United States Office of Air Quality EPA-450/3-80-028b
Environmental Protection Planning and Standards December 1980
Agency Research Triangle Park NC 27711
_
4
&ERA Organic Chemical
Manufacturing
Volume 7: Selected
Processes
-------
EPA-450/3-80-028b
Organic Chemical Manufacturing
Volume 7: Selected Processes
Emission Standards and Engineering Division
U S. Environmental Protection Agenc»
Region 5, Library (PL-12J)
77 West Jackson Boulevard,
Chicago, IL 60604-3590
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
-------
U.S. Environmental Protection
-------
Ill
This report was furnished to the Environmental Protection Agency by IT Enviro-
science 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 Environmen-
tal 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 North
Carolina 27711, or from National Technical Information Services, 5285 Port
Royal Road, Springfield, Virginia 22161.
D124R
-------
V
CONTENTS
Page
INTRODUCTION vii
Product Report
1. NITROBENZENE 1~i
2. ANILINE 2~1
3. CUMENE 3~i
4. TOLUENE DIISOCYANATE 4~i
5. CRUDE TEREPHTHALIC ACID, DIMETHYL TEREPHTHALATE,
AND PURIFIED TEREPHTHALIC ACID 5-i
6. PHENOL/ACETONE 6~1
7. LINEAR ALKYLBENZENE 7~1
-------
Vll
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 1
Volume 2
Volume 3
Volume 4
Volume 5
Volume 6-10
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: nitrobenzene, aniline, cumene, toluene diisocyanate, terephthalic
acid, dimethyl terephthalate, phenol, acetone, and linear alkylbenzene. 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 I
NITROBENZENE
F. D. Hobbs
C. W. Stuewe
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
February 1981
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.
D15A
-------
CONTENTS OF REPORT 1
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. References II-6
III. PROCESS DESCRIPTIONS III-l
A. Introduction III-l
B. Nitration of Benzene III-l
C. Process Variations III-4
D. References III-5
IV. EMISSIONS IV-1
A. Nitrobenzene Model Plants IV-1
B. Sources and Emissions IV-1
C. Effects of Process Variations on Emissions IV-6
D. References IV-7
V. APPLICABLE CONTROL SYSTEMS V-l
A. Process Sources V-l
B. Fugitive Sources V-3
C. Storage Sources V-4
D. Secondary Sources V-4
E. Control Devices Used by Industry V-4
F. References V-5
VI. IMPACT ANALYSIS VI-1
A. Environmental and Energy Impact VI-1
B. Control Cost Impact VI-3
C. Reference VI-9
VII. SUMMARY VII-1
-------
1-v
APPENDICES OF REPORT 1
Page
A. PHYSICAL PROPERTIES OF NITROBENZENE A-l
B. AIR-DISPERSION PARAMETERS B-l
C. FUGITIVE-EMISSION FACTORS C'1
D. COST ESTIMATE DETAILS AND CALCULATIONS D-l
E. EXISTING PLANT CONSIDERATIONS E-l
-------
1-vii
TABLES OF REPORT 1
Number
II-l Aniline Usage and Growth
II-2 Nitrobenzene Capacity
IV-1 Uncontrolled Benzene and Total VOC from Nitrobenzene
Model Plants
IV-2 Storage Parameters for Determining Model-Plant Emissions
V-l Controlled Benzene and Total VOC Emissions from
Nitrobenzene Model Plant
VI-1 Environmental Impact of Controlled Model Plants
VI-2 Cost Factors Used in Computing Annual Costs
VI-3 Emission Control Analyses for Nitrobenzene Model Plants
VII-1 Summary of Emissions for the Model Plants
A-l Physical Properties of Nitrobenzene
B-l Air-Dispersion Parameters for 30,000-Mg/yr
Nitrobenzene Model Plant
Page
II-2
II-3
IV-2
IV-5
V-2
VI-2
VI-4
VI-6
VII-2
A-l
B-l
Number
II-l
III-l
VI-1
VI-2
D-l
FIGURES OF REPORT 1
Nitrobenzene Manufacturing Locations
Process Flow Diagram for Manufacture of Nitrobenzene
Installed Capital Cost vs Plant Capacity for Emission
Control
Net Annual Cost or Savings vs Plant Capacity for
Emission Control
Precision of Capital Cost Estimate
Page
II-4
III-2
VI-7
VI-8
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)
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
IO'3
io"6
Example
1 Tg = 1 X IO12 grams
1 Gg = 1 X IO9 grams
1 Mg = 1 X IO6 grams
1 km = 1 X IO3 meters
1 mV = 1 X 10~3 volt
1 \ig = 1 X 10~6 gram
-------
II-l
II. INDUSTRY DESCRIPTION
A. REASON FOR SELECTION
Nitrobenzene was selected for consideration because preliminary estimates indi-
cated that its production caused relatively high emissions of volatile organic
compounds (VOC).1 The main constituent of these emissions is benzene, which was
included as a hazardous pollutant by the EPA in the Federal Register on June 8,
1977. Also, the growth rate of nitrobenzene production is expected to be higher
than the average growth rate for the industry.
Nitrobenzene is a relatively nonvolatile liquid under ambient conditions (see
Appendix A for pertinent physical properties). Most emissions from its produc-
tion are due to the volatility of benzene, the primary feed material.
B. USAGE AND GROWTH
Approximately 97% of all nitrobenzene produced is consumed in the manufacture of
aniline. Therefore the consumption pattern for aniline is the dominant factor
in the usage of nitrobenzene and its production growth. Table II-l lists the
end uses of aniline, with the percentage of production used for each end use,
and the expected growth rates. The use of nitrobenzene as a solvent accounts
for most of the remaining consumption.
O
Nitrobenzene production in 1978 was reported to have been 261,000 Mg, which is
51% of the capacity on-line at that time.3 Nitrobenzene production would uti-
lize 60% of the estimated 1982 capacity,3'4 with an average annual growth of 7%
assumed.
Five producers were operating seven nitrobenzene plants at the first of 1979.
Table II-2 lists the producers and their capacities, and Fig. II-l shows their
locations. All these plants produce nitrobenzene by nitrating benzene with
nitric acid mixed with sulfuric acid.3 Several recent developments have
affected the status of nitrobenzene capacity: Cyanamid reactivated its Bound
Brook, NJ, plant in 1978 and announced plans to bring a new nitrobenzene
facility of unspecified capacity on-stream in 1979; Dupont expanded the capaci-
ties at their Beaumont, TX, and Gibbstown, NJ, facilities by a total of about
-------
II-2
Table II-l. Aniline Usage and Growth'
End Use
Percentage of
Production (1978)
Average Rate
Growth (%/yr)
Polymeric isocyanates
Rubber chemicals
Dyes and intermediates
Hydroquinone
Drugs, pesticides, and
52
29
4
3
12
8
2 — 3
3
4.5
6
miscellaneous
See ref 3.
-------
II-3
Table II-2. Nitrobenzene Capacity
Plant
American Cyanamid
Du Pont
First Chemical
Mobay
Total
Location
Bound Brook, NJ
Willow Island, WV
Beaumont, TX
Gibbstown , NJ
Pascagoula , MS
New Martinsville , WV
Geismar, LA
Capacity (Mg/yr)
As of 1977
48,000
34,000
159,000
110,000
152,000
85,000
170,000
758,000
See refs 3 and 4.
bCyanamid's Bound Brook plant was on standby in 1977 but was
reactivated in 1978; this amount is included in the total.
clncludes 61,200-Mg/yr capacity brought on-stream in 1977.
-------
II-4
1. American Cyanamid, Bound Brook, NJ
2. American Cyanamid, Willow Island, WV
3. Du Pont, Beaumont, TX
4. Du Pont, Gibbstown, NJ
5. First Chemical, Pascagoula, MS
6. Mobay, New Martinsvilie, WV
7. Rubicon, Geismar, LA
Fig. II-l. Nitrobenzene Manufacturing Locations
-------
II-5
40,000 mg/yr (about 20,000 mg/yr at each plant) in 1978; First Chemical expanded
capacity by about 92,000 Mg in 1977; Mobay is to increase capacity by 25,000 Mg
by 1980; and Rubicon increased capacity by about 136,000 Mg during 1978. Allied
at Moundsville, WV, and Monsanto at Sauget, IL, discontinued nitrobenzene produc-
tion during the mid-1970s.
-------
II-6
C. REFERENCES*
1. T. C. Gunn and K. L. Ring, "Benzene," p. 618.5023V in Chemical Economics Handbook,
Stanford Research Institute, Menlo Park, CA (May 1977).
2. "Manual of Current Indicators -- Supplemental Data," p. 241 in Chemical Economics
Handbook, Stanford Research Institute, Menlo Park, CA (October 1978).
3. E. M. Klapprath, "Aniline and Nitrobenzene," pp. 614.5030A—J in Chemical Economics
Handbook, Stanford Research Institute, Menlo Park, CA (January, 1979).
4. "Chemical Research Services", p. 745 in, 1980 Directory of Chemical Producers,
United States of America, Stanford Research Institute, Menlo Park, CA.
*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 head-
ing.
-------
III-l
III. PROCESS DESCRIPTIONS
A. INTRODUCTION
Nitrobenzene is produced commercially by the direct nitration of benzene with a
mixture of nitric acid, sulfuric acid, and water. ' About 97% of the nitro-
benzene is used captively to produce aniline. There are no known foreign proc-
esses significantly different from the one used in the United States.
B. NITRATION OF BENZENE
Nitrobenzene is produced by the highly exothermic reaction
C..H.. + HNO. 2 4^ CCHCN00 + H00
b fa o > b b Z £
(benzene) (nitric (nitrobenzene) (water)
acid)
4
The heat released from this reaction is about 1.8 MJ/kg. The quantity of organic
by-products formed, primarily nitrated phenols, is only about 0.02 wt % of the
nitrobenzene produced. Typically, these phenolic materials are discharged with
the wastewater effluent.
A typical continuous-process flow diagram for the basic process is shown in Fig. III-l.
Benzene is nitrated at 55°C under atmospheric pressure by a mixture of concen-
trated nitric (Stream 1) and sulfuric (Stream 2) acids in a series of continuous
stirred-tank reactors. The exothermic heats of nitration and dilution are removed
2
by cooling coils. Yields of 96 to 98% of theory are reported.
The crude reaction mixture (Stream 3) flows to the separator, where the organic
phase is decanted from the aqueous waste acid.
The acid phase (Stream 4) is contacted in the extractor with fresh benzene from
r _ _ Q
storage (Stream 5) to extract most of the dissolved nitrobenzene and nitric
acid before the stream is stored in the waste-acid tank.
-------
REACTORS
STCRAG5.
RECYCLE. BEKJZ1EWE.
RECYCLE BEKJZEKJE.
