&EPA
United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EPA-450/3-80-028a
December 1980
Air
Organic Chemical
Manufacturing
Volume 6: Selected
Processes
-------
EPA-450/3-80-028a
Organic Chemical Manufacturing
Volume 6: Selected Processes
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
December 1980
-------
-111-
This report was furnished to the Environmental Protection Agency by IT
Enviroscience, Inc., 9041 Executive Park Drive, Knoxville, Tennessee 37923,
in fulfillment of Contract No. 68-02-2577. The contents of this report are
reproduced herein as received from IT Enviroscience. The opinions, findings,
and conclusions expressed are those of the authors and not necessarily those
of the Environmental Protection Agency. Mention of trade names or commercial
products is not intended to constitute endorsement or recommendation for use.
Copies of this report are available, as supplies permit, 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, Virginia 22161.
-------
-v-
CONTENTS
Paqe
INTRODUCTION Vll
Product Report Page
1. CYCLOHEXANE 1-i
2. CYCLOHEXANOL/CYCLOHEXANONE 2-i
3. CHLOROBENZENES 3-i
4. MALEIC ANHYDRIDE 4-i
5. ETHYLBENZENE/STYRENE 5-i
6. CAPROLACTAM 6-i
7. ADIPIC ACID 7-i
-------
-vii-
INTRODUCTION
A. SOCMI PROGRAM
Concern over widespread violation of the national ambient air quality standard
for ozone (formerly photochemical oxidants) and over the presence of a number
of toxic and potentially toxic chemicals in the atmosphere led the Environ-
mental Protection Agency to initiate standards development programs for the
control of volatile organic compound (VOC) emissions. The program goals were
to reduce emissions through three mechanisms: (1) publication of Control Tech-
niques Guidelines to be used by state and local air pollution control agencies
in developing and revising regulations for existing sources; (2) promulgation
of New Source Performance Standards according to Section lll(b) of the Clean
Air Act; and (3) promulgation, as appropriate, of National Emission Standards
for Hazardous Air Pollutants under Section 112 of the Clean Air Act. Most of
the effort was to center on the development of New Source Performance Stan-
dards .
One program in particular focused on the synthetic organic chemical manufactur-
ing industry (SOCMI), that is, the industry consisting of those facilities
primarily producing basic and intermediate organics from petroleum feedstock
meterials. The potentially broad program scope was reduced by concentrating on
the production of the nearly 400 higher volume, higher volatility chemicals
estimated to account for a great majority of overall industry emissions. EPA
anticipated developing generic regulations, applicable across chemical and
process lines, since it would be practically impossible to develop separate
regulations for 400 chemicals within a reasonable time frame.
To handle the considerable task of gathering, assembling, and analyzing data to
support standards for this diverse and complex industry, EPA solicited the
technical assistance of IT Enviroscience, Inc., of Knoxville, Tennessee (EPA
Contract No. 68-02-2577). IT Enviroscience was asked to investigate emissions
and emission controls for a wide range of important organic chemicals. Their
efforts focused on the four major chemical plant emission areas: process
vents, storage tanks, fugitive sources, and secondary sources (i.e., liquid,
solid, and aqueous waste treatment facilities that can emit VOC).
121A
-------
-ix-
B. REPORTS
To develop reasonable support for regulations, IT Enviroscience gathered data
on about 150 major chemicals and studied in-depth the manufacture of about
40 chemical products and product families. These chemicals were chosen consid-
ering their total VOC emissions from production, the potential toxicity of emis-
sions, and to encompass the significant unit processes and operations used by
the industry. From the in-depth studies and related investigations, IT Enviro-
science prepared 53 individual reports that were assembled into 10 volumes.
These ten volumes are listed below:
Volume I
Volume II
Volume III
Volume IV
Volume V
Volume VI-X
Study Summary
Process Sources
Storage, Fugitive, and Secondary Sources
Combustion Control Devices
Adsorption, Condensation, and Absorption Devices
Selected Processes
This volume is a compilation of individual reports for the following chemical
products: cyclohexane, cyclohexanol, cyclohexanone, chlorobenzenes, maleic
anhydride, ethylbenzene, styrene, caprolactam, adipic acid. The reports
generally describe processes used to make the products, VOC emissions from the
processes, available emission controls, and the costs and impacts of those
controls (except that abbreviated reports do not contain control costs and
impacts). Information is included on all four emission areas,- however, the
emphasis is on process vents. Storage tanks, fugitive sources, and secondary
sources are covered in greater detail in Volume III. The focus of the reports
is on control of new sources rather than on existing sources in keeping with
the main program objective of developing new source performance standards for
the industry. The reports do not outline regulations and are not intended for
that purpose, but they do provide a data base for regulation development by EPA.
C. MODEL PLANTS
To facilitate emission control analyses, the reports introduce the concept of a
"model plant" (not in abbreviated reports). A model plant by definition is a
representation of a typical modern process for production of a particular chem-
ical. Because of multiple production routes or wide ranges in typical production
-------
-xi-
capacities, several model plants may be presented in one product report.
The model plants can be used to predict emission characteristics of a new
plant. Of course, describing exactly what a new plant will be like is diffi-
cult because variations of established production routes are often practiced by
individual companies. Nonetheless, model plants provide bases for making new-
plant emission estimates (uncontrolled and controlled), for selecting and siz-
ing controls for new plants, and for estimating cost and environmental impacts.
It is stressed that model-plant analyses are geared to new plants and therefore
do not necessarily reflect existing plant situations.
-------
REPORT 1
CYCLOHEXANE
J. W. Blackburn
IT Enviroscience
9041 Executive Park Drive
Knoxville, Tennessee 37923
Prepared for
Emission Standards and Engineering Division
Office of Air Quality Planning and Standards
ENVIRONMENTAL PROTECTION AGENCY
Research Triangle Park, North Carolina
September 1980
This report contains certain information which has been extracted from the
Chemical Economics Handbook, Stanford Research Institute. Wherever used, it
has been so noted. The proprietary data rights, which reside with Stanford
Research Institute, must be recognized with any use of this material.
D113A
-------
CONTENTS OF REPORT 1
Page
I- ABBREVIATIONS AND CONVERSION FACTORS 1-1
II. INDUSTRY DESCRIPTION II-l
A. Introduction II-l
B. Usage and Growth II-l
C. Domestic Producers II-l
D. References II-7
III. PROCESS DESCRIPTIONS III-l
A. Introduction III-l
B. Cyclohexane Production by Hydrogenation of Benzene III-l
C. Cyclohexane Production by Separation from Petroleum Liquids III-5
D. References III-9
IV. EMISSIONS IV-1
A. Benzene Hydrogenation Process IV-1
B. Petroleum Separation Process IV-5
C. References IV-12
V. APPLICABLE CONTROL SYSTEMS V-l
A. Benzene Hydrogenation Process V-l
B. Petroleum Separation Process V-4
D. References V-6
APPENDICES OF REPORT 1
A. PHYSICAL PROPERTIES OF BENZENE, CYCLOHEXANE, AND HYDROGEN A-l
B. AIR-DISPERSION PARAMETERS B-l
C. SAMPLE CALCULATIONS FOR PROCESS DEPRESSURIZATION LOSSES C-l
D. FUGITIVE-EMISSION FACTORS D-l
E. SAMPLE CALCULATIONS FOR HANDLING LOSSES E-l
F. EXISTING PLANT CONSIDERATIONS F-l
-------
1- v
TABLES OF REPORT 1
Table No. Page
II-l Cyclohexane Usage and Growth II-3
II-2 Cyclohexane Capacity II-4
IV-1 Cyclohexane Capacity Range by Producing Sites IV-2
IV-2 Emissions Related to Depressurization of Process Equipment IV-4
IV-3 Model Plant Storage Data IV-6
IV-4 Benzene and Total VOC Uncontrolled Emissions for Model Plant I IV-7
IV-5 Benzene and Total VOC Uncontrolled Emissions for Model Plant II IV-7
IV-6 Benzene and Total VOC Uncontrolled Emissions for Model Plant III IV-8
IV-7 Storage Data for Model Plant Using Petroleum Separation Process IV-10
IV-8 Total VOC Uncontrolled Emissions Model Petroleum Separation IV-10
Process
V-l Controlled Benzene and Total VOC Emissions for Model Plant I V-2
V-2 Controlled Benzene and Total VOC Emissions for Model Plant II V-2
V-3 Controlled Benzene and Total VOC Emissions for Model Plant III V-3
V-4 Controlled Total VOC Emissions for Model Petroleum Separation V-5
VI-1 Summary of Emissions for Benzene Hydrogenation Model Plants VI-2
VI-2 Summary of VOC Emissions for Petroleum Separation Process VI-3
A-l Physical Properties of Benzene, Cyclohexane, and Hydrogen A-l
B-l Air-Dispersion Parameters B_-i
F-l Control Devices Currently Used by the Cyclohexane Industry in p_2
the Unites States
-------
1-vii
FIGURES OF REPORT 1
Figure No.
II-l Cyclohexane and Its Relationship to Manufacture of Other II-2
Organic Chemicals
II-2 Locations of Plants Manufacturing Cyclohexane II-5
III-l Process Flow Diagram for Benzene Hydrogenation Process Model III-2
Plant (Uncontrolled Emissions)
III-2 Process Flow Diagram for Petroleum Separation Process Model III-6
Plant (Uncontrolled Emissions)
-------
1-1
ABBREVIATIONS AMD CONVERSION FACTORS
EPA policy is to express all measurements used in agency documents in metric
units. Listed below are the International System of Units (SI) abbreviations
and conversion factors for this report.
To Convert From
Pascal (Pa)
Joule (J)
Degree Celsius (°C)
Meter (m)
Cubic meter (m3)
Cubic meter (m3)
Cubic meter (m3)
Cubic meter/second
(m3/s)
Watt (W)
Meter (m)
Pascal (Pa)
Kilogram (kg)
Joule (J)
To
Atmosphere (760 mm Hg)
British thermal unit (Btu)
Degree Fahrenheit (°F)
Feet (ft)
Cubic feet (ft3)
Barrel (oil) (bbl)
Gallon (U.S. liquid) (gal)
Gallon (U.S. liquid)/min
(gpm)
Horsepower (electric) (hp)
Inch (in.)
Pound-force/inch2 (psi)
Pound-mass (Ib)
Watt-hour (Wh)
Multiply By
9.870 X 10"5
9.480 X 10~4
(°C X 9/5) + 32
3.28
3.531 X 101
6.290
2.643 X 102
1.585 X 104
1.340 X 10"3
3.937 X 101
1.450 X 10"4
2.205
2.778 X 10~4
Standard Conditions
68°F = 20°C
1 atmosphere = 101,325 Pascals
PREFIXES
Prefix
T
G
M
k
m
Symbol
tera
giga
mega
kilo
milli
micro
Multiplication
Factor
Example
1012
109
106
103
io"3
io"6
1 Tg =
1 Gg =
1 Mg =
1 km =
1 mV =
1 pg =
1 X 10 1 2 grams
1 X IO9 grams
1 X IO6 grams
1 X IO3 meters
1 X IO"3 volt
1 X IO"6 gram
-------
II-l
II. INDUSTRY DESCRIPTION
A. REASON FOR SELECTION
Cyclohexane production was selected for study because preliminary estimates
indicated that product growth rates and volatile organic compound (VOC) emissions
would be high. In addition cyclohexane manufacture was of special interest
because benzene is commonly used as a feedstock. Cyclohexane and benzene, the
two major organic components in cyclohexane manufacture, are both colorless
liquids; their pertinent physical properties are given in Appendix A.
B. USAGE AND GROWTH
Cyclohexane is the first intermediate in a series of chemicals that leads to
the production of nylon 6,6 and nylon 6; the intermediates and their relation-
ships to cyclohexane are shown in Fig. II-l. Table II-l shows cyclohexane end
uses and expected growth rates.
The annual growth rate of cyclohexane production by the benzene hydrogenation
process is reported to be 6 to 6 1/2%.1 Production by the petroleum separation
process is expected to decrease annually by 1 1/2%.2 if an overall cyclohexane
production growth rate of 5%/yr occurs, as expected, for the period 1976 to
1982.
The 1977 domestic cyclohexane production capacity was reported to be 1395 giga-
grams (Gg); 71% of the capacity was utilized.1 Based on an overall growth rate
of 5%, over 90% of the domestic capacity will be utilized by 1982.
C. DOMESTIC PRODUCERS
In the late 1960s, 17 producers of cyclohexane in the United States (including
Puerto Rico) shared a market of only a few large customers. Competition forced
a number of the producers to close their plants, and today only 9 producers
with 11 plants produce cyclohexane.1 Table II-2 gives the producers, along
with the estimated capacities and the process used; Fig. II-2 shows the manu-
facturing locations.
The availability of low-cost, high-purity hydrogen is a prerequisite for cyclo-
hexane production by either process. Consequently, all cyclohexane production
-------
II-2
CYCLOWEXAklE
i
OXIME
CAPROLACTAM
(o
SiYLCW
/EXPORTS^
-W AklD
CYCLOHEXAVUOL
ACID
MVLOU 4^6* COMOK1OMEP
(HMDA')
Fig. II-1. Cyclohexane and Its Relationship to Manufacture
of Other Organic Chemicals
-------
II-3
Table II-l. Cyclohexane Usage and Growth
Percent of Production Average Annual Growth
End Use (1976) (%) (1976-1981)
Adipic acid
Exports
Capro lac tarn
1 , 6-Hexamethylenediamine (HMDA)
Miscellaneous
Average
53
18
23
3
3
5.0
8.5 — 10.0
2.0
1.5
6.0 — 6.5
aSee ref 1.
-------
Table II-2. Cyclohexane Capacity0
Producer
American Petrofina, Inc.
Cosden Oil & Chemical Co., sub.
Commonwealth Oil Refining Co. , Inc.
Corco Cyclohexane, Inc. , sub.
Exxon Chemical Co., USA
Division of Exxon Corp.
Gulf Oil Chemicals Co.
Division of Gulf Oil Corp.
Phillips Petroleum Co.
Phillips Puerto Rico Core, Inc. , sub.
Sun Company, Inc.
Texaco, Inc.
Union Oil of California
Union Pacific Corp.
Champlin Petroleum Co., sub.
Total
With expansions
Location
Big Spring, TX
Penuelas, PR
Baytown, TX
Port Arthur, TX
Borger, TX
Sweeny, TX
Guayama , PR
Tulsa, OK
Port Arthur , TX
Beaumont , TX
Corpus Christi, TX
1976 Capacity
(Gg)
35
117
147
106
117
250
212
(expanding to 265)
59
(expanding to 88)
117
88
65
1313
1395
Process
b
b
b
b
c
b,c
b
b
b
b
b
See ref 1.
DBenzene hydrogenation,
'Petroleum separation.
-------
II-5
1. Cosden - Big Spring, TX
2. Corco - Penuelas, PR
3. Exxon - Baytown, TX
4. Gulf - Port Arthur, TX
5. Phillips - Borger, TX
6. Phillips - Sweeny, TX
7. Phillips - Guayama, PR
8. Sun - Tulsa, OK
9. Texaco - Port Arthur, TX
10. Union Oil - Beaumont, TX
11. Champlin - Corpus Christi, TX
-------
II-6
plants are located in petrochemical complexes. However, with such processes as
fuel desulfurization competing for hydrogen in today's petrochemical complexes,
hydrogen is becoming an increasingly scarce and valuable chemical. Growth in
cyclohexane production may require new facilities to increase hydrogen production
through partial oxidation of hydrocarbon streams.2 For this reason cyclohexane
will continue to be produced only in petrochemical complexes.
-------
II-7
D. REFERENCES*
1. J. L. Blackford, "Cyclohexane," pp 638.5061A, 638.5062W in Chemical Economics
Handbook, Stanford Research Institute, Menlo Park, CA (February 1977).
2. R. W. Smith, "What's Happening with Cyclohexane?" Chemical Engineering Progress
73(9), 25—28 (1977).
*Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of
that paragraph, that reference number is indicated on the material involved.
When the reference appears on a heading, it refers to all the text covered by
that heading.
-------
III-l
III. PROCESS DESCRIPTIONS
A. INTRODUCTION
Two processes are used commercially to manufacture cyclohexane: catalytic
hydrogenation of benzene, which accounts for approximately 85% of the cyclo-
hexane capacity in the United States,- and separation from petroleum liquids,
which constitutes the remaining 15%.
B. CYCLOHEXANE PRODUCTION BY HYDROGENATION OF BENZENE
1. Basic Process
The reaction utilized to make cyclohexane from benzene is:
C6H6 + 3H2 > C6H12
(benzene) (hydrogen) (cyclohexane)
Figure III-l shows a model-plant* flow diagram for the manufacture of cyclo-
hexane by benzene hydrogenation. High-purity benzene is stored in large tanks
near the production plant. Benzene (stream 1) is fed to the reactors in paral-
lel, whereas hydrogen (stream 5) is fed into the reactors in series. Part of
the cyclohexane separated in the flash separator is recycled (stream 3) and fed
to the reactors in series. Recycling helps to control the reactor temperature,
since the reaction is highly exothermic. Typical temperatures and pressures in
the reaction section are 150 to 260°C and 2.1 to 3.5 MPa. The temperature is
also controlled by generating steam, which is used elsewhere in the petrochemical
complex. Both platinum and nickel catalysts are presently used to produce
cyclohexane.
The cyclohexane (stream 9) leaving the flash separator is sent to a distillation
column (stabilizer) for removal of methane, ethane, other light hydrocarbons,
and soluble hydrogen gas from the cyclohexane product. These impurities
(stream 10) are routed to the complex-wide fuel-gas storage system and used as
fuel in process heaters. Flow of this stream for one manufacturer is reported
to be 0.018 kg of gas/kg of cyclohexane produced at capacity. Composition is
reported to be 42.2 mole % H_, 44.7 mole % single-carbon compounds, 11.9 mole %
*See page 1-2 for a discussion of model plants.
-------
Fig. III-l. Process Flow Diagram for Benzene Hydrogenation Process
Model Plant (Uncontrolled Emissions)
TO FUE\_
BEKIZ.EWJE
-^- LIQUIDS-BYPRODUCT'S
TO PETROCHEMCAL
COMPLEX
i AQUEOUS STREAM
PROM PUA.K1TS
WATER WASH
(B)
- BYPRODUCTS
TO PUEL,
SYSTEM
- BYPRODUCTS
TO FU&-
SY5TGM
BE'-ZEME
TAMK PA
CATM.YST
CYCLOMEXAME CYCuowEXAme BAR&E
TAkiK. FARM
(PROM W
OtJLX)
"
HYDRO^SKAATIOM
-------
III-3
two-carbon compounds, with the remainder being oxygen, argon, and nitrogen.1
Cyclohexane (stream 11) purified in the stabilizer may be greater than 99.9%
pure.2'3 The residual benzene content is typically less than 500 mg/liter.2
This pure product is stored in large tanks to await shipment.
Gas from the flash separator, largely hydrogen (60 to 80 mole %), is not pure
enough for direct reuse. This stream (6) is purified to greater than 90 mole %
before being recycled to (stream 5) the reactor. Typical processes used for
hydrogen purification are absorption and stripping of the hydrogen gas and
cryogenic separation; some plants use a combination of the two processes.
Organic liquids (stream 12) separated from the hydrogen in the hydrogen purifica-
tion unit are sent to other petroleum processing units in the petrochemical
complex. The separated gases (stream 13) are used as fuel gas.
Depending on the type of hydrogen purification used, inert impurities present
in the gas from the flash separator can be purged from the system before the
gas enters the hydrogen purification equipment. This stream (8) is sent to the
fuel gas system. One manufacturer reports that the flow of this stream is
0.96 kg of gas/kg of cyclohexane produced at capacity.1 The composition of the
stream is the same as that of stream 10. The flow and composition of stream 8
will vary for each manufacturer because hydrogen may be purified by several
different methods.
There are no process emissions during normal operation. During shutdowns
individual equipment vents are opened as required during final depressurization
of equipment. Except for the feed streams the concentration of benzene in the
process equipment is low; therefore little or no benzene emissions would be
expected during a shutdown.1
Fugitive leaks can emit benzene, cyclohexane, methane, or other hydrocarbons.
Leaks from heat exchangers into cooling water or steam production can be a
potential fugitive loss. Fugitive losses are of special significance because
of the high diffusivity of hydrogen at elevated temperature and pressure and
the extremely flammable nature of the liquid and gas processing streams.
-------
III-4
Storage emission sources (labeled B on Fig. III-l) include benzene and cyclo-
hexane storage. In plants where absorption is used for hydrogen purification,
an additional organic compound (toluene, glycol amine, etc.) may be stored, but
the quantities of these solvents are usually small relative to the quantities
of benzene and cyclohexane. Handling emissions (C, Fig. III-l) relate to
transfer of cyclohexane to shipping vessels.
The potential sources of secondary emissions (K, Fig. III-l) are catalyst
handling and absorber wastewater (where an aqueous solution is used to purify
the recycled hydrogen). Plants comprising at least 16% of the total cyclohexane
capacity use an aqueous solution to purify hydrogen. Caution is taken to
remove the organic from the spent catalyst before it is replaced. The spent
catalyst is sold for metal recovery.
2. Process Variations
The most significant variation in the cyclohexane process is the type of catalyst
used. Present-day plants use either platinum or nickel on an inert support as
a catalyst. Operating temperatures and pressures vary in relation to which
catalyst is used. Nickel catalysts are poisoned by sulfur compounds; therefore
benzene and hydrogen feeds must be sulfur-free. Also, nickel catalysts require
monitoring and replacement at more frequent intervals than do platinum catalysts.
One manufacturer using a nickel catalyst replaces all the catalyst beds every
four years,3 whereas a manufacturer using a platinum catalyst reports catalyst
lifetimes in excess of ten years.2 Nickel catalysts also require a small supple-
mental heater to supply high-temperature hydrogen to convert the nickel catalyst
from an oxidized form to a reduced form before startup of a new bed.
Other variations in cyclohexane production include the number of reactors, the
type of reactor cooling equipment, and the level of process control used. One
manufacturer uses steam-driven compressors instead of electrically powered
compressors because the local power company is unreliable and plant safety
would be threatened unless hydrogen flow were maintained through the reactors.
The steam-driven compressor assures that there will be hydrogen flow during
power interruption.3
-------
III-5
C. CYCLOHEXANE PRODUCTION BY SEPARATION FROM PETROLEUM LIQUIDS
The petroleum separation process is used at two sites to produce cyclohexane
from petroleum fractions,- these sites prouce 15% of the cyclohexane produced
domestically. Cyclohexane of low purity can be commercially obtained by conven-
tional distillation; however, the presence of benzene and isomers of cyclohexane
(methylcyclopentane, hexanes, etc.) with vapor pressures similar to that of
cyclohexane makes the distillation of cyclohexane with purity greater than 85%
very difficult. To produce high-purity cyclohexane, additional process steps
are incorporated.
Figure III-2 is a flow diagram of the process used to manufacture high-purity
(99%) cyclohexane from petroleum streams.4'5 A petroleum fraction rich in
cyclohexane (stream 1) is fed to a distillation column, where benzene and
methylcyclopentane are removed (stream 2) and routed to a hydrogenation unit.
The bottoms (stream 3) from the column containing cyclohexane and other hydro-
carbons are combined with another petroleum stream (4) and sent to a catalytic
reformer, where the cyclohexane is converted to benzene. The hydrogen generated
in this step (stream 5) is used in the hydrogenation step or elsewhere in the
petrochemical complex.
The benzene-rich stream (6) leaving the catalytic reformer is sent to a distil-
lation column, where compounds that have vapor pressures higher than benzene
(pentanes, etc.) are removed (stream 7) and used as by-products. The benzene-
rich stream (8) that is left is sent to another distillation column, where the
benzene and methylcyclopentane (stream 9) are removed. The remaining hydrocarbons
(largely dimethyIpentanes) are used elsewhere in the petrochemical complex as
by-products (stream 10).
Stream 9 (benzene and methylcyclopentane) is combined with stream 2 and sent to
a hydrogenation unit similar to the one discussed in the hydrogenation process
(Sect.III-B). Hydrogen is fed to this unit, and the benzene is converted to
cyclohexane (stream 11). Isomers of cyclohexane such as methylcyclopentane are
converted to cyclohexane in an isomerization unit, and the effluent (stream 12)
from this equipment is separated in a final distillation step. Pure cyclohexane
(stream 14) is separated from isomers of cyclohexane (stream 13) and compounds
with lower vapor pressures (stream 15).
-------
Fig. III-2. Process Flow Diagram for Petroleum Separation Process
Model Plant (Uncontrolled Emissions), (refs-4,5)
{"}
PETROLEUM
PETROCHeM.
T
" T
-TO
PUEL SA-b 1
=>Y=>T£M ^ ^
£
CATALVTIC
PETROLEUM BYPRODUCTS
" (PeUTAM&b1) TO
PETROCHEM\CA>_
UQLJIOS - PETROLEUM
-^BYPRODUCTS TO
COMPLEX
\
T
" T
^
Di-bTIU-A-TiOW
^
TO FXJEU
in,
•*
/
TO PUEL,
PXJAU
REACTORS
' ' TTO
PETROUEUM
FROM peTP.oc.HEMicA.i-
COMPV.SK.
TO PETROCHEMICAL COMPLE.X
5MUTOOWM
PETROUEVJM «bEPARA.T\OU
,. ^
'9104-10
-------
Fig. III-2. (Continued)
ISOMERIZER
,+
5PEUT
PETROLEUM BYPRODUCT
(HEXAWE'b £ MCTWLcvcLOP
TO P6TROCWEM>CAJL COMPLEX
DISTILLATIOW
PRODUCT
y/ORWU<^
STORAGE
CYCV.OWEXAUE
PRODUCT
TAWK. FARM
STORAGE
BARGEE - SHIP
PETROLEUM BYPRODUCT
(ue^s vouwvv-e TWA
TO PETROCHEMICAL COMPLEX
- PBOCEAS EMI^iSlOU (PROM lUO\VlOOta_ 6OOIPMEUT VEMTi DURlUt
" " " ~ OK11-V)
PETROLEUM
CYCLOVAEXAVJE
-z. , a_
JWB
9IOA-IO
-------
III-8
There are no process emissions during normal operation. During emergency shut-
downs individual equipment vents are opened as required.
Equipment leaks can be sources of benzene, cyclohexane, methane, or other
petroleum compounds emissions. Leaks from heat exchangers into cooling water
or steam production can be a potential fugitive loss. Fugitive losses are of
special significance because of the high diffusivity of hydrogen at elevated
temperatures and pressures and the extremely flammable nature of the liquid and
gas processing streams.
Storage and handling emission sources (B and C, Fig. III-2) include cyclohexane
storage and shipping. Storage of the petroleum feed stream is not included.
A potential source of secondary emissions (K, Fig. III-2) is catalyst handling.
Caution is taken to remove the organic from the spent catalyst before it is
replaced. The spent catalyst is sold for metal recovery.
-------
III-9
D. REFERENCES*
1. K. Pardue, Cosden Oil and Chemical Co., Big Spring, TX, letter dated January
1978 to D. R. Goodwin, EPA, in response to EPA request for information on the
cyclohexane process (on file at EPA, ESED, Research Triangle Park, NC.)
2. J. W. Blackburn, Trip Report on Site Visit to Exxon Chemical Company,
Baytown, TX, Sept. 15, 1977 (on file at EPA, ESED, Research Triangle Park, NC)
3. J. W. Blackburn, Trip Report on Visit to Phillips Puerto Rico Core, Inc.,
Guayama, PR, Sept. 20, 1977 (on file at EPA, ESED, Research Triangle Park, NC)
4. F. A. Lowenheim and M. K. Moran, pp 298—303 in Faith, Keyes, and Clark's
Industrial Chemicals, 4th ed., Wiley-Interscience, New York, 1975.
5. Telephone conversation on Jan. 6, 1978, between J. W. Blackburn and M. F.
Potts, Process Engineering, Phillips Petroleum Company, Bartlesville, OK,
relative to nonconfidential aspects of the petroleum separation process for
cyclohexane manufacture.
*Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of
that paragraph, that reference number is indicated on the material involved.
When the reference appears on a heading, it refers to all the text covered by
that heading.
-------
IV-1
IV. EMISSIONS
Emissions in this report are usually identified in terms of volatile organic
compounds (VOC). VOC are currently considered by the EPA to be those of a
large group of organic chemicals, most of which, when emitted to the atmosphere,
participate in photochemical reactions producing ozone. A relatively small
number of organic chemicals have low or negligible photochemical reactivity.
However, many of these organic chemicals are of concern and may be subject to
regulation by EPA under Section 111 or 112 of the Clean Air Act since there are
associated health or welfare impacts other than those related to ozone forma-
tion.
A. BENZENE HYDROGENATION PROCESS
The size of a cyclohexane plant is largely determined by the quantities of
benzene and hydrogen available at a petrochemical complex. Each complex produces
different petrochemicals and may use different primary petroleum feeds. The
availability of benzene and hydrogen will also differ from site to site.
Therefore it is not possible to specify the size of a typical plant that is
expected to be constructed in the future.
1. Model Plants I, II, and III*
Since the capacities of actual cyclohexane plants vary widely (Table IV-1),
three model plants with different capacities must be used. The three model
plants for which emissions are described have capacities of 50, 150, and 250 Gg/hr,
respectively, based on 8760 hr of operation per year.** The benzene hydrogenation
process used in these plants is shown in Fig. III-l and described in Sect. III.
Platinum catalyst technology is used in all three plants. Air-dispersion para-
meters for each of the plants are given in Appendix B.
Benzene storage is sized to provide a 12-day supply. Working cyclohexane
storage is 4 days, and tank-farm cyclohexane storage is 30 days. These storage
*See p 1-2 for a discussion of model plants.
**Process downtime is normally expected to range from 5 to 15%. If the hourly
production rate remians constant, the annual production and annual VOC emissions
will be correspondingly reduced. Control devices will usually operate on the
same cycle as the process. From the standpoint of cost effectiveness calcula-
tions the error introduced by assuming continuous operation is negligible.
-------
IV-2
Table IV-1. Cyclohexane Capacity Range by Producing Sites'
Capacity Range
(Gg/yr)
30 — 100
100 — 200
over 200
Sites
4
5
1
10
Combined Capacity
(Gg/yr)
276
642
265
1183
Percent of Total
Capacity
23.2
54.3
22.4
100.0
For benzene hydrogenation process only.
-------
IV-3
capacities are consistent with data received from cyclohexane manufacturers.1'2
Actual tank capacities are standard for typical API Standard 650 cone-roof
tanks. The number of valves, pumps, and compressors used in the model plants is
discussed in Sect. 2.b. The number of fugitive emission sources is based on data
received from cyclohexane manufacturers1'2 and is expected to be typical of the
three model plants.
2. Sources and Emissions
a. Process Emissions (Shutdown) There are no continuous process emissions from
any of the cyclohexane model plants. Gaseous streams separated from the process
that contain hydrogen and hydrocarbons (streams 8, 10, and 13, Fig. III-l) are
sent to the fuel gas system at the petrochemical complex. This gas is used as
fuel in process heaters supplying energy for other petrochemical processes.1—6
Liquid hydrocarbons separated from recycled hydrogen (stream 12, Fig. III-l)
are by-products that are used elsewhere at the petrochemical complex. No gas
or liquid streams are discharged to the atmosphere on a continuous basis.
Some process emissions arise from shutdown of the process equipment. When
equipment is shut down for maintenance, the high-pressure vapor is released to
the fuel-gas system until the pressure in the equipment is the same as the
pressure in the fuel-gas system. From this point on, the remaining pressure is
relieved by discharging the vapor to a flare or to the atmosphere through a
blowdown tank. The blowdown tank collects any liquid that may form during the
depressurization process. During an emergency shutdown individual equipment
vents are opened as required to relieve the pressure.
Table IV-27 lists the estimated depressurization-related emissions resulting
from equipment shutdown and the assumptions on which the estimates are based.
Two cases are presented. For Case I the emission is mostly hydrogen (equilibrium
temperature is 25°C). For Case II the emission is totally cyclohexane vapor
(equilibrium temperature is 81°C). Sample calculations relating to Table IV-2
are presented in Appendix C. Benzene is not present in depressurization-
related emissions.4
b. Fugitive Emissions Process pumps, valves, and compressors are potential
sources of fugitive emissions. Each model plant is estimated to have 15 pumps
-------
IV-4
a,b
Table IV-2. Emissions Related to Depressurization of Process Equipment
Model Model Model
Plant I Plant II Plant III
Capacity (Gg/yr) 50 150 250
Equipment internal void
column (m3) 11 33 55
Emissions (kg of VOC/yr)
Case I
Equilibrium temperature,
25°C; vapor composition,
12.8 mole % cyclohexane 25 74 123
Case II
Equilibrium temperature,
81°C; vapor composition,
100 mole % cyclohexane 162 485 808
asee ref 7.
Assumptions: 1. Initial pressure before depressurization, 517 kPa; final
pressure after depressurization, 101 kPa. 2. Gases are ideal at these
pressures and temperatures. 3. Cyclohexane is the only organic present.
4. No condensate collection devices (blowdown) are used. 5. Equipment
void volume is proportional to capacity. 6. One complete depressuriza-
tion each year.
-------
IV-5
handling light liquids (2 on benzene), 150 process valves on gas or vapor (15
on benzene), 150 process valves on light liquids (15 on benzene), 7 relief
valves on gas or vapor (1 on benzene), 8 relief valves on light liquids (1 on
benzene), and 1 hydrogen compressor. All pumps utilize mechanical seals. The
factors used to establish the emission rates are shown in Appendix D.
c. Storage and Handling Emissions Storage and handling emissions result from the
handling of benzene and cyclohexane. For the model plants the sources are
shown as B and C on the flow diagram, Fig. III-l. The storage tank specifica-
tions and conditions for all model plants are given in Table IV-3. The emissions
listed in Tables IV-4, IV-5, and IV-6 are based on fixed-roof tanks, one-half
full, and a 12.1°C diurnal temperature variation,- the emission equations from
AP-428 were used. Weather conditions correspond to data from the Houston, TX,
area. Emissions calculated for fixed-roof tanks include working and breathing
losses. Equations from AP-42 were used for these calculations also; however,
breathing losses were divided by 4 to account for recent evidence indicating
that the AP-42 breathing loss equation overestimates emissions.9
Handling emissions from loading cyclohexane in ships and barges were calculated
with the equations in AP-42.8 Sample calculations are presented in Appendix E.
d. Secondary Emissions Secondary emissions resulting from catalyst replacement
are small because of precautions taken to clean the organic from the catalyst
before it is removed. Catalyst reclamation is performed off-site.1'2
The only process wastewater generated in cyclohexane manufacture is the effluent
from absorbers in which aqueous streams are used to purify spent hydrogen.3'4'6
These streams are sent to API separators. One plant reports 0.16-kg/hr VOC
emitted by this secondary source.3
B. PETROLEUM SEPARATION PROCESS
The petroleum separation process shown in Fig. III-2 is expected to have
emissions similar to those from the benzene hydrogenation process.10 The major
difference is the use of seven major processing steps instead of the three used
in the benzene hydrogenation method. As in benzene hydrogenation, organic
streams separated in the processing equipment are routed to other units in the
petrochemical complex as by-products or are burned in process heaters as fuel gas.
-------
IV-6
Table IV-3. Model-Plant Storage Data
Content
Benzene
Cyclohexane
Cyclohexane
Benzene
Cyclohexane
Cyclohexane
Benzene
Cyclohexane
Cyclohexane
Number of
Tanks
Model Plant
1
1
1
Model Plant
1
3
3
Model Plant
1
3
3
Volume
. T. a
(m )
Turnovers
per year
Bulk Liquid
Temperature (°C)
I , 50-Gg/yr Capacity
1,928
799
4,784
II, 1 50-Gg/yr
5,693
799
4,784
III, 250-Gg/yr
10,675
1,196
7,971
29.5
80.4
13.4
Capacity
30.0
80.4
13.4
Capacity
26.7
89.5
13.4
20.3
20.3
20.3
20.3
20.3
20.3
20.3
20.3
20.3
Per tank.
-------
IV-7
Table IV-4. Benzene and Total VOC Uncontrolled Emissions for Model Plant I
Source
Process emissions,
shutdown
Storage
Handling
c
Fugitive
Secondary
Total
Stream
a
Designation
A
B
C
J
K
b
Emission Ratio
Benzene Total VOC
3.2
425 1308
147
154 1419
13
Emission
Benzene
2.42
0.88
3.30
Rate (kg/hr)
Total VOC
0.02
7.47
0.84
8.1
0.07d
16.50
See Fig. III-l.
g of emission per Mg of cyclohexane produced; emission ratios are valid only
for the model plant operating at capacity.
°See Appendix D.
j
Only for plants employing aqueous absorption for hydrogen purification.
Table IV-5. Benzene and Total VOC Uncontrolled Emissions for Model Plant II
Source
Process emissions
shutdown
Storage
Handling
Fugitive
Secondary
Total
Stream
Designation
A
B
C
J
K
Emission Ratio
Benzene Total VOC
3.2
421 1305
148
51.4 473
13
Emission
Benzene
7.21
0.88
8.09
Rate (kg/hr)
Total VOC
0.05
22.3
2.53
8.1
0.22d
33.20
See Fig. III-l.
bg of emission per Mg of cyclohexane produced; emission ratios are valid only
for the model plant operating at capacity.
"See Appendix D.
*0nly for plants employing aqueous absorption for hydrogen purification.
-------
IV-8
Table IV-6. Benzene and Total VOC Uncontrolled Emissions for Model Plant III
Source
Process emissions,
shutdown
Storage
Handling
c
Fugitive
Secondary
Total
Stream
a
Designation
r
A
B
C
J
K
Emission Ratio Emission Rate (kg/hr)
Benzene Total VOC Benzene
3.2
420 1256 12.0
148
30.8 284 0.88
13
12.88
Total VOC
0.09
35.9
4.22
8.1
0.37d
48.68
See Fig. III-l.
°g of emission per Mg of cyclohexane produced; emission ratios are valid only
for the model plant operating at capacity.
•"i
"See Appendix D.
Only for plants employing aqueous absorption for hydrogen purification.
-------
IV-9
1. Model Plant
The model plant producing cyclohexane by petroleum separation has a capacity of
100 Gg/yr of cyclohexane based on 8760 hr of operation per year. Storage
requirements for the working cyclohexane tanks are 4 days of retention time;
tank-farm cyclohexane storage is 30 days of retention time.1'2 (Storage for
this process may sometimes be combined with storage for other processes in
those locations where cyclohexane is manufactured by both processes.) No
storage is assumed for the petroleum feed.
The number of valves, pumps, and compressors is discussed in Sect.B.Z.b. These
sources were estimated from data supplied on the benzene hydrogenation process.1'2
2. Sources and Emissions
a. Process Emissions (Shutdown) The emissions from this source are estimated
from the values given in Table IV-2, with allowance made for an increased
internal void volume for the equipment related to this process. An internal
void volume of 77 m results in emissions of 177 kg of VOC per depressurization
for Case I and in 1132 kg of VOC per depressurization for Case II, as defined
in the table. Assumptions for these calculations are also given in Table IV-2.
One complete equipment depressurization per year is assumed and the worst case,
Case II, is used.
b. Fugitive Emissions Process pumps, valves, and compressors are potential
sources of fugitive emissions. Each model plant is estimated to have 35 pumps
handling light liquids, 200 process valves on gas or vapor, 500 process valves
on light liquids, 10 relief valves on gas or vapor, 25 relief valves on light
liquids, and 1 hydrogen compressor. All pumps utilize mechanical seals. The
factors used to establish the emission rates are shown in Appendix D.
c. Storage and Handling Emissions Storage and handling emissions result from the
handling of petroleum liquids and cyclohexane. The sources for the model plant
for this process are shown as B and C on the flow diagram, Fig. III-2. The
storage tank specifications and conditions are given in Table IV-7. The emissions
given in Table IV-8 are based on fixed-roof tanks, one-half full, and a 12.1°C
diurnal temperature variation; emission equations from AP-428 were used,
together with the breathing loss equation adjustment discussed in Sect. A.2.c.
-------
Per tank.
IV-10
Table IV-7. Storage Data for Model Plant
Using Petroleum Separation Process
Content
Petroleum liquid
Cyclohexane
Cyclohexane
Number of
Tanks
None
2
3
Volume
(m3)a
799
3843
Turnovers
per year
80.4
11.1
Bulk Liquid
Temperature ( °
20.3
20.3
C)
Table IV-8. Total VOC Uncontrolled Emissions from
Plant Using Petroleum Separation Process
Source
Process emissions ,
shutdown
Storage
Handling
Fugitive0
ijiH'ondary
Total
Stream
a
Designation
A
B
C
J
K
. b
Emission Ratio
11
873
148
1367
Emission Rate
(kg/hr)
0.13
9.97
1.69
15.6
27.39
See Fig. III-2.
g.of emission per Mg of cyclohexane produced; emission ratios are valid only
for the model plant operating at capacity.
•^
"See Appendix D.
-------
IV-11
Handling emissions from loading cyclohexane in ships and barges were calculated
with the equations from AP-42.8
d. Secondary Emissions Secondary emissions result only from catalyst handling
and are small because of precautions taken by the manufacturers before removal
of the catalyst. Catalyst reclamation is performed off-site.1'2
No process wastewater is expected. Aqueous streams are not used to purify
hydrogen.
-------
IV-12
C. REFERENCES*
1. J. W. Blackburn, Trip Report on Visit to Exxon Chemical Company, Baytown,
TX, Sept. 15, 1977 (on file at the EPA, ESED, Research Triangle Park, NC).
2. J. W. Blackburn, Trip Report on Visit to Phillips Puerto Rico Core, Inc.,
Guayama, PR, Sept. 20, 1977 (on file at the EPA, ESED, Research Triangle Park,
NC).
3. M. P. Zanotti, Gulf Oil Col, USA, Port Arthur, TX, letter dated Jan. 26, 1978,
to D. R. Goodwin, EPA, in response to EPA request for information on the cyclo-
hexane process (on file at EPA, ESED, Research Triangle Park, NC).
4. K. Pardue, letter dated Jan. 24, 1978, to D. R. Goodwin, EPA, Cosden Oil and
Chemical Co., Big Spring, TX, in response to EPA request for information on the
cyclohexane process (on file at EPA, ESED, Research Triangle Park, NC).
5. R. L. Chaffin, Champlin Petroleum Co., Corpus Christi, TX, letter dated Jan.
25, 1978, to D. R. Goodwin, EPA (on file at EPA, ESED, Research Triangle Park,
NC).
6. W. W. Dickinson, Sun Petroleum Production Co., Tulsa, OK, letter dated Jan. 26
1978, to D. R. Goodwin, EPA (on file at EPA, ESED, Research Triangle Park, NC).
7. R. C. Reid, J. M. Prausnitz, and T. K. Sherwood, The Properties of Gases and
Liquids, 3d ed., McGraw-Hill, New York, 1977.
8. C. C. Masser, "Storage of Petroleum Liquids," pp 4.31 to 4.3-17 in Compilation
of Air Pollutant Emission Factors, AP-42, Part A, 3d ed. (August 1977).
9. Letter dated May 30, 1979, from E. C. Pulaski, TRW, Inc. to Richard Burn, EPA.
10. J. W. Blackburn, IT Enviroscience, Inc., telephone conversation on Jan. 6, 1978,
with M. F. Potts, Phillips Petroleum Co., Bartlesville, OK.
*Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of
that paragraph, that reference number is indicated on the material involved.
When the reference appears on a heading, it refers to all the text covered by
that heading.
-------
V-l
V. APPLICABLE CONTROL SYSTEMS
A. BENZENE HYDROGENATION PROCESS
1. Process Sources
The gas vented to the atmosphere when equipment is depressurized during shutdown
can be routed to a flare. A flare could be dedicated for this emission or
another available flare may be used. Flare efficiencies are dependent on
design, gas flow, and size. When in smokeless operation and operating with a
gas flow of 10% to 100% of the maximum smokeless design, the efficiencies
should be greater than 98% for flares with tips greater than 12 in. and greater
than 99% for flares with tips less than 12 in. An efficiency factor of 98%
will be assumed in this study since the flare utilized may have been designed
for a variety of applications. The cost effectiveness of flares is not presented
in Sect. VI because the flare would probably not be dedicated solely to cyclohexane
process duty.
2. Fugitive Sources
Controls for fugitive sources will be discussed in a future document covering
fugitive emissions from the synthetic organic chemicals manufacturing industry.
Controlled fugitive emissions have been calculated based on the factors given
in Appendix D and are included in Tables V-l, 2, and 3.
3. Storage and Handling Sources
Storage guidelines for all organic compounds will be covered in a future EPA
document. Control for storage losses involves the use of floating-roof tanks*
or retrofitting floating roofs to existing fixed-roof tanks. Emissions listed
in Tables V-l through V-3 assume a contact-type internal floating roof with
secondary seals is used and will reduce fixed-roof-tank emissions by 85%.3
Control of cyclohexane loading emissions is not practiced in the cyclohexane
industry. No reduction in loading emissions is assumed for this study. Control
systems using vapor-recovery, dedicated flares, and other technologies have
been used in the chemical industry to control organic emissions resulting from
chemical loading.
*Consist of internal flaoting covers or covered floating roofs as defined in
API-2519, 2d ed., 1976 (fixed-roof tanks with internal floating device to
reduce vapor loss).
-------
V-2
Table V-I. Controlled Benzene and Total VOC Emissions for
Model Plant I (Internal-Floating-Roof Tanks)
Emission Ratio Emission Rate (kg/hr)
sure dm
Source Designation
Process emissions,
shutdown
Storage
Handling
Fugitive
Secondary
Total
A
B
C
J
K
Benzene Total VOC
0.06
63.6 196
147
35.0 314
13
Benzene Total VOC
0.0003
0.36 1.12
0.84
0.20 1.79
0.07
0.56 3.82
aSee Fig. III-l.
b
g of emission per Mg of cyclohexane produced; emission ratios are valid only for the
model plant operating at capacity.
Q
See Appendix B.
d
Only for plants employing aqueous absorption for hydrogen purification.
Table V-2. Controlled Benzene and Total VOC Emissions for
Model Plant II {Internal-Floating-Roof Tanks)
Emission Ratio Emission Rate (kg/hr)
Source
Process emissions,
shutdown
Storage
Handling
Fugitive
Secondary
Total
ot-j-ecuu — .....
Designation3 Benzene
A
B 63.2
C
J 11.7
K
Total VOC
0.06
196
148
105
13
Benzene Total VOC
0.001
1.08 3.35
2.53
0.20 1.79
0.22d
1.28 7.89
aSee Fig. III-l.
g of emission per Mg of cyclohexane produced; emission ratios are valid only for the
model plant operating at capacity.
£•
See Appendix D.
d
Only for plants employing aqueous absorption for hydrogen purification.
-------
V-3
Table V-3. Controlled Benzene and Total VOC Emissions for
Model Plant III (Internal-Floating-Roof Tanks)
Emission Ratio (g/Mg)
Emission Rate (kg/hr)
Source
process emissions,
shutdown
Storage
Handling
Fugitive
Secondary
Total
•J L-U-ecUll
Designation Benzene
A
B 63.0
C
J 7.0
K
Total VOC
0.06
188
148
62.7
13
Benzene Total VOC
0.002
1.80 5.38
4.22
0.20 1.79
0.37d
2.00 11.76
asee Fig. III-l.
g of emission per Mg of cyclohexane produced; emission ratios are valid only for the model
plant operating at capacity.
"See Appendix D.
Only for plants employing aqueous absorption for hydrogen purification.
-------
V-4
4. Secondary Sources
Secondary emissions are small. Plants using aqueous absorption to purify the
hydrogen employ API separators to prevent water pollution. Emissions listed in
Tables V-l through V-3 are based on data from industry.4
B. PETROLEUM SEPARATION PROCESS
Control devices listed for the benzene hydrogenation process are also applicable
to the petroleum separation process. Table V-4 gives the controlled emissions
from the petroleum separation process, with internal-floating-roof tanks the
major control device.
-------
V-5
Table V-4. Model-Plant Controlled Total VOC Emissions for
Petroleum Separation Process
(Internal-Floating-Roof Tanks)
Source
Process emissions,
shutdown
Storage
Handling
c
Fugitive
Secondary
Total
a
Designation
A
B
C
J
K
Emission Ratio
(g/Mg)b
0.22
135
148
335
Emission Rate
(kg/hr)
0.003
1.55
1.69
3.82
7.06
See Fig. III-2.
g of emission per Mg of cyclohexane produced; emission ratios are valid only
for the model plant operating at capacity.
•*
'See Appendix D.
-------
V-6
C. REFERENCES*
1. V. Kalcevic, IT Enviroscience, Inc., Flares and the Use of Emissions as Fuels
(in preparation for EPA, ESED, Research Triangle Park, NC).
2. C. C. Masser, "Storage of Petroleum Liquids," pp 4.3-1 to 4.3-17 in Compilation
of Air Pollutant Emission Factors, AP-42, Part A, 3d ed. (August 1977).
3. Letter dated Aug. 15, 1979, from W. T. Moody, TRW, Inc. to Dave Beck, EPA.
4. Nonconfidential information received January 1978 from Gulf Oil Company—USA,
Port Arthur Refinery, XX (on file at ESED, EPA, Research Triangle Park, NC).
^Usually, when a reference number is located at the end of a paragraph, it
refers to the entire paragraph. If another reference relates to certain
portions of that paragraph, that reference number is indicated on the material
involved.
-------
VI-1
VI. SUMMARY
Two site visits and additional contact with the industry indicated that emissions
expressed in the literature are nearly 10 times as high as emissions found in
this study, which possibly reflects a greater concern for reuse of spent hydrogen
and petroleum by-products in today's market. This increased concern has grown
from a change in the economics of the petrochemical industry during the last
decade. Cyclohexane manufacture is currently one of the lowest emitters of the
eight benzene-consuming processes studied by IT Enviroscience.
Cyclohexane is predominantly used in the production of caprolactam, adipic
acid, and hexamethylenediamine intermediates in the manufacture of nylon 6 and
nylon 6,6. The demand for cyclohexane is expected to grow at a rate of 5% per
year through 1982. The present capacity of the nine United States cyclohexane
producers should be adequate throughout this period.
The predominant cyclohexane manufacturing process is the catalytic hydrogena-
tion of benzene, which is capable of producing 99.9% pure cyclohexane. The
process generates significant quantities of energy, usually as steam, and also
as fuel gas. Organic-laden spent hydrogen is recovered by hydrogen purification
processes and is reused in cyclohexane and other hydrogen-consuming processes.
Because of the scarcity of hydrogen in today's petrochemical refinery, every
effort is made to prevent hydrogen losses as emissions. This also results in
very low organic emissions. Potential emissions are eliminated by recycling
the process streams as by-products or by using them as fuel.
In a cyclohexane plant organic emissions can occur from benzene and cyclohexane
storage and cyclohexane loading operations. If fugitive sources are not controlled,
fugitive losses can be a significant organic source. Table VI-1 summarizes the
uncontrolled and controlled emissions from the model plants using the benzene
hydrogenation process; Table VI-2 summarizes the model-plant's uncontrolled and
controlled emissions for the petroleum separation process.
The combined VOC emissions from all domestic cyclohexane production would be
336 kg/hr if uncontrolled and 78 kg/hr if controlled.
-------
Table VI-1. Summary of Emissions for Benzene Hydrogenation Model Plants'
(Internal-Floating-Roof Storage Emission Control)
Emissions (kg/hr) for Emissions (kg/hr) for
Model Plant I Model Plant II
Emissions (kg/hr) for
Model Plant III
Uncontrolled Controlled Uncontrolled Controlled Uncontrolled Controlled
Emission Source
Process
Storage
Handling
Fugitive
Secondary
Total
Benzene VOC Benzene
0.02
2.42 7.47 0.36
0.84
0.88 8.1 0.20
0.07b
3.30 16.50 Q.56
VOC Benzene
0.0003
1.12 7.21
0.84
1.79 0.88
0.07b
3.82 8.09
VOC Benzene
0.05
22.3 1.08
2.53
8.1 0.20
0.22b
33.20 1.28
VOC Benzene
0.001
3.35 12.0
2.53
1.79 0.88
0.22b
7.89 12.88
VOC Benzene
0.09
35 . 9 1 . 80
4.22
8.1 0.20
0.37b
48.68 2.00
VOC
0.002
5.38
4.22
1.79
0.37b
11.76
Capacities: Model Plant I, 50 Gg/yr; Model Plant II, 150 Gg/yr; Model Plant III, 250 Gg/yr.
Only for plants employing aqueous absorption for hydrogen purification.
i
ro
-------
VI-3
Table VI-2. Summary of VOC Emissions for Petroleum
Separation Process (Internal-Floating-Roof Tank Emission Control)
Uncontrolled Controlled
Emission Source (kg/hr) (kg/hr)
Process 0.13 0.003
Storage 9.97 1.55
Handling 1.69 1.69
Fugitive3 15.6 3.82
Secondary
Total 27.39 7.06
a
See Appendix D.
-------
VI-4
The manufacturers presently use a mixture of fixed-roof and floating-roof
storage tanks for organic storage. Use of floating-roof tanks is a significant
control measure. The actual national emissions therefore would range between
the uncontrolled and controlled values stated above. An assumption of 50%
floating-roof usage would indicate the current national VOC emissions from
cyclohexane to be about 200 kg/hr. Exact calculation of national emissions is
not possible since the actual storage facilities for each manufacturer are not
presently known. Information regarding the emission control measures presently
practiced in the cyclohexane industry is given in Appendix F.
-------
A-l
APPENDIX A
Table A-l. Physical Properties of Benzene, Cyclohexane/ and Hygrogen
Benzene
Cyclohexane
Hydrogen
Synonyms
Molecular fomula
Molecular weight
physical state
Vapor pressure
Vapor specific gravity
Boiling point
Melting point
Liquid specific gravity
Water solubility
Octanol/water
partition coefficient
Benzol, coal naphtha,
phenylhydride
C..H,.
6 6
78.11
Liquid
95.9 mm Hg at 25°C
2.77
80.1°C at 760 mm Hg
5.5°C
0.8787 at 200C/4°C
1.79 g/liter
2.28
Hexahydrobenzene,
hexanaphthene,
hexamethylene
C6H12
84.16
Liquid
98.14 mm Hg at 25°C
(96.97 mm Hg at 25°C)'
2.90
80.7°C at 760 mm Hg
6.3°C
0.77855 at 20°C/4°C
<1 g/liter
H2
2.01
Gas
From J. Dorigan, B. Fuller, and R. Duffy, "Benzene," p AI-102 in Chemistry , Production
and Toxicity of Selected Synthetic Organic Chemicals (Chemicals A-c) , MTR-7248, Rev 1,
Appendix I, MITRE Corp., McLean, VA (September 1976).
b
c
J. Dorigan, B. Fuller, and R. Duffy, "Cyclohexane," ibid, p AI-318.
From R. C. Reid ejt a^. , The Properties of Gases and Liquids, McGraw-Hill, New York, 1977.
-------
APPENDIX B
Table B-l. Air-Dispersion Parameters
Emission Source
Model Plant I, 50 Gg/yr
Storage
Uncont ro 1 led
Benzene
Cyclohexane
Cyclohexane
Storage
Controlled
Benzene
Cyclohexane
Cyclohexane
Fugitive
Uncontrolled
Controlled
Handling
Model Plant II, 150 Gg/yr
Storage
Uncontrolled
Benzene
Cyclohexane
Cyclohexane
Controlled
Benzene
Cyclohexane
Cyclohexane
Number
of
Sources
1
1
1
1
1
1
1
3
3
1
3
3
Emission
Rate
(g/sec)
0.67
0.45
0.95
0.10
0.068
0.14
2.25
0.50
0.23
2.00
0.45
0.95
0.30
0.068
0.14
Tank
Height
(m)
14.63
9.75
12.19
14.63
9.75
12.19
12.19
9.75
12.19
12.19
9.75
12.19
Tank
Diameter
(m)
12.95
10.21
22.35
12.95
10.21
22.35
24.38
10.21
22.35
24.38
10.21
22.35
Discharge
Temperature
(K)
Ambient
Ambient
Ambient
Ambient
Ambient
Ambient
Ambient to 530
Ambient to 530
Ambient
Ambient
Ambient
Ambient
Ambient
Ambient
Ambient
Flow
Rate
(m-Vsec)
w
H
-------
Table B-l. (Continued)
Number
of
Emission Source Sources
Fugitive
Uncontrolled
Controlled
Handling
Model Plant III, 250 Gg/yr
Storage
Uncontrolled
Benzene 1
Cyclohexane 3
Cyclohexane 3
Controlled
Benzene 1
Cyclohexane 3
Cyclohexane 3
Fugitive
Uncontrolled
Controlled
Handling
Model Plant, Petroleum Separation Process
Storage
Uncontrolled
Cyclohexane 2
Cvclcrtxexarve "*
Emission Tank Tank
Rate Height Diameter
(g/sec) (m) (m)
2.25
0.50
0.23
3.33 14.63 30.48
0.68 12.19 11.18
1.52 17.07 24.38
0.50 14.63 30.48
0.10 12.19 11.18
0.23 17.07 24.38
2.25
0.50
0.23
0.66 9.75 10.21
0.45 14.63 18.29
Discharge Flow
Temperature Rate
(K) (m3/sec)
Ambient to 530
Ambient to 530
Ambient
Ambient i
to
Ambient
Ambient
Ambient
Ambient
Ambient
Ambient to 530
Ambient to 530
Ambient
Ambient
Ambient
-------
Table B-l. (Continued)
Emission Source
Controlled
Cyclohexane
Cyclohexane
Fugitive
Uncontrolled
Controlled
Handling
Number Emission Tank Tank Discharge Flow
of Rate Height Diameter Temperature Rate
Sources (g/sec) (m) (m) (K) (m3/sec)
2 0.098 9.75 10.21 Ambient
3 0.067 14.63 18.29 Ambient
4.33 . Ambient to 530
1.06 Ambient to 530
0 . 47 Ambient
w
i
CO
-------
C-l
APPENDIX C
SAMPLE CALCULATIONS FOR PROCESS DEPRESSURIZATION LOSSES
CALCULATION FOR PROCESS LOSSES DUE TO SHUTDOWN DEPRESSURIZATION
Case I ; Equilibrium flash temperature: 25 °C
Initial pressure 75 psia
3875 mm Hg
0.517 MPa
Final pressure 14.7 psia
760 mm Hg
0.101 MPa
Composition of escaping gas at equilibrium:
vapor pressure of cyclohexane at 25 °C
Y (mole fraction) = - - total final pressure *
/ \
vapor pressure of cyclohexane = e [15.7527 - _ 2766.63 \
\ T (K) - 50.50/
vapor pressure of cyclohexane at 25°C = 96.97 mm Hg.
y (mole fraction) = 9^7 = 0.128.
Moles of escaping gas at equilibrium
359 sc£ x ^§JS x 76Q ™ Hg = ?6>8 frl^^e (at equilibrium conditions)
Ib mole 273 K 3878 mm Hg
1160 ft3 (equipment volume) = 15>1 lb_nole released.
