"1
IO1
10"1
lO'1
lO'1
10"1
IO1
10"1
% of total
emissions
0.7%
33.9%
18.0%
15.6%
0.1%
5.5%
2.9%
2.5%
0.5%
25.4%
13.5%
11.7%
98.6%
35.1%
33.0%
32.6%
70.2%
No fugitive or
storage controls
Concentration
pg/m3
9.90 x lo"1
9.90 x 10"1
9.90 x io"1
9.90 x io"1
1.47 x 10"1
1.47 x 10'1
1.47 x 10'1
1.47 x lo"1
6.77 x IO"1
6.77 x IO"1
6.77 x IO"1
6.77 x IO"1
3.25 x lO1
2.7 x 10"1
3.32 x io"1
3.31 x 10"1
1.03 x Hf1
3.43 x IO1
2.09
2.48
2.85
% of total
emissions
2.9%
47.4%
40.0%
34.7%
0.4%
7.0%
5.9%
5.2%
2.0%
32.4%
27.3%
23.8%
94.8%
12.9%
13.4%
13.3%
36.2%
Fugitive and .
storage controls '
Concentration
pg/m3
9.90
9.90
9.90
9.90
1.47
1.47
1.47
1.47
6.77
6.77
6.77
6.77
3.25
2.7 >
3.32
3.31
1.03
3.27
4.56
8.41
1.21
-Hf2
xlO"2
xlO"2
. ID'2
x ID'2
* 10'2
xlO'2
x 10'2
x 10"2
xlO'2
x 1Q
x 10'2
-101
' 10' [
xlO'1
x 10"1
> 10'1
xlO1
x 10"1
x 10'1
front
% of total
emissions
0.3%
21.7%
11.8%
8.2%
0.04%
3.2%
1.8%
1.2%
0.2%
14.9%
8.1%
5.6%
99.4%
59.2%
39.5%
39.4%
85.4%
i nuecO
-------
TABLE 4-3. (continued)
0.7 km
No fugitive or
storage controls
Source/alternative
Storage tank A
A
B
C
D
Storage tank B
A
B
C
D
Fugitive sources
A
B
t C
D
Product recovery scrubber
A
Incinerator vent
B
Adsorber system vent A
C
Adsorber system vent B
C
Adsorption system vent
D
TOTALS
A
B
C
D
Concentration
ug/m3
6.0 x io"1
6.0 x io"1
6.0 x io"1
6.0 x lo"1
8.53 x Hf2
8.53 x 10"2
A
8.53 x Hf*
8.53 x IO"2
3.91 x 10"1
_i
3.91 x 10 L
3.91 x Hf1
3.91 x 10"1
2.93 * IO1
2.82 x IO"1
2.60 x 10'1
2.60 x Hf1
9.11 x IO"1
3.04 x IO1
1.35
1.60
1.98
% of total
emissions
2.0%
44.4%
37.5%
30.3%
0.28%
6.3%
5.3%
4.3%
1.3%
29.0%
24.4%
19.8%
96.4%
20.8%
16.3%
16.3%
45.9%
Fugitive and b
storage controls '
Concentration
ug/m3
6.0 x
6.0 x
6.0 x
6.0 x
8.53 x
8.53 x
8.53 x
8.53 x
3.91 x
3.91 x
3.91 x
3.91 x
2.93 x
2.82 x
2.60 x
2.60 x
9.11 x
2.94 x
3.90 x
6.28 x
1.02
ID'2
nf2
lO'2
lO'2
IO"3
10'3
_0
10 J
ID'3
1C'2
_?
10 /
io-2
io'2
IO1
lO'1
10'1
10" !
10" l
IO1
lO'1
lO'1
% of total
emissions
0.2%
15.4%
9.5%
5.9%
0.03%
2.2%
1.4%
0.8%
0.13%
10.0%
6.2%
3.8%
99.6%
72.4%
41.4%
41.4%
89.4%
1.0
No fugitive or
storage controls
Concentration
ug/ni3
3.51 x
3.51 x
3.51 x
3.51 x
4.87 x
4.87 x
4.87 x
4.87 x
2.16 x
2.16 x
2.16 x
2.16 x
2.42 x
2.43 x
1.87 x
1.87 x
7.40 x
2.48 x
8.58 x
9.89 x
1.37
lO'1
lO'1
lO'1
10'1
lO'2
ID'2
-y
10 *
lO'2
lO'1
-1
10 L
10'1
10" l
IO1
10" l
10'1
10'1
ID' l
IO1
10'1
lO'1
% of total
emissions
1.4%
40.9%
35.9%
25.6%
0.2%
5.7%
4.9%
3.6%
0.87%
25.2%
21.8%
15.8%
97.6%
28.3%
18.9%
18.9%
54.8%
km
Fugitive and
storage control sa%^
Concentration
ug/m3
3.51 x
3.51 x
3.51 x
3.51 x
4.87 x
4.87 x
4.87 x
4.87 x
2.16 x
2.16 *
2.16 x
2.16 '
2.42 •
2.43 >
1.87 -
1.87 «
7.40 •
2.43 •
3.04 «
4.35 x
8.1 * ]
10'2
ID'2
ID'2
lO'2
10'3
ID'3
-I
10 J
ID'3
ID'2
„ p
10
1C'2
io'2
IO1
.
10- l
10'1
10'1
10' A
IO1
10"1
10" *
% of total
emissions
0.14%
11.6%
8.1%
4.3%
0.02%
1.6%
1.1%
0.6%
0.09%
7.1%
5.0%
2.7%
99.6%
79.9%
42.9%
42.9%
92.5%
(continued)
-------
TABLE 4-3. (continued)
1.5
No fugitive or
storage controls
Source/a 1 terna t i ve
Storage tank A
A
6
C
D
Storage tank B
A
B
C
D
Fugitive sources
A
B
C
Product recovery scrubber
A
Incinerator vent
B
Adsorber system vent A
C
Adsorber system vent B
C
Adsorption system vent
D
TOTALS
A
B
C
D
Concentration
pg/m1
1.87 x
1.87 x
1.87 x
1.87 x
2.54 x
2.54 x
2.54 x
2.54 x
1.12 x
1.12 x
1.12 x
1.12 x
1.71 x
1.80 x
1.20 x
1.20 x
5.24 x
1.74 x
5.05 x
5.64 x
8.48 x
10'1
10'1
10'1
10'1
10'2
ID'2
lO'2
10'2
10'1
1(fl
10" l
10'1
101
10'1
10'1
10'1
10'1
101
10'1
10"1
10'1
% of total
emissions
1.1%
37.0%
33.2%
22.1%
0.15%
5.0%
4.5%
3.0%
0.64%
22.2%
19.9%
13.2%
98.3%
35.7%
21.3%
21.3%
61.8%
km
Fugitive and .
storage controls '
Concentration
pg/m3
1.87 x
1.87 »
1.87 x
1.87 x
2.54 x
2.54 x
2.54 x
2.54 x
1.12 x
1.12 x
1.12 x
1.12 x
1.71 x
1.80 x
1.20 x
1.20 x
5.24 x
1.71 x
2.12 x
2.72 x
5.56 x
10'2
lO'2
10'2
10'2
10'3
lO'3
lO'3
lO'3
nf2
10'2
lO'2
lO'2
101
10'1
10'1
10'1
10'1
101
10'1
10'1
10'1
% of tota^
emissions
0.11
8.8%
6.9%
3.4%
0.01%
1.2%
0.9%
0.5%
0.07%
5.3%
4.1%
2.0%
99.8%
84.7%
44. IX
44.1%
94.2%
2.0
No fugitive or
storage controls
Concentration
pg/m3
1.18 x
1.18 *
1.18 x
1.18 x
1.59 x
1.59 x
1.59 x
1.59 x
7.02 x
7.02 x
7.02 x
7.02 x
1.27 x
1.42 x
8.44 x
8.44 x
3.85 x
1.29 x
3.46 x
3.73 x
5.89 x
10'1
10'1
10'1
10"1
10'2
lO'2
10'2
lO'2
10'2
lO'2
10'2
10'2
101
10'1
lO'2
10'2
10'1
101
10'1
10'1
10'1
% of total
emissions
0.9%
34.1%
31.6%
20.0%
0.12%
4.6%
4.3%
2.7%
0.54%
20.3%
18.8%
11.9%
98.5%
41.0%
22.6%
22.6%
65.4%
km
Fugitive
-------
TABLE 4-3. (continued)
10.0 km
No fugitive or
storage controls
Source/a 1 ternative
Storage tank A
A
B
C
0
Storage tank B
A
B
C
D
Fugitive sources
A
B
-P» C
ID D
Product recovery scrubber
A
Incinerator vent
B
Adsorber system vent A
C
Adsorber system vent B
C
Adsorption system vent
D
TOTALS
A
B
C
D
Concentration
ug/m3
8.79 x io"3
8.79 x io"3
8.79 x io"3
8.79 x IO"3
1.14 x io"3
1.14 x IO"3
1.14 x IO"3
1.14 x IO"3
5.05 x IO"3
5.05 x IO"3
5.05 x IO"3
5.05 x IO"3
1.37
2. 19 x IO"2
8.14 x IO"3
8.14 x IO"3
4. 13 x IO"2
1.39
A
3.69 x 10 i
3.13 x IO"2
5.63 x IO"2
% of total
emissions
0.63%
23.8%
28.1%
15.6%
0.08%
3.1%
3.6%
2.0%
0.36%
13.7%
16.1%
9.0%
98.6%
59.4%
26.0%
26.0%
73.4%
Fugitive and b
storage controls '
Concentration
ug/m3
8.79
8.79
8.79
8.79
1.14
1.14
1.14
1.14
5.05
5.05
5.05
5.05
x IO"4
x IO"4
x IO"4
x IO"4
x IO"4
x IO"4
xio'4
x IO"4
xlO'4
x IO"4
x IO"4
x ID'4
1.37
2.19
8.14
8.14
4.13
1
2.34
1.78
4.28
x 10"2
x 10" 3
x 10" 3
x 10"2
.38
-y
x 10 Z
x 10"2
x 10" 2
% of total
emissions
0.06%
3.4%
4.9%
2.1%
0.01%
0.49%
0.64%
0.27%
0.04%
2.2%
2.8%
1.2%
99.9%
93.6%
45.8%
45.8%
96.5%
20.0
No fugitive or
storage controls
Concentration
ug/m3
3.05 x
3.05 x
3.05 x
3.05 x
3.97 x
3.97 x
3.97 x
3.97 x
1.75 x
1.75 x
1.75 x
1.75 x
4.98 x
5.89 x
2.91 x
2.91 x
1.50 x
5.03 x
1.11 x
1.10 x
2.02 x
10'3
io-3
10'3
io-3
ID'4
lO'4
ID'4
1C"4
lO'3
ID"3
10"3
lO'3
lO'1
ID'3
lO'3
ID'3
ID''
10- 1
-y
10 i
ID'2
ID'2
% of total
emissions
0.61%
27.5%
27.7%
15.1%
0.08%
3.6%
3.6%
2.0%
0.35%
15.6%
15.9%
8.7%
99.0%
53.1%
26.4%
26.4%
74.4%
km
Fugitive and .
storage controls '
Concentration
ug/m3
3.05 x
3.05 x
3.05 x
3.05 x
3.97 x
3.97 x
3.97 x
3.97 x
1.75 x
1.75 *
1.75 x
1.75 x
4.98 x
5.89 x
2.91 x
2.91 x
1.50 x
4.98 x
6.41 x
6.34 x
1.55 x
ID"4
1C'4
ID'4
lO'4
ID'5
10" 5
ID'5
ID'5
ID'4
io-4
10'4
io-4
10" 1
ID'3
ID'3
ID'3
io-.2
lO'1
-1
10 J
ID'3
lO'2
% of total
emissions
0.06%
4.8%
4.8%
2.0%
0.01%
0.62%
0.63%
0.26%
0.04%
2.7%
2.8%
1.1%
99.9%
91.9%
45.9%
45.9%
96.6%
-------
Notes for Table 4-3.
NOTE: Data adjusted for a benzene conversion rate of 90 percent.
a
b
aStorage tanks A and B are assumed to be controlled to the 90-percent level.
Fugitive sources are assumed to be controlled to the 90-percent level.
Percentage of total emissions is the percentage of the total emitted for that particular source and
alternative combination.
Concentration ppmv = (0.000314) * concentration in pg/n3 (at 25° C and at 1 atm).
SOURCE: H. E. Cramer Co., Inc. Dispersion Model Analysis of the Air Quality Impact of Benzene Emissions
from a Maleic Anhydride Plant for Four Emission Control Options. Salt Lake City, Utah. (Pre-
pared for Source-Receptor Analysis Branch, U.S. Environmental Protection Agency. Research
Triangle Park, N.C.). EPA Contract No. 68-02-2507. August 1978.
I
•—«
o
-------
calculated the maximum benzene concentrations from the combined emissions
for averaging times of I hour, 3 hours (uncontrolled case only), 8 hours,
and 24 hours. These data are shown in Table 4-4.
The short-term program (ISCST) assessed the. averaging times of 1,3,8,
and 24 hours. The long-term program (ISCLT) was used to calculate annual
averages. For the model features used in this study, the ISCST program
corresponds to the Single-Source (CRSTER) Model, modified to include the
effects of separation of individual sources and the effects of volume and
area source emissions. The ISCLT program, which is a sector-averaged model
similar to the Air Quality Display Model (AQDM) or the Climatological
Dispersion Model (COM), makes the same basic model assumptions as the ISCST
program. The ISCLT program uses STAR summaries (statistical tabulation of
the joint frequency of occurrence of wind-speed and wind-direction cate-
gories, classified according to the Pasquill stability categories) to
calculate annual average concentrations. For the same source data and
meteorological data base, ISCLT calculates annual average concentrations
that are equivalent to those obtained from ISCST using a year of sequen-
tial, hourly meteorological data.
The dispersion estimates used a model plant with a capacity of
22,700 Mg/yr. The layout of the emission points was assumed to be as
depicted in Figure 4-1. The emission points are identified by name and
number in Table 4-5. Fugitive emissions and emissions from benzene storage
were assumed to be uncontrolled, and estimated emission rates (at design
-4
capacity) were taken from Reference 3; these were 0.45 gin/sec (9.9 x 10
Ib/sec), 0.05 gm/sec (1.1 x io"4 Ib/sec), and 0.22 gin/sec (4.8 x io"4
Ib/sec) for Tank A, Tank B, and fugitive emissions, respectively. These
types of emissions will be covered in separate documents but were Included
here to Illustrate the relative Influence of the product recovery absorber
and refining vacuum system vents on the ambient benzene concentration.
Two maleic anhydride plants are located in the Pittsburgh, Pennsylvania,
area. Also, meteorological conditions that maximize ground level concentra-
tions produced by emissions from stacks of the type studied here (neutral
stability in combination with moderate-to-strong winds that persist within
a narrow angular sector for a number of hours) are common In the Pittsburgh
area. Surface-air and upper-air meteorological data from the Greater
4-11
-------
TABLE 4-4. MAXIMUM 1-HOUR, 3-HOUR, 8-HOUR, 24-HOUR,
AND ANNUAL AVERAGE GROUND LEVEL BENZENE CONCEN-
TRATIONS PRODUCED BY THE COMBINED EMISSIONS FROM
A MALEIC ANHYDRIDE PLANT AT ANY DISTANCE DOWNWIND
AND AT 0.1, 1.0, 10.0, AND 20.0 km
Concentration (ug/ro )a
Distance
(km)
0.3
0.1
1.0
10.0
20.0
0.7
0.1
1.0
10.0
20.0
0.3
0.1
1.0
10.0
20.0
0.3
0.1
1.0
10.0
20.0
0.3
0.1
1.0
10.0
20.0
Alternative
A
1.31 x
1.11 x
9.93 x
1.46 x
7.63 x
9.68 x
2.80 x
8.14 x
7.28 x
2.90 x
6.48 x
3.55 x
4.41 x
3.92 x
1.48 x
5.82 x
1.95 x
3.08 x
1.52 x
5.79
6.10 x
2.16 x
3.90 x
2.18
7.92 x
1-hour
103
103
102
102
101
3-hour
10*
102
102
101
101
8- hour
102
102
102
101
101
24-hour
10*
102
102
101
Annual
101
101
101
10- *
Alternative
B
concentrations
3.73xl02
4.32X101
3.28
1.71
concentrations
concentrations
_
1.55xl02
1.34X101
8.68X10"1
3. 29x10" X
concentrations
8.83X101
6.09
3.37X10"1
1.28X10"1
concentrations
-
2.07X101
1.26
4.89X10"1
1.75X10"2
Alternative
C
_
3.73xl02
5.23X101
3.26
1.70
-
.
-
1.56xl02
1.79X101
9.05X10"1
2.55X10"1
8.83X101
7.24
3.52X10"1
1.28X10"1
2.07X101
1.48
4.94xlO"2
1.74X10"2
Alternative
D
4.45xl02
6.31X101
6.03
3.11
-
-
1.62xl02
2.50X101
1.62
e.oixio"1
1.02xl02
1.27X101
6.31X10"1
2.34X10"1
-
2.07X101
2.03
8.89xlo"2
3.18X10"2
a Concentration in ppmv - (0.000314) x concentration in ug/m (at 25° C
[77° F] and 1 atm). .
SOURCE: H. E. Cramer, Co., Inc. Dispersion Model Analysis of the Air
Quality Impact of Benzene Emissions Prom a Maleic Anhydride
Plant for Four Emission Control Options. Salt Lake City, Utah.
(Prepared for Source-Receptor Analysis Branch, U.S. Environmental
Protection Agency. Research Triangle Park, N.C.) EPA Contract
"umber 68-02-2507. August 1978.
4-12
-------
MALEIC ANHYDRIDE PLANT
INCINERATOR OR ADSORBER VENTS
I
I—»
U)
"~"\ (
•1 ]
J X3
STORAGE
TANK A
i
v STORAGE PRODUCT
9 TANK B 1 RECOVERY
2 I SCRUBBER
| 80
X4 X5
1
i
1
X6
IX
9.10+12
X7
0
N
J
I
P
0 - STACK SOURCES
• = VOLUME SOURCES
X - AREA SOURCES
10 20m
Figure 4-1. Layout of the maleic anhydride plant showing source locations.
-------
TABLE 4-5.
IDENTIFICATION OF MODEL SOURCE NUMBERS BY SOURCE
NAME AND EMISSION CONTROL ALTERNATIVE
Model
source no.
Source name
Emission control
alternatives
1
2
3
4
5
6
7
8
9
10
11
12
Benzene storage tank A
Benzene storage tank B
Fugitive emissions (subarea source)
Fugitive emissions (subarea source)
Fugitive emissions (subarea source)
Fugitive emissions (subarea source)
Fugitive emissions (subarea source)
Product recovery scrubber
Incinerator vent (with heat recovery)
Adsorber system vent A
Adsorber system vent B
Adsorber system vent
All
All
All
All
All
All
All
A—uncontrolled
B—99 percent
C—99 percent
C--99 percent
D--97 percent
SOURCE: H. E. Cramer Co., Inc. Dispersion Model Analysis of the Air
Quality Impact of Benzene Emissions From a Maleic Anhydride Plant
for Four Emission Control Options. Salt Lake City, Utah. (Pre-
pared for Source-Receptor Analysis Branch, U.S. Environmental
Protection Agency. Research Triangle Park, N.C.) EPA Contract
Number 68-02-2507. August 1978.
4-14
-------
Pittsburgh Airport were used in the dispersion-model calculations for these
reasons and because the data were readily available. The maleic anhydride
plant was assumed to be located in the Pittsburgh urban area, and ISC Urban
Mode 2 was used in the dispersion-model calculations to account for the
effects of urban roughness elements and heat sources. Because of the high
frequency of occurrence of west winds in the Pittsburgh area, an east-west
orientation was assumed for the maleic anhydride plant (see Figure 4-1).
The ISC Model contains numerous model options and features that are
required to simulate the air quality impact of emissions from the wide
variety of sources found in industrial source complexes. Because of the
requisite complexity of the ISC Model, it is usually economically unfeas-
ible to use the ISCST program to calculate hourly concentrations for a year
of sequential, hourly meteorological data--the procedure generally followed
with the Single-Source (CRSTER) Model. However, with knowledge of the
critical meteorological conditions for emissions from maleic anhydride
plants, it was possible to select ten 24-hour periods representative of
"worst-case" dispersion conditions and to execute ISCST for these days only
to obtain maximum 1-hour, 3-hour (Alternative A only), 8-hour, and 24-hour
average benzene concentrations. First, the ratio of the vector mean wind
speed to the scalar mean wind speed was computed for each day of 1964 from
the Greater Pittsburgh Airport hourly surface observations. This ratio has
a value near unity for days of persistent wind directions that are usually
associated with maximum 8-hour and 24-hour average concentrations. The
days with ratio values near unity were examined to select 10 days with
neutral and/or stable conditions and either persistent light-to-moderate
winds ("worst-case" meteorological conditions for the storage tanks and
fugitive emissions) or persistent moderate-to-strong winds ("worst-case"
meteorological conditions for the stack emissions). Those selected were
days 5, 10, 26, 112, 118, 122, 265, 316, 343, and 358. The principal
meteorological input to the ISCLT program for the annual concentration
calculations was the 1964 annual STAR summary for the Greater Pittsburgh
Airport.
The objective of the dispersion study was to estimate maximum ground
level benzene concentrations at or beyond the property boundaries and at
downwind distances of 0.1, 1.0, 10.0, and 20.0 km (0.06, 0.62, 6.21, and
12.43 mi). Maximum ground level concentrations produced by fugitive and
4-15
-------
storage tank emissions can be expected to occur in the immediate vicinity
of the plant production area. However, the dispersion coefficients used by
the ISC Model, the Pasquill-Gifford curves, begin at a downwind distance of
100 m (109.4 yd). Consequently, the property boundaries were assumed to be
100 m (109.4 yd) from the edge of the plant production area, and receptors
were spaced at 30-m (32.8-yd) intervals around the edge of the property
boundaries. Preliminary calculations for the buoyant stack emissions
indicated that the maximum ground level concentrations should occur within
2 km of the stacks. Therefore, additional receptors were placed at distances
from the boundaries, 0.1., 0.2, 0.4, 0.6, 0.9, 1.4, 1.9, 9.9, and 19.9 km
(0.06, 0.12, 0.25, 0.37, 0.56, 0.87, 1.18, 6.15, and 12.37 mi) along radials
originating at the center of Source 5 and passing through each of the
receptors at the property boundary.
The results of the dispersion analysis are summarized in Tables 4-1,
4-2, 4-3, and 4-4. The ISC Model calculations for the uncontrolled case,
Alternative A, indicate that emissions from the product recovery absorber
and refining vacuum system vent are principally responsible for the maximum
ground level benzene concentrations produced by the combined emissions from
the plant. At distances where maximum concentrations occur, the product
recovery absorber emissions account for more than 85 percent of the calcu-
lated concentrations. Additionally, for some distances and averaging
times, the product recovery absorber accounts for more than 98 percent of
the calculated maximum concentrations. However, the product recovery
absorber of Alternative A does not necessarily dominate the maximum con-
centrations at the assumed boundaries of the plant property. For a given
averaging time, meteorological conditions such as stability and wind direc-
tion determine whether the fugitive and storage tank emissions or the
product recovery absorber emissions are primarily responsible for the
magnitude and location of the maximum concentration at the property boun-
daries.
For Alternatives B, C, and D, the benzene emissions from the product
recovery absorber are reduced by 97 to 99 percent. With these reduced
emissions, the ISC Model calculations show that the fugitive and storage
tank emissions (which are not controlled) wield a much greater influence on
the maximum ground level benzene concentrations than in Alternative A. The
4-16
-------
maximum concentrations calculated for all averaging periods are located at
the assumed boundaries of the plant property, and the fugitive and storage
tank emissions are principally responsible for the maximum concentrations.
Because the fugitive and storage tank emissions dominate the air quality
near the plant for Alternatives B, C, and D, the effects on ambient air
quality near the plant of different emission reductions are minimal without
control of fugitive and storage emissions. Disregarding the contribution
from fugitive and storage emissions, however, Table 4-1 shows that the
maximum concentration of benzene from the process vents after 99 percent
control is less than 0.1 percent of the ambient concentration resulting
from these sources when uncontrolled. When Alternative D is used (97
percent control by carbon adsorption), the resulting benzene concentration
is about 0.1 percent of the concentration before control.
Table 4-1 and Table 4-4 emphasize the positive impact on air quality
from application of the alternative control systems. Table 4-6 summarizes
the reduction in benzene emissions for both a 97-percent and 99-percent
control level on a pi ant-by-plant basis, using a benzene conversion rate of
90 percent. However, some adverse effects on air quality are usually
associated with each control technique incorporated in the regulatory
options. These adverse impacts are compared to the benefits for each
technique in the next sections.
4.1.1 Carbon Adsorption (99 Percent Control of Benzene)
The carbon adsorption system described for this level of control in
Chapter 2 will reduce benzene emissions from 1,540 Mg/yr (1,700 tons/yr) to
50 Mg/yr (55 tons/yr) for the model plant. Of this emission reduction,
about 1,450 Mg (1,600 tons) of benzene per year are recovered for recycling
3
to the process. The remaining benzene and the other VOC (~460 Mg/yr [~507
tons/yr]) are picked up in the scrubber or decanter water and are removed
-3
3
as effluent wastewater. The amount of benzene sent to wastewater treat-
ment is approximately 40 Mg/yr (44 tons/yr), or about 1.4 g/sec (3.1 x 10
Ib/sec).
All plants should be able to recycle the benzene-containing waste-
4
waters as makeup water for the absorbers. Wastewaters typically include
those from water of dehydration, water removed from scrubber solution, and
3
the condensed vacuum jet stream. If wastewaters from the carbon adsorp-
tion units are recycled, some plants will have to discharge previously
4-17
-------
TABLE 4-6. REDUCTION IN BENZENE EMISSIONS FOR SELECTED CONTROL LEVELS1
Ashland, Neal
W. Va.
Koppers, Bridge-
vi lie, Pa.
Monsanto,
St. Louis, Mo.
OENKA,
Houston, Tex.
Reichhold,
Elizabeth, N.J.
Reichhold,
Morris, 111.
Tenneco,
Fords, N.J.
U.S. Steel,
Neville Island,
Pa.
Totals
Capa-
city
(Mg/yr)
27,200
15,400
38.10012
22,700
13,600
20,000
11,800
38,500
Total
uncon-
trolled
emis-
sions2
(Mg/yr)
1,849
1,047
2,589
1,543
924
1,359
802
2,616
12,729
97% control
Existing
control
device
efficien-
cy (%)
0
99
0
97
97
90
0
90
Emissions
with
existing
control**
(Mg/yr)
1,849
26
2,589
68
41
154
802
295
5,824
Mg/yr
Allow-
able
benzene
emissions4
(Mg/yr)
126
72
177
105
63
93
55
179
870
Mg/yr
Incre-
mental
benzene
removed5
(Mg/yr)
1,723
0
2,412
0
0
61
747
116
5-, 059
Mg/yr
99% control
Allow-
able
benzene
emissions6
(Mg/yr)
61
35
85
50
30
45
26
86
420
Mg/yr
Incre-
mental
benzene
removed7
(Mg/yr)
65
0
92
18
11
48
29
93
356
Mg/yr
Emission rates are based on: Lawson, J. F. Emission Control Options for the Synthetic Organic Chemicals
Manufacturing Industry Maleic Anhydride—Product Report. Hydroscience, Inc. (Prepared for Office of Air
Quality Planning and Standards, U.S. Environmental Protection Agency. Research Triangle Park, N.C.) EPA
Contract Number 68-02-2577. March 1978.