CRUDE MlTRO&E.KlZ.EVje.
i EXTRACTOR
A©
WA'bTE
ACID
WATER
DILUTE
MoiOH
ACID
CCJ*
r
Ci) v
WASTE. -
WATER
TREAT MEWT
WARMER
VATER
TREA-TMEKIT
i©
STRIP PtP.
H
H
H
TO WZSO4 COWCJEKlTRATlOM
;~ TTT-I
-^
viow niaaram for Manufacture of Nitrobenzene
— - -^
-------
III-3
Benzene extract (Stream 6), two recovered and recycled benzene streams (7 and
8), and as much additional benzene (Stream 9) as is required make up the benzene
charge to the reaction step.
It is common practice to recover the benzene from the waste acid by distillation
in the acid stripper for recycle (Stream 8) to the reactor. The stripped acid
g
(Stream 10) is usually reconcentrated on-site but may be sold. Water carried
overhead with the benzene is forwarded (Stream 11) to the washer.
Crude nitrobenzene from the separator (Stream 12) is washed first with water and
then dilute caustic soda to remove the mineral acids and organic acids, such as
the nitrophenols. The washer and neutralizer effluents are discharged to waste-
water treatment. ' Following neutralization, the organic layer (Stream 13)
is fed to the nitrobenzene stripper, where water and most of the benzene and
c q
other low boilers are carried overhead. The organic phase, primarily ben-
zene, is decanted and recycled (Stream 7) to the reactor, and the aqueous phase
£• rj
is sent to the washer. ' Stripped nitrobenzene (Stream 14) is cooled and then
transferred to nitrobenzene storage for subsequent use as feed to an on-site
aniline process.
Typically, many of the process steps are padded with nitrogen gas to reduce the
chances of fire or explosion. ~~ ' This nitrogen padding gas and other inert
gases are purged from vents associated with the reaction and separator (Vent A),
the condenser on the acid stripper (Vent B), the washer and neutralizer
(Vent C), and the condenser on the nitrobenzene stripper (Vent D).
Fugitive emissions of benzene and nitrobenzene can occur when leaks develop in
valves, pump seals, and other equipment. Leaks can also occur from corrosion by
the sulfuric and nitric acids and hinder control of fugitive emissions.
All transfers of the product are by pipeline and there are no handling emis-
sions.
Storage emissions (G on Fig.III-1) occur from tanks storing benzene, waste acid,
and nitrobenzene.
-------
III-4
Three potential sources of secondary emissions (J on Fig.III-1) are the aqueous
waste from the washer, the caustic effluent from the neutralizer, and the waste
acid from the acid stripper.
C. PROCESS VARIATIONS
Another practiced process variation is to not strip residual benzene out of the
waste acid before sale or reconcentration of this acid. This can significantly
affect emissions unless the acid reconcentration process is adequately con-
trolled.
-------
III-5
D. REFERENCES*
1 D F Schiefferle, C. Hanson, and L. F. Albright, "Heterogeneous Nitration of
Benzene," p. 176 in Industrial and Laboratory Nitrations, edited by L. F.
Albright and C. Hanson, American Chemical Society Symposium Series 22,
Washington, 1976.
2 H J Matsuguma, "Nitrobenzene and Nitrotoluene," pp. 834 and 837 in Kirk-
Othmer Encyclopedia of Chemical Technology, 2d ed., vol 13, edited by Anthony
Standen et al., Wiley-Interscience, New York, 1967.
3 S Cooke, "Aniline and Nitrobenzene Salient Statistics," p. 614.5030C in
Chemical Economics Handbook, Stanford Research Institute, Menlo Park, CA
(January 1975).
4 H P L Kuhn W. J. Taylor, Jr., and P. H. Groggins, "Nitration," Chap. 4,
p. 85 in Unit Processes in Organic Syntheses, 5th ed., edited by P. H. Groggins,
McGraw-Hill, New York, 1958.
5 C Hanson, T. Kaghazchi, and M. W. T. Pratt, "Side Reactions During Aromatic
Nitration," p. 147 in Industrial and Laboratory Nitrations, edited by L. F.
Albright and C. Hanson, American Chemical Society Symposium Series 22,
Washington, 1976.
6 C W Stuewe, IT Enviroscience, Trip Report on Visit to E. I. du Pont de
Nemours & Co., Beaumont, TX. Sept. 7,8, 1977 (data on file at EPA, ESED,
Research Triangle Park, NC).
7 C W Stuewe IT Enviroscience, Trip Report on Visit to Rubicon Chemicals,
Geismar, LA, July 19,20, 1977 (data on file at EPA, ESED, Research Triangle
Park, NC).
8. D. W. Smith, E. I. du Pont de Nemours & Co., letter to D. R. Goodwin, EPA,
Feb. 3, 1978.
9. R. Barker, First Chemical Corporation, letter to D. R. Goodwin, EPA,
Jan. 20, 1978.
10. L. P. Hughes, Mobay Chemical Corp., letter to D. R. Goodwin, EPA, Jan. 31, 1978.
^Usually, when a reference is located at th- 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, par-
ticipate 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. NITROBENZENE MODEL PLANTS
Three model plant capacities -- 30,000, 90,000, and 150,000 Mg/yr -- were
selected to represent current domestic nitrobenzene manufacturing facilities.
The model process* (Fig. III-l) best represents today's nitrobenzene manufac-
turing and engineering technology.
Typical raw material, waste acid, and product storage capacities were selected
for the three model-plant capacities. The number of valves and pumps selected
was based on data from an existing facility. Characteristics of the model
plants important to air dispersion are given in Appendix B.
B. SOURCES AND EMISSIONS
]. General
Sources and emission rates for the model plants are summarized in Table IV-1.
Process and secondary emissions are based on data obtained from plant-site
visits and information submitted to the EPA. " Storage emissions were calcu-
lated with the equations in AP-42. However, breathing losses were divided by 4
to account for recent evidence indicating that the AP-42 breathing loss equation
overestimates emissions. Fugitive emissions were determined by
*See p. 1-2 for a discussion of model plants.
-------
Table IV-1. Uncontrolled Benzene and Total VOC from Nitrobenzene Model Plants
Emission Rates
For 30,000-Mg/yr Model plant
Source
Reactor and separator
Waste-acid stripper
Wash and neutralization
Nitrobenzene stripper
Small benzene storage
Waste -acid storage
Benzene storage
Nitrobenzene storage
Fugitive
Secondary
Total
Uncontrolled emissions
kg of benzene or total
°The small storage tank
Stream
Designatior
Ratiob (kg/Mg)
i —
(Fig. III-l) Benzene Total VOC
A
B
C
D
G
G
G
G
H
J
are emissions
VOC per Mg of
0.960 0.965
0.170 0.170
0.0081 0.0107
0.170 0.171
0.076 0.076
0.052 0.052
0.294 0.294
0.0024
1.9 2.98
0.10 0.33
3.73 5.05
from the process employing
nitrobenzene produced.
contains approximately one day's supply
Rate (kg/hr)
Benzene
3.29
0.582
0.0277
0.582
0.262
0.177
1.01
6.5
0.342
12.8
no additional
For 90,000-Mg/yr
Ratiob (kg/Mg)
Total VOC Benzene
3.30
0.582
0.0366
0.586
0.262
0.177
1.01
0.0083
10.2
1.10
17.3
control
of benzene; the large
0.960
0.170
0.0081
0.170
0.073
0.051
0.283
0.63
- 0.10
2.45
devices other
Total VOC
0.965
0.170
0.0107
0.171
0.078
0.051
0.283
0.0019
0.99
0.33
3.05
Model Plant
For 150,000-Mg/yr Model Plant
Rate (kg/hr)
Benzene
9.86
1.75
0.0832
1.75
0.797
0.526
2.91
6.5
1.03
25.2
Total VOC
9.91
1.75
0.110
1.76
0.797
0.526
2.91
0.0197
10.2
3.39
31.4
Ratiob
Benzene
0.960
0.170
0.0081
0.170
0.077
0.048
0.281
0.38
0.10
2.19
(kg/Mg)
Total VOC
0.965
0.170
0.0107
0.171
0.077
0.048
0.281
0.0018
0.596
0.33
2.65
Rate
Benzene
16.4
2.91
0.139
2.91
1.31
0.830
4.81
6.5
1.71
37.5
(kg/hr)
Total VOC
16.5
2.91
0.183
2.93
1.31
0.830
4.81
0.031
10.2
5.65
45.4
than that necessary for economical operation.
tank is referred to as the main
storage tank
•
H
1
-------
IV-3
estimating the number of valves and pumps for the model plants based on informa-
tion from an existing facility and applying the factors listed in Appendix C.
Handling losses are not considered, since it is assumed that the nitrobenzene
will be used on-site for production of aniline.
2. Process Emissions
There are four vents for process emissions from the model plants, two of which
are combined vents from associated equipment. All these vents are necessary for
removal of inert gases from the process. Nitrogen padding of benzene is used
for safety purposes and contributes to inert gases in the process and resultant
emissions. Benzene constitutes the bulk of emissions from the process, as shown
in Table IV-1, with less nitrobenzene being emitted because of its low volatil-
ity.
a. Reactor and Separator Vent -- This vent (Vent A, Fig. III-l) combines emissions
from the reactors and from the separator. Oxides of nitrogen are generated by
side reactions involving nitric acid and must be purged from the process, along
with nitrogen padding gas.
b. Acid Stripper Vent -- Organics are stripped from the waste acid for recycle to
the process, and noncondensables are vented (Vent B, Fig. III-l) from the asso-
ciated condenser.
c. Washer and Neutralizer Vent -- The washer removes mineral acids from the nitro-
benzene, and the neutralizer removes the remaining acids, primarily organic
acids. The combined vent (Vent C, Fig. III-l) for these two operations removes
nitrogen padding gas and some water vapor from the process.
d. Nitrobenzene Stripper Vent -- Benzene is stripped from the nitrobenzene, and
noncondensables, primarily nitrogen padding gas, are vented (Vent D, Fig. III-l)
from the associated condenser.
3. Storage Emissions
Emissions result from the storage of benzene, waste acid (which contains ben-
zene), and nitrobenzene. The sources of storage emissions for the model plants
are shown on the flow diagram, Fig. III-l (Source G). Storage tank conditions
-------
IV-4
for the model plants are given in Table IV-2. The uncontrolled storage emis-
sions in Table IV-1 were calculated with the equations from AP-42 with the
breathing loss adjustment6 as mentioned above and the assumption that fixed-roof
tanks are used; on the average these tanks are half full and have a 12°C diurnal
temperature variation. It was also assumed that the waste-acid and nitrobenzene
storage tanks are operated at nearly constant levels, with only six turnovers
per year, and that waste-acid stripping does not remove all the benzene from
that material before storage.