76.8 ft3/lb-mole
Cyclohexane released
15.1 (Ib-mole of gas released) X 0.128 (mole fraction of cyclohexane)
= 1.933 Ib-moles of cyclohexane,
= 163 Ib of cyclohexane,
= 74 kg of cyclohexane.
Case II: Pure cyclohexane vapor released (temperature = 81 °C)
Initial pressure 75 psia
3875 mm Hg
0.517 MPa
*Reid et al.., The Properties of Gases and Liquids, McGraw-Hill, New York, 1977.
-------
C-2
Final pressure 14.7 psia
760 mm Hg
0.101 MPa
Composition of escaping gas
, vapor pressure at 81 °C
y (mole fraction) = *~ ^ ^ Rg -
y (mole fraction) - 1.0.
Moles of escaping gas at equilibrium
359 scf „ 354 K „ 760 mm Hg n. „_ _
- ^ X 3878 mm Hg " 91'23 ft
1160 ft3 . _ ^_ n,
°f
91.23
Amount of cyclohexane relased
12.72 Ib-moles of gas X "^l6 °^ ^yclohexane 1Q68 cyclohexane
y Ib-mole of gas J
= 485 kg of cyclohexane •
-------
D-l
APPENDIX D
FUGITIVE-EMISSION FACTORS*
The Environmental Protection Agency recently completed an extensive testing
program that resulted in updated fugitive-emission factors for petroleum re-
fineries. Other preliminary test results suggest that fugitive emissions from
sources in chemical plants are comparable to fugitive emissions from correspond-
ing sources in petroleum refineries. Therefore the emission factors established
for refineries are used in this report to estimate fugitive emissions from
organic chemical manufacture. These factors are presented below.
Source
Uncontrolled
Emission Factor
(kg/hr)
Controlled
Emission Factorc
(kg/hr)
Pump seals
Light-liquid service
Heavy-liquid service
Pipeline valves
Gas/vapor service
Light-liquid service
Heavy-liquid service
Safety/relief valves
Gas/vapor service
Light-liquid service
Heavy-liquid service
Compressor seals
Flanges
Drains
0.12
0.02
0.021
0.010
0.0003
0.16
0.006
0.009
0.44
0.00026
0.032
0.03
0.02
0.002
0.003
0.0003
0.061
0.006
0.009
0.11
0.00026
0.019
Based on monthly inspection of selected equipment; no inspection of
heavy-liquid equipment, flanges, or light-liquid relief valves;
10,000 ppmv VOC concentration at source defines a leak; and 15 days
allowed for correction of leaks.
Light liquid means any liquid more volatile than kerosene.
*Radian Corp., Emission Factors and Frequency of Leak Occurrence for Fittings
in Refinery Process Units, EPA 600/2-79-044 (February 1979).
-------
E-l
APPENDIX E
SAMPLE CALCULATIONS FOR HANDLING LOSSES
1. Data
1976 capacity shipped = 342,105 bbls X 3 = 1,026,315 bbls/yr.
By tanker = 90% X 1,026,315 = 924,000 bbls .
tanker capacity = 300,000 bbls 3 tankers/yr.
By barge = 1,026,315 - 900,000 = 126,315 bbls
barge capacity =13,000 bbls 10 Barges/yr.
2. Product Data
Chemical, cyclohexane
Molecular weight = 84.16.
Specific gravity = 0.779 (20/4), 0.7834 (60°F/60°F).
Vapor pressure = 60 mm Hg at 14.7°C,
100 mm Hg at 25.5°C.
40°F 50°F 60°F 70°F 80°F 90°F 100°F
0.677 psia 0.928 psia 1.218 psia 1.605 psia 2.069 psia 2.610 psia 3.249 psia
Ambient average temperature = 79.3°F.
3. Loading Loss, Equations, and Terms
L = 12.46 ~ lb/103 gal of liquid loaded.
S = 0.2 for ships; 0.5 for barges.
P = true vapor pressure: assume 2.069 psia at 80 °F.
M = Molecular weight = 84.16 Ib/lb-mole.
T = bulk temperature: assume 79°F + 460° = 539°R.
3.1. For 3 t ankers /yr:
(12.46) (0.2) (2.069) '(84.16) _ on ,,,-nnn
L (tankers) = - - - — i - -* • - - — * - - = 0.80 lb/1000 gal.
3.2. For 10 barges/yr:
T ,v, ^ (12.46) (0.5) (2.069) (84.16)
L (barge) = jrrr
o m IK
bbl J
*
Supplement No. 7 for Compilation of Air Pollutant Emission Factors, API Bulletin 42,
2d ed., pp 4.3-6 to 4.3311, American Petroleum Institute, Washington, DC (April 1977)
-------
E-2
4. Total Emissions (Loading)
Tankers = 15.20 tons/yr
Barges = 5.49 tons/yr
20.69 tons/yr
-------
F-l
APPENDIX F
EXISTING PLANT CONDITIONS
Table F-l lists the control devices reported to be in use by industry.1—6
Further information on the processes follows:
A. EXXON AT BAYTOWN, TX2
1. Process Description
The nameplate capacity of Exxon's cyclohexane production plant is 3000 bbl/day
of 99.95 mole % cyclohexane. The plant was started in 1958, with a process
expansion in 1967.
The process was licensed from UOP. The hydrogen and recycled cyclohexane flow
in series through the reactors and benzene is added in parallel to each of the
reactors. The process uses a platinum metal catalyst, which has a life of
about 10 years.
Variations between the Exxon process and a standard process are a "methanator"
reactor, which catalytically converts trace quantities of CO in the feed hydrogen
to C02; five cyclohexane reactors; a scrubber that removes cyclohexane vapor
from the spent-hydrogen stream and uses benzene as the scrubbing fluid; a
second scrubber that removes benzene from the spent-hydrogen stream and uses
toluene as the scrubbing fluid; and the ability to vary the sources and sinks
of feed hydrogen and spent hydrogen as the overall plant-wide hydrogen demands
vary.
The necessity to vary the hydrogen feed stream and spent-hydrogen stream is
important. The Exxon chemical plant is located in close proximity to an Exxon
petroleum refinery. These two plants are separated adminstratively, but they
transfer chemical raw materials and products from one to the other on a con-
tinuous basis. All the benzene feed to the cylcohexane unit is produced by
extraction and purification of a refinery side stream containing high concen-
trations of aromatics. The hydrogen for the cyclohexane unit originates from
reformers in the refinery. However, hydrogen purity for cyclohexane production,
as well as for other hydrogen-consuming processes, must be higher than that
obtained directly from the reformers. Therefore during the past decade cryogenic
-------
Table F-l. Control Devices Currently Used
by the Cyclohexane Industry in the United States
Company
Cosden Oil
Corco Cyclohexane
Exxon Chemical
Gulf Oild
Phillips Petroleum
Borger, TX
Sweeny, TX
Guayama , PR
Sunf
Texaco
Union Oil of California
g
Champlin Petroleum
Floating-Roof Tanks
Benzene Storage
N.D.b
N.D.
Yes
N.D.
N.D.
N.D.
No
N.D.
N.D.
N.D.
N.D.
(Open Top and Internal)
Cyclohexane Storage
N.D.
N.D.
Yes
N.D.
N.D.
N.D.
Yes
N.D.
N.D.
N.D.
N.D.
Flares (Process or Fugitive)
Yes
N.D.
No
N.D.
N.D.
N.D.
Yes
N.D.
N.D.
N.D.
Yes
See ref 1.
bNo data.
"See ref 2.
See ref 3.
a
"See ref 4.
See ref 5.
gSee ref 6.
i
to
-------
F-3
hydrogen purification units have been added to increase the hydrogen purity
from 70 to 85 mole % to the 95 mole % required for cyclohexane and other processes.
In a petrochemical complex such as Exxon's the supply and demand of the valuable
chemical, hydrogen, often helps to determine the overall product mix of the
plant. Production of compounds requiring hydrogen is lowered when the hydrogen
supply is short. The sources of feed hydrogen and the "consumers" of spent
hydrogen at the Baytown plant are flexible so that cyclohexane operation can be
continued during changes in the hydrogen supply situation. Every attempt is
made to recycle hydrogen-rich gases to processes directly or by purification
units. Gases containing hydrogen but rich in organics are candidates for the
plant-wide fuel gas system. In this case heat is recovered from the gas.
Direct loss of hydrogen resulting from continuous or intermittent atmospheric
process emissions or from the lack of heat recovery is avoided. Therefore loss
of organics in these process-related streams is minimized. Exxon estimates
that about one-fifth of the spent hydrogen from cyclohexane production is
routed to the adjacent refinery complex, with the bulk sent to a localized
cryogenic purifier. The direct recycle of spent hydrogen into the hydrogen feed
to the cyclohexane plant occurs when there is a shortage of hydrogen on a
plant-wide basis.
The off-gas from the stabilizer still, which is sent to the plant-wide fuel gas
system, is composed predominently of H2 and CH4 staturated with cyclohexane.
Startup/Shutdown
The cyclohexane unit does not require annual shutdowns for maintenance; but
when the unit does have to be shut down, the procedure is to stop the benzene
feed to the reactors. The hydrogen and cyclohexane recycled flow is continued
until the reactor temperatures are at a safe level. During this period the
spent hydrogen is sent to the fuel gas manifold, where it remains until the
pressure in the system is the same as the manifold back-pressure. The remaining
pressure (V75 psi) is vented to the atmosphere through a condensible blowdown
drum. This drum catches some of the organic that would otherwise be emitted to
the atmosphere. This emission has not been quantified.
-------
F-4
In an emergency condition, which is estimated to occur about twice a year, only
the piece of equipment affected is depressurized. Again, the pressure is
relieved until it is the same as the fuel-gas manifold back-pressure. The
remainder is vented to the atmosphere. No quantitative data involving this
intermittent emission were obtained.
B. PHILLIPS AT GUAYAMA, PR4
1. Process Description
Most of the information regarding the process description is confidential and
therefore is not included. However, the following information was given in a
non-confidential manner:
The cyclohexane process at Guayama is a Phillips process and was developed
within the corporation. It is similar to the hydrogenation production units at
Phillips Petroleum, Borger, TX, and Sweeny, TX, locations. The original designs
for the American plants were based on a similar benzene hydrogenation plant in
Antwerp, Belgium.
The Guayama plant was built in 1967 and was started up in 1968. There have
been two expansions, the most recent one occurring in 1976. The present capacity
is 6000 bbl/day of 99.5 mole % cyclohexane.
The cyclohexane unit at Guayama is part of a petrochemical operation using
petroleum naphtha as the fundamental feed. A desulfurization unit removes
sulfur from the organic streams, allowing the use of a nickel catalyst to
hydrogenate benzene to cyclohexane. Nickel catalysts are poisoned by sulfur
compounds.
The catalyst life for this process is about 4 years. In other words, one of
the four catalyst-containing reactors is replaced each year. Exothermic runaway
reactions are possible due to the type of catalyst used, and therefore elaborate
temperature monitoring procedures are used to predict the changes in catalyst
performance and impending runaway reactions. The probability of a runaway or
exothermic reaction increases as the catalyst ages. When it is suspected that
a bed is involved, it is removed from the system and the catalyst is replaced.
-------
F-5
When a bed is to be replaced, it is isolated from the system and allowed to
cool. Before the catalyst is removed from the reactor, it is purged with
nitrogen under a 5-psia vacuum supplied by two jets. The off-gas from this
system is sent to a flare. The nitrogen is used to remove organics from the
catalyst and to prevent a hydrogen explosion. The bed is then removed and sold
for nickel recovery. New catalyst in reduced form is loaded into the reactor
and is then oxidized with hot air and steam. A small process heater using fuel
gas supplies the hot air and steam for the oxidizing process. It consumes less
than 4 million Btu/hr when in operation and is used from 10 to 20 days a year.
Complete combustion is assumed by Phillips.
Hydrogen is in short supply at Guayama, and the cyclohexane unit has a lower
priority for hydrogen than other processes. For this reason production fluc-
tuates with hydrogen availability. Spent hydrogen from the cyclohexane process
is purified in a cryogenic system, also developed by Phillips. Purified hydrogen
is routed back to the cyclohexane unit and elsewhere in the plant.
Startup/Shutdown
The public electrical power system in Guayama is unreliable, and plant personnel
estimate that power failures occur about once a month. When a power failure
occurs, the hydrogenation reaction may be exothermic. The benzene feed must be
stopped, and the hydrogen flow must be maintained to sustain internal cooling
and prevent a runaway reaction. The compressor supplying hydrogen to the
reactors is driven by a steam turbine instead of an electric motor. The proba-
bility of power and steam being lost simultaneously is small and occurrences of
reactor runaways are minimized. When a runaway reaction is not controlled by
all these precautions, it is vented to a flare. No quantitative data are
available on this emission, but it is felt to be very low.
GULF OIL AT PORT ARTHUR, TEXAS3
Gulf Oil lists the following description of the off-gas from a water absorber
used to purify spent hydrogen:
-------
F-6
Temperature: ambient
Flow: 0.76 Ib/hr
Composition Amount (wt %)
Nitrogen 2.5
Hydrogen 41.5
Methane 9.0
Ethane 16.2
Propane 18.1
Isobutane 1.3
n-Butane 8.0
Isopentane 1.7
n-Pentane 1.7
100.0
D. SUN PETROLEUM AT TULSA, OK.5
Information from Sun Petroleum shows no emissions.
E. COSDEN OIL AND CHEMICAL COMPANY AT BIG SPRING, XX1
Information from Cosden gives flow and composition information for pressure-
relief valves, fuel-gas streams, and off-gas from process heaters. Cosden also
claims to have no measurable levels of benzene in their streams except for the
reactor feed streams.
F. CHAMPLIN PETROLEUM AT CORPUS CHRISTI, TX.6
Little data were given by Champlin. Off-gases from column reboilers are sent
to smokeless flares. Fuel-gas streams are recovered.
-------
F-7
G. REFERENCES*
1. K. Pardue, Cosden Oil and Chemical Co., Big Spring, TX, letter dated January
1978 to D. R. Goodwin, EPA, in response to EPA request for information on
cyclohexane prodess (on file at EPA, ESED, Research Triangle Park, NC).
2. J. W. Blackburn, IT Enviroscience, Inc., Trip Report on Visit to Exxon Chemical
Company, Baytown, TX, Sept. 15, 1977 (on file at the EPA, ESED, Research
Triangle Park, NC).
3. M. P. Zanotti, Gulf Oil Company—USA, Port Arthur TX, letter dated Jan. 26,
1978 to D. R. Goodwin, EPA, in response to EPA request for information on
cyclohexane process (on file at EPA, ESED, Research Triangle Park, NC).
4. J. W. Blackburn, IT Enviroscience, Inc., Trip Report on Visit to Phillips Puerto
Rico Core, Inc., Guayama, PR, Sept. 20, 1977 (on file at the EPA, ESED, Research
Triangle Park, NC).
S. Sun Petroleum Production Co., Tulsa Refinery (on file at ESED, EPA, Research
Triangle Park, NC).
6. W. W. Dickinson, Champlin Petroleum Co., Corpus Christi TX, letter dated
Jan. 26, 1978 to D. R. Goodwin, EPA, in response to EPA request for information
on cyclohexane process (on file at EPA, ESED, Research Triangle Park, NC).
*Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of that
paragraph, that reference number is indicated on the material involved. When
the reference appears on a heading, it refers to all the text covered by that
heading.
-------
2-i
REPORT 2
CYCLOHEXANOL/CYCLOHEXANONE
J. W. Blackburn
V. Kalcevic
W. D. Bruce
IT Enviroscience
9041 Executive Park Drive
Knoxville, Tennessee 37923
Prepared for
Emission Standards and Engineering Division
Office of Air Quality Planning and Standards
ENVIRONMENTAL PROTECTION AGENCY
Research Triangle Park, North Carolina
December 1980
This report contains certain information that has been extracted from the
Chemical Economics Handbook, Stanford Research Institute. Wherever used, it
has been so noted. The proprietary data rights that reside with Stanford
Research Institute must be recognized with any use of this material.
Dl-E
-------
2-iii
CONTENTS OF REPORT 2
Page
I. ABBREVIATIONS AND CONVERSION FACTORS 1-1
II. INDUSTRY DESCRIPTION II-l
A. Selection of Cyclohexanol/Cyclohexanone II-l
B. Cyclohexanol/Cyclohexanone Usage and Growth II-l
C. Domestic Producers II-2
D. References II-6
III. PROCESS DESCRIPTIONS III-l
A. Introduction III-l
B. Cyclohexane Oxidation Process III-3
C. Phenol Hydrogenation Process III-5
D. References III-8
IV. EMISSIONS IV-1
A. Cyclohexane Oxidation Process IV-1
B. Phenol Hydrogenation Process IV-4
C. References IV-10
V. APPLICABLE CONTROL SYSTEMS V-l
A. Cyclohexane Oxidation Process V-l
B. Phenol Hydrogenation Process V-3
C. References V-6
VI. IMPACT ANALYSIS VI-1
A. Environmental and Energy Impacts VI-1
B. Control Cost Impact VI-4
C. References VI-12
VII. SUMMARY VII-1
-------
2-v
APPENDICES OF REPORT 2
A. PHYSICAL PROPERTIES OF CYCLOHEXANOL, CYCLOHEXANONE, BENZENE,
CYCLOHEXANE, AND PHENOL
B. AIR-DISPERSION PARAMETERS
C. FUGITIVE-EMISSION FACTORS
D. COST-ESTIMATE SAMPLE CALCULATIONS
E. EXISTING PLANT CONSIDERATIONS
-------
2-vii
TABLES OF REPORT 2
Number Page
II-l Cyclohexanol/Cyclohexanone Capacity II-3
IV-1 Uncontrolled Emissions from Model Plant — Cyclohexane IV-2
Oxidation Process
IV-2 Model Plant Storage Tank Data — Cyclohexane Oxidation IV-5
Process
IV-3 Benzene and VOC Uncontrolled Emissions from Model Plant — IV-7
Phenol Hydrogenation Process
IV-4 Molar Composition of the Hydrogenation Reactor Vent — Phenol IV-7
Hydrogenation Process
IV-5 Model Plant Storage Tank Data — Phenol Hydrogenation Process IV-9
V-l Controlled Emissions from Model Plant Using Cyclohexane Oxidation V-2
Process
V-2 Controlled Emissions from Model Plant — Phenol Hydrogenation V-4
Process
VI-1 Environmental Impact of Controlled Cyclohexane Oxidation Model VI-2
Plant
VI-2 Environmental Impact of Controlled Phenol Hydrogenation Model VI-3
Plant
VI-3 Annual Cost Parameters VI-5
VI-4 Emission Control Cost Estimates for Ethylene Dichloride Model VI-8
Plants
VII-1 Emission Summary for Cyclohexane Oxidation Model Plant VII-2
VII-2 Emission Summary for Phenol Hydrogenation Model Plant VII-3
A-l Physical Properties of Cyclohexanol A~l
A-2 Physical Properties of Cyclohexanone A-1
A-3 Physical Properties of Benzene A-2
A-4 Physical Properties of Cyclohexane A-3
A-5 Physical Properties of Phenol A-4
B-l Air-Dispersion Parameters for Cyclohexane Oxidation B-l
Model Plant
B-2 Air-Dispersion Parameters for Phenol Hydrogenation B-2
Model Plant
-------
2-ix
FIGURES OF REPORT 2
Number Page
II-l Manufacturing Locations of Cyclohexanol/Cyclohexanone II-4
III-l Process Flow Diagram for Uncontrolled Model Plant Producing III-4
Cyclohexanone/Cyclohexanol by Cyclohexane Oxidation
III-2 Process Flow Diagram for Uncontrolled Model Plant Producing III-6
Cyclohexanol/Cyclohexanone by Phenol Hydrogenation
VI-1 Installed Capital Cost vs Plant Capacity for Emission Control VI-7
(Thermal Oxidation)
VI-2 Net Annual Cost vs Plant Capacity for Emission Control VI-9
(Thermal Oxidation)
VI-3 Cost Effectiveness vs Plant Capacity for Emission Control VI-10
(Thermal Oxidation)
D-l Precision of Capital Cost Estimates D-2
-------
1-1
ABBREVIATIONS AND CONVERSION FACTORS
EPA policy is to express all measurements used in agency documents in metric
units. Listed below are the International System of Units (SI) abbreviations
and conversion factors for this report.
To Convert From
Pascal (Pa)
Joule (J)
Degree Celsius (°C)
Meter (m)
Cubic meter (m3)
Cubic meter (m3)
Cubic meter (m3)
Cubic meter/second
(m3/s)
Watt (W)
Meter (m)
Pascal (Pa)
Kilogram (kg)
Joule (J)
To
Atmosphere (760 mm Hg)
British thermal unit (Btu)
Degree Fahrenheit (°F)
Feet (ft)
Cubic feet (ft3)
Barrel (oil) (bbl)
Gallon (U.S. liquid) (gal)
Gallon (U.S. liquid)/min
(gpm)
Horsepower (electric) (hp)
Inch (in.)
Pound-force/inch2 (psi)
Pound-mass (Ib)
Watt-hour (Wh)
Multiply By
9.870 X 10"6
9.480 X 10"4
(°C X 9/5) + 32
3.28
3.531 X 101
6.290
2.643 X 102
1.585 X 104
1.340 X 10"3
3.937 X 101
1.450 X 10"4
2.205
2.778 X 10"4
Standard Conditions
68°F = 20°C
1 atmosphere = 101,325 Pascals
PREFIXES
Prefix
T
G
M
k
m
M
Symbol
tera
giga
mega
kilo
milli
micro
Multiplication
Factor
Example
1012
IO9'
106
103
io"3
io"6
1 Tg =
1 Gg =
1 Mg =
1 km =
1 mV =
1 ug =
1 X 10 12 grams
1 X IO9 grams
1 X IO6 grams
1 X IO3 meters
1 X IO"3 volt
1 X IO"6 gram
-------
II-l
II. INDUSTRY DESCRIPTION
A. SELECTION OF CYCLOHEXANOL/CYCLOHEXANONE
Cyclohexanol and cyclohexanone are currently synthesized industrially in either
of two basic processes -- cyclohexane oxidation or phenol hydrogenation.
Generally, both cyclohexanol and cyclohexanone are formed simultaneously in
either process. The ratio of cyclohexanol to cyclohexanone in the product
depends on the catalyst used and the process operating conditions. Since
manufacture of some products, primarily adipic acid, does not require separa-
tion of cyclohexanol/cyclohexanone product into its individual constitutuents,
this product report addresses both compounds simultaneously.
Cyclohexanol/cyclohexanone was selected as a product for study for the follow-
ing reasons: Preliminary data from emission inventory questionnaires indicate
that current emission factors for volatile organic compounds (VOC) are about
0.044 kg/kg of cyclohexanol/cyclohexanone produced by the cyclohexane oxidation
-4
process, and about 9.2 X 10 kg/kg of cyclohexanol/cyclohexanone produced by
the phenol hydrogenation process. Based on the total 1976 capacity for manu-
facture of cyclohexanol/cyclohexanone by cyclohexane oxidation, a preliminary
estimate of 15.9 Gg/yr of VOC emissions was obtained for this segment of the
industry. Similarly, gross emissions of VOC resulting from the phenol hydrogena-
tion process were estimated to be 0.11 Gg/yr.
Cyclohexanol is a colorless, crystalline solid at normal room conditions (21.1°C
and 101.3 kPa), but cyclohexanone is a colorless liquid at the same conditions.
Solutions of cyclohexanone and cyclohexanol may be solids or liquid at ambient
conditions, depending on the relative amounts of each constituent. Other
pertinent physical property data for cyclohexanol and cyclohexanone are given
in Appendix A.
B. CYCLOHEXANOL/CYCLOHEXANONE USAGE AND GROWTH
Cyclohexanol/cyclohexanone is used largely for the manufacture of adipic acid
and caprolactam. Adipic acid is a monomer for nylon 6,6 and a raw material for
manufacture of its co-monomer, hexamethylenediamine. Caprolactam is the sole
monomer for the production of nylon 6.
-------
II-2
Current overall capacity for cyclohexanol/cyclohexanone is 1147 Gg/yr. Reported
production data are believed to be understated, but based on adipic acid plant
capacity, production, and growth it is estimated that 1979 production was
1065 Gg of combined cyclohexanol/cyclohexanone and that growth will be 1—2%
per annum through 1984.*
More than 90% of the cyclohexanol produced domestically is used in the manu-
facture of adipic acid, which is consumed primarily for nylon 6,6. The remain-
der of the cyclohexanol produced is used in applications other than nylon, such
as the use of cyclohexanol as a stabilizer and dye solvent in the textile
2
industry and for manufacture of phthalate esters for use in plasticizers.
Approximately 95% of the cyclohexanone produced is consumed in caprolactam
manufacture for nylon 6 and is oxidized as the cyclohexanol/cyclohexanone mixture
to adipic acid. Use as a solvent accounts for the remainder of the cyclohexanone
2
produced.
C. DOMESTIC PRODUCERS
There are seven companies producing cyclohexanol/cyclohexanone, with a combined
2
capacity of 1147 Gg/yr for nine plants. About 80% of the capacity is based on
the oxidation of cyclohexane and the remainder on hydrogenation of phenol.
Table II-l lists cyclohexanol/cyclohexanone producers, their plant locations,
production capacities, and the basic raw material. Figure II-l shows the plant
locations on a map.
Companies producing cyclohexanol/cyclohexone are the following:
1. Allied Chemical Corp. produces mainly cyclohexanone, which is used captively
in the production of caprolactam. A palladium catalyst is used for hydro-
genation of captive phenol. Cyclohexanone and cyclohexanol are separated
c i 2'3
for sale.
2. Celanese Corp. oxidizes purchased cyclohexane, with the majority of the
product used captively to produce adipic acid for nylon 6,6 fibers.
3. Dow Badische Co. oxidizes cyclohexane, and the product is used for capro-
lactam production.
*IT Enviroscience estimate.
-------
II-3
Table II-l. Cyclohexanol/Cyclohexanone Capacity
Company
Allied Chemical Corp.
Celanese Corp.
Dow Badische Co.
Monsanto Co.
Nipro, Inc.
Union Carbide Corp.
Du Pont
Production Capacity
Location (Gg/yr) (Jan 1979) Raw Material
Hopewell, VA
Bay City, TX
Freeport, TX
Pensacola/ FL
Luling, LA
Augusta , GA
Taft, LA
Orange, TX
Victoria, TX
191
45
141
227
23
157
9
123
231
1147
Phenol
Cyclohexane
Cyclohexane
Cyclohexane
Phenol
Cyclohexane
Phenol
Cyclohexane
Cyclohexane
See ref 3.
-------
11-4
(1) Allied Chemical Corp., Hopewell, VA
(2) Celuncse Corp., Bay City, TX
(3) Dow Badische Co., Freeport, TX
(4) Monsanto Co., Pensacola, FL
(5) Monsanto Co., Luling, LA
(6) Nipro, Inc., Augusta, GA
(7) Union Carbide Corp., Taft, LA
(8) Du Pont, Orange, TX
(9) Du Pont, Victoria, TX
Fig. II-l. Manufacturing Locations of Cyclohexanol/Cyclohexanone
-------
II-5
4. Monsanto Co. at Pensacola, FL, oxidizes purchased cyclohexane, which is
used for adipic and manufacture. At Luling, LA, Monsanto hydrogenates
phenol to cyclohexanol, the bulk of which is converted to adipic acid at
Pensacola, FA.
5. Nipro, Inc., oxidizes cyclohexane, separates the cyclohexane/cyclohexanol
product, dehydrogenates the cyclohexanol to cyclohexanone, and uses the
4
cyclohexanone products to manufacture caprolactam.
6. Union Carbide hydrogenates phenol and then dehydrogenates the product, followed
by separation for cyclohexanone recovery and cyclohexanol recycle.
7. Du Pont Co. oxidizes purchased cyclohexane; the product is used captively
to produce adipic acid.
-------
II-6
D. REFERENCES*
1. Koon Ling Ring et al., "CEH Marketing Research Report on Adipic Acid,"
pp. 608.5031A-- 608.5031F and 608.5032A--608.5032D in Chemical Economics
Handbook, Stanford Research Institute, Menlo Park, CA (April 1980).
2. "CEH Salient Statistics on Cyclohexanol and Cyclohexanone," pp. 638.7020A—
638.7020D in Chemical Economics Handbook, Stanford Research Institute, Menlo
Park, CA (January 1979).
3. F. L. Piquet, letter dated Feb. 14, 1979, to EPA from Allied Chemical, Hopewell,
VA.
4. W. D. Bruce, IT Enviroscience, Inc., Trip Report for Visit to Nipro, Inc., Augusta^.
Georgia, Apr. 18, 1978 (on file at EPA, ESED, Research Triangle Park, NC).
5. F. D. Bess, letter to EPA from Union Carbide Corp., Taft, LA, May 5, 1978, in
response to EPA request for information on the cyclohexanol/cyclohexanone
process.
*Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of
that paragraph, that reference number is indicated on the material involved.
When the reference appears on a heading, it refers to all the text covered by
that heading.
-------
III-l
III. PROCESS DESCRIPTIONS
A. INTRODUCTION
Two basic processes, cyclohexane oxidation and phenol hydrogenation, are cur-
rently utilized industrially for domestic production of cyclohexanol and cyclo-
hexanone.
Cyclohexane Oxidation Process (reaction not balanced):
H
Catalyst
: =»-
(Cyclohexane)
(Air or
Oxygen)
H OH
(Cyclohexanol)
?
HJr/"\rH.
* M.
H.
H2°
(Cyclohexanone) (Water]
Phenol Hydrogenation Process:'
OH
2H
(Phenol)
(Hydrogen)
Catalyst
H
(Cyclohexanone)
or
(Phenol)
3H,
(Hydrogen)
Catalyst
H ^ JDH
H
(Cyclohexanol)
-------
III-2
In the cyclohexane oxidation process, liquid-phase catalytic oxidation of
cyclohexane yields a mixture (called KA oil) of cyclohexanol and cyclohexanone
(reaction conditions: 145 to 165°C and 0.8 to 1.0 MPa). Among those producers
using the cyclohexane oxidation process the main process variation is the type
of catalyst employed. Variations in the cyclohexane oxidation process are
briefly described below.-
1. Conventional Process
A 1:1 ratio of cyclohexanol to cyclohexanone is generated using a cobalt naph-
thenate or cobalt stearate catalyst. Conversion is kept below 10% to promote
selectivity to cyclohexanol and cyclohexanone rather than to other oxidation
products. Selectivity to the product KA oil (ketone-alcohol) is 65 to 75%, and
from 35 to 25% to by-products.
2. Scientific Design Process
By using a boric acid or metaboric acid catalyst a 12% conversion and a product
selectivity of 90 to 95% KA oil are realized. The ratio of alcohol to ketone
in the product is 9 or 10 to 1. Monsanto is presently the only domestic company
using this process.
3. Stamicarbon Modification
In this process a cobalt octoate catalyst is used, and product selectivity is
90 to 95% KA oil. The process was developed by Dutch State Mines of the Nether-
lands .
About 80% of cyclohexanol/cyclohexanone product capacity is based on the cyclo-
hexane oxidation process. Future increases in capacity will probably favor
2
cyclohexane oxidation because cyclohexane costs less. One disadvantage of
this synthesis route is that by-products, such as monobasic and dibasic car-
boxylic acids and cyclohexyl esters, are formed.
Both cyclohexanol and cyclohexanone are made by the catalytic hydrogenation of
molten phenol. A nickel catalyst favors cyclohexanol production and a palladium
catalyst favors cyclohexanone. Allied Chemical Corp. uses a palladium catalyst
and produces cyclohexanone. Monsanto Co. at Luling, LA, produces cyclohexanol.
Union Carbide Corp. produces cyclohexanone, but uses a dehydrogenation step
2--4
plus recycle of the cyclohexanol.
-------
III-3
Fewer by-products, are produced by the phenol hydrogenation process than by
cyclohexane oxidation; it should be noted, however, that benzene is among the
by-products that are formed. Also, phenol hydrogenation is much more selective
2 5
toward the desired end product, cyclohexanol or cyclohexanone. ' With a
nickel catalyst, selectivity to cyclohexanol is typically 97 to 99%. To minimize
product purification requirements, sufficient reactor residence time is allowed
to permit 99+% conversion of phenol to product. '
B. CYCLOHEXANE OXIDATION PROCESS
Figure III-l is a flow diagram illustrating the model plant* for the manufacture
of cyclohexanol/cyclohexanone by cyclohexane oxidation. The process illustrated
6—— 8
is basically conventional.
Production of cyclohexanol/cyclohexanone begins in the oxidation reactor.
Cyclohexane (Stream 1) is preheated and combined with the catalyst (Stream 2)
before it enters the multistaged tower oxidation reactor. Compressed air is
preheated and fed to the reactor (Stream 3). Cyclohexanol/cyclohexanone (KA
oil) product from the reactor (Stream 4) is collected in a receiver. The
product (Stream 5) then enters a neutralization reactor, in which carboxylic
acid by-products are neutralized. The organic and aqueous phases are then
separated. KA oil product (Stream 6) is sent to a stripping column for removal
of cyclohexane. The aqueous phase (Stream 7), which contains carboxylic acid
salts, is stripped of organics prior to storage, and is either sold or ther-
mally oxidized. Recovered cyclohexane (Stream 8) from the stripping column is
recycled for another pass through the oxidation reactor. KA oil product (Stream 9)
from the bottom of the stripping column enters a KA oil recovery column.
Residue from the KA recovery column (Stream 10) is thermally oxidized. Over-
head product from the column (Stream 11) goes to a phase separator. The oil
phase (Stream 12) is removed and then sent to a reactor for saponification of
cyclohexyl esters with caustic. The caustic is removed from the KA oil product
by a water wash. Crude KA oil product (Stream 13) is removed from the extraction
column. Depending on the end use, the product may be further purified or be
used directly as a reactant in another process, such as nitric acid oxidation
of the KA oil to adipic acid. The aqueous phase (Stream 14) from the extraction
column is recycled to the neutralization reactor.
*See page 1-2 for a discussion of model plants.
-------
uore •. -,TO«»aE OP MATER'"-
HA-,
T«e.
^ I L.P iceueecs
ACiD WATER T
Fig. III-l. Process Flow Diagram for Uncontrolled Model Plant
Producing Cyclohexanol/Cyclohexanone by Cyclohexane Oxidation
-------
III-5
Off-gas vented from the cyclohexane oxidation reactor (Stream 15) is cross-
exchanged with cyclohexane feed for heat recovery. Organics condensing from
the gas stream (Stream 16) enter a phase separator, where the oil phase is
separated and recycled to the reactor. The aqueous phase is combined with
other aqueous waste (acid water storage) and leaves the system (Stream 17).
Nitrogen and other gases (Stream 18) leaving the gas-liquid separator pass
through a scrubber for additional cyclohexane recovery. Part of the high-
pressure off-gas leaving the scrubber is recycled as diluent to compressed air
supplied to the cyclohexane oxidation reactor (Stream 19); the remainder is
vented (vent B). Cyclohexane and absorption solvent (Stream 20) enter a stripper
for removal of cyclohexane and water as overhead product (Stream 21). This
overhead product enters a phase separator for separation of the cyclohexane,
which is then recycled to the system (Stream 22). The aqueous phase leaves the
system as an aqueous waste stream (Stream 23). Bottoms from the stripping
column (Stream 24) are recycled to the scrubber.
Off-gas from the phase separators (Streams 25, 26, and 27) is combined and is
sent to a scrubber for removal of cyclohexane (Stream 28). The low-pressure
off-gas is discharged at vent B. Hydrocarbon solvent, containing cyclohexane,
is sent to a stripper for cyclohexane recovery (Stream 29).
C. PHENOL HYDROGENATION PROCESS
1. Model System
Figure III-2 is a very simplified flow diagram for the manufacture of cyclo-
hexanol/cyclohexanone by the hydrogenation of phenol. It shows the three basic
steps of phenol purification, catalytic hydrogenation, and product distillation
of the crude reaction product. Actual production facilities can differ signifi-
cantly from this diagram, especially in the number and sequence of distillation
units and in additional process steps, such as dehydrogenation. In general,
the emission sources should be similar.
There are three potential emission sources indicated on the diagram: the reactor
off-gas from the hydrogenation step (vent A), the distillation column vacuum
source vents (vents B), and the storage tank vents (vents C). There are two
organic residue streams that could be sources of secondary emissions: the
-------
PHEUOU
Fig. III-2. Process Flow Diagram for Uncontrolled Model Plant
Producing Cyclohexanone/Cyclohexanol by Phenol Hydrogenation
-------
III-7
phenol purification distillation bottoms (Stream D) and the residues from catalyst
recovery (Stream E). Sources of contaminated process waters that could be
sources of secondary emissions are the wastewaters from the distillation vacuum
sources (Streams F).
2. Process Variations
Allied Chemical Corp. uses a paladium catalyst that produces essentially
cyclohexan
variation.
4
cyclohexanone directly, and the simplified diagram most nearly depicts this
Union Carbide Corp. evidently uses a catalyst that gives a mixed cyclohexanol/
cyclohexanone product. The crude product is dehydrogenated to give mainly
cyclohexanone, and the unconverted cyclohexanol is recycled. Phenol feed purifi-
9
cation is not practiced.
Monsanto at Luling, LA, is reported to be producing cyclohexanol.
-------
III-8
D. REFERENCES*
1. Faith, Keyes, and Clark, Industrial Chemicals, 4th ed., Wiley-Interscience,
New York, 1975.
2. Koon Ling Ring, "CEH Marketing Research Report on Adipic Acid," pp. 608.5031 A--
608.5032 F in Chemical Economics Handbook, Stanford Research Institute, Menlo
Park, CA (April 1980).
3. Amin Khalil Rafie, "Cyclohexanol and Cyclohexanone Salient Statistics," pp. 638.7020
—638.7020 D in Chemical Economics Handbook, Stanford Research Institute, Menlo
Park, CA (January 1979).
4. F. L. Piquet, letter to EPA from Allied Chemical, Hopewell, VA, February 14,
1979.
5. R. F. Bradley, "CEH Marketing Research Report on Caprolactam," pp. 625.2031 A—
625.2032 W in Chemical Economics Handbook, Stanford Research Institute, Menlo
Park, CA (July 1977).
6. D. E. Danby and C. R. Campbell, "Adipic Acid," pp. 510—531 in Kirk-Othmer
Encyclopedia of Chemical Technology, 3d ed., vol. 1, M. Grayson et al, editors,
Wiley, New York, 1978.
7. V. D. Luedeke, "Adipic Acid," pp. 129-146 in Encyclopedia of Chemical Processing
and Design, vol. 2, edited by McKetta and Cunningham, Dekker Publishing Co.,
New York, 1971.
8. W. D. Bruce, IT Enviroscience, Inc., Trip Report on Visit to Nipro, Inc., Augusta^
Georgia, April 18, 1978 (data on file at EPA, ESED, Research Triangle Park,
NC).
9. F. D. Bess, Union Carbide Corp., letter to EPA, Taft, LA, May 5, 1978, in
response to EPA request for information on the cyclohexanol/cyclohexanone
process.
*Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of
that paragraph, that reference number is indicated on the material involved.
When the reference appears on a heading, it refers to all the text covered by
that heading.
-------
IV-1
IV. EMISSIONS
Emissions in this report are usually identified in terms of volatile organic
compounds (VOC). VOC are currently considered by the EPA to be those of a
large group of organic chemicals, most of which, when emitted to the atmosphere,
participate in photochemical reactions producing ozone. A relatively small
number of organic chemicals have low or negligible photochemical reactivity.
However, many of these organic chemicals are of concern and may be subject to
regulation by EPA under Section 111 or 112 of the Clean Air Act since there are
associated health or welfare impacts other than those related to ozone formation.
A. CYCLOHEXANE OXIDATION PROCESS
1. Model Plant
The model plant for the cyclohexane oxidation process for manufacture of cyclo-
hexanol/cyclohexanone (Fig. III-1) has a capacity of 100 Gg/yr, based on an
annual operation of 8760 hr.* Although not an actual operating plant, it has a
capacity corresponding to that of the average size process train currently
being employed in cyclohexanol/cyclohexanone manufacture and represents current
technology for manufacture of the product. The cyclohexane oxidation process
is the most widely used process for manufacture of cyclohexanol/cyclohexanone.
Storage tank requirements are discussed under "Storage Emissions." Character-
istics of the model plant that are important in air-dispersion modeling are
given in Table B-l, Appendix B.
2. Sources and Emissions
a. General -- Emission rates and sources for the model plant for manufacture of
cyclohexanol/cyclohexanone by cyclohexane oxidation are summarized in Table IV-1,
Emissions are classified as process, storage and handling, secondary, and fugi-
tive.
*Process downtime is normally expected to range from 5 to 15%. If the hourly
rate remains constant, the annual production and annual VOC emissions will be
correspondingly reduced. Control devices will usually operate on the same
cycle as the process. From the standpoint of cost-effectiveness calculations,
the error introduced by assuming continuous operation is negligible.
-------
IV-2
Table IV-1. Uncontrolled Emissions from Model Plant
Cyclohexane Oxidation Process
S tream
Emission Designation
Source (Fig. III-l)
High-pressure A
scrubber vent
Low-pressure B
scrubber vent
Storage and C
handling
Fugitive D
Secondary E
Total
Emissions
Rate (kg/hr) . . Ratio* (kg/kg)
Total VOC
193
30.1
9.6
19.6
1.04
253
CO Total VOC
486 1.69 X 10~
_3
111 2.64 X 10
8.41 X 10~4
1.72 X 10~3
9.11 X 10"5
597 2.22 X 10~2
CO
4.26 X 10~2
_3
9.72 X 10
5.23 X 10~2
a
-------
IV-3
Data contained in trip reports, in the GCA report, and from responses to EPA's
request for information from companies not visited constitute the basis for
specification of model plant emissions. Process emission factors for the
model plant were calculated by averaging the emission-factor data from actual
operating plants. Secondary VOC emissions, such as those from biological
treatment ponds, were obtained from calculations based on flow rates and composi-
tions of wastewater streams from existing plants. Storage emissions were
estimated with equations given in AP-42. However, breathing losses were divided
by 4 to account for recent evidence indicating that the AP-42 breathing-loss
8 9
equation overestimates emissions. ' Fugitive emission calculations were based
on actual pump-count data and estimates of the number of compressors, values,
and pressure-relief valves handling organic liquids or solutions of organics in
the process. The factors given in Appendix C were applied to the data to
calculate fugitive emissions of VOC.
b. Process Emissions — Process emissions occur from the high-pressure and low-
pressure scrubber vents (Vents A and B, Fig. III-l). Emissions from vents A
and B consist of VOC and carbon monoxide (CO). The VOC emissions from these
vents are mostly cyclohexane, but cyclohexanol, cyclohexanone, and other hydro-
carbons are also present. The largest single VOC emission (193 kg/hr) from the
model plant occurs from the high-pressure scrubber vent, and the second largest
VOC emission (30.1 kg/hr) occurs from the low-pressure scrubber vent.
All cyclohexane oxidation plants are equipped with high- and low-pressure
scrubbers, which are used mainly for economic reasons to recover hydrocarbons
and are an integral part of the process. They are usually about 90% efficient
in removing VOC, but are not regarded in this report as emission control devices.
In refs. 1--6 little information is given concerning the effects of startup,
shutdown, and process upsets. Comments in refs. 3 and 5 indicate that increases
in emission during upset conditions are insignificant.
c. Fugitive Emissions — Process pumps, process valves, and pressure-relief valves
handling VOC are sources of fugitive emissions. The model plant for cyclohexane
oxidation is estimated to have 75 pumps, 1900 process valves, and 55 pressure-
relief valves handling VOC. The pumps and process valves are assumed to be
-------
IV-4
used 50% in light-liquid service and 50% in heavy-liquid service. The relief
valves are assumed to be used 50% in vapor service and 50% in light-liquid
service. The fugitive-emission factors in Appendix C were applied to this
valve and pump count to determine the fugitive emissions given in Table IV-1.
d. Storage and Handling -- Virtually no emissions result from handling cyclohexanol/
cyclohexanone since it is used captively. Model-plant storage emission sources
are shown on the flow diagram (Fig. III-l, Source C). A list of model-plant
storage tanks is given in Table IV-2. Estimates of storage tank sizes, turnovers
per year, and bulk liquid temperature were influenced by the data in refs. 6
and 7. Uncontrolled emissions calculated in Table IV-1 are based on fixed-roof
tanks, half full, and a 11°C diurnal temperature variation.
e. Secondary Emissions — In the cyclohexane oxidation model plant secondary VOC
emissions can occur from aqueous wastewater effluent and from storage and
thermal oxidation of residue resulting from the KA oil recovery still (stream 10,
Fig. III-l). Secondary VOC emissions from the wastewater are estimated to be
1.04 kg/hr. Emissions from the wastewater are composed primarily of cyclohexane.
Secondary emissions caused by storage and thermal oxidation of the residue from
the KA oil recovery column are estimated to be negligible.
B. PHENOL HYDROGENATION PROCESS
1. Model Plant
The model plant for the phenol hydrogenation process for manufacture of cyclo-
hexanol/cyclohexanone has a capacity of 100 Gg/yr, based on an annual operation
of 8760 hr. Although slightly larger than the estimated average size of an
actual process train, the capacity of the model plant was set at 100 Gg/yr to
allow direct comparison of emission rates of cyclohexane oxidation and phenol
hydrogenation model plant processes.
Raw-material and intermediate storage tank capacities were estimated for a
100-Gg/yr model plant. As in the case of the cyclohexane oxidation process,
final product storage is considered to be part of either adipic acid or capro-
lactam manufacture and is not included in this report. Storage tank require-
ments are discussed under "Storage Emissions." Characteristics of the model
-------
Table IV-2. Model Plant Storage Tank Data for Cyclohexane Oxidation Process
Storage
Tank
Feed tank
Intermediate
storage3
Intermediate
storage
Intermediate
storage
Catalyst
makeup
Solvent used
in scrubbers
Contents
Cyclohexane
Cyclohexane
Cyclohexane
Cyclohexanol/cyclohexanone
Catalyst in cyclohexane
Hydrocabon solvent
No. of
Tanks
Required
1
1
1
1
1
1
Tank
Size Turnovers
(m3) Per Year
6243 18.2
35.8 6*>
213 6b
143 6b
35.8 130
35.8 247
Bulk
Temperature
30
80
68
127
24
25
aThe uncontrolled state for this tank is defined as a fixed-roof tank with vapor control by chilled brine
condenser. Condenser efficiency is estimated to be 97%.
bThese tanks operate at approximately constant level and the number of turnovers indicated is an attempt
to account for slight level variations.
-------
IV-6
plant that are important in air dispersion modeling are given in Appendix B,
Table B-2.
2. Sources and Emissions
a. General -- Emission rates and sources for the phenol hydrogenation model plant
are summarized in Table IV-3. Comparison of Table IV-3 with Table IV-1 indi-
cates that the total overall uncontrolled emissions for the phenol hydrogena-
tion process are smaller than those for the cyclohexane oxidation process.
Data obtained from trip reports and from responses to EPA for information
11 12
constitute the basis for specification of model plant emissions. '
b. Process Emissions -- The largest emission from the phenol hydrogenation model
plant is off-gas from the hydrogenation reactor (Vent A, Fig. III-2). Nonmethane
VOC emissions from vent A are composed of cyclohexanol, cyclohexanone, benzene,
and other hydrocarbons. Table IV-4 gives the molar composition of this stream.
The emission rate for VOC from vent A is 17.4 kg/hr, of which 0.91 kg/hr is
benzene. The only other significant vent is the distillation column (Vents B,
Fig. III-3).
The phenol hydrogenation reactor is equipped with a vent condenser and an
entrainment separator, but these devi<
and are not emission control devices.
entrainment separator, but these devices are an integral part of the process
11
The manufacturers using the phenol hydrogenation process employ different
processes with different catalysts. This accounts for the differences in
11 12
emissions in the reactor off-gas as reported by the manufacturers. ' The
12
process emissions in Table IV-3 are based on the largest producer.
Process upsets, startups, and shutdowns do not appreciably increase process air
11 12
emissions from vent A. ' Startup of the phenol columns increases emissions
slightly, because the columns operate under vacuum and the gas in the column
must be removed until normal operating pressure is obtained. The volume of gas
ejected is approximately equal to the internal volume of the distillation
i !2
column.
-------
IV-7
Table IV-3. Benzene and VOC Uncontrolled Emissions from
Model Plant — Phenol Hydrogenation Process
Emissions
Emission
Source
Hydrogenation
reactor vents
Distillation vents
Storage and
handling
Fugitive
Secondary
Total
Stream
„__. ^_ Rate
(Fig. III-2) Benzene
A 0.91
B
C
D
E
0.91
(kg/hr)
Total VOC
17.4
0.67
0.36
5.0
23.4
Ratio (kg/kg)
Benzene Total VOC
-5 -3
7.97 X 10 1.52 X 10
-5
5.87 X 10
3.15 X 10~5
4.38 X 10~4
-5 -3
7.97 X 10 2.05 X 10
Kg of emission per kg of cyclohexanol/cyclohexanone produced.
Considered to be negligible.
Table IV-4. Molar Composition of the Hydrogenation Reactor
Vent — Phenol Hydrogenation Process
Component
Composition (mole %)
Cyclohexanone
Cyclohexanol
Cyclohexane
Phenol
Benzene
Other
Total
0.0916
0.0027
0.0028
0.0006
0.0059
99.90
100.00
See ref 12.
Includes methane and other gases, such as hydrogen, which do
not comprise VOC.
-------
IV-8
c. Fugitive Emissions -- The model plant is estimated to have 75 pumps, 1000 process
valves, and 20 pressure-relief valves handling VOC. The pumps and process
valves are assumed to be in heavy-liquid service and the pressure-relief valves
in vapor service. The fugitive-emission factors shown in Appendix C were
applied to this valve and pump count to determine the fugitive emissions given
in Table IV-3.
d. Storage and Handling Emissions — Comparison of the data in Tables IV-1 and
IV-3 indicates that uncontrolled storage and handling losses from the phenol
hydrogenation model plant are much smaller than the corresponding losses from
the cyclohexane oxidation process. High volatility of cyclohexane, as opposed
to the low volatility of phenol, is a major factor in this case.
Model-plant storage emission sources are shown on the flow diagram (Fig. III-2,
Source B). A list of model-plant storage tanks is given in Table IV-5. Esti-
mates of storage tank sizes, turnovers per year, and bulk liquid temperature
were influenced by the data given in refs. 11 and 12. Uncontrolled storage
emissions for the tanks listed in Table IV-5 are based on fixed-roof tanks,
half full, and a 11°C diurnal temperature variation.
e. Secondary Emissions -- Secondary VOC emissions can result from wastewater
effluent generated during hydrogenation catalyst reactivation and from the
vacuum ejector on the phenol stills. Also, some waste organic tars from the
phenol hydrogenation (resulting from catalyst reactivation) and phenol purifi-
cation operations are disposed of by transfer to a waste treatment system. If
the treatment system is a biological treatment pond, some VOC air emissions may
occur from the pond. Total secondary air emissions from the phenol hydrogenation
model plant are estimated to be negligible.
-------
IV-9
Table IV-5. Model Plant Storage Tank Data
Phenol Hydrogenation Process
Storage
Tank
Feed
Phenol
aging
Phenol
purified
Phenol
recycle
Crude
product
Contents
Phenol
Phenol
Phenol
Phenol
Cyclohexanol/
cyclohexanone
No. of
Tanks
Required
1
1
1
1
1
Tank
Size
(m3)
4654
143
143
143
284
Turnovers
per Year
18.2
6*
6*
6*
6*
Bulk Temperature
(°C)
60
60
60
60
45
*
These tanks operate at approximately constant level, and the number of turnovers indicated
reflects an attempt to account for slight level variations.
-------
IV-10
C. REFERENCES*
1. Celanese Chemical Company, letter dated Apr. 21, 1978, in response to EPA's
request for information on emissions data on cyclohexanol/cyclohexanol and
adipic acid production facilities.
2. Dow Badische Company, letter dated May 12, 1978, in response to EPA's request
for information on emissions data on cyclohexanol/cyclohexanone and caprolactam
producton facilities.
3. D. R. Durocher et al., Screening Study to Determine Need for Standards of
Performance for New Adipic Acid Plants. Final Report, GCA-TR-76-16-G
(July 1976).
4. Du Pont Company, letter dated Apr. 12, 1978, in response to EPA's request for
information on emissions data on cyclohexanol/cyclohexanone and adipic acid
production facilities.
5. W. D. Bruce, IT Enviroscience, Inc., Trip Report for Visit to Monsanto Textiles
Company, Pensacola, FL, Feb. 8, 1978 (on file at EPA, ESED, Research Triangle
Park, NC).
6. W. D. Bruce, IT Enviroscience, Inc., Trip Report for Visit to Nipro, Inc., Augusta,
GA, Apr. 18, 1978 (on file at EPA, ESED, Research Triangle Park, NC).
7. Dow Badische Comany, emissions data in Emissions Inventory Questionnaire sub-
mitted to Texas Air Control Board, Mar. 19, 1976.
8. C. C. Masser, "Storage of Petroleum Liquids," Sect. 4.3 in Supplement No. 7 for
Compilation of Air Pollutant Emission Factors, AP-42, 2d ed., EPA, Research
Triangle Park, NC (April 1977).
9. E. C. Pulaski, TRW, Inc., letter dated May 30, 1979, to R. Burr, EPA.
10. J. J. Cudahy and R. L. Standifer, IT Enviroscience, Inc., Secondary Emissions
(June 1980) (EPA/ESED report, Research Triangle Park, NC).
11. Union Carbide Corporation, letter dated May 5, 1978, in response to EPA's
request for information on emissions data on cyclohexanol/cyclohexanone
production facilities.
12. W. D. Bruce, IT Enviroscience, Inc,, Trip Report for Visit to Allied Chemical
Company, Hopewell, VA, Feb. 21, 1978 (on file at EPA, ESED, Research Triangle
Park, NC).
*Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of
that paragraph, that reference number is indicated on the material involved.
When the reference appears on a heading, it refers to all the text covered by
that heading.
-------
V-l
V. APPLICABLE CONTROL SYSTEMS
A. CYCLOHEXANE OXIDATION PROCESS
1. Process Emissions
Emissions of VOC and CO occur from the high-pressure and low-pressure scrubbers,
as indicated in Table IV-1. Information from manufacturers indicates that in
one case a flare is used as a control device and in another only the high-pres-
sure emission is fed to an existing boiler.
The vent gases from the scrubbers can be thermally oxidized to effectively
control the VOC and CO in them; however, because of the large percentage of
nitrogen and other noncombustible gases normally present, supplemental fuel
must be added to ensure proper combustion. For the model plant the high-
pressure and low-pressure scrubber emissions (Vents A and B, Fig. III-l) are
assumed to be processed in a single thermal oxidizer with a reduction of 99% in
o
VOC and CO emissions when operating at 871°C and 0.75-sec residence time.
The controlled emission values are given in Table V-l.
In addition to a thermal oxidizer the scrubber's vent gas emissions could be
q
controlled equally effectively by feeding them to an existing boiler or by
using a catalytic oxidizer. These control systems could have special attrac-
tions, especially in retrofit situations. In using an existing boiler the
high-pressure vent, which is an order of magnitude larger, has sufficient
pressure to transport the emissions to the use site. The low-pressure vent may
need compression or special controls to be handled. A cyclohexanol/cyclohexanone
and adipic acid plant requires a large amount of steam. A catalytic oxidizer
used on the high-pressure vent could be combined into a system using a power
recovery turbine, thereby resulting in a significant energy savings.
2. Fugitive Emissions
Controls for fugitive emissions from the synthetic organic chemicals manufac-
12
turing industry are discussed in a separate EPA document. Emissions from
pumps, process valves, and pressure-relief devices can be controlled by an
appropriate leak-detection system and with repair and maintenance as needed.
Controlled fugitive emissions were calculated with the appropriate factors
given in Appendix C and are included in Table V-l.
-------
Table V-l. Controlled Emissions from Model Plant Using Cyclohexane Oxidation Process
Vent or
Source
Emission Designation
Source (Fig. III-l)
High-pressure A
scrubber vent
Low-pressure B
scrubber vent
Storage and handling C
Fugitive D
Secondary E
Total
Control
Device Emission
or Reductio
Technique (%)
Thermal 99
oxidizer
Thermal 99
oxidizer
Internal float- 85
ing roof on
cyclohexane tanks
Detection and
repair of
major leaks
None
n Rate
VOC
1.93
0.30
1.80
6.90
1.04
11.97
Emissions
(kg/hr) , Ratio3 (ka/kcr)
CO VOC CO
4.86 1.69 X 10"4 4.26 X 10~4
1.11 2.64 X 10~5 9.72 X 10~5
-4
1.55 X 10
-4
6.04 X 10
-5
9.11 X 10
-3 -4
5.97 1.05 X 10 5.23 X 10
Kg of emission per kg of cyclohexanol/cyclohexanone produced.
-------
V-3
3. Storage and Handling Emissions
Options for control of storage emissions are covered in another EPA report.
Emissions listed in Table V-l are based on the assumption that only the large
cyclohexane storage tank and one of the cyclohexane intermediate tanks (Table IV-2)
are fitted with contact-type internal floating roofs* with secondary seals. This
14
is estimated to reduce fixed-roof-tank emissions by 85%.
4. Secondary Emissions
No control system has been identified for secondary emissions from the model
plant. Control of secondary emissions is discussed in another EPA report.
B. PHENOL HYDROGENATION PROCESS
1. Process Emissions
Emissions of VOC, including benzene, occur from the hydrogenation reactor, as
indicated in Table IV-3. Information from cyclohexanol/cyclohexanone manufac-
turers using the phenol hydrogenation process indicates that the only type of
control technique in use for emissions from these vents consists of burning the
reactor off-gas in a boiler.
For the phenol hydrogenation model plant it is assumed that the hydrogenation
reactor off-gas is used as a fuel gas. When the off-gas is used as a fuel, the
VOC destruction efficiency c<
value is given in Table V-2.
Q
VOC destruction efficiency can be greater than 99.9%. The controlled emission
Control method alternatives that have been suggested are carbon adsorption,
17
refrigeration condensation, and cyrogenic separation. These methods may
have application in specific cases and would have the advantage that contained
hydrogen and organics could be recovered for recycle.
2. Fugitive Emissions
Controls for futitive emissions from the synthetic organic chemical manufacturing
12
industry are discussed in a separate EPA document. Emissions from pumps,
process valves, and pressure relief devices can be controlled by an appropriate
*Consist of internal floating covers or covered floating roofs as defined in API
25-19, 2d ed., 1976 (fixed-roof tanks with internal floating device to reduce
vapor loss).
-------
Table V-2. Controlled Emissions from Model Plant — Phenol Hydrogenation Process
Emission
Source
Hydro genat ion
reactor vent
Distillation vents
Storage and
handling
Fugitive
Secondary
Total
Vent or Control
Source Device Emission
Designation or Reduction
(Fig. III-2) Technique (%)
A Route to plant fuel 99.9
gas supply
B Route to plant fuel 99.9
gas supply
C None
D Detection and re-
pair of major
leaks
E None
Emissions
Rate (kg/hr)
Total VOC Benzene
0.0174 0.0009
0.0007
0.36
3.0
3.38 0.0009
Ratio3 (kg/kg)
Total VOC Benzene
1.52 X 10~6 7.97 X 10~8
5.87 X 10~8
3.15 X 10~5
2.63 X 10~4
2.96 X 10~4 7.97 X 10~8
Kg of emission per kg of cyclohexanol/cyclohexanone produced.