2Total uncontrolled emissions = [(2.6 kg/hr x 8,760 hr) * (190 kg/hr x 8,000 hr)]
x production capacity Mg/yr .
model plant capacity 22,700 Mg/yr
Based on a benzene conversion rate of 94.5 percent.
Emissions with existing control = {(2.6 kg/hr x 8,760 hr) + [190 kg/hr x (l-control device efficiency)
, production capacity Mg/yr .
a nnn
B'uuu
model plant capacity 22,700 Mg/yr
••Allowable benzene emissions with 97 percent control {[(2.6 kg/hr x 8,760 hr) + (345.4 kg/hr x 0.03
x 8,000 hr)] * (production capacity/22,700 Mg/yr)}.
Based on a benzene conversion rate of 90 percent.
Incremental benzene removed with 97 percent control = (emissions with existing controls) - (allowable
benzene emissions at 97% control).
Allowable benzene emissions with 99 percent control = {[(2.6 kg/hr x 8,760 hr) + (345.4 kg/hr x o.Ol
x 8,000 hr)] x (production capacity/22,700 Mg/yr)}
Based on a benzene conversion rate of 90 percent.
Incremental benzene removed with 99 percent control = (emissions with 97% controls) - (allowable benzene
emissions at 99% control).
4-18
-------
recycled wastewaters either to the plant's wastewater treatment plant or to
a municipal wastewater treatment plant. These wastewaters discharged to
treatment facilities will have lower benzene concentrations than the waste
stream from the carbon adsorbers but will contain additional organic com-
pounds; e.g., maleic acid.
Depending upon the method of wastewater treatment practiced by the
facility, some volatilization of benzene to the atmosphere can occur. A
worst-case assumption is that all of this benzene is emitted to the atmos-
_o
phere at a constant rate of 1.4 g/sec (3.1 x 10 Ib/sec). This rate is
roughly 3 percent of the benzene emission rate from the uncontrolled product
recovery absorber of a model plant, which was shown in Table 4-1. The
actual quantity of benzene emitted has not been determined because it is
highly dependent upon the wastewater treatment technique because benzene is
biodegradable. Furthermore, at least two facilities employ an aqueous
7 8
waste incinerator for the maleic anhydride production facilities. * The
aqueous stream from the carbon adsorber system could conceivably be combined
with the other liquid wastes sent to the incinerator. Assuming a conserva-
tive benzene destruction efficiency of 90 percent in this type of incinera-
tor, the benzene emissions from this source would be about 0.14 g/sec (3.1
-4
x 10 Ib/sec), or less than 1 percent of the emissions from an uncontrolled
product recovery absorber.
4.1.2 Carbon Adsorption (97 Percent Control of Benzene)
The carbon adsorption system described for this level of control will
reduce benzene emissions from 1,540 Mg/yr (1,700 tons/yr) to 100 Mg/yr
(110 tons/yr) based on the model plant. Like the system discussed in
Section 4.1.1, some fraction of this controlled benzene is picked up in the
decanter water and removed as effluent wastewater. In a manner analogous
to the previous system, about 1,430 Mg (1,580 tons) of benzene per year are
recovered and recycled to the process. Therefore, about 40 Mg/yr or 44
tons/yr (~1.4 g/sec [3.1 x io"3 Ib/sec]) of benzene would be carried in the
aqueous waste from the adsorber system.
Consistent with the assumptions in Section 4.1.1, a worst-case benzene
emission rate from this source is 1.4 g/sec (3.1 x 10 Ib/sec), or less
than 3 percent of the emissions from an uncontrolled product recovery
absorber. Similarly, control in a liquid incinerator would result in a
4-19
-------
benzene emission of less than 0.14 g/sec (3.1 x 10 Ib/sec). Recycle of
the aqueous stream could also be practiced.
4.1.3 Thermal Incineration (99 Percent Control of Benzene)
The incinerator system described in Chapter 2 reduces benzene emis-
sions from 1,540 Mg/yr (1,700 tons/yr) to 50 Mg/yr (55 tons/yr) for the
model plant. No organics are recovered for recycle. Pollutants generated
by the combustion process, particularly nitrogen oxides (NO ), may have a
A
negative impact on the environment. Compared with industrial furnaces and
boilers, however, fume incinerators tend to have low NO emission factors.
^
One reference reports NO effluent concentrations from thermal incinerators
of 20 to 30 ppmv.9 Even at temperatures as high as 870° C (1,598° F),
estimated NO concentrations would be on the order of 50 to 200 ppmv.
^
Preliminary results from recent tests on an incinerator operating at 760° C
(1,400° F) showed an average NO emission rate of about 0.27 g/sec (6.0 x
10"4 Ib/sec).10
The waste gas stream going to the incinerator contains approximately 2
percent CO by volume. Because of sufficient air and an incinerator tempera-
ture of 870° C (1,598° F) (this exceeds the autoignition temperature of CO,
which is 651° C [1,204° F]), essentially all CO is oxidized. The inciner-
ator thus achieves a net reduction in CO emissions from maleic anhydride
plants. For the incinerator conditions described in Chapter 2, it is esti-
3
mated that greater than 99 percent control of CO would be achieved.
In locations where natural gas may be unavailable for the incinerator,
fuel oil will be used as supplemental fuel. Depending upon the type of
fuel used, there may be SOp and particulate emissions from this control
technique. It is predicted that 87,000 bbl of oil per year will be used to
attain 99 percent control by thermal incineration. If this oil were 0.3
percent sulfur by weight, 64 Mg of S0« would be released to the atmosphere
per year.
4.1.4 Thermal Incineration (97 Percent Control of Benzene)
An incinerator with this efficiency will reduce benzene emissions from
1,540 Mg/yr (1,700 tons/yr) to 100 Mg/yr (110 tons/yr) for the model plant.
As with the incinerator discussed in the previous section, negative environ-
mental impacts may result from NO , SO and particulate matter emissions,
/N XX
but these impacts would tend to be less than impacts of the incinerator
operated at a higher temperature. It is predicted that 29,000 bbl of oil
4-20
-------
per year will be used to attain 97 percent control by thermal incineration.
If the oil were 0.3 percent sulfur by weight, 26 Mg of S02 would be released
to the atmosphere per year. The efficiency of CO destruction would also be
reduced at the lower temperature of 760° C (1,400° F).
4.1.5 n-Butane Process Conversion (100 Percent Control of Benzene)
Although few data are yet available on the quantity and composition of
the emissions, preliminary information indicates that total uncontrolled
VOC emissions are higher in the n-butane than in the benzene oxidation
o
process. n-Butane used as a feedstock may result in increased photochemi-
cal smog; it has been found to be photochemically oxidized to peroxides,
11 12
which are the precursors of the various types of photochemical smog. '
At present, there is no nationwide requirement to control the VOC emissions
from the n-butane oxidation process for maleic anhydride production.
Future regulations will be prepared for emissions from oxidation systems,
such as n-butane. Trace quantities of benzene may be formed in the n-butane
13 14
oxidation process. ' However, one company using n-butane has found no
benzene emissions from its reactors at lower detection limits of 1 ppmv.
If benzene is not used for new maleic anhydride production, benzene
supply intended for that production could be added to gasoline as an octane
booster. Total benzene emissions from gasoline marketing would then in-
crease. Several factors make this scenario seem unlikely, however.
Currently, benzene demand often exceeds the available supply, a situ-
ation expected to continue through the 1980's. Consequently, the benzene
that would be used at new sources, if there were no standard, is not part
of current production output. If new maleic anhydride sources wanted to
use benzene as their feedstock, an increase in benzene demand would result.
Typically, when demand for benzene fluctuates, supply is adjusted by
changing the level of production from the most expensive source. If ben-
zene were not prohibited as a feedstock for new sources and there were
little, if any, slack in the benzene supply, the additional benzene required
would probably be supplied by hydrodeakylation (HDA). HDA is the most
expensive benzene production method, and changes in benzene demand can be
accommodated by changing the volume of benzene production from HDA. HDA
production currently accounts for 25 to 30 percent of the benzene produced
per year. Because existing maleic anhydride plants currently use 3
4-21
-------
percent of the benzene produced, the HDA process should be able to accommo-
date fluctuations in demand for benzene caused by maleic anhydride producers.
If chemical process industries know that benzene will not be used as
feedstock for maleic anhydride production, they will presumably adjust
their projections of future benzene demand. Additional benzene production
capacity would be adjusted to reflect this decrease in demand for benzene
as a feedstock for new maleic anhydride sources.
In addition, toluene, the feedstock for benzene in the HDA process, is
a better octane booster than benzene. Toluene's octane number ([research
17 18
plus motor]/2) is 102.9 as compared to 97 for benzene. Unleaded premium
has an octane number of 93.0. Thus, instead of producing benzene by HDA
and blending it into gasoline, it seems more logical to blend toluene into
the gasoline than to produce more benzene.
4.2 WATER POLLUTION IMPACT
No water effluents are discharged as a result of the application of
incinerator systems. A wastewater stream containing benzene is associated
with the carbon adsorption systems. However, the organic load of the
wastewater from carbon adsorption would be less than 10 percent of the
3 19
total liquid waste from a typical maleic anhydride production facility. '
The wastewater stream from the carbon adsorption systems could also be
treated along with the other plant effluent or recycled to the process.
The organic liquid effluent from the carbon adsorber systems will include
maleic acid, other organic acids, formaldehyde, other aldehydes, and be^-
zene. These materials are also contained in other waste liquid streams
from the process. The organic liquid effluent from the carbon adsorption
systems is a small percentage of the total liquid effluent from maleic
anhydride production facilities and is therefore estimated to have an
insignificant incremental impact on water pollution.
Insufficient data are available to assess the wastewater load from the
n-butane process relative to the load from the benzene process. The dif-
ference in wastewater load will depend upon the specific operation of the
recovery, dehydration, and refining steps. As discussed in Chapter 1,
differences between the processes are not well known. Catalysts presently
used in the n-butane conversion process produce a greater quantity of
water-soluble byproducts than catalysts used in the benzene conversion
4-22
-------
process. The latest catalysts developed yield liquid and solid wastes of
20
the same order of magnitude as catalysts from the benzene process.
Formation of small quantities of fumaric acid in dehydration cannot be
avoided. Formic acid and maleic anhydride are the bases for some powerful
adhesives, which could lead to plugged pipes and equipment. To prevent
this plugging, periodic water washes of equipment are necessary, possibly
leading to an impact on water pollution, but no greater than that from the
19
benzene process.
4.3 SOLID WASTE DISPOSAL IMPACT
One potential impact on solid waste disposal associated with the
alternative emission control systems is the handling of spent carbon from
the adsorption systems. Typically, rather than being disposed of in a
landfill, the spent carbon will be transported to a facility for reclama-
tion and regeneration. If it were disposed of in a landfill, the amount of
solid waste from this source would be approximately 7,620 kg/yr (8.4 tons/
yr) from the model plant, using the system achieving 99 percent control of
3
benzene. Assuming the same bed life for the system at 97 percent control
of benzene, the amount of solid waste would be on the order of 7,430 kg/yr
(8.2 tons/yr) for the model plant. However, in both cases it is likely
that the carbon can be reclaimed. Because n-butane conversion to maleic
anhydride is not as efficient as benzene conversion, it yields lower recov-
ery of the product as molten maleic anhydride and increased recovery by
absorption in water and conversion to maleic acid. This recovery rate not
only makes it more difficult to purify the maleic anhydride but requires
13 21
disposal of increased amounts of acidic organic materials, * possibly as
solid waste. The latest catalyst developed, not currently in use, will
yield quantities of liquid and solid wastes of the same order of magnitude
20
as those from the benzene process.
4.4 ENERGY IMPACT
A model uncontrolled maleic anhydride plant, using the benzene proc-
ess, will produce a small energy surplus of 15 kJ (14.2 Btu) per kilogram
of maleic anhydride produced. * For a model plant with a production
capacity of 22,700 Mg/yr, the energy surplus is 340 GJ/yr (3.2 x io8 Btu
[55 bbl/yr]). The energy impact of each control technique for a model
plant is discussed in the following sections. For comparison, energy
4-23
-------
equivalents, in barrels of fuel oil, are included in parentheses (assuming
140,000 Btu/gal or 6.2 GJ/bbl).
4.4.1 Carbon Adsorption (99 Percent Control of Benzene)
Energy (steam) is required to desorb the benzene from the carbon. For
the model plant, this energy (as steam) is 57,000 MJ/Mg (24,500 Btu/lb) of
in ^
benzene emission reduced, or 86,000 GJ/yr (8.2 x 10 Btu/yr). The elec-
trical energy required for recycle pumps and other equipment is 2,500 MJ/Mg
(1,080 Btu/lb) of benzene emission reduced, or 3,800 GJ/yr (3.6 x 10
Btu/yr) for the model plant. The total energy requirement for a model
plant is thus about 90,000 GJ/yr (8.5 x 1010 Btu/yr [14,500 bbl/yr]) more
than for an uncontrolled plant.
4.4.2 Carbon Adsorption (97 Percent Control of Benzene)
Less energy is required for this system than for the system operating
at a higher removal efficiency. For the model plant, the required energy
(as steam) is estimated to be 82,000 GJ/yr (7.8 x io10 Btu/yr), or 56,000
22
MJ/Mg (24,000 Btu/lb) of benzene emission reduced. The required electrical
energy is assumed to be the same as for the system operating at 99 percent
control of benzene. Therefore, the total requirement is about 86,000 GJ/yr
(8.2 x io1 Btu/yr [14,000 bbl/yr]) more than for an uncontrolled model plant.
4.4.3 Thermal Incineration (99 Percent Control of Benzene)
Supplemental fuel is required in the form of natural gas or fuel oil
to maintain suitable operating conditions. The net amount of energy re-
quired for the model plant ranges from 278,000 MJ/Mg (120,000 Btu/lb) of
benzene removed for an incinerator without any form of heat recovery to
62,000 MJ/Mg (27,000 Btu/lb) for an incinerator in which 50 percent of the
3
heat in the exit gas stream is recovered. On an annual basis, this corre-
sponds to an additional 420,000 GJ/yr (4.0 x io11 Btu/yr [68,000 bbl/yr])
for an incinerator without heat recovery and 95,000 GJ/yr (8.9 x 10
Btu/yr [15,200 bbl/yr]) for an incinerator with heat recovery as compared
3
to an uncontrolled model plant.
4.4.4 Thermal Incineration (97 Percent Control of Benzene)
Because it is expected that 97 percent control of benzene can be
achieved at a lower temperature than 99 percent control, the amount of
supplemental fuel required will be decreased. For a typical model plant,
the expected net energy requirement is about 45,000 GJ/yr (4.1 x 10
4-24
-------
Btu/yr [7,000 bbl/yr]) for an Incinerator with heat recovery, or 29,000
22
MJ/Mg (12,500 Btu/lb) of benzene emission reduced.
4.4.5 n-Butane Process Conversion
Maleic anhydride recovery from n-butane oxidation may be less than
from benzene oxidation, depending on the type of catalyst used. One source
states that it presently recovers a 50-percent molar yield of maleic anhy-
14
dride from benzene oxidation and expects less from n-butane oxidation.
Another source claims a 70-percent molar yield from benzene oxidation and a
21
55-percent molar yield from n-butane oxidation. One recent study places
the annualized costs of utilities for the benzene process at $350,000,
while annualized utilities costs for the n-butane process are $450,000, based
20
on typical U.S. power and fuel costs. Thus, the n-butane process appears
to use more energy than the benzene process.
4.4.6 Summary
All of the add-on control devices require more energy than the typical
benzene process produces. The thermal incinerator requires the most energy
at 99 percent control, although with heat recovery it requires just slightly
more than the carbon adsorption system operating at the same level of
control. At 97 percent control, the carbon adsorption system requires
about twice as much energy as the incinerator does at 97 percent control.
Incineration with heat recovery operating at 99 percent removal efficiency
for benzene uses less than 10 percent more energy than the carbon adsorp-
tion system operating at 97 percent control of benzene. However, the
thermal incinerator operating at 97 percent control uses significantly less
energy than the other methods. Furthermore, consideration of the energy
usage by the add-on control devices at a 99-percent level of control does
not significantly favor either carbon adsorption or incineration, provided
heat recovery is employed with the latter. Thermal incineration is favored
at 97 percent control. As expected, energy requirement decreases as control
efficiency decreases.
The total national energy requirement will depend upon the particular
control technique chosen for any alternative level of control. However,
the anticipated upper limit on energy usage is the case in which all plants
achieve 99 percent control of benzene by thermal incineration with 50 per-
cent heat recovery. The lower limit is the case in which all benzene-based
facilities convert to n-butane as the feedstock.
4-25
-------
TABLE 4-7. TOTAL NATIONAL INCREMENTAL ENERGY REQUIREMENT
Control system
Annual incremental
energy requirement (TJ/yr)
Carbon adsorption
• 97% control
• 99% control
Thermal incineration
(with 50% heat recovery)
• 97% control
• 99% control
n-Butane process (new facilities
only)
• 100 control
440 (71,000 bbl/yr)
510 (82,000 bbl/yr)
180 (29,000 bbl/yr)
540 (87,000 bbl/yr)
~0 (~
Note: Current assumed energy use:
Plant Control technique
DENKA
Reichhold, N.J.
Reichhold, 111.
U.S. Steel
Thermal incineration
Carbon adsorption
Carbon adsorption
Catalytic incineration
Annual energy
requirement (TJ/yr)
95 (15,200 bbl/yr)
52 ( 8,400 bbl/yr)
60 ( 9,600 bbl/yr)
0 ( 0 bbl/yr)
SOURCE: Lawson, J. F. Emission Control Options for the Synthetic Organic
Chemicals Manufacturing Industry Maleic Anhydride—Product Report.
Hydroscience, Inc. (Prepared for Office of Air Quality Planning
and Standards, U.S. Environmental Protection Agency. Research
Triangle Park, N.C.) EPA Contract Number 68-02-2577. March 1978.
Rolke, R. W., et al. Afterburner Systems Study, Shell Development
Company. Office of Air Programs, U.S. Environmental Protection
Agency. EPA-R2-72-062. August 1972.
Letter from Weber, Robert, Office of Air Quality Planning and
Standards, U.S. Environmental Protection Agency, to Vatavuk, W. M.,
Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency. July 7. 1978.
4-26
-------
TABLE 4-8. n-BUTANE AND BENZENE CHARACTERISTICS
Characteristics Benzene Butane
Lower explosive limit 1.3% 1.9%
Upper explosive limit 7.1% 8.5%
Explosion hazard Moderate Moderate
Spontaneous heating No No
Disaster hazard Dangerous (highly Moderately dangerous
flammable) (when heated, it emits
acrid fumes; can react
with oxidizing materi-
als)
SOURCE: Sax, Irving. Dangerous Properties of Industrial Materials, Third
Edition. New York, Van Nostrand Reinhold Company, 1968. p. 456,
494.
4-27
-------
Table 4-7 reflects the total national energy increase from current
control use to that required for a specific type of control and particular
control level. Accordingly, plants that currently achieve 97 or 99 percent
control are not included in the energy requirements for a 97-percent regula-
tory option. Also, energy consumed by plants to achieve less than 97 percent
control is subtracted from that energy required to achieve 97 percent.
Similarly, plants that achieve 99 percent control are not included in the
energy requirements for a 99-percent option, and energy used currently to
meet any other level of control is subtracted.
4.5 OTHER ENVIRONMENTAL IMPACTS
No other known significant environmental impacts are associated with
any of the alternative emission control systems discussed in Chapter 3.
4.6 OTHER ENVIRONMENTAL CONCERNS
4.6.1 Irreversible and Irretrievable Commitment of Resources
Emission control systems using carbon adsorption recover a valuable
raw material--benzene--for reuse, whereas an incineration system uses a
nonrenewable resource. Over the long term, the amount of recovered benzene
is substantial. Moreover, when benzene is recovered rather than burned,
there is an energy savings equivalent to the energy required to produce
benzene.
4.6.2 Safety Issues
Table 4-8 summarizes the safety-related characteristics of n-butane
and benzene. The table shows that if n-butane is substituted for benzene
as a feedstock, no significant changes in process safety are expected.
n-Butane is already used in similar situations within the petrochemical
industry as a feedstock. On a relative basis, n-butane would be safer than
benzene.
4.7 REFERENCES
1. Letter from Madden, G. I., E. I. Dupont de Nemours and Co., to Weber,
R. C., Office of Air Quality Planning and Standards, U.S. Environ-
mental Protection Agency. August 29, 1978.
«
2. H. E. Cramer Co., Inc. Dispersion Model Analysis of the Air Quality
Impact of Benzene Emissions from a Maleic Anhydride Plant for Four
Emission Control Options. Salt Lake City, Utah. (Prepared for Source-
Receptor Analysis Branch, U.S. Environmental Protection Agency. Research
Triangle Park, N.C.) EPA Contract Number 68-02-2507. August 1978.
4-28
-------
3. Lawson, J. F. Emission Control Options for the Synthetic Organic
Chemicals Manufacturing Industry Maleic Anhydride—Product Report.
Hydroscience, Inc. (Prepared for Office of Air Quality Planning
and Standards, U.S. Environmental Protection Agency. Research
Triangle Park, N.C.) EPA Contract Number 68-02-2577. March 1978.
4. Letter from Pierle, M. A., Monsanto Chemical Intermediates Co., to
Patrick, D. R., Office of Air Quality Planning and Standards. April 11,
1978.
5. Letter from'Hewett, P. S., Reichhold Chemicals, Inc., to Goodwin,
D. R., Office of. Air Quality Planning and Standards, U.S. Environ-
mental Protection Agency. August 30, 1978.
6. Mitre Corp. Dorigan, J., B. Fuller, and R. Duffy. Preliminary Scor-
ing of Selected Organic Pollutants, Appendix I—Chemistry, Production,
and Toxicity of Chemicals A through C. (Prepared for Office of Air
Quality Planning and Standards, U.S. Environmental Protection Agency.
Research Triangle Park, N.C.) EPA-450/3-77-008b. October 1976.
7. Lawson, J. F. Trip Report for Visit to Reichhold Chemicals, Inc.,
Morris, Illinois, July 28, 1977. Hydroscience, Inc. EPA Contract
Number 68-02-2577.
8. Lawson, J. F. Trip Report for Visit to Amoco Chemicals Corp., Chicago,
Illinois, January 24, 1978. Hydroscience, Inc. EPA Contract Number
68-02-2577.
9. Rolke, R. W., et al. Afterburner Systems Study, Shell Development
Company. Office of Air Programs, U.S. Environmental Protection Agency.
EPA-R2-72-062. August 1972.
10. Midwest Research Institute. Stationary Source Testing of a Maleic
Anhydride Plant at the DENKA Chemical Corp. Houston, Tex. EPA Con-
tract Number 68-02-2814.
11. Bufalini, J. S., B. W. Gay, and S. L. Kopczynski. Oxidation of n-Butane
by the Photolysis of N02. Environmental Science and Technology.
5(4). April 1971.
12. Manahan, S. E. Environmental Chemistry, Second Edition. Willard
Grant Press, 1975. p. 532.
13. Letter from Weber, Robert, Office of Air Quality Planning and Standards,
U.S. Environmental Protection Agency, to Gans, M., Scientific Design
Co. December 27, 1978.
14. Letter from Pierle, M. A., Monsanto Chemical Intermediates Co., to
Goodwin, D. R., Office of Air Quality Planning and Standards, U.S.
Environmental Protection Agency. March 22, 1978.
4-29
-------
15. Telecon. Weber, Robert, Office of Air Quality Planning and Standards,
U.S. Environmental Protection Agency, with Pierle, Michael, Monsanto
Chemical Intermediates Co. January 18, 1979.
16. Gunn, T. C. Benzene. Chemical Economics Handbook. Stanford Research
Inst. Menlo Park, Calif. November 1973.
17. Dosher, John R. Toluene's Role as Gasoline Octane Improver Will Not
Change Much. Oil and Gas Journal. 104-110. May 28, 1979.
18. Lorenz, H. R. Toluene is the Essence. Chemical Engineering Progress.
73(8):11-13. August 1977.
19. Letter from Hewett, P. S., Reichhold Chemicals, Inc., to Patrick,
D. R., Office of Air Quality Planning and Standards, U.S. Environmen-
tal Protection Agency. March 27, 1978.
20. Letter from Gans, M. , Scientific Design Co., to Warren, J., Research
Triangle Inst. February 2, 1979.
21. Gans, M. The Manufacture of Maleic Anhydride. Chemical Age of India
(Special on "Progress"). 28(7). July 1977.
22. Letter from Weber, Robert, Office of Air Quality Planning and Stand-
ards, U.S. Environmental Protection Agency, to Vatavuk, W. M. , Office
of Air Quality Planning and Standards, U.S. Environmental Protection
Agency. July 7, 1978.
4-30
-------
5. ECONOMIC IMPACT
5.1 SUMMARY OF ECONOMIC IMPACTS OF THE MALEIC ANHYDRIDE BENZENE NESHAP
Two regulatory options—97 percent and 99 percent—are analyzed for
existing facilities. One additional regulatory option—100 percent—is
analyzed for new facilities. The following economic issues are examined in
this chapter:
Capital budget impacts,
Shifts in competitive positions because of unequal control costs,
Price impacts on products that use maleic anhydride (MA),
Employment and balance of trade effects,
Annualized costs in fifth year and energy impacts, and
Impact of requiring n-butane at new MA facilities.
The results are summarized below.
5.1.1 Capital Budget Impacts
At the 99-percent control level, all firms except probably DENKA and
possibly Tenneco would finance the investment. At the 97-percent control
level, the only possible closure candidate is Tenneco.
5.1.2 Shifts in Competitive Position
At 97 or 99 percent control and with high levels of demand (100 percent
of capacity), no shift in intraindustry competition is expected. At 97 or
99 percent control and with low levels of demand (56 percent), Tenneco
would face the highest degree of domestic intraindustry competition, with
DENKA, Koppers (111.), and Amoco facing the lowest degree of that competi-
tion.
5.1.3 Price Impacts On Products That Use MA
The most significant price increase (assuming full cost pass through)
projected is in fumaric acid, ranging up to 4.4 percent at the 99-percent
control level and low level of MA demand. Polyester resins and malathion
will have price increases less than 1 percent under all scenarios.
5-1
-------
5.1.4 Employment and Balance of Trade Impacts
At the 99-percent control level--assuming both DENKA and Tenneco
close—roughly 45 jobs would be lost. At 97 percent control, if only
Tenneco closes, 12 jobs would be eliminated. Under either control level,
balance of trade impacts should be negligible. MA is imported only in
briquette form, which has a limited market.