4. Fugitive Emissions
Process pumps and valves are potential sources of fugitive emissions. Each
model plant is estimated to have 42 pumps (including 17 spares), 500 process
valves, and 20 pressure-relief valves based on data from an existing facility.
All pumps have mechanical seals. Twenty-five percent of these pumps and valves
are being used in benzene service. The fugitive emissions included in
Table IV-1 are based on the factors given in Appendix C.
5. Secondary Emissions
Secondary VOC emissions can result from the handling and disposal of process
waste liquid. For the model plants three potential sources of secondary emis-
sions from waste liquids are indicated on the flow diagram, Fig. III-l
(Source J). These sources are the sulfuric acid from the acid stripper, waste-
water from the nitrobenzene washer, and waste caustic from the nitrobenzene
neutralizer. Because of its low volatility most of the nitrobenzene in the
waste acid will make no contribution to secondary emissions except when the acid
is being concentrated for reuse. Any benzene remaining after the acid is
stripped would create a potential for secondary emissions. Emissions from this
source will be discussed more fully in a future EPA report on concentration of
sulfuric acid used in organic chemical processing. The combined wastewater from
the wash and neutralization steps contains benzene, nitrobenzene, and neutral-
ized organic acid by-products (primarily nitrophenates). The latter are non-
volatile and will not contribute to the VOC emission rate. Secondary emissions
of nitrobenzene from the wastewater directed to a clarifier and conventional
air-activated sludge treatment system will be low due to the low vapor pressure
at ambient temperatures and the biodegradability of the nitrobenzene. The loss,,
estimated by methods to be described in a future EPA report on secondary emis-
-------
IV-5
Table IV-2. Storage Parameters for
Determining Model-Plant Emissions
Content
Benzene
Benzene
a
Waste acid
a
Nitrobenzene
Benzene
Benzene
a
Waste acid
a
Nitrobenzene
Benzene
Benzene
. a
Waste acid
Nitrobenzene3
Tank Size Turnovers Bulk Liquid
(m3) per Year Temperature (°C)
For 30,000-Mg/yr Model Plant
946
95
151
473
For 90,000-Mg/yr Model Plant
2840
284
454
1420
For 150,000-Mg/yr Model Plant
4730
473
757
2360
24
236
6
6
24
236
6
6
24
236
6
6
20
20
45
40
20
20
45
40
20
20
45
40
Surge tanks normally operated at constant level.
-------
IV-6
sions, is 1.1% of the nitrobenzene in the untreated water. This is equivalent
to an emission rate of 5 X 10 kg of VOC per Mg of nitrobenzene produced. The
benzene and total VOC secondary emissions listed in Table IV-1 were calculated
on the assumption that the benzene and 1.1% of the nitrobenzene in the waste-
water effluent will become secondary emissions.
C. EFFECTS OF PROCESS VARIATIONS ON EMISSIONS
Waste acid, which is not stripped of residual benzene before being sold or re-
concentrated, can significantly affect secondary emissions. Based on solubility
data the potential emissions from this source could be as much as 1 kg of ben-
zene per Mg of nitrobenzene produced.
Most plants use nitrogen blanketing on many of the process steps. The effects
on emissions from not using nitrogen blanketing have not been defined.
-------
IV-7
D. REFERENCES*
1 C W Stuewe, IT Enviroscience, Trip Report for Visit to E. I. du Pont de
Nemours & Co., Beaumont. TX, Sept. 7,8, 1977 (data on file at EPA, ESED,
Research Triangle Park, NC).
2. R. Barker, First Chemical Corporation, letter to D. R. Goodwin, EPA,
Jan. 20, 1978.
3. D. W. Smith, E. I. du Pont de Nemours & Co., letter to D. R. Goodwin, EPA,
Feb. 3, 1978.
4. L. P. Hughes, Mobay Chemical Corporation, letter to D. R. Goodwin, EPA,
Jan. 31, 1978.
5. C. C. Masser, "Storage of Petroleum Liquids," pp. 4.3-1 to 4.3-11 in Supplement
No. 7 for Compilation of Air Pollutant Emission Factors, AP-42, 2d ed., EPA,
Research Triangle Park, NC (April 1977).
6. E.G. Pulaski, TRW, 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. PROCESS SOURCES
A number of control systems are feasible and were considered for control of the
combined process emissions. In-process storage emissions can be readily con-
trolled in conjunction with the process emissions and were so treated.
1. Vent Absorber
An absorber using nitrobenzene as the scrubbing solvent has been selected for
detailed study. Absorption of a volatile hydrocarbon in a less volatile hydro-
carbon is a common method for recovery of light hydrocarbons and can be used for
absorption of benzene in nitrobenzene. The use or intended use of this type of
1 2
control device has been reported by two producers.
The absorber system described on page D-5 in Appendix D is a preliminary design
for cost estimating purposes per the standard design methods described by
Treybal.3 The design has not been optimized. The absorbent and absorbed mate-
rials are used or produced in the process and therefore very little additional
processing equipment is required for recovery of most of the emitted VOC. As
designed the system utilizes the existing process capability for separation of
benzene and nitrobenzene by recycling the liquid bottoms stream from the ab-
sorber to the nitrobenzene stripper. It is assumed that the existing nitroben-
zene stripper capacity is sufficient to handle this additional load. Estimated
capital equipment costs would be increased if additional stripping capacity is
required. Nitrobenzene absorbent is drawn from storage and chilled to 15°C
before it enters the absorbing column. Exhaust gases from the nitrobenzene
scrubbing section pass through additional scrubbing sections, where they are
4
washed with water and dilute caustic solution to remove oxides of nitrogen.
The vent absorption system will reduce benzene and total VOC emissions by about
95% at a pressure of 1 X 10 Pa.
Controlled emissions, based on this control device, are given in Table V-l for
the 30,000-, 90,000-, and 150,000-Mg/yr model plants.
-------
V-2
Table V-l. Controlled Benzene and Total VOC Emissions for
Nitrobenzene Model Plants
Emission Data
Stream Control Emission a
Designation Device or Reduction - - .Ratio
(kg/Mg)
Source (Fiq. m-i) Technique (%1 B^n^»n» T~tFl1 ^
Reactor and separator A "
Waste-acid stripper B
Wash and neutralization c
Nitrobenzene stripper D
Small benzene storage G
Waste-acid storage G ,
Benzene storage G
Nitrobenzene storage G
Fugitive H
For 30,000-Mg/yr Model Plant
Vent absorber 94.6 0.0775
i Thermal oxidizer 99.0 0.0144
Floating roof 88 0.0441
None
Detect
rect
plus
and cor- 67.7 0.50
leaks
mech-
0.0780
0.0144
0.0441
0.0024
1.08
Rate (kg/hr)
0.237 0.267
0.0440 0.0494
0.151 0,151
O.,0083
1.70 3,7
anical seals
Secondary j
Total with vent absorber
Total with thermal oxidizer
None
0.10
0.72
0.66
0.33
1.53
1.47
0.342 1..10
2.43 5.23
2.24 5,01
For 90,000-Mg/yr Model Plant
Reactor and separator A >
Waste -acid stripper B
Wash and neutralization c
Nitrobenzene stripper D
Small benzene storage G
Waste-acid storage G
Benzene storage G
Nitrobenzene storage G
Fugitive H
|
Vent absorber 94.6 0.0776
^ Thermal oxidizer 99.0 0.0144
Floating roof 85 0.0425
None
Detect
rect
plus
and cor- 67.7 0.165
leaks
mech-
0.0781
0.0145
0.0425
0.0019
0.36
0.797 0.802
0.148 0.149
0.437 0.437
0.0197
1.7 3.7
•anical seals
Secondary j
Total with vent absorber
Total with thermal oxidizer
None
0.10
0.39
0.22
0.33
0.81
0.75
1.03 3.39
3.96 8.32
3.32 7.70
For 150,000-Mg/yr Model Plant
Reactor and separator A ^
Waste-acid stripper B
Wash and neutralization c
Nitrobenzene stripper D [
Small benzene storage G 1
Waste-acid storage G J
Benzene storage G
Nitrobenzene storage G
Fugitive H
Vent absorber 94.6 0.0774
Therma]
oxidizer 99.0 0.0143
Floating roof 85 0.0421
None
Detect
rec^
plus
and cor- 67.7 0.099
leaks
mech-
0.0779
0.0144
0.0421
0.0018
0.216
1.32 1.33
0.245 0.247
0.721 0.721
0.031
1.70 3.7
anical seals
Secondary j
Total with vent absorber
Total with thermal oxidizer
None
0.10
0.32
0.26
0.33
0.67
0.60
1.71 5.65
5.45 11.43
4.38 10.35
kg of benzene or total VOC per Mg of nitrobenzene produced.
-------
V-3
2. Thermal Oxidizer
Efficient control of benzene and total VOC is technically feasible with the use
of thermal oxidation. It is estimated that, with effective design, the removal
efficiency for VOC can be greater than 99%.
The details of the system necessary for cost estimation for the 90,000-Mg/yr
model plant are described in Appendix D. Two combustion chambers are included
to reduce NO emissions by reducing the NO to N . Heat recovery on such a
X ££ ^
small unit is not economical and was not included.
Controlled emissions, based on this control device, are given in Table V-l for
the 30,000-, 90,000-, and 150,000-Mg/yr model plants.
With adequate design consideration, efficient VOC removal can be accomplished by
thermal oxidation of the vent stream in an existing boiler, in a process equip-
ment heater, or in a liquid thermal oxidizer. Technical feasibility and eco-
nomics for such an approach would be highly dependent on the specifics of each
situation.
3. Chemical Absorber.
A system that consists of an absorption column that removes benzene by nitration
in a circulating mixture of nitric and sulfuric acids has been reported in use
with a design efficiency of greater than 99.9% for benzene removal. Subse-
quently it was reported that operating difficulties had been experienced with
the column and that it has been converted to a scrubber using nitrobenzene. A
chemical (nitration) absorber system similar to that reported is described on
page D-19 in Appendix D. The reaction products and remaining acids are returned
to the primary nitration step in the process. Exhaust gases pass into a
scrubber, where they are washed with water and dilute caustic solution to remove
acids and oxides of nitrogen.4 Conceptually, an absorbing reactor, for this
application should be technically feasible vith relatively attractive economics,-
however, the technical practicality has not been proved by actual operation.
B. FUGITIVE SOURCES
Control for fugitive sources will be discussed in a future document covering
fugitive emissions from the synthetic organic chemicals manufacturing industry
-------
V-4
(SOCMI). The controlled fugitive emissions given in Table V-l were calculated
with the factors listed in Appendix C. These factors are based on the assump-
tion that any major leaks will be detected and repaired.