Considered to be negligible.
I
*».
-------
V-5
leak-detection system and with repair and maintenance as needed. Controlled
fugitive emissions were calculated with the appropriate factors given in Appen-
dix C and are included in Table V-2.
3. Storage and Handling Sources
Options for control of storage emissions are covered in another EPA report.
For the model plant no controls for storage are assumed.
4. Secondary Emissions
No control system has been identified for the model plant. Control of secondary
emissions is discussed in another EPA report.
-------
V-6
C. REFERENCES*
1. Celanese Chemical Company, letter dated Apr. 21, 1978, in response to EPA's
request for information on emissions data on cyclohexanol/cyclohexanol and
adipic acid production facilities.
2. Dow Badische Company, letter dated May 12, 1978, in response to EPA's request
for information on emissions data on cyclohexanol/cyclohexanone and caprolactam
production facilities.
3. D. F. Durocher e_t al. , Screening Study to Determine Need for Standards of
Performance for New Adipic Acid Plants. Final Report, GCA-TR-76-16-G (1976).
4. Du Pont Company, letter dated Apr. 12, 1978, in response to EPA's request for
information on emissions data on cyclohexanol/cyclohexanone and adipic acid
production facilities.
5. W. D. Bruce, IT Enviroscience, Inc., Trip Report for Visit to Monsanto Textiles
Company, Pensacola, FL, Feb. 8, 1978 (on file at EPA, ESED, Research Triangle
Park, NC).
6. W. D. Bruce, IT Enviroscience, Inc., Trip Report for Visit to Nipro, Inc., Augusta,
GA, Apr. 18, 1978 (on file at EPA, ESED, Research Triangle Park, NC).
7. W. R. Chalker, E. I. du Pont de Nemours & Company, Inc., letter to EPA with
information on cyclohexanol/cyclohexanone process, March 12, 1979.
8. J. W. Blackburn, IT Enviroscience, Inc., Control Device Evaluation. Thermal
Oxidation, (July 1980) (EPA/ESED report, Research Triangle Park, NC).
9. V. Kalcevic, IT Enviroscience, Inc., Control Devices Evaluation. Flares
and the Use of Emission as Fuels (in preparation for EPA, ESED, Research
Triangle Park, NC) (August 1980).
10. J. A. Key, IT Enviroscience, Inc., Control Device Evaluation. Catalytic
Oxidation (October 1980) (EPA/ESED report, Research Triangle Park, NC).
11. C. J. Schaefer and T. M. Kenesson, Celanese Chemical Co., Inc., letter to the
EPA with information on cyclohexanol/cyclohexanone process, January 12, 1979.
12. D. G. Erikson and V. Kalcevic, IT Enviroscience, Inc., Fugitive Emissions (September
1980) (EPA/ESED report, Research Triangle Park, NC).
13. D. G. Erikson, IT Enviroscience, Inc., Storage and Handling (September 1980)
(EPA/ESED report, Research Triangle Park, NC).
14. W. T. Moody, TRW, Inc., letter dated Aug. 15, 1979, to Dave Beck, EPA.
15. J. J. Cudahy and R. L. Standifer, IT Enviroscience, Icn., Secondary Emissions
(September 1980) (EPA/ESED report, Research Triangle Park, NC).
-------
V-7
16. Union Carbide Corporation, letter dated May 5, 1978, in response to EPA's
request for information on emissions data on cyclohexanol/cyclohexanone pro-
duction facilities.
17. F. L. Piguet, Allied Chemical, letter with information on cyclohexanol/cyclo-
hexanone process dated Feb. 14, 1979, to EPA.
*Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of
that paragraph, that reference number is indicated on the material involved.
When the reference appears on a heading, it refers to all the text covered by
that heading.
-------
VI-1
VI. IMPACT ANALYSIS
A. ENVIRONMENTAL AND ENERGY IMPACTS
1. Cyclohexane Oxidation Process
Table VI-1 shows the environmental impact of reducing the total VOC emissions
by application of the described control systems (Sect. V) to the model plant
described in Sects. Ill and IV. Use of these control devices or techniques
results in the reduction of total VOC emissions by 2.11 Gg/yr for the model
plant and in the controlled emissions from the model plant being reduced to
0.10 Gg/yr.
a. High-Pressure and Low-Pressure Vents -- The use of a thermal oxidizer for the
control of vents A and B (Fig. III-l) reduces the model-plant VOC emissions by
an estimated 1.93 Gg/yr and CO emissions by 5.18 Gg/yr. The use of a thermal
oxidizer requires an estimated 18.2 GJ/hr of supplemental fuel; when heat
recovery by a 400-psig steam waste heat boiler is practiced, there is a net
energy savings of 3.1 GJ/hr.
b. Other Emissions (Storage and Fugitive) -- These sources are controlled in the
model plant by internal-floating-roof storage tanks and by detection and repair
of leaking components for fugitive emissions. Application of these controls
results in a VOC emission reduction of 0.18 Gg/yr for the model plant. Neither
of these controls consumes significant energy, nor do they have an adverse
environmental impact.
2. Phenol Hydrogenation Process
Table VI-2 shows the environmental impact of reducing the total VOC emissions
by application of the described control systems (Sect. V) to the model plant
described in Sects. Ill and IV. Use of these control devices or techniques
results in the reduction of total VOC emissions by about 176 Mg/yr for the
model plant and in the controlled emissions from the model plant being 29.6 mg/yr.
a. Hydrogenation Reactor Vent — The use of the emissions from vents A and B
(Fig. III-2) as fuel gases reduces the model-plant VOC emissions by an estimated
152 Mg/yr and benzene emissions by 8 Mg/yr. Using these emissions as fuel will
reduce a plant's other source energy demands.
-------
Table VI-1. Environmental Impact of Controlled Cyclohexane Oxidation Model Plant
Emission
Source
High-pressure scrubber
vent
Low-pressure scrubber
vent
Storage and handling
Fugitive
Secondary
Total
Stream or
Vent
Designation
(Fig. III-l)
A
B
C
D
E
Control Device Emission
or Technique Reduction (%)
Thermal oxidizer 99
Thermal oxidizer 99
Internal floating roof 85
on cyclohexane tanks
Detection and repair
of major leaks
None
Emission Reduction (Gg/yr)
VOC CO
1.67 4.22
0.26 0.96
0.07
0.11
<
LJ
I
2.11 5.18
-------
Table VI-2. Environmental Impact of Controlled Phenol Hydrogenation Model Plant
Emission
Source
Hydro gen at ion reactor
vent
Distillation vents
Storage and handling
Fugitive
Secondary
Total
Stream or
Vent
Designation
(Fig. III-2)
A
B
C
D
E
Control Device Emission
or Technique Reduction (%)
Use as fuel gas 99. 9
Use as fuel gas 99.9
None
Detection and repair of
major leaks
None
Emission Reduction
(Mg/yr)
Total VOC Benzene
152.3
5.8
17.5
175.6
8.0
8.0
I
U)
-------
VI-4
b. Other Emissions (Fugitive) -- These sources are controlled in the model plant
by detection and repair of leaking components. This results in a VOC emission
reduction of 17.5 Mg/yr. This control does not consume significant energy nor
have an adverse environmental impact.
B. CONTROL COST IMPACT
This section gives estimated costs and cost-effectiveness data for control of
VOC emissions resulting from the production of cyclohexanol/cyclohexanone.
Details of the model plants (Figs. III-l and III-2) are given in Sects. Ill and
IV. Cost-estimate sample calculations are included in Appendix D.
Capital cost estimates represent the total investment required to purchase and
install all equipment and material required to provide a complete emission
control system performing as defined for a new plant at a typical location.
These estimates do not include the cost of cyclohexanol/cyclohexanone production
lost during installation or startup, research and development, or land acquisition.
Bases for the annual cost estimates for the control alternatives include utilities,
operating labor, maintenance supplies and labor, recovery credits, capital
charges, and miscellaneous recurring costs such as taxes, insurance, and adminis-
trative overhead. The cost factors used are itemized in Table VI-3.
1. Cyclohexane Oxidation Process
a. Absorber Vents -- The estimated installed capital cost of a thermal oxidizer
designed to reduce by 99% the VOC emissions from the high-pressure and low-pressure
adsorber vents is $500,000 when heat recovery is not used or is $800,000 when
the heat is recovered by steam being generated in a waste-heat boiler. These
costs are based on a thermal oxidizer that is designed for a residence time of
0.75 sec at 871°C, is completely installed, and is equipped to burn supplemental
natural-gas fuel. See Appendix D for the cost-estimate sample calculations for
a thermal oxidizer, based on a complete installation as described in the control
device evaluation report on thermal oxidizers.
-------
VI-5
Table VI-3. Annual Cost Parameters
Operating factor
Operating labor
Fixed costs
Maintenance labor plus
materials, 6%
b
Capital recovery, 18%
Taxes, insurances,
administration charges, 5%
Utilities
Electric power
Natural gas
Heat recovery credits
(equivalent to natural gas)
8760 hr/yr
$15/man-hr
29% of installed capital cost
S8.33/GJ ($0.03/kWh)
§1.90/GJ ($2.00/thousand ft3
million Btu)
$l.90/GJ ($2.00/million Btu)
or
aProcess downtime is normally expected to range from 5 to 15%. If the hourly
rate remains constant, the annual production and annual VOC emissions will be
correspondingly reduced. Control devices will usually operate on the same
cycle as the process. From the standpoint of cost-effectiveness calculations,
the error introduced by assuming continuous operation is negligible.
Based on 10-year life and 12% interest.
-------
VI-6
The costs are also based on the thermal oxidizer system being located within
the battery limits of the production facility. If for safety or for a particular
retrofit situation the oxidizer needs to be located some distance from the
facility, considerable capital cost could be associated with the required
additional piping.
The vent gas rates vary directly with the production rate; therefore a plant
with half the capacity of the model plant will require a thermal oxidizer with
half the capacity of one for the model plant. Figure VI-1 was plotted to show
the variation of installed capital cost of a thermal oxidizer, both without
heat recovery and with a waste-heat boiler, versus plant capacity.
To determine the cost effectiveness of the thermal oxidizer, estimates were
made of the gross annual operating cost for both cases: without heat recovery
and with a waste-heat boiler. For the waste-heat boiler case the recovery credit
and the net annual operating cost were also calculated; see Table VI-4. The
variation of net annual costs with plant capacity for both cases is shown in
Fig. VI-2. The cost effectiveness for each case for controlling VOC was calcu-
lated from the net annual cost and its emission reduction; see Table VI-3. The
variation in cost effectiveness with plant capacity is shown in Fig. VI-3.
b. Storage -- The system for controlling storage emissions is internal-floating-roof
tanks for cyclohexane storage. Another EPA report covers storage and handling
emissions and their applicable controls for all the synthetic organic chemicals
manufacturing industry.
c. Fugitive Sources — A control system for fugitive sources is defined in Appendix C.
Another EPA report covers fugitive emissions and their applicable controls for
4
all the synthetic organic chemicals manufacturing industry.
d. Secondary Sources — No control system has been identified for controlling the
secondary emissions from the model plant.
-------
VI-7
40
1OO
Plant Capacity (Gg/yr)
400
(a) Thermal oxldizer without heat recovery
(b) Thermal oxidizer with waste-heat boiler
Fig. Vl-1. Installed Capital Cost vs Plant Capacity for
Emission Control (Thermal Oxidation)
-------
Table VI-4. Emission Control Cost Estimates for Cyclohexanol/Cyclohexanone Model Plants
Control
Thermal oxidizer
With heat recovery
Without heat recovery
Total
Installed
Capital
Cost
$800,000
500,000
Annual Operating
Gross
Annual
$588,000
483,000
Recovery
Credits
$334,000
0
Costs
(A)
Net
Annual
$254,000
483,000
Emission
CO
(Mg/yr)
.5180
5180
(B)
Reduction
VOC
(Mg/yr) (%)
1930 99
1930 99
(Oa
Cost Effectiveness
VOC
(per Mg)
$132
250
(C) = (A) v (B).
I
en
-------
VI-9
o
o
o
(0
O
o
75
a
c
c
0
z
1000
800
600
400
200
o
x»
o
5
I i I I i
,
40
100
Plant Capacity (Gg/yr)
400
(a) Thermal oxidizer without heat recovery
(b) Thermal oxidizer with waste-heat boiler
Fig. VI-2. Net Annual Cost vs Plant Capacity for
Emission Control (Thermal Oxidation)
-------
VI-10
o>
2
01
o
c
Q)
u
UJ
in
O
U
400
300
200
100
e
2
a.
i i i i i
40
100
Plant Capacity (Gg/yr)
400
fa) Thermal oxidizer without heat recovery
(bi Thermal oxidizer with waste-heat boiler
Fig. VI-3. Cost Effectiveness vs Plant Capacity for
Emission Control (Thermal Oxidation)
-------
VI-11
2. Phenol Hydrogenation Process
a. Hydrogenation Reactor Vent -- Off-gas for the phenol hydrogenation (stream A,
Fig. III-3) is sent to the plant fuel-gas system for combustion in a boiler.
The costs associated with the use of VOC emissions as fuel are very site specific.
If there is a satisfactory use for the fuel with a reasonable cost system,
there can be a cost-effectiveness savings. Specific cost-effectiveness calcula-
tions were not performed.
b. Storage -- Storage emissions from the phenol hydrogenation model plant are small
and no control system has been identified.
c. Fugitive Sources -- A control system for fugitive sources is defined in Appendix C.
Another EPA report covers fugitive emissions and their applicable controls for
A
all the synthetic organic chemicals manufacturing industry.
d. Secondary Sources — No control system has been identified for the secondary
emissions from the model plant.
-------
VI-12
C. REFERENCES*
1. J. W. Blackburn, IT Enviroscience, Inc., Control Device Evaluation. Thermal
Oxidation (July 1980) (EPA/ESED report, Research Triangle Park, NC).
2. W. R. Chalker, E. I. du Pont de Nemours & Co., Inc., letter to EPA with infor-
mation on cyclohexanol/cyclohexanone process, March 12, 1979.
3. D. G. Erikson, IT Enviroscience, Inc., Storage and Handling (September 1980)
(EPA/ESED report, Research Triangle Park, NC).
4. D. G. Erikson and V. Kalcevic, IT Enviroscience, Inc., Fugitive Emissions
(September 1980) (EPA/ESED report. Research Triangle Park, NC).
*Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of
that paragraph, that reference number is indicated on the material involved.
When the reference appears on a heading, it refers to all the text covered by
that heading.
-------
VII-1
VII. SUMMARY
Cyclohexanol/cyclohexanone are currently manufacturer by two basic processes,
cyclohexane oxidation or phenol hydrogenation. The cyclohexane oxidation
process produces a mixture, called KA oil, of cyclohexanol and cyclohexanone
that can be used directly as a raw-material feed to an adipic acid plant. The
phenol hydrogenation process can produce primarily either cyclohexanol or
cyclohexanone or a mixture of the two, depending on the type of catalyst used.
Each of the three producers using this process practices a different version.
About 80% of the listed 1979 plant capacity is based on the cyclohexane oxidation
process and 20% on the phenol hydrogenation process.
The annual growth rate of cyclohexanol/cyclohexanone manufacture is estimated
to be 1—2% through 1984. The 1979 listed capacity was 1147 Gg/yr and the
estimated production was 1065 Gg.
Emission sources and uncontrolled and controlled emission rates for the cyclo-
hexane oxidation process and the phenol hydrogenation process model plants are
given in Tables VII-1 and VII-2. The current VOC emissions projected for the
industry based on the estimated degree of control existing in 1980 is 11,700 Mg/yr.
These emission estimates are based on engineering judgement and data from
individual producers, state emission control agencies, and the open literature.
The following individual estimated projections were made:
Source 1980 VOC Emission (Mg/yr)
Process 10,900
Storage 140
Fugitive 580
Secondary 80
11,700
The predominant emission points for the cyclohexane process are the high-pressure
and low-pressure scrubber vents. The gases from these vents can be controlled
by a thermal oxidizer, which will reduce both the VOC emissions and the CO
emissions by 99%. The installed cost of a thermal oxidizer for the model plant
is $500,000 without heat recovery and $800,000 with heat recovery by use of a
-------
VII-2
Table VII-1. Emission Summary for Cyclohexane Oxidation Model Plant
Emission
Source
High-pressure
scrubber vent
Low-pressure
scrubber vent
Storage and
handling
Fugitive
Secondary
Stream or
Vent
Designation
(Fig. III-l)
A
B
C
D
E
Emission Rate
Uncontrolled
VOC CO
193 486
30.1 111
9.6
19.6
1.04
(kg/hr)
'Controlled
VOC
1.93
0.30
1.80
6.90
1.04
CO
4.86
1.11
Total 253 597 11.97 5.97
-------
VII-3
Table VII-2. Emission Summary for Phenol Hydrogenation Model Plant
Emission
Source
Hydrogenation
reactor vent
Distillation vents
Storage and
handling
Fugitive
Secondary
Total
Stream or
Vent
Designation
(Fig. III-2)
A
B
C
D
E
Emission Rate (kg/hr)
Uncontro lied
Benzene Total VOC
0.91 17.4
0.67
0.36
5.0
0.91 23.4
Controlled
Benzene Total VOC
0.0009 0.0174
0.0007
0.36
3.0
0.0009 3.38
-------
VII-4
waste-heat boiler. Supplemental fuel is required. The cost effectiveness for
VOC controlled is $250/Mg without heat recovery and is $132/mg with a waste
heat boiler.
The predominant emission point for the phenol hydrogenation process is the
hydrogenation reactor vent. This vent can be controlled by the emission being
used as a fuel gas, which will reduce both the total VOC and the small quantity
of contained benzene emissions by 99.9% or greater.
-------
A-l
APPENDIX A
PHYSICAL PROPERTIES OF CYCLOHEXANOL, CYCLOHEXANONE,
BENZENE, CYCLOHEXANE, AND PHENOL
Table A-l. Physical Properties of Cyclohexanol*
Synonyms
Molecular formula
Molecular weight
Physical state
Vapor pressure
Vapor specific gravity
Boiling point
Melting Point
Liquid specific gravity
Water solubility
He xahydropheno1
C6H12°
100.16
Colorless
crystalline solid
467 Pa at 34°C
3.46 (air = 1)
161.1°C
25.5°C
0.9493 at 20°C/4°C
3.6 wt % at 20°C
J. Dorigan, B. Fuller, and R. Duffy, p. AI-320 in Scoring of Organic Air
Pollutants. Chemistry, Production and Toxicity of Selected Synthetic Organic
Chemicals (Chemicals A—C), Rev. 1, Appendix I, MTR-7248, MITRE Corp., McLean,
VA (September 1976).
Table A-2. Physical Properties of Cyclohexanone*
Synonyms
Molecular formula
Molecular weight
Physical state
Vapor pressure
Vapor specific gravity
Boiling point
Melting point
Liquid specific gravity
Water solubility
Pimelic ketone, anone, sextone,
ketohexamethylene
C,-H, O
6 10
98.14
Colorless liquid
600 Pa at 25°C
No data
155.6°C
-45°C
0.9478 at 20°C/4°C
50 g/1 at 30°C
J. Dorigan, B. Fuller, and R. Duffy, p. AI-322 in Scoring of Organic Air
Pollutants. Chemistry, Production and Toxicity of Selected Synthetic
Organic Chemicals (A—C), Rev. 1, Appendix I, MTR-7248, MITRE Corp.,
McLean, VA (September 1976).
-------
A-2
Table A-3. Physical Properties of Benzene'
Synonyms
Molecular formula
Molecular weight
Physical state
Vapor pressure
Vapor specific gravity
Boiling point
Melting point
Liquid specific gravity
Water solubility
O ct ano1/wate r
partition coefficient
Benzol, coal naphtha, phenylhydride
C6H6
78.11
Liquid
95.9 mm Hg at 25°C
2.77
80.1°C at 760 mm Hg
5.5°C
0.8787 at 20°C/4°C
1.79 g/liter
2.28
From: J. Dorigan, B. Fuller, and R. Duffy, "Benzene," p AI-102 in Chemistry,
Production and Toxicity of Selected Synthetic Organic Chemicals (Chemicals A-C),
MTR-7248, Rev 1, Appendix I, MITRE Corp., McLean, VA (September 1976).
-------
A-3
Table A-4. Physical Properties of Cyclohexane
Synonyms
Molecular formula
Molecular weight
Physical state
Vapor pressure
Vapor specific gravity
Boiling point
Melting point
Liquid specific gravity
Water solubility
Hexahydrobenzene, hexanaphthene, hexaraethylene
C6H12
84.16
Liquid
98.14 mm Hg at 25°C (96.97 mm Hg at 25°C)
2.90
80.7°C at 760 mm Hg
6.3°C
0.77855 at 20°C/4°C
<1 g/liter
aFrom: J. Dorigan, B. Fuller, and R. Duffy, "Cyclohexane," p AI-318 in
Chemistry, Production and Toxicity of Selected Synthetic Organic Chemicals
(Chemicals A-C) , MTR-7248, Rev 1, Appendix I, MITRE Corp., McLean, VA
{September 1976).
-------
A-4
Table A-5. Properties of Phenol
Synonyms
Molecular formula
Molecular weight
Physical state
Vapor pressure
Vapor specific gravity
Boiling point
Melting point
Density
Water solubility
Carbolic acid, phenic acid, phenylic acid,
oxybenzene, phenyl hydroxide, hydroxy-
benzene
C6H6°
94.11
Solid
0.530 mm Hg at 25°C
3.24
181.9°C at 760 mm
42.5 to 43°C
1.0576 g/ml at 20°C/4°C
Soluble.
aFrom: J. Dorigan et ail., "Phenol," p. AIV-32 in Scoring of Organic Air Pol-
lutants. Chemistry,_Prod_uctj.on and Toxicity of SeJLected Organic Chemicals
(Chemicals O-2), MTR-724B, Rev 1, Appendix IV, Mitre Corp., McLean, VA
, (September 1976) .
-------
B-l
APPENDIX B
AIR-DISPERSION PARAMETERS
Table B-l. Air-Dispersion Parameters for Cyclohexane Oxidation
Model Plant (Capacity, 100 Gg/yr), Controlled and Uncontrolled
Emission
Source
Emission
Rate
(g/sec)
Tank
Height
-------
Table B-2. Air-Dispersion Parameters for Phenol Hydrogenation
Model Plant (Capacity, 100 Gg/yr), Controlled and Uncontrolled
Source
Process emissions
Hydrogenation reactor vent
Distillation .vents
a
Storage
Phenol feed
Phenol in-process <3 tanks)
Cyclohexanol
cyclohexanone crude
product
Fugitive
process0
Fugitive
Emission Hate (a/secl Tank Tank Stack Stack Discharge Flow Discharge
EHUSSlOn Rate ig/SeC) Iliiaht nlanuitjiir Ilaiaht nl 1 i T I t i •
jieigii& LUeunecef neagnc uiame^eir Teinpeiracure Rate Velocity
Total VOC Benzene
-------
C-l
APPENDIX C
FUGITIVE-EMISSION FACTORS*
The Environmental Protection Agency recently completed an extensive testing
program that resulted in updated fugitive-emission factors for petroleum re-
fineries. Other preliminary test results suggest that fugitive emissions from
sources in chemical plants are comparable to fugitive emissions from correspond-
ing sources in petroleum refineries. Therefore the emission factors established
for refineries are used in this report to estimate fugitive emissions from
organic chemical manufacture. These factors are presented below.
Source
Uncontrolled
Emission Factor
(kg/hr)
Controlled
Emission Factor'
(kg/hr)
Pump seals ,
Light-liquid service
Heavy-liquid service
Pipeline valves
Gas/vapor service
Light-liquid service
Heavy-liquid service
Safety/relief valves
Gas/vapor service
Light-liquid service
Heavy-liquid service
Compressor seals
Flanges
Drains
0.12
0.02
0.021
0.010
0.0003
0.16
0.006
0.009
0.44
0.00026
0.032
0.03
0.02
0.002
0.003
0.00'03
0.061
0.006
0.009
0.11
0.00026
0.019
aBased on monthly inspection of selected equipment; no inspection of
heavy-liquid equipment, flanges, or light-liquid relief valves;
10,000 ppmv VOC concentration at source defines a leak; and 15 days
allowed for correction of leaks.
Light liquid means any liquid more volatile than kerosene.
*Radian Corp., Emission Factors and Frequency of Leak Occurrence for Fittings
in Refinery Process Units, EPA 600/2-79-044 (February 1979).
-------
D-l
APPENDIX D
COST-ESTIMATE SAMPLE CALCULATIONS
This appendix contains sample calculations showing how the costs presented in
this report were estimated.
The accuracy of an estimate is a function of the degree of data available when
the estimate was made. Figure D-l illustrates this relationship. The contingency
allowance indicated is included in the estimated costs to cover the undefined
scope of the project.
Capital costs given in this report are based on a screening study, as indicated
by Fig. D-l, based on general design criteria, block flowsheets, approximate
material balances, and data on general equipment requirements. These costs
have an accuracy range of +30% to -23%, depending on the reliability of the
data, and provide an acceptable basis to determine the most cost-effective
alternative within the limits of accuracy indicated.
A. THERMAL OXIDIZER CONTROLLING EMISSIONS FROM CYCLOHEXANE OXIDATION MODEL PLANT
HIGH-PRESSURE AND LOW-PRESSURE SCRUBBER VENTS
This example is based on the total combined emissions from the high-pressure
and low-pressure scrubbers being 30,000 Ib/hr (arrived at by averaging data
from two producers) and the VOC and CO quantities listed in Table IV-1. The
following data are based on the assumption that the VOC is equivalent to cyclo-
hexane and that the inert gas is essentially nitrogen:
Emission Rate Molecular Emission Rate
Material (Ib/hr) Weight (Ib-moles/hr)
VOC (193 + 30.1) X 2.2 490 84 5.8
CO (486 + 111) X 2.2 1,310 28 46.8
Inert gases (by difference) 28,200 29 972.4
Total 30,000 1025
With the molar volume of 1 Ib-mole of gas at 0°C and 1 atm being 359 ft3, then
the emission flow = 1025 X 359 X 1/60 = 6133 scfm.
-------
1MFORMATIOM USED BY ESTIMATOR
ESTIMATE, TYPE
EST/MATED COST
^"WITH
5CRE.EWIUG,
(PRtUM. 6Wq. JG). DESI^W^
•
•
•
•
\
>
^
\\
\
\
\\
O i 2 3 4 ' -60 -4o -20 O 20 40
A.PPROX. CO
-------
D-3
Using the lower heating value of cyclohexane as 18,500 Btu/lb and the heating
value of CO as 4,343 Btu/lb, the heating value of emission gases = (490 X
18,500 + 1310 X 4343) X 1/60 X 1/6133 = 40 Btu/scf.
The oxidizer operating conditions are 1600°F at 0.75-sec residence time (Sect. V-A-1
of this report) with 400-psig steam waste-heat boilers. The costs for this system
were obtained from Appendix B of the Control Device Evaluation. Thermal Oxidation
report. On pages B-20 and B-22 are costs for thermal oxidizers operating at
1600°F and 0.75-sec residence time with feed gas heat contents of 20 and 50 Btu/scf,
respectively. Since the costs given in the report do not coincide with the
emission conditions of 6133 scfm and 40 Btu/scf, it was necessary to interpolate
between the values given by plotting the appropriate values for the 400-psig
steam waste-heat boiler case. The following values were obtained:
Total installed capital cost $800,000
Fixed costs 232,000
Utilities costs 320,000
Manpower costs 36,000
Gross annual operating cost $588,000
Credit for steam (334,000)
Net annual operating cost $254,000
From Table VI-1 of this report the reduction of VOC emissions from the high-
pressure and low-pressure scrubbers is 1.93 Gg/yr (1930 Mg/yr); the cost
effectiveness is then 254,000 X = 132 $/Mg of VOC.
1J. W. Blackburn, IT Enviroscience, Inc., Control Device Evaluation. Thermal
Oxidation (July 1980) (EPA/ESED report, Research Triangle Park, NC).
-------
E-l
APPENDIX E
EXISTING PLANT CONSIDERATIONS
A. PROCESS CONTROL DEVICES
Reported controls used by industry are the following:
1. Du Pont Co. at one of their two cyclohexane oxidation process plants has
piped the high-pressure scrubber emission off-gas to a plant boiler for
incineration.
2. Nipro, Co. at their Augusta, GA, cyclohexane oxidation process facilities
uses a flare to control the emissions from the absorbers.
3. Union Carbide Corp. at their Taft, LA, phenol hydrogenation process facilities
sends the off-gas after the hydrogenation and dehydrogenation steps to
3
boilers.
B. RETROFITTING CONTROLS
The primary difficulty associated with retrofitting may be in finding space to
fit the control device into the existing plant layout. Because of the costs
associated with this difficulty it may be appreciably more expensive to retrofit
emission control systems in existing plants than to install a control system
during construction of a new plant.
-------
E-2
C. REFERENCES*
W. R. Chalkers, E. I. du Pont de Nemours & Co., Inc., letter to EPA with
information on cyclohexanol/cyclohexanone process, March 12, 1979.
W. D. Bruce, IT Enviroscience, Inc., Trip Report for Visit to Nipro, Inc..
Augusta, GA, Apr. 18, 1978 (on file at EPA, ESED, Research Triangle Park,
NC).
F. D. Bess, Union Carbide Corporation, letter to EPA with information on
cyclohexanol/cyclohexanone process. May 5, 1978.
*Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of
that paragraph, that reference number is indicated on the material involved.
When the reference appears on a heading, it refers to all the text covered by
that heading.
-------
3-i
REPORT 3
CHLOROBENZENES
S. W. Dylewski
IT Enviroscience
9041 Executive Park Drive
Knoxville, Tennessee 37923
Prepared for
Emission Standards and Engineering Division
Office of Air Quality and Standards
ENVIRONMENTAL PROTECTION AGENCY
Research Triangle Park, North Carolina
September 1980
This report contains certain information which has been extracted from the
Chemical Economics Handbook, Stanford Research Institute. Wherever used, it
has been so noted. The proprietary data rights which reside with Stanford
Research Institute must be recognized with any use of this material.
D10D
-------
3-iii
CONTENTS OF REPORT 3
I. ABBREVIATIONS AND CONVERSION FACTORS
II. INDUSTRY DESCRIPTION II-l
A. Introduction II-l
B. Chlorobenzene Usage and Growth II-l
C. Domestic Producers II-l
D. Producing Companies II-l
References II-5
III. PROCESS DESCRIPTION III-l
A. Introduction III-l
B. Model Process for Manufacture of Chlorobenzene Compounds III-l
C. Emission Sources III-7
D. Process Variations III-8
References III-9
IV. EMISSIONS IV-1
A. Model Plant IV-1
B. Sources and Emissions IV-1
References IV-7
V. APPLICABLE CONTROL SYSTEMS V-l
A. Tail-Gas Treatment (Vent A) V-l
B. Atmospheric Distillation Vents V-l
C. Vacuum System Vents (Vent D) V-3
D. E-DCB Crystallization (Vent E) V-4
E. E-DCB Crystal Processing (Vent F) V-4
F. Fugitive Sources (Discharge G) V-4
G. Storage and Handling Sources V-4
H. Secondary Sources (Discharges K and L) V-5
I. Current Emission Control V-5
References v~6
-------
3-v
CONTENTS (Continued)
Paqe
VI. IMPACT ANALYSIS VI-1
A. Control Cost Impact VI-1
B. Environmental and Energy Impacts VI-7
References VI-9
VII. PRODUCT ASSESSMENT VII-1
A. Summary VII-1
References VII-4
APPENDICES OF REPORT 3
A - Physical Properties of Chlorobenzene, o-Dichlorobenzene, A-l
p_-Dichlorobenzene and Benzene
B - Atmospheric Dispersion Parameters B-l
C - Fugitive-Emission Factors C-l
D - Existing Plant Considerations D-l
E - Cost Estimate Procedure for Process Emission Control with E-l
Carbon Adsorption
F - List of EPA Information Sources F-l
-------
3-vii
TABLES OF REPORT 3
Number Page
II-l Monochlorobenzene Usage and Growth II-2
II-2 Dichlorobenzene Usage and Growth 11-2
II-3 Chlorobenzene Capacity 11-3
IV-1 Emissions from Uncontrolled Model Plant IV-3
IV-2 Model Plant Storage IV-6
V-l Emissions from Controlled Model Plant v~2
V-2 Control Devices Used by Industry V-6
VI-1 Cost Factors in Annual Costs VI-2
VI-2 Environmental Impact of Control VI-8
VII-1 Emission Summary Model Plant VII-2
VH-2 Industry Emission Rates: 1978 VII-3
VH-3 Emission Ratios: Model Plant, Industry VII-3
A-l Physical Properties of Chlorobenzene A"1
A-2 Physical Properties of o-Dichlorobenzene A-l
A-3 Physical Properties of ^-Dichlorobenzene A-2
A-4 Physical Properties of Benzene A-2
B-l Atmospheric Dispersion Parameters for Model Plant B-l
D-l Emission Control Devices Currently Used by Domestic Chlorobenzene D-2
Products Industry
D-2 Direct Emissions D~3
-------
3-ix
FIGURES OF REPORT 3
Number Page
II-l Locations of Chlorobenzenes Plants II-4
III-l Reaction Chemistry III-2
III-2 Process Flow Diagram III-3
VI-1 Capital Cost of Emission Control vs Plant Capacity VI-4
VI-2 Annual Cost vs Plant Capacity VI-5
VI-3 Cost Effectiveness vs Plant Capacity VI-6
-------
1-1
ABBREVIATIONS AND CONVERSION FACTORS
EPA policy is to express all measurements used in agency documents in metric
units. Listed below are the International System of Units (SI) abbreviations
and conversion factors for this report.
To Convert From
Pascal (Pa)
Joule (J)
Degree Celsius (°C)
Meter (m)
Cubic meter (m3)
Cubic meter (m3)
Cubic meter (m3)
Cubic meter/second
(m3/s)
Watt (W)
Meter (m)
Pascal (Pa)
Kilogram (kg)
Joule (J)
To
Atmosphere (760 mm Hg)
British thermal unit (Btu)
Degree Fahrenheit (°F)
Feet (ft)
Cubic feet (ft3)
Barrel (oil) (bbl)
Gallon (U.S. liquid) (gal)
Gallon (U.S. liquid)/min
(gpm)
Horsepower (electric) (hp)
Inch (in.)
Pound-force/inch2 (psi)
Pound-mass (Ib)
Watt-hour (Wh)
Multiply By
9.870 X 10"6
9.480 X 10"4
(°C X 9/5) + 32
3.28
3.531 X 101
6.290
2.643 X 102
1.585 X 104
1.340 X 10"3
3.937 X 101
1.450 X 10"4
2.205
2.778 X 10"4
Standard Conditions
68°F = 20°C
1 atmosphere = 101,325 Pascals
PREFIXES
Prefix
T
G
M
k
m
M
Symbol
tera
giga
mega
kilo
milli
micro
Multiplication
Factor
1012
109
106
103
10"3
io"6
Example
1 Tg =
1 Gg =
1 Mg =
1 km =
1 mV =
1 pg =
1 X 10 12 grams
1 X IO9 grams
1 X IO6 grams
1 X IO3 meters
1 X 10~3 volt
1 X IO"6 gram
-------
II-l
II. INDUSTRY DESCRIPTION
A. INTRODUCTION
Production of chlorobenzenes was selected for study because preliminary estimates
indicated that the manufacturing process emits significant quantities of benzene,
a substance that is listed as a hazardous pollutant by the EPA (Federal Register
June 8, 1977). The principal chlorobenzene product is monochlorobenzene;
however, o-dichlorobenzene and p_-dichlorobenzene are co-produced in significant
amounts (see Appendix A for pertinent physical properties).
B. CHLOROBENZENE USAGE AND GROWTH
Table II-l shows monochlorobenzene end products, and the percentage of total
consumption. Table II-2 gives usage for dichlorobenzenes. When this report
was first prepared in 1978 the consumption growth rate projected for chloro-
2-7
benzene varied from reference to reference, but it appeared that a 2% growth
rate through 1982 was a reasonable consensus. It now appears that the growth
rate will be essentially zero through 1982.
The present utilization of production capacity for all chlorobenzene products
is about 50%.1/2
C. DOMESTIC PRODUCERS
In January 1980, there were five domestic producers of chlorobenzenes, listed
in Table II-3 at the plant locations shown in Fig. II-l. All the domestic
capacity is based on the direct chlorination of benzene, for which cell chlorine
(85 to 95% Cl ) is generally used, some producers have used purified (condensed,
distilled, revaporized) chlorine to improve reaction control and to decrease
problems in emission control. Since the chlorobenzene capacity already greatly
2
exceeds the demand, no new plant expansions are anticipated.
p. PRODUCING COMPANIES
\. Dow Chemical USA
Dow, with a mono- and dichlorobenzene capacity of 73 Gg/yr, is the largest
producer of these products. Dow also has unlisted production capacities for
tri- and tetrachlorobenzene. The tetrachlorobenzene is consumed captively in
the manufacture of 2,4,5-TCP herbicide.2
-------
II-2
Table II-l. Monochlorobenzene Usage (1978)*
Percent of
End Use Consumption
Solvents 42
Nitrochlorobenzene (agricultural 29
(products)
DDT, silicones, etc. 8
Diphenyl oxide 11
Rubber intermediates 10
*
See ref 1.
Table II-2. Dichlorobenzenes Usage (1978)
Percent of
End Use . Consumption
o-Dichlorobenzene
3,4-dichloroaniline, etc. 70
TDI process solvent 15
Solvents 8
Dye manufacture 4
Pesticides, etc. 3
p-Dichlorobenzene
Space deodorant 55
Moth control 35
Other 10
aSee ref 1.
Toluene diisocyanate.
-------
II-3
Table II-3. Chlorobenzene Capacity*
Producer
Dow, Midland, MI
Monsanto, Sauget, IL
Montrose, Henderson, NV
PPG, New Martinsville, WV
Standard Chlorine, Delaware City, DE
Total
Monochloro-
benzene
45
68
32
41
68
254
1980 Capacity (Gg)
o-Dichloro-
benzene
14
3
9
£3
49
p-Dichloro-
benzene
14
5
14
2!
67
*See ref.1.
-------
m I 4
1. Dow, Midland, MI
2. ICC, Niagara Falls, NY
3. Monsanto, Sauget, IL
4. Montrose, Henderson, NV
5. PPG, New Martinsville, WV
6. Standard Chlorine, Delaware City, DE
XX-\. Locations of Plants Manufacturing Chlorobenzenes
-------
II-5
2. ICC Industries, Inc.
Solvent Chemical Company, a subsidiary of ICC Industries, Inc., ceased operation
of it's 11-Gg/yr mono- and dichlorobenzene plant at Niagara Falls, New York in
July 1978.8
3. Monsanto Company
Monsanto has a mono- and dichlorobenzene capacity of 76 Gg/yr, making it a
major producer. Much of the monochlorobenzene is used captively in the manu-
facture of nitrochlorobenzenes.
4. Montrose Chemical Corporation
All the monochlorobenzene from the Montrose 32-Gg/yr plant is used captively to
2
produce DDT.
5, PPG Industries, Inc.
Expansion of the PPG 64-Gg/yr mono- and dichlorobenzene facility to 113 Gg/yr
has been considered but may have been postponed.
£f Standard Chlorine Chemical Co., Inc.
Standard Chlorine is a major producer of mono- and dichlorobenzenes. Their
previously listed capacity of 125 Gg/yr was expanded in 1976 but the addi-
o
tional equipment was not started up because of lack of sales growth. Their
listed dichlorobenzene capacity is the largest among the producers. They also
have unlisted production capacities for tri- and tetrachlorobenzene.
-------
II-6
REFERENCES*
1. A. C. Gaessler, "CEH Product Review on Chlorobenzenes," pp 633.5030 C-L in
Chemical Economics Handbook, Stanford Research Institute, Menlo Park, CA (March
1980).
2. R. Bradley (Stanford Research Institute), private conversation with S. W.
Dylewski (IT Enviroscience, Inc.), May 2, 1978.
3. "Chemical Profile on o-Dichlorobenzene," p 9 in Chemical Marketing Reporter,
(Sept. 6, 1976).
4. "Dichlorobenzene Profits Scant; Allied Will Mothball Its Plant," p 13 in
Chemical Marketing Reporter (Nov. 18, 1977).
5. J. H. Ayers (Stanford Research Institute), private conversation with 0. D.
Ivins (IT Enviroscience, Inc.), April 20, 1978.
6. "Chemical Profile on p-Dichlorobenzene," p 9 in Chemical Marketing Reporter
(Mar. 15, 1976).
7. Milton Davis (Standard Chlorine Co.), private conversation with C. A. Peterson,
Jr. (IT Enviroscience, Inc.), July 28, 1977.
8. Letter dated Feb. 5, 1979, to Leslie Evans, EPA, from D. L. Rankin, Dover Chemical
Corp.,
^Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of that
paragraph, that reference number is indicated on the material involved. When
the reference appears on a heading, it refers to all the text covered by that
heading.
-------
III-l
III. PROCESS DESCRIPTION
A. INTRODUCTION
Processes for the manufacture of chlorobenzenes have developed over a long
period of time, with various chemistry and product separation methods being
used. The process currently used by industry is direct chlorination of benzene
in the presence of Fed catalyst to produce monochlorobenzene (MCB), according
to reaction I in Fig. III-l. The monochlorobenzene reacts with the remaining
chlorine to form dichlorobenzenes (DCB), according to reaction 2. Hydrogen
chloride is a by-product in both reactions. Only the two major isomers of
dichlorobenzene, ortho and para, are shown; however, a very small amount of the
meta-isomer is also formed. As chlorination is continued beyond reaction 2,
tri-, tetra-, penta-, and, finally, hexachlorobenzenes are formed. Usually,
trichlorobenzene is the only one of the more highly chlorinated products found
in significant amounts. The degree of chlorination of benzene can be controlled
by the choice of catalyst, temperature, and benzene:chlorine ratio in the reac-
tor feed. However, it is not economically possible to exclude the formation of
dichlorobenzenes. Therefore, a three-product process is operated to produce
monochlorobenzene and o- and p_-dichlorobenzenes simultaneously. Hydrogen
chloride, a by-product of the reaction, is processed under anhydrous conditions
before it is absorbed in water. The reaction and the recovery operations are
3—5
continuous.
MODEL PROCESS FOR MANUFACTURE OF CHLOROBENZENE COMPOUNDS
The model* continuous process for the manufacture of chlorobenzenes by chlorin-
ation of benzene and separation of the gaseous and liquid products to recover
monochlorobenzene and o- and p_-dichlorobenzenes is shown in Fig. III-2,
Sheet 1. Batch crystallization, solid-liquid separation, freezing, crushing,
and screening used in the production of p_-dichlorobenzene are shown on Sheet 2.
Most of the steps for the manufacture of monochlorobenzene are generally known
in the industry; however, the manner in which some of these steps are carried
out varies from company to company and some of the steps, such as the removal of
catalyst and the exclusion of water, are subjects of proprietary information.
Since corrosion rates can be exceedingly high, violation of the anhydrous condi-
tions can be devastating to operations. These steps are represented in
*See page 1-2 for a discussion of model.plants.
-------
Reaction 1:
III-2
Cl
+ Cl_ Catalyst
2
HC1
(Benzene) (Chlorine) (Monochlorobenzene) (Hydrogen Chloride)
Cl
+ Cl Catalyst
(Monochlorobenzene) (Chlorine)
Cl
o-Dichlorobenzene + HC1
(Hydrogen
Chloride)
(p-Dichlorobenzene)
Fig. III-l. Reactions of Chlorine with Benzene and Monochlorobenzene
-------
MCB - K-o
p. DCS - p=
o DCB - o-
TCB - T
MIXED DCS-
=J
I
vE^j'a ;
r>
POOCt^'^C,-
AX
I
£^~
m
i}
••I
J." "1
r: DCS 1
f 'OKI (^J
~"
re PRCC**^
a
ty
lj
,®
i^
i U-^ *t:
T^ "'""'"
iO"EK
*•
Hi ;\ ?.v-^--^
.«! ^j; | •— v ™-™*
.-t=-L i TO wtAVV-tv.0^,
PPOCtSS^&~ *
^
"& ,0
TCB
WOBiiE
Fig. III-2. Process Flow Diagram for Manufacture of Chlorobenzene Products
-------
©»
('bWT O
<&
If
TO VUkGUU
^CiUT iT
^
tt- DCtt - a - DiO*t_O»OBevjZJE»a&
TCO - TRi
tineo DCO- Mix
KD
A
®
L^>
~& f
[_J I r
4®
©
tt-oca
p-oca
^>
It®
=1 f®
p.DCS
VTORAGC
P-CX1B
Fig. III-2. (Continued)
-------
III-5
Fig. III-2 as a block entitled "Heavy-Ends Processing." Since there has been
very little growth in the industry in over 10 years, some of the processing
steps may represent old designs. The o- and p_-dichlorobenzene solvent-grade
products are not known commercially by those names. However, they symbolize the
lower quality products that are or can be used as intermediates or solvents.
1. Chlorination of Benzene
The first step in the continuous process for producing monochlorobenzene and o-
and p_~dichlorobenzenes (Fig- III-2, Sheet 1) is removal of the water by distilla-
tion from as-received benzene. The dried benzene (Stream 1) and dry recycled
benzene (Stream 2) plus the chlorine and the Fed catalysts are fed to the
chlorination reactor. The reaction temperature is controlled between 30 and
50°C. The higher the temperature, the larger the amount of dichlorobenzenes
formed.
Reactions 1 and 2 of Fig. III-l are exothermic; for example, the energy release
7 8
for reaction 1, calculated from heats of formation, is 0.69 MJ/kg of benzene. '
This energy is absorbed partly by cooling and partly by vaporization of some of
the reaction mixture components.
2. Recovery of Monochlorobenzene (MCB)
The gas stream from the chlorination reactor (Stream 3), which contains pre-
dominantly HCl, unreacted chlorine, inert gases from the chlorine feed, and
benzene and other VOC, flows to the organic absorber, where most of the benzene
and other VOC are removed. The overhead gas stream, which is low in VOC (Stream 4),
then goes to HCl absorption. (See "Recovery of By-Product HCl" below.) The
bottoms stream, containing some HCl, is sent to the HCl stripper.
The crude liquid reaction product (Stream 5) is sent to the crude chlorobenzene
distillation step. The overhead from the distillation (Stream 6) contains most
of the chlorobenzenes, most of the unreacted benzene, and small amounts of HCl.
The bottoms stream is sent to heavy-ends processing, where the catalyst is
neutralized and some remaining products and higher chlorinated by-products are
recovered and dried before being returned to the process (Stream 7). The waste
catalyst is discharged for land-fill disposal (Stream L), and the aqueous phase
(Stream K) is discharged to wastewater treatment. The HCl that is contained in
-------
III-6
the overheads from crude chlorobenzene distillation (Stream 6) and in the bot-
toms from organic absorber is recovered in the HCl stripper. The bottoms stream
(Stream 8) is then sent to benzene recovery. Part of the benzene-free stream
(Stream 9) is returned to the organic absorber, where it is chilled and used to
absorb the benzene and other VOC from the HCl stream. The remainder (Stream 10)
is sent to MCB distillation, where the MCB product (Stream 11) is taken over-
head and then sent to storage.
3. Recovery of Dichlorobenzenes (DCB)
The residue from MCB distillation contains the DCB isomers. These isomers are
fractionated by distillation, with the lower boiling p_-DCB and some of the o-DCB
going into the overhead along with the m-DCB (Stream 12). The balance of the
o-DCB remains in the bottoms (Stream 13).
The bottoms stream is sent to o-DCB distillation, where the o-DCB product is re-
covered as the overhead (Stream 14) and then sent to storage. The by-product
trichlorobenzene remains in the bottoms (Stream 15) and is also sent to storage.
The crude £-DCB in the overheads from isomer fractionation (Stream 12) is puri-
fied by batch crystallization. Part of the purified j>-DCB (Stream 16) is sent
to liquid storage, and the rest (Stream 17) is frozen, crushed, screened, and
packaged as p_-DCB crystals. During the packaging, ambient air comes in contact
with the product and causes some sublimation loss. To keep the p_-DCB content in
the atmosphere below the hygienic standards of permissible exposure (450 mg/m
g
maximum), the packaging is done in close-fitting hoods. The hoods are ex-
hausted at Vent F.
The mother liquor from the jD-DCB crystallization, containing g- and o-DCB and
essentially all the m-DCB isomer, is sent to DCB solvent-grade fractionation,
where the stream is fractionated into p_-DCB solvent-grade (Stream 18) and o-DCB
solvent-grade (Stream 19), which are sent to storage.
4. Recovery of By-Product HCl
Gases from the organic absorber (Stream 4) go to the HCl absorber, where, by
adiabatic absorption, essentially all the HCl is absorbed. The remaining VOC,
-------
III-7
unreacted chlorine, water vapor released during absorption of HC1, and inert
gases go into the overhead (Stream 20) and to tail-gas treatment, where the
water vapor is condensed in the water scrubber and the acid gases are neutral-
ized in the caustic scrubber. The inert gases and small amounts of unabsorbed
benzene and VOC are vented (Vent A). The organic layer of the liquid stream
from the water scrubber is decanted and recycled to benzene drying (Stream 21),
and the aqueous layer containing the dissolved VOC is recycled to the HCl
absorber (Stream 22). The aqueous salt stream containing some dissolved VOC is
sent to aqueous disposal (Stream K).
EMISSION SOURCES
The primary emission in the production of monochlorobenzene and o- and p_-
dichlorobenzenes results from the tail-gas treatment vent (Vent A), where the
inert gases originally contained in the chlorine feed are vented. This vent
stream also contains some benzene and chlorobenzenes. Normal practice in the
industry is not to provide an emission control device on this vent.
Other emissions include those from benzene drying, heavy-ends processing, ben-
zene recovery, and MCB distillation steps; those from the vacuum system (Vent D)
that services the three vacuum stills; and those from batch p_-dichlorobenzene
crystallization (Vent E). There is also a discharge from the exhaust fan of the
hoods in the p_-dichlorobenzene crystal-processing area (Vent F), where normally
there is no emission control device.
Fugitive emissions occur when leaks develop in valves, pump seals, and major
equipment (Discharge G). In a process that is as potentially corrosive as the
chlorination of benzene a water leak into the system that contains HCl can
rapidly create leaks that will significantly contribute to benzene and VOC
losses to the environment. Losses can also occur through seals and moving and
vibrating parts of the mechanical equipment used in £-dichlorobenzene crystal
processing.
Storage emission sources include benzene storage (Vent H) and various chloro-
benzene product storages (Vent I), including monochlorobenzene, crude p_-dichloro-
benzene, mixed dichlorobenzenes, and finished o- and p_-dichlorobenzenes.
-------
III-8
Handling emissions (Vent J) result from the loading of monochlorobenzene, o- and
g-dichlorobenzene products, and trichlorobenzene.
Secondary emissions can occur when wastewater streams containing dissolved
benzene and other VOC (Stream K) are treated in a waste-treatment plant and when
the catalyst waste (Stream L) is discharged to a controlled land fill.
D. PROCESS VARIATIONS
The chlorine used to produce mono- and dichlorobenzenes in the model plant is
99% pure, which is the purity being achieved by some of the producers. Chlorine
containing up to 15% or more inert gases is used by older plants, which causes
the amount of VOC emission from Vent A to be proportionately higher. If the
water scrubber has no decanter or if inadequate provision is made for phase
separation and the liquid discharges to wastewater treatment, the secondary.
emission potential will be greatly increased.
A caustic scrubber is not always used following the water scrubber. A caustic
scrubber may not significantly affect the VOC in the vent stream; however, it
will absorb the traces of HCl and chlorine and reduce the corrosiveness of the
vent stream to the surrounding steel equipment.
Variations in the removal of the catalyst and the high-boiling organic residues
from the process can affect emission rates. For example, if the material is hot
and under a positive pressure when discharged, the benzene and other VOC emis-
sions will be greater.
Freezing, crushing, screening, packaging, or weighing of p-dichlorobenzene
crystals, if conducted in open equipment, will result in increased sublimation
and loss of the product. Since the vapor concentration in the work area must be
maintained below industrial hygiene standards, a large movement of air will be
required. This would necessitate larger equipment for emission control. The
variations among manufacturers in the processing steps following the recovery of
monochlorobenzene can result in differences in emissions. Some manufacturers
market a mixed dichlorobenzene product, which requires fewer separation steps.
Such a process will have fewer emission points and a lower emission ratio.
-------
III-9
REFERENCES*
1. D. W. F. Hardie, "Chlorinated Benzenes," pp 253-267 Kirk-Othmer Encyclopedia
of Chemical Technology, vol. 5, 2d ed., Interscience, New York, 1964.
2. "Acid Recovery Cuts Costs of Benzene Chlorination," Chemical Processing 41(5),
30 (April 1978). —
3. S.W. Dylewski, IT Enviroscience, Inc., Trip Report on Visit to Natrium Plant of
PPG Industries, Inc., March 23, 1978, (on file at EPA, ESED, Research Triangle
Park, NC).
4. C.A. Peterson, Jr., IT Enviroscience, Inc., Trip Report on Visit Regarding Sauget,
Illinois, Plant of Mosanto Co., April 25, 1978 (on file at EPA, ESED, Research
Triangle Park, NC).
5. Response to EPA request for information on emissions from chlorobenzene manu-
facturers (data on file at EPA, ESED, Research Triangle Park, NC).
6. E. M. Klapproth, "CEH Product Review on Chlorobenzenes," pp 633.5030 C-L in
Chemical Economics Handbook, Stanford Research Institute, Menlo Park, CA
(July 1977).
7. F. D. Rossini et al., Selected Values of Physical and Thermodynamic Properties
of Hydrocarbons and Related Compounds, API Proj. 44, Carnegie Press, Pittsburgh,
1953.
6. D. R. Stull et al., The Chemical Thermodynamics of Organic Compounds, Wiley,
New York, 1969.
g. D. D. Irish, "Halogenated Hydrocarbons: II. Cyclic," pp 1337-1340 in Industrial
Hygiene and Toxicology, edited by F. A. Patty et al., vol II, 2d ed.,
Interscience, New York, 1963.
^Usually, when a reference is located at the end of a paragraph, it refers to the
entire paragraph. If another reference relates to certain portions of that
paragraph, that reference number is indicated on the material involved. When
that reference appears on a heading, it refers to all the text covered by that
heading.
-------
IV-1
IV. EMISSIONS
Emissions in this report are usually identified in terms of volatile organic
compounds (VOC). VOC are currently considered by the EPA to be those of a large
group of organic chemicals, most of which, when emitted to the atmosphere,
participate in photochemical reactions producing ozone. A relatively small
number of organic chemicals have low or negligible photochemical reactivity.
However, many of these organic chemicals are of concern and may be subject to
regulation by EPA under Section 111 or 112 of the Clean Air Act since there are
associated health or welfare impacts other than those related to ozone formation.
A. MODEL PLANT
The process emissions estimated for the model plant are based on the emissions
reported in response to EPA's request for information from selected companies,
on Monsanto and PPG Industries trip reports, on SRI information and an under-
standing of the process chemistry.
The yearly production capacity of the model plant developed for this study is
68 Gg of monochlorobenzene, 12.4 Gg of o-dichlorobenzene, and 15.6 Gg of p_-
dichlorobenzene, for a total of 96 Gg of chlorobenzene products and 36 Gg of
by-product HCl. One half of the £-dichlorobenzene is estimated to be sold as a
crystal product and half as a liquid. The size of the model plant,is typical of
the larger plants. The process, shown in Fig. III-2, is based on a knowledge of
today's chlorobenzene manufacturing industry with the application of current
engineering technology of continuous chlorination, separation in a single-train
process, and resource recovery. From an emission standpoint the model plant is
more representative of the best processes in operation today and of what would
be expected from future construction than representative of today's total indus-
try.
Benzene is chlorinated in the presence of FeCl3 catalyst to form the three major
products and HCl. Separation steps include distillation for benzene drying,
benzene recovery, HCl separation, and product separation; absorption for HCl
recovery; crystallization for g-dichlorobenzene purification; and solids pro-
cessing for production of £-dichlorobenzene crystals.
-------
IV-2
Process emissions estimated for the model uncontrolled plant are discussed in
Sect. IV.B. Estimates of potential fugitive-emission sources based on data
from existing facilities are discussed in Sect. IV.B.6. Raw-material, inter-
mediate, and product storage-tank capacities were estimated, and their emissions
are discussed in Sect. IV.B.7. Estimates of potential emissions from secondary
sources are discussed in Sect.IV.B.8. Characteristics of the model plant that
are important in air-disperson modeling are shown in Appendix B.
B. SOURCES AND EMISSIONS
The uncontrolled emission rates from all sources for the chlorobenzene process
are summarized in Table IV-1, and the process vent locations are shown in
Fig. III-2.
1. Tail-Gas Scrubber
The largest process emission of benzene and total VOC is from tail-gas treatment
(Vent A), which is the reactor vent for inert gas. The emissions from this
vent are directly related to the inert-gas content of the chlorine feed and to
the production rate. The chlorine feed for the model plant is considered to be
99 vol % chlorine, the balance being inert gases. At capacity operation the VOC
emission is estimated to be 6.7 kg/hr, of which 5.7 kg/hr is benzene.
During startup the venting would be somewhat higher because the inert gases
trapped in the system would have to be purged.
2. Atmospheric Distillation Vents
The dissolved gases and the uncondensed VOC from benzene drying, benzene re-
covery, heavy-ends processing, and monochlorobenzene distillation sections are
emitted from these vents (Vents B,C). The quantities are, to a large extent,
related to production. The VOC emission rate is estimated to be 4.3 kg/hr, of
which 4.0 kg/hr is benzene.
3. Vacuum System Vent
The air that leaks into the isomer fractionation, o-dichlorobenzene distilla-
tion, and solvent-grade dichlorobenzene fractionation columns is discharged from
the vacuum system at Vent D. This discharge carries with it uncondensed VOC.
-------
IV-3
Table IV-1. Bonzone and Total VOC Uncontrolled Emissions
from Model Plant for Chlorobenzene Products
Source
Tail -gas scrubber
treatment
Atmospheric distillation
vents
Atmospheric distillation
vent
Vacuum system vent
p-DCB crystallization
vent
p-DCB crystal processing
vent
Fugitive
Benzene storage
Other storage
Handling
Secondary
Stream
Designation
(Fig. III-2)
A
B
C
D
E
F
G
H
IfJ
K,L
Emission Ratio
(g/kg)*
Benzene Total VOC
0.52 0.61
0.32 0.35
0.04 0.04
0.46
0.016
0.084
0.58 1.97
0.41 0.41
0.14
0.02
0.019 0.030
Emission Rate
(kg/hr)
Benzene Total VOC
5.7 6.7
3.5 3.8
0.48 0.48
5.1
0.18
0.92
6.4 21.6
4.5 4.5
1.6
0.23
0.21 0.33
of benzene or total VOC per kg of chlorobenzene products.