5.1.5 Annualized Costs in Fifth Year and Energy Impacts
Annualized costs in the fifth year will be $2.1 million and $3.5 mil-
lion for the 97- and 99-percent control options, respectively. Energy
consumption, assuming thermal incineration with 50 percent heat recovery,
will be 180 TJ and 540 TJ for the 97- and 99-percent options, respectively.
5.1.6 Impact of Requiring n-Butane (100 Percent Control) at New MA Facili-
ties
Because new MA facilities are expected to be n-butane-based whether or
not a standard is promulgated, there is no impact of this requirement.
5.2 INDUSTRIAL ECONOMIC PROFILE
5.2.1 Maleic Anhydride Supply and Capacity
5.2.1.1 General. MA is produced by eight companies at 10 plants
(Figure 5-1) across 6 States: New Jersey, Pennsylvania, West Virginia,
Illinois, Missouri, and Texas. Total employment at these companies is
estimated at 330 workers.
Considerable excess MA capacity currently exists in the United States.
Market conditions have forced MA producers to operate plants at under 60
percent design capacity. Late in 1978, however, increased demand for MA
resulted in improved market conditions. Several companies found it finan-
cially viable to increase their plants' operating rates to near full capac-
ity. Although present demand has increased 11 percent since 1976, the
predicted demand in 1979 is only 60 percent of the overall name plate capac-
ity of 239 Gg (52.7 x 107 Ib).1
The overcapacity situation may continue for the next few years.
Investor confidence has been strong, however, and producers believe that
demand will equal capacity by the end of 1982 (assuming a continued 11-per-
cent growth rate). The confidence felt by chemical producers is best
demonstrated by the rapid growth in MA capacity in the past few years.
In 1975, annual capacity was rated at 156 Gg (3.4 x 108 Ib). In 1976 it
grew to 239 Gg (5.3 x io8 Ib)—an increase of 53 percent—and included
5-2
-------
K
E
Y
1.
2,
3.
4.
5.
Amoco, Joliet, III.
Ashland, Neat, W. Va.
Koppers, Bridgeville, Pa.
Koppers, Chicago, III.
Monsanto, St. Louis, Mo.
6.
7.
8.
9.
10.
DENKA, Houston, Tex.
Reichhold, Elizabeth, N.J
Raichhold. Morris, III.
Tenneco, Fords, N.J.
U.S. Steal. Neville Island,
Pa.
Figure 5-1. Manufacturing locations of maleic anhydride.
SOURCE: Lawson, J.F. Emission Control Options for the Synthetic Organic Chemicals Manufacturing
Industry Maleic Anhydride—Product Report. Hydroscience, Inc. (Prepared for Office of Air
Quality Planning and Standards, U.S. Environmental Protection Agency. Research Triangle
Park, N.C.) EPA Contract Number 68-02-2577. March 1978.
5-3
-------
two new 28-Gg plants (Ashland and Amoco), a large 18.2-Gg (4.0 x 10 lb)
expansion (U.S. Steel), and a small 3.6-Gg (7.9 x 10 lb) expansion
2
(Koppers). Furthermore, Monsanto recently announced plans to expand its
MA capacity by building a new n-butane-based facility at Pensacola, Florida.
In the United States, MA is produced from one of two feedstocks,
benzene and n-butane. At present, benzene supplies are tight and although
benzene consumption accounts for 83 percent of all feedstock used in the
industry, this number could be reduced significantly should producers
convert current benzene-based MA plants to n-butane. Industry sources have
indicated that steadily increasing benzene prices have prompted serious
consideration of feedstock conversion.
Another source of MA, though minor, is byproduct recovery from phthalic
anhydride. Presently, only Koppers in Chicago, with a total capacity of
5 Gg (1.1 x 10 Ib), produces MA from the effluent of its phthalic anhydride
plant. This source of MA depends on production of phthalic anhydride as
the primary product. No growth in this source of MA is expected.
5.2.1.2 The Individual MA-Producing Companies. Table 5-1 compares
the capacities of the individual MA plants with their estimated 1978 produc-
tion, and Figure 5-2 summarizes the relationship between each company's
captive and merchant sales. The eight MA-producing companies and their
capacities are:
Amoco Chemical Corporation
Amoco has the only United States plant totally dedicated to
the n-butane process.5 The facility has an annual capacity
of 27 Gg and is expandable to 41 Gg (9.0 x 107 Ib).1
Ashland Chemical Company
The Ashland facility is a new benzene-based plant with an
annual capacity of 27 Gg (5.9 x 107 lb) expandable to 41 Gg
(9.0 x 10' lb).5 Fifty percent of the annual capacity is
used captively to produce unsaturated polyester resins.
This plant can be switched from benzene to n-butane feed-
stocks. 1
Koppers Company, Inc.
Kopper's Chicago facility can recover 5 Gg (1.1 x 107 lb) of
MA per year from the effluent of their phthalic anhydride
plant, which started in 1975.5 The company's Bridgeyille
plant was mothballed in the spring of 1979. When this
16-Gg/yr (3.5 x 107 lb) benzene plant is operating, approxi-
mately 25 percent of the MA produced is used captively to
produce unsaturated.polyester resins and alkyd resins.
5-4
-------
TABLE 5-1. MALEIC ANHYDRIDE CAPACITY
Manufacturing locations
Estimated
production
1978 (Gg) (Ib)
Capacity
1978 (Gg) (Ib)
Process
1. Amoco, Joliet, 111.
2. Ashland, Neal, W.Va.
16 (3.5 x 107)
16 (3.5 x 107)
3. Koppers, Bridgeville, Pa. 10 (2.2 x 1Q/)
4. Koppers, Chicago, 111.
3 (6.6 x 10°)
5. Monsanto, St. Louis, Mo. 29 (5.9 x 10')
6. DENKA, USA,
Houston, Tex.
7. Reichhold, Elizabeth,
N.J.
14 (3.1 x 10')
11 (2.4 x 10')
8. Reichhold, Morris, 111. 16 (3.5 x 10')
9. Tenneco, Fords, N.J.
7 (1.5 x
10. U.S. Steel, Neville 22 (4.9 x 10')
Island, Pa.
27 (5.9 x 107)
27 (5.9 x 10')
16 (3.5 x 10')
5 (1.1 x 107)
48 (1.1 x 108)
23 (5.1 x 10')
18 (4.0 x 107)
27 (6.0 x 107)
12 (2.6 x 107)
36 (7.9 x
TOTAL
144 (31.8 x 10) 239 (52.7 x 10)
Oxidation of n-
butane
Oxidation of
benzene
Oxidation of
benzene
Byproduct of
phthalic
anhydride
manufacture
(80%) oxidation
of benzene
(20%) oxidation
of n-butane
Oxidation of
benzene
Oxidation of
benzene
Oxidation of
benzene
Oxidation of
benzene
Oxidation of
benzene
Estimated by assuming 56 percent capacity for each plant.
SOURCES: Blackford, J. C. Marketing Research Report on Maleic Anhydride.
Chemical Economics Handbook. Stanford Research Inst. Menlo Park,
Calif. July 1976.
5-5
-------
MG X 1000
tn
I
18
16
14
12
10
8
6
2
0
—
—
—
—
_
—
—
v//////////,
••1MB
§s
w
•••••1
!
i
1
»»•
1
•M
1
i
1
•*•*•!
W///////A
i""~l MERCHANT SALES
kNSM CAPTIVE SALES
•••«•
1
ASHLAND MONSANTO DENKA REICHHOLD TENNECO U.S. STEEL AMOCO
Figure 5-2. Captive and merchant sales of MA companies.
KOPPERS
-------
Monsarao company
Monsanto at 48 Gg/yr (1.1 x 108 Ib/yr) is the largest producer
of MA. The company uses n-butane feedstocks for about 20
percent of its capacity. Approximately 25 percent of the MA
is consumed captively to produce fumaric acid, maleate/
fumerate esters, styrene copolymers, and ethylene-maleic
anhydride copolymers.5 Monsanto has announced plans to
build a new n-butane-based MA facility at Pensacola, Florida.
DENKA, USA
Their 23-Gg/yr (5.1 x io7 Ib/yr) Houston facility was designed
by Scientific Design Company, Inc., and was purchased from
Petrotex Chemical Corporation by a Japanese firm, July 1,
1977.6 The feedstock now used is benzene.
Reichhold Chemicals, Inc.
Reichhold's combined production from both plants, at Elizabeth,
New Jersey, and Morris, Illinois, is 45 Gg/yr (9.9 x IO7 lb/
yr), 20 percent of which is used captively to produce unsatu-
rated polyester resins, alkyd resins, and plasticizers.5
Tenneco Chemical, Inc.
Less than 20 percent of their 12-Gg/yr (2.6 x io7 Ib/yr) MA
production is used captively to produce fumaric acid, dibutyl
maleate, and dodecanylsuccinic anhydride.5
United States Steel Corporation
Their MA capacity was recently expanded to 36 Gg/yr (7.9 x
IO7 Ib/yr). Approximately 20 percent of their MA production
is used captively to produce fumaric acid, dibutyl maleate,
and dioctyl maleate.5
The expansion capabilities of 14 Gg (3.1 x 10 lb) each for Amoco and
Ashland represent a current potential nationwide capacity of 267 Gg (58.9 x
IO7 lb).2
5.2.2 MA Usage and Demand
Maleic anhydride is an important raw material in the production of
polyester resins, agricultural chemicals, lubricants, fumaric acid, copoly-
mers, and other intermediate raw materials. Table 5-2 breaks out these
categories by percentage and growth rates. Demand for MA has increased
historically at a rate of 9 percent a year and is expected to grow at a
rate of 6 to 11 percent over the next 5 years. A major reason for the
excess capacity is investor confidence that MA demand will be strong.
The predominant end use of MA is the production of unsaturated poly-
ester resins, which go into reinforced plastic applications such as marine
5-7
-------
TABLE 5-2. MALEIC ANHYDRIDE USAGE AND GROWTH
End use
Unsaturated polyester resins
Agricultural chemicals
Lubricating additives
Fumaric acid
Copolymers
Maleic acid
Reactive plasticizers
Surf ace- active agents
Alkyd resins
Chlorendic anhydride and acid
Other
All MA products Total
SOURCES: Chemical Profile on
1978
demand
Gg/yr (Ib/yr)
77.7 (1.7 x 108)
13.4 (3.0 x 107)
8.7 (1.9 x 107)
6.7 (1.5 x 107)
5.9 (1.3 x 107)
4.2 (9.3 x 106)
4.0 (8.8 x 106)
3.2 (7.0 x 106)
1.5 (2.6 x 106)
1.2 (2.6 x 106)
7.5 (1.7 x 107)
134.0 (3.0 x 108)
Maleic Anhydride.
1978 demand
as % of
Production
58
10
7
5
4
3
3
2
1
1
6
Average
annual
% growth
1978-83
13
10
12
5
9
10
8
8
4
13
5
Total 100 Ave. 11
Chemical Market i
ng Reporter.
February 18, 1978.
Blackford, J. C. Marketing Research Report on Maleic Anhydride.
Chemical Economics Handbook. Stanford Research Inst. Menlo
Park, Calif. July 1976.
5-8
-------
craft, building panels, automobiles, tanks, and pipes. Of the 134 Gg (30.0
x 107 Ib) of maleic anhydride produced in 1978, polyester resins consumed
58 percent (78 Gg [17.0 x 107 Ib]). Industry forecasters expect polyester
production to grow at an annual rate of 10 to 15 percent. The growth rate
would double this industry's use of maleic anhydride by 1981 to about
156 Gg (34.4 x 10 Ib)—or more, if potential polyester markets in housing,
marine craft, and autos materialize. Automobile manufacturers, striving to
lighten products in order to meet Congressionally mandated fuel economy
standards, are turning more and more to polyester with above average maleic
content to replace heavier metals. Presently, no chemical can substitute
for MA, as in the production of polyester resins.
The agricultural chemicals' market, which is the second largest MA
market (10 percent of demand), is expected to grow at a rate almost as
rapid as polyester resins. This growth could be further accelerated by
MA's use as a feedstock for agricultural pesticides. Although other chemi-
cals can substitute for MA as an agricultural chemical, MA is highly competi-
tive in this market.
Other markets, such as lubricants, maleic anhydride copolymers, fumaric
acid, and reactive plasticizers are expected to grow at either modest or
rapid rates in the next 5 years.
In periods of excess capacity, MA is not considered a regional product.
More than 15 percent of U.S. markets lie in each of the following regions:
Middle Eastern, South Eastern, Western South Central, and Eastern North
Central. Polyester resins are primarily produced in the Central States,
while agricultural chemicals and fumaric acid are mostly produced in the
East. Although these geographic tendencies exist, MA is a homogeneous
product and can be sold in any of the markets mentioned above.
5.2.3 Prices
5.2.3.1 Price of MA. Historically, prices of maleic anhydride have
fluctuated widely (Figure 5-3); it sold at 37.4
-------
CENTS PER
MOLTEN
KILOGRAM
90
80
70
60
50
40
30
20
10
1955
1960
1965 1970
YEAR
1975
1980
Figure 5-3. Price fluctuations of maleic anhydride.
SOURCE: Blackford, J. C. Marketing Research Report on Maleic Anhydride. Chemical
Economics Handbook. Stanford Research Inst. Menlo Park. Calif. July 1976.
5-10
-------
90.2$ per bagged kilogram (41.0
-------
user and can neither buy in bulk (i.e., 45,000 kg [1 x 105 lb]) per tank
car) nor afford liquid storage facilities, even though briquettes normally
cost 6.5
-------
GIG AG RAMS
(Gg)
(thousands of
metric tons)
120
100
80
60
40
20
1974
DOMESTIC
PRODUCTION
U.S. DEMAND
IMPORTS
1975
1976
1977
1978
YEAR
Figure 5-4. Maleic anhydride: a comparison of imports to
U.S. production and demand.
SOURCE: Telecon. Epstein, E. A., Energy and Environmental Analysis, Inc.,
with Mr. Kendrik, DENKA. May 4,1978.
5-13
-------
U.S.
BRIQUETTES
9%
IMPORTS
(BRIQUETTES)
5%
U.S.
MERCHANT
PRODUCTION
66%
U.S. CAPTIVE
PRODUCTION
20%
Figure 5-5. U.S. consumption of MA by source in 1977.
SOURCE: Telecon. Epstein, E.A., Energy and Environmental Analysis, Inc., with Magnusson, Fred,
U.S. Department of Commerce. March 3, 1978.
5-14
-------
mers as well as decrease their surplus supplies of MA. These price increases
have brought foreign prices closer in line with domestic prices, thereby
reducing the likelihood that foreign prices will continue to undercut the
domestic market.
At present, equal quantities of butene (C4)-based MA from Korea and
Japan and benzene-based MA from Mexico and Italy are currently being imported.
Of these, the C.-based MA has the most favorable duty--6 percent ad valorem.
Some domestic producers fear that this percentage is not high enough to
keep large amounts of C4-based MA from being imported in the future at
considerably lower than current prices.
The import rate of duty on benzene-derived products is presently
3.8
-------
5.3.2 Summary of Technical Parameters Used as the Basis in Cost Analysis
This section summarizes the assumptions used in developing control
costs at the 97- and 99-percent control level for eight maleic anhydride
plants. It should be recognized, however, that these costs are only pre-
liminary estimates (±30 percent). When these costs were developed, plant
capacity was the only plant-specific parameter taken into consideration.
To develop definite costs for an actual installation, a detailed engineer-
ing evaluation is required. Such an evaluation is beyond the resources and
scope of this document.
The efficiencies and other parameters used in determining cost for the
control methods in this analysis are listed in Table 5-3. Because some of
these parameters vary from plant to plant, expressions have been derived in
terms of plant capacity, "P," for the sake of brevity. Because benzene-fed
maleic anhydride plants have similar process designs, it has been assumed
that the gas volumetric flow rate and the benzene emission rate vary in
proportion to plant capacity.
Other assumptions used in determining cost for the add-on systems for
each plant follow:
Intensive stream parameters, such as gas pressure and tempera-
ture, are the same for all plants;
No credit is given for control equipment used in an existing
plant, unless the controls already achieve the alternative in
question (i.e., 97 or 99 percent);
All control system installations are retrofits, whose costs
include retrofit penalties of 40 and 30 percent of the new plant
installation cost for incinerator and carbon adsorption systems,
respectively.6 (These moderate penalties, in turn, reflect the
fact that the costs of retrofitted systems are only somewhat
greater than those for completely new plant installations. The
primary retrofit difficulty may be finding adequate space to fit
the control system into the existing plant layout.)
In addition to developing control costs for the various maleic anhy-
dride plants, costs were developed for continuous monitoring of benzene
stack emissions. The device costed is a gas chromatograph with appropriate
auxiliary equipment including air sampler, data processor, and piping.
The add-on control costs have been based primarily on data available
from an EPA contractor, Hydroscience, Inc., and a compendium of costs for
13
selected air pollution control systems. Monitoring costs were obtained
14
from a vendor.
5-16
-------
TABLE 5-3. ESTIMATED TECHNICAL PARAMETERS USED IN DEVELOPING
CONTROL SYSTEM COSTS
Parameter Value3
1. Gas temperature 38° C (100° F)
2. Gas pressure 120 KPa (18 psia)
3. Gas volumetric flow rate 0.0536 P m3/min (1.89 P ACFM)b
4. Inlet benzene emission rate 0.0084 P kg/hr (0.0185 P Ib/hr)
5. Benzene control efficiency 97%, 99%
6. Plant capacities (See Tables 5-5 to 5-6)
7. Incinerator combustion 760° C (1,400° F), 0.5 sec for 97%
temperature and residence time 870° C (1,600° F), 0.5 sec for 99%
8. Design carbon loading 6 kg VOC/100 kg carbon (6 Ib VOC/
100 Ib carbon)
aEPA estimates.
P = plant capacity in megagrams of maleic anhydride per year.
SOURCES: Lawson, J. F. Emission Control Options for the Synthetic Organic
Chemicals Manufacturing Industry Maleic Anhydride--Product Report.
Hydroscience, Inc. (Prepared for Office of Air Quality Planning
and Standards, U.S. Environmental Protection Agency. Research
Triangle Park, N.C.) EPA Contract Number 68-02-2577. March 1978.
Letter from Vatavuk, William M., Office of Air Quality Planning
and Standards, U.S. Environmental Protection Agency, to Weber,
Robert, Office of Air Quality Planning and Standards, U.S.
Environmental Protection Agency. March 28, 1978.
5-17
-------
Two cost parameters—installed capital and total annualized—have been
evaluated in this analysis. The installed capital cost for each emission
control system includes the purchased costs of the major and auxiliary
equipment, costs for site preparation and equipment installation, and
design engineering costs. No attempt has been made to include costs for
research and development, possible lost production during equipment instal-
lation, or losses during startup. Capital and operating costs in this
section were developed in detail with cost parameters originally indexed to
the fourth quarter of 1977. Because of rapid escalation of benzene and
energy costs, the cost parameters were updated to the second quarter of
1979 without reoptimizing design parameters. Both the older and later data
are shown in Table 5-4.
The total annualized cost (TAG) consists of direct operating costs,
annualized capital charges, and recovery credits. Direct operating costs
(DOC) include fixed and variable annual costs, such as:
Labor and materials needed to operate control equipment;
Maintenance labor and materials;
Utilities, such as natural gas and electric power; and
Liquid waste disposal.
The annualized capital charges account for depreciation, interest,
administrative overhead, property taxes, and insurance. Depreciation and
interest have been computed by use of a capital recovery factor, the value
of which depends on the depreciable life of the control system and the
annual interest rate (10 percent is used for the latter). Administrative
overhead, taxes, and insurance have been fixed at an additional 4 percent
of the installed capital cost.
The recovery credits apply to the value of material or energy recovered
by the control system. With carbon adsorption systems, benzene is recovered
from the regenerated carbon beds, while a credit for natural gas is applied
to thermal incineration systems with primary heat recovery.
Finally, the total annualized cost is obtained by adding the direct
operating costs to the annualized capital charges and subtracting the
recovery credits from this sum.
5-18
-------
TABLE 5-4. COST PARAMETERS'
Parameter
Value
1. Operating factors
2. Maintenance:
Control systems
Monitoring system
3. Depreciation and interest
4. Taxes, insurance, overheads
4,500 and 8,000 hr/yr
5.0% of total installed cost
3.4% of total installed cost
16.28%b of total installed cost
4.0% of total installed cost
Fourth quarter 1977
Second quarter 1979
7.
8.
Utilities:
Electricity
Natural gas
Steam
Cooling water
Operating materials
Sodium hydroxide (50%)
Activated carbon
Liquid waste disposal
Credits
Primary heat recovery
Recovered benzene
Plant cost index
$0.03/kWh $0.04/kWh
$1.90/GJ ($2.00/MM Btu) $2.85/GJ ($3.00/MM Btu)
$5.50/Mg ($2.50/M Ib) $8.15/Mg ($3.70/M Ib)
$0.026/kL ($0.10/M gal) $0.037/kL ($0.14/M gal)
$0.20/kg ($0.09/lb)
$2.42/kg
$0.20/kg ($0.09/lb)
$1.90/kg ($0.85/lb)
$2.60/kL ($10/M gal) $2.60/kL ($10/M gal)
Value of natural gas
$0.17/L ($0.63/gal) $0.33/L ($1.25/gal)
210 237
aEPA estimates.
Based on a 10-year life and a 10-percent annual interest.
NOTE: Although a plant does not shut down for 3,500 hr/yr, when demand is
reduced, it is easier to estimate costs this way. This is not
expected to change the estimated costs appreciably.
SOURCES: Benzene Price Quotations. Chemical Marketing Reporter. June 18,
1979.
Economic Indicators. Chemical Engineering. 86(13):7. 1979.
Lawson, J. F- Emission Control Options for the Synthetic Organic
Chemicals Manufacturing Industry Maleic Anhydride—Product Report.
Hydroscience, Inc. (Prepared for Office of Air Quality Planning
and Standards, U.S. Environmental Protection Agency. Research
Triangle Park, N.C.) EPA Contract Number 68-02-2577. March 1978.
Letter from Vatavuk, William M., Office of Air Quality Planning and
Standards, U.S. Environmental Protection Agency, to Joe Lorber,
Hewlett-Packard. March 22, 1978.
5-19
-------
5.3.3 Control Costs for Maleic Anhydride Facilities
Costs for controlling benzene emissions have been developed for each
of the eight benzene-fed maleic anhydride plants for the two regulatory
options of 97 and 99 percent control. Only existing plant costs are
included because no new plants are anticipated before 1983. The maleic
anhydride industry has been operating at 56 percent capacity. Projected
increases in demand could therefore be accommodated by increases in ope-
rating rates over the next few years.
5.3.3.1 Costs to Achieve the 97-Percent Regulatory Option. The costs
for reducing benzene emissions by 97 percent are shown in Tables 5-5a and
5-5b for the carbon adsorption and thermal incineration control systems,
respectively.
Of the eight plants for which costs are shown, the best information
available indicates that three (DENKA, Reichhold in Elizabeth, N.J., and
Koppers) already reduce their benzene emissions by at least 97 percent.
Therefore, the control costs are indicated for only the five other plants.
Of the two control systems costed at 97 percent, the carbon adsorption
systems are inherently more complex and more expensive. But the adsorption
systems have lower direct operating costs, which are, at mid-1979 prices
for benzene, tempered by credits for recovered benzene. As a result, the
total annualized costs are somewhat lower for the adsorption systems at
full capacity and nearly equal at 4,500 hr/yr.
5.3.3.2 Costs to Achieve the 99-Percent Regulatory Option. Tables
5-6a and 5-6b display respective control costs to achieve 99 percent ben-
zene emission reduction via carbon adsorption and thermal incineration,
with heat recovery. Except for the Koppers plant, all benzene-fed plants
require additional control to achieve this level. Table 5-6b lists costs
for a thermal incinerator with primary heat recovery. As capacity
increases from 11,800 to 38,500 Mg/yr (26 to 85 million Ib/yr), the
installed cost increases from $0.78 million to $1.88 million--only slightly
more than the incineration capital costs in Table 5-5b, which were for 97
percent control. However, because of high fuel costs, the direct operating
costs for incineration at 99 percent control are significantly higher than
those for 97 percent control. These costs range from about $0.68 to
$2.15 million/yr at 8,000 hr/yr.
5-20
-------
TABLE 5-5a.
in
i
PO
EXISTING PLANT COSTS FOR ACHIEVING 97 PERCENT BENZENE EMISSION REDUCTION
CONTROL METHOD: CARBON ADSORPTION
Direct operating
cost ($/yr)a
Plant name and location
1.
2.
3.
4.
5.
6.
7.
8.
Ashland—Neal, W. Va.
Monsanto--St. Louis, No.
DENKA— Houston, Tex.f
Reichhold— Elizabeth, N.J.f
Reichhold—Morris, 111.
Tenneco — Fords, N.J.
U.S. Steel—Neville Island,
Pa.
Koppers—Bridgeville, Pa.f
Capacity
(Mg/yr)
27,200
38,100*
22.700
13,600
20,000
11,800
38,500
15,400
Total
installed
cost ($)a>c
1,650
2,240
0
0
1,250
780
2,260
0
ACC
($/yr)a
335
454
0
0
254
158
458
0
4,500d
hr/yr°
467
637
0
0
354
225
643
0
8,000
hr/yr
739
1.018
0
0
554
334
1,028
0
Benzene
credit
4,500
hr/yr
(374)
(523)
0
0
(274)
(162)
(528)
0
recovery
(*/yr
8,000
hr/yr
(664)
(930)
0
0
(488)
(288)
(939)
0
Total annual ized cost .
($/y) ($/Mg product)0
4,500
hr/yr
428
568
0
0
334
221
573
0
8.000
hr/yr
410
542
0
0
320
213
547
0
4,500
hr/yr
28.6
26.9
0
0
30.5
34.7
26.9
0
8,000
hr/yr
15.4
14.5
0
0
16.4
18.8
14.5
0
Dollars are stated in thousands.
Includes the cost of benzene continuous monitoring (approximately $9,000/yr per plant).
clnstalled costs are rounded to the nearest 10 thousand dollars; other costs are rounded to the nearest thousand.
Operating factor represents production at 56 percent of capacity.
eThis represents 80 percent of the total plant capacity. The rest -of the MA is produced from n-butane.
These plants already achieve 97 percent reduction.
SOURCES: Benzene Price Quotation. Chemical Marketing Reporter. June 18, 1979.
Blacker, Herbert G., and Thomas M. Nichols. Capital and Operating Costs of Pollution Control Equipment Modules—Vol. II--
Data Manual. ICARUS Corp. Silver Spring, Md. (Prepared for Office of Research and Monitoring, U.S. Environmental Protection
Agency. Washington, D.C.) EPA-R5-73-023b. July 1973.
Economic Indicators. Chemical Engineering. 86(13):7. 1979.