C. STORAGE SOURCES
Storage guidelines for SOCMI are given in a separate EPA document. Emissions
from the benzene daily-storage tank* and waste-acid storage tank are controlled
in conjunction with the process emissions that are controlled by the absorbing
reactor. The main benzene feed storage emissions are controlled by using float-
ing-roof tanks.** Storage emissions were calculated by asuming that a contact-
type internal floating roof with secondary seals will reduce fixed-roof-tank
emissions by 85%. Emissions from storage of nitrobenzene remain uncontrolled.
D. SECONDARY SOURCES
Potential secondary emissions originate with the waste acid, the wastewater from
the nitrobenzene washer, and the waste caustic from the nitrobenzene neutral-
izer. Benzene discharged with the wastewater effluent will create a secondary
emission because of its relatively high volatility. Because of its low volatil-
ity most of the nitrobenzene in the wastewater effluent will make no contribu-
tion to secondary emissions. The total estimated potential secondary emissions
from the model plants are listed in Table V-l. Secondary emissions are uncon-
7
trolled. A separate EPA report discusses emissions from secondary sources.
E. CONTROL DEVICES USED BY INDUSTRY
Control devices used by industry are covered in Appendix E.
*Small storage tank contains approximately one day's supply of benzene; the
larger tank is the main benzene storage tark.
**Consist of internal floating covers or covered floating roofs as defined in API
25-19, 2nd ed., 1976 (fixed-roof tanks with internal floating device to reduce
vapor loss).
-------
V-5
F. REFERENCES*
1. R. Barker, First Chemical Corp., letter to D. R. Goodwin, EPA, Jan. 20, 1978.
2. W. C. Anthon, Rubicon Chemicals, letter to David A. Beck, EPA, Apr. 14, 1978.
3. R. E. Treybal, Mass-Transfer Operations, Chaps. 6 and 8, McGraw-Hill, New York,
1955.
4. E. F. Spencer, Jr., "Pollution Control in the Chemical Industry," Chap 14,
p. 14-6 in Industrial Pollution Control Handbook edited by H. F. Lund,
McGraw-Hill, New York, 1971.
5. D. G. Erikson, IT Enviroscience, Storage and Handling (September 1980)
(EPA/ESED report, Research Triangle Park, ND).
6. William T. Moody, TRW, letter dated Aug. 15, 1979, to D. Beck, EPA.
7. J. J. Cudahy and R. L. Standifer, IT Enviroscience, 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. ENVIRONMENTAL AND ENERGY IMPACTS
Table VI-1 shows the effect on the environment of reducing benzene and total VOC
emissions by application of the described control systems to the model plants.
Individual effects are discussed below.
1. Control of Process Emissions and Emissions from In-Process Storage of Benzene
and Waste Acid
Process emissions and emissions from in-process storage of benzene and waste
acid can be controlled by using either a vent absorber or a thermal oxidizer.
a. Vent Absorber -- A vent absorber using nitrobenzene as the absorbent can be
installed for control of process emissions and emissions from in-process storage
of benzene and waste acid. This vent absorber reduces benzene and total VOC by
40.8 and 41.0 Mg/yr for the 30,000-Mg/yr model plant, 122.4 and 123.1 Mg/yr for
the 90,000-Mg/yr model plant, and 203.0 and 204.4 Mg/yr for the 150,000-Mg/yr
model plant. The electrical energy required for operation of the vent absorber
is small (less than 400 MJ/Mg of VOC recovered for the 90,000-Mg/yr model
plant).
b. Thermal Oxidizer -- As an alternative device, a thermal oxidizer can be
installed for control of process emissions and emissions from in-process storage
of benzene and waste acid. This thermal oxidizer reduces benzene and total VOC
by 42.7 and 42.9 Mg/yr for the 30,000-Mg/yr model plant, 128.1 and 128.8 Mg/yr
for the 90,000-Mg/yr model plant, and 212.5 and 213.9 Mg/yr for the 150,000-Mg/
yr model plant. The electrical energy required for operation of the thermal
oxidizer is small (less than 100 MJ per Mg of VOC reduced).
2. Benzene Storage
Retrofitting existing fixed-roof tanks with floating roofs or installing new
floating-roof tanks for control of emissions from the main benzene storage tanks
reduces benzene emissions by 11.4, 30.8, and 50.4 Mg/yr for the 30,000-,
90,000-, and 150,000-Mg/yr model plants, respectively. The use of floating-roof
storage tanks for emissions control does not consume energy and has no adverse
environmental or energy impact.
-------
Table VI-1. Environmental Impact of Controlled Model Plants
Stream Control Device
Designation or
Source (Fig. III-l) Technique
b .. ^
Reactor and separator A^
Waste-acid stripper B
Wash and neutralization C
Nitrobenzene stripper D
Small benzene storage G
b _
Waste-acid storage G ^
Benzene storage G
Nitrobenzene storage G
Fugitive H
Secondary J
Total with vent absorber
Total with thermal oxidii2r
Vent absorber
' Thermal oxidizer
Internal floating roof
None
Detect and correct
minor leaks plus
mechanical seals
None
Emission Reduction (Mg/yr)
30,000-Mg/yr Model Plant 90,000-Mg/yr Model Plant 150,000-Mg/yr Model Plant
Benzene Total VOC Benzene Total VOC Benzene Total VOC
40.8 (95%)° 41.0 (95%) 122.4 (95%) 123.1 (95%) 203.0 (95%) 204.4 (95%)
42.7 (99%) 43.9 (99%) 128.1 (99%) 128.8 (99%) 212.5 (99%) 213.9 (99%)
11.4 (85%) 11.4 (85%) 30.8 (85%) 30.8 (85%) 50.4 (85%) 50.4 (85%)
42.0 (13.8%) 56.9 (613.7%) 42.0 (73.8%) 56.9 (63.7%) 42.0 (73.8%) 56.9 (63.7%)
94.2 109.3 195.2 210.8 295.4 311.7
96.1 112.2 200.9 216.5 304.9 321.2
aAnnual reduction is based on 8760 hr of operation.
Combined for control.
cFigures in parentheses are the percent reduction of benzene and total VOC emissions.
-------
VI-3
3. Fugitive Emissions
Control of fugitive emissions is accomplished by detection and repair of major
leaks plus mechanical seals on pumps. This reduces benzene emissions by 42.0 Mg/
yr and total VOC emissions by 56.9 Mg/yr for each of the model plants. If each
of the seven domestic production facilities operating in 1979 had an average
number of pumps and valves equivalent to those in the model plants, the control
of fugitive emissions for the industry would reduce the total industry benzene
emissions by 294 Mg/yr and the total VOC emissions by 398 Mg/yr.
B. CONTROL COST IMPACT
This section presents estimated costs and cost-effectiveness data for control of
VOC emissions resulting from the production of nitrobenzene. Details of the
model plants are given in Sect. Ill, emission sources and emissions are dis-
cussed in Sect. IV, and cost estimate calculations are given in Appendix D.
Capital cost estimates represent the total investment required for purchase and
installation of all new equipment for a complete emission control system, per-
forming as defined for a typical location. These estimates do not include the
cost resulting from production lost during installation of control systems or
the costs for research and development.
The bases for annual cost estimates for the control alternatives include utili-
ties, operating labor, maintenance supplies and labor, recovery credits, capital
charges, and miscellaneous recurring costs such as taxes, insurance, and admin-
istrative overhead. The cost factors that were used are itemized in Table VI-2.
Emission recovery credits are based on the current equivalent raw material
market value of the material being recovered. Annual costs are for a 1-year
period beginning in December 1979.
1. Process Emissions
Process emissions, emissions from daily-use storage of benzene, and emissions
from waste acid storage are controlled by a vent absorber or a thermal oxidizer,
which are shown in Appendix D. The estimated capital cost of installing the
vent absorber is $41,500, $48,000, and $56,500 for the 30,000-, 90,000-, and
150,000-Mg/yr model plants, respectively. Utilities, related capital costs, and
recovery credits vary with the plant capacity, as shown in Table VI-3.
Installed capital and net annual cost variations with capacity are shown in
-------
VI-4
Table VI-2. Cost Factors Used in Computing Annual Costs
Item Factor
Electricity $0.00833/MJ ($0.03/kWh)
Operating time 8760 hr/yr
Operating labor $15/hr
Fixed costs
Maintenance labor plus materials, 6$
Capital recovery, 18% (10 yr life @ 12% int.)
29% installed capital
Taxes, insurance, administration charges, 5%
Liquid-waste disposal ' Minor,- not considered
Recovery credits
Benzene $220/Mg ($0.10/lb)
Nitrobenzene (raw material value) $220/Mg ($0.10/lb)
-------
VI-5
Figs. VI-1 and VI-2. The estimated capital cost of the installed thermal
oxidizer, $277,000, does not vary for the three model plants because the unit is
very small.
2. Storage
Model plant emissions from the small benzene storage tank and the waste-acid
storage tank are controlled in conjunction with process emissions by the chemi-
cal absorber. Benzene-feed storage emissions are controlled by the use of float-
ing-roof tanks. Another EPA report covers storage emissions and their appli-
cable controls for all the synthetic organic chemicals manufacturing industry.
3. Fugitive Sources
Controlled emission factors for fugitive sources are described in Appendix C. A
separate EPA document covers 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
plants. A separate EPA document discusses secondary sources and their control.
-------
Table VI-3. Emission Control Analyses for Nitrobenzene Model Plants
Annual Operating Costs (X 1000)
Item
Total Installed
Capital Cost
(X 1000)
Mid-1978
Capital
Related Recovery
Utilities Manpower Cost Credits
30,000-Mg/yr
c
Vent absorber
Thermal oxidizer
$ 41.5
277
$2.1 $12
$18.0 80
.0
.0
90,000-Mg/yr
Vent absorber
c
Thermal oxidizer
Vent absorber
Thermal oxidizer
$ 48
277
$ 56.5
277
$5.7 $13
$18.0 80
150,
$9.5 $16
$18.0 80
.9
.0
000-Mg/yr
.4
.0
(A)
Net
Annual
Cost
(B)
Emission Reduction
Benzene
(Mg/yr)
b
Total VOC
(Mg/yr)
Percent
(for both)
(Oa
Cost Effectiveness
for Total VOCT
(per Mg)
Model Plant
$ 9.5
$ 4.6
98.0
40
42
.8
.7
41.0
43.9
95
99
$ 112
2,232
Model Plant
$28.6
Model
$47.7
$(9.0)d
98.0
Plant
$(21.8)d
98.0
122
128
203
212
.4
.1
.0
.5
123.1
128.8
204.4
213.9
95
99
95
99
$ (73)
760
$ (107)
$ 458
d
d
(C)
(A) * (B).
bTotal VOC consists of more than 99% benzene for the vent absorbers and thermal oxidizers.
cControls process emissions and emissions from daily stored benzene and waste-acid storage.
Net annual savings.