-------
IV-4
The emission from this vent is essentially fixed for a given plant design and
does not vary significantly with the throughput rate.
4. p_-Dichlorobenzene Crystallization
The venting from the crude g-dichlorobenzene crystallization section (Vent E)
represents the displacement loss in batch processing. The vent lines of the
equipment in this section are interconnected to prevent further displacement
losses.
5. p_-Dichlorobenzene Crystal Processing
The process emissions from freezing, crushing, and screening operations are not
significant. (Fugitive emissions from the equipment, which can be significant,
are discussed later.) During packaging and weighing of the finished crystals a
sweep of air is drawn across the exposed crystals by an exhaust fan in order to
maintain the £-dichlorobenzene content of the air in the work area to less than
3 4
450 mg/m , which is the accepted standard for this material. The loss during
p_-DCB crystal processing results in a VOC emission of 0.92 kg/hr. The emissions
are directly related to throughput and are seasonal, higher during the summer
and lower during the winter.
6. Fugitive Emissions
Process pumps and valves are potential sources of fugitive emissions (Vent G).
The model plant is estimated to have 102 pumps handling VOC, 22 of which handle
benzene, 68 handle other light liquids and 12 handle heavy liquids. The estimated
number of valves is 792, 45 of which service benzene vapor, 61 service other VOC
vapor, 173 service benzene liquid, 421 service other light liquids and 92 service
heavy liquids. The estimated number of pressure relief valves is 12, 6 of which
wervice benzene, 5 service other light liquids, and 1 services heavy liquids.
The estimated number of flanges is 1848, of which 486 service benzene and 1328
service other VOC. The fugitive emission factors from Appendix C were applied
to this valve, pump, and flange count to determine the fugitive emissions shown
in Table IV-1. Significant fugitive emissions can occur if freezing, crushing,
or screening operations are conducted in open equipment or if there are loose-
fitting flexible joints between pieces of equipment. Data were not obtained to
estimate losses from these sources.
-------
IV-5
7. Storage and Handling Emissions
Emissions result from the storage and handling of benzene, monochlorobenzene, p_-
and o-dichlorobenzenes, trichlorobenzene residues, and in-process streams.
Sources of losses for the model plant are shown in Fig. III-2 (sources H, I,
and J). Storage tank conditions for the model plant are given in Table IV-2.
The storage- and handling-emission calculations in Table IV.-1 were based on
fixed-roof tanks, half full, and a diurnal temperature variation of 11°C and
with the use of the emission equations from AP-42. However, breathing losses
were divided by four to account for recent evidence indicating that the AP-42
breathing loss equation overestimates emissions.
Since benzene freezes at 5.5°C, p_-dichlorobenzene freezes at 53°C, and tri-
chlorobenzene residues freeze at about 20°C, the storage tanks containing these
materials are heated as necessary to maintain their temperatures above the
freezing point.
Emissions from barge loading of monochlorobenzene and tank car loading of o- and
p_-dichlorobenzenes and trichlorobenzene were calculated with the equations from
AP-42. VOC emissions from storage and handling are estimated to be 6.3 kg/hr,
or which 45 kg/hr is benzene.
8. Secondary Emissions
Secondary emissions of benzene and other VOC can result from the handling and
disposal of wastewater streams. For the model plant two potential sources of
secondary emissions from wastewater are indicated in Fig. III-2 (Source K): the
aqueous purge stream from the caustic scrubber in tail-gas treatment and the
aqueous effluent from heavy-ends processing. It is estimated that 50% of the
benzene and 25% of the other VOC contained in these streams are emitted.
The neutralized catalyst waste (Source L) contains high-boiling chlorinated
benzenes and decomposition products. Incineration of this waste may result in
some emission. Disposal in land fill may result in a minor emission. It is
estimated that <2% of the organic is emitted during disposal.
-------
IV-6
Table IV-2. Model Plant Storage Tank Data
Content
Benzene, wet
(bulk)
Benzene, wet
(working)
Benzene, dry
(working)
Monochlorobenzene
(working)
Monochlorobenzene
(bulk)
HC1 solution
(bulk)
P-Dichlorobenzene
(crude)
p-Dichlorobenzene
(working)
p-Dichlorobenzene
~ (bulk)
_o-Dichlorobenzene
(working)
o-Dichlorobenzene
(bulk)
Mixed dichlorobenzene
(working)
C3-DCB solvent grade
(bulk)
p-DCB solvent grade
(bulk)
Trichlorobenzene
(bulk)
Number
2
2
1
2
3
2
2
3
1
2
1
1
2
2
2
Tank Size
(m3)
1900
230
230
230
1900
1900
90
40
460
40
830
40
60
40
40
Turnovers
per Year
19
155
Constant
level
134
11
13
75
98
12
102
10
51
11
9
4
Bulk Liquid
Temperature (°C)
20
20
20
35
16
16
75
75
75
35
16
35
16
75
50
-------
IV-7
REFERENCES*
1. S. W. Dylewski, IT Enviroscience, Inc., Trip Report on Visit to Natrium Plant of
PPG Industries, Inc., March 23, 1978 (on file at EPA, ESED, Research Triangle
Park, NC).
2. C. A. Peterson, Jr., IT Enviroscience, Inc., Trip Report on Visit Regarding Sauget,
Illinois, Plant of Monsanto Co., April 25, 1978 (on file at EPA, ESED, Research
Triangle Park, NC).
3. Response to EPA request for information on emissions from chlorobenzene manu-
facturers (data on file at EPA, ESED, Research Triangle Park, NC).
4. D. D. Irish, "Halogenated Hydrocarbons: II. Cyclic," pp 1337-1340 in
Industrial Hygiene and Toxiciloqy. edited by F. A. Patty et al., vol. II, 2d
ed., Interscience, New York, 1963.
5. C. C. Masser, "Storage of Petroleum Liquids," pp 4.3-12 and 4.3-13 in
Supplement No.7 for Compilation of Air Pollutant Emission Factors, AP-42,
2d ed., EPA, Research Triangle Park, NC (April 1977).
6. Letter dated May 30, 1979, from E. C. Pulaski, TRW, Inc., to Richard Burr, EPA.
*Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of
that paragraph, that reference number is indicated on the material involved.
When the reference appears on a heading, it refers to all the text covered by
that heading.
-------
V-l
V. APPLICABLE CONTROL SYSTEMS
A. TAIL-GAS TREATMENT (VENT A)
The emissions from the caustic scrubber of the tail-gas treatment step can be
effectively reduced by a carbon adsorption (CA) system; however, the VOC con-
centration in Vent A is too high for efficient control by carbon adsorption.
Dilution with a carrier gas, which is provided by the combination of this
stream with other emissions as described later is this report, is necessary for
efficient control. The vent gas from the caustic scrubber, being essentially
saturated with water, is heated to reduce the relative humidity. Since chloro-
benzene products have a slight tendency to hydrolyze under the conditions of
operation, the material of construction must be capable of withstanding the
resulting corrosive environment. The process equipment must have safeguards to
ensure that malfunction or misoperation of the caustic scrubber system does not
allow free chlorine to pass into the CA system where it could cause a reaction
leading to fire or explosion.
After the vent gas from the tail-gas treatment has been conditioned, it passes
through one of the adsorber beds. The VOC-depleted gas is then released to the
atmosphere. When the first bed approaches breakthrough, the conditioned feed
gas is routed to the second bed. At this time regeneration of the first bed by
steam stripping is started. The VOC-laden effluent vapor is then condensed and
decanted. The organic layer is recycled to the process, and the VOC-saturated
aqueous layer is recycled to HC1 absorption. When essentially all the VOC are
stripped from the first bed, air supplied by a purge-gas fan is sent through
the first bed to cool it and make it ready for adsorption. The purge gas
discharged is sent to the adsorbing bed for removal of the small amount of VOC
that it contains. A removal efficiency of 98.7%* is considered to be attainable
under the above conditions, and has been used to project the final emissions
from the model controlled plant (Table V-l). The net annual cost of this
system is shown in Table VI-2 of Sect. VI.
B. ATMOSPHERIC DISTILLATION VENTS
*Based on a carbon adsorption effluent concentration of 12 ppm VOC.
-------
Table V-l. Benzene and Total VOC Controlled Emissions from
Model Plant for Chlorobenzene Products
Stream
Designation
Source (Fig. III-2)
Tail gas treatment
Atmospheric distillation
vents
Atmospheric distillation
vents
Vacuum system vent
p-DCB crystallization vent
p-DCB crystal processing
vent
Fugitive
Benzene storage
Other storage and handling
Secondary
A
B
C
D
E
F
G
H
I,J
K,L
Emission
Control Device Reduction
or Technique (%)
Carbon adsorption 98.7
Carbon adsorption 98.7
Vent to process 100
Carbon adsorption 98.7
None
Carbon adsorption 98.7
Detection and repair
of major leaks 71
Internal floating
roof 85
None
None
Emission Ratio
(g/kg) *
Benzene Total VOC
0,0067 0
0.0042 0
0
0
0
0.16 0
0.06 0
0
0.019 0
.0079
.0045
.006
.016
.001
.58
.06
.16
.030
Emission Rate
(ka/hr)
Benzene Total VOC
0.074 0
0.045 0
0
0
0
1.77 6
0.67 0
1
0.21 0
.087
.049
.066
.18 <
I
.012
.17
.67
.83
.33
C
g of benzene or total VOC per kg of chlorobenzene products.
-------
V-3
1. Benzene Drying (Vent B)
The emission from benzene drying is a small portion of the total emissions from
the process and can be combined with other emissions to be sent to the carbon
adsorber system for control. A removal efficiency of 98.7% is estimated to
be attainable for this stream.
2. Heavy-Ends Processing (Vent B)
The quantity of emissions from the process step is small but it may contain
benzene and may be subject to process upsets. If this stream is brought in
contact with caustic in heavy-ends processing, additional conditioning will not
be necessary before it is sent to the carbon adsorption system.
2. Benzene Recovery (Vent C)
The discharge from benzene recovery is estimated to contain a small quantity of
HCl remaining from the HCl stripping step. This discharge is compatible with
and connected to the gas stream going to the organic scrubber in the model
controlled plant. Any inert gas present is eventually processed through the
carbon adsorption system.
4. MCB Distillation (Vent B)
The discharge from MCB distillation contains a small amount of inert gas and
any benzene remaining from the benzene recovery step. This discharge can be
combined with other streams that feed the carbon adsorption system without the
need for additional conditioning. A removal efficiency of 98.7% is considered
to be attainable for this stream.
c VACUUM SYSTEM VENTS (VENT D)
The discharge from the vacuum system vent contains benzene, other VOC, and air
that has leaked into the three vacuum distillation volumns. This stream is
estimated to be slightly acidic and requires materials of construction capable
of withstanding the environment. This discharge can be combined with other
streams that feed the carbon adsorption system with no additional conditioning.
Again, the removal efficiency for this stream is considered to be 98.7%.
-------
V-4
D. JD-DCB CRYSTALLIZATION (VENT E)
The discharge from the £-DCB crystallization represents displacement loss due
to batch processing and is relatively low in significance; therefore control is
not considered necessary.
E. £-DCB CRYSTAL PROCESSING (VENT F)
The discharge from £-DCB crystal processing results from air movement in the
crystal packaging area that is needed to maintain compliance with industrial
2
hygiene standards regarding p_-DCB concentration in the work environment. This
discharge is incorporated with the other feeds to the carbon adsorption system
and provides the dilution needed for efficient adsorption and control of VOC.
The removal efficiency here is also estimated to be 98.7%.
F. FUGITIVE SOURCES (DISCHARGE G)
Control for fugitive sources is discussed in a separate report covering fugitive
emissions from the synthetic organic chemicals manufacturing industry (SOCMI).
The controlled fugitive emissions shown in Table V-l were calculated with the
factors given in Appendix C. These factors are based on the assumption that
any major leaks will be detected and repaired.
G. STORAGE AND HANDLING SOURCES
1. Benzene Storage (Vent H)
Control of benzene storage emissions from the SOCMI is discussed in a separate
EPA report. A floating roof is commonly used on storage tanks to control
emissions from chemicals in the vapor pressure range of benzene. The internal-
floating-roof tanks* used in the model controlled plant are assumed to reduce
the fixed-roof-tank emissions by 85%.
2. Other Storage and Handling (Vents I and J)
The storage of nonbenzene VOC results in a calculated average emission of
2.97 kg/hr from 23 storage tanks. The largest single emissions is 0.42 kg/hr.
This single nonbenzene VOC emission is not considered to be sufficiently large
to warrant control. The calculated losses during loading of products into
barges and tank cars result in an average VOC emission of 0.23 kg/hr.
*Consist of internal floating covers or covered floating roofs as defined in
API 25-19, 2d ed., 1976 (fixed roof tanks with internal floating device to
reduce vapor loss).
-------
V-5
H. SECONDARY SOURCES (DISCHARGES K AND L)
Secondary emissions can result from handling and treating the wastewater
(Discharge K) in typical wastewater treatment facilities. This emission may be
reduced by steam stripping the wastewater and recycling the benzene and other
VOC but no provision has been made in the model controlled plant for this
control method. It is assumed that the periodic disposal of spent carbon will
be by land-fill or equivalent and that VOC will be removed as much as possible
by steaming prior to disposal. The control of secondary emissions is further
discussed in a separate EPA report. No alternate system to incineration or
land-filling operation has been identified for the waste catalyst (Dis-
charge L).
I. CURRENT EMISSION CONTROL
The information available concerning the emissions and emission controls currently
used by existing chlorobenzene producers in the United States is presented in
Appendix D.
-------
V-6
REFERENCES*
1. H. S. Basdekis, IT Enviroscience, Control Device Evaluation - Carbon Adsorption
(in preparation for EPA, ESED, Research Triangle Park, NC.)
2. D. D. Irish, "Halogenated Hydrocarbons: II. Cyclic," pp 1337—1340 in
Industrial Hygiene and Toxicology, edited by F. A. Patty et al., vol. II,
2d ed., Interscience, New York, 1963.
3. Letter dated Aug. 15, 1979, from W. T. Moody, TRW, Inc., to David Beck, EPA.
4. C. C. Masser, "Storage of Petroleum Liquids," pp 4.3-12 and 4.3-13 in
Supplement No. 7 for Compilation of Air Pollutant Emission Factors, AP-42,
2d ed., EPA, Research Triangle Park, NC (April 1977).
*Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of that
paragraph, that reference number is indicated on the material involved. When
the reference appears on a heading, it refers to all the text covered by that
heading.
-------
VI-1
VI. IMPACT ANALYSIS
A. CONTROL COST IMPACT
This section presents estimated costs and cost-effectiveness data for control
of benzene and total VOC process emissions by use of a carbon adsorption emission
control system. Details of the model plant (Fig. III-2) are given in Sects. Ill
and IV. The capital and annual costs presented for the process emission controls
were obtained from the Control Device Evaluation Report for Carbon Adsorption
(CA). The procedure ui
detailed in Appendix E.
(CA). The procedure used to develop the costs for this control system is
Capital cost estimates represent the total investment required to purchase and
install all equipment and material required to provide a complete process
emission control system performing as defined for a new plant at a typical
location. These estimates do not include the cost of production lost during
installation or startup, of research and development, or of land acquisition.
The capital costs do not include the cost of safeguards in the process to
ensure that malfunction or misoperation of the caustic scrubber system does not
allow free chlorine to enter the CA system where it could cause a reaction
leading to fire or explosion. The costs also do not include incorporation of a
device, such as an induced draft fan, that may be desired to protect against
possible backflow into one of the 6 process waste gas vents joined together to
feed the carbon adsorption system. These items are considered highly site
specific.
The bases for the annual cost estimates for the carbon adsorption (CA) control
system includes utilities, operating labor, maintenance supplies and labor,
capital recovery charges, chemical recovery credits, and miscellaneous recurring
costs such as taxes, insurance and administrative overhead. Chemicals recovered
are taken as raw material value. Chlorobenzene recovered is credited with the
benzene and chlorine equivalent. Annual costs are for a 1-year period beginning
in December 1979. Cost and design factors used are itemized in Table VI-1.
Process Emissions A, B, D, and F
These streams vent the discharges from tail-gas treatment, benzene drying,
heavy-ends processing, monochlorobenzene distillation, vacuum system, and p_-DCB
-------
VI-2
Table Vl-1. Cost and Design Factors to Carbon
Adsorption Emission Control System
3 hr loading cycle
Carbon loading
Steam for regeneration
Steam for gas conditioning
Gas velocity
Bed depth
Pressure drop
Carbon
Benzene equivalent recovery credit
Chlorine equivalent recovery credit
CA minimum effluent concentration
Granular activated carbon replacement every 5 yrs
Operating factor
Steam cost
Electric power
Fixed cost
Maintenance labor plus materials, 6%
Capital recovery, 18%
Taxes, insurance, administration charges, 5%
6 kg VOC/100 kg carbon
1 kg/kg of carbon
35.2 kg/hr
30.5 meters /min (100 ft/min)
0.9 meters (3 ft)
5,257 pascals/meter
(6.5 in H20/ft)
4 X 10 mesh BPL carbon
480 kg/m3 (30 lb/ft3)
$0.22/kg ($0.10/lb)
$0.15 /kg ($0.07/lb)
12
($1.17/lb)
8760 hr/yrb
$5.50/Mg ($2.50/M Btu)
$8.33/GJ ($0.03/kWh)
29% installed capital
VOC
a
If it became necessary to replace the carbon every 2 yrs, the annual cost
would increase 1.4% ($l,270/yr).
3
Process downtime is normally expected to range from 5 to 15%. If the hourly
rate remains constant, the annual production and annual VOC emissions will
be correspondingly reduced. Control devices will usually operate on the same
cycle as the process. From the standpoint of cost-effectiveness calculations,
the error introduced by assuming continuous operation is negligible.
"Based on 10-year life and 12% interest.
-------
VI-3
crystal processing. The estimated installed capital cost of a carbon adsorption
system to reduce total VOC emissions, including benzene, from the model chloro-
benzene products plant is $325,000 (see Appendix E). This cost is based on the
installation of a 2-bed, carbon adsorption system, which includes blowers, con-
denser, decanter, interconnecting piping, valving, instrumentation, and all neces-
sary duct work and utility connections for a complete installation. Since the vent
gas rate varies directly with production, a plant twice the size of the model
plant would have twice the emissions from these vents. Figure VI-1 shows the
variation of capital cost of a carbon adsorption system with plant capacity.
To determine the cost effectiveness of controlling process emissions by carbon
adsorption, the costs obtained from the CA report have been adjusted to include
the additional capital and operating expenses for preconditioning the process
waste gas and the need for costly equipment linings because of the corrosiveness
of the waste gas. The results of the cost adjustment and emission reduction cost
effectiveness calculation are as follows:
Annual cost from CA report Fig. IV-4 $80,416
Fixed cost adjustment 36,250
R.M. credit (30,618)
Steam for conditioning 1,686
Net annual cost $87,734
Benzene emission reduction 79.5 Mg/yr
Total VOC emission reduction 143 Mg/yr
Cost effectiveness
Cost per Mg of benzene emission reduced $1104
Cost per Mg of total VOC reduced $613
The details for estimating the costs and cost effectiveness for the control of
emissions from chlorobenzene plants ranging in size from 50% of the model plant
capacity to a capacity 50% greater than the model plant are given in Appendix E.
The results are plotted on Figs. VI-1, VI-2, and VI-3.
-------
VI-4
500
400
8
4-1
•H
04
nj
U
•O
•H O
•H O
w£ 300
c •—
in
r-
tn
4J
(0
rH
O-
0)
T)
O
s
-------
VI-5
200
o
o
o
Ul
o
u
flj
3
c
100 —
40
70 80 90 100
Plant Capacity (Gg/yr)
Fig, Vl-2. Net Annual Cost vs Plant Capacity for
Carbon Adsorption Emission Control System
200
-------
VI-6
2500
2000
cr>
1500
c
QJ
•H
4J
U
4-1
4-1
W
ti 1000
o
U
500
Benzene
.Total VOC
I
I
I
40
50 60 70 80 90 100
Plant Capacity (Gg/yr)
200
Fig. VI-3. Cost Effectiveness vs Plant Capacity for
Carbon Adsorption Emission Control System
-------
VI-7
2. Benzene Storage
Model-plant benzene storage emissions are controlled by the use of internal
floating-roof tanks. Installed capital cost, net annual cost, and cost-
effectiveness data for retrofitting the model plant fixed-roof tanks and for
the incremental costs of new internal-floating-roof tanks (based on the capital
cost of new internal-floating-roof tanks minus capital cost of new fixed- roof
2
tanks) are given in a separate EPA report.
3. Fugitive Sources
Control emission factors for fugitive sources are described in Appendix C. A
separate EPA report will cover fugitive emissions and their applicable controls
for the synthetic organic chemicals manufacturing industry.
4. Secondary Sources
No control system has been defined for secondary emissions from the model
plant. Secondary sources and their control are discussed in a separate EPA
report.
B. ENVIRONMENTAL AND ENERGY IMPACTS
Table VI-2 shows the environmental impact of reducing benzene and total VOC
emissions by the application of the described control systems for the model
plant.
1f Carbon Adsorption System
This system is designed for the model plant to reduce the benzene emissions by
79 Mg/yr and total VOC emissions by 143 Mg/yr. Operation of this unit consumes
steam at the rate of 7122 GJ/yr including conditioning and electric power for
pumps at the rate of 522 GJ/yr.
2. Floating-Roof Storage
The use of floating-roof tanks for control of benzene emission from storage
does not consume energy and has no adverse environmental or energy impact.
-------
Table VI-2. Environmental Impact of Controlled Model Plant
Source
Steam Designation
(Fig. III-2)
Control Device
or Technique
Emission
Reduction (%)
Emission Reduction (Mg/vr)
Benzene Total VOC
Tail-gas treatment
Atmospheric distillation
vents
Atmospheric distillation
vent
Vacuum system vent
p_-DCB crystallization vent
£-DCB crystal processing
vent
Fugitive
C
D
E
F
G
Carbon adsorption 98.7
Carbon adsorption 98.7
Vent to process 100
Carbon adsorption 98.7
None
Carbon adsorption 98.7
Detection and repair of
major leaks 71
49
30
4.1
40
58
33
4.1
44
8.0
130
i
CO
Benzene storage
Other storage and handling
Secondary
K,L
Internal floating roof 85
None
None
33
156
33
310
-------
VI-9
REFERENCES*
1. H. S. Basdekis, IT Enviroscience, Control Device Evaluation - Carbon Adsorption
(in preparation for EPA, ESED, Research Triangle Park, NC).
2. D. G. Erikson, IT Enviroscience, Inc., Storage and Handling Report, (in pre-
paration for EPA, ESED, Research Triangle Park, NC).
3. D. G. Erikson, IT Enviroscience, Inc., Fugitive Emission Report, (in preparation
for EPA, ESED, Research Triangle Park, NC).
4. J. J. Cudahy, R. L. Standifer, IT Enviroscience, Secondary Emissions Report,
EPA, ESED, Research Triangle Park, NC, June 1980.
^Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of
that paragraph, that reference number is indicated on the material involved.
When the reference appears on a heading, it refers to all the text covered by
that heading.
-------
VII-1
VII. SUMMARY
The chlorobenzene products, monochlorobenzene, o-dichlorobenzene, and g-dichloro-
benzene, are produced in the United States by Fed -c;
of benzene, wherein HC1 is generated as a by-product.
benzene, are produced in the United States by FeCl -catalyzed direct chlorination
It is estimated that the present utilization of chlorobenzene products repre-
sents about 50% of the current domestic capacity of 370 Gg/yr. Zero growth is
projected for chlorobenzene products through 1982. No shortage of benzene is
2
expected during this period.
Emission sources and control levels for the model plant are summarized in
Table VII-1. The emissions from the total industry for 1980 are estimated to
be higher than would be projected from the model uncontrolled plant. Table
VII-2 shows the emissions estimated for the total industry for 1980 to be
48 kg/hr for benzene and 140 kg/hr for VOC. ~" The model plant reflects the
processes in operation today with the lowest emission levels and the likely
emission level of future processes. Since some of the older processes in
operation today inherently have higher emissions, the actual industry emissions
are expected to exceed the projection based on the model uncontrolled plant.
The areas of deviation from the model plant are noted on Table VII-3.
There are five major emission sources: tail-gas treatment vent, atmospheric
distillation vents, vacuum system vent, fugitive vents, and storage and handling
vents. The vent from tail-gas treatment, the atmospheric distillation vents,
and the vacuum system vent can be either returned to a compatible process step
or controlled by processing through a carbon adsorption system. The vent from
crystal processing can also be controlled by the above carbon adsorption system.
The resultant emission reduction would be about 98.79% for both benzene and
total VOC. The estimated capital cost of the carbon adsorption installation is
$325,000 and the annual cost, including a $30,618 credit for materials recov-
ered, would be $87,734. The cost effectiveness for benzene would be $1104/Mg,
and for total VOC emission reduction would be $613/Mg.
Benzene storage emissions can be controlled by means of internal-floating-roof
tanks. The emission reduction would be 85%.
-------
VII-2
Table VII-1. Emission Summary Model Plant
Emission
Emission Source
Tail- gas treatment
Atmospheric distillation
vents
Vacuum system vent
p_-DCB crystallization
vent
£-DCB crystal processing
vent
Fugitive
Storage and handling
Secondary
Total
Uncontrolled
Benzene
5.7
4.0
6.4
4.5
0.21
20.81
VOC
6.7
4.3
5.1
0.18
0.92
21.6
6.33
0.33
4575"
Rate (kg/hr)
Controlled
Benzene
0.074
0.045
1.77
0.67
0.21
2.77
VOC
0.087
0.049
0.066
0.18
0.012
6.77
2.5
0.33
9.99
-------
VII-3
Table VII-2. Industry Emission Rates at 1980 Production Levels
Emission Rate (kg/hr)
Benzene
Total VOC
Model plant uncontrolled
emissions scaled to
1978 demand
Industry emissions
estimated for 1978
39
48
87
140
Table VII-3. Emission Ratios for Model Plant and Industry
Emission Source
process
Fugitive
Storage and handling
Secondary
Uncontrolled Model Plant
Emission Ratio (a/ka)
Benzene
0.88
0.58
0.41
0.02
1.89
Total VOC
1.56
1.97
0.57
0.03
4.13
Industry Emissions
Ratio (g/kg) Estimated for 1978
Benzene
1.24a
0.58
0.3
0.18b
T.3
Total VOC
3.9a
1.97
0.5
0.28b
6.65
Some industry processes have greater emissions from the heavy-ends discharge
(see refs 3,4) .
Some industry processes lose benzene and chlorobenzene into the wastewater
because of poor phase separation and vacuum jet discharge (see refs 3,5).
-------
VII-4
REFERENCES*
A. C. Guessler, "CEH Product Review on Chlorobenzenes," pp. 633.5030 C-L in
Chemical Economics Handbook, Stanford Research Institute, Menlo Park, CA
(March 1980).
T. C. Gunn et al., "CEH Marketing Research Report on Benzene," pp. 718.5021A-E
in Chemical Economics Handbook, Stanford Research Institute, Menlo Park, CA
(May 1977).
S. W. Dylewski, Hydroscience, Inc., Trip Report on Visit to Natrium Plant of PPG
Industries, Inc., March 23, 1978 (data on file at EPA, ESED, Research Triangle
Park, NC).
C. A. Peterson, Jr., Hydroscience, Inc., Trip Report on Visit Regarding Sauget,
Illinois, Plant of Monsanto Co., April 25, 1978 (data on file at EPA, ESED,
Research Triangle Park, NC).
Response to EPA's request for information on emissions from chlorobenzene manu-
facturers (data on file at EPA, ESED, Research Triangle Park, NC).
*Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of that
paragraph, that reference number is indicated on the material involved. When
the reference appears on a heading, it refers to all the text covered by that
heading.
-------
A-l
APPENDIX A
PHYSICAL PROPERTIES OF CHLOROBENZENE, £-DICHLOROBENZENE,
p-DICHLOROBENZENE AND BENZENE
Table A-l. Physical Properties of Chlorobenzene*
Synonyms
Molecular formula
Molecular weight
Physical state
Vapor pressure
Vapor specific gravity
Boiling point
Melting point
Density
Water solubility
Phenyl chloride, monochlorobenzene,
chlorobenzol
112.56
Liquid
1618 Pa at 25°C
3.88
131. 7°C
-45.6°C
1.1058 g/ml at 20°C/4°C
Insoluble
*From: J. Dorigan et al. , "Chlorobenzenes," p. Al-248 in Appendix 1, Rev. 1
(Chemicals A—C), to Scoring of Organic Air Pollutants. Chemistry^Production
and ToKicity of Selected Synthetic Organic Chemicals, MTR-7248, MITRE Corp.,
McLean, VA (September 1976).
Table A-2. Physical Properties of o-Dichlorobenzene*
Synonyms
Molecular formula
Molecular weight
Physical state
Vapor pressure
Vapor specific gravity
Soiling point
Melting point
Density
Water solubility
DCB, dichlorobenzol
C.H.C1,,
642
147.0
Liquid
193 Pa at 25°C
5.05
180.5°C
-17°C
1.305 at 20°C/4°C
Insoluble
*From.- J. Dorigan et al. , "o-Dichlorobenzene," p. AII-38 in Appendix II, Rev. 1
(Chemicals D—E), to Scoring of Organic Air Pollutants. Chemistry, Production
5>nd Toxicity of Selected Synthetic Organic Chemicals, MTR-7248, MITRE Corp.,
McLean, VA (September 1976).
-------
A-2
Table A-3. Physical Properties of g-Dichlorobenzene*
Synonyms DCB, dichlorobenzol
Molecular formula C H Cl
Molecular weight 147.0
Physical state Solid
Vapor pressure 1333 Pa at 54-8°C
Vapor specific gravity
Boiling point 173.4°C
Melting point 53°C
Density 1.2475 g/ml at 55°C/4°C
Water solubility Insoluble
*From: J. Dorigan et al., "£-Dichlorobenzene," p. AII-40 in Appendix II,
Rev. 1 (Chemicals D--E), to Scoring of Organic Air Pollutants. Chemistry,
Production and Toxicity of Selected Synthetic Chemicals, MTR-7248, MITRE Corp.,
McLean, VA (September 1976).
a
Table A-4. Physical Properties of Benzene
Synonyms Benzol, phenylhydride, .coal naphtha
Molecular formula C^H^
o o
Molecular weight 78.11
Physical state Liquid
V.ipoi pressure 95.9 mm at 25°C
Vapor density 2.77
Boiling point 80.1°C at 760 mm
Melting point 5.5°C
Density 0.8787 at 20°C/4°C
Water solubility Slight (1.79 g/liter)
aj. Dorigan, B. Fuller, and R. Duffy, "Benzene," p AI-102 in Scoring of Organic
Air Pollutants. Chemistry, Production and Toxicity of Selected Organic
Chemicals (Chemicals a-c), MTR-7248, Rev 1, Appendix I, MITRE Corp.,
McLean, VA (September 1976).
-------
B-l
APPENDIX B
ATMOSPHERIC DISPERSION PARAMETERS
Table B-l. Atmospheric Dispersion Parameters for Model Plant
Parameters
Source
Kiiilssion Rate (CB
ri-UCB -M (2)
o-DCB (2)
o-DCB
o-DCU SO (2)
Mixed UCB
TCB (2)
Carbori absorber
Secondary emissions
1.57 1.B6
0.053 0.053
None
O.'Jl? 1.01
0.133 0.133
1.40
0.05
0.26
O.-lB(ea)
0.12(ea)
0.056
0.05(ea)
0.075(ea)
0.029{ea)
O.OlS(ea)
0.075
O.OOS(ea)
O.OOJ(ca)
0.00')
U.LH'i(ed)
O.OU2
O.U02(ea)
1.78 6.01
0.058 0.092
Controlled
0.05
O.Ob(ea)
0.02
-------
C-l
APPENDIX C
FUGITIVE-EMISSION FACTORS*
The Environmental Protection Agency recently completed an extensive testing
program that resulted in updated fugitive-emission factors for petroleum re-
fineries. Other preliminary test results suggest that fugitive emissions from
sources in chemical plants are comparable to fugitive emissions from correspond-
ing sources in petroleum refineries. Therefore the emission factors established
for refineries are used in this report to estimate fugitive emissions from
organic chemical manufacture. These factors are presented below.
Source
Pump seals k
Light-liquid service
Heavy-liquid service
Pipeline valves
Gas/vapor service
Light- liquid service
Heavy- liquid service
Safety/ relief valves
Gas/vapor service
Light-liquid service
Heavy- liquid service
Compressor seals
Flanges
Drains
Uncontrolled
Emission Factor
(kg/hr)
0,12
0,02
0.021
0.010
0.0003
0.16
0.006
0.009
0.44
0.00026
0.032
Controlled
Emission Factor
(kq/hr)
0.03
0.02
0.002
0.003
0.0003
0.061
0.006
0.009
0.11
0.00026
0.019
aBased on monthly inspection of selected equipment; no inspection of
heavy-liquid equipment, flanges, or light-liquid relief valves;
10,000 ppmv VOC concentration at source defines a leak; and 15 days
allowed for correction of leaks.
Light liquid means any liquid more volatile than kerosene.
*Radian Corp., Emission Factors and Frequency of Leak Occurrence for Fittings
in Refinery Process Units, EPA 600/2-79-044 (February 1979).
-------
D-l
APPENDIX D
EXISTING PLANT CONSIDERATIONS
Table D-l lists process control devices reported to be in use by industry. To
gather information for the preparation of this report two site visits were made
to manufacturers of chlorobenzene products. One site visit included a tour
through the facilities, the other only involved a discussion of the process.
Data from two other facilities was received in response to request for informa-
tion by the EPA.3'4
PPG Industries, Inc. - New Hartinsville. WV
Variations between the PPG process and the model process are: (1) the water
leaving the water scrubber (about 120 gpm) is discharged to the river while the
vent is emitted to the atmosphere. This results in any VOC that are present in
the vent from the HC1 gas stream resulting in direct emissions or potential
secondary emission; (2) Atmospheric distillation vent pases through a blowdown
kettle which is vented through a steam jet, through a crude product surge tank,
and finally is vented to the atmosphere through a water-cooled surface condenser;
(3) the vacuum distillation system is vented to the atmosphere through a steam
jet and water-cooled contact condenser; (4) the p-dichlorobenzene crystalization
and crystal processing steps are directly vented to the atmosphere; and the
benzene storage tanks are vented directly to the atmosphere. The emissions are
shown in Table D-2 and their sources are shown in Fig. D-l. The cost or
practicability of retrofit improvement to this process has not been studied.
Monsanto Company, Sauget, IL
Variations between the Monsanto process and the model process are: (1) the
water leaving the water scrubber (peak flow of 150 gpm) is discharged to waste-
water treatment while the vent is emitted to the atmosphere. This results in
any VOC that is present in the vent from the HCl gas stream resulting in direct
emission or potential secondary emission; (2) Vents from chlorobenzene distilla-
tion is emitted to the atmosphere; (3) the separation of o-dichlorobenzene from
£-dichlorobenzene is accomplished by crystalization with the o-dichlorobenzene
obtained as product of high purity and the g-dichlorobenzene fraction is a
mixture of isomers. The product is transferred or sold without a pj-dichloro-
benzene crystal purification and handling step and thus does not result in
losses attendant to this type of processing. The emissions are shown in
Table A-3.2 The cost or practicability of retrofit improvement to this process
has not been studied.
-------
Table D-l. Emission Control Devices Currently Used by Domestic
Chlorobenzene Products Industry
PPG
a
Source Industries
Control
Monsanto
Company'3
Devices Used
Montrose Corp.
of Calif.0
Dow ,
Q
Chemical
Tail-gas treatment
Atmospheric
distillation
Vacuum system
p-DCB crystallization
Water scrubber
Vent condensers
and steam jet
after-condenser
Steam jet
after condenser
None
Water scrubber
Vent condensers
N.R.
N.R.
Water scrubber
Vent condensers
_
See ref 1-
b* c •>
See ref ^«
£*
See ref 3.
d _ .
See ref 4.
STo be installed May 1978.
Not reported.
gNot applicable.
^Reported to be on crystallization vent.
Water scrubber,
caustic scrubber
Vent condensers,
venturi scrubber,
carbon adsorption
Steam jet after- Venturi scrubber
condenser
g
N.A.
Carbon adsorption
h
p-DCB crystal processing
Storage and handling
None
None
N.R.
None
N.A.
N.R.
Carbon adsorption
Vent condenser on
benzene storage
D
I
See ref 4.
-------
•TABLE p-2
DIRECT EMISSIONS
Pollutant Flow Rate lb/hr/100 Tons/day Products
Stream
No.
1
5
9
11
15
20
12
13
22
ECD No. Benzene MCB o-DCB p-DCB m-DCB Gas CPH
101 0.017
102 8.34 24.3 0.95 1.34
103 — 4.0 — — __ 57
104 0.004 0.0117 0.0002 0.0042 — .0011
105 ~ ~ 0.06 6.25 0.26 0.274
106 — — 1.098 - — .04G
Receiver Vent .01
Receiver Vent — 0.0018
Roof Fan1 0.12 0.14 5.54 98.61
8.49
28. 38
0
7.65
106.2
0.62
No Emission Control Device present
-------
D-4
Montrose Chemical Corporation - Henderson, NV
Variations between the Montrose Chemical process and the model process are-.
(1) the water leaving the water scrubber (quantity unknown) is presumed to go
to wastewater treatment while the vent stream containing benzene (3.9 Ib/hr)
and chlorinated benzenes (0.76 Ib/hr) is emitted to the atmosphere; (2) the
vent stream from distillation passes through steam jet ejectors, water-cooled
after-condensers, and are emitted to the atmosphere; (3) the dichlorobenzenes
are presumed to be sold prior to any further processing. The cost or practi-
cability of retrofit improvement to this process has not been studied.
Dow Chemical USA - Midland, MI
Variations between the Dow Chemical process and the model process are: (1) the
vent stream from the caustic scrubber is discharged to the atmosphere, however,
the manufacturer stated they planned to install an adsorption bed (presumed to
be carbon) in May 1978; (3) the vent streams from distillation pass through
steam jet vacuum systems and discharge to the atmosphere; (4) the vent stream
from £-dichlorobenzene does go through a carbon adsorption bed, however, not in
combination with other process vent streams. The cost or practicability of
retrofit improvement to this process has not been studied.
Retrofitting Controls
The primary difficulty associated with retrofitting may be in finding space to
fit the control device into the existing plant layout. Because of the costs
associated with this difficulty it may be appreciably more expensive to retrofit
emission control systems in existing plants than to install a control system
during construction of a new plant.
-------
D-5
REFERENCES*
1. S. W. Dylewski, IT Enviroscience, Inc., Trip Report on Visit to Natrium Plant
of PPG Industries, Inc., March 23, 1978 (on file at EPA, ESED, Research Triangle
Park, NC).
2. C. A. Peterson, Jr., IT Enviroscience, Inc., Trip Report on Visit Regarding
Sauget, Illinois, Plant of Monsanto Co., April 25, 1978 (on file at EPA,
ESED, Research Triangle Park, NC).
3 H. J- Wurzer, Montrose Chemical Corp. of California, letter dated March 7,
1978, to EPA.
4, J. S. Beale, Dow Chemical Company, letter dated March 14, 1978, to EPA.
*Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of
that paragraph, that reference number is indicated on the material involved.
When the reference appears on a heading, it refers to all the text covered by
that heading.
-------
E-l
APPENDIX E
COST ESTIMATE PROCEDURE
Emission to Carbon Adsorption (CA)
Benzene = x _ _ x _ =
hr kg 60 mm 78 Ibs Ib mole
2 43 kn
Monochlorobenzene j" * at 113 Ib/lb mole = 0.28 scfm
Dichlorobenzene 4'8^ kg at 147 Ib/lb mole = 0.44 scfm
hr -
Total VOC 2.22 scfm
Air = 5531 kg/hr at 29 Ib/lb mole 2510.56 scfm
Total waste gas to CA 2513 scfm
From Fig. IV- 1 of the control device evaluation report for carbon adsorption
the December 1979 installed capital for 2,500 scfm is $200,000. The pretreat-
ment system requires a 46 ft2 Karbate heaterchanger to reduce the relative
humidity. Since the waste gas is corrosive, adequate protection will be re-
quired such as glass -lined pumps, and the use of liners such as epoxy, Haveg,
or equivalent. These extra items are estimated to increase the capital cost by
$125,000. Total installed capital = $325,000.
With an average VOC molecular weight of MOO and an estimated loading capacity
of 6 Ib of VOC/100 Ib of carbon, the carbon requirement from CA report Fig. II-l
is 4 Ib of carbon/1000 scf. The total carbon requirement is therefore:
4 Ib C 2513 scf 60 min 3 hr _ 1809 Ib of C
1000 scf min hr cycle ~ cycle
As indicated by CA report Fig. IV-1, the capital cost includes the carbon for
two beds. CA Fig. IV-4 indicates the annual cost to be $32/scfm or $80,416.
The annual cost adjustments for fixed costs associated with added capital,
steam for gas conditioning and raw material recovery credits are included in
the following cost summary.
-------
Carbon Adsorption Control Cost Summary
Capital from CA report (Fig. IV-1)
Extra capital
Total capital
Fixed cost for extra capital
Steam conditioning
Benzene credit
Chlorobenzene credit (as raw material)
Annual cost: CA Fig. IV- 4
Net annual cost
Benzene emission reduction 98.7%
Total VOC reduction 98.7%
Cost effectiveness
$/Mg of benzene reduced
$/Mg of VOC reduced
Model Plant
2500 scfm
$200,000
125,000
$325,000
$ 36,250
1,686
(18,468)
(12,150)
80,416
$ 87,734
79.5 Mg/yr
143 Mg/yr
$1104
$ 613
1250 scfm
$150,000
82,469
$232,469
$ 23,916
843
(9,234)
(6,075)
53,750
$ 63,200
39.7 Mg/yr
72 Mg/yr
$1,592
$ 884
3750 scfm
$240,000
159,428
$399,428
$ 46,234
2,529
(27,702)
(18,225)
101,250
$104,086
119 Mg/yr
214 Mg/yr
$875
$486
-------
F-l
APPENDIX F
LIST OF EPA INFORMATION SOURCES
F. J. Basile, Jr. (Monsanto Co.), letter to C. A. Peterson (Hydroscience, Inc.),
March 3, 1978.
C. R. Dilmore, Jr. (PPG Industries), letter to S. W. Dylewski (Hydroscience, Inc.),
February 8, 1978.
H. J. Wurzer (Montrose Chemical Corp. of California), letter to EPA,
March 7, 1978.
J. S. Beale (Dow Chemical Co.), letter to EPA, March 14, 1978.
-------
4-i
REPORT 4
MALEIC ANHYDRIDE
R. E. White
J. F. Lawson
IT Enviroscience
9041 Executive Park Drive
Knoxville, Tennessee 37923
Prepared for
Emission Standards and Engineering Division
Office of Air Quality Planning and Standards
ENVIRONMENTAL PROTECTION AGENCY
Research Triangle Park, North Carolina
September 1980
This report contains certain information which has been extracted from the
Chemical Economics Handbook, Stanford Research Institute. Wherever used,
it has been so noted. The proprietary data rights which reside with
Stanford Research Institute must be recognized with any use of this material.
D110A
-------
4-iii
CONTENTS OF REPORT 4
Page
I. ABBREVIATIONS AND CONVERSION FACTORS 1-1
II. INDUSTRY DESCRIPTION II-l
A. Maleic Anhydride II-l
B. MA Usage and Growth II-l
C. Domestic Producers II-l
D. References II-7
III. PROCESS DESCRIPTIONS III-l
A. Introduction III-l
B. Benzene Oxidation Process III-l
C. n-Butane Oxidation III-5
D. Phthalic Anhydride By-Product Process III-5
E. Foreign Processes (Mixed Butenes) III-5
F. References III-7
IV. EMISSIONS IV-1
A. Benzene Oxidation Process IV-1
B. n-Butane Process IV-7
C. Phthalic Anhydride By-Product Process IV-7
D. References IV-8
V. APPLICABLE CONTROL SYSTEMS V-l
A. Benzene Oxidation Process V-l
B. Other Processes V-7
C. References V-8
VI. IMPACT ANALYSIS VI-1
A. Control Cost Impact VI-1
B. Environmental and Energy Impacts VI-11
C. References VI-14
VII. SUMMARY VII-1
-------
4-v
APPENDICES OF REPORT 4
A. PHYSICAL PROPERTIES OF MALEIC ANHYDRIDE A-l
B. AIR-DISPERSION PARAMETERS B-l
C. FUGITIVE-EMISSION FACTORS C-l
D. DETAILED COST ESTIMATES AND CALCULATIONS FOR ADSORPTION AND D-l
INCINERATION
E. EXISTING PLANT CONSIDERATIONS E-l
-------
4-vii
TABLES OF REPORT 4
Number
II-l Maleic Anhydride Usage and Growth II-2
II-2 Maleic Anhydride Capacity II-3
IV-1 Benzene and Total VOC from Uncontrolled Emissions from IV-3
Production of Maleic Anhydride in Model Plant (22,700-Mg/yr
Capacity)
IV-2 Waste Gas Composition—Product Recovery Absorber (Weighted IV-4
Average)
IV-3 Model-Plant Storage IV-6
V-l Benzene and Total VOC from Emissions Controlled by Carbon V-2
Adsorption in the Production of Maleic Anhydride in Model Plant
(22,700-Mg/yr Capacity)
V-2 Benzene and Total VOC from Emissions Controlled by Incineration V-4
in the Production of Maleic Anhydride in Model Plant (22,700-Mg/yr
Capacity)
VI-1 Carbon Adsorption System Emission Control Cost Estimate for VI-3
Model Plant (Main Process Vent and Refining Vacuum Vent)
VI-2 Cost Factors Used to Compute Annual Costs VI-4
VI-3 Incineration System Emission Control Cost Estimate for Model VI-7
Plant (Main Process Vent and Refining Vacuum Vent)
VI-4 Environmental Impact - Model Plant Emission Controlled VI-12
VII-1 Summary of Uncontrolled and Controlled Emissions from Model VII-2
Plant (22,700-Mg/yr Capacity)
A-l Physical Properties of Maleic Anhydride A-l
A-2 Physical Properties of Benzene A-l
A-3 Physical Properties of Formic Acid A-2
A-4 Physical Properties of Formaldehyde A-3
A-5 Physical Properties of Maleic Acid A-3
B-l Air-Dispersion Parameters for Model Plant (22,700-Mg/yr Capacity) B-l
D-l Carbon Adsorption System Emission Control Costs for Main D-5
Process Vent and Refining Vacuum Vent
D-2 Incineration System Emission Control Costs for Main Process D-7
Vent and Refining Vacuum Vent
E-l Control Devices Currently Used by the Maleic Anhydride Industry E-2
in the United States
-------
4-ix
FIGURES OF REPORT 4
Number
II-l Manufacturing Locations of Maleic Anhydride II-4
III-l Maleic Anhydride - Model Plant III-2
VI-1 Cost vs Waste-Gas Flow Rate - Carbon Adsorption VI-5
VI-2 Cost vs Waste-Gas Flow Rate - Incineration Without Heat VI-8
Recovery
VI-3 Cost vs Waste-Gas Flow Rate - Incineration with 50% Heat VI-9
Recovery
VI-4 Cost Effectiveness vs Waste-Gas Flow Rate VI-10
D-l Precision of Capital Cost Estimates D_2
-------
1-1
ABBREVIATIONS AND CONVERSION FACTORS
EPA policy is to express all measurements used in agency documents in metric
units. Listed below are the International System of Units (SI) abbreviations
and conversion factors for this report.
To Convert From
Pascal (Pa)
Joule
-------
II-l
II. INDUSTRY DESCRIPTION
A. MALEIC ANHYDRIDE
Maleic anhydride (MA) production was selected for the following reasons:
preliminary estimates indicated that emissions of volatile organic compounds
(VOC) are relatively high; the predominant manufacturing process emits large
quantities of benzene, which was listed as a hazardous pollutant by the EPA in
the Federal Register on June 8, 1977; and the product growth is expected to be
higher than the industry average.
HA is solid at ambient conditions (see Appendix A for pertinent physical prop-
erties) although it generally exists in the process as a liquid (molten MA) or
as maleic acid. The predominant emission, benzene, however, is a volatile
liquid at ambient conditions but is emitted as a gas.
B. MA USAGE AND GROWTH
Table II-l shows MA end uses and the expected growth rate. The major end use
is the production of unsaturated polyester resins, which are used in reinforced-
plastic applications such as marine craft, building panels, automobiles, tanks,
and pipes.
The domestic MA production capacity for 1979 was reported1 to be 214,000 Mg,
with approximately 70% of this capacity being utilized in 1978. Based on the
assumption of an 8% annual growth in MA consumption, production will reach
approximately 96% of the present capacity by 1982. No shortage of benzene, the
major raw material, is expected during this period.
C. DOMESTIC PRODUCERS
As of 1978 there were eight domestic facilities producing MA in ten plants.
Table II-2 lists the producers and the processes being used; Fig. II-l shows the
plant locations. Approximately 83% of the 214,000-Mg/yr domestic capacity is
based on the oxidation of benzene. Oxidation of n-butane accounts for another
15% of capacity, and the remaining 2% is from phthalic anhydride production,
which gives MA as a by-product.1 The projected growth rate for the n-butane
oxidation process through 1982 is 24%, primarily through conversion, compared
to an overall maleic anhydride expected growth rate of 8%. Data regarding the
-------
II-2
Table II-l. Maleic Anhydride Usage and Growth
End Use
Unsaturated polyester resins
Fumaric acid
Agricultural chemicals
Alkyd resins
Lubricating additives
Copolymers
Maleic acid
Chlorendic anhydride and acid
Other
Production (%)
56
5
8
1
9
6
4
2
9
Average Annual
Growth (%)
9
4
6
0
9
8
8
10
17
100
See ref 1.
-------
II-3
Table II-2. Maleic Anhydride Capacity'
Production
Company Location
Amoco Joliet, IL
Ashland Neal, WV
Koppers Bridgeville, PA
Chicago, IL
Monsanto St. Louis, MO
Denka (Petro-Tex) Houston, TX
Reichholdf Elizabeth, NJ
Morris , IL
Tenneco Fords , NJ
U. S. Steel Neville Island, PA
Total
a
See ref 1.
Oxidation of n-butane.
c
Oxidation of benzene.
d!5,000-Mg plant shut down in April 1979.
6By-product of phthalic anhydride manufacture.
fSee ref 6.
Capacity as of 1979
(10 Mg) Process
27 b
27 c
od
5 e
48 C(80%) b(20%)
23 c
14 c
20 c
12 c
38 c
214
-------
II-4
1. Amoco, Joliet, IL
2. Ashland, Neal, W. VA
3. Ko£3pers, Bridgeville, PA
4. Koppers, Chicago, IL
5. Monsanto, St, Louis, MO
6. Denka (Petro-Tex), Houston, TX
7. Reichhold, Elizabeth, NJ
8. Reichhold, Morris, IL
9. Tenneco, Fords, NJ
10. U.S. Steel, Neville Island, PA
Fig. II-l. Manufacturing Locations of Maleic Anhydride
-------
II-5
economic incentives for switching to n-butane oxidation are not available. No
growth in the quantity of MA recovered during phthalic anhydride production is
expected.2
In 1960 work began on developing a catalyst suitable for producing MA from
butane/butene (C ) streams available from naphtha cracking. This effort was
curtailed during the 1961—1967 period, when the MA market was depressed and
low-cost benzene was available. In 1967 demand for MA increased, and work was
renewed in Japan by Kasei Mizuishima; in 1974 announcements concerning the
production of MA from C s were made by Petro-Tex, Chem Systems, BASF, Bayer,
Alusuisse/UCB, and Mitsubishi.2 Presently, Amoco and Monsanto are producing MA
from an n-butane feedstock.1 The main drawback of the n-butane process is the
unavailability of a catalyst that provides competitive yields.3
Companies that product MA are listed below:
1. Amoco Chemicals Corporation
Amoco has the only domestic plant totally dedicated to the n-butane
process.1 It has an annual capacity of 27,000 Mg and is reportedly expan-
sible to 41,000 Mg.3
2. Ashland Chemical Company
The Ashland facility is a benzene-based plant with an annual capacity of
27,000 Mg and is expansible to 41,000 Mg.1 This plant can be switched
from benzene to n-butane feedstocks.3
3. Koppers Company, Inc.
Koppers announced that their 15,000-Mg/yr benzene oxidation plant was shut
down in April 1979. Their Chicago facility can recover 5000 Mg of MA per
year from the effluent of their phthalic anhydride plant, which was started
in 1975.1
4. Monsanto Company
Monsanto with a capacity of 48,000 Mg/yr is the largest producer of MA.
Some MA is consumed captively to produce fumaric acid, maleate/fumerate
esters, styrene copolymers, and ethylene-maleic anhydride copolymers.
Monsanto plans to start up a 46,000-Mg butane-based plant in Pensecola,
FL, in 1983.l
-------
II-6
5. Denka USA
Denka's 23,000-Mg/yr Houston facility was designed by Scientific Design
Company, Inc., and was purchased from Petro-Tex Chemical Corporation on
July 1, 1977.4 They have a current permit from the Texas Air Control Board
to operate an n-butane reactor.5
6. Reichhold Chemicals, Inc.
Reichhold's combined production from their Elizabeth, NJ, and Morris, TL,
plants is 34,000 Mg/yr, some of which is used captively to produce unsat-
urated polyester resins, alkyd resins, and plasticizers.1;6
7. Tenneco Chemicals, Inc.
A small part of Tenneco's 12,000-Mg/yr MA production is used captively to
produce fumaric acid, dibutyl maleate, and dodecanyl-succinic anhydride.1
8. United States Steel Corporation
The MA capacity of U.S. Steel was expanded to 38,000 Mg/yr.3 Some of
their MA production is used captively to produce fumaric acid, dibutyl
maleate, and dioctyl maleate.1
9. Allied Chemicals Corporation
Allied ceased production in 1974 of MA at their Moundsville, WV, plant,
which had a capacity of 900 Mg/yr.1.
The expansion capabilities of 14,000 Mg each for Amoco and Ashland plus the
46,000-Mg new Monsanto plant represent a potential nationwide capacity of
288,000 Mg per year.
-------
II-7
D. REFERENCES*
1. G. T. Gerry, "CEH Marketing Research Report on Maleic Anhydride,"
pp 672.5031A—672.5033F in Chemical Economics Handbook, Stanford Research
Institute, Menlo Park, CA (November 1979).
2. "Mitsubishi Chemical Details Its C4-Based Maleic Process," p 30,
European Chemical News (Apr. 5, 1974).
3. "Maleic Makers Build on Hopes for Polyester," Chemical Week, pp 37 and 38
(Feb. 2, 1977).
4. Personal communication Nov. 17, 1977, between J. F. Lawson, IT Enviroscience,
Inc., and R. E. Hinkson, Denka USA.
5. Permit Exemption Request from R. D. Pruessner, Petro-Tex Chemical, to Charles
Barden, Texas Air Control Board, Feb. 23, 1976.
6. P.S. Hewett, Reichhold Chemicals, Inc., letter dated Mar. 27, 1978, to D. P.
Patrick, EPA.
*Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of
that paragraph, that reference number is indicated on the material involved.
When the reference appears on a heading, it refers to all the text covered by
that heading.
-------
III-l
III. PROCESS DESCRIPTIONS
A. INTRODUCTION
As discussed in Sect. II, the two major processes used to manufacture maleic
anhydride (MA) in the United States are benzene oxidation and butane oxidation.
A small amount of MA is recovered as a by-product of phthalic anhydride produc-
tion. The only significant foreign process for MA production not used in the
United States starts with a butene mixture feedstock. This process is operated
in France1 and Japan.2 There are no known plans to introduce this process
domestically.
B. BENZENE OXIDATION PROCESS
1- Basic Process
MA is produced by the following chemical reaction:
o
//
H-C - C
+ 9/20 »• H S0 + 2H 0 -I- 2C02
H-C - C
^
0
(benzene) (oxygen) (MA) (water) (carbon dioxide)
The process flow diagram shown in Fig. III-l represents a typical process,
which is continuous; however, some plants operate batchwise. The emissions in
either case are judged to be the same.3
A mixture of vaporized benzene and air enters a tubular fixed-bed reactor,
where the catalytic oxidation of benzene is carried out at a temperature of
350 to 400°C. The catalyst contains approximately 70% vanadium pentoxide
supported on an inert carrier,- most of these systems also contain 25 to 30%
molybdenum oxide. The reaction is highly exothermic, releasing 24.4 MJ/kg of
reacted benzene, with the excess heat being used to generate steam. MA yields
range from 60 to 67% of theoretical.3
The reactor feed mixture contains an excess of air because benzene is explosive
in air at concentrations above 1.5 vol %. The resulting large volume of reactor
exhaust (stream 3) dictates the size of subsequent product recovery equipment.
-------
AIR
©
BEuiSVJe.
^TORWie
U=sn
i\\ I V
COMPR£ ^i CEACTCP.(S)
RCD
-------
III-3
The stream passes through a cooler, partial condenser, and separator, in which
40% of the MA is condensed and separated as crude MA (stream 4).4 The remaining
vapor (stream 5) enters the product recovery absorber, where it is contacted
with water or aqueous maleic acid. The absorber product (stream 6) is a
40 wt % aqueous solution of maleic acid. The absorber vent (A) exhausts to
the atmosphere or is directed to an emission control device.3
The 40% maleic acid (stream 6) is dehydrated by azeotropic distillation with
xylene. Any xylene retained in the crude MA (stream 9) is removed by the
xylene stripping column, and the crude MA (stream 10) from this column is
combined with the crude MA (stream 4) from the separator.
Crude MA is aged, which causes any color-forming impurities to polymerize.
After aging, the crude MA (stream 11) is fed to the fractionation column, which
yields molten MA as the purified overhead product (stream 12). A small percen-
tage is taken an additional step for sale as briquets. The fractionation
column bottoms containing the color-forming impurities are removed as liquid
residue waste (stream 13). This stream either becomes part of the untreated
effluent or is fed to a liquid incinerator.3
The vacuum lines from the dehydration column, xylene stripper, and fraction-
ation column are joined (stream 14) to the vacuum system. The refining vacuum
system vent (B) can exhaust to the atmosphere, recycle (stream 5) to the product
recovery absorber, 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
. O
residue waste (stream 13).
Essentially all process emissions will exit through the product recovery absorber
(vent A). These emissions will include any unreacted benzene, which can constitute
3 to 7% of the total benzene feed.5 The only other process emission source is
the refining vacuum system vent (vent B), which can contain small amounts of MA
and xylene.
Fugitive emissions throughout the process can contain benzene, xylene, MA, and
maleic acid. Corrosion problems due to leaks caused by maleic acid can increase
-------
III-4
fugitive emissions. As with most organic chemical processes, leaks into cooling
water could occur and allow volatile organic compounds (VOC) to escape as a
fugitive emission.