Guthrie, Kenneth, M. Process Plant Estimating Evaluation and Control. Los Angeles, Craftsman Book Company of America, 1974.
Letter from Hewett, P. S., Reichhold Chemicals, Inc., to Patrick, D. R., Office of Air Quality Planning and Standards,
U.S. Environmental Protection Agency. March 27, 1978.
Kinkley, M. T., and R. B. Neverill. Capital and Operating Costs of Selected Air Pollution Control System. CARD, Inc. Niles,
EPA-450/3-76-014. May 1976.
Lawson, J. F. Emission Control Options for the Synthetic Organic Chemicals Manufacturing Industry Maleic Anhydride—Product
Report. Hydroscience, Inc. (Prepared for Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency.
Research Triangle Park, N.C.) EPA Contract Number 68-02-2577. March 1978.
Letter from Vatavuk, William M., to Weber, Robert, Office of Air Quality Planning and Standards, U.S. Environmental Protection
Agency. March 28, 1978.
Letter from Vatavuk, William M., to Weber, Robert, Office of Air Quality Planning and Standards, U.S. Environmental Protection
Agency. June 27, 1978.
111.
-------
TABLE 5-5b. EXISTING PLANT COSTS FOR ACHIEVING 97 PERCENT BENZENE EMISSION REDUCTION
CONTROL METHOD: THERMAL INCINERATION WITH PRIMARY HEAT RECOVERY
in
ro
Direct operating
cost ($/yr)a
Plant name and location
1.
2.
3.
4.
5.
6.
7.
8.
Ashland--Neal, W. Va.
Monsanto--St. Louis, Mo.
OENKA--Houston, Tex.f
Reichhold--Elizabeth, N.J.f
Reichhold— Morris, 111.
Tenneco--Fords, N.J.
U.S. Steel—Neville Island,
Pa.
Koppers—Bridgeville, Pa.
Capacity
(Hg/yr)
27,200
38,100e
22,700
13,600
20,000
11,800
38,500
15,400
Total
installed
cost ($)a'c
1,320
1,730
0
0
1,070
730
1,750
0
ACC
($/yr)a
260
351
0
0
216
149
355
0
4,500,
hr/yr
701
966
0
0
503
302
976
0
8,000
hr/yr
1,176
1,632
0
0
876
531
1.650
0
Benzene
credit
4,500
hr/yr
(528)
1739)
0
0
(388)
(229)
(747)
0
recovery
($/yr)a
8,000
hr/yr
(938)
(1.314)
0
0
(690)
(407)
(1.328)
0
Total annual ized cost .
($/yr)a ($/Mg product)0
4,500
hr/yr
433
578
0
0
331
222
584
0
8,000
hr/yr
498
669
0
0
402
273
677
0
4,500
hr/yr
28.9
27.3
0
0
30.2
34.8
27.3
0
8,000
hr/yr
18.6
17.8
0
0
20.6
23.9
17.8
0
aOoliars are stated in thousands.
Includes the cost of benzene continuous monitoring (approximately $9,000/yr per plant).
Installed costs are rounded to the nearest 10 thousand dollars; other costs are rounded to the nearest thousand.
Operating factor represents production at 56 percent of capacity.
eThis represents 80 percent of the total plant capacity. The rest of the MA is produced from n-butane.
This plant already achieves 97 percent reduction.
SOURCES: Economic Indicators. Chemical Engineering. 86(13):7. 1979.
Guthrie, Kenneth, M. Process Plant Estimating Evaluation and Control. Los Angeles, Craftsman Book Company of America, 1974.
Letter from Hewett, P. S., Reichhold Chemicals, Inc., to Patrick, D. R., Office of Air Quality Planning and Standards,
U.S. Environmental Protection Agency. March 27, 1978.
Kinkley, M. T., and R. B. Neverill. Capital and Operating Costs of Selected Air Pollution Control System. CARD, Inc. Niles, 111.
EPA-450/3-76-014, May 1976.
Lawson, J. F. Emission Control Options for the Synthetic Organic Chemicals Manufacturing Industry Maleic Anhydride—Product
Report. Hydroscience, Inc. (Prepared for Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency.
Research Triangle Park, N.C.) EPA Contract Number 68-02-2577. March 1978.
Letter from Vatavuk, William M., Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency, to
Weber, Robert, Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency. March 28, 1978.
Letter from Vatavuk, William M., Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency, to
Weber, Robert, Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency. June 27, 1978.
-------
TABLE 5-6a.
EXISTING PLANT COSTS FOR ACHIEVING 99 PERCENT BENZENE EMISSION REDUCTION
CONTROL METHOD: CARBON ADSORPTION
Ol
N>
CJ
Plant name and location
1.
2.
3.
4.
5.
6.
7.
8.
Ashland— Neal, W. Va.
Monsanto--St. Louis, Mo.
DENKA— Houston, Tex.
Reichhold—Elizabeth, N.J.
Re ichhold— Morris, 111.
Tenneco--Fords, N.J.
U.S. Steel— Neville Island,
Pa.
Koppers — Bridgeville, Pa.
Total
Capacity installed ACC
(Mg/yr) cost ($)a'c ($/yr)a
27,200
38,100*
22,700
13.600
20,000
11,800
38,500
15,400
1,680
2,270
1,440
900
1.280
810
2,290
0
341
460
292
183
260
164
464
0
Direct
cost
4,500
hr/yr°
481
657
409
261
365
232
663
0
operating
($/yr)a
8,000
hr/yr
764
1,052
645
402
573
355
1,063
0
Benzene
credit
4,500
hr/yr
(382)
(536)
(319)
(191)
(281)
(166)
(541)
0
recovery
($/yr)a
8,000
hr/yr
(680)
(952)
(568)
(340)
(500)
(295)
(962)
0
Total annual ized cost .
($/yr)a ($/Mg product)0
4,500
hr/yr
440
581
382
253
344
230
586
0
8,000
hr/yr
425
560
369
245
333
224
565
0
4,500
hr/yr
29.3
27.5
30.6
34.2
31.4
36.0
27.5
0
8,000
hr/yr
16.0
14.9
16.7
18.7
17.1
19.7
14.9
0
aDollars are calculated by thousands.
Includes the cost of benzene continuous monitoring (approximately $9,000/yr per plant).
clnstalled costs are rounded to the nearest 10 thousand dollars; other costs are rounded to the nearest thousand.
Operating factor represents production at 56 percent of capacity.
This represents 80 percent of the total plant capacity. The rest of the MA is produced from n-butane.
f *
This plant already achieves 99 percent reduction.
SOURCES: Benzene Price Quotation. Chemical Marketing Reporter. June 18, 1979.
Blacker, Herbert G. , and Thomas M. Nichols. Capital and Operating Costs of Pollution Control Equipment Modules--Vol. II--Data
Manual. ICARUS Corp. Silver Spring, Md. (Prepared for Office of Research and Monitoring, U.S. Environmental Protection Agency.
Washington, O.C.) EPA-R5-73-023b. July 1973.
Economic Indicators. Chemical Engineering. 86(13):7. 1979.
Guthrie, Kenneth M. Process Plant Estimating Evaluation and Control. Los Angeles, Craftsman Book Company of America, 1974.
Letter from Hewett, P. S., Reichhold Chemicals, Inc., to Patrick, D. R., Office of Air Quality Planning and Standards,
U.S. Environmental Protection Agency. March 27, 1978.
Kinkley, M. T., and R. B. Neverill. Capital and Operating Costs of Selected Air Pollution Control System. CARD, Inc. Niles,
111. EPA-450/3-76-014. May 1976.
Lawson, J. F. Emission Control Options for the Synthetic Organic Chemicals Manufacturing Industry Maleic Anhydride—Product
Report. Hydroscience, Inc. (Prepared for Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency.
Research Triangle Park, N.C.) EPA Contract Number 68-02-2577. March 1978.
Letter from Vatavuk, William M., Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency, to
Weber, Robert, Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency, March 28, 1978.
Letter from Vatavuk, William M., Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency, to
Weber, Robert, Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency, March 8, 1978.
-------
TABLE 5-6b. EXISTING PLANT COSTS FOR ACHIEVING 99 PERCENT BENZENE EMISSION REDUCTION
CONTROL METHOD: THERMAL INCINERATION WITH PRIMARY HEAT RECOVERY
Direct operating
cost ($/yr)a
Plant name and location
1.
2.
3.
4.
5.
6.
7.
8.
Ashland--Neal , W. Va.
Monsanto--St. Louis, Mo.
DENKA--Houston, Tex.
Reichhold--Elizabeth, N.J.
Reichhold--Morris, 111.
Tenneco--Fords, N.J.
U.S. Steel— Neville Island,
Pa.
Koppers—Bridgeville, Pa.
Capacity
(Mg/yr)
27,200
38,100e
22,700
13,600
20,000
11,800
38,500
15,400
Total
installed
cost ($)a'c
1,400
1,860
1,220
nsn
1,130
780
1,880
0
ACC ,
($/yr)a
284
378
247
173
229
158
382
0
4,500,,
hr/yra
899
1,246
757
468
672
411
1,259
0
8,000
hr/yr
1,526
2,124
1,280
782
1,133
683
2,146
0
Energy recoverv
credit ($/yr)B
4,500
hr/yr
(625)
(876)
(522)
(313)
(460)
(271)
(885)
0
8,000
hr/yr
(1.111)
(1,557)
(927)
(557)
(817)
(482)
(1,573)
0
Total annual ized cost ..
($/yr)a ($/Mg product)
4,500
hr/yr
558
748
482
328
441
298
756
0
8,000
hr/yr
699
945
600
398
545
359
955
0
4,500
hr/yr
37.1
35.3
38.5
44.1
40.0
46.3
35.3
0
8,000
hr/yr
26.0
25.0
26.8
29.9
27.7
30.4
25.0
0
Dollars are calculated by thousands.
Includes the cost of benzene continuous monitoring (approximately $9,000/yr per plant).
' clnstalled costs are rounded to the nearest 10 thousand dollars; other costs are rounded to the nearest thousand.
•^ Operating factor represents production at 56 percent of capacity.
eThis represents 80 percent of the total plant capacity. The rest of the MA is produced from n-butane.
This plant already achieves 99 percent reduction.
SOURCES: Economic Indicators. Chemical Engineering. 86(13):7. 1979.
Guthrie, Kenneth M. Process Plant Estimating Evaluation and Control. Los Angeles, Craftsman Book Company of America,
1974.
Letter fror.. Hewett, P. S., Reichhold Chemicals, Inc., to Patrick, D. R., Office of Air Quality Planning and Standards,
U.S. Environmental Protection Agency. March 27, 1978.
Kinkley, M. T., and R. B. Neverill. Capital and Operating Costs of Selected Air Pollution Control System. GARO, Inc. Miles,
111. EPA-450/3-76-014. May 1976.
Lawson, J. F. Emission Control Options for the Synthetic Organic Chemicals Manufacturing Industry Maleic Anhydride—Product
Report. Hydroscience, Inc. (Prepared for Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency.
Research Triangle Park, N.C.) EPA Contract Number 68-02-2577. March 1978.
Letter from Vatavuk, William M., Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency, to
Weber, Robert, Offic' of Air Quality Planning and Standards, U.S. Environmental Protection Agency. March 28, 1978.
Letter from Vatavuk, William M., Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency, to
Weber, Robert, Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency. March 8, 1978.
-------
As in the incineration systems for 97 percent control, a heat exchanger
is built into the incinerator unit to preheat the incoming air, while
cooling the exhaust gas. In this way, about 50 percent of the total heat
duty of the incinerator is recovered, as shown in Table 5-6b. Depending on
plant size and operating mode, these credits amount to 70 to 80 percent of
the direct operating costs.
The carbon adsorption system costs shown in Table 5-6a also represent
achievement of a 99-percent control efficiency. Ranging from $0.81 to
$2.29 million, the installed costs for the 99-percent adsorbers are moder-
ately higher than those for the incineration systems. However, lower
direct operating costs are generated for the adsorption systems for incin-
eration, and a larger fraction (80 to 90 percent) of these operating costs
is offset by credits for recovered benzene at today's prices. The end
result is that the total annualized cost, for 99 percent control at full
capacity, is conspicuously less for adsorption than for incineration: a
range of $224,000 to $565,000/yr as opposed to a range of $359,000 to
$955,000/yr. As production is cut back, the advantage of adsorption is
gradually lost because the process is capital intensive. Yet even at
4,500 hr/yr, adsorption is preferable.
5.3.3.3 Comparison of Control Levels. When an incinerator design is
revised from 760° C (1,400° F) for 97 percent control to 870° C (1,600° F)
for 99 percent control, construction materials, insulation thickness, and
firebox volume undergo changes that affect capital costs. Fuel expense and
maintenance necessarily increase. The addition of a heat exchanger for
nominal, 50-percent primary recovery incurs more capital expense, but this
expense should be more than offset by fuel savings. (Strictly speaking,
the optimal degree of primary heat recovery depends on scale, temperature,
and anticipated fuel costs. No attempt has been made here to tailor recov-
ery to the different situations.) The costs shown in Tables 5-5b (97 per-
cent) and 5-6b (99 percent) reflect engineering estimates of these factors.
Unlike an incinerator, a carbon adsorption unit is not normally designed
to a specified efficiency. The underlying science and engineering are
inexact, and the adsorptive capacity of the carbon diminishes unpredict-
ably with use. If an adsorption system is built to operate as efficiently
as possible, the net annualized operating costs will be insensitive to
5-25
-------
control level. It is almost certain, however, that the efficiencies required
here are greater than the economic optimum and that 99 percent control
costs more than 97 percent control. Certainly, design and operating changes
exist that would increase both control efficiency and cost. To represent
this effect meaningfully, in the absence of a strict design method, the
amount of carbon required is assumed to be directly proportional to the
desired control efficiency. (Clearly, this assumption fails at 100 percent.)
The capital and operating cost differentials between Tables 5-5a and 5-6a
result from this assumption, but the end result is an annualized cost that
is almost unaffected by the control level.
5.3.3.4 Monitoring. Table 5-7 contains costs for continuous monitor-
ing of benzene stack emissions from both the 97- and 99-percent emission
controls using a gas chromatograph with a flame ionization detector. The
installed cost of the chromatograph and its auxiliaries is $35,000. Depend-
ing on the operating factor, the direct operating cost varies from about
$1,450 to $l,680/yr. When the annualized capital charges are added, the
total annualized cost amounts to $8,600 and $8,800/yr for the 4,500- and
8,000-hr/yr cases, respectively. These costs are relatively low when com-
pared to the control costs in Tables 5-5 and 5-6, where the annualized
monitoring cost has been rounded to $9,000 for all cases.
5.3.4 Cost Effectiveness of the Alternative Emissions Limits
For each of the control methods costed to achieve one or more of the
alternative emission limits considered, it is informative to compare the
total annualized cost with the amount of benzene removed. A convenient
yardstick for expressing this comparison is the cost-effectiveness ratio,
the quotient of the annualized cost and the quantity of benzene removed
annually, expressed in dollars per megagram of benzene removed.
These ratios and other important cost data appear in the cost summary
table, Table 5-8. Three kinds of data are shown: total annualized cost,
the amount of benzene removed annually by the control method, and the
cost-effectiveness ratio. Because the total annualized cost is expressed
in dollars per megagram of maleic anhydride produced annually, two costs
are given: one computed at 4,500 hr/yr operating factor, the other at full
capacity, or 8,000 hr/yr.
5-26
-------
TABLE 5-7. COSTS FOR CONTINUOUS MONITORING OF
BENZENE STACK EMISSIONS3
Operating factor (hr/yr)
Cost
Installed ($)
j i rer ,
Annual
Total
op< ; .";•> utq
iZwU -dp itu
annual i zed
($ ';. -
1 CVjfr)
($/yr) (rounded)5
4,500
35,000
1.450
7.100
8,600
8,000
35,000
1/bO
7,100
8,800
These costs apply to all existing plants using benzene feed.
Includes: gas chromatograph with flame ionization detector, automatic
gas sampling value, air sampler, post-run calculator, and gas regulators.
SOURCE: Letter from Vatavuk, William M., Office of Air Quality Planning and
Standards, U.S. Environmental Protection Agency, to Lorber, Joe,
Hewlett-Packard. March 22, 1978. (Updated to mid-1979 by rules
of thumb.)
5-27
-------
TABLE 5-8. COST SUMMARY FOR EXISTING MALEIC ANHYDRIDE PLANTS
cn
i
ro
oo
97 Percent Benzene Emission Reduction
Plant name and location
Ashland
Monsanto (St. Louis, Mo.)
DENKA (Houston, Texas)
Reichhold (Elizabeth, N.J.)
Reichhold (Morris, 111.)
Tenneco (Fords, N.J. )
U.S. Steel (Neville Island, Pa.)
Koppers (Pa. )d
Thermal
4,
TAC
($/
Mg
prod-
uct)3
28.9
27.3
-
-
30.2
34.8
27.3
incineration with primary
500 hr/yr
Ben- Cost-
zene- effec-
removed tive-
(MgA ness
yr)b ($/Mg)c
998 443
1,398 420
-
-
734 463
433 533
1,413 420
heat recovery
8,000 hr/yr
TAC
18.6
17.8
-
-
20.6
23.9
17.8
Ben-
zene
re-
moved
1,773
2,484
-
-
1,304
769
2,510
99 Percent Benzene
Plant name and location
Ashland (Neal , W. Va. )
Monsanto (St. Louis, Mo.)
DENKA (Houston, Texas)
Reichhold (Elizabeth, N.J.)
Reichhold (Morris, 111.)
Tenneco (Fords, N.J.)
U.S. Steel (Neville Island, Pa.)
Koppers (Pa. )d
Thermal
4,
TAC
($/
Mg
prod-
uct)3
37.1
35.3
38.5
44.1
40.0
46.3
35.3
incineration with primary
500 hr/ yr
Ben- Cost-
zene effec-
removed tive-
(Mg/. ness
yr)D ($/Mg)c
1,018 557
1,397 542
832 590
499 675
773 614
433 709
1,412 542
Cost-
effec-
tive-
ness
286
273
-
-
314
367
273
Emission
TAC
28.6
26.9
-
-
30.5
34.7
26.7
Carbon adsorption
4,500 hr/yr
Ben-
zene
re-
moved
998
1,398
-
-
734
433
1,413
Cost-
effec-
tive-
ness
438
411
-
-
467
531
411
TAC
15.4
14.5
-
-
16.4
18.8
14.5
Ben-
zene
re-
moved
1,773
2,484
-
-
1,304
769
2,510
8,000 hr/yr
Cost-
effec-
tive-
ness
236
220
-
-
292
290
220
Reduction
heat recovery
8,000 hr/yr
TAC
26.0
25.0
26.8
29.9
27.7
30.4
25.0
Ben-
zene
re-
moved
1,810
2,484
1,480
887
1,304
764
2,510
Cost-
effec-
tive-
ness
391
384
411
454
425
479
384
TAC
29.3
27.5
30.6
34.2
31.4
36.0
27.5
4,500 hr/yr
Ben-
zene
re-
moved
1,018
1.397
832
499
773
433
1,412
Carbon adsorption
Cost-
effec-
tive-
ness
441
422
470
525
482
552
422
TAC
16.0
14.9
16.7
18.7
17.1
19.7
14.9
8,000
Ben-
zene
re-
moved
1,810
2,484
1,480
887
1,*304
769
2,510
hr/yr
Cost-
effec-
ti ve-
ness
240
229
255
286
262
303
229
Includes costs for continuous monitoring.
Product of the inlet emission rate and the control efficiency.
cQuotient of the total annualized cost ($/yr) and the
benzene removed (Mg/yr).
Koppers is expected to incur no costs under either
control level.
-------
SOURCES: Blacker, Herbert G., and Thomas M. Nichols. Capital and Operating Costs of Pollution Control Equipment Modules--Vol. II--
Data Manual. ICARUS Corp. Silver Spring, Md. '(Prepared for Office of Research and Monitoring, U.S. Environmental
Protection Agency. Washington, D.C.) Report Number EPA-R5-73-023b. July 1973.
Benzene Price Quotations. Chemical Marketing Reporter. June 18, 1979.
Economic Indicators. Chemical Engineering. 86(13):7. 1979.
Guthrie, Kenneth M. Process Plant Estimating Evaluation and Control. Los Angeles, Craftsman Book Company of America, 1974.
Letter from Hewett, P. S., Reichhold Chemicals, Inc., to Patrick, 0. R., Office of Air Qualify Planning and Standards,
U.S. Environmental Protection Agency. March 27, 1978.
Kinkley, M. T., and R. B. Neverill. Capital and Operating Costs of Selected Air Pollution Control System. Hydroscience,
Inc. EPA-450/3-76-014. May 1976.
Lawson, J. F. Emissions Control Options for the Synthetic Organic Chemicals Manufacturing Industry Maleic Anhydride—Product
Report. Hydroscience, Inc. (Prepared for Office of Air Quality, Planning and Standards, U.S. Environmental Agency.
Research Triangle Park, N.C. EPA Contract Number 68-02-2577. March 1978.
Letter from Vatavuk, William M., Office of Air Quality, Planning, and Standards, U.S. Environmental Protection Agency, to
Lorber, Joe, Hewlett-Packard. March 22, 1978.
Letter from Vatavuk, William M., Office of Air Quality, Planning, and Standards, U.S. Environmental Protection Agency, to
Weber, Robert, Office of Air Quality, Planning, and Standards, U.S. Environmental Protection Agency. March 28, 1978.
Letter from Vatavuk, William M., Office of Air Quality, Planning, and Standards, U.S. Environmental Protection Agency,
to Weber, Robert, Office of Air Quality, Planning, and Standards, U.S. Environmental Protection Agency. June 27, 1978.
01
I
f\>
vo
-------
Each "benzene-removed" number is the product of the inlet emission
loading to the control device (see Table 5-3) and the control efficiency.
Figures 5-6 and 5-7 depict the cost-effectiveness ratios for the
4,500- and 8,000-hr/yr operating factors, respectively. To display the
data better, the vertical axes on these figures have been expanded.
In Figures 5-6 and 5-7, note that the 99-percent curves lie above the
97-percent curves for incineration but not for adsorption. These results
can be explained by considering the two control techniques employed—
thermal incineration and carbon adsorption.
For incineration, the lower combustion temperature employed at the
97-percent control level (as opposed to the 99-percent level) leads to
conspicuously lower capital and operating costs for the system, with only
slight changes in the amount removed. Consequently, cost effectiveness
improves somewhat with adoption of the 97-percent control level, although
it might decrease at lower control levels.
For carbon adsorption, the similarity between the 97- and 99-percent
control level costs is due to the assumptions described earlier, employed
in estimating costs at 97 percent control. Neither capital nor operating
costs, modified by benzene credits, change significantly between the two
cases. This would not be true if one sought more stringent control than
99 percent.
Finally, note that in both Figures 5-6 and 5-7 the curves slope grad-
ually downward with increasing plant capacity, indicating the expected
economy of scale. For plant sizes above 40,000 Mg/yr (8.8 x 10 Ib/yr),
the cost-effectiveness ratios may seem to approach asymptotic values. If
design parameters are held constant as plant capacity increases, the influ-
ence of the fixed costs on the total annualized cost becomes less and less
pronounced, while the variable costs—those nearly proportional to produc-
tion capacity--become more influential. In other words, the total annual-
ized cost becomes nearly proportional to the production capacity. Then,
because the amount of benzene removed is exactly proportional to production
capacity, the ratio of these quantities approaches a constant value.
It is normal in process design, however, that such parameters as the
liguid/gas ratio in the reactor vent scrubber or the fractional heat recovery
in the incinerator, assume different optimal values at different scales.
5-30
-------
860 -
840 -
820 -
800
780 -
760 -
_ 740 -
Q
UJ
| 720
at
ff
700
N
680
660
| 640
o
Ul
It 620
UJ
t-
8 600
580
560
540
520
500
I: CARBON ADSORPTION • 99 PERCENT CONTROL
II: THERMAL INCINERATION - 99 PERCENT CONTROL
III: CARBON ADSORPTION - 97 PERCENT CONTROL
IV: THERMAL INCINERATION - 97 PERCENT CONTROL
10
40
20 30
PLANT CAPACITY (Gg/yr)
Figure 5-6. Cost effectiveness of alternative control systems—operating factor 4,500 hours,
« with monitoring.
5-31
-------
O
520
500
480
460
440
420
400
ui
UJ
CD
O
§
380
360
340
320
300
280
260
240
220
200 -
I: CARBON ADSORPTION - 99 PERCENT CONTROL
II: THERMAL INCINERATION - 99 PERCENT CONTROL
III: CARBON ADSORPTION • 97 PERCENT CONTROL
IV: THERMAL INCINERATION - 97 PERCENT CONTROL
III, I
10
20 30
PLANT CAPACITY (Gg/yr)
40
Figure 5-7. Cost effectiveness of alternative control systems—operating factor 8,000 hours,
with monitoring.
5-32
-------
Ordinarily, the consequence is that the large-scale version is more "effi-
cient" in a variety of ways, and both fixed and variable costs increase
less rapidly than productive capacity.
5.3.5 Control Cost Comparison
Before the accuracy and representativeness of control costs can be
ascertained, they must be compared with costs obtained from other data
sources. In doing this, one can either compare the installed capital
costs, the annualized costs, or both. However, because the capital costs
influence the annualized costs (via the annualized capital charges) and
because the other terms in the annualized cost contain more site-specific
variability (utilities, for instance), it is preferable to limit the com-
parison to the installed costs.
Even for a control system sized for a particular emission point, the
installed cost may vary considerably from site to site. Such factors as
the cost of installation labor (electricians, pipefitters, etc.), the
requirement of special installation materials (e.g., extra insulation for
systems installed in colder climates), and the presence or absence of
excess utility capacity considerably influence the total installed cost.
With this in mind, however, capital cost comparisons can be made among
a range of control system sizes. Such comparisons are best made graphically;
that is, installed costs adjusted to the same reference date (second quarter
1979, in this case) are plotted against some technical parameter relevant
to the control system. In this section, installed costs are compared among
various sizes of carbon adsorbers and thermal incinerators with primary
heat recovery. The technical parameters used in the comparisons are the
gas volumetric flow rates at the control system inlet and outlet, respec-
tively.
For the carbon adsorbers, capital costs developed for the existing
plants (Tables 5-5a and 5-6a) are compared with costs developed by an EPA
contractor and with an actual cost for a system installed at a maleic
anhydride plant. These costs have been plotted against the inlet volu-
metric flow rate on logarithmic scales (Figure 5-8). Note that two EPA
curves appear in the figure. The top EPA curve applies to a 99-percent
efficient adsorber, while the one beneath it is a unit designed for 97
percent efficiency.
5-33
-------
Consider first the 99-percent adsorbers. Figure 5-8 shows that the
contractor's costs, after adjustment to the same time base, exceed the EPA
figures over the entire range of 730 to 2,000 m /min. However, the differ-
ences between the costs range only from 5 to 21 percent, which falls within
the ±30 percent accuracy range assigned to the costs.