-------
VI-7
4-1
in
O
u
•H
a
cO
u
•O
0)
cO
-p
in
c
en
r^
CTi
a)
o
a;
Q
(X $1000)
300
200
100
90
80
70
60
50
40
3C
(1)
I I I
20
30
40
50
60 70 80 90 100
200
Plant Capacity (Gg/yr)
(1) Thermal oxidizer for benzene, total VOC, and No
(2) Adsorption system for benzene and total VOC
Fig. VI-1. Installed Capital Cost vs Plant Capacity for
Emission Control
-------
VI-8
tn
CP
a
co
4-1
w
o
o
0)
N
•i-l
iH
nj
3
C
-P
0)
2
(X $1000)
45
20
30
40 50 60 70 80 90 100
Plant Capacity (Gg/yr)
200
(1) Thermal oxidizer for benzene and total VOC.
(2) Absorption system for benzene and total VOC.
Fig. VI-2. Net Annual Cost or Savings vs Plant Capacity for Emission Control
-------
VI-9
C. REFERENCE*
1. D. G. Erikson, IT Enviroscience, Storage and Handling (September 1980) (EPA/ESED
report, Research Triangle Park, NC)
2. D. G. Erikson, IT Enviroscience, Fugitive Emissions (September 1980) (EPA/ESED
report, Research Triangle Park, NC).
3. J. J. Cudahy and R. L. Standifer, IT Enviroscience, 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 relptes 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
All domestic nitrobenzene production is based on nitrating benzene with nitric
acid mixed with sulfuric acid. Approximately 97% of all nitrobenzene produced
is consumed in the manufacture of aniline.1 The two chemicals are expected to
grow at an average annual rate of about 7%.
Emission sources and control levels for the model plants are summarized in
Table VII-1.
Projected emissions for the domestic nitrobenzene industry in 1979 are based on
the following assumptions:
1. The 1978 production estimated in Sect. II increased by 7% during 1979 to
244,000 Mg.
2. The 90;000-Mg/yr model-plant emission rates, excluding fugitive emissions,
are typical for the composite industry.
3. For the purpose of projecting fugitive emissions, the average number of
pumps and valves for the seven domestic nitrobenzene manufacturing plants
is the same as that for the model plants.
A weighted average of the following individual emission control estimates for
process, in-process storage, raw material and product storage, secondary, and
fugitive emissions indicates that the domestic nitrobenzene industry is approxi-
mately 50% controlled:
Percent
Controlled
Process emissions 50
In-process storage emissions 38
Raw material and product 53
storage emissions
Secondary emissions °
Fugitive emissions 80
T. C. Gunn and K. L. King, "Benzene," p. 618.5023V in Chemical Economics Handbook,
Stanford Research Institute, Menlo Park, CA (May 1977).
-------
Table VII-1. Model Plant Emission Summary
Emission Rate (kq/hr)
30,000-Mg/yr Model
Plant
Uncontrolled Controlled
Benzene Total VOC Benzene
Reactor and separator
Waste-acid stripper
Wash and neutralization
Nitrobenzene stripper
Small benzeno storage
Waste-acid storage
Benzene storage
Nitrobenzene storage
Fugitive
Secondary
Total with vent
absorber
Total with thermal
oxidizer
3.29 3.30 ^
0.582 0.582
0.0277 0.0366
0.582 0.586
0.262 0.262
0.177 0.177^
1.01 1.01
0.0083
6.5 10.2
0.342 1.10
12.8 17.3
12.8 17.3
0.2373
0.044b
0.151
1.70
0.342
2.43
2.24
Total VOC
0.267d
0.044b
0.151
0.0083
3.70
1.10
5.23
5.01
90 ,000-Mg/yr Model Plant
Uncontrolled
Controlled
Benzene Total VOC Benzene Total VOC
9.86 9.91 v
1.75 1.75
0.0832 0.110
1.75 1.76
0.797 0.797
0.526 0.526J
2.91 2.91
0.0197
6.5 10.2
1.03 3.39
25.2 31.4
25.2 31.4
0.7973 0.0802a
> h b
0.148 0.149
0.437 0.437
0.0197
1.70 3.70
1.03 3.39
3.96 8.32
3.32 7.70
150,000-Mg/yr Model Plant
Uncontrolled
Benzene Total VOC
16.4 16.5 ^
2.91 2.91
0.139 0.183
2.91 2.93
1.31 1.31
0.830 0.830J
4.81 4.81
0.031
6.5 10.2
1.71 5.65
37.5 45.4
37.5 45.4
Controlled
Benzene Total VOC
1.323 1.333
/ h h
0.245 0.274
0.721 0.721
0.031
1.70 3.70
<-<
1.71 5.65 |5
H
5.45 11.43 |
NJ
4.38 10.35
Controlled by vant absorber.
Controlled by thermal oxidizer.
-------
VII-3
For the process, storage, and secondary emissions the projections are based on
data reported from producers representing 83% of domestic capacity. The fugi-
tive-emission projection is based on the estimate that all equipment handling
nitrobenzene would be controlled because of the extreme toxicity of that mate-
rial and the necessity for worker protection and that all equipment not handling
nitrobenzene is uncontrolled in respect to the fugitive-emission calculations.
From these data the emission projections for the domestic nitrobenzene industry
in 1979 were 434 Mg of benzene and 619 Mg of total VOC.
The predominant emission points are the reactor and separator vent and the
storage tanks. The emissions from the reactor and separator vent and other
process emissions can be controlled in conjunction with emissions from the ben-
zene daily-storage tank and from the waste-acid storage tank by a vent absorber
using nitrobenzene as the absorbent or by a thermal oxidizer. These control
devices result in removal efficiencies of 95% and 99% respectively. The capital
cost of the vent absorber is $41,500, $48,000, and $56,500 for the 30,000-,
90,000-, and 150,000-Mg/yr model plants, respectively. Due to the small duty
requirements, the thermal oxidizer capital cost is constant at $277,000 for all
three model plant sizes. Benzene storage emissions from the main storage tanks
can be controlled by using covered floating-roof tanks in a new plant or by
retrofitting existing fixed-roof tanks with floating-roof tanks. The emission
reductions resulting from the use of floating roof is 85% of the fixed-roof-tank
emissions.
-------
A-l
APPENDIX A
Table A-l. Physical Properties of Nitrobenzene and Benzene
Nitrobenzene
Benzene
Synonyms
Molecular formula
Molecular weight
Physical state
Vapor pressure
Vapor density
Boiling point
Melting point
Density
Water solubility
Oil of mirbane, nitrobenzol,
mononitrobenzene, artificial
oil of bitter almonds, sol-
vent black 5, nigrosine
spirit soluble B
C6H5N°2
123.11
Solid or oily liquid
0.284 mm Hg at 25 °C
4.25
210.8°C at 760 mm Hg
5.7°C
1.2037 g/ml at 20°C/4°C
Slight (0.09 g/100 ml of
H O at 20°C)b
Benzol, phenylhydride,
coal naphtha
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 g/ml at 20°C/4°C
Slight (1.79 g/100 ml
of H20)
Except for the last item, the data in this table are from: J. Dorigan ert al.,
"Scoring of Organic Air Pollutants - Chemistry, Production, and Toxicity of
Selected Synthetic Organic Chemicals (Chemicals F-N)," MTR-7248, Rev. 1,
Appendix III, p. A-III-264, Mitre Corp., Metrek Division (September 1976).
bj. Dorigan et al., "Scoring of Organic Air Pollutants - Chemistry, Production,
and Toxicity~~of Selected Synthetic Organic Chemicals (Chemicals A-C),"
MTR-7248, Rev. 1, Appendix I, p. AI-102, Mitre Corp., Metrek Division
(September 1976).
CH.P.L. Kuhn, W. J. Taylor, Jr., and P. H. Groggins, "Nitration," Chap. 4,
p. 110, in Unit Processes in Organic Syntheses, edited by P. H. Groggins,
5th ed., McGraw-Hill, New York, 1958.
-------
APPENDIX B
AIR-DISPERSION PARAMETERS
Table B-l. Air-Dispersion Parameters for 90,000-Mg/yr Nitrobenzene Model Plant
Source
Reactors and separator
Waste-acir stripper
Wash and neutralization
Nitrobenzene stripper
Small benzene storage tank
Waste-acid storage
Benzene storage
Nitrobenzene storage
Fugitive
Secondary
Vent absorber
Thermal oxidizer
Benzene storage
Nitrobenzene storage
a
Fugitive
Secondary
Emission
Benzene
2.74
0.486
0.0231
0.486
0.221
0.146
0.808
1.81
0.286
0.221
0.0411
0.121
0.472
0.286
Rate (g/sec)
Total VOC
2.75
0.486
0.0306
0.489
0.221
0.146
0.808
0.0055
2.83
0.942
0.223
0.0411
0.121
0.0055
1.03
0.942
Height Diameter
(m) (m)
Uncontrolled
20 0.038
20 0.025
11 0.031
20 0-025
7.3 7.0
9.8 7.7
12.2 17.2
12.2 12.2
Controlled
20 0.076
20 0.305
12.2 17.2
12.2 12.2
Discharge Flow Discharge
Temperature Rate Velocity
(K) (m3/sec) (m/sec)
328 3.99 X 10~
-4
305 9.67 X 10
-3
318 2.83 X 10
-4
305 9.72 X 10
293
318
293
313
298 9-22 X 10"3
477 1.88 X 10~
293
313
293-328
_
3.5
2.0
3.8
2.0
2.0
2.6
to
I
Distributed
of 40 m by 80 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 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
Q.00'03
0.061
0.006
0.009
0.11
0.00026
0.019
3Based 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.
3Light liquid means any liquid more volatile than kerosene.
*Radian Corp., Emission Factors and Frequency of Leak Occurrence for Fittings
in Refinery Process Units, LPA 600/2-79-044 (February 1979).
-------
D-l
APPENDIX D
COST ESTIMATE DETAILS AND CALCULATIONS
A. GENERAL
This appendix contains the details of the estimated costs presented 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.
B. ABSORPTION OF PROCESS EMISSIONS
Capital and operating cost estimates for the model-plant vent absorption systems
described in Sect. V were determined as follows. The example given below is for
model-plant 2 (90,000 Mg/yr capacity).
Basis:
Plant, 90,000-Mg/yr
Vent composition and rate, as follows:
Component Rate (Ib/hr) Composition (wt %)
Benzene 34.47 29.0
Nitrobenzene 0.18 0.1
N2 79.75 67.0
NO (N02) 3.28 2.8
/S
H20 1.25 1.1
Total 118.93 100.0
-------
U-bED BY ESTIMATOR.
MlU. PROB.
CO
-------
D-3
The specified system consists of a packed tower with the necessary instruments
and controls, a solvent feed pump, a refrigerated solvent cooler and the corre-
sponding refrigeration equipment, a tower bottoms-discharge pump, and a blower
to overcome tower pressure drop.