Storage and handling emission sources (labeled C on Fig. III-l) include benzene,
xylene, MA, and crude-MA storage, plus emissions from the briquetting operation.
There are four potential sources of secondary emissions (labeled K on Fig.III-l):
spent reactor catalyst, excess water from the dehydration column, vacuum system
water, and fractionation column residues. The small amount of residual organics
in the spent catalyst after it is washed has a low vapor pressure and produces no
significant emissions. Xylene is the principal organic contamination in the
excess water from the dehydration column and the vacuum system water. Residues
from the fractionation column are relatively heavy organics with a molecular
weight greater than 116 and produce no significant secondary emissions.
2. Process Variations
In place of the partial condensation system (cooler, partial condenser, and
separator) shown in Fig. III-l, a so-called switch condenser system can be
incorporated. This utilizes a series of condensers that are alternately cooled
to freeze solid MA on the surface and then heated to melt the MA for pumping to
crude-MA storage. Switch condensing can remove up to 60% of the MA from the
process, compared to 40% removed by the partial condensation system.6 The
removal of additional MA 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 being used for dehydration.
Several other agents can be used, including isoamyl butyrate, di-isobutyl
ketone, anisole, and cumene.3
A vacuum evaporation system, which replaces the dehydration column and xylene
stripper, is used by at least one plant to remove water and to dehydrate the
maleic acid to form MA.6 Since an azeotropic agent is no longer required,
xylene is eliminated from process emissions.
-------
III-5
C- n-BUTANE OXIDATION
All process data concerning the n-butane oxidation process are currently pro-
prietary and unavailable. A benzene oxidation process can be converted to
n-butane oxidation by changing the catalyst system; this conversion can be done
for much less than the cost of a new plant. However, the converted process
would be less efficient than a new n-butane process because reaction conditions
are not optimum. The lowered efficiency might be serious enough to warrant
other major design changes.7
A major advantage of this process is that there are no benzene emissions.
Other VOC emissions should not differ to a great extent from the benzene oxida-
tion process.5
D. PHTHALIC ANHYDRIDE BY-PRODUCT PROCESS
Phthalic anhydride is manufactured from naphthalene and ortho-xylene. Maleic
anhydride is recovered as a by-product from the plant effluent.3 The emissions
associated with MA recovery are believed to be insignificant and are not being
investigated at this time.
E. FOREIGN PROCESSES (MIXED BUTENES)
1- Basic Process
The only significant deviation from benzene oxidation for MA production in
foreign processes is the use of feedstocks of 65 to 80% n-butenes, with the
remainder being mostly butanes or isobutene. The general process description
is very similar to that shown in Fig. III-l for benzene oxidation and therefore
is not repeated here.1'2
The exhaust from the main process vent contains unreacted butane, butene, and
carbon monoxide and various secondary products. Except for the absence of
benzene the emissions should be about the same as those for the benzene oxida-
tion process.5
2. Process Variations
The most significant process variation is the use of a fluidized catalyst bed
rather than a fixed bed. This variation provides good temperature control
-------
111-6
within the bed and thus allows optimum ratios of butene/air. In contrast
optimum benzene/air ratios cannot be used with fixed-bed systems because excess
air is necessary for the processes 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, the product yields with the fluidized
bed are not as good as with the fixed bed.8
-------
III-7
D- REFERENCES*
1- D. Lenz and M. De Bouille, "The Bayer Process for the Production of Maleic
Anhydride from Butenes," Revue de 1 "Association Francaise des Techniciens
du Petrole 236(20-3), 17 (1976).
2- Uemura Shinji and Kamimura Shiego, "Production of Anhydrous Maleic Acid from C
Distillate," Petroleum Association Journal, vol 16, No. 8 (1973). ^
3- R. T. Gerry e_t a 1., "CEH Marketing Research Report on Maleic Anhydride,"
pp 672.5031A—672.5033F in Chemical Economics Handbook, Stanford Research
Institute, Menlo Park, CA (November 1979).
4. J. F. Lawson, IT Enviroscience, Inc., Trip Report for Visit to Denka Chemical
Corp., Houston, TX, Nov. 17, 1977 (on file at EPA, ESED, Research Triangle
Park, NC).
5. W. A., Lewis, Jr., G. M. Rinaldi, and T. W. Hughes, Monsanto Research Corp.,
Source Assessment: Maleic Anhydride Manufacture (received January 1978).
6. J. F. Lawson, IT Enviroscience, Inc., Trip Report for Visit to Reichhold
Chemicals, Inc., Morris, IL., July 28, 1977 (on file at EPA, ESED, Research
Triangle Park, NC).
7. "Maleic Makers Build on Hopes for Polyester," Chemical Week, pp 37 and 38
(Feb. 2, 1977).
8- "Mitsubishi Chemical Details Its C4-Based Maleic Process," p 30,
European Chemical News (Apr. 5, 1974).
^Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of that
paragraph, that reference number is indicated on the material involved. When
the reference appears on a heading, it refers to all the text covered by that
heading.
-------
IV-1
IV. EMISSIONS
Emissions in this report are usually identified in terms of volatile organic
compounds (VOC). VOC are currently considered by the EPA to be those of a
large group of organic chemicals, most of which, when emitted to to the atmos-
phere, participate in photochemical reactions producing ozone. A relatively
small number of organic chemicals have low or negligible photochemical reactivity.
However, many of these organic chemicals are of concern and may be subject to
regulation by EPA under Section 111 or 112 of the Clean Air Act since there are
associated health or welfare impacts other than those related to ozone formation.
The process emission estimates for the benzene oxidation model plant are based
on the emissions which were reported in the Houdry study, SRI, a trip report on
a visit to Reichhold, and an understanding of the process chemistry and yields.
A- BENZENE OXIDATION PROCESS
!• Model Plant
The model plant* for this study has a capacity of 22,700 Mg/yr (50 x 106 Ib/yr),
based on 8760** annual hours of operation. Although not an actual operating
plant, it is typical of most plants. The model benzene oxidation process,
shown in Fig. III-l, best fits today's maleic anhydride manufacturing and
engineering technology. Single-process trains as shown are typical for the
large plants except for the reaction area, where multiple reactors are common.
The model process uses partial condensation and azeotropic drying with xylene.
Typical raw-material, intermediate, and product storage-tank capacities were
estimated for a 22,700-Mg/yr plant. The storage-tank requirements are covered
under storage emissions. Estimates of potential fugitive sources were based on
an equipment count from existing facilities. Characteristics of the model
plant .important to air-dispersion modeling are shown in Appendix B.
*See p 1-2 for a discussion of model plants,
**Process downtime is normally expected to range from 5 to 15%. If the hourly
rate remains constant, the annual production and annual VOC emissions will
be correspondingly reduced. Control devices will usually operate on the same
cycle as the process. From the standpoint of cost-effectiveness calculations
the error introduced by assuming continuous operation is negligible.
-------
IV-2
2. Sources and Emissions
All emission rates and sources for the benzene oxidation process are summarized
in Table IV-1; part of the data in the table are from refs 1—4.
a. Main Process Vent The largest vent is the main process vent (A, Fig. III-l)
from the product recovery absorber. All plants have this vent. The emission
is influenced by the excess air fed to the reactor to maintain the benzene
concentration below the explosive limit. The composition of this emission for
the model plant is shown in Table IV-2.1—3 The majority of the unreacted
benzene is contained in this stream.
Process upsets that cause more benzene release will affect benzene emissions
since the absorber can remove benzene only up to its solubility level in water.
These upsets can cause short-duration benzene and VOC emissions of 3 to 5 times
the normal amounts. Process startup also increases benzene emissions 3 to 5
times the normal amount because of incomplete benzene reaction. Shutdown will
not affect emissions because benzene is shut off as the first step in the
shutdown procedure. As a result the level of unreacted benzene emitted from
the reactor is immediately reduced.5
b. Refining Vacuum Vents The refining vacuum system vent (B, Fig. III-l) exhausts
the noncondensibles from the three vacuum columns used to dehydrate and fraction-
ate MA. The emissions from this vent are affected by the use of inert gases
bled into the system for vacuum control and by process leaks into the system.
The VOC emission will be maleic acid or xylene and is estimated to be relatively
insignificant, as is indicated in Table IV-1. Process upsets, startups, and
shutdowns do not affect the VOC emissions from this vent.3
c. Fugitive Emissions Process pumps and valves are potential sources of fugitive
emissions. The model plant is estimated to have 15 pumps handling VOC, 3 of
which handle benzene. All remaining pumps handle essentially heavy liquids.
The estimated number of valves is 500, with 75 controlling benzene vapor, 100
controlling benzene liquid, and 325 controlling heavy liquids. The fugitive-
emission factors from Appendix C were applied to this valve and pump count to
determine the fugitive emissions shown in Table IV-1.
-------
IV-3
Table IV-1. Benzene and Total VOC from Uncontrolled
Emissions from Production of Maleic Anhydride in
Model Plant (22,700-Mg/yr Capacity)
Emission Source
Product recovery
absorber0
Refining vacuum system
c,
Storage and handling
pugitivef
Secondary
Total
Designation
(Fig. IV-1)
A
B
C
J
K
Emission Ratio
(kg/kg X 10~3)a
Benzene
67.0
1.23
1.13
69 X 10~3
Total VOC
86.0
0.1
1.27
1.24
0.11
89 X 10~3
Emission Rate
(kg/hr)b
Benzene
173
3.2
2.93
179.0
Total VOC
224
0.28
3.3
3.23
0.3
230.0
kg of emission per kg of maleic anhydride produced.
Emission rates are annual averages at 8760 hr/yr.
c
See refs 1—3.
See ref 3.
a
"See ref 4.
f
See Appendix C.
-------
IV-4
Table IV-2. Waste Gas Composition—Product Recovery Absorber'
(Weighted Average)
Component
Oxygen
Nitrogen
Carbon dioxide
Carbon monoxide
Water
Benzene
Maleic acid
Formaldehyde
Formic acid
Amount (wt % )
16.67
73.37
3.33
2.33
4.00
0.23
0.01
0.05
0.01
Emission Rate
(kg/hr)
12,603
55,470
2,517
1,761
3,025
173
7
37
7
75,600
See refs 1—3.
-------
IV-5
Storage and Handling Emissions—Emissions result from the storage and handling
of benzene, maleic anhydride, and xylene. For the model plant the sources (C)
are shown on the flow diagram in Fig. III-l. Storage-tank data for the model
plant are given in Table IV-3. The emissions in Table IV-1 were calculated
based on fixed-roof tanks, one-half full, on a 11°C diurnal temperature varia-
tion, and on the emission equations from AP-42.4 However, breathing losses
were divided by 4 to account for recent evidence6 indicating that the AP-42
breathing-loss equation overpredicts emissions.
Benzene freezes at 5.5°C; therefore storage tanks are generally heated to
maintain the temperature above freezing. Maleic anhydride freezes at 52.8°C;
therefore the finished product is normally stored at 60°C and in-process
material is stored at 60 to 105°C.
The equations in AP-424 were used to calculate the emissions from loading bulk
maleic anhydride into tank cars and trucks. These emissions, included in the
storage and handling emissions (Table IV-1), are 0.02 kg/hr. Maleic anhydride
dust is produced in the briquetting operation but is not a significant atmos-
pheric emission.
Secondary Emissions Secondary emissions of VOC can result from the handling
and disposal of process-waste liquid and solid streams. For the model plant
four potential sources (K) are indicated on the flow diagram. Fig. III-l.
The spent catalyst removal and reclamation do not present a significant emission
potential because the catalyst is thoroughly purged and washed before it is
removed. Reclamation is normally done off-site.2
For the model plant, aqueous effluent from the vacuum system is estimated to be
417 kg/hr, of which 0.05 kg/hr is xylene. The fractionation column residue
stream is 27 kg/hr of low-vapor-pressure organics. Even untreated, these
streams do not represent significant emissions. In at least one operating
facility, incineration is used to destroy the organics in these liquid
streams.2 With it assumed that there is a well-designed liquid incinerator with
99% destruction of organics, the secondary VOC emissions from this incinerator
would be 0.27 kg/hr. Whether liquid incineration or terminal wastewater treat-
-------
IV-6
Table IV-3. Model-Plant Storage
Content
Benzene
Maleic anhydride
Xylene
Tank Size
(m3)
2460
75
380
190
380
150
Turnovers
Per Year
12.5
410
40
80
40
12
Bulk
Liquid
Temp. (°C)
13
13
77
71
60
13
-------
IV-7
ment is used, the emission will be no greater than 0.27 kg/hr; this emission
rate is included in uncontrolled emissions, Table IV-1.
B- n-BUTANE PROCESSES
As stated earlier, no benzene emissions are associated with the n-butane
process. In all other respects the VOC emissions are believed to be about the
same as those from the benzene oxidation process, although there are no public
data to support this statement.
c- PHTHALIC ANHYDRIDE BY-PRODUCT PROCESSES
The emissions associated with MA recovery are believed to be insignificant and
are not being investigated at this time.
-------
IV-8
D. REFERENCES*
1. P. L. Morse, Maleic Anhydride, Interim Al, 46A1, A private report by Process
Economic Program, Stanford Research Institute, Menlo Park, CA (November 1973).
2. J. F. Lawson, IT Enviroscience, Inc., Trip Report for Visit to Reichhold
Chemicals, Inc., Morris, LA, July 28, 1977 (on file at EPA, ESED, Research
Triangle Park, NC).
3. J. W. Pervier et al., Houdry Division of Air Products, Inc., Survey Reports
on Atmospheric Emissions from the Petrochemical Industry, Volume III,
EPA-450/3-73-005-C (April 1974).
4. C. C. Masser, "Storage of Petroleum Liquids," pp 4.3-1 to 4.3-17 in
Compilation of Air Pollutant Emission Factors, AP-42, Part A, 3d ed.
(August 1977).
5. Personal communication between J. F. Lawson, IT Enviroscience, Inc., and G. R.
Wood, Monsanto Chemical Co., Oct. 20, 1977.
6. E. C. Pulaski, TRW, Inc., letter dated May 30, 1979 to Richard Burr, EPA.
*Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of that
paragraph, that reference number is indicated on the material involved. When
the reference appears on a heading, it refers to all the text covered by that
heading.
-------
V-l
V. APPLICABLE CONTROL SYSTEMS
A. BENZENE OXIDATION PROCESS
1. Main Process Vent
A carbon adsorption system or an incineration system can be used to effectively
control process emissions. It is assumed that the vent from the refining
vacuum system is combined with the main process vent. Therefore treatment of
the product recovery absorber vent stream will control all process emissions.
a. Carbon Adsorption In order to use carbon adsorption the exhaust gas stream
must be scrubbed with caustic to remove organic acids and water-soluble organics.
Benzene is essentially the only remaining VOC. The stream is then conditioned
by heating to reduce the relative humidity. Three carbon beds are specified for
the model plant. The exhaust stream passes through two parallel beds while the
other bed is being regenerated with steam. The steam condensate is decanted to
separate the benzene for recycle to the process, and the benzene-saturated
aqueous layer is recycled to the product recovery absorber.
Based on engineering experience with similar applications it was concluded that
a carbon adsorption system can be designed and operated at a sustained removal
efficiency of greater than 99%. A removal efficiency of 99% was used to project
the final emissions from the controlled model plant (Table V-l). The cost of
this system is shown in Table VI-1 in Sect. VI. One carbon adsorption system
is currently being used to recover benzene from an MA operation, but reportedly
has achieved only 85% removal efficiency.1
The application of carbon adsorption to control VOC emissions is discussed in
the control device evaluation report on carbon adsorption 2
b. Incineration The direct fire incineration system for the model-plant waste
gas stream has a knockout demister tank to protect the incinerator by preventing
liquid stream from reaching the firing area, an incinerator with a combustion
chamber of sufficient volume to give a retention time of 0.75 sec, and a stack
designed for a velocity of 10 m/sec.3 Supplemental natural gas and makeup air
are required to maintain the necessary combustion temperatures. Heat recovery
-------
Table V-l. Benzene and Total VOC from Emissions Controlled by Carbon Adsorption in the
Production of Maleic Anhydride in Model Plant (22,700-Mg/yr Capacity)
Emission Source
Stream
Designation
(Fig. IV-1)
Control Device
or Technique
Emission
Reduction
Emission Ratio
(kg/kg X 10~5)a'b
Benzene Total VOC
Emission Rate
(kg/hr)
Benzene
Total VOC
Product recovery
absorber
Refining vacuum system
Q
Storage and handling
Fugitive
Secondary
Total
B
C
C
C
J
K
Carbon adsorption 99
Vent through absorber 99
Floating-roof tanks 85
(benzene)
Scrubber (MA) 99+
Xylene (none) 0
Detection and correc-
tion of major leaks
None
67
18
21
Trace
67
18
34
11
106 X 10~5 130 X 10 5
1.73
0.48
0.54
2.75
1.73
0.48
0.88 <
K)
0.3
3.39
g of emission per kg of maleic anhydride produced.
The VOC emission from carbon adsorption is essentially benzene since all other VOC are removed by the caustic scrubber
prior to the carbon adsorber.
pproximately 65% of the storage emissions does not vary with production rate.
Fugitive emissions do not vary with production rate.
-------
V-3
can be used for steam generation or for preheating the feedstream and thereby
reducing natural-gas requirements. Since the waste gas contains corrosive
organic acids, stainless steel is specified for construction ahead of the
combustion chamber.
Based on engineering experience with similar incineration applications it was
concluded that a properly designed and operated incinerator will result in
sustained benzene and VOC removal efficiencies of greater than 99%. A removal
efficiency of 99% was used to project final emissions from the controlled model
plant (Table V-2).
A temperature of 87l°C (1600°F) is specified to ensure complete combustion of
the waste gas. While it is conceivable that greater than 99% VOC removal could
be obtained at lower temperatures, it cannot be dependably predicted. This
determination is consistent with government air-pollution engineering manuals.3,4
Although the manuals contain no data on combustion temperatures above SOO°C,
extrapolation of the data presented combined with similar incineration experience
justifies the projection of greater than 99% removal at 871°C. The high carbon
monoxide content of the waste gas must be considered. Similar incineration
experience indicates that the high temperature specified is necessary to obtain
acceptable oxidation of the carbon monoxide, as well as complete VOC oxidation.
The cost of incorporating this system is shown in Table VI-3.
An incineration system used for VOC removal from an MA process reportedly
operates with a combustion temperature of 760°C and achieves a removal effic-
iency of 93%.5 An incinerator designed for 871°C can be operated at lower
temperatures if it is determined by operating experience that a lower tempera-
ture will still provide adequate removal of benzene, VOC, and carbon monoxide.
The application of incineration for VOC emission control is covered in a control
device evaluation report for thermal oxidation.6
c. Catalytic Incineration One company is using a catalytic incinerator to control
emissions from the product recovery absorber. Very little design data are
available. The maximum practical achievable VOC removal efficiency that is
-------
Table V-2. Benzene and Total VOC from Emissions Controlled by Incineration in the
Production of Maleic Anhydride in Model Plant (22,700-Mg/yr Capacity)
Emission Source
Product recovery
absorber
Refinery vacuum vent
Storage and handling
c
Fugitive
Secondary
Total
Stream
Designation
(Fig. 2)
A
B
C
C
c
J
K
Emission
Control Device Reduction
or Technique (%)
Incineration
Vent through incin-
erator
Floating-roof tank
(benzene)
Scrubber (MA)
Xy lene (none )
Detection and correc-
tion of major leaks
None
99 \
99 '
85 )
J
99 '
Emission Ratio Emission Rate
(kg/kg X 10 5)a (kg/hr)
Benzene Total VOC Benzene Total VOC
67 86 1.73 2.24
18 18 0.48 0.48
<
21 34 0.54 0.88 *
11 0.3
106 146 2.75 3.89
akg of emission per kg of maleic anhydride produced.
Approximately 65% of the storage emissions does not vary with production rate.
°Fugitive emissions do not vary with production rate.
-------
V-5
achievable is reported to be less than 95%. The high catalyst volume or
high-temperature requirement makes the unit uneconomical for greater removal
efficiencies.7
The application of catalytic oxidation for VOC emission control is covered by a
control device evaluation report for catalytic oxidation.8
d- n-Butane Process Since the n-butane oxidation process has no potential benzene
emissions, conversion to the n-butane process for MA production is an option
for benzene emission control. No data are presently available on the control
device options for the n-butane process. Also, no data are currently available
for comparing the economics of MA production by n-butane oxidation with that by
benzene oxidation.
2- Refining Vacuum Vent
The refining vacuum vent is controlled by joining the waste stream ahead of the
product-recovery-absorber control device or by joining the product-recovery-
absorber feed stream. The incremental costs are relatively small since only
piping additions are required, and no added utilities, manpower, or other
operating costs are involved. Emissions from the refining vacuum system vent
are included in all control system calculations.
Fugitive Sources
Controls for fugitive sources are discussed in a separate EPA report9 covering
fugitive emissions for the entire synthetic organic chemicals manufacturing
industry. Control of emissions from pumps and valves can be attained by an
appropriate leak detection system followed by repair maintenance. Controlled
fugitive emissions have been calculated with the use of the factors given in
Appendix C and are included in Tables V-l and V-2. The factors are based on
the assumption that major leaks are detected and corrected.9
Storage and Handling Sources
*• Benzene storage—Control of benzene storage emissions is covered in a separate
EPA document.10 Information on MA manufacturing locations indicates that benzene
is stored in floating-roof tanks at three locations and in fixed-roof tanks at
-------
V-6
the others.7 A floating roof* is commonly used to control storage-tank emissions
for VOCs in the vapor pressure range of benzene and is used in the model plant
instead of fixed-roof tanks. Controlled storage emissions were calculated by
assuming that a contact type of internal floating roof with secondary seals
would reduce fixed-roof-tank emissions by 85%.ll
An alternative control method may be possible in MA plants: the fixed-roof
storage-tank vents could be tied into the main-process-vent control. The
resulting emission reduction should be 99%, which is better than that of a
floating-roof tank. However, this modification may add a safety hazard because
benzene storage vapors may be in the flammable range.
Maleic Anhydride Storage and Handling The vapor pressure of MA at storage
conditions in the model plant is 550 to 1170 Pa,- this results in calculated
average emissions from fixed-roof tanks of only 0.3 kg/hr. At two production
plants the tank vents (as.well as air handling from the briquetting operation)
are treated in aqueous scrubbers.1,12 The effluent water is discharged to the
sewer. The scrubbers are primarily for plant housekeeping to prevent solid MA
buildup rather than for emission reduction. It is estimated that scrubber
efficiency under these conditions should be at least 99%. The controlled VOC
emissions given in Tables V-l and V-2 were calculated based on this efficiency.
At one plant the crude maleic anhydride is stored at 130°C and the vent is
treated in a xylene scrubber with a stated efficiency of 50%.13 The xylene ef-
fluent can be recycled to the process for MA recovery. This control option may
not be adequate without an aqueous scrubber to prevent the resulting MA emission
from creating a housekeeping problem associated with a condensed MA solid.
Xylene Storage The calculated emissions from xylene storage in fixed-roof
tanks for the model plant average less than 0.03 kg/hr, which is insignificant
relative to other emissions from the process; therefore it was assumed that
control is not needed.
^Consist of internal floating covers or covered floating roofs as defined in
API-2519, 2d ed., 1976 (fixed-roof tanks with internal floating device to
reduce vapor loss).
-------
V-7
Secondary Sources
Secondary emissions are also insignificant. In plants where a liquid incinerator
is used to destroy organic effluents the design must be adequate to ensure
essentially complete combustion.
Current Emission Control
The control devices currently being used by domestic MA producers are given in
Appendix E.
OTHER PROCESSES
Data are not currently available to size emission control devices for other
processes. The only other significant domestic process, n-butane oxidation, is
considered to be a control option for benzene emissions.
-------
V-8
C. REFERENCES*
1. J. L. Lawson, IT Enviroscience, Inc., Trip Report for Visit to Reichhold
Chemicals, Inc., Morris, IL., July 28, 1977 (on file at EPA, ESED, Research
Triangle Park, NC) .
2. H. S. Basdekis, IT Enviroscience, Inc., Control Device Evaluation. Carbon
Adsorption (in preparation for EPA, ESED, Research Triangle Park, NC).
3. Air Pollution Engineering Manual, U.S. Environmental Protection Agency, pp
and 181 (May 1973) .
4. Ibid., p 709.
5. Personal communication Nov. 17, 1977, between J. F. Lawson, IT Enviroscience,
Inc., and R. E. Hinkson, Denka USA.
6. J. W. Blackburn, IT Enviroscience, Inc., Control.Device Evaluation. Thermal
Oxidation (July 1980) (EPA/ESED report, Research Triangle Park, NC).
7. W. A. Lewis, Jr., G. M. Rinaldi, and T. W. Hughes, Monsanto Research Corp.,
Source Assessment: Maleic Anhydride Manufacture (received January 1978).
8. J. A. Key, IT Enviroscience, Inc., Control Device Evaluation . Catalytic
Oxidation ( in preparation for EPA, ESED, Research Triangle Park, NC).
9. D. G. Erikson, IT Enviroscience, Inc., Fugitive Emissions (in preparation for
EPA, ESED, Research Triangle Park, NC).
10. D. G. Erickson, IT Enviroscience, Inc., Storage and Handling (in preparation
for EPA, ESED, Research Triangle Park, NC).
11. Letter dated Aug. 15, 1979, from William T. Moody, TRW, Inc., to David Beck,
EPA.
12. Houdry Division of Air Products, Inc., Engineering and Cost Study of Air
Pollution Control for the Petrochemical Industry-Maleic Anhydride, prepared f°r
EPA.
13. Permit Exemption Request from R. E. Pruessner, Petro-Tex Chemical, to Charles
Borden, Texas Air Control Board, Feb. 23, 1976.
*Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of that
paragraph, that reference number is indicated on the material involved. When
the reference appears on a heading, it refers to all the text covered by that
heading.
-------
VI-1
VI. IMPACT ANALYSIS
A- CONTROL COST IMPACT
The purpose of this section is to present estimated costs and cost-effectiveness
(CE) ratios for the control of benzene and total VOC emissions resulting from
the production of maleic anhydride. Details of the model plant (Fig. III-l)
have been covered in Sects. Ill and IV. Criteria have been established for
control of the VOC process emissions (main process vent combined with refining
vacuum) by carbon adsorption and incineration. The capital and annual costs
presented for these control systems were obtained from the control device
evaluation report for carbon adsorption1 and for thermal oxidation.2
Basis for Capital Cost Estimate The capital cost estimates represent the
total investment required to purchase and install all equipment and material to
provide a complete control system performing as defined for a new installation
at a typical location. These estimates do not include the cost of MA production
lost during installation, the costs for research and development, or the cost
for purchase of the land required.
Basis for Annual Cost Estimates Estimates for annual costs for control alter-
natives include utilities, operating labor, maintenance supplies, chemicals or
raw materials, recovery credit, capital recovery, and miscellaneous recurring
costs such as taxes, insurance, and administrative overhead. Recovery credits
are based on the market value for the material recovered. Chemical or other
raw-material costs are based on the market value for the material required.
Carbon loading <6 kg of VOC/100 kg of carbon
Steam for regeneration 20 kg/kg of organic adsorbed
Steam for gas conditioning [825 kg/hr (1819 lb/hr)]
50% caustic solution for scrubber 37 kg/hr
Granular activated carbon Replaced every 5 years
(initial charge nearly 3
times the minimum charge
required)
Gas velocity 30.5 m/min (100 fpm)
Bed depth 0.9 m (3 ft)
-------
VI-2
Pressure drop
Carbon
Benzene recovery credit
Benzene removal efficiency
5,257 Pa/m (6.5 in H20/ft)
4 X 10 mesh BPL carbon, 480 kg/m"
(30 lb/ft3)
$0.22/kg ($0.10/lb)
99%
The cost estimates and cost effectiveness for the control of benzene and total
VOC emissions from the model-plant main process vent and refining vacuum vent
are shown on Table VI-1. The cost factors used to compute the annual costs are
shown in Table VI-2. A sample of the calculations used in conjunction with the
control device report for carbon adsorption1 are shown in Appendix D. All cost
estimates are adjusted to December 1979.
Corrosion-resistant type 316 stainless steel is required throughout the caustic
scrubber. The costs of the scrubber and heater are included in the cost estimate.
It is assumed that all the nonbenzene VOC is removed by the caustic scrubber
pretreatment. Nearly identical capital and cost-effectiveness figures have
been reported for an existing carbon adsorption system operating at 85% VOC
removal efficiency.3 These data indicate that the cost effectiveness is not
appreciably less for a system operating at a lower efficiency. Costs of systems
for removal efficiencies of more than 99% are not included because such higher
efficiencies would be practically impossible to obtain.
Figure VI-1 is a plot of the installed capital and net annual costs for carbon
adsorption systems developed from the model-plant calculations and of costs
calculated for a plant with one-half the capacity of the model plant plus a
plant with a capacity that is 50% larger than that of the model plant. The
calculated results for all three plant sizes are listed in Appendix D.
Incineration Estimated costs are based on the incineration system described
in Sect. V. A well-designed and operated incinerator will have a VOC removal
efficiency greater than 99%. The cost savings of operating at a lower efficiency
do not appear to be justified. Higher removal efficiencies might increase the
cost significantly because of the disproportionately large increase in combustl
temperature and residence time required. One incinerator was reported to
operate at 760°C with a VOC removal efficiency of 93%.4
-------
VI-3
Table VI-1. Carbon Adsorption System Emission Control Cost
Estimate for Model Plant (Main Process Vent and
Refining Vacuum Vent)
Installed capital (excluding pretreatment) $ 780,000
Caustic scrubber pretreatment installed capital 542,000
Total installed capital $1,322,000
Utilities, per year 258,000
Raw materials and chemicals, per year 80,000
Manpower, per year 24,000
Capital recovery, per year 383,000
Benzene recovery credit, per year (330,000)
Net annual cost $415,000
Benzene emissions reduced, Mg/yr 1,500
Total VOC emissions reduced, Mg/yr 1,947
Cost effectiveness per Mg of benzene reduced 277
Cost effectiveness per Mg of total VOC reduced 213
-------
VI-4
Table VT-2. Cost Factors Used to Compute Annual Costs
Operating factor
Operating labor
8760 hr/yr
$15/hr
Maintenance labor plus materials, 6%
Capital recovery, 18% (10-yr life
at 12% interest)
Taxes, insurance, administration, 5%
29% of installed capital
Utilities
Electric power
Steam
Cooling water
Natural gas
Heat recovery credits (equivalent
to natural gas)
Activated carbon (replacement)
$8.33/GJ ($0.03/kWh)
$5.50/Mg ($2.50/thousand Ib or
million Btu)
$0.026/m3 ($0.10/thousand gal)
$1.90/GJ ($2.00/million Btu)
?1.90/GJ (§2.00/million Btu)
$2.58 kg ($1.17/lb)
-------
VI-5
5000
1000
o
o
o
CO
4J
V)
0
u
1000
jg ^Includes capital
a.
H
0)
1
cost recovery
456
8 9 10
20
30
40
Waste-Gas Flow Rate (m /s)
Fig. VI-1. Cost vs Waste-Gas Flow Rate - Carbon Adsorption
-------
VI-6
Heat recovery from incineration can be used to generate steam or to preheat the
incinerator feed stream to reduce supplemental fuel requirements. For this
comparative study it was assumed that 50% of the energy in the exit stream is
recovered. Credit for the value of this heat is computed to be equivalent to
natural gas at $1.90/GJ, not for the steam that may be generated. If the
recovered heat is used to preheat the waste gas feed stream, the conservative
assumption is made that the natural gas (utilities) requirement can be reduced
by the full value of the heat recovered. For this preliminary estimate it was
assumed that the capital costs for a preheat system or a steam generation
system to augment an existing system are equivalent.
The basis for the estimated costs for the incineration system described in
Sect. V are as follows:
Removal efficiency 99%
Waste-gas heat content 484 kJ/m (13 1
Combustion temperature 871°C (1600°F)
Waste-gas flow 16 m /sec (34,(
Residence time 0.75 sec
Cost-effectiveness estimates were obtained for the model plant from the control
device evaluation report for thermal oxidation,2 Appendix B, p B-18. The cost
factors used to compute the annual costs are shown in Table VI-2. The estimated
December 1979 installed capital costs, net annual costs (including capital
recovery), and cost effectiveness per Mg of VOC removed are shown in Table VI-3.
Figure VI-2 is a plot of the installed capital and net annual costs for incinera-
tion systems developed from the model-plant estimates plus projected costs for
a plant with one-half the capacity of the model plant plus a plant with a
capacity that is 50% larger than that of the model plant. The costs and cost
effectiveness projected for all three plant sizes are listed in Appendix D.
Figure VI-3 is a plot of the same information for an incinerator system with
50% heat recovery.
Cost Effectiveness Figure VI-4 is a plot of cost effectiveness versus capacity
for each of the benzene and total VOC removal control systems considered. The
data for Fig. VI-4 are listed in Tables D-l and D-2.
-------
VI-7
Table VI-3. Incineration System Emission
Control Cost Estimate for Model Plant
(Main Process Vent and Refining Vacuum Vent)
No Heat Recovery 50% Heat Recovery
Installed capital
Net annual costs (including
capital recovery)
Benzene emission reduced, Mg/yr
Total VOC emissions reduced, Mg/yr
Cost effectiveness per Mg of benzene
reduced
Cost effectiveness per Mg of total VOC
reduced
$980,000
$2,200,000
1500
1943
$1467
$1132
$1,300,000
$700,000
1500
1943
$467
$360
-------
VI-8
5000
o
o
o
rH
•CO-
CO
-p
CO
o
u
1000
100
1 I J
*Includes capital
rHj cost recovery
1
8 9 10
20
30
40
Waste-Gas Flow Rate. Prior to Incineration (m /s)
Fig. VI-2. Cost vs Waste-Gas Flow Rate - Incineration
Without Heat Recovery
-------
5000
VI-9
o
o
o
to
-p
en
o
U
1000
100
I I I
8 9 10
*Includes capital
cost recovery
4J
0>
a
20
30
40
Waste-Gas Flow Rate (m /s)
Fig. VI-3.
Cost vs Waste-Gas Flow Rate - Incineration
with 50% Heat Recovery
-------
VI-10
3000 i—
a
t/>
in
W
Q)
C
0)
>
•H
4J
U
cu
W
4J
0)
o
U
Benzene
1000
100
Incineration with-
out heat recovery
Incineration with
50% heat recovery
Carbon adsorption
I
I I I I
10
Waste-Gas Flow Rate (m /s)
Fig. VT-4. Cost Effectiveness vs Waste-Gas Flow Rate
100
-------
VI-11
Fugitive Emissions Fugitive emissions factors are listed in Appendix C, A
fugitive emissions report covers the applicable fugitive emissions controls for
all the synthetic organic chemicals manufacturing industry.5
Storage and Handling Storage control costs and cost effectiveness are covered
by a storage and handling report for the entire synthetic organic chemicals
manufacturing industry.6
Secondary Emissions No significant secondary emissions sources exist for the
model plant; therefore no control is required.
Other Processes No data are available for determining the cost impact of
converting to the n-butane process as a benzene emission control option. Data
are also unavailable for determining the cost of any control devices required
to control emissions from an n-butane oxidation process.
ENVIRONMENTAL AND ENERGY IMPACTS
Table VI-4 shows the environmental impact of reduced benzene and VOC emissions
by application of the described control systems to the model plant. From an
energy standpoint a typical uncontrolled MA process will produce a heat surplus
of approximately 15 kJ/kg of MA.7 Control device environmental and energy
impacts are discussed below.
Carbon Adsorption
The carbon adsorption system described to control process emissions from the
model plant will reduce benzene emissions by 1500 Mg/yr and total VOC emissions
by 1947 Mg/yr. Energy (steam) is required to condition and desorb the benzene
from the carbon. For the model plant this energy (as steam) is 56,000 MJ/Mg of
benzene emissions reduced and 43,000 MJ/Mg of total VOC removed. For the total
domestic MA production capacity this is equivalent to 800 million MJ/yr. The
benzene recovered is recycled to the process rather than being used as fuel.
The electrical energy required is 2300 MJ/Mg of benzene emissions reduced and
1700 MJ/Mg of VOC removed. For the total domestic MA production capacity this
is equivalent to 33 million MJ/yr.
-------
Table VI-4. Environmental Impact - Model Plant Emission Controlled
Emission Source
Main process vent
Refining vacuum vent
Storage
b
Benzene
MA
Xylene
Fugitive0
Secondary
Stream
Aj
B '
C
C
C
J
K
Control Device
or Technique
Carbon adsorption or
Incineration
Floating-roof tank
Scrubber
None
Detection and correction
of major leaks
None
Emission
Reduction
99
99
85 *\
99 >
J
^65
Emissions Controlled
(Mg/yr)
Benzene
1500
1500
23.8
20.9
1545
Total VOC
19473
1943
24.7
20.6
1993a
1989
VOC removal with narbon adsorption, which is slightly higher because of the VOC removed by the caustic
scrubber.
Approximately 65% of the storage emissions does not vary with production rate.
i^
"Fugitive emissions do not vary with production rate.
to
-------
VI-13
Incineration
The incineration system described to control process emissions from the model
plant will reduce benzene emissions by 1500 Mg/yr and total VOC emissions by
1943 Mg/yr. No organics are recovered for recycle to the process with incin-
eration. Stack gases from incineration can have a negative impact on the
environment (NO , C02, CO), particularly if the carbon monoxide is not ade-
X
quately oxidized.
Supplemental fuel is required to maintain suitable operating conditions. The
net amount of energy required for the model plant ranges from 666 GJ/Mg of
benzene removed for an incinerator without heat recovery to 70 GJ/Mg for an
incinerator with 50% heat recovery. For total VOC removal the net energy
requirement ranges from 514 GJ/Mg without heat recovery to 54 GJ/Mg with 50%
heat recovery. The net energy required for all domestic MA production projects
ranges from 9.4 million GJ/yr without heat recovery to 1.0 million GJ/yr with
50% heat recovery. Because the potential exists for some companies to have
excess steam on-site, it may not be practical for all companies to utilize the
heat recovery option,
Other Emissions (Fugitive, Storage, and Secondary)
The control methods described for emissions from these sources are floating-
roof storage tanks, scrubbing for product vents, and correction of leaks for
fugitive emission control. Application of these systems results in a VOC
emission reduction of 45 Mg/yr for the model plant. The energy impact result-
ing from application of these systems is minimal because the only energy required
is electricity for pumps.
-------
VI-14
C. REFERENCES*
1. H. S. Basdekis, IT Enviroscience, Inc., Control Device Evaluation. Carbon
Adsorption (in preparation for EPA, ESED, Research Triangle Park, NC).
2. J. W. Blackburn, IT Enviroscience, Inc., Control Device Evaluation. Thermal
Oxidation (July 1980) (EPA/ESED report, Research Triangle Park, NC).
3. J. F. Lawson, IT Enviroscience, Inc., Trip Report for Visit to Reichhold
Chemical, Inc., Morris, IL, July 28, 1977 (data on file at EPA, ESED, Research
Triangle Park,, NC).
4. Personal communication Nov. 17, 1977, between J. L. Lawson, IT Enviroscience,
Inc., and R. E. Hinkson, Denka USA.
5. D. G. Erickson, IT Enviroscience, Inc., Fugitive Emissions (in preparation for
EPA, ESED, Research Triangle Park, NC).
6. D. G. Erickson, IT Enviroscience, Inc., Storage and Handling (in preparation
for EPA, ESED, Research Triangle Park, NC).
7. P. L. Morse, Maleic Anhydride, Interim Al, 46A1, A private report by Process
Economic Program, Standford Research Institute, Menlo Park, CA (November 1973).
*Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of that
paragraph, that reference number is indicated on the material involved. When
the reference appears on a heading, it refers to all the text covered by that
heading.
-------
VI I-1
VII. SUMMARY
Maleic anhydride is produced in the United States predominantly by catalytic
oxidation of benzene. In recent years producers have been looking very care-
fully at the feasibility of n-butane as a feedstock. Late in 1976 Amoco began
operation of a 27,000-Mg/yr plant using butane feedstock. In 1974 Monsanto
also dedicated 20% of its capacity toward development of butane technology and
during 1979 announced expansion of their butane facilities. Denka USA has
notified the regulatory agencies of its intent to engage in butane process
development. Approximately 83% of the 214,000-Mg/yr domestic capacity is based
on benzene oxidation. Butane oxidation accounts for another 15% of capacity,
and the remaining 2% is produced as a by-product of phthalic anhydride manu-
facture.
The maleic anhydride annual growth rate is estimated to be 8% through 1982.
Based on 1977 domestic MA production capacity there is sufficient capacity at
present to meet the growth rate through 1982. No shortage of either benzene or
n-butane is expected during this period.
The emission sources and control levels for the model plant are summarized in
Table VII-1. The predominant emission points are the main process vent and
benzene storage facilities. Control of these sources alone can reduce emissions
by 99%.
The model-plant main process vent emits benzene at a rate of 173 kg/hr and
total VOC at 224 kg/hr. These process emissions can be controlled by either
carbon adsorption or incineration with a destruction efficiency of 99%. The
installed capital cost to control process emissions from the model plant ranges
from $980,000 for an incineration system without heat recovery to $1,300,000
for an incinerator system with heat recovery. The carbon adsorption installed
capital cost is estimated to be $780,000. The net cost effectiveness per Mg of
reduction is as follows:
Benzene (per mg) Total VOC (per mg)
Carbon adsorption $ 277 $ 213
Incineration without heat recovery 1467 1132
Incineration with 50% heat recovery 467 360
-------
VI I-2
Table VII-1. Summary of Uncontrolled and Controlled
Emissions from Model Plant (22,700-Mg/yr Capacity)
Emission Rate (kg/hr)
Uncontrolled
Emission Source
Process
Storage and handling
Fugitive
Secondary
Total
Benzene
173
3.2
2.93
179
VOC
224
3.3
3.23
0.3
230
Controlled
Benzenea
1.73
0.48
0.54
2.8
VOC
2.24b
0.48
0.88
0.3 .
3.9
1.73C
0.48
0.88
0.3
3.4
aBenzene controlled emissions are identical for carbon adsorption and incinera-
tion.
Based on incineration option.
CBased on carbon adsorption; VOC emissions are lower because of the VOC re-
moved by caustic scrubbing prior to carbon adsorption.
-------
VII-3
The carbon adsorption system requires an estimated 84 million MJ of steam
energy per year for the model plant and 3 million MJ of electrical energy per
year. The incineration system requires an estimated 1000 million MJ of supple-
mental fuel energy per year for an incineration system without heat recovery
and a net requirement of 105 million MJ/yr for a system with 50% heat recovery.
With either control system the vacuum vent system is tied to the control device
feed and all estimates include these emissions.
Benzene storage can be controlled by use of a floating-roof tank with an emis-
sion reduction of 85% of the fixed-roof-tank emissions.
A carbon adsorption system is currently being used to recover benzene from an
MA operation, but reportedly has achieved only 85% efficiency.1 One MA producer
using incineration for VOC emission control reports a removal efficiency of
93%.2
Since the n-butane oxidation process has no inherent benzene emissions, it is
an option for benzene emission control. However, data regarding the potential
economic incentive for switching to n-butane oxidation are not presently avail-
able. It's increased popularity is evidenced by the fact that a growth rate of
24.3% is projected for n-butane oxidation through 1982 compared to 8% for total
maleic anhydride production.
1J. F. Lawson. IT Enviroscience, Inc., Trip Report for Visit to Reichhold
Chemicals, Inc.. Morris, IL, July 28, 1977 (on file at EPA, ESED, Research
Triangle Park, NC).
2Personal communication Nov. 17, 1977, between J. F. Lawson, IT Enviro-
science, Inc., and R. E. Hinkson, Denka USA.
-------
A-l
APPENDIX A
Table A-l. Physical Properties of Maleic Anhydride
Synonyms Toxilic anhydride,
cis-butene dioicanhydride,
2,5-furandione
Molecular formula C H 0
Tt <& J
Molecular weight 98.06
Physical state Solid
Vapor pressure 1 mm at 44°C
Vapor density 3.4
Boiling point 197—199°C
Melting point 60°C
Density 1.48 at 20°C/4°C
Water solubility Reacts with water
a
J. Dorigan, B. Fuller, and R. Duffy, "Maleic Anhydride," pp AIII-8 in
Scoring of Organic Air Pollutants. Chemistry, Production and Toxicity of
Selected Organic Chemicals (Chemicals f-n), MTR-7248, Rev 1, Appendix III,
MITRE Corp., McLean, VA (September 1976).
The Merck Index, 8th ed., Merck & Co., Rahway, NJ, 1968.
Table A-2. Physical Properties of Benzene*
Synonyms Benzol, phenylhydride,.coal naphtha
Molecular formula C H
6 6
Molecular weight 78.11
Physical state Liquid
Vapor pressure 95.9 mm at 25°C
Vapor density 2.77
Boiling point 80.1°C at 760 mm
Melting point 5.5°C
Density 0.8787 at 20°C/4°C
Water solubility Slight (1.79 g/liter)
a
J. Dorigan, B. Fuller, and R. Duffy, "Benzene," p AI-102 in Scoring of Organic
Air Pollutants. Chemistry, Production and Toxicity of Selected Organic
Chemicals (Chemicals a-c), MTR-7248, Rev 1, Appendix I, MITRE Corp.,
McLean, VA (September 1976).
-------
A-2
Table A-3. Physical Properties of Formic Acid
Synonyms Methanoic acid, hydrogen carboxylic
acid
Molecular formula CHo°o
Molecular weight 46.3
Physical state Liquid
Vapor pressure 42.38 mm at 25°C
Vapor density 1.59
Boiling point 100.8°C
Melting point 8.3°C
Density 1.2201 at 20°C/4°C
Water solubility Soluble
J. Jorigan, B. Fuller, and R. Duffy, "Formic Acid," p AIII-16 in Scoring of
Organic Air Pollutants. Chemistry, Production and Toxicity of Selected Organic
Chemicals (Chemicals f-n), MTR-7248, Rev 1, Appendix III, MITRE Corp.,
McLean, VA (September 1976).
-------
A-3
Table A-4. Physical Properties of Formaldehyde
Synonyms Methanol, methyl aldehyde, forma-
lin
Molecular formula CH O
Molecular weight 30.03
Physical state Gas or liquid
Vapor pressure 1946.67 mm at 25°C
Vapor density Not given
Boiling point -21°C at 760 mm
Melting point -92°C
Density 0.815 at 20°C/4°C
Water solubility Soluble
3J. Dorigan, B. Fuller, and R. Duffy, "Formaldehyde," p AIII-12 in Scoring of
Organic Air Pollutants. Chemistry, Production and Toxicity of Selected
Organic Chemicals (Chemicals f-nj^ MTR-7248, Rev 1, Appendix III, MITRE,
Corp., McLean, VA (September 1976).
Table A-5. Physical Properties of Maleic Acida
Synonyms Maleinic acid, toxilic acid,
cis-butenedioic acid
Molecular formula C,H.O.
444
Molecular weight 116.07
Physical state Solid
Vapor pressure Essentially zero
Vapor density 4.0
Boiling point 135°C; decomposes
Melting point 139 to 140°C
Density 1.590 at 20°C/4°C
Water solubility Very soluble
3J. Dorigan, B. Fuller, and R. Duffy, "Maleic Acid," p AIII-116 in Scoring of
Organic Air Pollutants. Chemistry, Production and Toxicity of Selected
Organic Chemicals (Chemicals f-n), MTR-7248, Rev 1, Appendix III,
MITRE Corp., McLean VA (September 1976).
-------
B-l
APPENDIX
Table B-l. Air-Dispersion Parameters for
— Source
Process emissions
(uncontrolled)
Benzene
Total VOC
Incinerator
(with heat recovery)
Benzene
Total VOC
carbon adsorber
Benzene
Total VOC
Storage and handling emissions
Uncontrolled
Benzene
Maleic anhydride
Xylene
Storage and handling emissions
Controlled
Benzene
pi
Maleic anhydride
Xylene
fugitive emissions13
Uncontrolled
Benzene
Total von
Emission
Rate
(g/sec)
48.1
61.9
0.48
0.62
0.48
0.48
0.41
0.04
0.04
0.02
0.02
0.01
0.01
0.01
Trace
0.01
0.81
0.90
Height
(m)
27.4
14.0
15.2
12
6
9
6
9
6
12
6
9.1
6
B
Model Plant (22 , 700-Mg/yr capacity)
Discharge
Diameter Temp.
(m) (K)
1 311
2 533
1.5 333
16 Ambient
4 Ambient
7 350
6 344
7 333
5 Ambient
16 Ambient
4 Ambient
0.6 298
5 Ambient
Flow Discharge
Rate Velocity
(m3/sec) (m/sec)
20.3 25.8
31.4 10.0
18.0 10.2
3.34 11.8
-------
B-2
Table B-l. (Continued)
Source
Emission Discharge Flow Discharge
Rate Height Diameter Temp. Rate Velocity
(g/sec) (m) (m) (K) (m /sec) (m/sec)
Fugitive emissions (cent.)
Controlled
Benzene
Total VOC
Secondary emissions
(uncontrolled)
(total VOC)
0.15
0.24
0.08
Ambient
Storage scrubber.
fugitive emissions are evenly distributed over a rectangular area 150 m by 30 m.
-------
C-l
APPENDIX C
FUGITIVE-EMISSION FACTORS*
The Environmental Protection Agency recently completed an extensive testing
program that resulted in updated fugitive-emission factors for petroleum re-
fineries. Other preliminary test results suggest that fugitive emissions from
sources in chemical plants are comparable to fugitive emissions from correspond-
ing sources in petroleum refineries. Therefore the emission factors established
for refineries are used in this report to estimate fugitive emissions from
organic chemical manufacture. These factors are presented below.
Source
Pump seals ^
Light-liquid service
Heavy-liquid service
Pipeline valves
Gas/vapor service
Light-liquid service
Heavy- liquid service
Safety/relief valves
Gas/vapor service
Light-liquid service
Heavy-liquid service
Compressor seals
Flanges
Drains
Uncontrolled
Emission Factor
(kg/hr)
0.12
0.02
0.021
0.010
0.0003
0.16
0.006
0.009
0.44
0.00026
0.032
Controlled
Emission Factor
< kg/hr)
0.03
0.02
0.002
0.003
0 . Ou"03
0.061
0.006
0.009
0.11
0.00026
0.019
a
Based on monthly inspection of selected equipment; no inspection of
heavy-liquid equipment, flanges, or light-liquid relief valves;
10,000 ppmv VOC concentration at source defines a leak; and 15 days
allowed for correction of leaks.
bLight liquid means any liquid more volatile than kerosene.
*Radian Corp Emission Factors and Frequency of Leak Occurrence for Fittings
in Refinery Process Units, EPA 600/2-79-044 (February 1979).
-------
D-l
APPENDIX D
DETAILED COST ESTIMATES AND CALCULATIONS FOR CARBON ADSORPTION INCINERATION
This appendix contains an explanation and sample calculations for the estimated
costs used in this report.
Capital costs shown are based on an accuracy range of +30% to -23%. This range
is a function of the degree of detailed data available when the estimate was
made. The evaluation made in this report is a screening study based on general
design criteria, block flowsheets, approximate material balances, and general
equipment requirements. Figure D-l illustrates the relationship between the
degree of accuracy of an estimated cost and the amount of data available.
The allowance indicated on this chart to cover the undefined scope of the
project has been included in the estimated costs.
This type of estimate is an acceptable basis to provide a screening estimate
to indicate the most cost-effective alternative, within the limits of accuracy
indicated.
-------
ESTIMATE. TYPE.
USED BY ESTIMATOR
SCRE-EKllkiG,
(PRE.LIM. EKIG,. STUDY)
PHASE. H
(PREUM. PROC. EKJ(^.)
PHASE nr
(COMPLETE PROCESS
Efjq. DESI^U')
•
•
•
•
^\
\
. PROS-
CO>
11
'
f
ll
' i
///
"
/ /
/I/'
if'
1
\
/ LI
/'
,
~J
-fcO ~4o -2O O Eo ^O 4>O
RAUGjE. - PROBA.BLE-
ACTUAL. PROJECT
j
/
/
•
'I
I/I
II
I
1
j /
' / t
/ /
•
/
'
O /o zo -50 4c
"/« ALLO\VAkiC£
-TO /MCLUDE,
Co-bT
a
i
Fig. D-l. Precision of Capital Cost Estimates
-------
D-3
A- CARBON ADSORPTION
The following procedure was used to prepare the emission control cost estimates
for Table VI-1. As shown by Table IV-2 the emission stream to be controlled
contains 173 kg of benzene per hour and 224 kg of total VOC per hour. Essentially
all VOC except benzene will be removed by the caustic scrubber pretreatment.
Emissions to Carbon Adsorption (CA) System
173 kg 2.2 Ib hr Ib mole 359 ft3 on „
Benzene = — ___ —- -—g- __ = 29.20 scfm
Air = 75,427 kg/hr = 34,236 scfm
Total waste gas to CA = 34,266 scfm
From Fig. IV-1 of Control Device Evaluation. Carbon Adsorption1 (CA) the
December 1979 installed capital cost for 34,000 scfm is $780,000, including
67,500 Ib of carbon. The pretreatment caustic scrubbing system requires a
stainless steel scrubber 8 ft in diameter by 36» ft high, with 402 ft3 of sad
dles, a stainless steel 100-gpm circulation pump, a 500-gal sump tank, and a
900-ft2 heat exchanger, plus instrumentation for level, pH, and temperature
control. The installed capital estimate for the pretreatment system is
$542,000. Total installed capital cost is $1,322,000.
Annual Cost
Utilities - The following items comprise the cost of utilities:
Regeneration . 3* Ib of bensene g
Conditioning = 1,819 Ib of steam/hr = $ 38,641/yr
Blower electricity (App. A, ref 1) = $4,925 X 3°'°°° = $ 29,550/yr
XH. S. Basdekis, IT Enviroscience, Inc., Control Device Evaluation. Carbon
Adsorption (in preparation for EPA, ESED, Research Triangle Park, NC).
-------
c.
d.
3.
a.
c.
d.
Cooling water =
D-4
4.2 gal 67 M Ib of steam $0.10
Ib of steam yr M gal
= $ 28,140
Total utilities
$258,203
Raw Materials and Chemicals The cost of raw materials and chemicals includes
carbon replacement and caustic:
Carbon replacement = 22'5°Vb i_beds $1^7
bed 5 yr Ib
Caustic =
1b 8760_hr $_CK09
hr yr Ib
Total
Manpower - 1600 hr/yr X $15/hr
Fixed Cost (Including Capital Recovery) - 0.29 X $1,322,000
Benzene Recovery Credit 381 Ib of benzene 8760 hr
hr
yr
X
X 0.99
$0.10
Ib of benzene
Net annual cost
Cost Effectiveness
Benzene Emissions Reduced (Table IV-2) 173 kg/hr X 8760 hr/yr X
Total VOC Reduced 5 kg/hr X 8760 hr/yr
1000
X 0.99 X 1/1000
+ 1500 Mg/yr
Cost Effectiveness of Benzene $415,609/1500 Mg
Cost Effectiveness of VOC $415,609/1947 Mg
= $ 15,795
= $ 64,649
$ 80,444
= $ 24,000
= $383,380
=($330,418)
$415,609*
= 1500 Mg/yr
= 1947 Mg/yr
= $277/Mg
= $213/Mg
Table D-l lists the carbon adsorption control cost estimates for the model
plant plus plants with 50% greater and lesser capacity.
B. INCINERATION
The following procedure was used to prepare the estimated cost projections for
Table VI-3 plus those in Figs. VI-2 and VI-3.
*After the annual cost is adjusted for pretreatment capital recovery cost,
caustic cost, and steam conditioning cost, the net annual cost is in agree-
ment with that of Figs. IV-4 and IV-5 of the carbon adsorption report.
-------
Table D-l. Carbon Adsorption System Emission Control Costs
for Main Process Vent and Refining Vacuum Vent
Costs
Caustic
Installed Scrubber Total
Raw Materials
Emission Reduction Cost Ettoc t iv
(Hq/yr) (per Ma;
Benzene Recovery
Capital*1 Installed Installed Utilities and Chemicals Manpower Fixed Credit Net Annual Benzene Total VOC Benzene Total voc
$480,000 $332,000 $ 812,000 $129,100 $ 40,221
$780,000 $542,000 $1,322,000 $258,203 $ 80,444
$970,000 $641,000 $1,611,000 $387,304 $120,666
17,000-scfm (8 m /s) Plant
$15,000 $235,480 ($165,203)
34,000-scfm (16 m /s) Model Plant
$24,000 $383,000 ($330,418)
51,000-scfm (24 n\3/s) Plant
$29,000 $467,190 ($495,632).
$254,598 750 974
$415,609 1500 1947
$508,528 2250 2921
$339 $261
$277 $213
$17.;
'Excluding pretreatment.
Including capital recovery.
Ct
en
-------
D-6
Waste gas Btu content (model-plant data from Table IV-2)
CO = i'76* k? -^flb hr^ 4.200 Btu = 271,200 Btu/min
hr kg 60 mm Ib
voc = Ea- > L^. , u = 84Q Btu/min
hr kg 60 mm . Ib - • -
419,040
419 '°4° BtU ,A f =12.1Btu/scf
mm 34,266 scf
The procedure used for designing thermal oxidizer systems is described on
p III-7 of the control device evaluation report for thermal oxidation.2 As is
discussed on p III-7 and indicated on Fig. III-3 of that report, all tables in
Appendix B of the report are based on the premise that the waste-gas contains no
oxygen for combustion and that with the waste-gas Btu contents below 80 Btu/scf
the volume of combustion air required is approximately equal to the volume of
waste gas. Since the waste gas from a maleic anhydride process contains more
than enough oxygen required for combustion, no additional combustion air is
required. The cost estimates were projected by using the tables on p B-18 of
the thermal oxidation report and using the costs listed for a waste-gas flow of
one-half the specified flow. The tables on p B-18 apply for all waste gases
with a combustion temperatures of 1600°F, a residence time of approximately
0.75 sec, and a Btu content of approximately 13 per scf. Table D-2 lists the
cost estimates for the model plant plus those for plants with 50% greater and
lesser capacity.
The costs of incineration utilities are attributed essentially to natural gas
used as supplemental fuel. The supplemental fuel requirements discussed in
Sect. VI were also obtained from Table B-18 discussed above. The supplemental
fuel requirement was obtained by transposing the annual utility cost to GJ of
natural gas by using the $1.90/GJ factor from Table VI-2.
2J. W. Blackburn, IT Enviroscience, Inc., Control Device Evaluation. Thermal
Oxidation (July 1980) (EPA/ESED report, Research Triangle Park, NC).
-------
Table D-2. Incineration System Emission Control Costs
for Main Process Vent and Refining Vacuum Vent
Cost
plant Installed Net Annual
34,000-scfm (16 m /s)
model plant
Mo heat recovery S 980,000 $2,200,000
50% heat recovery 1,300,000 700,000
17 ,000-scfni C8 m3/s) plant
No heat recovery S 600,000 $1,100,000
50-% heat recovery 830,000 420,000
51,000-scfm (24 nt /s ) plant
No heat recovery $1,200,000 $2,800,000
50% heat recovery 1,600,000 1,600,000
""includes capital recovery.