Figure 5-9 displays the installed costs for thermal incineration
systems with primary heat recovery. The EPA cost at this flow rate, 5,920
m3/min (2.1 x 105ft3/min), is $1.40 million, which is $850,000 (60 percent)
less than the industry figure. This discrepancy results primarily from a
design difference between the two units. The plant system contains a
boiler for heat recovery, while the EPA system uses a simple gas-to-gas
heat exchanger. As expected, the former system is more costly.
On the other hand, EPA costs are significantly higher than a contrac-
tor's trend line shown in Figure 5-9, both of which represent 99 percent
control efficiency. The cost difference varies from 74 to 76 percent over
the flow rate range of 2,200 to 8,000 m3/min (78,000 to 280,000 ft3/min).
After an examination of itemized costs from these two data sources, it
appears that most of the discrepancy is caused by the costs of the respec-
tive incinerator chambers. In addition, the EPA figures include costs for
ductwork and stacks, while the contractor costs do not. Differences aris-
ing from variation in scope are not unusual.
5.4 ECONOMIC IMPACT OF REGULATORY OPTIONS
5.4.1 Introduction
This section discusses the economic impact of two possible regulatory
options for existing MA plants--97 percent and 99 percent benzene emission
reductions. Subsection 5.4.2 describes the impacts on individual MA manu-
facturers because of increased capital budget requirements, shifts in
competitive position due to unequal control costs, and other possible costs
borne by MA producers to meet other environmental standards. Subsection
5.4.3 describes the impact on the price of products that use MA. Sub-
section 5.4.4 describes effects on employment and balance of trade. Finally,
Subsection 5.4.5 summarizes the annualized costs incurred 5 years from now.
5-34
-------
o
u
Q
ui
V)
Z
10'
,6-
CONTRACTOR
(99 PERCENT EFFICIENCY)
EPA (99 PERCENT EFFICIENCY)
EPA (97 PERCENT EFFICIENCY)
7
102
103
INLET GAS FLOW RATE (ACTUAL m3/min)
Figure 5-8. Installed costs of carbon adsorbers.
SOURCES: Lawson, J.F. Emission Control Options for the Synthetic Organic Chemicals Manufacturing
Industry Maleic Anhydride—Product Report. Hydroscience, Inc. (Prepared for Office of Air
Quality Planning and Standards, U.S. Environmental Protection Agency. Research Triangle
Park, N.C.) EPA Contract Number 68-02-2577. March 1978.
Letter from Lawson, John F., Hydroscience, Inc., to Weber, Robert, Office of Air Quality
Planning and Standards, U.S. Environmental Protection Agency. November 1, 1977.
5-35
-------
10
7_
i
Q
ui
10°-
EPA (99 PERCENT EFFICIENCY)
CONTRACTOR
(99 PERCENT EFFICIENCY)
103
10s
OUTLET GAS FLOW RATE (ACTUAL m3/min)
Figure 5-9. Installed costs of thermal incinerators with primary heat recovery.
SOURCE: Lawson, J.F. Emission Control Options for the Synthetic Organic Chemicals Manufacturing
Industry Maleic Anhydride-Product Report. Hydroscience, Inc. (Prepared for Office of Air
Quality Planning and Standards, U.S. Environmental Protection Agency. Research Triangle
Park. N.C.) EPA Contract Number 68-02-2577. March 1978.
Pruessner, R.O., and L.D. Broz. Air Pollution Control: Hydrocarbon Emission Reduction
Systems. Chemical Engineering Progress. Petrotex Chemical Corp.:73. Houston, Texas.
August 1977. p. 69-73.
5-36
-------
5.4.2 Impact on Manufacturers
The direct economic effect of a benzene control standard would fall on
MA manufacturers that use benzene as a feedstock. This impact, which
depends upon the level of benzene control required, principally involves
two areas—capital budget requirements and competition from domestic and
foreign producers. The substitutes for MA are few and relatively unimpor-
tant in its primary markets, so little interindustry impact is expected.
5.4.2.1 Capital Budget Requirements. Two control levels for reducing
benzene emissions are being considered: 97 percent and 99 percent reduc-
tion. At the 97-percent emission level, five plants will require no invest-
ment in controls because they already meet the standards. These plants are
Koppers (Pa.), DENKA, Reichhold (N.J.)—which presently achieve this level
of control—and Koppers (111.) and Amoco, which do not use benzene as a
feedstock. Of these five plants, the Amoco and the two Koppers plants also
meet the 99-percent emission level. Table 5-9 shows the capital budget
requirements for plants requiring either carbon adsorption or incineration
to achieve the standards being considered. The investment costs given in
the table do not credit partial controls already in place on MA plants
presently not meeting a possible standard. Thus, for plants presently
meeting some level of benzene control, capital outlay assumptions for the
97- or 99-percent emission reduction standard may represent a "worst case."
In discussion of capital budget requirements, the main question is
whether MA producers will have the financial resources from which to fund
the initial capital investment. Funding can occur either externally, by
sales of debt or equity, or internally, by using current capital budgets or
allocating funds from the capital budgets of other divisions in the parent
company. In addition to these financing techniques, costs may be passed
through to consumers by raising the price of MA or other products produced
by the firm to recover the capital expenses over a time period. Finally,
although the method or combination of methods firms may employ to finance
the additional capital expenditures cannot be predicted, it is likely that
the expenditures will be treated as part of the parent company's overall
capital budget.
In addition to the question of how the control costs might be financed
is the question of whether such an investment is worthwhile. This decision
5-37
-------
TABLE 5-9. ESTIMATED TOTAL INVESTMENT COST FOR ACHIEVING BENZENE EMISSION REDUCTION
tn
i
to
00
Plant name & location
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
Ashland
Neal, W. Va.
Monsanto
St. Louis, Mo.
DENKA
Houston, Tex.
•
Reichhold
Elizabeth, N.J.
Reichhold
Morris, 111.
Tenneco
Fords, N.J.
U.S. Steel
Neville Island, Pa.
Koppers
Bridgeville, Pa.
Koppers
Chicago, 111.
Amoco
Joliet, 111.
Type & level of control
CA
CA
I
I
CA
CA
I
I
CA
CA
I
I
CA
CA
I
I
CA
CA
I
.1
CA
CA
I
I
CA
CA
I
I
CA/I
CA/I
CA/I
97%
99%
97%
99%
97%
99%
97%
99%
97%
99%
97%
99%
97%
99%
97%
99%
97%
99%
97%
99%
97%
99%
97%
99%
97%
99%
97%
99%
97-99%
97-99%
97-99%
Control equipment
investment cost
(103 dollars)
1650
1680
1320
1400
2240
2270
1730
1860
0
1440
0
1220
0
900
0
850
1250
1280
1070
1130
780
810
730
730
2250
2290
1750
1880
0
0
0
Monitoring equipment
investment cost
(103 dollars)
30
30
30
30
30
30
30
30
UNK
30
UNK
30
UNK
30
UNK
30
30
30
30
30
30
30
30
30
30
30
30
30
UNK
UNK
UNK
Total investment
cost
(103 dollars)
1680
1710
1350
1430
2270
2300
1760
1890
0
1470
0
1250
0
930
0 -
880
1280
1310
1100
1160
810
840
760
760
2280
2350
1780
1910
0
0
0
NOTES: CA = carbon adsorption; I = incineration.
UNK indicates that the cost of monitoring equipment for those firms now meeting certain standard levels is unknown.
-------
depends on the importance of MA to total company sales, capital availability,
and the rate of return and risks relative to competing investment opportun-
ities. Of the seven companies that could be affected by the alternative
benzene standards being considered, five appear able to meet the additional
capital budget requirements at both the 97- and 99-percent control levels.
This assessment is based on analyses of physical and financial characteris-
tics of individual plants as well as communications with industry. The
other two companies may either choose not to or cannot finance such an
investment.
One of these companies, DENKA, expressed concern that it could not
-jc -1C
finance additional benzene pollution controls. ' The company has already
made a major capital outlay for benzene controls within the past 4 years
and is still recovering from this investment. It is a two-product company,
with MA sales representing 33 percent of its total company sales. An
investment of another 1.3 to 1.5 million dollars in new pollution control
equipment could cause DENKA to go out of business, according to a company
spokesperson. '
Although not confirmed by a company spokesperson, another possible
candidate for closure is Tenneco. This opinion is based on several factors.
First, Tenneco's plant, built in the early 1960's, is one of the oldest
domestic MA plants. The number of remaining viable operating years at the
plant, which may be few, will be a critical factor in an investment decision.
Second, based on the economic analysis performed in this section, cost
increments associated with a benzene standard would be highest for Tenneco
under both control scenarios. Tenneco may therefore be at a cost disadvan-
tage relative to its MA competitors. Third, it may be more profitable for
the company to invest in products that are a more critical part of its
q
product mix, or yield a higher return. Tenneco1s MA sales as a percentage
of total company sales are the lowest of all affected companies, 0.05 percent
(see Table 5-10). Finally, Tenneco's MA plant is located in a highly
industralized area of New Jersey. Many chemical companies have already
left this area and relocated in the South, where production and raw material
costs are presently less expensive.
In summary, DENKA's possible closure response to a benzene standard is
a stated company position and has been confirmed by an independent analysis.
5-39
-------
TABLE 5-10. RATIO OF MA SALES TO PARENT COMPANY SALES'
1.
2.
3.
4.
5.
6.
7.
8.
Ashland
Monsanto
DENKA
Reichhold
Tenneco
U.S. Steel
Koppers
Amoco
Sales of MA
(1978 $ millions)
9.39
13.16
NAb
11.60
4.07
13.30
6.91
9.32
Parent company sales
(1978 $ millions)
5,426
5,019
NA
754
8,762
11,050
1,582
14,962
MA sales as percent
of total parent
company sales
0.17%
0.26%
33.0%c
1.5%
0.05%
0.12%
0.44%
0.06%
A 56-percent capacity figure was assumed in calculating MA sales for each
company.
NA = not available.
cBased on calculations by EPA.
5-40
-------
It can be assigned a high degree of probability. Tenneco's closure pos-
sibility, however, has not been verified by a company spokesperson and
therefore cannot be given as much confidence as the conclusions regarding
DENKA's possible closure.
Table 5-11 compares the cost of the least expensive control option to
total capital expenditures of the MA companies. Among the firms, Reichhold--
which had the highest capital costs for control when expressed as a percen-
tage of its capital budget-Ms expected to continue MA production. The
capital budget requirement may affect Reichhold to a greater degree than
the other companies (except for DENKA); nevertheless, Reichhold would
probably fund needed control equipment and attempt to pass its increased
costs to the consumer.
5.4.2.2 Intraindustry Competition. Intraindustry competition,
prompted by a 97- or 99-percent benzene emissions standard, would chiefly
involve the price differentials in MA merchant sales created by control
costs being passed through to the consumer. Other potentially contributing
factors are the interplay of transportation costs and the prospect of
import competition. This section summarizes possible effects of the above
factors on competition within the MA industry due to imposition of a ben-
zene control standard.
5.4.2.3 Effect of Cost Pass-Through On Market Competition Due to
Benzene Emissions Control. After funding initial capital to install con-
trols, companies may be able to raise the price of MA to recover expenses
incurred. The price increment each manufacturer would like to pass through
to sales depends on the total investment costs and the quantity of product
produced (i.e, the number of units through which price increments can be
passed).
The MA industry is expanding production from approximately 56 percent
toward the 100-percent production figure predicted for the post-1983 period.2
Price increments due to cost pass-through should thus be viewed according
to two production scenarios—56 percent and 100 percent capacity utilization.
Tables 5-12 and 5-13 depict possible price increments of MA based on 100
percent cost pass-through and production capacities of 56 percent and 100
percent, respectively. In each case, the least expensive control option is
employed, and costs are passed through according to both the production
5-41
-------
TABLE 5-11. COMPARISON OF CONTROL COSTS TO
TOTAL COMPANY CAPITAL EXPENDITURES
1978 Control costs as percent of total
capital company capital expenditures
Ashland
Monsanto
DENKA
Reichhold
Tenneco
U.S. Steef
Koppers
Amoco
cApeiiu i tui es
(106 dollars)
310
480
NA
24
1,008
668
144
1,744
97% control3 'b
0.21
0.18
NA
2.3
0.03
0.13
0
0
99% control a'b
0.23
0.20
NA
4.3
0.03
0.14
0
0
aAverage annual capital cost assuming 2-year installation time and financing.
Assumed the least expensive control between carbon adsorption and
incineration.
SOURCE: Annual Reports and Form 10-K filed with the Securities and Exchange
Commission Control Costs.
5-42
-------
TABLE 5-12.
01
4^
OJ
POSSIBLE COST PASS-THROUGH UNDER CASE
PRODUCTION CAPACITY3
ASSUMPTION OF 56 PERCENT
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
Company
Ashland
Monsanto
DENKA
Reichhold (N.J.)
Reichhold (111.)
Tenneco
U.S. Steel
Koppers (Pa. )
Koppers (111.)
Amoco
97% benzene
Increment
(
-------
TABLE
CJl
I
5-13. POSSIBLE COST PASS-THROUGH UNDER CASE
PRODUCTION CAPACITY3
ASSUMPTION OF 100 PERCENT
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
Company
Ashland
Monsanto
DENKA
Reichhold (N.J.)
Reichhold (111.)
Tenneco
U.S. Steel
Koppers (Pa. )
Koppers (111.)
Amoco
97% benzene
emissions control level
Increment Total cost
(
-------
figures of each company and annual!zed capital control costs extended over
a 10-year period. For the case of 56 percent capacity utilization, cost
increments are based on the estimated actual price of molten MA (73.4
-------
benzene control costs through fully can raise product prices by up to 4.5
percent in the 97-percent control case and by up to 4.8 percent in the
99-percent control case for an individual plant. In the high-demand,
full-capacity utilization market (Table 5-13), potential cost increments
would be lower, varying from 1.9 to 2.1 percent for the 97- and 99-percent
control cases, respectively. Under both control and capacity utilization
assumptions, Tenneco may be forced to raise prices the most, assuming it
chooses to pass costs through.
It must be emphasized that, when intraindustry competition is viewed
relative to possible MA price increases from cost pass-through, the major
arena of competition involves merchant, rather than captive, markets.
While costs may be passed through both markets, the greatest direct effect
is seen in the highly competitive merchant market. Conversely, price
increments in the captive MA market emerge only in the final product,
having been diluted by costs of other constituents also necessary to pro-
duce the final product. The effect of this "dilution" is substantial (see
Section 5.4.3). The following subsections, therefore, focus on the effects
of cost differentials on MA merchant sales in both a low- and high-demand
market.
5.4.2.3.1 Price differentials in a low-demand MA market. Raising
prices in a low-demand market is not an auspicious prospect. Price elas-
ticities governing the present MA market are simply not known. However,
the MA industry has indicated that it would be reluctant to raise prices
under low-demand market conditions; conversely, because most MA manufac-
turers are part of larger parent firms having other products, an alterna-
tive choice to total MA cost pass-through is to pass costs through par-
tially or totally to other products. Some members of the MA industry have
noted that they may employ this option, assuming the choice is made to
install benzene emissions control if a standard were promulgated. ' ' ' '
Tables 5-14 and 5-15 further examine the effect of cost pass-through
for two benzene control levels, based on the 56-percent production rate
that approximates present MA operating conditions. Numbers in the table
indicate the price differential (in cents per kilogram) between company and
competitor, assuming 100 percent of cost pass-through. Negative numbers
indicate a price advantage over a competitor, and the accompanying percent-
5-46
-------
en
TABLE 5-14. DEPICTION OF POSSIBLE COMPETITIVE ADVANTAGES DUE TO COST PASS-THROUGH
UNDER THE 56-PERCENT PRODUCTION CAPACITY ASSUMPTION
(97% Benzene Emissions Control Level)
NOTES: Top number in grid represents price differential between companies; negative sign indicates company has lower price than competitor.
Number in parentheses depicts potential advantage in terms of cost differential expressed in percent (current list price of
MA is basis: 73.4
-------
TABLE 5-15.
in
->
CD
DEPICTION OF POSSIBLE COMPETITIVE ADVANTAGES DUE TO COST PASS-THROUGH UNDER THE
56-PERCENT PRODUCTION CAPACITY ASSUMPTION
(99% Benzene Emissions Control Level)
Name-]
plate
city pet- ' ' ! Reichhold
(Gg) itor Ashland Monsanto \ DENKA (N.J.)
Company
1 Reichhold
j (111.)
Tenneco U.S.
97 Achla^rf ^^^^^H -n i n i n i n i n c
Koppers
Steel ' (Pa.)
Koppers
(111.)
i
| Amoco
n i -i ? -3.2 -3.2
(-4.4)
(-4.4)1
NOTES: Top number in grid represents price differential between companies; negative sign indicates company has lower price than competitor.
Number in parentheses depicts potential advantage in terms of cost differential expressed in percent (current list price of MA is
basis: 73.«Ag [3.3t/kg]).
-------
age expresses that advantage in terms of a price differential divided by
the assumed base price of MA prior to cost pass-through. Positive table
numbers indicate that a company is marketing its MA at a higher price than
its competitor. For example, Table 5-14 shows .that Ashland1s product costs
0.1$ more per kilogram (0.05
-------
and Amoco--need not incur any new expense to meet a possible benzene emis-
sion standard. Because of these circumstances, MA companies financially
affected by the standard may choose partial or total pass-through of ben-
zene control costs to other company products as a means of offsetting the
costs associated with required controls, while not placing themselves at a
competitive disadvantage. Nevertheless, companies may not choose this
option for a number of reasons.
Even in a depressed MA market situation, companies may choose instead
to raise MA prices. Seven out of eight MA facilities will be required to
install hardware under the 99-percent regulatory option (five for the
97-percent option), and each of these will want to pass control costs
through to the consumer. If all or several companies with large capacities
independently decide to pass costs through to MA, the market system may be
able to sustain the higher MA prices. In this case, Amoco and Koppers
would impose a limited threat of competition because neither company could
produce MA in quantities sufficient to satisfy the entire domestic MA
demand. Furthermore, Amoco, a large MA producer, has invested large sums
of money into the research and development of its n-butane process. Accord-
ing to an Amoco spokesman, the company would readily raise prices to offset
high R&D costs if a favorable market existed. Thus, in the absence of
import competition, higher MA prices may be a possible option for MA manu-
facturers even in a low-demand market.
5.4.2.3.2 Price differentials and the future MA market (at 100 per-
cent capacity utilization). For at least the next 5 years, demand for the
MA market is expected to continue to increase, corresponding to a 100-
2
percent production rate of present MA capacity. As this high-demand
market is approached, increasing the price of MA will become more auspi-
cious, and cost pass-through, thus, more possible. Furthermore, as the
base price of MA increases to meet market demand (regardless of a benzene
standard), price increments due to control cost pass-through become less
significant. Table 5-13 shows that the highest MA price increments under
both the 97- and 99-percent .control level are 1.9 and 2.1 percent, respec-
tively.
MA price differentials created from passing benzene control costs
through to the consumer may not occur in an industry running at 100 percent
5-50
-------
production capacity. The largest producers are likely to be the price
leaders as demand increases; thus, Monsanto and U.S. Steel may set prices
in the high-demand market of 1983. Monsanto and U.S Steel, the largest MA
producers, are both affected by the two possible benzene control level
requirements. However, it is likely that as market demand increases to
total MA production capacity, prices will rise to accommodate passing
benzene control costs through with no concomitant loss in sales. (This
assumes, of course, no import competition; see following sections.)
5.4.2.3.3 Effect of transportation on intraindustry impacts. Trans-
portation charges can affect the delivered cost of MA. In 1978, costs for
transporting MA by train averaged 0.005$/kg/mi (0.002
-------
TABLE 5-16. EFFECT OF TRANSPORTATION UNDER A 97-PERCENT CONTROL STANDARD
(Industry at 56% Production Capacity Utilization;
Companies Possibly Entering Competitor's Market)
tn
en
INJ
Name-
plate!
; capa- Corn-
Company
NOTES: NIC = no transportation credit.
A/B = numbers in table; where A is the number of miles designated, companies may be able to penetrate their competitor's market.
These figures are based on price differentials between plants operating at 56 percent capacity with the 97-percent carbon absorption
control.
A O.OK/kg/mi transportation cost was assumed in the calculations; B is the number of miles between designated companies.
-------
tn
in
CO
TABLE 5-17. EFFECT OF TRANSPORTATION UNDER A 99-PERCENT CONTROL STANDARD
(Industry at 56% Production Capacity Utilization;
Companies Possibly Entering Competitor's Market)
NOTES: NIC = no transportation credit.
A/B = numbers in table; where A is the number of miles designated, companies may be able to penetrate their competitor's market.
These figures are based on price differentials between plants operating at 56 percent capacity with the 99-percent thermal incin-
eration control. A O.OW/kg/mi transportation cost was assumed in the calculations; B is the number of miles between designated
companies.
-------
Consequently, if between 600 and 800 miles are subtracted from the distance
shown between two companies, it is evident that many companies already
compete in common market regions. Another unknown factor within each
plant's radius is distribution of its major market.
Tables 5-16 and 5-17 were calculated using the differential between
companies at 56 percent capacity, assuming both a 97-percent and a 99-
percent control level for benzene emissions, respectively. The transporta-
tion effects at 100 percent capacity utilization were not included because,
at that utilization rate, companies would be selling to a high-demand
market in which lack of buyers should not be an issue remedied by transport
to other markets. At the 56-percent production capacity market, however,
lack of buyers could be an issue.
Table 5-16 shows that, under a 97-percent benzene control level, the
companies having the greatest potential for regional market penetration
(due to MA price differentials) are DENKA, Reichhold (N.J.), U.S. Steel.
Koppers, and Amoco. Of these, Amoco presents the greatest potential
problem because of its large, residual production capacity, which can be
used to capture part of a competitor's markets.
Table 5-17 shows that, under a 99-percent benzene emissions control,
financially affected companies face potential regional market penetration
by a greater number of competitors. In this situation, however, Amoco
again presents the greatest threat because of its potentially large trans-
port credit and its large residual production capacity.
Although both tables show that the competitive position of MA compan-
ies due to price differentials can be enhanced by transportation, this
situation is unprecedented. Transportation costs currently limit the
regional market of a producer to a radius of about 483 to 644 km (300 to
400 mi). Under a benzene emissions standard, assuming industries decide
to pass control costs through, transportation may be used to sell MA beyond
the traditional 483- to 644-km (300 to 400 mi) market radius and still
meet competitive prices in certain cases, as illustrated in the tables.
Whether companies will pass costs through and employ this method to achieve
market penetration of a competitor cannot be determined. If MA producers
elect to absorb pollution control cost and retain current prices, transpor-
tation costs could then serve to restrict market radii to below 483 or 644 km
5-54
-------
(300 or 400 mi). In this case, regional market penetration of a com-
petitor would not be profitable and would have low probability of occurring.
5.4.2.4 Effect of Imports. Imports have a small effect on the present
MA market. In 1977, they accounted for only 2.3 percent of the United
States MA demand.12 The effect imports have on the future MA market is
uncertain, however.
Maleic anhydride is produced in two forms: molten and briquette.
Molten MA is the preferred form for large consumers, accounting for 90
percent of the domestic MA output. The remaining 10 percent of MA is
converted into briquettes, the form of all imported MA. Because bri-
quettes can be converted to the molten form only at a cost and because
briquettes have a limited shelf life, imported MA competes primarily with
domestic MA briquettes and not with the "molten market."
Imports could pose problems for United States MA briquette producers,
although small quantities of imported MA have not interfered with domestic
production. At times, imports have sold for 9
-------
5.4.2.5 Summary of Impact on Manufacturers. Tables 5-18 and 5-19
summarize the major impacts on the manufacturer arising from capital budget
requirements and intraindustry competition under two benzene control levels--
97 and 99 percent, respectively. Impacts of capital budget requirements
indicate that Tenneco may be and DENKA will be most severely affected.
Impacts of intraindustry competition indicate possible competition arising
from benzene control cost pass-through in the future, high-demand market.
Overall, the prospect of offsetting add-on benzene control costs and augment-
ing profits is enhanced in the future, high-demand market of 1983.
5.4.3 Effect on Product Prices
5.4.3.1 Cost Pass-Through to the Final Consumer. MA is purchased by
manufacturers for use as an intermediate in the production of various end
products. Should MA producers raise their product prices to compensate for
pollution abatement costs, purchasers of MA will feel the greatest economic
impact of these increments. It is therefore likely that MA users, like MA
producers, will choose to raise prices of their products to offset higher
production costs. Price increments of these products should not be as high
as potential MA price increments because MA accounts for only a fraction of
the wholesale price of each product. Thus, the impact of increased MA
prices should be diluted in the final product.
MA is used as an intermediate in the production of several products,
mainly polyester resins, agricultural chemicals, and fumaric acid. Tables
5-20, 5-21, and 5-22 show how the increased MA prices could affect the
prices of these products. Product price increments depend on the percent-
age of MA in the product and the percentage of its wholesale price attri-
buted to MA. These factors contribute to the varying increments associated
with each product.
The predominant end use of MA is the production of unsaturated poly-
ester resins (58 percent of MA demand) that go into reinforced plastic
applications such as marine craft, building panels, automobiles, tanks, and
pipes. Currently, these resins sell for 55$ to $3.30/kg (25$ to $1.50/lb).
Should MA prices increase, polyester resins could rise from 0.2
-------
TABLE 5-18. SUMMARY OF IMPACT OF 97 PERCENT BENZENE CONTROL LEVEL ON MA COMPANIES
Capital budget requirements
Intraindustry impact
At 56% production capacity (present) At 100% production capacity (1983)
Ashland
Monsanto
OENKA
in
on
Reichhold, N.J.
Reichhold, 111.
Tenneco
U.S. Steel
Koppers, Pa.
Koppers, 111.
Amoco
Could fund costs and continue
production
Could fund costs and continue
production
Would not need to install a
total control system
May not be able to meet capital
budget requirements from internal
resources
May choose to discontinue produc-
tion
Would not need to install a total
control system
Could fund costs and continue
production
Could fund costs and continue
production
Could fund costs but may
discontinue MA production
Could fund costs and continue
production
Would not need to install a
total control system
Could fund costs and continue
production
Would not need to install controls
Would not need to install controls
Would have to pass costs through MA
or other product line
Could face competition with Monsanto,
U.S. Steel, Koppers, and Amoco
Would have to pass costs through MA
or other product line
Could face competition from U.S.
Steel, Koppers, and Amoco
Would have to pass costs through MA
or other product line
Should have no new regional competitors
Would have to pass costs through MA
or other product line
Could face competition from Koppers
and Amoco
Would have to pass costs through MA
or other product line
Could face competition from Ashland,
U.S. Steel, Reichhold, Koppers,
Monsanto, and Amoco
Would have highest production costs
due to benzene control investment
Could face competition from Reich-
hold, Koppers, U.S. Steel, and Ashland
Would have to pass costs through MA
or other product line
Could face competition from Koppers
(111.) and Amoco
Would have to pass costs through MA
or other product line
Could face competition from Ashland
or U.S. Steel
Would not have to pass costs or
raise prices
Would not have to pass costs or
raise prices
Could pass costs through
without sales loss
Could pass costs through
without sales loss
Could pass costs through
without sales loss
Could pass costs through
without sales loss
Could pass costs through
without sales loss
Could pass costs through
without sales loss
Could pass costs through
without sales loss
Possible price leader
Could pass costs through
without sales loss
Same as in present market
Would likely raise prices
in response to demand
Same as in present market
Would likely raise prices
in response to demand
-------
TABLE 5-19. SUMMARY OF IMPACT OF 99 PERCENT BENZENE CONTROL LEVEL ON MA COMPANIES
Capital budget requirements
Intraindustry impact
At 56% production capacity (present) At 100% production capacity (1983)
01
i
en
00
Ashland
Monsanto
OENKA
Reichhold, N.J.