As designed the system uses nitrobenzene, chilled to 15°C, as the scrubbing
solvent and existing process capability for the separation of the absorbed
benzene by recycling the liquid bottoms stream from the absorber to an existing
nitrobenzene stripper. It is assumed that the existing stripper capacity is
sufficient to handle the additional load. Estimated control equipment costs
would be increased if additional stripping capacity is required.
Following is a summary of the design parameters used to estimate the capital and
operating costs. The absorber parameters were developed by standard design
methods described by Treybal.
Absorber tower, 10 in. dia, 15 ft packed height, 1/2-in. Raschig rings
Refrigeration, 1 ton at 15°C
Blower, 30 cfm, 8-in. WC
Pumps, 2 gpm
Solvent (nitrobenzene) rate, 452 Ib/hr at 15°C
Steam (for stripping), 0.5 Ib of steam/lb of stripper feed
Capital cost estimates were developed by the summation of installed costs for
the major individual components of each system. These installed capital costs
are based on IT Enviroscience experience, adjusted to a December 1979 base. On
this basis the installed capital cost for the absorption system is estimated to
be $48,000. The cost of utilities (stream and electrical power) is estimated to
be $5700/yr, and the fixed cost is estimated to be $13,900/yr ($48,000 X 29%).
With an estimated credit for recovered benzene of $28,600 ($0.10/lb) the absorp-
tion system would provide an estimated savings of $9000/yr.
"""R. E. Treybal, Mass-Transfer Operations, Chaps. 6 and 8, McGraw-Hill, New York,
1955.
-------
D-4
C. INCINERATION OF PROCESS EMISSIONS
A preliminary estimate was made of the size and cost for a thermal oxidizer to
incinerate the process VOC and NO emissions. The following design basis was
X
used for the estimate:
Model-plant capacity 90,000 Mg/yr
Waste-gas composition (Ib/hr)
Benzene 34.47
NB 0.18
N2 79.75
NO (N02) 3.28
A
H20 1.25
118.93
238 acfm at 60°F (including combustion air)
225 scfm at 32°F
Waste gas fuel valve 47 Btu/scf
The incinerator system must include a small combustion chamber for reducing NO
to N2 by the waste-gas stream being burned in a reducing atmosphere, with less
than theoretical air used for complete combustion. This chamber is followed by
the main combustion chamber, where additional air is introduced to oxidize the
organics. Some auxiliary fuel is required for flame stability, but the cost of
the small quantity of fuel is relatively insignificant.
It is estimated that the first combustion chamber will operate at approximately
2000°F and the second chamber at approximately 1600°F, which are adequate for
2
VOC destruction. The control device evaluation report for thermal oxidation
was used to determine the preliminary estimate for the thermal oxidizer. The
cost estimates presented in the thermal oxidation report do not cover any
thermal oxidizer sized to handle a waste-gas stream of less than 500 scfm, and
none are designed with two combustion chambers. The 500-scfm incinerator was
the smallest standard incinerator listed by any of the vendors contacted. For
this preliminary estimate it is reasonable to assume that the cost of an inciner-
2J. W. Blackburn, IT Enviroscience, Control Device Evaluation. Thermal Oxidation
Supplement (September 1980) (EPA/ESED report, Research Triangle Park, NC).
-------
D-5
ator with two combustion chambers in series sized to handle a waste-gas stream
of 225 scfm will be approximately the same as that for the smallest units
quoted. Although for the smallest units the duty specifications do not have a
large bearing on installed capital, the most appropriate duty specifications are
listed on the table of p. B-21 of the thermal oxidation report. On this basis
the installed cost for the thermal oxidizer is estimated to be $277,000. The
auxiliary fuel cost is considered to be negligible, the manpower requirement is
estimated to be $18,000/yr, and the fixed cost is estmated to be $80,000/yr
($277,000 X 29%). The total annual operating cost is estimated to be $98,000.
-------
E-l
APPENDIX E
EXISTING PLANT CONSIDERATIONS
A. CURRENT INDUSTRY
Information on control devices used by nitrobenzene producers was secured from
four producers for five nitrobenzene plants representing about 89% of the indus-
try capacity.
1. Dupont, Beaumont, TX
A water scrubber is used to control benzene-contaminated vent emission, and
benzene storage emissions are controlled by use of a floating-roof tank.
Streams of oxides of nitrogen contaminated with benzene are controlled by
incineration. A refrigerated vapor condenser is used for control of emis-
sions from the waste-acid tanks.
2. Dupont, Gibbstown, NJ
Streams of oxides of nitrogen contaminated with benzene are controlled by
condensation and a benzene-contaminated vent emission is controlled by
2
water scrubbing.
3. First Mississippi, Pascagoula, MS
An absorbing reactor, reported as being highly efficient, was initially
utilized, but it was subsequently indicated that the reactor was converted
to an absorption column, with nitrobenzene used as the scrubbing liquor,
for control of all process emissions.
4. Mobay, New Martinsville, WV
4
No control devices were reported.
5. Rubicon, Geismar, LA
An absorption column in which nitrobenzene is used as the scrubbing liquor
is used for control of all process emissions. A water scrubber is used for
control of emissions from a benzene-contaminated vent.
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 costs asso-
ciated 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. An absorption control system using nitro-
-------
E-2
benzene as the absorbing liquor could be especially difficult to retrofit if
existing nitrobenzene stripping capacity is insufficient for the increased
demand.
-------
E-3
E. REFERENCES*
1 C. W. Stuewe, IT Enviroscience, Trip Report on Visit to E. I. du Pont de Nemours
& Co., Beaumont, TX, Sept. 7, 8, 1977 (on file at EPA, ESED, Research Triangle
Park, NC).
2. W. Smith, E. I. du Pont de Nemours & Co., letter to D. R. Goodwin, EPA, Feb. 3,
1978.
3. R. Barker, First Chemical Corp., letter to D. R. Goodwin, EPA, Jan. 20, 1978.
4. L. P. Hughes, Mobay Chemical Corp., letter to D. R. Goodwin, EPA, Jan. 31, 1978.
5. C. W. Stuewe, IT Enviroscience, Trip Report on Visit to Rubicon Chemicals,
Geismar, LA, July 19, 20, 1977 (on file at EPA, ESED, Research Triangle Park,
NC).
6. W. C. Anthon, Rubicon Chemicals, letter to David A. Beck, EPA, Apr. 14, 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.
-------
2-i
REPORT 2
ANILINE
F. D. Hobbs
C. W. Stuewe
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
October 1980
This report contains certain information whirh 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.
D76N
-------
2-iii
CONTENTS OF REPORT 2
Page
I. ABBREVIATIONS AND CONVERSION FACTORS I"1
II. INDUSTRY DESCRIPTION II-1
A. Reason for Selection II~
B. Usage and Growth
C. References
III. PROCESS DESCRIPTION III-l
A. Introduction III-l
B. Nitrobenzene Hydrogenation Process III-l
C. Process Variations
III-5
ences
IV. EMISSIONS
D. References
IV-1
IV-1
A. Emissions
IV-4
B. References
V. APPLICABLE CONTROL SYSTEMS V~1
A. Emission Control Options v~1
B. References
VI. SUMMARY VI~1
A. Industry Capacity and Estimated Production VI-1
B. Estimated Emissions VI-1
C. References VI~2
APPENDICES OF REPORT 2
Page
A. PHYSICAL PROPERTIES OF ANILINE
B. EXISTING PLANT CONSIDERATIONS
-------
2-v
Number
II-l
II-2
IV-1
A-l
B-l
TABLES OF REPORT 2
Aniline Usage and Growth
Aniline Capacity
Uncontrolled Emissions from 100,000-Mg/yr Aniline Process Plant
Physical Properties of Aniline
Process Control Devices Used by Industry
Page
II-2
II-3
IV-2
A-l
B-2
Number
II-l
III-l
FIGURES OF REPORT 2
Aniline Manufacturing Locations
Process Flow Diagram for Manufacture of Aniline
Page
II-4
III-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)
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
Symbol
tera
giga
mega
kilo
milli
micro
Multiplication
Factor
1012
109
106
103
io"3
io"6
Example
1
1
1
1
1
1
Tg =
Gg =
Mg =
km =
mV =
pg =
1
1
1
1
1
1
X
X
X
X
X
X
10
10
10
12 grams
9
6
IO3
10
10
"""
™
grams
grams
meters
3 volt
s gram
-------
II-l
II. INDUSTRY DESCRIPTION
A. REASON FOR SELECTION
Aniline was selected for study because it is an important intermediate in the
synthetic organic chemicals manufacturing industry (SOCMI) and has a relatively
high production rate. The interrelationship between the manufacture of aniline
and nitrobenzene also was a consideration. It is estimated that 97% of the
nitrobenzene produced domestically is converted to aniline, as is cited in a
previous report.1 Nitrobenzene production results in emissions of significant
amounts of benzene, a substance listed as a hazardous pollutant by the EPA
(Federal Register, June 8, 1977). Aniline production also will create benzene
emissions if benzene remains with the nitrobenzene feed as an impurity. Emis-
sions of aniline itself are restricted because of its relatively low volatility
(see Appendix A for pertinent physical properties of aniline).
B. USAGE AND GROWTH
The end uses and expected growth rates of aniline are given in Table II-l. The
predominant use of aniline is as an intermediate in the manufacture of diphenyl-
methane diisocyanate (MDI) and its polymeric derivative polymethylenepolyphenyl
isocyanate (PMPPI), which are important in the production of polyurethane foams.2
The expected annual growth of 8% for this application of aniline could be higher
if government regulations require certain standards for insulation in residential
housing; on the other hand, it could be lower if a planned MDI plant based on
nitrobenzene instead of aniline proves to be commercially successful.3 Other
uses of aniline3 are as an intermediate in the production of rubber-processing
chemicals, hydroquinone, pesticide intermediates, dyes, and pharmaceuticals.
The current domestic aniline capacity is reported to be about 528,000 Mg/yr
(capacity increased about 153,000 Mg/yr during 1978 and 1979), with 1978
production utilizing about 53% of that capacity. The projected capacity will
increase to about 567,000 Mg/yr by 1983, and, based on predicted growth rates,
production will utilize about 66% of the capacity.3
Six producers were operating eight domestic aniline plants as of January 1,
1979. Table II-2 lists the producers and their capacities, and Fig. II-l shows
their locations. Several recent developments have affected the status of
-------
II-2
Table II-l. Aniline Usage and Growth*
Percentage of
Production
(1978)
Diphenylmethane diisocyanate (MDI)
Rubber chemicals
Dyes
Hydroquinone
Drugs, pesticides, and miscellaneous
52
29
4
3
12
1978 — 1983
Average Rate Growth
8.0
2.0 — 3.0
3.0
4.5
6.0
*See ref 3.