Emission Reduction (t',g/yr) Cost Effectiveness (per i;q)
Benzene
1500
1500
750
750
2250
2250
Total VOC Benzene
1941 $1467
1943 467
971 $1467
971 560
2914 $1244
2914 466
Total VOC
$1132
360
$1130
432
$ 961
360
Cost Effectiveness
(per scfin)
$64.7
20.5
$64.7
24.7
$54.9
20.6
D
-------
E-l
APPENDIX E
EXISTING PLANT CONSIDERATIONS
Table E-l1 lists process control devices reported to be in use by industry. To
gather information for the preparation of this report four site visits were
made to manufacturers of maleic anhydride. Trip reports have been cleared by
the companies concerned and are on file at EPA, ESED, in Research Triangle
Park, NC.2—5 Some of the pertinent information concerning process emissions
from these existing maleic anhydride plants is presented in this appendix.
A. PROCESS EMISSIONS FROM EXISTING PLANTS
1- Reichhold Chemicals, Morris, IL
The process was licensed from Lurgi and the plant was built by Badger in 1971.
Variations between the Lurgi process and the model process are the following:
a series of switch condensers serve as primary recovery in lieu of crude-product
separators; a carbon adsorption system recovers benzene from the secondary-prod-
uct-recovery absorber; double-effect evaporators are used to dehydrate the
maliec acid formed during secondary product recovery; and a thermal oxidizer is
used to incinerate the process organic effluents. The total annual cost for
operating all pollution control facilities is $865,000 (1976 dollars).
The carbon adsorption system cycle consists of 2 hr on line and 1 hr of steam
regeneration. The process gas is then vented to the atmosphere. During regene-
ration the benzene-water vapor is condensed and separated. The benzene layer
is recycled to the process, and the water layer is pumped to the process thermal
oxidizer for use as quench water.
The vacuum system from the dehydration and purification sections is vented to
the atmosphere. The vacuum systems are two-stage jets with barometric con-
densers. The water discharge is used for lean-acid scrubber makeup. No vent
emission data are available.
Some key operating parameters associated with the operation of the carbon
adsorption benzene recovery system are as follows:
-------
Table E-l. Control Devices Currently Used by the Maleic Anhydride Industry in the United States
Emission Point
Company
Amoco Chemicals Corp.
Ashland Oil, Inc.
Denka Chemical Corp.
Koppers Company, Inc.
Monsanto Co.
Reichhold Chemical, Inc.
. Elizabeth, NJ
Morris , IL
Tenneco , Inc .
U.S. Steel Corp.
Flaking,
Pelletizing
and Packaging
NR
b
Scrubber
Scrubber
S crubber
Scrubber
b
c
Scrubber
Product
Recovery
Absorber
NR
Scrubber
Incineration
Incineration
c
Carbon adsorber
Carbon adsorber
c
Catalytic incinera-
tor
Vacuum
System
Vent
NR3
NR
Scrubber
NR
b
NR
c
NR
NR
Storage
Tank
Vents
NR
Floating- roof tanks
Floating-roof tanks
Return vents
Scrubber
Conservation vents w
i
Scrubber, conservation
vents
Scrubber, conservation
vents
Floating-roof tanks
Not reported.
Vents to product recovery absorber.
"No control.
-------
E-3
Vapor flow rate, cfm 43,000 at 100°F
Inlet temperature, °C 43 to 46
Efficiency, % Approximately 85
Efficiency range, % 65 to 95
Adsorber vent composition, Ib/hr
Maleic anhydride 15
Benzene 18
Oxygen 31,200
Nitrogen 147,300
Carbon dioxide 7040
Carbon monoxide 3360
Water vapor 4800
Aqueous layer composition from condenser separator
Flow, gpm 20 to 22
Benzene concentration, ppm 2 to 800
Formaldehyde concentration, % 0.3 to 0.6
The above information was extracted from a trip report.2
From a secondary emissions standpoint there may be some need for concern re-
garding the benzene concentration in the water layer being pumped to the thermal
oxidizer for use as quench water. This water could conceivably average approxi-
mately 400-ppm benzene, which would essentially all be stripped from the quench
water and be emitted with the thermal oxidation flue gas.
It is IT Enviroscience's belief that the most logical step to take for improve-
ment of the 85% efficiency being experienced with the carbon adsorption system
would be to further cool the bed before it is put on-line. This would require
the addition of a cooling fan. Further improvement might require an additional
carbon bed. The cost or practicality of this retrofit improvement has not been
studied.
Monsanto, St. Louis, MO3
Variations between the Monsanto air-oxidation process and the model process are
the following: the dehydration column operates at atmospheric pressure rather
than under a vacuum; the xylene stripper is eliminated; and the fractionation
-------
E-4
column vacuum jet goes to a vent scrubber. The process emission sources are
the product scrubber vent and the dehydrator decanter. The product scrubber
has no control device for emission reduction. The dehydrator, which operates
at atmospheric pressure, has a decanter. The decanter separates xylene and
water and has a nitrogen purge of 100 scfh. The dehydrator decanter is reported
to emit 0.00005 Ib of xylene/lb of maleic anhydride produced. The emissions
from the product scrubber are reported as follows:
Emission Amount (Ib/lb of MA Produced)
Benzene 0.062—0.104
Xylene Trace
CO 0.548—0.564
MA Trace
3. Denka Chemical Corp., Houston, TX4
The maleic anhydride production facility is essentially in agreement with the
model plant. It consists of a single train with multiple reactors in parallel.
The facility was built in 1962, with a major expansion in 1971. The emissions
from the dehydration, stripping, and fractionation systems are vented through
steam jets. The steam jets have surface-type condensers. The main process
emission is the product absorber off-gas, which is incinerated.
The incinerator normally operates at 1400°F with a residence time of 0.7 sec.
Hydrocarbon emissions are reportedly reduced by 91% at 1350°F and by 96% at
1500°F. Denka calculates the benzene level in the incinerator feed to be 0.001
to 0.0025 wt %. Information on benzene destruction is not available. Denka
reports that their present fume incinerator cannot be operated at higher tem-
peratures to improve VOC distruction efficiency.6
6. Amoco Corp., Chicago, IL5
The discussion of Amoco's butane-oxidation process at Joliet, IL, was held in
their Chicago office without a tour of the facility. The butane-oxidation
process is reportedly very similar to the benzene-oxidation process from an
emissions standpoint. The one major distinction of the butane-oxidation process
is that, since no benzene is present, there are no benzene emissions. The
-------
E-5
facility consists of a single train with multiple parallel reactors. The plant
was built in 1976.
Product recovery is by partial condensation, product absorption, and fractiona-
tion. The combined process emissions are treated in a thermal oxidizer. Amoco
test analyses of the waste gas to the thermal oxidizer and flue gas from the
thermal oxidizer at 60—65% of production capacity are as follows:
Component Waste Gas Feed (Ib/hr) Flue Gas (Ib/hr)
N2 139,000 139,000
02 37,000 31,500
Water 4,300 6,400
Carbon oxides 2,800 7,500
Organics 1,300 No data
The thermal oxidizer operates at 1500°F.
5- Ashland Chemical Co.7
No site visit was made to Ashland's maleic anhydride plant at Neal, WV. In a
letter submitted by Ashland as a review of the report Ashland stated that the
benzene conversion efficiency of 93.3% displayed for the model plant is well
below that which is achieved at Ashland. They point out that performance of
all existing plants will be upgraded as catalyst development proceeds. They
suggest that 2 to 3% unconverted benzene is a better basis for emissions esti-
mates. Ashland also contends that, with the lower estimate for benzene emissions,
catalytic incineration is an appropriate method of control.
B. RETROFITTING CONTROLS
The primary difficulty associated with retrofitting may be in finding space to
fit the control device into the existing plant layout. Because of the costs
associated with this difficulty it may be appreciably more expensive to re-
trofit emission control systems in existing plants than to install a control
system during construction of a new plant.
-------
E-6
C. REFERENCES*
1. W. A. Lewis, Jr., G. M. Rinaldi, and T. W. Hughes, Monsanto Research Corp.,
Source Assessment: Maleic Anhydride Manufacture (received January 1978).
2. J. F. Lawson, IT Enviroscience, Inc., Trip Report for Visit to Reichhold Chemicals^.
Inc., Morris, IL, July 29, 1977 (on file at EPA, ESED, Research Triangle Park, NC)-
3. J. F. Lawson, IT Enviroscience, Inc., Trip Report for Visit to Monsanto Industrial
Chemicals Company, St. Louis, MO, Oct. 20, 1977 (on file at EPA, ESED,
Research Triangle Park, NC).
4. J. F. Lawson, IT Enviroscience, Inc., Trip Report for Visit to Denka Chemical
Corporation, Houston, TX, November 17, 1977 (on file at EPA, ESED, Research
Triangle Park, NC).
5. J. F. Lawson, IT Enviroscience, Inc., Trip Report for Visit to Amoco Chemicals
Corporation, Jan. 24, 1978 (on file at EPA, ESED, Research Triangle Park, NC).
6. R. E. Hinkson, Denka Corporation, letter dated Apr. 18, 1978, to D. R. Patrick,
EPA.
7. R. C. Sterritt, Ashland Chemical Company, letter dated Apr. 21, 1978, to
D. R. Patrick, EPA.
*Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of
that paragraph, that reference number is indicated on the material involved.
When the reference appears on a heading, it refers to all the text covered by
that heading.
-------
5-i
REPORT 5
ETHYLBENZENE AND STYRENE
J. A. Key
F. D. Hobbs
IT Enviroscience
9041 Executive Park Drive
Knoxville, Tennessee 37922
Prepared for
Emission Standards and Engineering Division
Office of Air Quality Planning and Standards
ENVIRONMENTAL PROTECTION AGENCY
Research Triangle Park, North Carolina
September 1980
This report contains certain information which has been extracted from the
Chemical Economics Handbook, Stanford Research Institute. Wherever used, it has
been so noted. The proprietary data rights which reside with Stanford Research
Institute must be recognized with any use of this material.
D110L
-------
5-iii
CONTENTS OF REPORT 5
Page
I. ABBREVIATIONS AND CONVERSION FACTORS 1-1
JI. INDUSTRY DESCRIPTION H'1
A. Reason for Selection II-l
B- Usage and Growth II-l
c- Domestic Producers II-3
D- References 11-10
111• PROCESS DESCRIPTION III-l
A. Introduction III-l
B- Other Processes III-l
C- Styrene Co-production with Propylene Oxide III-6
0- References I3:i-8
1V- EMISSIONS IV-1
A- Styrene from Benzene and Ethylene IV-1
B. Other Processes IV~7
C- References IV'8
V- APPLICABLE CONTROL SYSTEMS V'1
A. Styrene from Benzene and Ethylene v~l
B- Other Processes v~5
C- References v~6
Vl- IMPACT ANALYSIS VI"1
A- Control Cost Impact VI~1
B- Environmental and Energy Impacts VI-9
c- References VI"12
VI1- SUMMARY
-------
5-v
APPENDICES OF REPORT 5
A. PHYSICAL PROPERTIES OF BENZENE, STYRENE, AND ETHYLBENZENE
B. AIR-DISPERSION PARAMETERS
C. FUGITIVE-EMISSION FACTORS
D. COST ESTIMATE DETAILS AND CALCULATIONS FOR MODEL-PLANT
EMISSIONS CONTROL
E. LIST OF EPA INFORMATION SOURCES
F. EXISTING PLANT CONSIDERATIONS
-------
5-vii
TABLES OF REPORT 5
Number
II-l Styrene Usage and Growth II-2
II-2 Ethylbenzene Capacity II-4
II-3 Styrene Capacity II-5
IV-1 Estimates of Uncontrolled Benzene and Total VOC Emission from IV-4
Model Plant
IV-2 Model-Plant Storage Data IV-5
V-l Estimates of Controlled Benzene and Total VOC Emissions from V-2
Model Plant
VI-1 Annual Cost Parameters VI-2
VI-2 Emission Control Cost Estimates for Styrene Model Plant VI-3
VI-3 Environmental Impact of Controlled Model Plant VI-10
VII-1 Emission Summary for Model Plant VII-2
A-i Physical Properties of Ethylbenzene A-1
A-2 Physical Properties of Styrene ^_^
A-3 Physical Properties of Benzene A-2
B-l Atmospheric Dispersion Parameters for 300,000-Mg/yr Model Plant B-l
F-l Control Devices Currently Used by the Domestic Styrene Industry F-2
F-2 Reported Uncontrolled Emission from Various Alkylation Reactor F-3
Vents
F-3 Reported Uncontrolled Emissions from Various Column Vents F-4
F-4 Reported Uncontrolled Emissions from Two Vents on Two Vacuum- F-5
Column Jets at Sun Oil
-------
5-ix
FIGURES OF REPORT 5
Number page
II-l Locations of Plants Manufacturing Ethylbenzene II-6
II-2 Locations of Plants Manufacturing Styrene II-7
III-l Process Flow Diagram of Uncontrolled Model Plant for III-3
Production of Styrene from Benzene and Ethylene by Dehydro-
genation of Ethylbenzene
v*-l Installed Capital Cost vs Plant Capacity for Emission Controls VI-5
VI-2 Net Annual Cost or Savings vs Plant Capacity for Emission Controls VI-6
VI-3 Cost Effectiveness vs Plant Capacity for Emission Controls VI-7
-------
1-1
ABBREVIATIONS AND CONVERSION FACTORS
EPA policy is to express all measurements used in agency documents in metric
units. Listed below are the International System of Units (SI) abbreviations
and conversion factors for this report.
To Convert From
Pascal (Pa)
Joule (J)
Degree Celsius (°C)
Meter (m)
Cubic meter (m3)
Cubic meter (m3)
Cubic meter (m3)
Cubic meter/second
(m3/s)
Watt (W)
Meter (m)
Pascal (Pa)
Kilogram (kg)
Joule (J)
To
Atmosphere (760 mm Hg)
British thermal unit (Btu)
Degree Fahrenheit (°F)
Feet (ft)
Cubic feet (ft3)
Barrel (oil) (bbl)
Gallon (U.S. liquid) (gal)
Gallon (U.S. liquid)/min
(gpm)
Horsepower (electric) (hp)
Inch (in.)
Pound-force/inch2 (psi)
Pound-mass (Ib)
Watt-hour (Wh)
Standard Conditions
68°F = 20°C
1 atmosphere = 101,325 Pascals
PREFIXES
Multiply By
9.870 X 10~6
9.480 X 10"4
(°C X 9/5) + 32
3.28
3.531 X 101
6.290
2.643 X 102
1.585 X 104
1.340 X 10~3
3.937 X 101
1.450 X 10"4
2.205
2.778 X 10"4
Prefix
T
G
M
k
m
M
Symbol
tera
giga
mega
kilo
milli
micro
Multiplication
Factor
1012
109
106
103
10~3
10"6
1 Tg =
1 Gg =
1 Mg =
1 km =
1 mV =
1 \ig =
Example
1 X 10 12 grams
1 X 109 grams
1 X 106 grams
1 X 103 meters
1 X 10"3 volt
1 X 10"6 gram
-------
II-l
II. INDUSTRY DESCRIPTION
A. REASON FOR SELECTION
Ethylbenzene and styrene were selected for in-depth study because preliminary
estimates indicated that their production causes relatively high volatile
organic compound (VOC) emissions. These emissions include significant quanti-
ties of benzene, which was listed as a hazardous pollutant by the EPA in the
Federal Register on June 8, 1977. Ethylbenzene and styrene were combined for
consideration because ethylbenzene is produced almost exclusively as an inter-
mediate for styrene production.
Benzene, ethylbenzene, and styrene are liquids at ambient conditions (see
Appendix A for pertinent physical properties).
B. USAGE AND GROWTH
Since more than 99% of the ethylbenzene produced is used as an intermediate for
styrene production, the growth of styrene production is the dominant considera-
tion for this report. Table II-l shows styrene end uses, percentages of produc-
tion, and expected growth rates. Virtually all styrene is consumed in polymer
? 1
manufacture, with more than half used to manufacture polystyrenes. Packaging
applications account for more than one-third of the polystyrene consumed; other
diversified end uses are toys, sporting goods, appliances and cabinets, housewares,
2
electrical parts, and disposable serviceware and flatware.
The current domestic ethylbenzene capacity is 5,070,000 Mg/yr, — with the
1976 production being 55% of this capacity.4 Production will be 78% of current
capacity by 1982, based on a projected 6% annual growth in ethylbenzene con-
2
sumption.
The current domestic styrene capacity is 3,986,000 Mg/yr, with the 1976 produc-
tion being about 72% of this capacity.6 Production levels will be nearly equal
to current capacity by 1982, based on a projected 6% annual growth in styrene
2
production.
-------
II-2
Table II-l. Styrene Usage and Growth'
End Use
Percentage of
Production (1976)
Average Annual
Growth (%) {1976—1980)
Polystyrene
Styrene copolymer resins
Styrene-butadiene elastomers
Unsaturated polyester resins
Miscellaneous
Exports
54
17
9
6
1
13
7
7
2.5
9
12.5
-25
See refs 1 and 2.
-------
II-3
C. DOMESTIC PRODUCERS
At the end of 1977 there were 15 producers operating 18 ethylbenzene plants and
157 9
12 producers operating 14 styrene plants in the United States. Tables II-2 ' ' —
and II-3 list the producers, locations, capacities, and processes; Figs. II-l
and II-2 show the plant locations. Approximately 95% of the ethylbenzene capa-
city is based on benzene alkylation, with the remainder based on extraction of
mixed xylene streams. About 89% of the styrene capacity is based on ethylbenzene
dehydrogenation. The remaining capacity is based on ethylbenzene oxidation
followed by hydroperoxidation of propylene and dehydration to styrene. The
latter process was brought on-stream by Oxirane Chemical Company in 1977 and
7
yields propylene oxide as a co-product.
The companies that produce ethylbenzene/styrene are listed below:
1. American Hoechst Corporation
Benzene is purchased for alkylation to ethylbenzene, which is captively
consumed in styrene manufacture. The styrene is 50% captively consumed
2
for polystyrene production.
2. Cosden Oil and Chemical Company
Ethylbenzene is separated from mixed xylenes, but additional feed must be
purchased to operate the styrene plant at capacity. All styrene produced
2 3
is captively consumed in polystrene production. '
3. Atlantic Richfield Company
Ethylbenzene is separated from mixed xylenes at Houston, TX, and produced
by benzene alkylation at Port Arthur, TX. Ethylbenzene supplies are
transferred to Kobuta, PA where polystyrene and styrene-butadiene co-polymers
2 3
are manufactured. '
4. Charter Company
Ethylbenzene is separated from mixed xylenes and sold. '
5. Commonwealth Oil Refining Company, Inc.
Ethylbenzene is separated from mixed xylenes and sold. '
-------
II-4
Table II-2. Ethylbenzene Capacityc
Producer
American Hoechst Corp.
Cosden Oil and Chemical Co.
Atlantic Richfield Co.
Charter Co.
Commonwealth Oil Refining Co. , Inc.
Cos-Mar, Inc.
Dow Chemical USA
El Paso Natural Gas Co.
d
Gulf Oil Corp.
Monsanto Co.e
Oxirane Chemical Co.
Standard Oil Co.
Sun Oil Co.
Tenneco
Union Carbide Corp.
Total
Location
Baton Rouge, LA
Big Spring, TX
Houston, TX
Port Arthur, TX
Houston , TX
Penuelas , PR
Carville , LA
Freeport , TX
Midland, MI
Odessa, TX
St . James , LA
Alvin, TX
Texas City, TX
Channelview, TX
Texas City, TX
Corpus Christi, TX
Chalmette, LA
Seadrift, TX
1977 Capacity
(103 Mg)
526
47
50
227
18
73
689
782
249
125
313
23
771
526
447
61
16
154
5097
Proces^
b
c
c
b
c
c
b
b
b
b
b
b
b,c
b
b
b,c
c
b
D
Benzene alkylation.
•^
"Mixed xylene stream recovery.
Assumes that Gulf increased capacity as planned in the latter part of 1977.
a
"Not listed in all refs.
-------
II-5
Table II-3. styrene Capacity"
Producer
American Hoechst Corp.
Cosden Oil and Chemical Corp.
Atlantic Richfield Co.
Cos -Mar, inc.
D°w Chemical USA
PI r, d
fil Paso Natural Gas Co.
Gulf oil corp.e
Monsanto Co.
Oxirane Chemical Co.
Standard oil Co.
Sun oil Co.
Union Carbide Corp.
Total
Location
Baton Rouge/ LA
Big Springs, TX
Houston, TX
Kobuta, PA
Carville, LA
Freeport, TX
Midland, MI
Odessa, TX
St. James/ LA
Texas City, TX
Channelview, TX
Texas City, TX
Corpus Christ! , TX
Se adrift, TX
1977 Capacity
(10 3 Mg)
408
50
54
100
590
680
181
68
272
590
454
363
36
_136
3982
Process
c
c
c
c
c
c
c
c
c
c
f
c
c
c
See refs 5 and 10.
Plans to bring a 408.2 X 103-Mg/yr styrene plant on-stream at Bayport, TX, in 1980.
c
Ethyiiienzene dehydrogenation.
Will increase its capacity by 40.8 X 103 Mg/yr in 1978.
eAssumes that Gulf Oil Corp. increased its capacity by 34.0 X 103 Mg/yr in the latter part
of 1977,
Hydroperoxidation of propylene.
-------
II-6
1. American Hoechst, Baton Rouge, LA
2. Cosden Oil and Chemical, Big Spring,
3. Atlantic Richfield, Houston, TX
4. Atlantic Richfield, Port Arthur, TX
5. Charter, Houston, TX
6. Commonwealth Oil, Penuelas, PR
1. Cos-Mar, Carville, LA
8. Dow, Freeport, TX
9. Dow, Midland, MI
10. El Paso National Gas, Odessa,
TX 11. Gulf, Donaldsonville, LA
12. Monsanto, Alvin, TX
13. Monsanto, Texas City, TX
14. Oxirane, Channelview, TX
15. Standard Oil, Texas City, TX
16. Sun Oil, Corpus Christi, TX
17. Tenneco, Chalmette, LA
18. Union Carbide, Seadrift, TX
Fig. II-l. Locations of Plants Manufacturing Ethylbenzene
-------
II-7
1. American Hoechst, Baton Rouge, LA 8.
2. Cosden Oil and Chemical, Big Spring, TX 9.
3. Atlantic Richfield, Houston, TX 10.
4. Atlantic Richfield, Kobuta, PA 11.
5. Cos-Mar, Carville, LA 12.
6. Dow, Freeport, TX 13.
7. Dow, Midland, MI 14.
El Paso National Gas, Odessa, TX
Gulf, Donaldsonville, TX
Monsanto, Texas City, TX
Oxirane, Channelview, TX
Standard Oil, Texas City, TX
Sun Oil, Corpus Christi, TX
Union Carbide, Seadrift, TX
Fig. I1-2. Locations of Plants Manufacturing Styrene
-------
II-8
6. Cos-Mar, Inc.
Benzene is purchased and alkylated to ethylbenzene. The plant capacity
3
was increased by 362,800 Mg/yr in 1976. About 45% of the styrene produced
2
is captively consumed.
7. Dow Chemical USA
Ethylbenzene is produced by benzene alkylation and captively consumed in
the manufacture of styrene. More than 96% of the styrene is captively
consumed in the production of polystyrene, acrylonitrile-butadiene-styrene
(ABS), styrene-acrylonitrile (SAN), and styrene-butadiene.
8. El Paso Natural Gas Company
Ethylbenzene is produced by benzene alkylation and is captively consumed
for manufacture of styrene, but some is sold. Styrene capacity will be
2
increased by 40,800 Mg/yr in 1978. All styrene produced is sold.
9. Gulf Oil Corporation
Ethylbenzene is produced by benzene alkylation and captively consumed in
styrene manufacturing. Styrene capacity was increased to 272,100 Mg/yr
in 1977.1
10. Monsanto Company
About 96% of the ethylbenzene is produced by benzene alkylation, with the
remainder being extracted from mixed xylenes. ' All ethylbenzene is
captively consumed in styrene manufacture. About 35% of the styrene is
2
captively consumed in polystyrene, ABS, and SAN production.
11. Oxirane Chemical Company
Ethylbenzene is produced by benzene alkylation and is captively consumed
3 2
in the manufacture of styrene. All styrene produced is sold.
12. Standard Oil Company
Ethylbenzene is produced by benzene alkylation and is captively consumed
in styrene manufacture. Polystyrene manufacture captively consumes 40%
of the styrene.
-------
II-9
13. Sun Oil Company
Benzene alkylation produces about 65% of the ethylbenzene, with the remainder
being extracted from mixed xylenes. Ethylbenzene is consumed captively in
3 2
styrene manufacture. All styrene produced is sold.
14. Tenneco, Inc.
Ethylbenzene is extracted from mixed xylenes and sold. '
15. Union Carbide Corporation
Ethylbenzene is produced by benzene alkylation and is captively consumed
in styrene manufacture. The company manufactures polystyrene and SAN
2
resins from the styrene.
-------
11-10
D. REFERENCES*
1. T. C. Gunn and K. L. Ring, "Benzene," pp 618.5022 V—Y in Chemical Economics
Handbook, Stanford Research Institute, Menlo Park, CA (May 1977).
2. S. L. Sober, "Styrene," pp 694.3052 A,B in Chemical Economics Handbook, Stanford
Research Institute, Menlo Park, CA (January 1977).
3. S. K. Paul and S. L. Soder, "Ethylbenzene Salient Statistics," pp 645.3000
A—H in Chemical Economics Handbook, Stanford Research Institute, Menlo Park,
CA (January 1977).
4. "Manual of Current Indicators Supplement Data," p 211 in Chemical Economics
Handbook, Stanford Research Institute, Menlo Park, CA (December 1977).
5. C. V. Sleeth, "Styrene Monomer," Chemical and Engineering Progress. 73(11),
31—35 (1977). —
6. "Manual of Current Indicators Supplement Data," p 235 in Chemical Economics
Handbook, Stanford Research Institute, Menlo Park, CA (December 1977).
7. J. L. Blackford, "Propylene Oxide," p 690.8022 C in Chemical Economics Handbook,
Stanford Research Institute, Menlo Park, CA (November 1976).
8. "Ethylbenzene," Chemical Marketing Reporter, p 9 (July 1, 1975).
9. T. M. Nairn, Cosden Oil and Chemical Co., letter dated June 30, 1978, to EPA.
10. H. M. Walker, Monsanto Company, Chocolate Bayou Plant, Alvin, TX, Texas Air
Control Board 1975 Emissions Inventory Questionnaire, July 9, 1976.
*Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of
that paragraph, that reference number is indicated on the material involved.
When the reference appears on a heading, it refers to all the text covered by
that heading.
-------
III-l
III. PROCESS DESCRIPTION
A- INTRODUCTION
More than 95% of domestic ethylbenzene production is by benzene alkylation with
ethylene. The remainder is recovered by distillation from mixed xylene streams
that result from naphtha re-forming or cracking in petroleum refineries.
Refinery recovery processes, however, are not within the scope of this study.
More than 99% of the ethylbenzene produced is used as an intermediate for
making styrene, often in an integrated ethylbenzene-styrene plant. Except for
a new plant brought on-stream in July 1977 by Oxirane Corporation, all domestic
styrene is produced by catalytic dehydrogenation of ethylbenzene. The Oxirane
ethylbenzene oxidation process is also used in Spain and Japan; however, most
foreign styrene production is by dehydrogenation of ethylbenzene. Future new
plants will use either the conventional ethylbenzene dehydrogenation process or
the Oxirane ethylbenzene oxidation process, depending on the economics of both
styrene and the propylene oxide that is co-produced with styrene in the Oxirane
process.
B- STYRENE FROM BENZENE AND ETHYLENE
^-- Basic Process '
The primary reactions in the production of styrene are catalytic alkylation of
benzene with ethylene to produce ethylbenzene, and catalytic dehydrogenation of
the ethylbenzene to produce styrene:
1. Catalytic alkylation of benzene with ethylene to ethylbenzene
Benzene Ethylene Ethylbenzene
2. catalytic dehydrogenation of ethylbenzene to styrene
-CH=CH_ . „
2 + H2
Hydrogen
-------
III-2
The first step in the process (Fig. III-l) is benzene drying to remove water
from both feed and recycled benzene. The dry benzene (stream 1) is fed to the
alkylation reactor along with ethylene, aluminum chloride catalyst, and recycled
polyethylbenzenes. The reactor effluent goes to a settler, where crude ethyl-
benzene is decanted and the heavy catalyst-complex layer is recycled to the
reactor. Reactor vent gas is routed though a condenser and scrubbers in the
alkylation reaction section to recover aromatics and to remove HCl.
The crude ethylbenzene (stream 2) from the settler is washed with water and
caustic to remove traces of chlorides and is then fed to the ethylbenzene
purification section. The first step in purification is separation of recycled
benzene (stream 3) from the crude ethylbenzene in the benzene recovery column.
In the second step the product ethylbenzene (stream 4) is separated from the
heavies in the ethylbenzene recovery column. Finally, the heavies are distilled
in the polyethylbenzene column to separate the polyethylbenzenes, which are
recycled (stream 5), from the residue oil.
Fresh ethylbenzene (stream 4) from the ethylbenzene purification section is
combined with recycled ethylbenzene (stream 6) from the styrene purification
section and fed to the heat exchanger in the dehydrogenation reaction section.
Here the ethylbenzene is vaporized and superheated by heat exchange with the
reactor effluent. Next, the superheated ethylbenzene vapor (stream 7) is mixed
with superheated steam, and the combined stream is fed to the catalytic dehydro-
genation reactor. The reactor effluent (stream 8) is cooled by heat exchange
with the ethylbenzene feed. The mixture goes through a water-cooled condenser,
where the steam and crude-styrene vapor are condensed and then flow to the
separator. After separation, the noncondensible hydrogen-rich gas (stream 9)
is sent to the recovery section, the process water condensate is decanted
(stream 10) and goes to a stripper, and the remaining crude styrene (stream 11)
is sent to the styrene purification section. The hydrogen-rich gas (stream 9)
is compressed and then cooled in the recovery section to recover aromatic
organics, which are returned to the process. After removal of the organics,
the hydrogen-rich gas (stream 12) is usually sent to the steam superheater,
where it is used as fuel. The process water condensate (stream 10) is fed to
the process water stripper, where dissolved aromatic organics are removed and
returned to the process. The purified water is sent to the plant boiler for
use as boiler feed water.
-------
BE.WZ.Ewe
COUJMW COUJMM COi-UMM
PumPlC/CnOki •=;
H
M
V
UJ
WATER
•ST RIPPER
)- F=\JC(ITIVE EMl^blOKJ^ - OVERALL PUVWT
- SECOMDAB-r EMl'b'ilOki POTElJTlJia_
Fig. III-l. Process Flow Diagram of Uncontrolled Model Plant for
Production of Styrene from Benzene and Ethylene by Dehydrogenation of'Ethylbenzene
-------
III-4
In the styrene purification section, benzene and toluene (stream 13) are first
separated from the crude styrene in the benzene-toluene column, and benzene
(stream 15) is recovered and sent to the alkylation section after separation in
the benzene recovery column. The recycled ethylbenzene (stream 6) is separated
from the styrene and tars in the ethylbenzene recycle column and then reprocessed.
Finally, the product styrene (stream 14) is separated from the tars in the
styrene finishing column.
Any inert gases fed with the ethylene or produced in the alkylation reactor,
along with some unreacted benzene, other organics, and hydrogen chloride, are
exhausted from vent A of the reactor or from vent A of the treating section.
The benzene and organics are recovered, and the hydrogen chloride is removed
A
from the exhaust gas before the remaining inert gases are vented. Normally
67
these vent gases are routed to an emission-control device. ' Other emission
sources are the vents (B) from the columns for benzene drying, ethylbenzene
purification, and styrene purification and the emergency vent (C) between the
6 7
separator and the recovery section. '
Storage emission sources (vents D through F) include benzene, crude and finished
ethylbenzene, crude and finished styrene, polyethylbenzene, benzene-toluene,
residue oil, and tar storage. Handling emissions (vent G) result from barge
loading of styrene and benzene-toluene.
Fugitive emissions occur when leaks develop in valves, pump seals, etc. When
the process pressures are higher than the cooling-water pressure, benzene and
other organics can leak into the cooling water and result in fugitive emissions
from the cooling tower.
Secondary emissions occur when wastewater streams, such as the aluminum chloride
catalyst solution and spent caustic from the treating section (discharge K),
containing benzene and other organics, are treated in a waste-treatment plant.
2. Process Variations
Ethylene used to produce styrene is usually 95 to 99% pure. However, a dilute
8 11
ethylene stream containing as little as 10% ethylene can be used. —
-------
III-5
The pressure in the alkylation reactor can range from near-ambient to 2758 kPa,
and temperatures may range from 80 to 400°C, depending on the pressure and the
catalyst used. Aluminum chloride is the most common catalyst used in low-pres-
sure liquid-phase processes. Some high-pressure vapor-phase processes using
solid catalysts are in operation, such as the Alkar process by UOP and the
newer Mobil/Badger process. The Mobil/Badger process was used in a demon-
stration unit at Foster Grant's Baton Rouge, LA, plant and in a converted
ethylbenzene unit at Cosden Oil and Chemical's Big Spring, TX, plant. The
Alkar process uses boron trifluoride as a catalyst for the alkylation of ben-
zene with ethylene. El Paso Natural Gas Co. of Odessa, TX, uses a dilute
ethylene stream (50%) to feed this process and sends the off-gases from the
alkylation reactor to their boilers as fuel, thereby effectively reducing .
emission of volatile organic compounds (VOC) and recovering the fuel value of
the non-ethylene hydrocarbons in the ethylene feed. It has been reported that
this process does not produce by-products or sludge and that the catalyst lasts
9 11
for several years. —
Another process variation is operation of the alkylation reactor at pressures
greater than atmospheric with high-purity ethylene. Consequently natural gas
has to be added at times to maintain the desired pressure in the reactor, and
little, if any, emissions will vent from the reactor. The inert gases fed with
the ethylene and natural gas and any inert gases produced as by-products in the
reactor go with the alkylate to a degassing step, where they are removed when
9
the pressure on the alkylate is reduced.
The dehydrogenation reactor can be operated either at constant temperature,
which requires that heat be added in the reactor; or it can be operated under
adiabatic conditions in which all the heat is supplied by diluting the ethyl-
benzene feed entering the reactor with superheated steam. '
The sequence of columns in the styrene purification section can be varied from
that shown in Fig. III-l. The first separation takes a mixture of ethylbenzene,
toluene, and benzene overhead, leaving the styrene and tars as bottoms. The
overhead stream is then fed to a column where the benzene and toluene are
separated from the ethylbenzene. The bottoms from the first separation go to
-------
III-6
another column, where the styrene is separated from the tars. In one process,
only one condenser is used to condense the overhead vapor from three distilla-
tion columns: the benzene drying column, the benzene recovery column, and the
column separating benzene from toluene.
C. OTHER PROCESSES STYRENE CO-PRODUCTION WITH PROPYLENE OXIDE
Ethylbenzene can be oxidized with air to make ethylbenzene hydroperoxide, which
in turn is reacted with propylene to form methyl phenol carbinol and propylene
oxide. The carbinol is dehydrated to styrene. The chemical reactions are
shown below:
1.
Ethylbenzene
+ 0,
Oxygen
OOH
CH-CH.
Ethylbenzene
Hydroperoxide
2.
OOH
-CH-CH.
Ethylbenzene
Hydroperoxide
H-CH,
Propylene
Methyl Phenyl
Carbinol
+ CH A
v;
Propylene
Oxide
3.
H-CH.
Methyl Phenyl Carbinol
-CH=CH.
Styrene
+ HO
2
Water
-------
III-7
About 2.5 kg of styrene is produced per kilogram of propylene oxide. ' '
Actual process data are currently not available. Since air is used for oxida-
tion of ethylbenzene, the nitrogen from the air has to be vented; this is a
possible source of VOC emissions.
-------
III-8
D. REFERENCES*
1. T. C. Gunn and K. L. Ring, "Benzene," pp 618.5022 V—Y in Chemical Economics
Handbook, Stanford Research Institute, Menlo Park, CA (May 1977).
2. S. L. Soder, "Styrene," pp 694.3052 A, B in Chemical Economics Handbook, Stanford
Research Institute, Menlo Park, CA (January 1977).
3. J. L. Blackford, "Propylene Oxide," p 690.8022 C in Chemical Economics Handbook,
Stanford Research Institute, Menlo Park, CA (November 1976).
4. "Styrene," Hydrocarbon Processing, 56(11), 226—228 (1977).
5. J. W. Pervier et al., Survey Reports on Atmospheric Emissions from the
Petrochemical Industry, Houdry Division of Air Products, Inc.,
EPA-450/3-73-005-D (April 1974).
6. J. A. Key, IT Enviroscience, Inc., Trip Report for Visit to Dow Chemical, USA,
Freeport, TX, July 28 and 29, 1977 (on file at EPA, ESED, Research Triangle
Park, NC).
7. J. A. Key, IT Enviroscience, Inc., Trip Report for Visit to Cos-Mar Plant, Cosden
Oil and Chemical Company, Carville, LA, July 16, 1977 (on file at EPA, ESED,
Research Triangle Park, NC).
8. E. M. Carlson and M. G. Erskine, "Ethylene," pp 648.5054 H—I in Chemical
Economics Handbook, Stanford Research Institute, Menlo Park, CA (February 1975).
9. "Ethylbenzene," Hydrocarbon Processing 56(11), 151, 152 (1977).
10. "Better Path to Ethylbenzene,11 Chemical Engineering, 84 120, 121 (Dec. 5, 1977).
11. K. E. Coulter, H. Kehde, and B. F. Hiscock, "Styrene," pp 55—85 in Kirk-Othmer
Encyclopedia of Chemical Technology, edited by A. Standen e_t al., vol 19, 2d
ed., Wiley-Interscience, New York, 1969.
12. C. V. Sleeth, "Styrene Monomer," Chemical and Engineering Progress 73(11),
31—35 (1977). —
13. L. R. Roberts, Texas Air Control Board Permit No. 2993, issued to Oxirane
Chemical Co., Channelview, TX, for construction of propylene oxide—styrene
monomer plant, Apr. 3, 1975.
*Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of that
paragraph, that reference number is indicated on the material involved. When
the reference appears on a heading, it refers to all the text covered by that
heading.
-------
IV-1
IV. EMISSIONS
Emissions in this report are usually identified in terms of volatile organic
compounds (VOC). VOC are currently considered by the EPA to be those of a
large group of organic chemicals, most of which, when emitted to the atmosphere,
participate in photochemical reactions producing ozone. A relatively small
number of organic chemicals have low or negligible photochemical reactivity.
However, many of these organic chemicals are of concern and may be subject to
regulation by EPA under Sections 111 or 112 of the Clean Air Act since there
are associated health or welfare impacts other than those related to ozone
formation. It should be noted that although ethane may be included in VOC
emission totals in this report it does not, based on current research data,
participate in ozone-forming reactions to an appreciable extent.
The process emissions estimated for the styrene model plant are based on the
emissions reported in response to EPA's requests for information from selected
? 3 A.
companies, the Cos-Mar and Dow trip reports, a Houdry study, and SRI informa-
tion and on engineering judgement. One source estimates that 30% of the 1976
capacity of the domestic styrene industry was from smaller, older, less efficient
plants. These plants would be expected to have more emissions than newer, more
efficient, and modern plants. Data received from styrene producers show wide
variations in estimates of emissions, even from the more modern plants.
STYRENE FROM BENZENE-ETHYLENE
Model Plant
The model plant* for this study has a capacity of 300,000 Mg/yr, based on
8760 annual hours of operation.** Though not an actual operating plant, it is
typical of many recently built plants. The model styrene process (Fig. III-l)
*See p 1-2 for a discussion of model plants.
**Process downtime is normally expected to range from 5 to 15%. If the hourly
rate remains constant, the annual production and annual VOC emissions will
be correspondingly reduced. Control devices will usually operate on the same
cycle as the process. From the standpoint of cost-effectiveness calculations
the error introduced by assuming continuous*operation is negligible.
-------
IV-2
best fits today's styrene manufacturing and engineering technology. Single-
process trains as shown are typical for the large plants built recently. The
model process uses alkylation of benzene with ethylene in the presence of
aluminum chloride catalyst to produce ethylbenzene. After purification, the
ethylbenzene is dehydrogenated in the presence of steam over a solid catalyst
to produce styrene. The crude styrene, separated from the hydrogen-rich gas
and condensed steam, is purified to make the final product.
Typical raw material, intermediate, and product-storage tank capacities were
estimated for a 300,000-Mg/yr plant. Storage-tank requirements are given in
Sect. IV.A.2.d. Estimates of potential fugitive sources, based on data from
existing facilities, are given in Sect. IV.A.Z.e. Characteristics of the model
plant that are important in air-dispersion modeling are shown in Appendix B.
2. Sources and Emissions
Emission rates and sources for the styrene process are summarized in Table IV-1.
a. Alkylation Reaction Vent The alkylation reaction section vent gas (A, Fig. III-l)
consists of unreacted ethylene and impurities such as methane and ethane. Some
hydrogen chloride, unreacted benzene, and other aromatic organic vapors leave
the reactor with the vent gas. The alkylation reaction section includes a
condenser and a polyethylbenzene scrubber for recovering benzene and other
aromatics from the vent gases. With high-purity ethylene (greater than 99.9%)
the vent gases are almost nonexistent during normal operation. The amount of
vent gases increases directly as the impurities (methane and ethane) in the
ethylene increase. Data as to the purity of the ethylene were not available;
so engineering judgement was used to estimate these emissions for the model
plant. During the first few minutes of startup the quantity of vent gases
temporarily increases, or until the alkylation reaction is fully established.
b. Column Vents The gases from the benzene drying column and the columns in the
ethylbenzene purification and styrene purification sections (vents B, Fig.
III-l) are the noncondensibles that are dissolved in the feed to the columns,
VOC that are not condensed, and, for the columns operated under vacuum, air
that leaks into the column and is removed by the vacuum-jet systems.- At one
plant the gases from the process-water stripper condenser are vented to the
-------
IV-3
atmosphere. The data available generally give the emissions for a group of
distillation columns. Even when emissions from a single column are shown, the
values are not consistent (see Appendix F), so that an estimate of the emissions
from a single distillation step is questionable. ]
used to estimate the emissions shown in Table IV-1.
from a single distillation step is questionable. Engineering judgement was
c- Emergency Vent on Separator The hydrogen-rich gas from the separator also
contains methane, ethane, ethylene, carbon dioxide, carbon monoxide, and aromatic
organics, all of which are formed by side-reactions in the dehydrogenation
reactor. It is vented only during an emergency shutdown of the recovery section.
Normally aromatic organics are recovered and sent back to the process, and
usually the hydrogen-rich gas is sent to the steam superheater, where it is
burned as fuel. One plant vents it to the atmosphere and some plants send it
to a flare.1 Although benzene and VOC can be vented at significant rates
during an emergency, the annual emission from the emergency vent is relatively
insignificant, as is indicated in Table IV-1.
Storage and Handling Emissions Emissions result from the storage and handling
of benzene, ethylbenzene, styrene, and by-product and recycled streams. Sources
for the model plant are shown in Fig. Ill-i (sources D through G). Storage-tank
conditions for the model plant are given in Table IV-2. The emissions in
Table IV-1 were calculated based on fixed-roof tanks, half full, and a 8°C diurnal
temperature, with use of the emission equations from AP-42. However, calculated
breathing losses were divided by 4 to account for recent evidence indicating that
•7
the AP-42 breathing-loss equation overpredicts emissions.
Since benzene freezes at 5.5°C, storage tanks are heated to maintain the tempera-
ture above freezing if th'e ambient temperature drops below this point for a
significant period. The styrene tar is heated to 115°C to reduce its viscosity.
The styrene storage tanks are cooled to help inhibit polymerization. These
temperature controls will reduce the emissions caused by breathing. No data
were available on the amount of this reduction.
\
Emissions from loading of styrene and toluene barges were calculated with the
equations from AP-42. These emissions, included in Table IV-1, are 0.8 kg/hr.
-------
IV-4
Table IV-1. Estimates of Uncontrolled Benzene and Total VOC Emissions
from Model Plant
Emissions
Source
Vent
Designation
(Fig. III-l
Ratio (g/kg)'
Rate (kg/hr)
Benzene Total VOC Benzene • Total VOC
Alkylation reaction
section
0.29
0.88
10
30
Column vents
Emergency vent on
separator
Storage and handling
Fugitive
Secondary
Total
B
C
D-G
H
K
1
0
0
0
0
2
.7
.003
.57
.11
.067
.7
2
0
0
0
0
4
.6
.008
.67
.51
.088
.8
58
0.091
20
3.8
2.3
94
90
0
23
17
3
160
.27
.0
g of emission per kg of product produced.
"See refs 1—3; estimated ethane content is 67%.
1-1
"See refs 3 and 4.
Average rate for entire year, based on 80-kg/hr benzene and 240-kg/hr total
VOC vented for 5 hr twice a year (from refs 3 and 4).
-------
IV-5
Table IV-2. Model-Plant Storage Data
Content
Benzene
Crude ehtylbenzene
Polyethylbenzene
Residue oil
Ethylbenzene
Crude styrene
Toluene
Styrene (3 tanks)
Styrene (2 tanks)
Styrene tar
Tank Size
(m3)
10,000
300
80
160
3,000
900
900
1,000
8,000
160
Turnovers
per Year
40
6
6
60
6
6
30
110
20
60
Bulk Liquid
Temperature (°c)
25
30
50
65
35
50
25
15
15
115
-------
IV-6
Emissions from loading residue oil and tar into tank trucks are not significant
because of their low vapor pressures.
e. Fugitive Emissions Process pumps and valves are potential sources of fugitive
emissions. The model plant is estimated to have 50 pumps handling VOC, with 10
in heavy liquid service and 40 in light liquid service (10 of these handle
benzene). The estimated number of pipeline valves is 1000, with 160 in gas/vapor
service (includes 32 handling benzene), 700 in light liquid service (includes
140 handling benzene), and 140 in heavy liquid service. The estimated number
of safety/relief valves is 25 with 12 "in gas/vapor service (includes 3 on
benzene) 12 in light liquid service (includes 3 on benzene) and 1 in heavy
liquid service. The fugitive emission factors from Appendix C were applied to
this valve and pump count to determine the fugitive emissions shown in Table
IV-1.
f. Secondary Emissions Secondary VOC emissions can result from the handling and
disposal of process-waste liquid streams. Five potential sources (K) are
indicated in Fig. III-l for the model plant.
For the model plant the total VOC-associated aqueous effluent is estimated to
be 12,000 kg/hr and to contain 6 kg/hr VOC, of which 5 kg/hr is benzene. The
aqueous effluent consists of the aluminum chloride solution and the spent
caustic from the treating section, plus other miscellaneous wastewater streams
such as the vacuum-system water discharges. ' The secondary emission from the
treatment of this aqueous waste is assumed to be 2.7 kg of VOC per hr, of which
2.3 kg is benzene. These values are included in the uncontrolled secondary
emissions shown in Table IV-1.
The residue oil from the polyethylbenzene column and the tar from the styrene
finishing column can be used as fuel or can be sold, depending on market condi-
23 4
tions. The total of the two streams is estimated to be 2000 kg/hr. Based
on the assumption that these streams are properly conditioned (by diluting with
low-viscosity fuel oil or by heating) so as to reduce the viscosity and ensure
proper atomization into an industrial boiler and that they otherwise meet the
Q
conditions specified in AP-42. The resulting emissions are calculated to be
0.3 kg of VOC per hr. This value is included in the uncontrolled emissions.
Table IV-1.
-------
IV-7
B. STYRENE CO-PRODUCTION WITH PROPYLENE OXIDE
No data are available on the actual process or emissions when styrene is co-pro-
duced with propylene oxide.
-------
IV-8
C. REFERENCES*
1. Responses to EPA request for information on emissions from ethylbenzene and
styrene manufacturing; see Appendix E.
2. J. A. Key, IT Enviroscience, Inc., Trip Report on Visit to Cos-Mar Plant,
Cosden Oil and Chemical Company, Carville, LA, June 16, 1977 (on file at EPA,
ESED, Research Triangle Park, NC).
3. J. A. Key, IT Enviroscience, Inc., Trip Report on Visit to Dow Chemical, USA,
Freeport, TX, July 28 and 29, 1977 (on file at EPA, ESED, Research Triangle
Park, NC).
4. J. W. Pervier et al., Houdry Division of Air Products, Inc., Survey Reports
on Atmospheric Emissions from the Petrochemical Industry, EPA-450/3-73-005-d
(April 1974).
5. S. L. Soder and R. E. Davenport, "Ethylene," p 648.5054 N in Chemical Economics
Handbook, Stanford Research Institute, Menlo Park, CA (January 1978).
6. C. C. Masser, "Storage of Petroleum Liquids," pp 4.3-12 and 4.3-13 in Compilation
of Air Pollutant Emission Factors, AP-42, Part A, 3d ed. (August 1977).
7. E. C. Pulaski, TRW, letter dated May 30, 1979, to Richard Burr, EPA.
8. T. Lahre, "Fuel Oil Combustion," pp 1.3-1 to 1.3-5 in Compilation of Air
Pollutant Emission Factors, AP-42, Part A, 3d ed. (August 1977).
^Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of that
paragraph, that reference number is indicated on the material involved. When
the reference appears on a heading, it refers to all the text covered by that
heading.
-------
V-l
V. APPLICABLE CONTROL SYSTEMS
A. STYRENE FROM BENZENE-ETHYLENE
1. Alkylation Reaction Section
Some styrene producers control the vent gases from the alkylation reaction
section by incinerating the process emissions from that section for use as a
portion of the fuel for a fired heater or boiler. Since the vent stream would
have already passed through a condenser and scrubbers for recovery of the
aromatics and removal of HC1, no additional treatment would be needed before the
vent stream is burned. When the ethylene feed stream to the alkylation reaction
section is dilute, there is a strong economic incentive to use the methane,
etc., as fuel.1'2 To accomplish this a pipe is needed to transport the vent
gas to the heater or boiler and a special burner and controls are required to
prevent variations in the vent-gas flow and composition from disturbing the
heater or boiler operation. For the model plant this stream is piped to the
steam superheater.
Since the VOC burned is essentially equivalent to natural gas, the controlled
VOC emissions in the steam superheater flue gas were calculated with the emission
factors in AP-42, and resulted in a VOC reduction efficiency of greater than
99.9% (see Table V-l).
2. Column Vents
The gases from the column vents are controlled by incineration in a flare
and/or by a refrigerated vent condenser in many styrene plants. For the model
plant the same flare is also used to control emissions from the styrene storage-
tank vents as discussed in Sect. A.4.C. A flare-tip pressure drop of 3 in. of
water was selected to provide low back pressure to the columns and storage
tanks. A new plant can be designed so that a back pressure of this magnitude
will not cause problems such as upsets in column operation or overpressure of
the storage tanks. The flare system includes the manifold piping from the
vents, provisions for a continuous purge-gas flow plus an oxygen-monitoring
system to warn of possible explosive concentrations in the manifold piping, and
a flare consisting of a knockout drum to remove liquid from the vent gases, a
water seal, flare tip, pilot lights, steam ring, and all necessary controls,
-------
Table V-l. Estimates of Controlled Benzene and Total VOC Emissions
from Model Plant
Emissions
Stream
Designation
Source (Fig. III-l)
Alkylation reaction section A
Column vents B
Emergency vent on separator C
Storage tanks D
Storage tanks E
Storage tank Fl
Other storage and F&G
handling
Fugitive H
Secondary K
Total
Control Device
or Technique
Process heater
Flare
Flare
Internal-floating-
roof tanks
Pad vented to
flare
Refrigerated
vent condenser
None
Detection and correction
of major leaks
None
Emission
Ratio
(%) Benzene
99.9+ 0.00004
99 0.017
99 0.000027b
85 0.085
99 0.000031
80 0.00013
0
74 0.029
0.067
0.20
(g/kg) a
Total VOC
0.0012
0.026
0.000080b
0.090
0.00050
0.00025
0.015
0.14
0.088
0.36
Rate (kg/hr)
Benzene
0.0014
0.58
0.00091b
2.9
0.0011
0.0043
0
0.98
2.3
6.8
Total VOC
0.0041
0.90
0.0027b
3.1
0.017
0.0087
0.53
4.7
3.0
12
g of emission per kg of styrene produced.
""Average rate for entire year, based on emergency venting for 5 hr twice a year.
I
NJ
-------
V-3
etc., to provide for automatic ignition and for steam application to ensure
smokeless operation. An emission reduction of 99% was used to calculate the
controlled benzene and VOC emissions given in Table V-l for the column vents.
4
Emergency Vent on Separator
The emergency vent on the separator releases hydrogen-rich gas that is control-
led in many styrene plants by incineration in the same flare system used to
control the gases from the column vents. For the model plant a separate flare
is sized to handle only the emergency venting of gases from the separator when
the compressor is shut down. A flare-tip pressure drop of 18 in. of water was
used to size this flare. The decision to use one flare or multiple flares
depends on the design philosophy of each company, and so a discussion of the
factors involved is beyond the scope of this report. An emission reduction of
99% was also used to calculate the values given in Table V-l for controlled
benzene and VOC emissions from the separator emergency vent. This reduction is
based on smokeless operation of the flare by use of assist steam. Because the
gases to this flare contain over 90% hydrogen by volume, the capacity of the
flare for smokeless operation with assist steam is believed to be approximately
the same as the capacity calculated for a flare-tip pressure drop of 18 in. of
water.
Storage and Handling Sources
Benzene Storage Control of benzene storage emissions will be discussed in a
future EPA document. Information from styrene manufacturers indicates that
benzene is stored in floating-roof tanks at seven locations and in fixed-roof
5 9
tanks at the others. — A floating roof is commonly used to control storage-
tank VOC emissions in the vapor-pressure range of benzene and is used in the
model plant for storage emission control. The benzene emissions given in
Table V-l were calculated by assuming that a contact type of internal-floating
roof* with secondary seals will reduce fixed-roof tank emissions by approximately
85%.1
*Consist of internal floating covers or covered floating roofs as defined in
API-2519, 2d ed, 1976 (fixed-roof tanks with internal floating device to
reduce vapor loss).
-------
V-4
b. Ethylbenzene and Toluene Storage The vapor pressure of ethylbenzene under
storage conditions in the model plant is 2.5 kPa and that of toluene is 4.0 kPa.
The use of internal-floating-roof tanks for the model plant reduces the emissions
from 1.0 to 0.15 kg of VOC per hour as calculated by assuming 85% control.
The emissions from these internal-floating-roof tanks are included in the
emissions for floating-roof tanks given in Table V-l.
c. Styrene, Crude Styrene, and Tar Storage Styrene, crude styrene, and tar
stored in fixed-roof tanks will result in calculated average emissions of
1.7 kg VOC per hour. At two production plants the styrene storage tanks are
blanketed with an inert gas to exclude air. ' For the model plant these
emissions are controlled by the same flare system used to control the vent gas
4
from the columns as previously discussed. Using an emission reduction of 99%
gives a controlled emission of 0.017 kg/hr from the styrene, crude styrene, and
tar storage tanks, as given in Table V-l.
d. Crude-Ethylbenzene Storage The crude-ethylbenzene contains 40—60% benzene
and is often warmer than ambient temperature. The temperature and level are
relatively constant because normally the flows in and out are equal; therefore
the breathing and working losses are small. A refrigerated vent condenser is
used as the control on the model-plant crude-ethylbenzene storage vent. The
emission reduction is estimated to be 80% and the calculated controlled emissions
to be 0.0043 kg of benzene per hour and 0.0087 kg of VOC per hour.
e. Other Storage and Handling The storage of polyethylbenzene and residue oil in
fixed-roof tanks results in a calculated average emission of less than 0.016 kg/hr,
which is insignificant relative to other emissions from the process; therefore
control is assumed not to be needed. The loading of barges with styrene and
toluene results in an average emission of less than 0.51 kg VOC/hr as calculated
12
with the AP-42 equations for submerged loading of barges. The total uncontrolled
emissions from these sources are 0.53 kg/hr of VOC (see Table V-l).
5. Fugitive Sources
Controls for fugitive sources are discussed in another EPA report. Controlled
fugitive emissions calculated from use of the factors given in Appendix C are
included in Table V-l. The factors are based on the assumption that major
-------
V-5
leaks are detected and corrected and that the estimated emission reduction is
74%.
Secondary Sources
Secondary emissions caused by burning of the residue oil and tar streams as
fuel are insignificant. The secondary emissions from the aqueous effluent from
the treating section (streams K, Fig. III-l) are difficult to control. Control
14
of secondary emissions is discussed in another EPA report. No control system
has been identified for the secondary emissions from the model plant.
OTHER PROCESSES STYRENE CO-PRODUCTION WITH PROPYLENE OXIDE
Data are not currently available concerning control devices for the Oxirane
process for styrene co-production with propylene oxide by the oxidation of
ethylbenzene.
-------
V-6
B. REFERENCES*
1. "Styrene," Hydrocarbon Processing 56(11), 226 — 228 (1977).
2. "Ethylbenzene," Hydrocarbon Processing 56(11), 151,152 (1977).
3. T. Lahre, "Fuel Oil Combustion," p 1.3-2, and "Natural Gas Combustion,"
p 1.4-2 in Compilation of Air Pollutant Emission Factors, AP-42, Part A,
3d ed. (August 1977) .
4. V. Kalcevic, IT Enviroscience, Inc., Control Device Evaluation. Flares and the
Use of Emissions as Fuels (in preparation for EPA, ESED, Research Triangle
Park, NC).
5. Responses to EPA request for information on emissions from styrene and ethyl-
benzene manufacture, see Appendix E.
6. J. A. Key, IT Envirosicence, Inc., Trip Report on Visit to Cos-Mar Plant, Cosden
Oil Chemical Company, Carville, LA June 16, 1977 (on file at EPA, ESED,
Research Triangle Park, NC) .
7. J. A. Key, IT Enviroscience, Inc., Trip Report on Visit to Dow Chemical, USA,
Freeport, TX, July 28-29, 1977 (on file at EPA, ESED, Research Triangle Park, NC)
8. Responses to Texas Air Control Board 1975 Emission Inventory Questionnaire,
see Appendix E .
9. Responses to Louisiana Air Control Commission 1975 Emission Inventory
Questionnaire, see Appendix E.
10. William T. Moody, TRW, letter dated Aug. 15, 1979, to D. Beck, EPA.
11 D. G. Erikson, IT Enviroscience, Inc., Storage and Handling (in preparation
for EPA, ESED, Research Triangle Park, NC) .
12. Supplement No. ^f^r Compilation of Air Pollutant Emission Factors, AP-42,
2d ed., pp 4.5-5 to 4.4-6, EPA, OAQPS, Research Triangle Park, NC, April 1977.
13. D. G. Erikson, IT Enviroscience, Inc., Fugitive Emissions (in preparation for
EPA, ESED, Research Triangle Park, NC).
14. J. J. Cudahy and R. L. Standifer, IT Enviroscience, Inc., Secondary Emissions
(June 1980) (EPA/ESED, report, Research Triangle Park, NC) .
^Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of
that paragraph, that reference number is indicated on the material involved.
When the reference appears on a heading, it refers to all the text covered by
that heading.
-------
VI-1
VI. IMPACT ANALYSIS
A. CONTROL COST IMPACT
The purpose of this section is to present estimated costs and cost-effectiveness
data for control of benzene and total VOC emissions resulting from the production
of styrene. Details of the model plant (Fig. III-l) are given in Sects. Ill
and IV. Sample calculations are included in Appendix D.
The estimated capital costs represent the total investment, including all
indirect costs such as engineering and contractors' fees and overheads that
will be required for purchase and installation of all equipment and material to
provide a facility as described. These are battery-limit costs and do not
include the provisions for bringing utilities, services, or roads to the site,
the backup facilities, the land, the research and development required, or the
process piping and instrumentation interconnections that may be required within
the process generating the waste gas. The costs are based on a new-plant
installation; no retrofit cost considerations are included. Those costs are
usually higher than the cost for a new-site installation for the same system
and include, for example, demolition, crowded construction working conditions,
scheduling construction activities with production activities, and longer
interconnecting piping. These factors are so site-specific that no attempt has
been made to provide costs. For specific retrofit cases rough costs can be
obtained by using the new-site data and adding as required for a defined specific
retrofit situation.
The annual costs include utilities, operating labor, maintenance supplies and
labor, recovery credits, capital charges, and miscellaneous recurring costs
such as taxes, insurance, and administrative overhead. The cost factors that
were used are itemized in Table VI-1. Recovery credits are based on the market
value of the raw materials for the material being recovered.
Styrene Process
«. Alkylation Reaction Section Vent The estimated cost of a special burner
system installed in the steam superheater to thermally oxidize benzene and
total VOC emissions from the model-plant alkylation reaction section vent is
$41,000 (see Table VI-2). This cost is based on the installation of piping, a
-------
VI-2
Table VI-1. Annual Cost Parameters
Operating factor
Operating labor
Fixed costs
Maintenance labor plus
materials, 6%
b
Capital recovery, 18%
Taxes, insurances,
administration charges, 5%
Utilities
Electric power
Steam
Natural gas
Heat recovery credits
(equivalent to natural gas)
8760 hr/yr
Negligible
29% of installed
capital cost
$8.33/GJ ($0.03/kWh)
$5.50/Mg ($2.50/thousand Ib or million Btu)
$1.90/GJ ($2.00/thousand ft or million Btu)
$1.90/GJ ($2.00/million Btu)
Process downtime is normally expected to range from 5 to 15%. If the hourly
rate remains constant, the annual production and annual VOC emissions will be
correspondingly reduced. Control devices will usually operate on the same
cycle as the process. From the standpoint of cost-effectiveness calculations
the error introduced by assuming continuous operation is negligible.
on 10-year life and 12% interest.
-------
Table VI-2. Emission Control Cost Estimates for Styrene Model Plant
Emission Control Device
Source or Technique
Total
Installed
Capital
Cost
Annual
Gross
Annual
Operating
Recovery
Credits
Costs
(A)
Net
Annual
(B)
Emission
Reduction
Benzene Total VOC
(Mg/yr) (Mg/yr) (%)
(C)
Cost
Effectiveness
(per Mg)a
Benzene Total VOC
Alkylation Process heater $41,000 $12,000 $38,000 ($26,000)C 88 263 99.9+ (295)° (99)C
reaction section (A)
Column vents Flare 51,000 19,600 None 19,600 500 795 99 39 25
and styrene
storage tanks (B&E)
Emergency Flare 60,000 18,700 None 18,700 0.79 2.4 99 24,000 7,900
vent on
separator (C) H
_—. __ w
a(C) = (A) * (B).
Vent designations shown on Fig. III-l.
c
Savings.
-------
VI-4
compressor, a special burner, and all control instrumentation necessary to burn
the low-pressure vent-gas stream in the steam superheater. Since the vent-gas
rate varies directly with production, a plant twice the size of the model plant
would have twice the emissions from this vent. Curve a of Fig. VI-1 was plotted
to show the variation of installed capital cost of special burner systems with
plant capacity.
To determine the cost effectiveness of the special burner system, an estimate
was made of the gross annual cost (see Appendix D for sample calculations) and
a recovery credit was calculated from the heating value of the vent gas. For
the model plant the recovery credit is $38,000, resulting in a negative net
annual cost (savings) of $26,000 (see curve A, Fig. VI-2). The variation of
cost effectiveness with plant capacity is shown by curve A of Fig. VI-3 for
control of benzene and by curve b for control of total VOC. Both show a nega-
tive cost effectiveness (or savings) at all plant capacities.
b. Column Vents The vent streams (B, Fig. III-l) from the columns are controlled
by connecting the vents to a flare system that is also used to control the
emissions from the styrene storage-tank vents as discussed previously. The
installed capital cost of this flare system was estimated to be $51,000 (see
Table VI-2). The variation of the estimated installed capital cost of the
flare system with plant capacity is shown by curve b of Fig. VI-1. The basis
for these estimates is the installation of a complete flare system as described
in the control device evaluation report on flares. The estimated annual costs
and the cost effectiveness for VOC and for benzene emission control are given
in Table VI-2. Curve b of Fig. VI-2 is a plot of net annual cost of the flare
system vs plant capacity, and curves c and d of Fig. VI-3 show the variation of
cost effectiveness with plant capacity for benzene and for total VOC emission
control.
c. Emergency Vent on Separator Emissions from the emergency vent (C, Fig. III-l)
on the separator are controlled by a second, larger, flare system. The installed
capital cost of this flare system for the model plant was estimated to be
$60,000 (see Table VI-2). The variation of the estimated installed capital
cost of this flare system with plant capacity is shown by curve c of Fig. VI-1.
These estimates are based on the installation of a complete flare system as
-------
VI-5
0)
r^
en
10O
90
80
70
60
a>
XI
E
d)
^
O
O
O
-------
50
40
VI-6
o
0
a
E
4)
«
Q)
Q
O
O
O
w
O
O
3
C
e
<
0)
Z
30
20
10
„ 10
01
o
o
(O
O)
c
'>
CO
crt
« 20
30
40
50
100
(a) Burner system
(b) Flare for column and styrene-tank vents
(c) Flare for emergency vent on separator
200 300
Plant Capacity (Gg/yr)
400
500
Fig. VI-2. Net Annual Cost or Savings vs Plant
Capacity for Emission Controls
-------
VI-7
50,000
40,000
30,000
20,000
01
5
10,OOO
in
-------
VI-8
described above. The estimated annual costs and the cost effectiveness for VOC
and benzene emission control are given in Table VI-2. Curve c of Fig. VI-2 is
a plot of net annual cost of the flare system versus capacity, and curves e and
f of Fig. VI-3 show the variation of cost effectiveness for total VOC and for
benzene emission control with plant capacity.
d. Storage and Handling Storage and handling emissions are described below:
Benzene and toluene Model-plant benzene and total VOC emissions are controlled
by the use of floating-roof tanks for storage of benzene and toluene. The
costs for internal floating-roof tanks are contained in a companion IT Enviro-
2
science report covering storage tanks.
Ethylbenzene The control of VOC emissions from ethylbenzene storage is also
by use of an internal-floating-roof tank instead of a fixed-roof tank. The
installed capital costs, annual costs, and cost effectiveness for controlling
VOC are contained in a companion IT Enviroscience report covering storage
tanks.
Crude styrene, finished styrene, and tars The emissions from the model-plant
crude-styrene storage, styrene day tanks, styrene storage, and tar storage are
controlled by connecting the vents to the small flare system that also controls
the emissions from the column vents. The cost impact and cost effectiveness of
this flare system are discussed in Sect. VI.A.l.b.
e. Fugitive Sources A control system for fugitive sources is defined in Appendix C.
A future document will cover fugitive emissions and their applicable controls
for all the synthetic organic chemicals manufacturing industry.
f. Secondary Sources No control system has been identified for the secondary
emissions from the model plant.
2. Other Processes Styrene Coproduction with Propylene Oxidtf
No data are available to determine the cost of any control devices required to
control emissions from the Oxirane process for styrene co-production with
propylene oxide.
-------
B-
VI-9
ENVIRONMENTAL AND ENERGY IMPACTS
Styrene Process
Table VI-3 shows the environmental impact of reducing benzene and VOC emissions
by application of the described control systems to the model plant. From an
energy standpoint a typical uncontrolled styrene process will require about
10,200 kJ/kg of styrene. Individual impacts are discussed below.
Alkylation Reaction Section Vent The special burner system installed in the
steam superheater to control emissions from the alkylation reaction section
vent reduces benzene emissions by 88 Mg/yr and total VOC emissions by 263 Mg/yr
for the model plant. By utilizing the fuel value of the vent gas, 2300 MJ/hr
of energy is recovered in the model plant. The energy effectiveness is (76)
MJ/kg of VOC destroyed and (228) MJ/kg of benzene destroyed.
Column Vents, Emergency Vent on Separator, Crude-Styrene Storage, Styrene Day
Tanks, Styrene Storage, and Tar Storage The two flare systems reduce benzene
emissions from these sources by 500 Mg/yr and total VOC emissions by 800 Mg/yr
for the model plant. Generation of NO , CO, and smoke from flaring these
X
emissions can have a negative impact on the environment if steam injection is
not controlled carefully to ensure complete combustion.
Natural gas is required for the pilot lights and purge gas, and steam is used
to ensure smokeless operation. The energy consumed is equivalent to 326 MJ/hr
for the model plant, which is less than 0.1% of the total energy used by the
model plant. The electrical energy required for the flare system and the
oxygen monitor is negligible.
Other Emissions (Fugitive, Other Storage and Handling, and Secondary)—Control
methods described for these sources are floating-roof storage tanks, refrigerated
vent condenser, and repair of leaking components for fugitives. Application of
these systems results in a VOC reduction of 260 Mg/yr for the model plant. The
use of floating-roof storage tanks for emission control does not consume energy
and has no adverse environmental or energy impacts.
-------
Table VI-3. Environmental Impact of Controlled Model Plant
Emission
Source
Alkylation reaction section
Column vents
Emergency vent on separator
Storage tanks
Storage tanks
Storage tank
Other storage and
handling
Fugitive
Secondary
Stream or
Vent
Designation
(Fig. Ill-l)
A
B
C
D
E
Fl
F&G
H
K
Emission Reduction
Control Device
or Technique (%)
Process heater 99.9+
Flare 99
Flare 99
Internal-floating- 85
roof tanks
Pad vented to 99
flare
Refrigerated 80
vent condenser
None
Detection and correction 74
of major leaks
None
Benzene
(Mg/yr)
88
500
0.79
140
0.92
0.15
24
Total VOC
(Mg/yr)
263
780
2.4
150
15
<
0.30 H
o
110
Total
750
1300
-------
VI-11
2. Styrene Co-production with Propylene Oxide
Emission control systems for the Oxirane process for co-producing styrene with
propylene oxide have not been described.
-------
VI-12
C. REFERENCES*
1. V. Kalcevic, IT Enviroscience, Inc., Control Device Evaluation. Flares and
the Use of Emissions as Fuels (in preparation for EPA, ESED, Research Triangle
Park, NC).
2. D. G. Erikson, IT Enviroscience, Inc., Storage and Handling (in preparation
for EPA, ESED, Research Triangle Park, NC).
3. D. G. Erikson, IT Enviroscience, Inc., Fugitive Emissions (in preparation for
EPA, ESED, Research Triangle Park, NC).
4. "Styrene," Hydrocarbon Processing," 56(11), 227 (1977).
*Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of
that paragraph, that reference number is indicated on the material involved.
When the reference appears on a heading, it refers to all the text covered by
that heading.
-------
VII-1
VII. SUMMARY
Styrene is produced in the United States predominantly by the catalytic dehydro-
genation of ethylbenzene. Ethylbenzene is manufactured by the catalytic alkyla-
tion of benzene with ethylene, usually in the same facility. In 1977 Oxirane
started up a process that co-produces styrene and propylene oxide that accounts
for about 11% of the current domestic styrene capacity of 3,986,000 Mg/yr.
Approximately 5% of the ethylbenzene capacity is located at petroleum refineries
2
where ethylbenzene is distilled from mixed xylene streams.
The styrene and ethylbenzene production annual growth rate is estimated to be
6% through 1982. The 1977 domestic styrene capacity is sufficient to meet the
growth rate through 1982. No shortage of either benzene or ethylene is expected
during this period.
Emission sources and control levels for the model plant are summarized in
Table VII-1. The emissions projected for the domestic styrene industry in 1978
based on the estimated degree of control existing in 1978 are 2460 Mg/yr for
benzene and 5670 Mg/yr for total VOC. These emission estimates are based on
engineering judgement and data from individual styrene producers, state and
local emission control agencies, and the open literature. The following
individual projections were estimated:
1978 Emissions (Mg/yr)
Source
Process
Storage and handling
Fugitive
Secondary
Totals
Benzene
1000
1200
60
200
2460
VOC
2750
2300
340
280
5670
1C. V. Sleeth, "Styrene Monomer," Chemical and Engineering Progress 73(11),
31—35 (1977). —
2S. K. Paul and S. L. Soder, "Ethylbenzene - Salient Statistics," pp 645.3000 A—H
in Chemical Economics Handbook. Stanford Research Institute, Menlo Park, CA
(January 1977).
3S. L. Soder, "Styrene," pp 694.3052 A, B in Chemical Economics Handbook, Stanford
Research Institute, Menlo Park, CA (January 1977).
-------
VI I-2
Table VII-1. Emission Summary for Model Plant
Emission Rate (kg/hr)
Uncontrolled
Emission source
Alkylation reaction vent
Column vents
Emergency vent on
separator
Storage and handling
Fugitive
Secondary
Total
Benzene
10
5£
C.09ia
20
3.8
2.3
94
VOC
30
90
0.27a
23
17
3.0
160
Controlled
Benzene
0.0014
0.58
0.00091a
2.9
0.98
2.3
6.8
VOC
0.0041
0.90
0.0027a
3.7
4.7
3.0
12
^Average rate for entire year, based on emergency venting for 5 hr twice
a year.
-------
VII-3
The predominant emission points are the alkylation reaction vent, the column
vents, and the storage-tank vents. The alkylation reaction vent gas can be
controlled by piping it to the steam superheater for use as fuel, which would
result in a removal efficiency of greater than 99% for both benzene and VOC.
The cost of piping and a special burner for the model plant is estimated to be
$41,000. The recovery credits are greater than the annual cost; therefore the
net annual cost is negative; that is, a saving is realized. The column vents can
be controlled by a flare with an estimated destruction efficiencey of 99%. The
same flare can control the emissions from the styrene storage tanks, and its
total installed cost is estimated at $51,000.
Benzene, ethylbenzene, and toluene storage emissions can be controlled
by using covered floating-roof tanks. The emission reduction is estimated
to be 85%.
-------
A-l
APPENDIX A
Table A-l. Physical Properties of Ethylbenzene
Property
Value
Synonyms
Molecular formula
Molecular weight
Physical state
Vapor pressure
Vapor specific gravity
Boiling point
Melting point
Density
Water solubility
Ethylbenzol, phenylethane.
C8H10
106.17
Liquid
1284 Pa
3.66
136.2°C
-94.97°C
0.8670 g/ml at 20°C/4°C
b
0.014 g/100 ml of water
From: J. Dorigan, B. Fuller, and R. Duffy, "Ethylbenzene,"
p A 11-246 in Scoring of Organic Air Pollutants. Chemistry,
Production and Toxicity of Selected Synthetic Organic
Chemicals (Chemicals D-E) , MTR-7248, Rev 1, Appendix II,
MITRE Corp., McLean, VA (September 1976).
Handbook of Chemistry and Physics.
Table A-2. Physical Properties of Styrene
Property
Value
Synonyms
Molecular formula
Molecular weight
Physical state
Vapor pressure
Vapor specific gravity
Boiling point
Melting point
Density
Water solubility
Phenylethylene, vinylbenzen«
C8H8
104.14
Liquid
807 Pa
3.59
146°C
-30.63°C
0.9045 g/ml at 25°C/25°C
Insoluble
From: J. Dorigan, B. Fuller, and R. Duffy, "Styrene,"
p AIV-156 in Scoring of Organic Air Pollutants. Chemistry,
Production and Toxicity of Selected Synthetic Organic
Chemicals (Chemicals 0-Z), MTR-7248, Rev. 1, Appendix IV,
MITRE Corp., McLean, VA' '(September 1976) .
-------
A-2
a
Table A-3. Physical Properties of Benzene
Synonyms Benzol, phenylhydride, .coal naphtha
Molecular formula C^H..
6 6
Molecular weight 78.11
Physical state Liquid
Vapor pressure 95.9 mm at 25°C
Vapor density 2.77
Boiling point S0.1°C at 760 mm
Melting point 5.5°C
Density 0.8787 at 20°C/4°C
Water solubility Slight (1.79 g/liter)
J. Dorigan, B. Fuller, and R. Duffy, "Benzene," p AI-102 in Scoring of Organic
Air Pollutants. Chemistry, Production and Toxicity of Selected Organic
Chemicals (Chemicals a-c), MTR-7248, Rev 1, Appendix I, MITRE Corp.,
McLean, VA (September 1976).
-------
B-l
APPENDIX B
Table B-l. Atmospheric Dispersion Parameters for
Source
Alkylation reaction
section (uncontrolled)
Steam superheater
300,000-Mg/yr Model Plant
Emission
Rate Height Diameter
(g/sec) (m) (m)
2.8 (benzene)
8.3 (total VOC) 20 0.1
0.001 60 3
Discharge Flow
Temp. Rate
(K) (m3/sec)
320 .0.015
550 35
Discharge
Velocity
(m/sec)
1.9
4.5
Column vents (7) (uncontrolled) 2.3 ea (benzene)
Emergency vent on
separator (uncontrolled)
Storage and handling .
(Uncontrolled)
Benzene
Crude ethylbenzene
Polyethylbenzene
Residue oil
Ethylbenzene
Crude styrene
Toluene
Styrene (day) (3)
Styrene storage (2)
Tar
Flare
Storage and handling
(cantrolled)
Benzene
Crude ethylbenzene
Polyethylbenzene
Residue oil
Ethylbenzene
Crude styrene3
Toluene
Styrene (day) (3)a
3.6 ea (total VOC) 6 0.05(7)
0 (normal)
5.4 12.2 31.0
0.012 6.1 7.9
0.002 6.1 4.3
0.003 6.1 5.5
0.12 12.2 13.3
0.06 9.1 11.0
0-16 9.1 11.0
0.033 (ea) 9.1 11.6
0.13 (ea) 12.2 29.0
0.068 6.1 5.5 .
0.16 (benzene) 12 0.1
0.26 (total VOC)
0.81 12.2 31.0
0.002 6.1 7.9
0.002 6.1 4.3
0.003 6.1 5.5
0.018 12.2 18.3
0-023 9.1 11.0
320 Neg.
Ambient
305
320
340
310
320
Ambient
290
290
390
1250
Ambient
305
320
340
310
Ambient
Neg.
Styrene storage (2)'
Tara
Fugitive emissions
(uncontrolled)
Fugitive emissions
(controlled)
Secondary emissions
(uncontrolled)
1.04 (benzene)
4.8 (total VOC)
0.27 (benzene)
1.3 (total VOC)
0.63 (benzene)
0.84 (total VOC)
Ambient
Controlled by flare.
Fugitive emissions are evenly distributed over a rectangular area 50 m X 200 m.
-------
C-l
APPENDIX C
FUGITIVE-EMISSION FACTORS*
The Environmental Protection Agency recently completed an extensive testing
program that resulted in updated fugitive-emission factors for petroleum re-
fineries. Other preliminary test results suggest that fugitive emissions from
sources in chemical plants are comparable to fugitive emissions from correspond-
ing sources in petroleum refineries. Therefore the emission factors established
for refineries are used in this report to estimate fugitive emissions from
organic chemical manufacture. These factors are presented below.
Source
Uncontrolled
Emission Factor
(kg/hr)
Controlled
Emission Factor0
(kg/hr)
Pump seals ,
Light-liquid service
Heavy-liquid service
Pipeline valves
Gas/vapor service
Light-liquid service
Heavy-liquid service
Safety/relief valves
Gas/vapor service
Light-liquid service
Heavy-liquid service
Compressor seals
Flanges
Drains
0.12
0.02
0.021
0.010
0.0003
0.16
0.006
0.009
0.44
0.00026
0.032
0.03
0.02
0.002
0.003
0.00'03
0.061
0.006
0.009
0.11
0.00026
0.019
aBased on monthly inspection of selected equipment; no inspection of
heavy-liquid equipment, flanges, or light-liquid relief valves;
10,000 ppmv VOC concentration at source defines a leak; and 15 days
allowed for correction of leaks.
liquid means any liquid more volatile than kerosene.
ARadian Corp., Emission Factors and Frequency of Leak Occurrence for Fittings
in Refinery Process Units, EPA 600/2-79-044 (February 1979).
-------
D-l
APPENDIX D
COST ESTIMATE AND ENERGY SAMPLE CALCULATIONS
This appendix contains an explanation and sample calculations for the estimated
costs presented in this report.
The accuracy of an estimate is a function of the degree of data available when
the estimate is made. Figure D-l illustrates this relationship. The allowance
indicated on this chart was included in the estimated costs to cover the unde-
fined scope of the project.
The capital costs given in this report are based on a screening study, as
indicated by Fig. D-l, which used general design criteria, block flowsheets,
approximate material balances, and data on general equipment requirements.
These costs have an accuracy range of +30% to -23% and provide an acceptable
basis for determining the most cost-effective alternative.
A- USE OF ALKYLATION REACTOR VENT GASES AS FUEL-GAS SOURCE
This example is based on the estimated vent gases in the alkylation reaction
section having an emission rate of 50 kg/hr (110 Ib/hr) of total organics at
116°F. The VOC emission rate is 30 kg/hr based on the following composition:
Methane 40 wt%
Ethane 40 wt%
Benzene 20 wt%
The net heating value is 19,800 Btu/lb and the molecular weight is 24.5. The
data for propylene in the control device evaluation report on flares and the
use of emissions as fuel can be used, but it is necessary to adjust the emission
rate to a propylene equivalent because the molecular weight of propylene (42)
is greater than that of the vent gas.
110 Ib/hr X ^J^ = 190 Ib/hr of propylene.
1- Installed Capital Cost
Figure V-9 in the control device evaluation report on flares shows that the
installed capital cost a fuel-gas system for 190 Ib/hr of propylene is $41,000.
-------
D-2
2. Gross Annual Cost
Figure V-10 from the above-cited report indicates that the gross annual operating
cost for the above system is $12,000.
3. Credit
From Table VI-1 of this report the fuel credit is $2.00/million Btu. For the
vent gas with a net heating value of 19,800 Btu/lb this is equal to a credit of
$0.0396/lb of vent gas. For 100% of annual capacity operation the credit is
110 X 8760 X 0.0396 = $38,000/yr.
4. Annual Cost Summary
Gross $12,000
Credit (38,000)
Net ($26,000) savings
5. Cost Effectiveness
Cost effectiveness is the net annual operating cost ($26,000 savings) divided
by the annual benzene or VOC destroyed at 99.9+% efficiency.
a. Annual VOC Destroyed Based on a VOC emission rate of 30 kg/hr (0.003 Mg/hr)
0.03 X 8760 X 0.999 = 263 Mg/yr.
b. Cost Effectiveness (VOC) ($26,000) _ ($99)/Mg of voc destroyed.
263
c. Annual Benzene Destroyed Based on a benzene emission rate of 10 kg/hr (0.01
Mg/hr)
0.01 X 8760 X 0.999 = 88 Mg/yr.
V. Kalcevic, IT Enviroscience, Inc., Control Device Evaluation. - Flares and
the Use of Emissions as Fuels (in preparation for EPA, ESED, Research Triangle
Park, NC).
-------
D-3
d. Cost Effectiveness (Benzene) {$26,000) _ ($295)/Mg of benzene destroyed.
oo
6. Energy Effectiveness
Energy effectiveness is the net energy consumed (the energy consumed by the
compressor minus the energy recovered by use of the vent gas as fuel) divided
by the annual VOC destroyed at 99.9+% efficiency.
a. Electric Power Consumption The electrical power consumption for compressing
the vent gas must be calculated. For the model system the adiabatic horsepower
required was calculated to be 2.4 (Eq. 6-23 on page 6-16 of Perry's Chemical
Engineers Handbook, 5th ed., McGraw-Hill, was used). Based on the assumption
of an 85% electric motor efficiency and an 85% compressor efficiency and con-
verting horsepower to kilowatts (1 hp = 0.746 kW), the electric power consump-
tion rate is
2'4 X 0^85 X 05 X °-746 = 2'48 kW-
Annual energy used: convert kWh to joules (1MJ = 2.778 X 10~ kWh):
2-48 X 876°1 = 7.82 X 104 MJ/yr.
2.778 X 10
Annual energy recovered: convert Btu to joules (1 MJ = 9.48 X 10 Btu):
110 X 19,800 X 8760 = 2^ x 1Q7 MJ/yr
9.48 X 10
Net annual energy:
7.83 X 104 - 2.01 X 107 = -2.005 X 10? MJ/yr (savings).
7
v, Energy Effectiveness (VOC) -2.005 X 10 _ _?6 MJ/,kg of voc destroye(j (savings).
263 X 10
7
c Energy Effectiveness (Benzene) (2.005 X 10 ) _ _22Q MJ^kg Qf benzene destroyed
88 X 10 (savings).
-------
D-4
B. FLARE ON COLUMN VENT AND STYRENE STORAGE-TANK VENT EMISSIONS
This example is based on the estimated emissions from the column vents and
styrene storage- tank vents. The maximum emission is estimated to be 1300 Ib/hr
at 105°F and to have a molecular weight of 31. The average emission is estimated
to be 570 Ib/hr containing 200 Ib of VOC per hr of which 125 Ib/hr is benzene.
1. Installed Capital Cost
The equation given in Appendix A of the flare report was used to calculate the
flare-tip diameter (2.95 in.) that would have a pressure drop of 3 in. HO when
the vent gases are flared as described above at 1300 Ib/hr. Figure B-l in
Appendix B of the same report shows that the installed capital cost of a total
flare system of that size is $51,000.
2. Gross Annual Operating Cost
From Table VI-1 of this report the total fixed costs, including capital recovery,
is 29% of the installed capital cost:
51,000 X 0.29 = $14,800/yr.
From Fig. IV-4 of the flare report the natural gas used for the pilots is
60 scfh and for purging is 4 scfh. From Table VI-1 the cost of gas is $2.00
per thousand ft :
(60 + 4) X 8760 X = $1100/yr.
From Sect. IV-A-1 of the flare report it is estimated that 0.3 Ib of steam is
required per pound of emission; from Table VI-1 the cost of steam is $2.50/
thousand Ib. The average emission is 570 Ib/hr:
? SO
0.3 X 570 X 8760 X = $3700/yr.
The annual cost summary is as follows:
-------
D-5
Fixed $14,800
Natural gas 1,100
Steam 3,700
Total $19,600
Cost Effectiveness
Cost effectiveness is the gross annual operating cost, $19,600, divided by the
annual VOC or benzene destroyed at 99% efficiency. From Table VI-3 of this
report the total VOC reduction for the column vents and the styrene storage-tank
vents is 780 + 15 = 795 Mg/yr, and the total benzene destroyed is 500 + 0.92 =
500 Mg/yr:
= $25/Mg of VOC destroyed.
£1Q CQQ
'— = $39/Mg of benzene destroyed.
-------
E-l
APPENDIX E
LIST OF EPA INFORMATION SOURCES
H. M. Brennan, Amoco Chemicals Corp., letter to EPA, Mar. 6, 1978.
William G. Kelly, Atlantic Richfield Co., letter to EPA, Feb. 23, 1978.
L. T. Bufkin, American Hoechst Corp., letter to EPA, Jan. 26, 1978.
Charles R. Kuykendall, El Paso Products Co., letter to EPA, Jan. 31, 1978.
Frank E. Berry, Gulf Oil Chemicals Co., letter to EPA, Jan. 27, 1978.
Harry M. Keating, Monsanto Chemical Intermediates Co., letter to EPA, Apr. 28,
1978.
F. D. Bess, Union Carbide Corp., letter to EPA, May 5, 1977.
Theodore M. Nairn, Jr., Cosden Oil & Chemical Co., Texas Air Control Board
1975 Emissions Inventory Questionnaire.
Ray Warren, El Paso Products Co., Texas Air Control Board 1975 Emissions
Inventory Questionnaire.
Alvin J. Pokorny, Foster Grant Co., Louisiana Air Control Commission Emissions
Inventory Questionnaire (Oct. 13, 1975).
Harry M. Walker, Monsanto Co., Chocolate Bayou plant, Texas Air Control
Board 1975 Emissions Inventory Questionnaire.
L. R. Roberts, Texas Air Control Board Permit No. 2993, issued to Oxirane
Chemical Company (Channelview) for construction of a propylene oxide-styrene
monomer plant, Apr. 3, 1975.
J. L. Laird, Sun Oil Co. of Pennsylvania, Texas Air Control Board 1975 Emissions
Inventory Questionnaire.
Val D. Dutcher, Union Carbide Corp., Texas Air Control Board 1975 Emissions
Inventory Questionnaire.
R. C. Parnell, Amoco Chemicals Corp., EPA Questionnaire.
E. R. Hendrick, Monsanto Co., Texas City plant, EPA Questionnaire (Oct. 6, 1972).
Frank Berry, Dow Chemical Co., EPA Questionnaire (Aug. 25, 1972).
W. E. Holmes, Dow Chemical USA, Texas Division, EPA Questionnaire (Aug. 17, 1972)
James M. Black, Cos-Mar Co., EPA Questionnaire (Sept. 8, 1972).
Milton K. Dawson, Sinclair-Koppers Company, Kobuta plant, EPA Questionnaire
(Sept. 8, 1972).
J. D. Martin, Union Carbide Corp., EPA Questionnaire (Aug. 23, 1972).
-------
F-l
APPENDIX F
EXISTING PLANT CONSIDERATIONS
Table F-l lists process control devices reported in use by industry. Tables F-2
through F-4 give data on reported uncontrolled emissions from several producers
that show the variability in composition and flow.
As is described in Sect. Ill of this report, several variations of the processes
for production of ethylbenzene and styrene are possible. Some of these variations
influence the amount and rate of emissions. Such variations and the resulting
influence on emissions should be considered before it is decided to retrofit
control devices into existing plants.
The primary difficulty associated with retrofitting may be in finding space to
fit the control device into the existing plant layout. Because of the costs
associated with this difficulty it may be appreciably more expensive to retrofit
emission control systems in existing plants than to install a control system
during construction of a new plant.
-------
Table F-l. Control Devices Currently Used by the Domestic Styrene Industry'
Emission Source
Company
American Hoechst Corp.
Amoco Chemicals Corp.
Atlantic Richfield Co.
Alkylation
Reaction Section
Flare or absorber
Flare
Absorber
Column Vents
Vent condensers
Flare
Vent condensers
Emergency Vent
on Separator
NRC
Flare
NR
Storage Tank Vents
Vent condenser
NR
Floating-roof tanks
conservation vents
and
b
Cos-Mar, Inc.
Dow Chemical, USA
Process heater
Flare and process
heater
Absorber
aSee Appendix A.
bSome vents go directly to the atmosphere.
°Not reported.
^Dilute ehtylene feed stock to reactor and off-gases used as fuel.
6Gases vent to atmosphere after going through recovery section.
f During startup some vents go through a scrubber to the atmosphere.
conservation vents*3
Floating-roof tanks,
flares, and conservation
vents*3 ,
El Paso Natural Gas Co.
Gulf Oil Corp.
Monsanto Co.
Union Carbide Corp.
Boilers0
b
Flare
Absorbers
Process heater
Flare and vent
condenser
Flare
Vent condensers
Process heater and
vent condensers
NRC
Flare
Flare
NR
Conservation vents
Floating-roof tanks and
conservation vents'3
Floating-roof tanks
Vent condenser and
floating-roof tanks
-------
F-3
Table F-2. Reported Uncontrolled Emission from
Various Alkylation Reactor Ventsa
Company
Dow
Mons anto
(reactor vent)
Monsanto
(wash system vent)
Arco
Gulfc
Component
Hydrocarbons
Hydrogen chloride
C.. and C_ hydrocarbons
Inert gases
HC1 and aromatics
C and C hydrocarbons
Aromatics (includes benzene)
(Benzene)
Inert gases
Methane and ethane
Benzene
Carbon dioxide
Methane and ethane
Benzene
Hydrogen chloride
Composition
(wt %)
45 — 50
45 — 55
20
80
Trace
20.78
6.88
(5.19)
72.42
58
13
29
81
19
Some
Emission Ratio
(g/kg)
1.1 — 1.5
1.1 — 1.6
0.08
0.31 -
0.078
0.026
0.020
0.27
6.0
1.3
3.0
0.23
0.06
See Appendix E.
""g of emission per kg of ethylbenzene produced.
'Based on design data; not measured.
-------
F-4
Table F-3. Reported Uncontrolled Emissions from
Various Column Vents
Company
Monsanto Co.
El Paso
Natural Gas
Sun Oil
Amoco
Arco
Component
Organics (includes benzene)
(Benzene)
C , C , C , C / C hydrocarbons
j. ^ j 4 D
Inert gases
Benzene
Toluene
Ethylbenzene
Styrene
Water
Benzene
Other organics
Water
Inert gases
Benzene
Other organics
Inert gases
Aromatics
Other organics
Inert gases
Composition
(wt %)
46
(44)
27
27
1.7
0.8
0.8
0.8
96.1
36
21
8
35
27
9
64
58
0.88
41
Emission Ratio
(g/kg)b
5.1
' (4.9)
3.0
3.0
0.6 '
0.3
0.3
0.3
34.5
0.29
0.16
0.07
0.28
0.4
0.1
1.0
7.8
0.1
5.5
See Appendix E.
g of emission per kg of styrene produced.
'Based on design data; not measured.
Based on design calculations for 3 vacuum-column vents only; no data on
atmospheric column vent or on vents from columns in ethylbenzene purification
section.
-------
Table F-4. Reported Uncontrolled Emissions from
Two Vents on Two Vacuum-Column Jets
at Sun Oil Co.
Composition (wt %)
Component
Benzene
Other organics
Water
Inert gases
From Styrene Column
As of 4-21-76, 3:00 pm
29
11
10
50
As of 4-20-76,
46
23
12
19
From EB and BT Columns
4:00 pm As of 4-21-76, 2:00 pm
47
22
10
21
1974
Average
19
19
4
58
See Appendix E.
The discharges from the vacuum jets on the ethylbenzene (EB) column and on the benzene-toluene (BT) column are
piped together for one emission (BT) source.
i
Ul
-------
6-i
REPORT 6
CAPROLACTAM
H. S. Basdekis
IT Enviroscience
9041 Executive Park Drive
Knoxville, Tennessee 37923
Prepared for
Emission Standards and Engineering Division
Office of Air Quality Planning and Standards
ENVIRONMENTAL PROTECTION AGENCY
Research Triangle Park, North Carolina
August 1980
This report contains certain information that has been extracted from the
Chemical Economics Handbook, Stanford Research Institute. Wherever used, it
has been so noted. The proprietary-data rights, which reside with Stanford
Research Institute, must be recognized with any use of this material.
D51A
-------
6-iii
CONTENTS OF REPORT 6
Page
I- ABBREVIATIONS AND CONVERSION FACTORS 1-1
II. INDUSTRY DESCRIPTION II-l
A. Reason for Selecting Caprolactam II-l
B. Caprolactam Usage and Growth II-l
C. Domestic Producers II~3
D. References II-6
III. PROCESS DESCRIPTION III'I
A. Introduction III-l
B. Caprolactam Production Processes III-l
C. References III-7
IV. EMISSIONS IV-1
A. Model Plant for Caprolactam Production IV-1
B. Sources and Emissions IV-1
C. Process Variations IV-3
D. References IV-4
V. APPLICABLE CONTROL SYSTEMS V-l
A. Conventional Process V-l
B. Process Variation V-3
C. References V-4
VI. IMPACT ANALYSIS VI-1
A. Typical Plant VI-1
B. Industry VI~!
C. References VI~2
APPENDICES OF REPORT 6
A. PHYSICAL PROPERTIES OF CAPROLACTAM A-l
B. EXISTING PLANT CONSIDERATIONS B-l
-------
6-v
TABLES OF REPORT 6
Number
n-i
II-2
IV-1
V-l
A-l
B-l
Domestic Caprolactam Consumption
Caprolactam Producers
Uncontrolled Emissions from Model Plant
Controlled Emissions from Model Plant
Physical Properties of Caprolactam
Emission Control Devices in Current Use
Page
II-2
II-4
IV-2
V-2
A-l
B-l
Number
II-l
Hl-1
FIGURES OF REPORT 6
Locations of Caprolactam Plants
Process Flow Diagram
Page
II-5
III-3
-------
1-1
ABBREVIATIONS AND CONVERSION FACTORS
EPA policy is to express all measurements used in agency documents in metric
units. Listed below are the International System of Units (SI) abbreviations
and conversion factors for this report.
To Convert From
Pascal (Pa)
Joule (J)
Degree Celsius (°C)
Meter (m)
Cubic meter (m3)
Cubic meter (m3)
Cubic meter (m3)
Cubic meter/second
(m3/s)
Watt (W)
Meter (m)
Pascal (Pa)
Kilogram (kg)
Joule (J)
To
Atmosphere (760 mm Hg)
British thermal unit (Btu)
Degree Fahrenheit (°F)
Feet (ft)
Cubic feet (ft3)
Barrel (oil) (bbl)
Gallon (U.S. liquid) (gal)
Gallon (U.S. liquid)/min
(gpm)
Horsepower (electric) (hp)
Inch (in.)
Pound-force/inch2 (psi)
Pound-mass (Ib)
Watt-hour (Wh)
Multiply By
9.870 X 10"6
9.480 X 10"4
(°C X 9/5) + 32
3.28
3.531 X 101
6.290
2.643 X 102
1.585 X 104
1.340 X 10"3
3.937 X 101
1.450 X 10"4
2.205
2.778 X 10"4
Standard Conditions
68°F = 20°C
1 atmosphere = 101,325 Pascals
PREFIXES
Prefix
T
G
M
k
m
M
Symbol
tera
giga
mega
kilo
milli
micro
Multiplication
Factor
1012
109
106
103
10"3
10"6
Example
1 Tg =
1 Gg =
1 Mg =
1 km =
1 mV =
1 ug =
1 X 10 12 grams
1 X 109 grams
1 X 10s grams
1 X 103 meters
1 X 10"3 volt
1 X 10"6 gram
-------
II-l
II. INDUSTRY DESCRIPTION
REASON FOR SELECTING CAPROLACTAM
Caprolactam production was selected for study because the total estimated
emissions of volatile organic compounds (VOC) from its manufacture were projected
to be high and include some benzene, which is used as a solvent in the process.
Caprolactam is a solid at normal room conditions ; however , the material may be
stored or transferred in molten form. Physical property data are given in
Appendix A.
CAPROLACTAM USAGE AND GROWTH
Table II-l gives a breakdown of Caprolactam uses. Almost all the Caprolactam
manufactured in the United States is used in the production of nylon fibers,
resins, and films. In 1976, 354 Gg of Caprolactam was produced domestically.
The total domestic consumption was 337 Gg; exports amounted to 13.2 Gg; and
imports were negligible. Of the 337 Gg consumed, 311 Gg was used to produce
nylon fibers, and 25 Gg went into production of nylon resins and films. The
remainder was consumed in a variety of miscellaneous applications, such as the
formulation of specialty coatings, textile stiff eners, adhesives, floor-polish
additives, and brush bristles.1
Through 1981 Caprolactam consumption is expected to increase at the rate of 5
to 6% per year. Caprolactam consumption for nylon fiber, which is greatly
influenced by the use of nylon for rugs and carpets, is expected to increase at
5 to 6% per year. Consumption of Caprolactam for nylon resins is expected to
increase at 7 to 9% per year because of the increased demand for nylon resins
in the automotive and electronics industries.1
Caprolactam availability and cost depend on the availability and cost of the
raw materials phenol and benzene. The price of Caprolactam during 1977 was
at 52 to 54
-------
II-2
Table II-l. Domestic Caprolactam Consumption3
1976 Caprolactam Percent of
End Use Consumption (Gg) Total Consumption
Nylon-6 fibers
Nylon-6 resins and films
Miscellaneous
Total
310.7
25.0
1.4
337.1
92.2
7.4
0.4
100.0
aSee ref 1.
-------
II-3
DOMESTIC PRODUCERS
Currently three plants are being operated by three producers manufacturing
caprolactam in the United States.1 Table II-2 lists the producers, locations,
capacities, process used, and raw-material source. Figure II-l shows the plant
locations.
Caprolactam is manufactured from cyclohexanone in a three- step process. Two
variations on the conventional process are practiced, the BASF process and the
DSM/HPO (Stamicarbon) process, which differ mainly in the production of the key
intermediate and in the amount of by-product ammonium sulfate produced. All
three plants have captive production for the cyclohexanone feed, which is
manufactured, together with cyclohexanol, as a co-product by either cyclohexane
oxidation or phenol reduction.2 — 5
Allied Chemical has the largest capacity and consumes most of its caprolactam
in the production of nylon-6 fibers and resins. Nipro sells all the caprolactam
that it produces. Badische consumes about one-third of its caprolactam captively
and markets the remainder.1
-------
II-4
Table II-2. Caprolactam Producers
Company
Location
Annual Production
Capacity
(Gg) (May 1977)
Process Used
Caprolactam
Raw Material
Source
Allied Chemical
Fibers Division
Badische Co.
Nipro, Inc.
Total
Hopewell, VA
Freeport, TX
Augusta, GA
19ia
159a
140b
490
Conventional
BASF
DSM/HPO0 .
(Stamicarbon)
Phenol
Cyclohexane
Cyclohexane
See ref 1.
bSee ref 2.
Nipro also uses the conventional process for part of caprolactum production.
-------
II-5
V\V ^ \
1. Allied Chemical Corp., Hopewell, VA
2. Badische Co., Freeport, TX
3. Nipro, Inc., Augusta, GA
Fig. II-l. Locations of Plants Manufacturing Caprolactam
-------
II-6
D. REFERENCES*
1. R. F. Bradley, "Caprolactam," pp. 625.2031C—D and 625.2032A—T in Chemical
Economics Handbook, Stanford Research Institute, Menlo Park, CA (July 1977).
2. W, D. Bruce, IT Enviroscience, Inc., Trip Report for Visit to Nipro, Inc.,
Augusta, GA, April 1978 (on file at EPA, ESED, Research Triangle Park, NC).
3. F. D. Bess, Union Carbide Corp., Chemicals and Plastics Division, Taft, LA, letter
to EPA dated May 5, 1978, in response to EPA request for information on the cyclo-
hexanol/cyclohexanone process.
4. W. D. Bruce, IT Enviroscience, Inc., Trip Report for Visit to Allied Chemical Co.,
Hopewell, VA, Feb. 21, 1978 (on file at EPA, ESED, Research Triangle Park, NC).
5. W. D. Bruce and J. W. Blackburn, IT Enviroscience, Inc., Cyclohexanol/Cyclohexanone
(in preparation for EPA, ESED, Research Triangle Park, NC).
^Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of
that paragraph, that reference number is indicated on the material involved.
When the reference appears on a heading, it refers to all the text covered by
that heading.
-------
III-l
III. PROCESS DESCRIPTION
A-
INTRODUCTION
The three domestic processes for producing caprolactam utilize the same synthesis
route in that they start with cyclohexanone, but they differ in the use or prepa-
ration of key hydroxylamine salt intermediates. The conventional process uses
hydroxylamine sulfate prepared by the Rashig process as the key intermediate.
This process is used by Allied Chemical for all its production and by Nipro for
about 20% of its production. The BASF process, which is used by Badische, also
uses hydroxylamine sulfate, but it is prepared by an alternative route that
produces less ammonium sulfate by-product. In the DSM/HPO (Stamicarbon) process,
used by Nipro for about 80% of its production, hydroxylamine phosphate is the
key intermediate.1
CAPROLACTAM PRODUCTION PROCESSES
Conventional Process
In the conventional method ammonium nitrile, which is obtained by oxidation of
ammonia and absorption in ammonia carbonate solution, and ammonia are used to
absorb sulfur dioxide to produce hydroxylamine disulfonate. Subsequent hydrolysis
yields hydroxylamine sulfate and ammonium sulfate.1 Hydroxylamine sulfate is
then reacted with cyclohexanone to produce caprolactam, as shown by the following
reactions:
Reaction 1: Conventional Cyclohexanone Oximation
C
H2C CH2
H2CV
H2C
H(NH3OH)2S04 + NH4OH
(cyclohexanone) (hydroxyl- (aqueous
amine ammonia)
sulfate)
NOH
II
/\
H2C CH2
H2C CH2
H2C
(cyclohexanone
oxime)
H(NH4)2S04 + 2H20
(ammonium (water)
sulfate)
-------
III-2
Reaction 2: Beckmann Rearrangement
HC
NOH
II
/c\
H2S04'S03
(cyclohexanone (fuming sulfuric
oxime) acid)
H2C —CH2
H2C NH-H2S04
H2C jC=0
C#T
(caprolactam—
sulfuric acid)
Reaction 3: Neutralization
H2C-
HoC
H9C —
CH2
NH-H2S04
C=0
CH2
2(NH4)OH
H2C
/
H2C
H2C
C=0
CH2
(NH4)2S04 + 2H20
(caprolactam—
sulfuric acid
mass)
(aqueous
ammonia)
(caprolactam)
(ammonium (water)
sulfate)
The flow diagram for the manufacture of caprolactam by the conventional process
is shown by Fig. III-l.2—4 The indicated process vents are discussed in Sect. IV.
The cyclohexanol/cyclohexanone feed (stream 1) enters a still for separation of
cyclohexanone from cyclohexanol. The bottom product (stream 2) from the still
is rich in cyclohexanol, which is dehydrogenated to cyclohexanone in a dehydrogena-
tion reactor.5 Cyclohexanone product from the reactor contains some unconverted
cyclohexanol and is recycled (stream 3) to the cyclohexanone purification
still.
The overhead product (stream 4) from the cyclohexanone purification still goes
to an oximation reactor, where cyclohexanone is reacted with .hydroxylamine
sulfate in the presence of aqueous ammonia to form cyclohexanone oxime. The
hydroxylamine sulfate step is not shown since no organic compounds are involved.
-------
P s'x.veuT
(UUUT1STASE) '
I Fg *.!
COUV , IAUUO^UM -"<-P-°°'^
AIZ.^' "iUi.rATE
^, 4;^ £
ikrui
i ifc^
BCCVCUtD
HAPBOLACTUM
PUEIF\C*--riOU
J> J
r>rt-o.
p;oo
sic,.
TAE
STCOGE.
(MOU VOLATILE;
TO EFC.
TC.EAT
Fig. III-l. Process Flow Diagram for Manufacture of Caprolactam - Conventional Process
-------
III-4
The aqueous ammonia is present to neutralize the sulfuric acid formed during
omxiation, and this reaction yields ammonium sulfate and water. Reaction 1
illustrates the cyclohexanone oximation reaction.1
Product (stream 5) from the oximation reactor enters a phase separator, in which
the aqueous phase, containing ammonium sulfate, is separated from the cyclohexa-
none oxime—rich phase. Cyclohexanone oxime (stream 6) from the separator is
reacted with oleum (fuming sulfuric acid containing dissolved sulfur trioxide)
to yield caprolactam sulfate; see reaction 2.
Caprolactam sulfate (stream 7) from reaction 2 is reacted with aqueous ammonia
in the neutralization reactor to form caprolactam and additional ammonium sulfate;
see reaction 3. Benzene or toluene (stream 8) is also added to the neutralization
reactor as a solvent for extraction of the caprolactam product. The two-phase
mixture (stream 9) leaving the neutralization reactor is sent to a phase separator
for separation of the aqueous and solvent layers.
The solvent layer (stream 10) leaving the separator is mixed with recycled water
(stream 11) and sent to a solvent recovery still. The overhead product (stream 12)
from the still, a two-phase azeotrope of solvent and water, enters a phase sepa-
rator. The solvent layer is partly utilized for recycle to the distillation
column, and the remainder (stream 13) is recycled to the neutralization reactor.
The aqueous phase (stream 14) from the separator is stripped to remove the remaining
solvent. The aqueous phase (stream 15) from the stripping column goes to a
wastewater treatment system. The solvent overhead (stream 16) containing some
product is combined with stream 13 for recycle to the neutralization reactor.
Crude caprolactam, the bottom product (stream 17) from the solvent recovery
still, is fed to a caprolactam purification step, in which water, traces of
solvent, and traces of other impurities are removed. Caprolactam final product
(stream 18) from the purification step is stored in the molten state in a heated
storage tank.
The aqueous phase (stream 19) from the oximation separator and the aqueous phase
(stream 20) from the neutralization and extraction separator are collected in a
storage tank. An aqueous solution of ammonium sulfate (stream 21) from the
-------
III-5
tank is processed in the ammonium sulfate recovery system. Stream 21 goes to
vacuum crystallization and then to a centrifuge for removal of the mother liquor.
The crystalline ammonium sulfate product (stream 22) from the centrifuge is fed
to the next crystallization stage. Ammonium sulfate (stream 23) from the last
crystallization stage is dried, cooled, and then screened to separate the fine
and coarse product crystals. Final ammonium sulfate product (streams 24 and
25) is collected in storage bins. Scrubbers for particulate control are shown
on all three lines leading to vent I.
The details of caprolactam purification are not covered here because these steps
are considered to be confidential by the producers interviewed for this study.
Process Variations
BASF Process This process, which is used by Badische, varies from the conventional
process in the method by which hydroxylamine sulfate is prepared. Nitric oxide
is prepared by oxidation of ammonia with oxygen in the presence of steam and is
then reduced with hydrogen over a platinum catalyst on carbon suspended in a
dilute sulfuric acid solution to yield hydroxylamine sulfate.1 The production
of ammonium sulfate is practically eliminated in this step. In the rest of the
process the conventional process route is followed. The amount of ammonium
sulfate generated while in contact with VOC is the same for both processes,-
therefore the process variation is not expected to significantly affect organic
emissions.6
DSM/HPQ (Stamicarbon) Process In this process variation, which is used by
Nipro, the hydroxylamine phosphate solution is prepared by hydrogenation of
nitrate ions with the aid of a noble-metal catalyst in a buffered phosphate
solution. Hydroxylamine phosphate solution is reacted with a stream containing
cyclohexanone in the presence of toluene, which extrj^ts the product, cyclo-
hexanone oxime. The cyclohexanone oxime is then separated from the toluene by
distillation. The DSM/HPO method completely eliminates the production of ammonium
sulfate in phosphate preparation and in cyclohexanone oximation. The remainder
of the process is conventional. Benzene emissions are eliminated by the use of
toluene as solvent. No significant additional organic emissions result from
this deviation from the conventional process, as indicated by the emissions
data supplied by Nipro.2
-------
III-6
Nipro also indicates a variation in their technique for purification of caprolactam.
Crude caprolactam product is purified by ion exchange, followed by hydrogenation.
No VOC process emissions are indicated by Nipro for this segment of their caprolactam
process.2
-------
III-7
C. REFERENCES*
1. R. F. Bradley, "Caprolactam," pp 625.2031C—D and 625.3032A—T in Chemical
Economics Handbook, Stanford Research Institute, Menlo Park, CA (July 1977).
2. W. D. Bruce, IT Enviroscience, Inc., Trip Report for Visit to Nipro, Inc.,
Augusta, GA, Apr. 18, 1978 (on file at EPA, ESED, Research Triangle Park, NC).
3. Significant Organic Products, p 121, EPA-440/1-75/045.
4. M. Taverna and M. Chiti, "Compare Routes to Caprolactam," Hydrocarbon Processing,
pp 134—137 (November 1970).
5. McKetta and Cunningham, "Adipic Acid," pp 129—146 in Encyclopedia of Chemical
Processing and Design, vol 2, Dekker Publishing Co., New York, 1971.
6. Dow Badische Company, emissions data in Emissions Inventory Questionnaire,
submitted to Texas Air Control Board, Mar. 19, 1976.
*Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of
that paragraph, that reference number is indicated on the material involved.
When the reference appears on a heading, it refers to all the text covered by
that heading.
-------
IV-1
IV. EMISSIONS
Emissions in this report are usually identified in terms of volatile organic
compounds (VOC). VOC are currently considered by the EPA to be those of a
large group of organic chemicals, most of which, when emitted to the atmosphere,
participate in photochemical reactions producing ozone. A relatively small
number of organic chemicals have low or negligible photochemical reactivity.
However, many of these organic chemicals are of concern and may be subject to
regulation by EPA under Section 111 or 112 of the Clean Air Act since there are
associated health or welfare impacts other than those related to ozone formation.
A- MODEL PLANT FOR CAPROLACTAM PRODUCTION
The model plant for caprolactam manufacture has a capacity of 70 Gg/yr, based
on 8760* hr of operation per year. Although not an actual operating plant, it
has a capacity corresponding to that of the average-size process train now
being employed in caprolactam manufacture. The model plant shown by Fig. III-l
represents current technology for manufacture of caprolactam by the conventional
process.
B- SOURCES AND EMISSIONS
*• General
Uncontrolled emission rates and sources for the model plant are summarized in
Table IV-1. The emission rates were obtained by averaging information from the
three producers.1—4 Storage and handling and fugitive emissions from SOCMI
will be covered by separate EPA documents.5'6 Potential storage emission
sources are indicated on Fig. III-l by the letter J and for secondary emissions
by the letter K.
2- Process Emissions
The major emissions of benzene and total VOC occur from the cyclohexanone puri-
fication vent (A, Fig. III-l) and from the phase-separation, solvent recovery,
and stripping vents (F, Fig. III-l). Benzene occurs as an impurity in the feed
*Process downtime is normally expected to range from 5 to 15%. If the hourly
rate remains constant, the annual production and annual VOC emissions will be
correspondingly reduced. Control devices will usually operate on the same
cycle as the process. From the standpoint of cost-effectiveness calculations
the error introduced by assuming continuous operation is negligible.
-------
IV-2
Table IV-1. Uncontrolled Emissions from Model-Plant Caprolactam Manufacture by
Conventional Process
Emissions
E
Emission Source (
Cyclohexanone purification
Dehydrogenation reactor
Oximation reactor and separator
Rearrangement
Neutralization reactor
Phase separation, solvent recovery,
and stripping
Caprolactum purification
Ammonium sulfate drying0
Ammonium sulfate cooling, screening,
storage, and loadingc
Total
Stream . .a
Ratio (g/kg)
lesignation
Fig.III-1) Benzene Total VOC
A 0.16 3.1
B NAb
C 0.
D 0.
E 0.0043 0.
F 0.21 2.
G 0.
H 0.
1 0.
0.37 5.
024
00011
04
0
13
6
055
95
Rate (kg/hr)
Benzene Total VOC
1.27 24.8
NA
0.19
0.0009
0.034 0.32
1.68 16.0
1.04
4.8
0.44
2.98 47.59
g of emission per kg of caprolactam produced.
Data not available.
'Water scrubbers needed for ammonium sulfate recovery are part of the process and are
not considered to be a control device.
-------
IV-3
to the cyclohexanone purification distillation column. The total VOC emission
can vary, depending on the vacuum system design and operation. Benzene is used
as a solvent to extract the caprolactam product, and subsequent recovery is the
source of the emission from vents F.
Hydrogen produced in the cyclohexanol dehydrogenation reaction is emitted from
vent B and the stream can contain significant VOC. Data on the composition and
flow are not available. Normally this stream is burned as fuel.
The second largest process VOC emission source is the ammonium sulfate crystalli-
zation dryer vent (H). Figure III-'l indicates the use of water scrubbers on
all lines to vents H and I. These scrubbers are for particulate removal and
ammonium sulfate reovery and are considered to be part of the process.
The other VOC process emissions occur from the oximation reaction and separator
vent (C), rearrangement vent (D), neutralization reactor vent (E), caprolactam
purification vent (G), and ammonium sulfate cooling, screening, storage, and
loading vent (I).
5- Secondary Emissions
Secondary VOC emissions can result from the handling and disposal of such
process waste liquid streams as the residue from the neutralization reactor and
the streams from the solvent stripping column and the caprolactam purification
and ammonium sulfate crystallization steps. The model-plant wastewater flow
rate is estimated to be 550 Mg/hr, with a total carbon composition of 480 ppm.
The components of the wastewater include caprolactam, cyclohexanone oxime,
cyclohexanol, and cyclohexanone.1'3 Evaluation of the potential secondary
emissions from the SOCMI is covered by a secondary emissions report.7
c- PROCESS VARIATIONS
The VOC emissions from the process and DSM/HPO (Stamicarbon) processes are very
similar to the conventional process emissions except that with the DSM/HPO
process benzene is replaced by toluene as the solvent.
-------
IV-4
C. REFERENCES*
1. W. D. Bruce, IT Enviroscience, Inc., Trip Report for Visit to Allied Chemical
Company, Hopewell, VA, Feb. 21, 1978 (on file at EPA, ESED, Research Triangle Park, NC)
2. F. L. Piguet, Allied Chemical, Hopewell, VA, letter dated Sept, 28, 1979, to
D. R. Patrick (EPA).
3. W. D. Bruce, IT Enviroscience, Inc., Trip Report for Visit to Nipro, Inc.,
Augusta, GA, Apr. 18, 1978 (on file at EPA, ESED, Research Triangle Park, NC).
4. Badische Company, emissions data on cyclohexanol/cyclohexanone and caprolactam
production facilities supplied in response to EPA-114 letter request, May 12, 1978.
5. D. G. Erikson, IT Enviroscience, Inc., Storage and Handling (in preparation
for EPA, ESED, Research Triangle Park, NC).
6. D. G. Erikson, IT Enviroscience, Inc., Fugitive Emissions (in preparation
for EPA, ESED, Research Triangle Park, NC).
7. J. J. Cudahy and R. L. Standifer, IT Enviroscience, Inc., Secondary Emissions
(June 1980) (EPA/ESED report, Research Triangle Park, NC).
^Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of
that paragraph, that reference number is indicated on the material involved.
When the reference appears on a heading, it refers to all the text covered by
that heading.
-------
V-l
V. APPLICABLE CONTROL SYSTEMS
Control devices currently being used to control VOC emissions from the manufacture
of caprolactam are shown in Appendix B. The projected effect of using these
control devices to reduce the uncontrolled emissions from the conventional
process model plant (Table IV-1) is shown by Table V-l. The costs and cost
effectiveness for these applications have not been determined.
A. CONVENTIONAL PROCESS
1- Cyclohexanone Purification Vent
One of the large sources of of benzene and total VOC emissions is the cyclohexa-
none purification vent (A, Fig. III-l). A 90% reduction of emissions by use of
a vent gas condenser is reported by one producer.1 A 90% reduction efficiency
was also used to project the emission reduction for the model plant (Table V-l).
2. Dehydrogenation Reactor Vent
The emission from this vent (B) is primarily composed of hydrogen formed from
the conversion of cyclohexanol to cyclohexanone and is normally used as fuel.