Reichhold, 111.
Tenneco
U.S. Steel
Koppers, Pa.
Koppers, 111.
Amoco
Could fund costs and continue
production
Could fund costs and continue
production
May not be able to meet capital
budget requirements from internal
resources
May choose to discontinue produc-
tion
Could fund costs and continued
production
Could fund costs and continue
production
Could fund costs but may choose
to discontinue production
Could fund costs and continue
production
Would not need to install a
total control system
Could fund costs and continue
production
Would not need to install controls
Would not need to install controls
Would have to pass costs through MA
or other product line
Could face competition from Koppers
(111.) and Amoco
Would have to pass costs through MA
or other product line
Could face competition from Koppers
(111.) and Amoco
Would have to pass costs through MA
or other product line
Should have no new regional competitors
Would have to pass costs through MA
or other product line
Could face competition from Koppers
(111.) and Amoco
Would have to pass costs through MA
or other product line
Could face competition from Koppers,
(111.) and Amoco
Would have highest production costs
due to benzene control investment
Could face competition from U.S.
Steel, Koppers (111.), and Amoco
Would have to pass costs through MA
or other product line
Could face competition from Koppers
(111.) and Amoco
Would have to pass costs through MA
or other product line
Would not have to pass costs or
raise prices
Would not have to pass costs or
raise prices
Could pass costs through
without sales loss
Could pass costs through
without sales loss
Possible price leader
Could pass costs through
without sales loss
Could pass costs through
without sales loss
Could pass costs through
without sales loss
Could pass costs through
without sales loss
Could pass costs through
without sales loss
Possible price leader
Could pass costs through
without sales loss
Same as in present market
Would likely raise prices
in response to demand
Same as in present market
Would likely raise prices
in response to demand
-------
in
01
TABLE 5-20. PRICE INCREASES OF POLYESTER RESINS DUE TO INCREASED MA PRICES3
X control
of benzene
emissions
97
99
X production
capacity of
MA plant
56
100
56
100
Present whole-
sale price of
polyester resins
in t/kg («/1b)
55-330 (25-150)
55-330 (25-150)
X of wholesale
price of polyester
resins attributed
to MA
15
15
Price increase
of polyester
resins
in
-------
TABLE 5-21. PRICE INCREASES OF FUMARIC ACID DUE TO INCREASED MA PRICES'
en
i
or>
o
% control % production
of benzene capacity of
emissions MA plant
97 56
100
99 56
100
aPrice increments of fumaric
Present whole-
sale price of
fumaric acid
in t/kg (t/lb)
94 (43)
94 (43)
acid were calculated
% of wholesale
price of fumaric
acid attributed
to MA
73
73
from the worst price increases
Price increase
of fumaric
acid •
in «/kg (
-------
TABLE 5-22. PRICE INCREASES OF MALATHION DUE TO INCREASED MA PRICESa
X control
of benzene
emissions
97
99
X production
capacity of
MA plant
56
100
56
100
Present whole-
sale price of
malathion
in t/kg (f/lb)
242 (110)
242 (110)
X of wholesale
price of malathion
attributed
to MA
11
11
Price increase
of malathion
in t/kg (t/lb)
1.3 (0.6)
0.6 (0.3)
1.6 (0.7)
0.8 (0.4)
X increase
of malathion
prices
0.5
0.2
0.7
0.3
New wholesale
price of
malathion
in
-------
These increments are relatively small since MA accounts for only 15 percent
of the total sales price of these resins; they represent only a 0.3- to
0.9-percent increase over the wholesale price of resins.
Agricultural chemicals, such as malathion, are the second largest
users of MA (10 percent of MA demand). The current wholesale price of
malathion is $2.42/kg ($1.10/lb),* and MA accounts for 11 percent of this
price. Should MA prices go up, the price of malathion could rise from 0.6$
to 1.6$/kg (0.3$ to 0.7$/lb). Such an increment is relatively insignificant
compared to the total cost of the product, representing an increase of only
0.2 to 0.7 percent over the present price of malathion.
Another market for MA is fumaric acid (5 percent of MA demand), which
is primarily used as a food acidulant. MA accounts for approximately 73
percent of the total price of fumaric production. This large percentage is
reflected in the maximum potential price increase of fumaric acid (1.6$ to
4.1$/kg [0.7$ to 1.9$/lb]). This increase represents a 1.70- to 4.40-percent
increase over the current sale price of 94$/kg (43$/lb).
5.4.4 Employment and Balance of Trade
Imposing a benzene control standard could affect employment in the MA
industry and could alter the present balance of trade. The MA industry
presently employs approximately 330 individuals; under a 99-percent control
standard, one manufacturer--DENKA--has indicated that capital expenditures
for controls would prompt closure. This company represents 10 percent of
the total domestic capacity. If it closes, approximately 33 workers could
lose their jobs. Tenneco is also considered a closure candidate, although
with less certainty than DENKA. If Tenneco did close, roughly 12 employees
could lose their jobs.
Regarding balance of trade, foreign competition exacerbated by benzene
control costs for U.S. producers could result in United States producers
losing up to 12 percent of their market to imports. This number is based
on the present 1977 import penetration of 2.3 percent in the total domestic
market and the potential for United States briquette sales, the only form
that is imported—10 percept of the total domestic market--to lose to
^Increments reflect the maximum range of increments that could occur using
the worst price increases associated with the least expensive add-on
control options at 97 and 99 percent benzene control.
5-62
-------
foreign competition. The assumption of a 12-percent market loss to imports
is highly conservative; more likely, imports will penetrate little or no
more of the domestic market over the next 5 years (see Section 5.4.2.4).
5.4.5 Fifth-Year Impacts
This section summarizes the following aggregate economic impacts
occurring 5 years after the standard is proposed:
Total annualized costs,
Net increase in national energy consumption, and
Inflationary impact on the cost of MA. It is assumed that the
plants will be running at full capacity at that time.
For the 97-percent regulatory option, total annualized costs would be
about $2.1 million, including annualized capital costs and operation and
maintenance costs of control and monitoring equipment. The additional con-
sumption of energy with 50 percent heat recovery credit would be approxi-
mately 180 TJ/yr (29,000 bbl/yr). Furthermore, the price of MA would rise
1.2 percent over the current list price.* This is equivalent to about
1.0
-------
5.5.2 Impact on Licensors
There is only one domestic licensor of maleic anhydride technology,
licensing both the benzene and n-butane processes. Abroad, there are
five licensing companies. Of these five, only one licenses both the
benzene- and n-butane-based technology. The other four license only the
benzene-based process.
If n-butane were used as an alternate feedstock for new plants, it
would be difficult to predict the effect on domestic producers. On one
hand, it is likely the domestic licensor will maintain its foreign business
because an EPA requirement would not affect usage of benzene-based technol-
ogy abroad. However, the company's domestic business will depend on the
competitive status of its n-butane and butene catalysts if benzene replace-
on
ment were mandated. Which companies are developing these catalysts—and
the success of their experimental work—is presently unclear. Catalyst
technology is usually a closely held company secret.
5.5.3 Impact on the Price and Availability of Feedstocks
At present, it is more economically favorable to use n-butane than
benzene as a feedstock in a new facility. Use of the n-butane feedstock
has been estimated to result in as much as a 7.3
-------
economics tied closely to gasoline and other fuels. One of the main sources
of benzene production is from aromatics produced as a byproduct of the
petroleum refining process. At present, little new refining capacity is
being added in the industry because of an anticipated no-growth situation
in domestic gasoline production. In addition, U.S. dependence on foreign
oils has affected benzene production. Foreign oil is not as napthenic-rich
as domestic crude and, therefore, has a lower aromatic concentration. As
the aromatics concentration from which benzene is derived decreases, the
economics of extraction become less attractive. Either less benzene is
produced or higher benzene prices result.
At present, benzene consumed in U.S. MA production comprises 3.7 per-
21
cent of the total benzene consumption. This percentage is expected to
21
increase to 4.4 percent by the year 1990. However, this study does not
account for EPA's possible regulation of new maleic anhydride plants.
n-Butane is primarily derived from natural gas as a coproduct of
liquefied petroleum gas. Most n-butane produced in the United States is
consumed directly at the refineries that produce it. The exact amount of
refinery-consumed n-butane is difficult to calculate because refineries do
not keep accurate accounts of the quantity of that n-butane that goes into
22
the production of other refinery products.
In spite of the difficulty associated with accurately calculating
n-butane supply and demand at the refineries, figures on the n-butane
unused by these refineries are available. Government statistics show at
22
present an oversupply of n-butane on the market. Because of this situa-
tion, the U.S. Department of Commerce recently considered the prospect of
22
exporting domestic n-butane.
The future availability of n-butane is linked directly to the future
availability of natural gas and liquefied petroleum gas. At present,
neither natural gas nor LPG production is expected to increase significant-
ly over the next decade. However, even if increases are minimal, supplies
of n-butane should adequately meet the domestic n-butane demand at new MA
plants because refineries are expected to remain the prime users of n-butane
??
and minimal increases are anticipated in their consumption pattern.
5-65
-------
5.5.4 Impact on the Economic Life of Existing Plants
A requirement of no detectable benzene emissions at new MA plants
would have no impact on the economic life of existing plants. As mentioned
in Section 5.2.3.2, use of n-butane feedstock has led to as much as a
7.3
-------
12. Letter from Epstein, E. A., Energy and Environmental Analysis, Inc.,
to Magnusson, Fred, U.S. Department of Commerce. June 29, 1979.
13. Kinkley, M. T., and R. B. Neverill. Capital and Operating Costs of
Selected Air Pollution Control System. Card, Inc. Niles, 111. (Pre-
pared for Office of Air Quality Planning and Standards, U.S. Environ-
mental Protection Agency.) EPA Contract Number 68-02-2577. May 1977.
14. Letter from Vatavuk, William M., Office of Air Quality Planning and
Standards, U.S. Environmental Protection Agency, to Lorber, Joe,
Hewlett-Packard. March 22, 1978.
15. Letter from Hinkson, B., DENKA, to Tippitt, William, U.S. Environ-
mental Protection Agency. June 12, 1979.
16. Letter from Hinkson, B., DENKA, to Warren, John, Research Triangle
Inst. May 22, 1979.
17. Telecon. E. A. Epstein, Energy and Environmental Analysis, Inc., with
Dr. Mackay, U.S. Steel. April 5, 1978.
18. Letter from Epstein, E. A., and Thomasian, J., Energy and Environ-
mental Analysis, Inc., to Farley, Charles, Monsanto Industrial
Chemicals Co. April 26, 1978.
19. Letter from Farley, Charles P., Monsanto Chemical Intermediates Co.,
to Epstein, E. A., Energy and Environmental Analysis, Inc. May 10,
1978.
20. Letter from Warren, J., Research Triangle Inst., to Gans, M., Scienti-
fic Design Co., Inc. February 2, 1979.
21. Pace Co. MACRO/MICRO Economic Analysis of United States Petrochemical
Demand to 1990, Volume II. April 1, 1976.
22. Letter from Epstein, E. A., Energy and Environmental Analysis, Inc.,
to Higgins, Terry, Economic Regulatory Administration, U.S. Department
of Energy. April 4, 1978.
23. Letter from Epstein, E. A., Energy and Environmental Analysis, Inc.,
to Wyatt, S., Tippitt, W., and Basala, A., U.S. Environmental Protec-
tion Agency. May 10, 1979.
5-67
-------
5-68
-------
APPENDIX A
EVOLUTION OF THE BID
A-l
-------
TABLE A-l. EVOLUTION OF THE BID
Date
Company, consultant,
or agency
Nature of action
Location
s*
i
PO
June 8, 1977
July 28-29, 1977
Sept. 28-29, 1977
October 20, 1977
October 25/1977
November 17, 1977
December 21, 1977
January 24, 1978
February 2, 1978
March 13-17, 1978
March 20-24, 1978
May 24, 1978
June 2, 1978
June 28, 1978
August 15, 1978
August 23-24, 1978
October 31, 1978
April 10, 1979
April 25-26, 1979
May 3, 1979
August 2, 1979
September 1979
October 1979
EPA
Reichhold Chemicals, Inc.
Manufacturing Chemists
Association Air Quality
Committee
Monsanto Chemical
Intermediates Co.
Koppers Co., Inc.
DENKA Chemical Corp.
Monsanto Chemical
Intermediates Co.
Amoco Chemicals Corp.
Monsanto Chemical
Intermediates Co.
Reichhold Chemicals, Inc.
DENKA Chemical Corp.
Reichhold Chemicals, Inc.
EPA
EPA
Koppers Co., Inc.
EPA
Koppers Co., Inc.
EPA
DENKA Chemical Corp.
Ashland Chemical Corp.
Tenneco
EPA
EPA
Publication in the Federal Register
of decision to list benzene as a
hazardous air pollutant under Section
112 of the Clean Air Act.
Plant visit.
Meeting with engineers from EPA and
Hydroscience, Inc., to review progress
on the study of the maleic anhydride
industry and control technology.
Plant visit.
Section 114 request for information.
Plant visit.
Section 114 request for information.
Plant visit.
Test data supplied to EPA.
Carbon adsorption test by EPA.
Incinerator test by EPA.
Plant visit and pretest survey.
Draft SSEIS mailed to working group.
Meeting with working group to review
comments on draft SSEIS.
Section 114 request for information.
NAPCTAC meeting.
Test data supplied to EPA.
Steering committee meeting.
Plant visit.
Plant visit.
Section 114 request for information.
Steering committee meeting.
Delivery of AA Concurrence Package.
Morris, 111.
Washington, D.C.
St. Louis, Mo.
Pittsburgh, Pa.
Houston, Tex.
St. Louis, Mo.
Chicago, 111.
St. Louis, Mo.
Morris, 111.
Houston, Tex.
Elizabeth, N.J.
Durham, N.C.
Durham, N.C.
Pittsburgh, Pa.
Alexandria, Va.
Pittsburgh, Pa.
Washington, D.C.
Houston, Tex.
Neal, W. Va.
Washington, D.C.
Washington, D.C.
-------
APPENDIX B
INDEX TO ENVIRONMENTAL IMPACT CONSIDERATIONS
B-l
-------
APPENDIX B
INDEX TO ENVIRONMENTAL IMPACT CONSIDERATIONS
Agency guidelines
Background and description
a. Process affected
b. Industry affected
c. Availability of control
technology
Alternatives considered
a. No action or postponing action
• Environmental impacts
b. 97 percent control impact
• Air pollution
• Water pollution
• Solid waste disposal
• Energy impact
• Economic impact
Location within the BID
The process to be affected is
described in Section 1.2.1.
A description of the industry
to be affected is given in
Sections 1.1 and 5.2.
Information on the availability
of control technology is given
in Chapter 2.
The environmental impacts of
not implementing any standard
are discussed in Section 4.1.
The air pollution impacts of a
97-percent standard are dis-
cussed in Sections 4.1.2 and
4.1.4.
The water pollution impact of
a 97-percent standard is
discussed in Section 4.2.
The impact of a 97-percent con-
trol standard on solid waste
disposal is discussed in Section
4.3.
The energy impacts of a 97-
percent control standard are
discussed in Sections 4.4.2 and
4.4.4.
The economic impacts of a 97-
percent control standard are
discussed in Sections 5.3 and
5.4.
B-2
-------
Agency guidelines
Location within the BID
c. 99 percent control impact
• Air pollution
• Water pollution
• Solid waste disposal
• Energy
• Economic
d. Conversion to n-butane for
existing plants
- Feasibility
• Environmental impact
e. n-Butane use for new plants
• Air pollution impact
• Water pollution impact
• Solid waste disposal
The air pollution impacts of a
99-percent control standard
are discussed in Sections 4.1.1
and 4.1.3.
The water pollution impacts of
a 99-percent control standard
are discussed in Section 4.2.
The solid waste disposal impacts
of a 99-percent control standard
are discussed in Section 4.3.
The energy impacts of a 99-
percent control standard are
discussed in Sections 4.4.1
and 4.4.3.
The economic impacts of a 99-
percent control standard are
discussed in Sections 5.3 and
5.4.
The feasibility of converting
existing plants to the n-butane
process is discussed in Sections
1.2.2 and 2.1.4.
The environmental impacts of the
n-butane process are discussed
in Section 4.1.5.
The air pollution impacts of
requiring new plants to use
n-butane are discussed in Sec-
tion 4.1.5.
The water pollution impacts of
requiring new plants to use
n-butane are discussed in Sec-
tion 4.2.
The solid waste disposal impacts
of requiring new plants to use
n-butane are discussed in Sec-
tion 4.3.
B-3
-------
Agency guidelines Location within the BID
• Economic impact The economic impacts of requir-
ing new plants to use n-butane
are discussed in Section 5.5.
• Energy impact The energy impacts of requiring
new plants to use n-butane are
discussed in Section 4.4.5.
B-4
-------
TABLE B-l. MATRIX OF ENVIRONMENTAL IMPACTS OF ALTERNATIVE CONTROL SYSTEMS
Alternative
A.
B.
97%
99%
control
control
Impact on Impact Other air Water Solid waste Energy Space Noise Economic
benzene on HC impacts impact impact impact impact impact impact
+3 +3 -1 -1 -1 -1 -1 -1 -2'
+4 +4 -1 -1 -1 -3 -1 -1 -3
A. Substitution
of feedstock,
such as n-butane
(new plants only)
-2
-1
-1
CD
1
cn
Delayed standard
No standard
-2
-3
KEY + Beneficial impact
Adverse impact
0
1
2
3
4
5
-2 0
-3 0
No impact
Negligible impact
Small impact
Moderate impact
Large impact
Larger impact
-------
-------
APPENDIX C
EMISSION SOURCE TEST DATA
C-l
-------
APPENDIX C
EMISSION SOURCE TEST DATA
C.I INTRODUCTION
The purpose of this appendix is to present and summarize data gathered
during the development of a standard for benzene emissions from the produc-
tion of maleic anhydride. The facilities tested are described and the
source testing methods are identified. Any reference to commercial products
and processes by name does not constitute endorsement by the U.S. Environ-
mental Protection Agency.
C.2 SUMMARY
Two separate yet similar facilities producing maleic anhydride were
tested to determine percent reduction of emissions. Each plant was tested
by a different contractor. Pollutants analyzed and procedures used were:
Benzene Draft EPA Method
Total hydrocarbons (THC) Draft EPA Method
*Carbon dioxide EPA Method 10/Orsat
*0xygen EPA Method 10/Orsat
*Carbon monoxide EPA Method 10/Orsat
*Methane Draft EPA Method
*Ethane Draft EPA Method
Total organic acids (TOA) tLAAPCD Method
Total aldehydes LAAPCD Method
Formaldehyde LAAPCD Method
Temperature Thermocouple
Duct pressure EPA Method 2
Volumetric flow rate EPA Method 2
NO EPA Method 7
A
Although production facilities are similar in the two plants, the
means of emissions control are different. Plant A controls emissions by use
*Data not presented in text.
tLAAPCD--Los Angeles Air Pollution Control District.
C-2
-------
of a carbon adsorption system, while Plant B uses an incinerator. The
available data indicate that incineration provides slightly higher removal
efficiencies for benzene and total hydrocarbons, while carbon adsorption is
acceptable and is more consistent in removing emissions.
C.3 DESCRIPTION OF FACILITIES
C.3.1 Plant A
This plant uses a single-train, multiple reactor process with a capacity
of 20,000 Mg/yr of maleic anhydride. At the time samples for this study
were taken, the plant was operating at about 40 percent of its design
capacity.
Maleic anhydride is produced at Plant A (Table C-l) by the vapor phase
oxidation of benzene in a tubular reactor. The resulting reactor exhaust
passes through a series of switch condensers that are first cooled to
freeze the maleic anhydride from solution and later heated to melt the
maleic anhydride for pumping to storage. After the freezing process, the
remaining exhaust gas enters the product recovery absorber, which scrubs
the exhaust with water or aqueous maleic acid. The liquid effluent from
the absorber is about 40 percent, by weight, maleic acid. Vented emissions
from the absorber are directed to the carbon adsorption system. Essen-
tially all process emissions exit through the product recovery absorber.
The single inlet sampling port is located in the 1.07-m I.D. duct
leading into the three carbon adsorption units. The outlet sampling port
is located in the 1.07-m I.D. stainless steel exhaust stack, 11.6 m from
the bottom of its 29.3-m height.
«\ 3
Integrated gas samples were obtained in Tedlar bags (0.113 m3 [4 ft ]),
heated to a temperature of 49° to 74° C (120° to 165° F). Sampling commenced
at the onset of a 1-hour desorption cycle and lasted for three full cycles
(3 hours).
Benzene and total hydrocarbon concentrations were determined in the
field by gas chromotography with flame ionization detection. Methane and
ethane concentrations were not determined because of inadequate resolution
by retention time. Total hydrocarbons were determined as propane.
Total aldehydes were determined using the LAAPCD Method, involving
sample reaction with sodium bisulfate, pH adjustment freeing bisulfate ions
in an amount equivalent to the aldehydes present, and titration with a
C-3
-------
TABLE C-l: PLANT A--EMISSIONS SUMMARY
Sample
designation
A
B
c
Benzene:
Concentration
Emission rate
Concentration
Emission rate
Concentration
Emission rate
(ppm)
(kg/hr)
(ppm)
(kg/hr)
(ppm)
(kg/hr)
Inlet
861
83
977
92
915
86
Total hydrocarbon concentration
Concentration
Emission rate
Concentration
Emission rate
Concentration
Emission rate
(ppm)
(kg/hr)
(ppm)
(kg/hr)
(ppm)
(kg/hr)
Total organic acids (as
Concentration
Emission rate
Concentration
Emission rate
Concentration
Emission rate
Total aldehydes
Concentration
Emission rate
Concentration
Emission rate
Concentration
Emission rate
Formaldehyde:
Concentration
Emission rate
Concentration
Emission rate
Concentration
Emission rate
(ppm)
(kg/hr)
(ppm)
(kg/hr)
(ppm)
(kg/hr)
1,610
87
1,600
85
1,360
73
maleic
17
2
<4
<0
6
0
Outlet %
.2
.3
.9
81
7
63
6
42
4
.4
.87
.9
.04
.2
.01
reduction
90.
QC
5
A
•f
(as propane):
.9
.3
.0
acid):
.6
.53
.00
.56
.58
.93
147
8
110
5
67
.02
.87
.3
90.
93.
95.
9
1
1
3.61
0
0
<0
<0
1
0
.39
.06
.24
.03
.16
.16
97.
94.
82.
8
0
4
(as formaldehyde):
(ppm)
(kg/hr)
(ppm)
(kg/hr)
(ppm)
(kg/hr)
(ppm)
(kg/hr)
(ppm)
(kg/hr)
(ppm)
(kg/hr)
88
3
112
4
74
2
70
2
85
3
54
2
.0
.3
.1
.8
.7
.9
.6
.5
.1
.2
.0
<9
<0
13
0
109
4
1
0
1
0
1
0
.21a, 3.6b
.34a, 0.13b
.Oa, 2.4b
.47a, 0.09b
a, 2.5b N.
.03a, 0.09b
.8a, 1.8b
.07a, 0.07b
.8a, 1.4b
.06a, 0.05b
.la, 1.3b
.04a, 0.05b
N.A.
96.
88.
97.
A.a'd,
97.
97.
97.
98.
97.
97.
a.c
Ob
4a
9b
96. 7b
5a
5b
9a
4b
9a
6b
aFlask analysis of bag sample.
bSample from continuous isokinetic impinger sampling train.
cNot available due to "less than" values of outlet.
dNot available due to outlet concentration being greater than inlet concentra-
tion.
C-4
-------
standard iodine solution. Formaldehyde is determined by sample reaction
with sodium bisulfate, reaction with chromotropic acid, and colorimetric
determination of a uniquely colored compound formed in the second reaction.
Samples for formaldehyde and total aldehydes were taken from the bag samples
and from a continuous isokinetic impinger sampling train. Both results are
presented.
Analysis for To.tal Organic Acids (TOA) was conducted in accordance
with the LAAPCD Air Pollution Source Testing Manual. November 1963. This
method involves "field fixing" samples with sodium hydroxide and transfer
to the lab for ether extraction of organic acids and titration with a
standard base.
The EPA Method 3 (Orsat) was used for determining C02> 02, and CO.
For Plant A, in addition to analysis of the gas stream at the inlet
and outlet of the emission control system, water samples from the drain to
the product recovery absorber were analyzed for benzene, TOA, formaldehyde,
and total aldehydes.
C.3.2 Plant B
This plant has a maximum production capacity of 23,000 Mg/yr of maleic
anhydride. At the time samples for this study were taken, the plant was
operating at about 70 percent of its design capacity. Plant personnel did
not think the lower production rate would seriously affect the validity of
the results.
Maleic anhydride is produced at Plant B (Table C-2) by the vapor phase
oxidation of benzene in a tubular reactor. The resulting reactor exhaust
gas passes through a partial condenser that separates out a portion of the
crude maleic anhydride. The remaining exhaust gas enters the product
recovery absorber, which scrubs the exhaust with water or aqueous maleic
acid and produces an aqueous solution containing about 40 percent, by
weight, maleic acid. The exhaust gas from the absorber'is directed to an
incinerator. Essentially all process emissions exit through the product
recovery absorber.
A single inlet sampling port is located in the 0.91-m ID inlet duct to
the incinerator. An outlet sampling port is located at each of eight
7.5-cm ID incinerator outlet ducts. The outlet ports were located down-
stream of any physical disturbances.
C-5
-------
TABLE C-2. PLANT B—EMISSIONS SUMMARY
Sample
designation
A
r
lr
Inlet
Benzene:
Concentration (ppm) 780
Emission' rate (kg/hr) 154
Concentration (ppm) 820
Emission rate (kg/hr) 161
Concentration (ppm) 940
Emission rate (kg/hr) 185
Outlet
11.1
2.2
11.8
2.4
14.4
2.9
% reduction
QQ C
JO. V
QQ C
7O . w
QQ C
7O * 3
Total hydrocarbon concentration (as propane):
p
i*
p
Lr
Concentration (ppm) 1,520
Emission rate (kg/hr) 169
Concentration (ppm) 1,880
Emission rate (kg/hr) 209
Concentration (ppm) 2,090
Emission rate (kg/hr) 232
Total organic acids (as maleic acid):
Concentration (mg/m3) 153
Emission rate (kg/hr) 9.3
Concentration (mg/m ) 208
Emission rate (kg/hr) 12.6
Concentration (mg/m ) 281.1
Emission rate (kg/hr) 17.0
Total aldehydes (as formaldehyde):
Concentration (mg/m ) 33.9
Emission rate (kg/hr) 2.0
Concentration (mg/m ) 49.7
Emission rate (kg/hr) 3.0
Concentration (mg/m ) 90.1
Emission rate (kg/hr) 5.4
Formaldehyde:
Concentration (mg/m ) 11.0
Emission rate (kg/hr) 0.67
Concentration (mg/m ) 21.8
Emission rate (kg/hr) 1.32
Concentration (mg/m ) 47.8
Emission rate (kg/hr) 2.90
Oxides of nitrogen:
Concentration (mg/m ) N.O.