-------
II-3
a
Table II-2. Aniline Capacity
Capacity
(Mq/yr as of 1979)
Plant and Location u^y/_y
American Cyanamid, Bound Brook, NJ '
American Cyanamid, Willow Island, WV '
118,000
Du Pont, Beaumont, TX
73,000
Du Pont, Gibbstown, NJ e
First Chemical, Pascagoula, MS '
Mallinckrodt, Raleigh, NC
m 45,000
Mobay, New Martinsville, WV
127,000g
Rubicon, Geismar, LA
528,000
Total
See ref 3.
bCapacity brought back on-stream during 1978.
'Includes a 13,000-Mg/yr increase in capacity scheduled for late in 1978 or early
in 1979.
Includes a 13,000-Mg/yr increase in capacity scheduled for late in 1978.
Includes a 70,000-Mg/yr increase in capacity during 1977.,
Capacity figures not available (see ref 4); aniline produced as a by-product
in the synthesis of para-aminophenol.
Includes a 100,000-Mg/yr increase in capacity scheduled for early in 1979.
-------
II-4
1.
2.
3.
4.
American Cyanamid, Bound Brook,NJ 5.
American Cyanamid, Willow Island, WV 6.
Du Pont, Beaumont, TX 7.
Du Pont, Gibbstown, NJ 8.
First Chemical, Pascagoula, MS
Mallinckrodt, Raleigh, NC
Mobay, New Martinsville, WV
Rubicon, Geismar, LA
Fig. II-l. Aniline Manufacturing Locations
-------
II-5
aniline capacity. American Cyanamid's plant at Bound Brook, NJ, had been on
standby since 1974, but was brought back on-stream in 1978. The capacity of
the American Cyanamid plant at Willow Island, WV, is to be increased by about
27,000 Mg/yr in early 1980. It was reported that the capacity of both du Pont
plants was to be increased by 13,000 Mg/yr by late 1978 or early 1979. First
Chemical increased its capacity by 70,000 Mg/yr in 1977. No capacity figures
were located for the Mallinckrodt facility, where aniline is produced as a
by-product of para-amenophenal. Rubicon scheduled an increase in aniline
capacity of 100,000 Mg/yr for early 1979.3 5
An area of change in aniline production involves the methods of production.
Most current domestic production of aniline is based on catalytic hydrogenation
of vaporized nitrobenzene. However, it is reported4 that a liquid-phase
process is used commercially in the United States. The producer using this
process was not identified. Also, it is reported that beginning in 1981 Mobay
will recover aniline as a by-product from the production of iron oxide.3 Also, a
process based on the vapor-phase ammonolysis of phenol is used to produce
aniline in the foreign market.2 Since no further information has been obtained
concerning these processes, they are not covered in this report.
-------
II-6
C. REFERENCES*
1. F. D. Hobbs and C. W. Stuewe, IT Enviroscience, Inc., Nitrobenzene Product
(in preparation for EPA, ESED, Research Triangle Park, NC).
2. M. Cans, "Which Route to Aniline?" Hydrocarbon Processing 5_5(11), 145—150
(November 1976). —
3. E. M. Klapproth, "CEH Product Review on Aniline and Nitrobenzene," pp. 614.5030A—I
in Chemical Economics Handbook, Stanford Research Institute, Menlo Park, CA
(January 1979).
4. W. Lb'wenbach, J. Schlesinger, Nitrobenzene/Aniline Manufacture: Pollutant Pre-
diction and Abatement, MTR-7828, Metrek Division of the MITRE Corp. (May 1978).
5. S. N. Robinson, Mallinckrodt, Inc., letter to Robert E. Rosenteel, EPA,
July 28, 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.
-------
III-l
III. PROCESS DESCRIPTION
A. INTRODUCTION
Vapor-phase hydrogenation of nitrobenzene is the predominant domestic method of
aniline production, although liquid-phase hydrogenation is reported to be in
current use. Also, one producer reportedly plans to begin recovering aniline
from a process involving reaction of iron with nitrobenzene in the presence of
a hydrochloric acid catalyst.1 This process will yield aniline as a by-product
of the iron oxide product.1 Ammonolysis of chlorobenzene was once a significant
route to aniline, but no current domestic aniline production is based on this
process. Ammonolysis of phenol is used by foreign aniline producers.2 This
report presents details of the process based on vapor-phase hydrogenation of
nitrobenzene, the predominant domestic method of aniline production.
B. NITROBENZENE HYDROGENATION PROCESS
The vapor-phase hydrogenation of nitrobenzene is accomplished by the use of a
metal catalyst such as copper on a carrier according to the reaction
C6H5N02 + 3H2 Cu C6H5NH2 + 2H20
(nitrobenzene) (hydrogen) (copper) (aniline) (water)
The flow diagram shown in Fig. III-l represents a typical continuous process.
Nitrobenzene (stream 1) is vaporized and fed with 300% excess hydrogen (stream 2)
to a fluidized-bed reactor, which is held at about 270°C. Excess heat from the
exothermic reaction is removed by internal cooling coils in the reactor. Product
gases are filtered free of catalyst by internal filters in the top of the reactor.
Product gases (stream 3) are passed through a condenser. Condensed materials
(stream 4) are sent to a decanter, and noncondensables (stream 5) are recycled
to the reactor. Condensables form two phases in the decanter: a lower phase
(stream 6), which is crude aniline containing about 0.5% nitrobenzene and 5%
water, and an upper aqueous phase (stream 7). The crude aniline phase is passed
to a dehydration column that operates under vacuum. The aniline in the upper
aqueous phase is recovered either by stripping or by extraction with nitrobenzene
for recycle while the water is sent to wastewater treatment. Overheads from
the dehydration column (stream 8) are condensed and recycled to the decanter.
The bottoms (stream 9), which contain the aniline, are sent to the purification
-------
H
H
H
I
NJ
Fig. III-l. Process Flow Diagram for Manufacture of Aniline
-------
III-3
column. The column operates under vacuum. Overheads (stream 10) from the
purification column consist of product aniline. The bottoms (stream 11) are
tars, which are disposed of.3'4
Process emissions typically would originate from the purge of noncondensables
(Stream 5) during recycle to the reactor and from purge of inert gases from the
various items of separation and purification equipment (vents A).4
Fugitive emissions of nitrobenzene and aniline can occur when leaks develop in
valves, pump seals, and other equipment.
Storage emissions occur from tanks storing intermediate materials, final-product
aniline, and waste materials. Handling emissions occur from transfer of product
aniline for off-site shipment.
Potential sources of secondary emissions (D on Fig. III-l) are spent-catalyst,
wastewater, and tars.
C. PROCESS VARIATIONS
The following variations of the process shown in Fig. III-l are possible:
I. filtering catalyst fines from the product gases outside the reactor for
recycle of the catalyst,
2. using a nickel sulfide catalyst deposited on alumina in a fixed-bed reactor,
3. using liquid-phase processing with different catalysts,
4. purifying the crude aniline from the decanter (stream 6, Fig. III-l) by
first taking aniline and water overhead in a column, with heavies such as
nitrobenzene being removed in the column bottoms; the overheads would then
be distilled to separate the product aniline from water.5
Of these variations it is known that removal of catalyst from product gases
outside the reactor can have a significant influence on process emissions, as
is described in Sect. IV of this report. No information is available for
differences in emissions resulting from other variations. However, it is
believed that approximately 80% of the aniline currently produced in the United
States is manufactured by a process that is basically similar to the process
described in Sect. III-B of this report. Therefore the emissions discussed in
Sect. IV should accurately represent current practices.
-------
III-4
As was mentioned previously, ammonolysis of phenol is used by foreign producers
to manufacture aniline. No information is available on emissions from this
process for comparison to those from the vapor-phase hydrogenation of nitro-
benzene process.
-------
III-5
D. REFERENCES*
1. E. M. Klapproth, "CEH Product Review on Aniline and Nitrobenzene," pp. 614.5030AI
in Chemical Economics Handbook, Stanford Research Institute, Menlo Park, CA
(January 1979).
2. M. Cans, "Which Route to Aniline?," Hydrocarbon Processing 55(11), 145—150
(November 1976).
3. F. A. Lowenheim and M. K. Moran, Faith. Keyes, & Clark's Industrial Chemicals,
4th ed., pp. 113 and 114, Wiley-Interscience, New York, 1975.
4. C. W. Stuewe, IT Enviroscience, Inc., Trip Report on Visit to E. I. du Pont de
Nemours & Co, Beaumont, TX, Sept. 7,8, 1977 (on file at EPA, ESED, Research
Triangle Park, NC).
5. W. Lb'wenbach and J. Schlesinger, Nitrobenzene/Aniline Manufacture: Pollutant
Prediction and Abatement, MTR-7828, Metrek Division of the MITRE Corp.
(May 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
A. 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, parti-
cipate in photochemical reactions producing oxone. 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 oxone formation.
As is indicated on Fig. III-l, several process vents (vents A) are used to purge
inert gases from the production equipment. The uncontrolled total VOC process
emissions listed in Table IV-1 were calculated for a 100,000-Mg/yr production
plant at full capacity from information supplied by producers. The total process
emissions in Table IV-1 are a capacity-weighted average of the emissions reported
by producers. The benzene emissions were calculated from data reported by one
producer.1 A process variation that can significantly influence process emissions
is the manner in which the catalyst is handled. One producer reports filtration
of catalyst fines from the reaction gases outside the reactor for recycle.
This operation is reported1 to create an uncontrolled emission of 1.4 kg of VOC
per Mg of production. Another manufacturing location reports2 emissions from
catalyst handling to be 0.018 kg of VOC per Mg of production. However, in the
latter case it is not known whether the catalyst handling is for recycle of the
catalyst or for disposal of spent catalyst. Emissions from disposal of spent
catalyst would be classified as a secondary emission source.1—3
The storage emissions shown in Table IV-1 are a combination of reported emissions
from storage of crude aniline and waste materials2 and of calculated emissions
based on the estimated use of two aniline product day tanks and one final aniline
product tank. The calculations for emissions from these aniline tanks were
based on equations from AP-42,4 although breathing losses were divided by 4 to
account for recent evidence indicating that the AP-42 breathing loss equation
overestimates emissions.5 Emissions from loading aniline product into tank cars
and trucks based on submerged loading into clean vessels were calculated with
equations from AP-42.4
-------
IV-2
Table IV-1. Uncontrolled Process, Storage, and Handling
Emissions from a 100,000-Mg/yr Aniline Process Plant
Emission
Source
Process vents
Storage
Handling
Stream
Designation
(Fig. III-l)
A
B
C
Ratio
Benzene
0.0057d
Uncontrolled
(kg/Mg)b
Total VOC
0.095e
0.023
0.0012
Emissions
Rate
Benzene
0.065d
(kg/hr)c
Total VOC
1.08
0.26
0.014
aEmissions from plants employing no controls other than those necessary for
economical operation.
kg of emission per Mg of aniline produced.