Although data on the VOC content are not available, the destruction efficiency
is expected to exceed 99%. This is in agreement with the emission destruction
efficiency listed in AP-42 for natural gas burned in an industrial process
boiler.2
3. Neutralization and Solvent Processing Vents
The emissions from vents E and F contain benzene plus other VOC and are normally
controlled by vent stream condensers. For the controlled emissions listed in
Table V-l an efficiency of 70% was used for the emissions from vent E and 90%
for the emissions from vent F. These efficiencies were reported by one producer.1
4. Ammonium Sulfate Drying and Handling Vents
The emissions from vents H and I are from the water scrubbers needed to control
and recover ammonium sulfate. These water scrubbers are not considered to be
emission control devices even though they do achieve a VOC reduction of about
50%. A control device is not normally used to control the emissions from these
vents and is not included for the model plant.
-------
Table V-l. Controlled Emissions from Model-Plant Caprolactam Manufacture by Conventional Process
Emissions
Emission Source
Cyclohexanone purification
b
Dehydrogenation reactor
Oximation reactor and separator
Rearrangement
Neutralization reactor
Phase separation, solvent recovery,
and stripping
Caprolactum purification
c
Ammonium sulfate drying
Ammonium sulfate cooling, screening,
storage, and loading
Total
Stream
Designation
(Fig.III-1)
A
B
C
D
E
F
G
H
I
Control
Device or
Technique
Condenser
Used as fuel
None
None
Condenser
Condenser
None
None
None
Emission „ . , ., . a
Pcducti-n Ratl° (g/kg)
(%) Benzene
90 0.016 0
>99
0
0
70 0.0013 0
90 0.021 0
0
0
0
0.038 1
VOC
.31
.024
.00011
.012
.20
.13
.6
.055
.33
Rate (kg/hr)
Benzene VOC
0.13 2.5
0.19
0.0009
0.01 0.096
0.17 1.6
1.04
4.8
0.44
0.31 10.67
g of emission per kg of caprolactam produced.
DNo emission data available.
"Water scrubbers needed for ammonium sulfate recovery are part of the process and are not considered to be a control
device.
NJ
-------
V-3
5. Other Process Vents
Emissions from the other process vents (C, D, and G) are normally not controlled
and are left uncontrolled for the model plant. These emissions do not contain
benzene.
B- PROCESS VARIATIONS
There is only one significant emission change resulting from caprolactam manufacture
by another process. Benzene emissions from the neutralization reaction and
solvent processing steps are eliminated with the DSM/HPO (Stamicarbon) process
because toluene is used as the solvent instead of benzene.
-------
V-4
C. REFERENCES*
1. Badische Company, emission data in Emissions Inventory Questionnaire,
submitted to Texas Air Control Board, Mar. 19, 1976.
2. T. Lahre, "Natural Gas Combustion," pp. 1.4-1—1.4-3 in Compilation of Air
Pollutant Emission Factors, 3d ed., Part A, AP-42 (August 1977).
^Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of
that paragraph, that reference number is indicated on the material involved.
When the reference appears on a heading, it refers to all the text covered by
that heading.
-------
VI-1
VI. IMPACT ANALYSIS
A. TYPICAL PLANT
The environmental impact of the application of the described control system to
the model plant would be a VOC emission reduction of 323 Mg/yr and a benzene
emission reduction of 23 Mg/yr.
B. INDUSTRY1—3
The caprolactam industry appears to be using the level of control indicated for
the controlled model plant. The domestic production of caprolactam in 1979 is
projected to be 415 Gg. On this basis total industry emissions are projected
to be 16 Mg of benzene and 554 Mg of total VOC during 1979. This does not include
the fugitive, secondary, or storage and handling emissions that are expected to
be typical of SOCMI.
-------
VI-2
C. REFERENCES*
1. Badische Company, emission data in Emissions Inventory Questionnaire, submitted
to Texas Air Control Board, Mar. 19, 1976.
2. W. D. Bruce, IT Enviroscience, Inc., Trip Report to Nipro, Inc., Augusta, GA,
Apr. 18, 1978 (on file at EPA, ESED, Research Triangle Park, NC).
3. W. D. Bruce, IT Enviroscience, Inc., Trip Report to Allied Chemical Co., Hopewell,
VA, Feb. 21, 1978 (on file at EPA, ESED, Research Triangle Park, NC).
*Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of
that paragraph, that reference number is indicated on the material involved.
When the reference appears on a heading, it refers to all the text covered by
that heading.
-------
A-l
APPENDIX A
Table A-l. Physical Properties of Caprolactam
a,b
Synonyms
Molecular formula
Molecular weight
Physical state
Vapor pressure
Boiling point
Melting point
Liquid specific gravity
Water solubility
Cyclic lactam, 2-aza-
cycloheptanone,
6-aminohexanoic acid,
aminocaprotic lactam
C6H11N°
113.16
Solid
800 Pa at 120°C (6 mm Hg)
139°C at 1.60 kPa
69-71°C
1.02 at 75°C/4°C
Soluble
J. Dorigan, B. Fuller, and R. Duffy, "Caprolactam,"
p. AI-210 in Scoring of Organic Dry Pollutants.
Chemistry, Production and Toxicity of Selected
Synthetic Organic Chemicals (Chemicals A-C), MTR-7248,
Rev. 1, Appendix I, MITRE Corp., McLean, VA
(September 1976).
DP. G. Stecher, Ed., The Merck Index, 8th ed.,
Merck and Company, Inc., p 202 (1S68).
-------
B-l
APPENDIX B
EXISTING PLANT CONSIDERATIONS
Table B-l. Emission Control Devices in Current Use
Control Devices Used by
Emission Source
Cyclohexanone purification
Dehydrogenation reactor
Oximation reactor and separator
Rearrangement
Neutralization reactor
Phase separation, solvent recovery,
and stripping
Caprolactum purification
Ammonium sulfate drying
AnuTionium sulfate cooling, screening,
storage, and loading
Allied Chemical
Hopewell, VA
Condenser
Not applicable
None
None
Condenser
Condensers
Scrubber, dust
collector
None
None
Badische Co. .
Freeport, TX
Condenser
Scrubber and
incineration
None
None
Condenser
Condenser
None
None
None
c
Nipro , Inc .
Augusta, GA
Closed system
Used as fuel
None
None
No data
No data
None
None
None
See refs 1 and 2.
b,.
See ref 3.
See ref 4.
a
Water scrubbers needed for ammonium
considered to be a control device.
sulfate recovery are part of the process and are not
-------
B-2
Table B-l1—4 lists process control devices reported in use by industry. As is
described in the table, most of the control devices currently used by industry
are the same control devices used to reduce the uncontrolled emissions from the
conventional-process model plant.
Variations of the process for production of caprolactam are possible (see
Sect. III). Some of these variations influence the amount and rate of the emis-
sions. For example, in the neutralization reactor, toluene instead of benzene
is added for extraction of the caprolactam. Such variations and the resulting
influence on emissions should be considered before it is decided to retrofit
control devices into existing plants. The primary difficulty associated with
retrofitting may be in finding space to fit the control device into the existing
plant layout. Because of the costs associated with this difficulty it may be
appreciably more expensive to retrofit emission control systems in existing
plants than to install a control system during construction of a new plant.
*W. D. Bruce, IT Enviroscience, Inc., Trip Report for Visit to Allied Chemical
Company, Hopewell, VA, Feb. 21, 1978 (on file at EPA, ESED, Research Triangle
Park, NC).
2F. L. Piguet, Allied Chemical, Hopewell, VA, letter dated Sept. 28, 1979, to
D. R. Patrick (EPA).
3Badische Company emissions data on cyclohexanol/cyclohexanone and caprolactam
production facilities supplied in response to EPA-114 letter request, May 12, 1978.
4W. D. Bruce, IT Enviroscience, Inc., Trip Report for Visit to Nitro, Inc.,
Augusta, GA, Apr. 18, 1978 (on file at EPA, ESED, Research Triangle Park, NC).
-------
7-i
REPORT 7
ADIPIC ACID
H. S. Basdekis
J. W. Blackburn
W. D. Bruce
IT Enviroscience
9041 Executive Park Drive
Knoxville, Tennessee 37923
Prepared for
Emission Standards and Engineering Division
Office of Air Quality Planning and Standards
ENVIRONMENTAL PROTECTION AGENCY
Research Triangle Park, North Carolina
August 1980
This report contains certain information which has been extracted from the
Chemical Economics Handbook, Stanford Research Institute. Wherever used, it
has been so noted. The proprietary data rights which reside with Stanford
Research Institute must be recognized with any use of this material.
D2A
-------
7-iii
CONTENTS OF REPORT 7
Page
I- ABBREVIATIONS AND CONVERSION FACTORS 1-1
II. INDUSTRY DESCRIPTION II-l
A. Reason for Selection II-l
B. Usage and Growth II-l
C. Domestic Producers II-l
D. References II-5
III. PROCESS DESCRIPTION III-l
A. Introduction III-l
B. Nitric Acid Oxidation Process III-l
C. References III-7
IV. EMISSIONS IV-1
A. Model Plant for Oxidation of Cyclohexanol/Cyclohexanone IV-1
B. Sources and Emissions IV-1
C. References IV-7
V. APPLICABLE CONTROL SYSTEMS V-l
A. Process Sources V-l
B. Fugitive Sources V-3
C. Storage and Handling Sources V-3
D. Secondary Sources V-4
E. References V-5
VI. IMPACT ANALYSIS VI-1
A. Control Cost Impact VI-1
B. Environmental and Energy Impacts VI-2
VII- SUMMARY VII-1
-------
7-v
APPENDICES OF REPORT 7
Page
PHYSICAL PROPERTIES OF ADIPIC ACID, CYCLOHEXANE, CYCLOHEXANOL, A-l
AND CYCLOHEXANONE
ATMOSPHERIC DISPERSION PARAMETERS B-l
FUGITIVE EMISSION FACTORS C-l
EXISTING PLANT CONSIDERATIONS D-l
-------
7-via
TABLES OF REPORT 7
Table No. Page
" ^_«^_
ll-l Domestic Adipic Acid Consumption II-2
II-2 Adipic Acid Capacity II-2
IV-1 Model-Plant Uncontrolled Emissions IV-3
IV-2 Model-Plant Storage-Tank Data IV-5
V-l Model-Plant Controlled Emissions V-2
VII-l Model-Plant Emission Summary VII-2
A-l Physical Properties of Adipic Acid A-I
A-2 Physical Properties of Cyclohexanone A-1
A-3 Physical Properties of Cyclohexanol A~2
A-4 Physical Properties of Cyclohexane A~2
B-l Atmospheric-Dispersion Parameters for Adipic Acid Model
Plant (Capacity, 150 Gg/yr), Controlled and Uncontrolled B-l
D-l Emission Control Devices Used by Adipic Acid Producers D-2
-------
7-ix
FIGURES OF REPORT 7
Figure No. Page
II-l Location of Plants Manufacturing Adipic Acid II-3
III-l Process Flow Diagram for Manufacture of Adipic Acid by
Nitric Acid Oxidation Process in Model Plant III-3
-------
1-1
ABBREVIATIONS AND CONVERSION FACTORS
EPA policy is to express all measurements used in agency documents in metric
units. Listed below are the International System of Units (Si) abbreviations
and conversion factors for this report.
_ To Convert From
Pascal (Pa)
Joule (J)
Degree Celsius (°C)
Meter (m)
Cubic meter (m3)
Cubic meter (m3)
Cubic meter (m3)
Cubic meter/second
(m3/s)
Watt (W)
Meter (m)
Pascal (Pa)
Kilogram (kg)
Joule (J)
TO
Atmosphere (760 mm Hg)
British thermal unit (Btu)
Degree Fahrenheit (°F)
Feet (ft)
Cubic feet (ft3)
Barrel (oil) (bbl)
Gallon (U.S. liquid) (gal)
Gallon (U.S. liquid)/min
(gpm)
Horsepower (electric) (hp)
Inch (in.)
Pound-force/inch2 (psi)
Pound-mass (Ib)
Watt-hour (Wh)
Multiply By
9.870 X 10"6
9.480 X 10"4
(°C X 9/5) + 32
3.28
3.531 X 101
6.290
2.643 X 102
1.585 X 104
1.340 X 10"3
3.937 X 101
1.450 X 10~4
2.205
2.778 X 10~4
Standard Conditions
68°F = 20°C
1 atmosphere = 101,325 Pascals
PREFIXES
Prefix
T
G
M
k
m
M
Symbol
tera
giga
mega
kilo
milli
micro
Multiplication
Factor
1012
109
106
103
10"3
10"6
Example
1 Tg =
1 Gg =
1 Mg =
1 km =
1 mV =
1 X 1012 grams
1 X 10 9 grams
1 X 106 grams
1 X 103 meters
1 X 10"3 volt
1 X 10"6 gram
-------
II-l
II. INDUSTRY DESCRIPTION
A. REASON FOR SELECTION
Adipic acid production was selected for study because preliminary estimates
indicated that volatile organic compounds (VOC) and adipic acid particulate
emissions were significant. Also, although not of primary interest in this
study, significant NO emissions are generated.
A
Adipic acid is a white crystalline solid at ambient conditions; however, the pro-
duct may be stored or transferred in either solid or molten form. Other pertinent
physical property data are given in Appendix A.
B- USAGE AND GROWTH
Table II-l gives a breakdown of domestic adipic acid consumption. The predomi-
2
nant end use of adipic acid is in the manufacture of nylon 6,6. Other applica-
tions include plasticizers, synthetic lubricants, polyurethane resins, polyester
resins, and food additives.
It is estimated that adipic acid production will grow through 1981 at the rate of
4 to 5% per year, or somewhat less than the growth of nylon 6,6. A lower growth
rate for adipic acid, compared with that of nylon 6,6, can be partly attributed
to a shift in the use of raw materials for hexamethylenediamine (HMDA) produc-
tion, which can be based on precursors other than adipic acid. (HMDA and adipic
4
acid are co-monomers used in the production of nylon 6,6.)
Another factor affecting the growth of adipic acid is the availability of cyclo-
hexane, the primary raw material used in its production, which is obtained pri-
marily by benzene hydrogenation. In 1974 cyclohexane-derived adipic acid produc-
4
tion dropped 5 to 6% due to shortages of cyclohexane.
c- DOMESTIC PRODUCERS
As of mid-1978 four manufacturers were producing adipic acid in a total of five
2 4
plants; other pertinent data are given in Table II-2 and are shown in Fig. II-l.
-------
II-2
Table II-l. Domestic Adipic Acid Consumption'
End Use
Production (%)
Nylon fibers and plastics
Esters for plasticizers
Polyurethane resins
Miscellaneous
90.0
4.0
4.5
1.5
See ref 3.
Includes use of adipic acid in food additives, polyester resins,
lubricants, etc.
Table II-2. Adipic Acid Capacity
Company and Plant Location
Allied Chemical, Hopewell, VA
Celanese, Bay City, TX
Du Pont, Orange, TX
Du Pont, Victoria, TX
Monsanto Textiles, Pensacola, FL
Capacity (Gg/yr)
as of February 1978
13.6
63.5
181.4
317.5
290.3
866.3
Basic
Raw
Material
Phenol
Cyclohexane
Cyclohexane
Cyclohexane
Cyclohexane
See refs 3 and 4.
-------
II-3
(1) Allied Chemical Corp., Hopewell, VA
(2) Celanese Chemical Co., Bay City, TX
(3) Du Pont Co., Orange, TX
(4) DU Pont Co., Victoria, TX
(5) Monsanto Textiles Co., Pensacola, FL
Pig. II-l. Location of Plants Manufacturing Adipic Acid
-------
II-4
2
The overall domestic capacity for adipic acid is currently 866.3 Gg/yr, and the
demand projected for 1979 is 805.1 Gg. In approximately 95% of adipic acid pro-
duction cyclohexane is used as the raw material, with the remainder derived from
phenol. Since cyclohexane is less expensive than phenol, it will be used primar-
4
ily for future capacity increases.
2 3
The companies producing adipic acid are the following: '
1. Allied Chemical Corporation
Allied produces adipic acid from cyclohexanol and cyclohexanone derived from cap-
tive phenol. Adipic acid is also produced as a by-product of caprolactam manu-
facture. All adipic acid produced is sold.
2. Celanese Corporation
Celanese produces adipic acid from cyclohexane; they use some of their product
captively in the production of nylon 6,6 fibers and sell the remainder.
3. E. I. Du Pont de Nemours and Company, Inc.
The combined capacity of Du Font's adipic acid plants at Orange, TX, and
Victoria, TX, is 498.9 Gg/yr and represents 57.6% of the total existing domestic
adipic acid capacity. Du Pont markets some adipic acid and utilizes the rest
captively in the production of nylon 6,6 fibers and plastics.
4. Monsanto Textiles Company
Monsanto sells about 27 Gg/yr of adipic acid; the remainder is used captively to
produce plastics, adipate plasticizers, and nylon 6.6. Monsanto's adipic acid
capacity of 290.3 Gg/yr represents 33.5% of the total domestic capacity for the
product.
-------
II-5
D. REFERENCES*
j. D. F. Durocher et al., Screening Study to Determine Need for Standards of
Performance for New Adipic Acid Plants, GCA-TR-76-16-G, GCA Corp., Bedford,
MA (July 1976).
2 "Chemical Profile on Adipic Acid," p 9 in Chemical Marketing Reporter, Feb. 20,
1978.
3. K. L. Ring et al., "CEH Marketing Research Report. Adipic Acid," pp 608.5031A—
608.50330 in Chemical Economics Handbook, Stanford Research Institute, Menlo Park,
CA (April 1980).
4. J. L. Blackford, "CEH Marketing Research Report on Cyclohexane," p 638.5062K in
Chemical Economics Handbook, Stanford Research Institute, Menlo Park, CA
(February 1977).
5. Monsanto Textiles Co., letter dated June 11, 1979, to David R. Patrick (EPA).
*Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of
that paragraph, that reference number is indicated on the material involved.
When the reference appears on a heading, it refers to all the text covered by
that heading.
-------
III-l
III. PROCESS DESCRIPTIONS
A. INTRODUCTION
Starting with cyclohexane as the basic raw material, several synthesis routes to
adipic acid are possible. The route used by all adipic acid producers except
Allied Chemical is a liquid-phase, catalytic air oxidation of cyclohexane to
yield cyclohexanol and cyclohexanone (KA oil). This is followed by nitric acid
oxidation, with ammonium metavanadate and cupric nitrate used as catalysts, to
obtain adipic acid. The primary variation among producers is the catalyst they
use in the cyclohexane oxidation step. Allied makes cyclohexanol by catalytic
hydrogenation of phenol; however, the second step of adipic acid synthesis is
basically the same as that used by the other producers. Although the phenol
process for adipic acid is simple and produces fewer by-products, cyclohexane-
based processes are preferred, largely because of the lower cost of cyclohexane.
An alternate route for synthesizing adipic acid from cyclohexane (I. G. Farben
process) involves two air oxidation steps: cyclohexane is oxidized to cyclo-
hexanol and cyclohexanone; cyclohexanone and cyclohexanol are then oxidized to
adipic acid, with a mixed manganese-barium acetate used as the catalyst. Another
possible synthesis method is a direct one-stage air oxidation of cyclohexane to
adipic acid with a cobaltous acetate catalyst. Cyclohexane can be oxidized all
the way to adipic acid with nitric acid, but the yield is low and the process
requires a large quantity of nitric acid.
B. NITRIC ACID OXIDATION PROCESS
1. Basic Process
All adipic acid plants currently utilize basically the same process. Cyclohex-
anol or cyclohexanone or a mixture of the two (KA oil) is oxidized with nitric
acid in the presence of a catalyst. Chemical reactions with cyclohexanol and
cyclohexanone may be described3 as follows (the reactions are not balanced):
-------
Reaction 1:
III-2
0
A
HL-C C-H,
H0-C C-H,
2 \ / 2
C
H,,
+ HNO_
H,-C-CH,-COOH + NO + H.O
X I ^ A 4U
H2-C-CH2-COOH
(cyclohexanone) (nitric acid) (adipic acid) (nitrogen oxides)(water)
Reaction 2:
C
C-H,
H0-C C-H0
2 2
+ HNO,
H -C-CH -COOH + NO + HO
£t I £* A £*
H2-C-CH2-COOH
(cyclohexanol) (nitric acid)
(adipic acid) (nitrogen oxides)(water)
The nitrogen oxides generated by the reactions are nitric oxide, nitrogen dioxide,
and nitrous oxide. Some organic by-products are also generated, such as glutaric
acid, formic acid, acetic acid, and succinic acid.
Figure III-l is a flow diagram illustrating the basic process for adipic acid manu-
3 4
facture. ' Production of adipic acid begins with a two-stage oxidation of cyclo-
hexanol or cyclohexanone or a mixture of the two (KA oil) with a 50 wt % nitric
acid solution. The catalysts (cupric nitrate and ammonium metavanadate) are dis-
solved in the acid solution. Optimum catalyst concentrations in the acid are
4
about 0.25% copper and 0.1% vanadium.
-------
Fig. III-l. Process Flow Diagram for Manufacture of Adipic Acid by
Nitric Acid Oxidation Process in Model Plant
-------
III-4
The nitric acid (containing the catalyst) and cyclohexanol-cyclohexanone streams
(1) are fed to the first oxidation reactor in a ratio of about 40:1. The high
concentration of nitric acid solution is required to complete the oxidation and
4 5
to provide a sufficient heat sink for the highly exothermic reaction. ' The
4 5
reactor is operated at 70 to 80°C and 0.1 to 0.4 MPa. ' Reactors for the oxi-
dation either are stirred tanks or are circulating loops having a pump, heat ex-
changer, and gas-liquid separator. The reactor volume must be sufficient for
about 90% conversion of the cyclohexanol-cyclohexanone feed, and the surface area
4
must be sufficient for removal of the reaction heat.
The gas-liquid separator is required for removal of nitrogen and the nitrogen
oxides that emerge from the oxidation process. The presence of nitrogen oxides
(hence nitrous acid) is necessary for a smooth controllable reaction. The ni-
trogen oxide from the separator is oxidized by contact with air and enters
(stream 2) an absorber to be recovered as nitric acid. The effluent (stream 3)
from the first absorber is further scrubbed in a second absorber. Nitrogen and
water vapor, along with some nitric oxide and nitrogen dioxide, from the second
4 5
absorber are vented (vent A). '
The product from the first reactor passes (stream 4) through a preheater and a
gas-liquid separator before it enters (stream 5) a second oxidation reactor
4 5
operated at about 100°C and 0.1 to 0.4 MPa. ' The reactor effluent (stream 6)
then enters a bleacher, in which the dissolved nitrogen oxides are stripped from
the adipic acid—nitric acid solution with air. In addition to acting as the
stripping^agent the air, before entering the first absorber, further oxidizes
nitric oxide to nitrogen dioxide.
The adipic acid product solution emerging (stream 7) from the bleacher is col-
lected in the adipic acid solution feed tank. Material (stream 8) from the feed
tank undergoes vacuum crystallization at 30 to 70°C, followed by centrifugation
to remove the mother liquor. ' Further vacuum recrystallization of adipic acid
from water is necessary to obtain a product of sufficient purity. Wet adipic
acid (stream 9) from the last crystallization stage is dried and cooled and then
transferred to a storage bin. Emissions of adipic acid particulates occur from
the dryer (vent B), the cooler (vent C), and the product storage bin (grouped with
storage and handling emissions, which are all designated collectively with the
letter D).
-------
III-5
The mother liquor from the first crystallization step is recycled (stream 10) to
4
the first absorption tower. The mother-liquor stream (11) is continually purged
to a system for removal of other dibasic acids (primarily succinic and glutaric)
4 5
and for recovery of residual adipic acid, nitric acid, and catalysts. ' Various
schemes have been developed for recovery of residual adipic acid, nitric acid,
and catalysts.
One representative recovery scheme is to combine the dilute nitric acid (stream
12) from the second absorber, the mother liquor (containing nitric acid) (stream
13) from the second crystallization stage, and a purge stream (11) of mother
liquor from the first crystallization stage and process them for recovery of nitric
acid, adipic acid, and catalysts. The combined acid (stream 14) is vacuum-distilled
for water removal (stream 15) and recovery of nitric acid (stream 16) at approxi-
mately its azeotropic concentration in water. Stream 16 is sent to an evaporator
for further removal of the liquid phase. The overhead product (nitric acid)
(stream 17) from the evaporator is transferred to a storage tank, and the bottom
product (stream 18) from the evaporator is sent to a vacuum crystallizer for re-
covery of adipic acid. A portion of the filtrate from the centrifuge is taken
for purging the system of dibasic acids. The remainder is recycled to the nitric
acid and catalyst feed tank.
The filtrate (stream 19) from the last crystallization stage is utilized for dis-
solution (stream 20) of the recovered adipic acid. The recovered adipic acid
solution (stream 21) is recycled to the second crystallization stage.
Process Variations (Fig. III-l)
In some processes bleacher effluent (stream 7) is concentrated by vacuum strip-
ping to remove water and monobasic acids (e.g., acetic acid). The bottoms from
the still are vacuum-crystallized for recovery of product adipic acid. The mother
liquor is recycled for use in the oxidation reactors. These variations will not
affect process emissions since vapors are routed to absorber No. 1.
Buildup of dibasic acids, such as succinic and glutaric, is controlled by a portion
of the mother liquor being continually purged to a system for recovery of the
4
nitric acid, most of the adipic acid, and the catalyst.
-------
III-6
In Fig. lli-l stream 9 leaves the crystallization section and passes to a dryer
and a cooler and then to a storage bin. In some cases crystallization may be fol-
lowed by a heater to melt the adipic acid for storage in a molten state, especially
it is to be used on-site.
-------
III-7
C. REFERENCES*
1. M. E. O'Leary, "CEH Marketing Research Report on Adipic Acid," pp 608.5032A—C
in Chemical Economics Handbook, Stanford Research Institute, Menlo Park, CA,
January 1974.
2. K. Tanaka, "Adipic Acid by Single Stage," Hydrocarbon Processing 55(11),
114—119 (1974). —
3. D. F. Durocher et al., Screening Study to Determine Need for Standards of
Performance for New Adipic Acid Plants, GCA-TR-76-16-G, GCA Corp.,
Bedford, MA (July 1976).
4. McKetta and Cunningham, Encyclopedia of Chemical Processing and Design, vol 2,
pp. 129—146, Dekker Publishing Co., New York, 1971.
S. D. E. Danly and C. R. Campbell, "Adipic Acid," pp 510—531 in Kirk-Othmer
Encyclopedia of Chemical Technology, 3d ed., vol 1, edited by M. Grayson
e_t al., Wiley-Interscience, New York, 1978.
*Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of
that paragraph, that reference number is indicated on the material involved.
When the reference appears on a heading, it refers to all the text covered by
that heading.
-------
IV-1
IV. EMISSIONS
Emissions in this report are usually identified in terms of volatile organic
compounds (VOC). VOC are currently considered by the EPA to be those of a large
group of organic chemicals, most of which, when emitted to the atmosphere, partic-
ipate in photochemical reactions producing ozone. A relatively small number of
organic chemicals have low or negligible photochemical reactivity. However, many
of these organic chemicals are of concern and may be subject to regulation by EPA
under Section 111 or 112 of the Clean Air Act since there are associated health
or welfare impacts other than those related to ozone formation.
A. MODEL PLANT FOR NITRIC ACID OXIDATION OF CYCLOHEXANOL/CYCLOHEXANONE
The model plant* for this study has a capacity of 150 Gg/yr based on 8760 hr** of
operation annually and represents the approximate average capacity of the single
process train used by major domestic adipic acid producers. The process of nitric
acid oxidation of cyclohexanol and cyclohexanone represents current adipic acid
manufacturing and engineering technology.
Typical raw-material and product-storage-tank capacities were estimated for a
150-Gg/yr plant. Storage tank requirements are discussed under Sect. B-4.
Characteristics of the model plant that are important in atmospheric- dispersion
modeling are given in Appendix B. Emission considerations for existing plants
are given in Appendix D.
B. SOURCES AND EMISSIONS
1. General
Emissions from the model plant are classified as process, storage and handling,
secondary, and fugitive emissions. Process and secondary emissions from the model
plant are based on data obtained from plant-site visits and reports submitted to
19 3
the EPA Storage emissions were estimated with equations in AP-42. Fugitive
*See p 1-2 for a discussion of model plants.
**Process downtime is normally expected to range from 5 to 15%. If the hourly
production rate remains constant, the annual production and annual VOC emissions
will be correspondingly reduced. Control devices will normally operate on the
same cycle as the process. Therefore from the standpoint of cost effectiveness
calculations the error introduced by assuming continuous operation is negligible.
-------
IV-2
emissions were determined by estimating the number of pumps, process valves, and
pressure-relief valves for each unit operation and multiplying by the appropriate
factors given in Appendix C.
Emission rates and sources associated with the adipic acid model plant are listed
in Table IV-1. For each source the emissions are broken down into volatile organic
compounds (VOC) and particulates. For vent A (see Fig. III-l) only, N0x emissions
are listed since a data base is readily available and since they occur simultan-
eously with the VOC emissions.
2. Process Emissions
Process emissions occur from absorber No. 2, the adipic acid dryer, and the adipic
acid cooler (vents A, B, and C, Fig. III-l). Emissions from vent A are composed
of volatile organic compounds and nitrogen oxides. Organic emissions from vent A
are normally low because of a large excess of nitric acid, which is required in
the oxidation reactor to provide a heat sink for the exothermic reaction and to
drive the reaction to completion. To some extent monobasic carboxylic acid by-
products, such as acetic acid, leave with the oxidation reactor off-gases and go
to the absorbers. Small quantities of these compounds exit through vent A.
Process upsets causing a rapid increase in cyclohexanol and cyclohexanone emis-
sions from the oxidation reactors may cause an increase in VOC emissions from
vent A, even though oxidation of the organics by nitric acid solution occurs to
some extent in the absorbers. Hydrocarbon emissions are not increased during shut-
downs, because the cyclohexanol-cyclohexanone feed is stopped at the beginning of
the shutdown procedure.
Drying and cooling of the adipic acid product create particulate emissions from
vents B and C. According to actual process data particulate emissions are the
1 2
most significant uncontrolled emissions in the process. ' Process upsets, start-
ups, and shutdowns affect the particulate emission rate to the extent that they
affect the rate of product processed through the dryer and product cooler.
3. Fugitive Emissions
Process pumps and valves are potential sources of fugitive emissions. The model
plant is estimated to have 56 pumps (28 of which are spares) handling VOC. The
-------
Table IV-1. Model-Plant Uncontrolled Emissions
Uncontrolled Emissions
,7 4- ~r Ratioa (kg/kg) Rate (ka/hr)
Emission Emission Designation Adipic Acid N Total Adipic Acid
Source {Fig. III-l) VOC Particulateb * VOC Particulateb N°*
Absorber No. 2
Adipic acid
drying
Adipic acid
cooler
Storage and
handling
Fugitive
Secondary0
-4 -2
A 2.01 X 10 4.74 X 10 3.44 fill
—4
B 7.94 X 10 13.6
-2
C 7.83 X 10 1340.0
-5 -2
D 5.0 X 10 7.30 X 10 0.86 1250.0
-5
E 8.3 X 10 1.42
F
kg of emission per kg of adipic acid produced.
Even though particulates prior to filtration are classified with uncontrolled emissions, a bag filter is
normally used to prevent product loss. The bag filter could logically be considered a necessary part of the
process equipment, thereby essentially eliminating this uncontrolled emission.
'Emissions are less than 1 Mg/yr and are considered to be negligible.
f
U)
-------
IV-4
estimated number of process valves is 349, and the number of pressure-relief valves
is 22. The fugitive emission factors for heavy-liquid service shown in Appendix C
were applied to this valve and pump count to determine the fugitive emissions given
in Table IV-1.
4. Storage and Handling Emissions
VOC emissions result from storage and handling of cyclohexanol-cyclohexanone.
For the model plant the storage emission sources are shown on the flow diagram in
Fig. III-l (source D). The model-plant storage tanks are listed in Table IV-2.
Since cyclohexanol-cyclohexanone storage tanks in the adipic acid process are the
primary source of VOC storage emissions, storage-tank emission calculations were
based on cyclohexanol-cyclohexanone. Estimates of KA oil storage tank sizes, turn-
overs per year, and bulk liquid temperature were influenced by the data given in
4 7
EPA reports. — Emissions listed in Table IV-1 for these storage tanks are based
on fixed-roof tanks, half full, and a 10°C diurnal temperature change. Equations
from AP-42 were used for the calculations. However, breathing losses were divided
by 4 to account for recent evidence indicating that the AP-42 breathing loss equa-
g
tion overestimates emissions.
Particulate emissions resulting from transfer of adipic acid to a storage bin were
determined for the mode!
included in Table IV-1.'
determined for the model plant from data given for an existing process, and are
2
5. Secondary Emissions
In the model plant, secondary emissions occur only from aqueous effluent dis-
charged from the plant. The effluent is transferred by pipeline to a holding
pond and is then sent to deep well or by pipeline. The total aqueous effluent
is estimated to be 230 Mg/day. Of this, 59 Mg/day is nonvolatile organic acids,
such as succinic, glutaric, and adipic acid; 47 Mg/hr is nitric acid; and the
remainder is water.
To evaluate the significance of secondary organic emissions from the holding pond,
a hypothetical worst-case situation was devised. It was assumed that the aqueous
stream is composed of 25% succinic acid, the most volatile organic acid in the
stream, and that the remainder is water. A closed system, ideal solution, and
-------
Table IV-2. Model-Plant storage-Tank Datac
Tank
Process Segment Contents
KA oil feed Cyclohexanol,
cyclohexanone,
misc. hydrocarbons
Nitric acid Nitric acid.
No. of
Tanks
Required
(Fig. III-l)
1
1
2
1
1
Tank,Size
-------
IV-6
ideal gas in equilibrium with the aqueous solution were also assumed. With these
_Q
assumptions the mole fraction of succinic acid in the vapor phase is 2.0 X 10
(or 20 ppb by volume).
Additional calculations were performed in which a Henry's-law constant was esti-
mated for the vapor-liquid system and was compared with information given in an
g
article by Thibodeaux. Following this line of reasoning secondary VOC emissions
resulting from the wastewater stream were estimated to be less than 1 ton/yr. It
was then estimated that secondary emissions are negligible compared with other
emissions of organic compounds from the model plant.
-------
IV-7
C. REFERENCES*
1. W. D. Bruce, IT Enviroscience, Inc., Trip Report for Visit to Monsanto Textiles
Company, Pensacola, FL, Feb. 8, 1978 (data on file at EPA, ESED, Research
Triangle Park, NC).
2. D. R. Durocher et al., Screening Study to Determine Need for Standards of
Performance for New Adipic Acid Plants, GCA-TR-76-16-G, GCA Corp., Bedford, MA
(July 1976).
3. C. C. Masser, "Storage of Petroleum Liquids," pp 4.3-1—4.3-16 in Compilation
of Air Pollutant Emission Factors, AP-42, 3d ed. (August 1977).
4. Response by Monsanto Textiles, Pensacola, FL, to EPA questionnaire on adipic
acid, Air Pollution Control Engineering and Cost Study of the Petrochemical
Industry, OMB Approval No. 158 S 72019 (Aug. 31, 1972).
e>. Response by Celanese Chemical Co., Bay City, TX, to EPA questionnaire on adipic
acid, ibid. (Aug. 15, 1972).
6 Response by Du Pont, Orange City, TX, to EPA questionnaire on adipic acid, ibid.
(Sept. 18, 1972).
7. J. M. Mullins, Celanese Chemical Co., Bay City, TX, Texas Air Control Board
1975 Emissions Inventory Questionnaire (July 9, 1976).
8. E. C. Pulaski, TRW, letter dated May 30, 1979, to Richard Burr (EPA).
9. L. J. Thibodeaux, "Air Stripping of Organics from Wastewater. A Compendium,"
pp 358—378 in The Proceedings of the Second National Conference on Complete
Watereuse. Water's Interface with Energy, Air and Solids, Chicago, IL, May 4—8,
1975, sponsored by AIChE and EPA Technology Transfer.
*Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of
that paragraph, that reference number is indicated on the material involved.
When the reference appears on a heading, it refers to all the text covered by
that heading.
-------
V-l
V. APPLICABLE CONTROL SYSTEMS
A. PROCESS SOURCES
Process emissions from the model plant occur from vents A, B, and C. Vent A emits
both NO and volatile organic carbon compounds; vents B and C emit adipic acid
X
participate matter. (See Fig. III-l for vent locations and Table IV-1 for uncon-
trolled emissions.)
Emissions from vent A can be controlled with a thermal reduction unit, by a scrub-
ber, by combustion in a boiler, or by flaring. A recent study by the GCA Corpora-
tion indicates that thermal reduction is the best proven means for control of NO
x
emissions from adipic acid plants. Efficiency of NO removal for this type of
X
device is about 97% but varies with the inlet NO concentration. Thermal reduction
X
units can take any inlet concentration of NO and reduce the outlet NO concentra-
X X
tion to about 500 ppm by volume, even at high flow rates. The Monsanto plant is
presently running its thermal reduction unit at 1500 ppm because of problems
with the ceramic liner cracking but it has been run at 500 ppm. Comments in the
GCA report concerning the effectiveness of combustion techniques for removal of
hydrocarbons indicate that VOC emissions from vent A can be removed with almost
100% efficiency in a thermal reduction unit; this value was used for calculation
of model-plant VOC emissions from vent A (see Table V-l). The control devices
currently used by adipic acid producers is shown in Appendix D.
In the thermal reduction unit, off-gases containing nitrogen oxides are heated to
high temperatures and reacted with excess fuel (natural gas) in a reducing atmosphere.
Three steps are involved in this process: At ~2600°F nitrogen oxides are reacted
with excess fuel under reducing conditions to form water vapor, elemental nitrogen,
and carbon dioxide; the gases are cooled to 1400°F by a heat recovery unit; and
the excess fuel is combusted, usually in the presence of a catalyst.
Scrubbers are used in industry for NO emissions control, but the lower achiev-
A
able concentration limit with the best available systems is approximately 1000 ppm
(by volume). For removal of low concentrations of N0x (less than 1 vol %) at high
flow rates, NO removal efficiency is about 70%. In some adipic acid plants N0x
emissions are routed to a powerhouse boiler, a combustion technique similar to
thermal oxidation. In the GCA report this technique is estimated to be about 70%
-------
Table V-l. Model-Plant Controlled Emissions
Tyixj
of
Emission
Process
Process
Process
Storaqo and
handl ing
Fugitive
Secondary
Totals
Emission
Source
Absorber Ho. 2
Adipic acid dryer
Adipic acid cooler
Storage tanks
or bins
Pumps , valves ,
pressure-relief
valves
Holding pond
Vent or
Emission
Designation
(Fig. III-l)
A
B
C
D
E
F
Control
Device
or
Technique
Thermal reduction
unit
Wet scrubber
Bag filter
Floating roofs on
KA stg. tanks
Bag filter on adip-
ic acid stg. bin
Detection and correc-
tion of major
leaks
Hone
Controlled Emissions
Emission Emission Rate (kg/hr) Emission Ratio (kg/kg/)3
Reduction Par- Par-
(%) VOC ticulate £ VOC ticulates N°x
98 N
-------
V-3
efficient for removal of N0x- The use of flares is also reported, but they are
only about 70% efficient for the removal of NO -1
X
Removal of VOC from the stream leaving vent A with the scrubber, used for removal
of NO , would be inefficient since water is used as the scrubbing solvent, and
X
cyclohexanol, cyclohexanone, and cyclohexane have low water solubilities. If a
boiler or a flare is used for NO removal, about 99% VOC reduction is achievable.
X
Emissions from vents B and C are adipic acid particulates from the dryer and pro-
duct cooler, respectively, which are controlled industrially by wet scrubbers and/or
bag filters. The typical efficiency of bag filters for the removal of particulates
is greater than 99%, provided that the relative humidity of the stream is low.
High relative humidity results in caking and consequent operating problems with
the filters. The use of bag filters on humid air streams necessitates heating
the stream or diluting it with relatively dry air to prevent caking and related
problems. In contrast the efficiency of a wet scrubber may range from about 80
to 95%, depending on its design, but there are no potential operational problems
from caking or from high-energy consumption to prevent condensation. Model-plant
controlled emissions from the dryer are based on data for a wet scrubber with about
80% particulate removal efficiency. Model-plant controlled emissions from the
dryer product stream are based on a bag filter with a particulate removal efficiency
of 99.98%.1
B. FUGITIVE SOURCES
Controls for fugitive sources will be discussed in a separate document covering
fugitive emissions from the synthetic organic chemicals manufacturing industry.
Controlled fugitive emissions calculated with the factors given in Appendix C are
included in Table V-l. These factors are based on the assumption that any major
leaks will be detected and repaired.
C. STORAGE AND HANDLING SOURCES
2
Controls for storage emissions are discussed in another EPA document. Control
for VOC storage losses involves the use of new, floating-roof tanks,* or other
appropriate control devices. The VOC storage emissions listed in Table V-l for
*Consist of internal floating covers or covered floating roofs as defined in
API 25-19, 2d ed., 1976 (fixed-roof tanks with internal floating device to
reduce vapor loss).
-------
V-4
the model plant were calculated by assuming that 85% control can be provided by a
contact type of internal floating roof with secondary seals. If the storage tem-
perature were reduced to about 40°C, the need for a contact type of internal float-
ing roof may be eliminated.
An alternative technique (not currently employed) for controlling VOC storage emis-
sions in the adipic acid model plant consists of routing them to the thermal reduc-
tion unit. Combustion of the hydrocarbons would virtually eliminate VOC storage
emissions but would not have the advantage of recovery offered by floating- roof
tanks.
Emissions of adipic acid particulates result from the product storage bin. In
the model plant a bag filter is used to control particulate emissions from the
bin. A particulate removal efficiency of 99.96% was used for calculation of the
controlled emissions given in Table V-l.
D. SECONDARY SOURCES
Aqueous effluent from the adipic acid manufacturing process contains dibasic organic
3 4
acids, such as adipic, succinic, and glutaric. ' Since these compounds are essent-
ially nonvolatile, air emissions resulting from an aqueous solution of them are
insignificant, and no control device or technique is considered to be necessary
for VOC control.
-------
V-5
E. REFERENCES*
1. D. F. Durocher et al., Screening Study to Determine Need for Standards of
Performance for New Adipic Acid Plants, GCA-TR-76-16-G, GCA Corp., Bedford, MA
(July 1976).
2. D. G. Erikson, IT Enviroscience, Inc., Storage and Handling (in preparation
for EPA, ESED, Research Triangle Park, NC).
3. W. D. Bruce, IT Enviroscience, Inc., Trip Report for Visit to Monsanto Textiles
Company, Pensacola, FL, Feb. 8, 1978 (data on file at EPA, ESED, Research
Triangle Park, NC).
4. Response by Monsanto Textiles Co., Pensacola, FL, to EPA questionnaires on adipic
acid, Control Engineering and Cost Study of the Petrochemical Industry, OMB
Approval No. 158 S 72019 (Aug. 31, 1972).
*Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of
that paragraph, that reference number is indicated on the material involved.
When the reference appears on a heading, it refers to all the text covered by
that heading.
-------
VI-1
VI. IMPACT ANALYSIS
A. CONTROL COST IMPACT
This section gives the cost considerations for control of VOC emissions resulting
from the production of adipic acid. Details of the model plant (Fig. III-l) are
covered in Sects. Ill and IV.
1. Absorber No. 2—Vent A
A thermal reduction unit for NO control can also be used for elimination of VOC
x
emissions, such as those that occur from vent A, Fig. III-l, of the model adipic
acid plant. With a thermal reduction unit already in existence for NO control
X
no additional cost is incurred for VOC control. Consequently, cost-effectiveness
calculations for VOC removal are unnecessary.
2. Cyclone Separator Vent and Fluidized-Bed Cooler Vent (Vents B and C, Fig. III-l)
Emissions from vents B and C are entirely adipic acid particulates and are not
regarded as VOC emissions. The same is true for the vent on the adipic acid
product storage bin, located just after the fluidized bed cooler. For economic
reasons adipic acid plants employ control devices on these vents, and the control
devices are considered to be an integral part of the process. For the reasons
indicated in Sect. 1 cost-effectiveness calculations were not performed for the
bag filters on those vent streams.
3. Storage and Handling
Model-plant cyclohexanol-cyclohexanone storage emissions are controlled by the
use of floating-roof tanks. Installed capital cost, net annual cost, and cost-
effectiveness data are contained in another EPA report covering storage of chemi-
cals.
b. Fugitive Sources—A control system for fugitive sources is defined in Appendix C.
Another document will cover fugitive emissions and their applicable controls for
all the synthetic organic chemicals manufacturing industry.
c. Secondary Sources—No control system has been identified for the secondary emis-
sions from the model plant.
-------
VI-2
B. ENVIRONMENTAL AND ENERGY IMPACTS
Control of emissions from storage tanks is achieved through the use of floating-
roof tanks, which do not consume energy and have no adverse energy or environ-
mental impacts.
Control of fugitive emissions is achieved by the prompt correction of leaks on
pumps, agitators, and compressors. These control techniques do not have adverse
energy or environmental impacts.
Use of a thermal reduction unit for NO and VOC, of floating-roof storage tanks,
A
and of fugitive-emission control techniques results in a VOC emissions reduction
of 36.6 Mg/yr for the model plant.
-------
VII-1
VII. SUMMARY
Adipic acid is currently produced exclusively from nitric acid oxidation of cyclo-
hexanol and cyclohexanone. Other synthesis routes to adipic acid are feasible,
such as two-stage air oxidation from cyclohexane, but are not now being utilized
industrially.
The annual growth rate of adipic acid through 1981 is estimated to be 4 to 5%.
As indicated in Sect. II, the current capacity for adipic acid is 866.3 Gg/yr,
and the demand for 1979 is projected to be 805.1 Gg/yr. Slight capacity increases
by process modifications will be required for increased production capacity, with
the projected demand assumed to be accurate. If growth continues to increase at
the projected rate, some plant expansions will be required.
About 90% of the adipic acid manufactured is used in the production of nylon 6,6
fibers and plastics. Consequently, growth of adipic acid is strongly dependent
on the demand for nylon 6,6. Availability of basic raw materials, such as cyclo-
hexane and phenol, can also be influential economic factors.
Emission sources and control levels for the adipic acid model plant are sum-
marized in Table VII-1. The predominant VOC emissions points in the model plant
are from absorber No. 2 (vent A), fugitive sources (i.e., pumps, valves, and
pressure-relief valves), and storage tank sources. Projection of these emission
values for the entire domestic adipic acid industry at the estimated 1979 demand
would result in VOC emissions of 30.7 kg/hr for all uncontrolled plants. It is
estimated that VOC emissions from the adipic acid industry are currently 73% con-
trolled. The estimated current emission rate of VOC, based on the projected product
demand for 1979, is 8.3 kg/hr.
With model-plant process VOC emissions assumed to be controlled in existing equip-
ment used for NO control and fugitive emissions to be controlled by increased
A
H. E. O'Leary, "CEH Marketing Research Report on Adipic Acid," pp 608.5032A— K
in Chemical Economics Handbook, Stanford Research Institute, Menlo Park, CA
(January 1974).
2J. L. Blackford, "CEH Marketing Research Report on Cyclohexane," p 638.5062K in
Chemical Economics Handbook, Stanford Research Institute, Menlo Park, CA
(February 1977).
-------
VII-2
Table VII-1. Model-Plant Emission Summary0
Emission Source
Absorber No. 2
VOC
3.44
Emission
Uncontrolled
NOX Particulate
811
Rate (kg/hr)
Controlled
VOC NOX Particulate
16.4
Adipic acid dryer
Adipic acid cooler
Storage and handling
Fugitive
Total
0.86
0.86
5.16 811
13.6
1340
1250
2600
0.12
0.86
0.98
16.4
2.79
0.27
0.55
3.61
a
All emissions are based on a plant operation rate of 8760 hr/yr; secondary
emissions are less than 1 Mg/yr and are considered to be negligible.
-------
VI I-3
maintenance, the only additional significant VOC emissions to be controlled are
storage emissions. Cyclohexanol-cyclohexanone storage is controlled in the model
plant by floating-roof tanks, which would allow an overall VOC emission reduction
of approximately 85% to be attained.
-------
A-l
APPENDIX A
Table A-l. Physical Properties of Adipic Acid3
Synonyms Hexaneodioic acid, adipinic
acid; 1,4-butane—dicar-
boxylic acid
Molecular formula ^fi^in0^
Molecular weight 146.14
Physical state White, crystalline solid
Vapor pressure 133 Pa at 159.5°C
Vapor density 5.04 (air = 1)
Boiling point 337.5°C
Melting point 153°C
Density 1.360 at 25°Cb
Water solubility Slight (15 g/liter)
J. Dorigan, B. Fuller, and R. Duffy, p Al-42 in Scoring of Organic Air
Pollutants. Chemistry, Production and Toxicity of Selected Synthetic
Organic Chemicals (Chemicals A—C), MTR-7248, Rev. 1, Appendix I,
MITRE Corp., McLean, VA (September 1976).
b
With reference to water at 4°C.
Table A-2. Physical Properties of Cyclohexanone
Synonyms Ketohexamethylene,
cyclohexylketone
pimelic ketone,
hexanon
Molecular weight 98.14
Physical state Liquid
Vapor pressure 4.77 mm at 25°C
Vapor density 3.4 (air = 1)
Boiling point 155.6°C
Melting point -47°C
Density 0.9478 at 20°Cb
Water solubility 50g/liter
aj. Dorigan, B. Fuller, and R. Duffy, p AI-322 in Scoring of Organic Air
Pollutants. Chemistry, Production and Toxicity of Selected Synthetic
Organic Chemicals (Chemicals A—C), MTR-7248, Rev. 1, Appendix I,
MITRE Corp., McLean, VA (September 1976).
With reference to water at 4°C.
-------
A-2
Table A-3. Physical Properties of Cyclohexanol
Synonyms Hexahydrophenol,
hexalin
Molecular weight 100.17
Physical state Liquid
Vapor pressure 1.7 mm at 25°C
Vapor density 3.45 (air = 1)
Boiling point 161.5°C
Melting point 23°C
Density 0.9449 at 25°Cb
Water solubility Yes
J. Dorigan, B. Fuller, and R. Duffy, p AI-320 in Scoring of Organic Air
Pollutants. Chemistry, Production and Toxicity of Selected Synthetic
Organic Chemicals (Chemicals A—C), MTR-7248, Rev. 1, Appendix I,
MITRE Corp., McLean, VA (September 1976).
b
With reference to water at 4°C.
Table A-4. Physical Properties of Cyclohexane
Synonyms Hexahydrobenzene,
hexanaphthene,
hexamethylene
Molecular weight 84.16
Physical state Liquid
Vapor pressure 98.14 at 25°C
Vapor density 2.90 (air = 1)
Boiling point 80.7°C
Melting point 6.3°C
Density 0.77855 at 20°Cb
Water solubility Insoluble (
-------
Table B-l. Atmospheric-Dispersion Parameters for Adipic Acid Model Plant
(Capacity, 150 Gg/yr), Controlled and Uncontrolled
Emission
Source
Process absorber
vent A
(Fig. III-l)
Cyclohexanol/
cyclohexanone
storage tanks (5)
1
2
3
4
5
Fugitive
Emission
Rate
(g/sec)
57.3
1.24
12.50
0.13
0.13
0.33
23.7
Tank
Height
(m)
4.9
4.9
4.9
4.9
7.3
Tank Stack Stack
Diameter Height Diameter
(m) (m) (m)
Uncontrolled Emissions
30 1.05
4.6
4.6
4.6
4.6
6.7
Controlled Emissions
Discharge Flow
Temperature rate
(K) (m3/sec)
330
338
361
338
338
338
310--375
Discharge
Velocity
(m/sec)
Process absorber
vent A
(Fig. III-l)
Cyclohexanol/
cyclohexanone
storage tanks (5)
1
2
3
4
5
30
1.05
533
8.64
10
0.19
1.88
0.02
0.02
0.05
4.9
4.9
4.9
4.9
7.3
4.6
4.6
4.6
4.6
6.7
338
361
338
338
338
Fugitive
23.7
310—375
-------
C-l
APPENDIX C
FUGITIVE-EMISSION FACTORS*
The Environmental Protection Agency recently completed an extensive testing
program that resulted in updated fugitive-emission factors for petroleum re-
fineries. Other preliminary test results suggest that fugitive emissions from
sources in chemical plants are comparable to fugitive emissions from correspond-
ing sources in petroleum refineries. Therefore the emission factors established
for refineries are used in this report to estimate fugitive emissions from
organic chemical manufacture. These factors are presented below.
Source
Uncontrolled
Emission Factor
(kq/hr)
Controlled
Emission Factor'
(kg/hr)
Pump seals t
Light-liquid service
Heavy-liquid service
Pipeline valves
Gas/vapor service
Light-liquid service
Heavy-liquid service
Safety/relief valves
0.12
0.02
0.021
0.010
0.0003
0.03
0.02
0.002
0.003
O.Oti'03
Gas/vapor service
Light-liquid service
Heavy- liquid service
Compressor seals
Flanges
Drains
0.16
0.006
0.009
0.44
0.00026
0.032
0.061
0.006
0.009
0.11
0.00026
0.019
Based on monthly inspection of selected equipment; no inspection of
heavy-liquid equipment, flanges, or light-liquid relief valves;
10,000 ppmv VOC concentration at source defines a leak; and 15 days
allowed for correction of leaks.
Light liquid means any liquid more volatile than kerosene.
*Radian Corp., Emission Factors and Frequency of Leak Occurrence for Fittings
in Refinery Process Units, EPA 600/2-79-044 (February 1979).
-------
D-l
APPENDIX D
EXISTING PLANT CONSIDERATIONS
Table D-l ' lists process control devices reported in use by industry. Most
of them are the same control devices used to reduce the uncontrolled emissions
from the process model plant.
The largest source of uncontrolled VOC emission is from absorber No. 2 (vent A);
however, the NO reductions from the vent also result in reduction of VOC
A
emissions. In the model plant it was assumed that a thermal reduction unit
could be used to reduce NO emission, as well as VOC emissions. A thermal
A
reduction unit is currently being used by Monsanto and was designed to reduce
NO emissions to 500 ppm. The Monsanto unit is presently running at 1500 ppm
X
NO because of problems with the ceramic liner cracking, but it can and has
X
been run at 500 ppm. At the lower NO concentrations (500 ppm) the unit will
X
be down one-half the time because of thermal cracking, whereas if the unit runs
at 2500 ppm NO , then it will be operating almost full time. In Du Font's
X
plants powerhouse boilers are used to reduce NO emissions to the 2500—5000 ppm
A
range. The boilers are designed to provide steam, to operate while the adipic
acid process is in operation, and to recover most of the energy value of their
fuel. The thermal reduction unit wastes most of the supplemental fuel's energy
value because it has no provision for heat recovery or the waste heat recovery
unit is not as efficient as a boiler.
Celanese states that in a properly designed and operated plant with a high-
pressure absorber there should be essentially no volatile organic emission from
the adsorber. Three reasons are given: there is essentially total conversion
of the volatile organics in the reactor, the reactor off-gas is scrubbed by
D. F. Durocher ejt al., Screening Study to Determine Need for Standards of
Performance for New Adipic Acid Plants, GCA-TR-76-16-G, GCA Corp., Bedford, MA
(July 1976).
2F. L. Piguet, Allied Chemical, Hopewell, VA, letter dated June 29, 1979, to
D. R. Patrick (EPA).
3j. R. Cooper, Du Pont, Wilmington, DE, letter dated July 24, 1979, to
D. R. Patrick (EPA).
-------
Table D-l. Emission Control Devices Used by Adipic Acid Producers*
Company
Allied Chemical
Celanese
Du Pont, Orange, TX
Du Pont, Victoria, TX
Monsanto
See refs 1 and 2.
Absorber off-gas routed
Vent A Absorber Number 2
NOX VOC
Reduction Reduction
Control Device (%) (%)
b
c
Powerhouse 70 99
Powerhouse 70 99
Thermal reduction 97 99+
unit
to caprolactam process; no NO emission.
Vent B Dryer
Control Device
Wet scrubber
Wet scrubber
Bag filter
Wet scrubber
Vent C Cooler
Control Device
Not applicable
Bag filter
Bag filter
Wet scrubber
D
to
X
'Efficiency of nitric acid recovery section "v-95%; no further NO control required.
q —
Absorber off-gas routed to powerhouse, which also burns waste gases from other processes.
-------
D-3
incoming blend acid feed, and the reactor off-gas is further scrubbed in a high-
4
pressure absorber.
As is described in Sect. Ill of this report, variation of the process for the
production of adipic acid is possible. Some of these variations, for example,
in the design and operation of the reaction off-gas scrubbers, influence the
amount and rate of the emissions. Such variations and resulting influence on
emissions should be considered before it is decided to retrofit control devices
into existing plants.
The primary difficulty associated with retrofitting may be in finding space to
fit the control device into the existing plant layout. Because of the costs
associated with this difficulty it may be appreciably more expensive to retro-
fit emission control systems in existing plants than to install a control
system during construction of a new plant.
Celanese Chemical Company, Bay City, TX, letter dated June 25, 1979, to
D. R. Patrick (EPA).
-------
TECHNICAL REPORT DATA
(Please read Inunctions on the reverse before completing)
REPORT NO.
EPA-450/3-80-028a
2.
I. RECIPIENT'S ACCESSION NO,
TITLE AND SUBTITLE
Organic Chemical Manufacturing
Volume 6: Selected Processes
5. REPORT DATE
1
6. PERFORMING ORGANIZATION CODE
AUTHOR(S)
3. W. Blackburn,Bvf°talcevic, S. W. Dylewski, R. E. White
J. F. Lawson, J. A. Key, F. D. Hobbs, H. S. Basdekis,
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
IT Enviroscience, Inc.
9O41 Executive Park Drive
Suite 226
Knoxville, Tennessee 37923
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-02-2577
12. SPONSORING AGENCY NAME AND ADDRESS
DAA for Air 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
Final
14. SPONSORING AGENCY CODE
EPA/200/04
15. SUPPLEMENTARY NOTES
16. ABSTRACT
EPA is developing new source performance standards under Section 111 of
the Clean Air Act and national emission standards for hazardous air pollutants
under Section 112 for volatile organic compound emissions (VOC) from organic
chemical manufacturing facilities. In support of this effort, data were gathered
on chemical processing routes, VOC emissions, control techniques, control costs,
and environmental impacts resulting from control. These data have been analyzed
and assimilated into the ten volumes comprising this report.
This volume presents in-depth studies of several major organic chemical
products.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b. IDENTIFIERS/OPEN ENDED TERMS
c. COS AT i Field/Group
13B
!;£.. DISTRIBUTION STATEMENT
! Unlimited Distribution
19. SECURITY CLASS (This Report)
Unclassified
!1. NO. OF PAGES
404
20. SECURITY CLASS (TMs page)
Unclassified
22. PRICE
EP* Forn2220-l (Rcv.<-77) PREVIOUS EDITION >s
------- |