Emission rate (kg/hr) N.D.
Concentration (mg/m ) N.D.
Emission rate (kg/hr) N.D.
Concentration (mg/m ) N.D.
Emission rate (kg/hr) N.O.
24.3
2.7
23.6
2.8
26.5
3.0
362
21.8
89
5.8
57
3.6
4.7
0.3
6.2
0.4
4.7
0.3
0.60
0.04
0.00
0.00
0.60
0.04
15.7
1.0
17.4
1.1
13.6
0.9
QQ £
3O. *f
op 7
70. /
4ft 7
JO. /
N.A.a
C.A n
i^T. W
78 8
/ O • O
85 0
Ow • W
86 7
OW * f
94 4
Jr • "T
94 0
J~ • W
100
xw
98.6
N.D.
N.D.
N.D.
aNot available due to outlet concentration being higher than inlet
concentration.
NOTE: N.D. = not determined.
C-6
-------
Gas samples were obtained according to the EPA draft method for ben-
zene, in 70-L (2.5 cf) Mylar® bags, heated to 66° C (151° F). Total hydro-
carbons, benzene, methane, and ethane were determined by gas chromatography,
with flame ionization detection. Total hydrocarbons were determined as
propane. Carbon dioxide, oxygen, and carbon monoxide were analyzed by
nondispersive infrared spectroscopy, using Orsat analysis to determine
percentages.
Total organic acids in the samples were analyzed according to the
LAAPCD Method, which involves bubbling the sampled gas through a dilute
sodium hydroxide solution, acidification, ether extraction of the organic
acids, and titration with a standard base. The LAAPCD Method employed for
total aldehydes entails sample reaction with sodium bisulfate, pH adjust-
ment freeing bisulfate ions, and titration with a standard iodine solution.
Formaldehyde is determined by sample reaction with sodium bisulfate, reac-
tion with chromotropic acid, and colorimetric determination of a uniquely
colored compound formed in the second reaction.
Oxides of nitrogen (NO ) were determined according to EPA Reference
/\
Method 7, employing colorimetric analysis of a complex ion formed by reac-
tion of organic acids with the NO . Duct temperature was obtained by
^\
thermocouple, and duct pressure by water manometer, according to EPA Refer-
ence Method 2.
C-7
-------
APPENDIX D
EMISSION MEASUREMENT AND
CONTINUOUS MONITORING
D-l
-------
APPENDIX D
EMISSION MEASUREMENT AND CONTINUOUS MONITORING
D.I EMISSION MEASUREMENT METHODS
D.I.I General Background
For stack sampling purposes, benzene will, except in the case of sys-
tems handling pure benzene, exist in the presence of other organics.
Accordingly, methods for benzene analysis consist first of separating the
benzene from other organics, followed by measuring the quantity of benzene
with a flame ionization detector. However, among various stack-testing
groups concerned with measuring benzene, there is no uniformity of proce-
dures. Problems of uniformity could exist in the following areas:
Sample collection,
Sample introduction to gas chromatograph,
Chromatographic column and associated operating parameters, and
Chromatograph calibration.
Two possible approaches for benzene sample collection are grab samples
and integrated samples. Because benzene emission concentrations may vary
considerably during a short period, the integrated sample approach is more
advantageous since emission fluctuations due to process variations are
automatically averaged. In addition, the integrated approach reduces the
number of samples that need to be analyzed. For integrated samples, both
®
tubes containing activated charcoal and Tedlar bags have been used. How-
ever, charcoal sampling tubes were basically designed for sampling ambient
concentration levels of organics. Because source effluent concentrations
are expected to be higher (particularly since organics other than benzene
could be present), uncertainty would be involved in predicting sample
breakthrough or sample termination. Bag samples would also offer the best
D-2
-------
potential for precision because no intermediate sample recovery step would
be involved.
Based on the above considerations, collecting an integrated sample in
Tedlar® bags appears to be the best alternative, a conclusion shared by an
EPA-funded study whose purpose was to propose a general measurement tech-
nique for gaseous organic emissions. Another study of benzene stability
® 2
in Tedlar bags was undertaken to confirm the soundness of this approach.
Because this study showed no significant deterioration of benzene over a
period of 4 days, the integrated bag technique was deemed suitable. However,
anyone preferring to use activated charcoal tubes has this option, provided
that efficiency equal to or better than the bag technique can be demonstrated
and procedures to protect the integrity of the sampling technique are
followed.
A collected gas sample can be introduced to a gas chromatograph either
by using a gas-tight syringe or an automated sample loop. The latter
approach was selected for the reference method because it has less potential
for leakage and provides a more reproducible sample volume.
Several columns are mentioned in the literature that may be suitable
3 4
for separating benzene from other gases. ' Most notable are 1, 2, 3-tris
(2-cyanoethoxy) propane for separating aromatics from aliphatics and Ben-
tone 34 for separating aromatics. A program was undertaken to establish
whether various organics associated with benzene in stack emissions inter-
fered with the benzene peaks from the two columns. ' The study revealed
the former column to be suitable for analyzing benzene in gasoline vapors
and the latter column to be suitable for analyzing benzene emissions from
maleic anhydride plants. It should be noted that selection of these two
columns for inclusion in Method 110 does not mean that some other column(s)
may not work equally well. In fact, the method has a provision for using
other columns.
Calibration has been accomplished by two techniques, the most common
being the use of cylinder standards. The second technique involves inject-
(6)
ing known quantities of 99 mole percent pure benzene into Tedlar bags as
they are being filled with known volumes of nitrogen. The second technique
has been found to produce equally acceptable results; both are included in
Method 110.
D-3
-------
D.I.2 Field Testing Experience
(R)
Based on the study of benzene stability in Tedlar bags, of possible
interferences by various process-associated gases, and of calibration
methods, and as a result of a field study and tests conducted at sources of
benzene emissions, a new draft of Method 110 for determining compliance
with benzene standards or NESHAPS was prepared. This method is the same as
the originally investigated method, except that the audit procedure has
been refined and an appendix has been added to help verify benzene peak
resolution.
Two maleic anhydride manufacturing plants were tested during the
development test program. One of these plants employed a carbon adsorption
system, and the other employed a thermal oxidizer (incinerator) to control
organic emissions. Method 110 was used to collect and analyze for benzene
emissions. The SP 1200/Bentone 34 gas chromatographic column described in
th^ method was used to resolve the benzene. No major deviations from
Method 110 were required. At the plant employing the carbon adsorption
system, a liquid dropout was required to prevent intermittent entrained
liquid from being introduced into the integrated bag sample. This liquid
entrainment, caused when an undried steam-desorbed carbon bed was reintro-
duced into the control system, occurred 5 percent of the sampling time. A
rotameter was also installed at the inlet to the integrated bag sample at
this plant because the rigid container housing the bag sample could not be
adequately sealed. Neither of these two deviations are considered to have
affected data validity.
The sampling lines and bags were maintained at or slightly above the
source temperature during collection and analysis to prevent condensation
of organics that would normally be a vapor at the source temperature.
Organic acid and aldehyde emission data were also collected during the
test program. Data were collected to adequately determine emissions in
terms of concentration, mass rate, and control system mass removal effi-
ciency.
D.2 CONTINUOUS MONITORING
No emission-monitoring instrumentation, data acquisition, or data
processing equipment have been identified for measuring benzene from maleic
anhydride plant stack gases. However, the U.S. Environmental Protection
D-4
-------
Agency (EPA) has only recently begun to explore the development of specifi-
cations for a system for benzene monitoring; such a system, which would
employ a package of individual, commercially available items, is considered
feasible.
For a chromatographic system that measures benzene concentration, the
installed cost of the chromatograph and its auxiliaries, which include gas
chromatograph with dual-flame detector, automatic gas sampling valve, air
sampler, post-run calculator, and gas regulators, is $35,000. This figure
would increase by approximately $10,000 for the additional hardware neces-
sary to report a benzene mass emissions rate in terms of benzene feedstock.
Depending on the operating factor, the direct operating cost varies from
about $1,200 to $1,400 per year.
D.3 EMISSION TEST METHODS
The recommended emission test method for determining benzene emission
concentrations at maleic anhydride plants is Method 110. The method uses
the Method 106 train for sampling and a gas chromatograph/flame ionization
detector equipped with a column selected for separation of benzene from the
other organics present for analysis.
Subpart A of 40 CFR 61 requires that facilities subject to National
Emission Standards for Hazardous Air Pollutants be constructed to provide
sampling ports adequate for the applicable test methods, and platforms,
access, and utilities necessary to perform testing at those ports.
Assuming that the test location is near the analytical laboratory and
that sample collection and analytical equipment are available, the cost of
field collection, laboratory analysis, and reporting of benzene emissions
in triplicate from a single stack is estimated to be $2,500 to $3,500 for
an emission test effort. This figure assumes a cost of $25 per person-hour.
This amount would be reduced by approximately 50 percent per stack if
several stacks were tested.
If the plant established in-house sampling capabilities and conducted
its own tests and/or analyses, the cost per person-hour could be less.
D-5
-------
D.4 REFERENCES
1.
2.
Feairheller, W. R., A. M. Kemmer, B. J. Warner, and D. Q. Douglas.
Measurement of Gaseous Organic Compound Emissions by Gas Chromatography.
EPA Contract Number 68-02-1404. January 1978.
Knoll, Joseph E., Wade H. Penny, and M. Rodney Midgett. The Use of
Tedlar Bags to Contain Gaseous Benzene Samples at Source-Level Concen-
trations. U.S. Environmental Protection Agency. Research Triangle
Park, N.C. EPA Publication Number 600-78-057. September 1978.
Separation of Hydrocarbons. Bulletins 743A, 740C, and D. Supelco,
Inc. Bellefonte, Pa. 1974.
4.
Current Peaks. 10:1.
1977.
Carle Instruments, Inc. Fullerton, Calif.
Letter from Knoll, Joseph E., Environmental Monitoring Systems Labora-
tory, U.S. Environmental Protection Agency, to Grimley, William,
Emission Standards and Engineering Division, U.S. Environmental Protec-
tion Agency. October 18, 1977.
Letter from Knoll, Joseph E., Environmental Monitoring Systems Labora-
tory, U.S. Environmental Protection Agency, to Grimley, William,
Emission Standards and Engineering Division, U.S. Environmental Protec-
tion Agency. November 10, 1977.
D-6
-------
APPENDIX E
METHODOLOGY FOR ESTIMATING MORTALITY AND LIFETIME
RISK FROM EXPOSURE TO BENZENE EMISSIONS FROM
MALEIC ANHYDRIDE PLANTS
E-l
-------
APPENDIX E
METHODOLOGY FOR ESTIMATING MORTALITY AND LIFETIME
RISK FROM EXPOSURE TO BENZENE EMISSIONS FROM
MALEIC ANHYDRIDE PLANTS
E.I INTRODUCTION
The purpose of this appendix is to describe the methodology used in
estimating leukemia mortality and lifetime risk attributable to population
exposure to benzene emissions from maleic anhydride manufacturing plants.
The appendix is presented in three parts:
Part E.2, Summary and Overview of Health Effects, summarizes and
references reported health effects from benzene exposure. The
major reported health effect is leukemia. Mortalities cited in
the BID include only the estimated leukemia cases attributable to
exposure to benzene emissions from existing maleic anhydride
plants although other, sometimes fatal, effects are known to
result from benzene exposure.
Part E.3, Population Density Around Maleic Anhydride Plants,
describes the method used to estimate the population at risk;
i.e., persons residing within 20 km of existing maleic anhydride
plants.
Part E.4, Population Exposures, Risks, and Mortalities, describes
the methodology for estimating benzene emissions from a model
plant, calculating expected population exposures, and estimating
leukemia deaths attributable to benzene emissions from the eight
existing U.S. plants.
E.2 SUMMARY AND OVERVIEW OF HEALTH EFFECTS
E.2.1 Health Effects Associated with Benzene Exposure
A large number of occupational studies over the past 50 years have
provided evidence of severe health effects in humans from prolonged inhala-
tion exposure to benzene. Some 300 studies of the health effects of
benzene have recently been reviewed and analyzed in terms of application to
E-2
-------
low-level ambient benzene exposures that might occur in a population resid-
ing near a source of benzene emissions.
The reviewers concluded that benzene exposure by inhalation is strongly
implicated in three pathological conditions that may be of public health
concern at environmental exposure levels:
Leukemia (a cancer of the blood-forming system),
Cytopenia (decreased levels of one or more of the formed elements
in the circulating blood), and
Chromosomal aberrations.
Leukemia is a neoplastic proliferation and accumulation of white blood
cells in blood and bone marrow. The four main types are acute and chronic
myelogenous leukemia and acute and chronic lymphocytic leukemia. The
causal relationship between benzene exposure and acute myelogenous leukemia
and its variants in humans appears established beyond reasonable doubt.
The term "pancytopenia" refers to diminution of all formed elements of
the blood and includes the individual cytopenias: anemia, leukopenia,
thrombocytopenia, and aplastic anemia. In mild cases, symptoms of pancyto-
penia are such nonspecific complaints as lassitude, dizziness, malaise, and
shortness of breath. In severe cases, hemorrhage may be observed, and
death may occasionally occur because of hemorrhage or massive infection.
Patients with pancytopenia may subsequently develop fatal, acute leukemia.
Chromosomal aberrations include chromosome breakage and rearrangement
and the presence of abnormal cells. These aberrations may continue for
long periods in hematopoietic and lymphoid cells. The health significance
of these aberrations is not fully understood. However, aberrant cells have
been observed in individuals exposed to benzene who have later developed
leukemia. Some types of chromosomal aberrations may be inheritable.
In one study too recent to include in the review previously cited,
workers exposed to 2.1 ppm benzene for 4 years showed a statistically
significant increase in chromosomal aberrations (as high as tenfold) over
2
those in unexposed controls.
The review concluded that man may be the only species yet observed to
be susceptible to benzene-induced leukemia. Evidence for production of
leukemia in animals by benzene injection was considered nonconclusive.
E-3
-------
Moreover, benzene exposure by oral dosing, skin painting, or inhalation has
not been shown to produce leukemia or any other type of neoplastic diseases
in test animals, although other effects, including pancytopenia, have been
widely observed.
E.2.2 Benzene Exposure Limits
It should be noted that where the health effects described above have
been associated with benzene exposure, the exposure has been at occupa-
tional levels. That is, the benzene exposure levels associated with the
effects have been high (10 ppm up to hundreds of parts per million of
benzene, except in a few cases of exposures to 2 to 3 ppm benzene) or they
have been unknown.
Benzene exposure was first associated with health effects in occupa-
tional settings, so initial attempts to limit benzene exposures were aimed
at occupational exposures. With recognition of the toxic effects of ben-
zene and its greatly expanded use after 1920, several occupational exposure
limits were established in the United States. These limits, originally in
the range of 75 to 100 ppm, were successively lowered as more information
on benzene toxicity became known.
For example, the American Conference of Governmental Industrial Hygien-
ists (ACGIH) recommended a benzene threshold limit value of 100 ppm in
1946, 50 ppm in 1947, 35 ppm in 1948, 25 ppm in 1949, and 10 ppm in 1977.3'4
The National Institute for Occupational Safety and Health recommended an
c
exposure limit of 10 ppm in 1974 and revised it downward to 1 ppm in 1976.
The current Occupational Safety and Health Administration permissible
exposure limit is 10 ppm (a lower limit of 1 ppm is currently in litigation*).
Occupational exposure limits were initially established to protect
workers from adverse changes in the blood and blood-forming tissues. The
most recently recommended or pending limits of 1 ppm and 10 ppm are based
on the conclusion that benzene is leukemogenic in man (NIOSH and OSHA ) or
4
a suspected carcinogen in man (ACGIH ).
*A benzene standard with a limit of 1 ppm was proposed by OSHA May 27,
1977, (42 FR 27452) and promulgated February 10, 1978 (43 FR 5918). This
standard was struck down October 5, 1978, by the U.S. Fifth Circuit Court
of Appeals. The U.S. Department of Labor appealed the decision, and the
Supreme Court agreed to hear arguments on the case during its fall 1979
term.
E-4
-------
E.2.3 Health Effects at Environmental Exposure Levels
Little information is available on health effects of nonoccupational
exposures of the general populace to benzene. Virtually all of the studies
1 2
cited ' were on the working population (mostly males) exposed to higher
than ambient benzene levels on a work cycle. Applying these studies to
chronic (24 hours per day) low-level exposure to the general population
(including infants, the ill, and the elderly) requires extrapolation.
The recent analysis of benzene health effects concluded that the
evidence of increased risk of leukemia in humans on exposure to benzene for
various time periods and concentrations was overwhelming but that data were
not adequate for deriving a dose-response curve.
However, EPA's Carcinogen Assessment Group (CAG), acknowledging the
absence of a clear dose-response relationship, has estimated the risk of
o
leukemia in the general population from low-level benzene exposure. Data
from three epidemiological studies of leukemia in workers (mostly adult
white males) were used to estimate the risk of developing leukemia. A
no-threshold linear model was used to extrapolate this estimated risk to
the low levels (below 5 to 10 ppb) to which some populations may be exposed.
The annual risk factor derived for benzene-induced leukemia was 0.34
case per 10 ppb person-years. For example, if 3 million persons are
chronically exposed to 1 ppb benzene, the model predicts there will be 1.02
cases of leukemia (3 x 0.34) per year in that population. Use of a "linear"
model means that the model would predict the same number of leukemia deaths
among 3 million people exposed to 1 ppb benzene as among 1 million people
exposed to 3 ppb.
The risk factor (0.34 case per 10 ppb person-years) was used in
estimating the number of leukemia deaths as attributable to benzene emis-
sions from maleic anhydride plants. Other effects of benzene exposure
(including fatalities from causes other than leukemia) were not included in
the estimated number of deaths. The risk factor equated one leukemia case
to one death (that is, each case was presumed fatal).
Several sources of uncertainty are present in the risk factor. First,
the retrospective occupational exposure estimates may be inaccurate. CAG
calculated the 95-percent confidence intervals for this risk factor to be
0.17 to 0.66 if exposure estimates in the three studies extrapolated are
E-5
-------
precisely correct, and 0.13 to 0.90 if exposure estimates are off by a
factor of 2. Second, the composition of the exposed populations around
maleic anhydride plants may vary from that of the populations used as a
basis for the CAG estimate. Third, the true dose-response relationship for
benzene exposure may not be a linear no-threshold relationship at the low
concentrations to which the general population may be exposed. Fourth, the
risk factor includes only leukemia deaths and not other health risks. No
quantitative estimate of the error in the risk factor due to the latter two
uncertainties has been attempted.
E.3 POPULATION DENSITY AROUND MALEIC ANHYDRIDE PLANTS
The population groups "at risk" to benzene exposure (those residing
Q
around maleic anhydride plants) were determined from the 1970 Bureau of
the Census Master Enumeration District List (MED List) for the area sur-
rounding each plant site.
Each plant site was located by latitude and longitude on a grid system
having grids of 10 sq km. The population was determined from the MED List
for each grid block within 20 km of each plant site. There were thus
approximately 125 population grids for each plant site.
Circles of radii 0.1, 0.3, 0.5, 0.7, 1.0, 1.5, 2.0, 10, and 20 km were
overlaid on each grid, and the population within each annular (ring or
doughnut-shaped) area was determined. Where grid blocks overlapped two
annular areas, the population was assumed to be uniformly distributed and
was assigned proportionately to each area. The population in each annular
area was considered to be the population exposed to the estimated benzene
concentration at the midpoint of that area. The estimated total popula-
tions exposed as a function of distance from the plant site are reported in
Table E-l.
The method used contains potential sources of error. First, the
assumption of uniform population distribution within grid blocks and annular
areas may not be correct. For urban areas the assumption is probably
valid, but it may introduce some error for rural areas (which is the case
with one plant) 10 to 20 km from the site. Another source of error is the
use of 1970 population data. However, these are the latest available in
the form required. The contractor deriving the population figures made no
numerical estimates of probable errors.
E-6
-------
TABLE E-l. TOTAL POPULATION EXPOSED AS A FUNCTION OF DISTANCE
FROM PLANT SITE
Population
Plant
Monsanto
Reichhold (111.)
U.S. Steel
Reichhold (N.J.)
Tenneco
DENKA
Ashland
Koppers
Total population
Average popula-
tion density
(persons/sq km)
0.1-
0.3
1,300
0
300
0
200
500
0
0
2,300
1,150
0.3-
0.5
2,300
0
500
0
400
1,000
0
150
4,350
1,090
0.5-
0.7
3,500
100
700
0
600
1,500
0
200
6,600
1,100
at various distances
0.7-
1.0
7,400
300
1,500
200
1,300
3,100
0
500
14,300
1,120
1.0-
1.5
18,200
700
3,900
800
3,500
7,600
0
1,400
36,100
1,150
from plant (km)
1.5-
2.0
22,400
600
5,000
3,800
7,400
10,800
0
3,700
53,700
1,220
2.0-
10.0
747,000
11,700
294,000
651,000
310,000
481,000
19,800
297,000
2,811,500
1,170
10.0-
20.0
784,000
17,600
873,000
2,358,000
1,001,000
739,000
140,000
810,000
6,722,600
890
Total
population
1,586,100
31,000
1,178,900
3,013,800
1,324,400
1,244,500
159,800
1,112,950
9,651,450
960
SOURCE: Letter Report from Energy and Environmental Analysis, Inc., to U.S. Environmental Protection Agency.
-------
E.4 POPULATION EXPOSURES, RISKS, AND MORTALITIES
E.4.1 Summary of Methodology
The methodology for estimating leukemia deaths due to exposure to
benzene emissions from maleic anhydride plants follows.
»
A typical (or "model") maleic anhydride plant was developed, based on
a nominal maleic anhydride production capacity of 22,700 Mg/yr (metric tons
per year). Its benzene emissions were estimated to be 190 kg/hr from the
product recovery absorber, based on a 94.5-percent benzene conversion rate
in the reactor and no benzene emission controls. An additional 2.6 kg/hr
benzene was estimated to be emitted from storage, handling, and fugitive
sources.10'11
The Industrial Source Complex Dispersion (ISCD) model, urban mode 2,
was used to estimate mean annual benzene concentrations out to 20 km from
12
the model plant. Pittsburgh meteorological data were used in the disper-
sion model.
Dispersed benzene concentrations from the model plant were corrected
to reflect actual plant capacity and degree of benzene emission controls
9 12
currently exercised in each plant. '
The population (1970) around each actual plant location was correlated
with corrected benzene concentrations to yield benzene dose in 10 ppb
9
person-years per year. The methods for determining populations are described
in Part E.3 of this appendix.
o
From health effects data, the EPA Carcinogen Assessment Group derived
a leukemia risk estimate of 0.34 death per 10 ppb person-years from exposure
o
to benzene. The methodology for estimating the leukemia risk factor is
described in Part E.2.3 of this appendix.
The leukemia deaths per year attributable to exposure to benzene from
maleic anhydride plants were estimated by multiplying 3.4 x 10 ppb person-
years exposure times the exposure in "ppb person-years per year."
Leukemia deaths were estimated for benzene exposures from existing
maleic anhydride plants, assuming no control of fugitive, storage, or
handling emissions, but four different degrees of benzene emission control
on the product recovery absorbers. These four conditions, termed "control
alternatives," are:
E-8
-------
A. No regulation by any standard; that is, the current conversion
rate (94.5 percent) and the current state of benzene emission
control on the product recovery absorber in each plant (varies
from no control to 90, 97, or 99 percent control);
B. 97 percent reduction in benzene emissions;
C. 99 percent reduction in benzene emissions by thermal incinera-
tion; and
D. 99 percent reduction in benzene emissions by carbon adsorption.
For Alternatives B, C, and D, a benzene conversion rate of 90 percent
was assumed, resulting in an estimated uncontrolled benzene emission rate
of 345.4 kg/hr from the product recovery absorber of the model plant.
For all control alternatives, two control system failure scenarios
were assumed. First, it was assumed that benzene control systems on the
product recovery absorbers would fail for 3 hours 15 times per year and
would be limited to emissions of 250 kg of benzene per plant (on the basis
of the model plant capacity) during each failure. The uncontrolled benzene
emission rate from the model plant product recovery absorber is 190 kg/hr,
so the 250-kg emission 15 times per year would be equivalent to 20 hr/yr of
uncontrolled emissions. The second failure condition assumed that benzene
control systems would fail for 48 hr/yr, during which product recovery
absorber emissions would be 190 kg/hr (based on the model plant) for the
full 48 hours.
It was assumed further that the atmospheric dispersion of these unan-
ticipated emissions would follow the same pattern as during normal (con-
trolled) operations and that emissions from fugitive and storage sources
would be unaffected by the failures.
E.4.2 Estimates of Leukemia Deaths
The general equation for estimating the number of leukemia deaths
attributable to benzene emissions from a particular plant (e.g., plant X)
under either normal or uncontrolled operations is:
10-20 K
Dx = I (R)(Pi/10b)(B.)(C/22.7)(f), (1)
i = 0.1-0.3
E-9
-------
in which
DX= estimated number of leukemia deaths per year from benzene
emissions from the plant (e.g., plant X).
R = the risk factor (0.34 death per 10 ppb person-years).
P.j= population at risk, in area (i) around plant X (Table E-l).
B.j= mean annual average benzene concentration (ppb) in area (i)
around plant X for the control alternative selected (A,B,C, or D)
or for the uncontrolled condition (Table E-3).12
C = capacity of plant X (in thousands of metric tons of maleic anhy-
dride produced per year; the capacity of the model plant is
22,700 metric tons/yr).
f = fraction of plant X output produced from benzene (e.g., if 20 per-
cent of plant capacity is from n-butane feedstock, f = 0.80).
i = the particular area in which P. and B. occur (i progresses from
the area 0.1 to 0.3 km from the plant to the area 10 to 20 km
from the plant).
I = summation of deaths from all areas (i).
Values for P. are given in Table E-l. Values for R, C, f, current
control levels, and control levels under failure conditions are given in
Table E-2. Values of B. for all control levels of interest, as determined
from Table 32 of Reference 12, are given in Table E-3.
Leukemia deaths per year were estimated for each plant for both con-
trolled and uncontrolled conditions, using Equation 1. The total estimated
number of leukemia deaths per year attributable to benzene emissions from
all plants was determined for each condition by the equation:
Total estimated number of leukemia deaths/yr (D.)
= D1 + D2 + ... + D8 . (2)
The total numbers of estimated leukemia deaths attributable to benzene
emissions from maleic anhydride plants are given in Table E-4 on a plant-by-
pi ant basis, in deaths per year, for three control alternatives (current,
97 percent, and 99 percent), the uncontrolled condition (0 percent), and
the shutdown condition (storage, handling, and fugitive emissions only).
For purposes of Table E-4, Alternatives C and D are indistinguishable and
have been combined. The alternatives defined in Table E-4 thus differ
slightly from the nomenclature previously used (A,B,C, and D).
E-10
-------
TABLE E-2.
FACTORS IN ESTIMATING LEUKEMIA DEATHS FROM BENZENE
EMISSIONS FROM MALEIC ANHYDRIDE PLANTS
Plant
Monsanto
Reichhold (111.)
U.S. Steel
Reichhold (N.J.)
Tenneco
DENKA
Ashland
Koppers
/ Fraction
Capacity, 103 from
metric tons/yr benzene
(C) (f)
47.6
20
38.5
13.6
11.8
22.7
27
15.4
0.8
1.0
1.0
1.0
1.0
1.0
-1.0
1.0
Benzene control
uct recovery
Current
level
0
90
90
97
0
97
0
99
level at prod-
absorber (%)
Under control
system failure
conditions
0
0
0
0
0
0
0
0
Risk factor
(R)
0.34
0.34
0.34
0.34
0.34
0.34
0.34
0.34
-------
TABLE E-3.
ESTIMATED MEAN ANNUAL AVERAGE BENZENE CONCENTRATIONS (B.)
AROUND MODEL MALEIC ANHYDRIDE PLANTS* ]
Column (1)
(2)
(3)
(4) (5) (6)
Benzene concentration (ppb)
(7)
(8)
(9)
Unregulated plant.(94.5% con-
version) with PRA emissions
reduced by
Regulated plant (90% conver-
sion) with PRA emissions
reduced by
Distance from
plant (km)
0.1-0.3
0.3-0.5
0.5-0.7 "
0.7-1.0
1.0-1.5
1.5-2.0
2.0-10.0
10.0-20.0
0%C
5.27
6.12
5.7
4.8
3.6
2.55
0.49
0.135
90%
2.3
1.37
0.96
0.7
0.5
0.34
0.057
0.0155
97%
2.0
0.97
0.6
0.38
0.24
0.155
0.025
0.0062
99%
1.95
0.87
0.49
0.29
0.17
0.105
0.0155
0.0038
97%
2.11
1.1
0.73
0.51
0.33
0.22
0.037
0.0097
99%
(incin-
eration)
1.99
0.87
0.53
0.33
0.20
0.13
0.023
0.0057
99%
(adsorp-
tion)
2.12
0.99
0.61
0.38
0.22
0.14
0.021
0.0054
ioo%d
1.9
0.82
0.43
0.245
0.13
0.076
0.0103
0.0025
Values in Columns 2, 3, 4, 5, and 9 were derived by adding storage, handling, and fugitive source benzene
concentrations from Table 32 of Reference 12 to 100 percent, 10 percent, 3 percent, 1 percent, or 0 percent
of the "Alternative A" product recovery absorber (PRA) benzene concentrations, respectively, for each distance.
Resulting concentrations were plotted vs. distance on log-log paper, and the concentration at the midpoint
of each segment was determined. Micrograms/m were converted to ppb by dividing by 3.2. Values in
Columns 6, 7, and 8 were derived by adding the storage, handling, and fugitive source benzene concentrations
to the PRA concentrations for the appropriate control alternative (B, C, or D) from Table 32, except that:
(1) for all three options, PRA concentrations were increased by (1 - 0.9)/(1 - 0.945), or 1.82X, to account
for higher emissions due to changing the benzene conversion rate from 94.5 percent to 90 percent; and
(2) for Alternatives C and D (Columns 7 and 8), PRA concentrations were further increased by (1 - 0.99)/
(1 - 0.995), or 2X, to account for higher emissions due to changing the required control system efficiency
from 99.5 percent to 99 percent. Total concentrations were plotted vs. distance, on log-log paper, and
the concentration at the midpoint of each segment was determined and recorded in Table E-3.
Product recovery absorber.
cCondition of no control or control system failure.
Fugitive, storage, and handling emissions only.
SOURCE: H. E. Cramer Co., Inc. Dispersion Model Analysis of the Air Quality Impact of Benzene Emissions
from a Maleic Anhydride Plant for Four Emission Control Options. EPA Contract Number 68-02-2507.
August 1978.
-------
TABLE E-4.
I
I—•
CO
ESTIMATED LEUKEMIA DEATHS DUE TO BENZENE EMISSIONS
FROM MALEIC ANHYDRIDE PLANTS
Column (1)
Plant
Monsanto
Reichhold (111.)
U.S. Steel
Reichhold (N.J.)
Tenneco
DENKA
Ashland
Koppers
Total deaths/yrc
(2)
0 percent
(control
system
failure
condition)
0.3827
0.00425
0.1756
0.1326
0.0586
0.1437
0.0116
0.0632
0.9723
(3) (4)
Estimated leukemia deaths/yr
benzene emissi
Current control levels
With no
% control
current f ai 1 ures
control or shutdown
0 0.3827
90 0.00054
90 0.0214
97 0.00647
0 0.0586
97 0.0082
0 0.0116
99 0.00200
0.492
(5)
(6)
(7)
with product recovery absorber
ons reduced by
97%
by standard
With no
control
failures
or shutdown
0.033
0.000358
0.0140 .
0.00647°
0.00458
0.0082°
0.000845
0.00200°
0.0695
99%
by standard
With no
control
failures
or shutdown
0.0212
0.000222
0.00869
0.00593
0.00283
0.00744
0.000507
0.00200°
0.0488
100%a
(plant
shutdown
condition)
0.0122
0.000125
0.00447
0.002657
0.00141
0.00400
0.000224
0.001356
0.0264
fugitive, storage, and handling emissions only.
b
Where current control levels result in fewer estimated deaths than with the control levels required by
the standard, the lesser figure is used.
cThese figures are for each condition if it were to occur 8,760 hr/yr. Figures in Table E-5 assume
8,000 hr/yr of plant operations and so are lower.
-------
The death estimates shown in Table E-4 are for each condition if it
were to occur continuously, and so were weighted in accordance with the
following operating assumptions. Product recovery absorbers (PRA) were
assumed to operate 8,000 hr/yr. Storage, handling, and fugitive emissions
were assumed to occur 8,760 hr/yr; that is, they continued even during
plant shutdown. During the first failure condition, with 15 control system
failures of 3 hours each, the period of normal control system operations
would be 7,955 hr/yr. During failure, emissions would be equivalent to
20 hr/yr of uncontrolled operations, based on the model plant and a 94.5-
percent benzene conversion rate. During the second failure condition, with
48 hours of control system failure, the period of normal control system
operations would be 7,952 hr/yr.
The risk factor (0.34/10 ppb person-years) is based on continuous
o
exposure (8,760 hr/yr). The estimated leukemia deaths per year were
therefore determined by weighting deaths from Table E-4 in accordance with
the duration of each condition. For the first failure condition, this
would be: controlled operations (7,955 hr/yr), uncontrolled operations
(equivalent of 20 hr/yr), and storage, handling, and fugitive emissions
only (an additional 785 hr/yr). Thus, the total estimated number of leu-
kemia deaths per year was determined as:
Dy - [(D. under no-control conditions) x (20/8,760)]
+ [(D. under current or regulated condition; ,~ ..
current level, 97, or 99%) x (7,955/8,760)]
+ [(D. under shutdown conditions) x (785/8,760)] .
The equivalent of 20 hours of uncontrolled emissions instead of the
45-hour actually assumed emission period is used to put the emission rate
into the terms used in the dispersion model report in order to simplify the
calculations. However, fugitive, storage, and handling emissions continue
for the remaining 25 hours of this period as well as the 760 hr/yr when the
plant is shut down, for a total of 785 hr/yr.
Similarly, for the second failure condition, leukemia deaths per year
were determined as:
E-14
-------
DT = [(D. under no-control conditions) x (48/8,760)]
+ [(D. under current or regulated condition;
current level, 97, or 99%) x (7,952/8,760)]
+ [(D. under shutdown conditions) x (760/8,760)] .
t
Deaths per year resulting from operations based on 8,000 hr/yr of
product recovery absorber operations with no control failures and an addi-
tional 760 hr/yr of only storage, handling, and fugitive emissions were
determined by:
D-r = C(Dt under current or regulated condition;
current level, 97, or 99%) x (8,000/8,760)] (3c)
+ [(D. under shutdown conditions) x (760/8,760)] .
The total estimated leukemia deaths per year for the control options
of interest are summarized in Table E-5. It is apparent that the brief
control system failure conditions have little effect on annual exposure and
deaths. Also, even with continuous 99 percent control of product recovery
absorber benzene emissions, there will still be an estimated 0.0488 leukemia
death per year. From Table E-4 (last two columns), it can be seen that
0.0264 (or 54 percent) of these results from fugitive, handling, and storage
emissions, and 0.0224 death (46 .percent) from residual product recovery
absorber emissions.
E.4.3 Example of Leukemia Death Calculation
Values in Table E-4 were determined in accordance with the following
example, using the U.S. Steel plant. From Table E-2, R = 0.34, C = 38.5,
and f = 1 for all distances, so Equation 1 may be simplified to:
,. 10-20
D = (R)(C/22.7)(f)(l/10b) I P.B. , (4)
x i = 0.1-0.3 n n
7 10-20
or D = 5.77 x 10~' I P.B. .
* i = 0.1-0.3 1 1
P. values for each area (i) around this plant are taken from Table E-l.
Table E-2 shows that this plant has a current control level of 90 percent,
so Bi values for 90 percent control in each area (i) are taken from Table E-3.
IP.B. is calculated as follows:
E-15
-------
TABLE E-5. ESTIMATED LEUKEMIA DEATHS FROM BENZENE
EMISSIONS FROM MALEIC ANHYDRIDE PLANTS3
Estimated leukemia deaths
Product recovery
absorber control
alternative
With no
control.
failures
With 15
control
failures/yr
With
48 hr/yr
of control
-r • -l U
failures
No regulation by
EPA (current level
of control ^-Alter-
native A
97% benzene
removal (by
incineration or
adsorption)--Alter-
native B
99% benzene
removal (by
incineration or
carbon adsorption)--
Alternatives C and D
0.45
0.45
0.45
0.066
0.068
0.071
0.047
0.049
0.052
It should be recognized that considerable uncertainty is associated with
the cases of leukemia that appear in the table. First, the cases were
calculated based on an extrapolation of leukemia risk associated with a
healthy white male cohort of workers to the general population, which
includes men, women, children, infants, the aged, nonwhites, and the
unhealthy. Second, there are potential errors in estimating the benzene
levels to which people in the vicinity of maleic anhydride plants are
exposed. Also, the number of cases includes consideration of only one
effect of benzene; i.e., leukemia. Benzene has also been indicated to
cause aplastic anemia, cytopenias, and the development of chromosomal
aberrations. In addition, the benefits to the general population of
controlling other hydrocarbon emissions from maleic anhydride manufacture
are not quantified.
For comparison, 665.95 total leukemia deaths per year would be expected in
the exposed population of 9.65 million residing within 20 km of the eight
existing maleic anhydride plants. This figure is based on the overall U.S.
leukemia death rate of 6.9 per population of 100,000 for 1972, 1973, and
1975.13 The rate ranged from 6.7 to 7.6 among the six States where maleic
anhydride plants are located.
CAG has estimated that there are 90 leukemia deaths per year nationwide
because of benzene exposure.9
Calculation described in Equation 3c.
Calculation described in Equation 3a.
Calculation described in Equation 3b.
E-16
-------
. Pi (from Bi (from Table E-3)
(i) Table E-l) for 90% control Pi X Bi
0.1 - 0.3 300 2.3 690
0.3 - 0.5 500 1.37 685
0.5 - 0.7 700 0.96 672
0.7 - 1.0 1,500 0.7 1,050
1.0 - 1.5 3,900 0.5 1,950
1.5 - 2.0 5,000 0.34 1,700
2.0 - 10.0 294,000 0.057 16,758
10.0 - 20.0 873,000 0.0155 13.532
2PiBi = 37,037
DY = 5.77 x l(f7 (37,037) = 0.0214.
^\
This value (0.0214) >s shown for U.S. Steel in Column 4 of Table E-4,
assuming no control system failures. The same procedure is used to deter-
mine deaths assuming 100 percent control failure, by using B^ values for
0 percent control from Table E-3. The resultant value (0.1756) is in
Column 2 of Table E-4. All values in Table E-4 were determined with Equa-
tion 1 in a like manner, using B. values for the appropriate control level
and R, C, f, and P. values for the specific plant. Total deaths were
obtained by summing the deaths attributable to each plant, using Equation 2.
Totals are shown in the lower section of Table E-4.
Deaths listed in Table E-5 were determined from Equations 3a, 3b, and
3c. Deaths at, say, 97 percent control, with 15 control failures per year
were determined from Equation 3a as:
DT = 0.0695 (7,955/8,760) + 0.9723 (20/8,760)
+ 0.0264 (785/8,760) = 0.068.
With no control failures, deaths per year were determined from Equation 3c
as:
DT = 0.0695 (8,000/8,760) + 0.0264 (760/8,760) = 0.066.
Other deaths listed in Table E-5 were determined in the same manner.
E.4.4 Estimate of Leukemia Risk
The estimated leukemia deaths shown in "tables E-4 and E-5 are based on
estimates of mean annual average benzene concentrations around maleic
anhydride plants. Because atmospheric dispersion patterns are not uniform,
some population groups will receive above-average benzene exposures and
will therefore incur a higher risk (or probability) of contracting leukemia.
E-17
-------
Maximum annual risk is the estimated probability to someone, who is
constantly exposed to the highest maximum annual average benzene concentration
in the ambient air around a particular source for a year, of contracting
leukemia because of exposure to benzene emissions from that source. Maximum
lifetime risk can be found by multiplying the maximum annual risk by 70
years.
Maximum annual risks of leukemia deaths in exposed population groups
were estimated from the maximum annual average benzene concentrations esti-
12
mated for various distances from the model plant, using the equation:
Maximum annual risk of leukemia death for an 1nd1vidual/yr
= (B., the max annual benzene concentration, ppb)
6
x (R, the risk factor of 0.34 death per 10 ppb person-years)
x (10 , which is one person as a fraction of 1 million), or
Maximum annual risk of leukemia death for an 1ndiv1dual/yr
= (3.4 x I0"7)(max BI in ppb) , (5a)
or
= (1.06 x I0"7)(max B. in ug/m3) . (5b)
For the uncontrolled model plant, the estimated maximum annual average
3
benzene concentration (max B.) is 35.7 ug/m at 0.3 km from the plant
(Table 14 of Reference 12). The estimated maximum annual risk or probability
of leukemia death for an individual is 3.8 x 10 (1.06 x 10 x 35.7) per
year.
For the same plant with 97 percent control of benzene emissions from
the product recovery absorber, the maximum estimated risk 1s 2.2 x 10 per
person per year, which occurs 0.1 km from the plant (Table 22 of Reference 12).
This calculation is shown in Figure E-l. Calculations for maximum risk are
summarized in Table E-6.
Maximum annual risks of leukemia from product recovery absorber emis-
sions only were calculated in a similar manner, using maximum annual average
concentrations from Table 9 of Reference 12. Calculations are shown 1n
Table E-7. The maximum risk associated with the uncontrolled model plant
is 3.4 x 10 per person per year. With 97 percent benzene control, the
E-18
-------
Max B; = 9.09 + 2.77 + 8.83 + 0.0317 x 1.82
I
l-»
(O
MaxB-,
due to
MaxB, Max
8= Max a,
due to product hi
due to due to' recovery absorbei (PRA) B
crease in
due to
*t4W44i£ handfing fugitive cnangng
sources
rate from
94.5% to
90%
x 2
Change in
control
system
efficiency
from
99.5% to
99%
v 7,955] ,
8,760 J
Pei cent of
total annual
hours
h 0.497
Max &, at
0.1 km due
to PRA during
control system
malfunction
x 2° 1
8.760 .
Pei cent of total
hours that
occur during
control system
malfunction
= 20.796
Maximum Risk = (1.06 x 10-7){20.796) = 2.2 x 10 6
Figure E-1. Example calculation of maximum annual risk for 97 percent control.
-------
ro
o
TABLE E-6. EFFECT OF MODIFYING EMISSION CONTROL PARAMETERS ON ESTIMATED MAXIMUM LEUKEMIA RISK
Factors
Column (1)
Benzene conversion rate (%)
Control method3
Control level (%)
15 failures/yr with excess
emissions
48 hr/yr of failure with excess
emissions
Emissions at
(2)
94.5%
either
97
NO
NO
(3)
94.5%
inc.
99.5
NO
NO
3
jxn'nt of maximum risk (in pg/m )
(4)
94.5%
ads.
99.5
NO
NO
(5)
90%
either
97
NO
NO
(6)
90%
inc.
99
NO
NO
(7)
90%
ads.
99
NO
NO
Benzene from storage, handling,
and fugitive sources
Benzene from product recovery
absorber control system:
Total benzene concentration
20.69 20.69
20.69
20.69
20.69 20.69
0.0317b 0.00405b 0.03955 0.0576 0.01473 0.1436
20.7217b 20.6941b 20.7295b 20.7476 20.7047 20.8336
Risk:c probability of leukemia
death per person per year
2.2x10
-6
2.2X10"6 2.2X10"6 2.2x10
"6 2.2x!0"6 2.2xlO~6
Inc. = incineration; ads. = adsorption; either = either method.
Benzene concentrations (ug/m3) taken directly from Reference 12.
cRisk = (B., ug/m3) x (0.34, the risk per 106 ppb person years/yr) x (10~6, which is one person as
a fractio?i of 106) x (1/3.2, to convert ug/m3 to parts per billion). Risk = 0.106 x 10~6 x B..
Decimal places are retained only to show effects of parameters changes. For comparison, the maxi-
mum risk associated with emissions from the unregulated model plant is 3.8 x 10~6 per person per
year.
-------
TABLE E-7. CALCULATIONS FOR MAXIMUM ANNUAL LEUKEMIA RISK
FROM MODEL PLANT PRODUCT RECOVERY ABSORBER EMISSIONS ONLY3
1. UNCONTROLLED (32.0)b x (1.06 x 10~7)a = 3.4 x lo"6
2. 97% CONTROL (e.g., incineration)
x (1.91)c x (1.06 x 10"7) = 1.8 x 10"7
3. 99% CONTROL (e.g., incineration)
x (0.466)C x (1.06 x 10"7) = 4.5 x 10"8
Risk = (B., ug/m3) x (0.34, the risk peg :
which is one person as §7fraction of 10 )
ppb). Risk = 1.06 x 10 x B.. Decimal i
106 ppb person-years/yr) x (10 ,
x (1/3.2, to convert ug/m3 to
ppb). Risk = 1.06 x 10 ' x B.. Decimal places are retained only to show
effects of parameter changes.
Table 9, Cramer Report (Reference 12), in ug/m3.
These numbers represent the respective ambient concentrations, given in
Table 9 of the Cramer report. The Cramer report ambient concentrations
are multiplied by 1.82 because of the increase in allowable emissions
under the standard (based on a 90-percent benzene conversion rate) over
emissions in the Cramer model based on 94.5 percent benzene conversion.
For 99 percent control, the revised concentrations are then multiplied by
2 because the Cramer report is based on 99.5 percent control, while the
control level considered is 99 percent.
E-21
-------
maximum risk is 1.8 x 10 per person per year. For 99 percent control,
~8
the maximum risk is 4.5 x 10 per person per year.
E.4.5 Validity of Estimates
Several potential sources of error exist in the estimated factors used
in Equation 1 (R, P., B., C, f). Possible errors in the risk factor (R)
are discussed in Part E.2.3, and sources of error in populations "at risk"
(P.) are discussed in Part E.3. The validity of the other factors is
discussed below. Readers should note that the number of significant figures
carried in the decimals in Tables E-3 through E-7 is not an indication of
their accuracy but rather a means of enabling readers to duplicate the
calculations and to visualize the magnitude of difference among the options
considered.
E.4.5.1 Plant capacity (C. f). Plant capacities (C) in thousands of
metric tons per year were estimated by projecting output to 1982 and assum-
ing operation at full capacity. The factor (f) represents the fraction of
total plant output produced using benzene as the feedstock. Estimates of
(f) are based on current plant operations.
Because plant capacities are often nominal values, the capacity and
production estimates may be inaccurate either on the high or low side,
affecting (C). Plants may start using n-butane as a feedstock to a greater
or lesser degree, affecting (f). The degree of error in the estimates of
(C) and (f) cannot be defined in numerical terms.
E.4.5.2 Benzene Concentrations (B.)
The estimated benzene concentrations are derived from several factors,
as follows:
Configuration of the model plant,
Emission rates from the model plant, and
Dispersion patterns of the emissions.
Numerical error limits could not be calculated for these factors, but
their qualitative effects on the estimated number of leukemia deaths are
discussed below.
The configuration of the model plant assumes a given area (4,500 sq m),
with storage tanks and product recovery absorbers at specific locations and
heights, and with fugitive sources uniformly distributed at the plant site.
Emissions from the model plant have been estimated from several sources and
E-22
-------
uniform emission rates assumed. Current levels of control on the existing
recovery absorbers were determined for each plant in calculating emissions,
and these control levels were assumed to be the same for future operations.
When the model plant was applied to existing plant emissions, it was assumed
that benzene emissions varied in direct proportion to plant capacity.
It is unlikely that any plant duplicates the model plant precisely,
and it is recognized that the estimated benzene emissions will be in error
to some degree. For example, the benzene emissions used in the original
calculations (190 kg/hr from the product recovery absorber and 1.8 kg/hr
from storage and handling) were estimated to have 95 percent confidence
limits of ±58 percent and ±20 percent, respectively. Confidence limits
were not determined for fugitive emissions (0.8 kg/hr). No numerical
estimates on the other potential errors could be determined, nor could an
overall error range be estimated from all factors.
Additional sources of potential error occur in the atmospheric disper-
sion model for benzene emissions. First, the model used weather data from
the Greater Pittsburgh Airport to project dispersion patterns for all
existing plants. Pittsburgh has a high frequency of west winds, so the
model may tend to overstate maximum risks for plants in other areas.
Because the model (the Industrial Source Complex Dispersion Model, urban
mode 2) is intended to account for effects of urban roughness and heat
sources, it may introduce inaccuracies when applied to less urbanized
areas. Second, the model assumes there is no loss of benzene from atmos-
pheric reactions or ground level absorption. If such losses occur, the
actual concentration of benzene will be less than the estimated values.
Third, atmospheric dispersion patterns of benzene emitted during control
failures may not be the same as those over the course of a year. However,
any such deviation from average would have little effect on total death
estimates because of the short failure periods assumed. This can be seen
in Table E-4. It is estimated that benzene concentrations predicted by the
12
dispersion model may be inaccurate by a factor of 2.
A final source of error is that the model measures benzene dispersion
only to 20 km. If the linear risk model is accurate, exposures at distances
greater than 20 km, however small, may be important. If such exposures
occur, the estimated number of deaths will be higher than estimated here,
particularly for the unregulated plants.
E-23
-------
E.5 REFERENCES
1. U.S. Environmental Protection Agency. Assessment of Health Effects of
Benzene Germane to Low Level Exposure. EPA-600/1-78-061. September 1978.
2. Picciano, D. Cytogenetic Study of Workers Exposed to Benzene. In:
Environmental Research (in press). 1979.
3. National Institute for Occupational Safety and Health. Criteria for a
Recommended Standard—Occupational Exposure to Benzene. HEW Publica-
tion Number (NIOSH)74-137. 1974.
4. American Conference of Governmental Industrial Hygienists. Threshold
Limit Values for Chemical Substances and Physical Agents in the Work-
room Environment with Intended Changes for 1977. 1977.
5. National Institute for Occupational Safety and Health. Revised Recom-
mendation for an Occupational Exposure Standard for Benzene.
August 1976.
6. Occupational Safety and Health Administration. Occupational Safety
and Health Standards, 29 CFR 1910.1000, Table 1-2. Publication 2206.
1976.
7. 42 FR 27452. May 27, 1977.
8. U.S. Environmental Protection Agency. Carcinogen Assessment Group
(R. Albert, Chairman). Population Risk to Ambient Benzene Exposures.
August 1978.
9. Letter Report to EPA, from Energy and Environmental Analysis, Inc.
June 20, 1978.
10. PEDCo Environmental, Inc. Preliminary Assessment of Maleic Anhydride
Manufacturing Emissions. EPA Contract Number 68-02-2515, Task 23.
April 1978.
11. Lawson, J.F. Emissions Control Options for the Synthetic Organic
Chemicals Manufacturing Industry Maleic Anhydride—Product Report.
Hydroscience, Inc. (Prepared for Office of Air Quality Planning and
Standards, U.S. Environmental Protection Agency. Research Triangle
Park, N.C.). EPA Contract Number 68-02-2577. March 1978.
12. H. E. Cramer Co., Inc. Dispersion Model Analysis of the Air Quality
Impact of Benzene Emissions from a Maleic Anhydride Plant for Four
Emission Control Options. EPA Contract Number 68-02-2507. August 1978.
13. Public Health Service, U.S. Department of Health, Education, and
Welfare. Vital Statistics of the United States, 1973 (and 1975 pre-
liminary figures).
E-24
-------
TECHNICAL REPORT DATA
(Please read liutrucliont on the reverse before completing)
1. REPORT NO.
EPA-450/3-80-001a
4. TITLE AND SUBTITLE
Benzene Emissions from the Maleic Anhydride Industry
Background Information for Proposed Standards
7. AUTHOR(S)
3. RECIPIENT'S ACCESSION>NO.
6. REPORT DATE
February 1980
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Office of A1r Quality Planning and Standards
Environmental Protection Agency
Research Triangle Park, North Carolina 27711
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
DAA fdr A1r Quality Planning and Standards
Office of Air, Noise, and Radiation
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
13. TYPE OF REPORT AND PERIOD COVERED
F i nal
14. SPONSORING AGENCY CODE
EPA/200/04
16. SUPPLEMENTARY NOTES
T6. ABSTRACT
A National Emission Standard for the control of benzene emissions from
maleic anhydride plants is being proposed under the authority of section 112
of the Clean Air Act. The proposed standard would apply to both new and
existing sources. This document contains background information and
environmental and economic assessments of the regulatory alternatives
considered in developing the proposed standard.
17,
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. cos AT I Field/Group
Air Pollution
Pollution Control
National Emission Standards for Hazardous
A1r Pollutants
Maleic Anhydride Plants
Benzene
Hazardous Pollutants
A1r Pollution Control
13b
18. DISTRIBUTION STATEMENT
Unlimited
10. SECURITY CLASS (Thlf Report)
Unclassified
ZI.NO.OFPAQE^
194
20. SECURITY CLASS (nit page)
Unclassified
22. PRICE
EPA Form 2220-1 (t-73)
------- |