CBased on 8760 hr/yr operation. Process downtime is normally expected to
range from 5 to 15%. If the hourly rate remains constant, the annual produc-
tion 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 assum-
ing continuous operation is negligible.
See ref 1.
Q
See refs 1—3.
-------
IV-3
As shown in Fig. III-l, there are three potential sources of secondary emissions:
spent-catalyst handling, wastewaters, and tars from the purification column.
Secondary emissions and fugitive emissions were not estimated for this abbre-
viated report. Storage and handling, fugitive, and secondary emissions for the
entire synthetic organic chemicals manufacturing industry are covered by
separate EPA documents.6—8
-------
IV-4
B. REFERENCES*
1. D. W. Smith, E. I. du Pont de Nemours & Co., letter to D. R Goodwin EPA
Feb. 3, 1978.
2. C. W. Stuewe, IT Enviroscience, Inc., Trip Report for Visit to E. I. du Pont
de Nemours & Co., Beaumont, TX. Sept. 7. 8. 1977 (on file at EPA, ESED, Research
Triangle Park, NC).
3. R. Barker, First Chemical Corp., letter to D. R. Goodwin, EPA, Jan. 20, 1978.
4. C. C. Masser, "Storage of Petroleum Liquids," pp. 4.3-1—4.3.12 in Compilation
of Air Pollutant Emission Factors. AP-42, Part A, 3d ed (August 1977).
5. E. C. Pulaski, TRW, Inc., letter dated May 30, 1979, to Richard Burn, EPA.
6. D. G. Erikson, IT Enviroscience, Inc., Storage and Handling (September 1980)
(EPA/ESED report, Research Triangle Park, NC).
7. D. G. Erikson and V. Kalcevic, IT Enviroscience, Inc., Fugitive Emissions
(September 1980) (EPA/ESED report, Research Triangle Park, NC).
8. 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
A. EMISSION CONTROL OPTIONS
Various control devices can be used for control of emissions from process, storage,
and secondary sources. Industry reports the control options currently in use
to be condensation, water scrubbing, dilute sulfuric acid scrubbing, and thermal^ ^
oxidation.1—5 Condensation is used for control of emissions from distillation,
from catalyst filtration and recycle.* and from storage.3 Water scrubbing is
used to control process and storage sources.3'4 Thermal oxidation is used to
control emissions from the reactor purge vent1'4 and secondary sources.4 Addi-
tional details are provided in Appendix B.
It is estimated that aniline process emissions account for less than 0.002% of
the total SOCMI emissions. Emissions from the aniline process are estimated to
be relatively low because of the low volatility of the materials involved and the
control devices already in use. Benzene emissions can occur from the production
of aniline only as a result of benzene impurities contained in the nitrobenzene
fed to the process.
-------
V-2
B. REFERENCES*
1. C. W. Stuewe, IT Enviroscience, Inc., Trip Report for Visit to E. I. du Pont
de Nemours & Co, Beaumont, XX. Sept. 7, 8. 1977 (on file at EPA, ESED, Research
Triangle Park, NC).
2. D. W. Smith, E. I. du Pont de Nemours & Co., letter to D. R Goodwin EPA
Feb. 3, 1978.
3. L. P. Hughes, Mobay Chemical Corp., letter to D. R. Goodwin, EPA, Jan 31
1978.
4. W. L. Anthon, Rubicon Chemicals Inc., letter to D. A. Beck, EPA, Apr. 14, 1978.
5. S. N. Robinson, Mallinckrodt, Inc., letter dated July 28, 1980, to Robert E Rosen-
steel, 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. SUMMARY
A. INDUSTRY CAPACITY AND ESTIMATED PRODUCTION
As is shown in Sect. II of this report, six domestic aniline producers were
operating eight plants as of January 1, 1979. These producers have a listed
capacity of 528,000 Mg/yr, although no capacity figure was located for the
Mallinckrodt plant in Raleigh, NC. As is also shown in Sect. II, industry
production was about 280,000 Mg in 1978. Based on an annual growth rate of 6%,
the 1979 production was estimated to have been 297,000 Mg.
B. ESTIMATED EMISSIONS
Current process emissions were estimated by calculating a capacity-weighted
average emissions ratio from information supplied by three producers1—3 and
multiplying that emission ratio times the estimated 1979 production listed above.
This calculation indicates a total 1979 process emission of about 26 Mg of total
VOC, which includes about 2 Mg of benzene. Storage, secondary, and fugitive
emissions are not included in this estimate. Current process emission control
devices reported to be in use by industry are described in Appendix B.
-------
VI-2
C. REFERENCES*
1. C. W. Stuewe, IT Enviroscience, Inc., Trip Report for Visit to E. I. du Pont
de Nemours & Co., Beaumont, XX, Sept. 7, 8, 1977 (on file at EPA, ESED,
Research Triangle Park, NC).
2. D. W. Smith, E. I. du Pont de Nemours & Co., letter to D. R. Goodwin, EPA,
Feb. 3, 1978.
3. W. L. Anthon, Rubicon Chemicals Inc., letter to D. A. Beck, EPA, Apr. 14, 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.
-------
A-l
APPENDIX A
Table A-l. Physical Properties of Aniline*
Synonyms
Molecular formula
Molecular weight
Physical state
Vapor pressure
Vapor density
Boiling point
Melting point
Density
Water solubility
Benzeneamine, benzamine, aminobenzine ,
phylamine , aminophen, aniline oil
93.12
Liquid
0.67 mm Hg at 25°C
3.22
184°C
-6.3°C
1.02173 at 20°C/4°C
36.5 g/liter of HO
*J. Dorigan et al., Scoring of Organic Air Pollutants—Chemistry,
Production and "roxicity of Selected Synthetic organic Chemicals
(Chemicals A—C) , MTR 7248, Rev. 1, Appendix, I, p. AI-78,
MITRE Corp. ,' Metrek Division (September 1976) .
-------
B-l
APPENDIX B
EXISTING PLANT CONSIDERATIONS
A EXISTING PLANT CONSIDERATIONS
Table B-l1-6 lists process control devices reported in use by industry. As is
described in the table, many of the control devices are also used for control
of storage and/or secondary sources.
B RETROFITTING CONTROLS
As is described in Sect. Ill of this report, numerous variations of the process
for production of aniline are possible. Some of these variations influence the
amount and rate of the emissions. For example, filtration of catalyst from
reaction gases outside the reactor for recycle creates a significant emzssxon
source as is described in more detail in Sect. IV. 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.
-------
B-2
Table B-l. Process Control Devices Used by Industry
Producer and Location
c
Du Pont, Beaumont, TX
g
Du Pont , Gibbstown , NJ
First Chemical Corp. ,
i
Pascagoula, MS
Mobay Chemical Co., New
Martinsville, WV3
Rubicon Chemical, Geismar, LA
Halinckrodt, Raleigh, NC
Devices
d
Condenser
Thermal oxidation
h
Condenser
None reported
k
Condensers
Water scrubber
m
Water scrubber
Thermal oxidizer
Dilute sulfuric acid
scrubber
Percentage
Control
NRS
NR6
96
NR6
e
NR
99.9
P
e
NR
Controlled Process
Emissions
Rate (kq/Mq)
0.014
NR6
0.056
NRS
.TDe
NR
o.ooin
•pi
0.38n
G
NR
Devices listed specifically for control of secondary emissions are not included here but
are listed in Section V of this report.
bkg of emission per Mg of reported capacity for the specific controlled emissions.
Q
See ref 1.
dCondenser on two distillation vents.
SNot reported or too little information available for calculation.
fReactor vented to combustion device; no information given for efficiency or final emissions.
h
dSee ref 2.
'The condenser is reported to control emissions from purging a catalyst filtration and re-
cycle operation. Vacuum-pump liquid-ring seals also are reported to be used on 3 distilla-
tion columns to reduce emissions and were judged to be normal items of equipment; disposal
route for the liquid is not described.
"'"See ref 3.
-'see ref 4.
"process and storage emissions are controlled separately by condensers and combined for
control by a water scrubber.
See ref 5.
mA scrubber is used to control combined storage and process emissions.
the calculations given here.
^
PThe overall efficiency for the combined process and secondary sources is reported to be
greater than 99 percent.
-------
B-3
C. REFERENCES*
1 c W Stuewe, IT Enviroscience, Inc., Trip Report for Visit to E. I. du Pont de
' Nemours & co. Beaumont. TX, Sept. 7.8, 1977 (on file at EPA, ESED,
Research Triangle Park, NC).
2. D. W. Smith, E. I. du Pont de Nemours & Co., letter to D. R. Goodwin, EPA,
Feb. 3, 1978.
3. R. Barker, First Chemical Corp., letter to D. R. Goodwin, EPA, Jan. 20, 1978.
4. L. P. Hughes, Mobay Chemical Corp., letter to D. R. Goodwin, EPA, Jan. 31, 1978.
5. W. L. Anthon, Rubicon Chemicals Inc., letter to D. A. Beck, EPA, Apr. 14, 1978.
6. S. N. Robinson, Mallinckrodt, Inc., letter dated July 28, 1980, to Robert E. Rosen-
steel, 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.
-------
3-i
REPORT 3
CUMENE
C. A. Peterson
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 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.
D58I
-------
3-iii
CONTENTS OF REPORT 3
Page
I. ABBREVIATIONS AND CONVERSION FACTORS I_l
II. INDUSTRY DESCRIPTION II-l
A. Reason for Selection II-l
B. Cumene Usage and Growth II-l
C. Domestic Producers II-3
D. References II-6
III. PROCESS DESCRIPTIONS III-l
A. Introduction III-l
B. Catalysis III-l
C. References 111-12
IV. EMISSIONS IV-1
A. Solid Phosphoric Acid Catalyst Process IV-1
B. Aluminum Chloride Catalyst Process IV-4
C. Other Processes IV-11
D. References IV-12
V. APPLICABLE CONTROL SYSTEMS V-l
A. Solid Phosphoric Acid Catalyst Process V-l
B. Aluminum Chloride Catalyst Process V-3
C. Other Processes V-9
D. References V-10
VI. IMPACT ANALYSIS VI-1
A. Environmental and Energy Impacts VI-1
B. Control Cost Impact VI-6
C. References VI-8
VII. SUMMARY VII-1
-------
3-v
APPENDICES OF REPORT 3
Page
A. PHYSICAL PROPERTIES OF PROPANE, PROPYLENE, BENZENE, ETHYLBENZENE, A-l
CUMENE, m-DIISOPROPYLBENZENE, £-DIISOPROPYLBENZENE
B. AIR-DISPERSION PARAMETERS B-l
C. FUGITIVE-EMISSION FACTORS C-l
D. EXISTING PLANT CONSIDERATIONS D-l
-------
3-vii
TABLES OF REPORT 3
Page
Number —*-
II-l
Cumene Production and Growth
11-2 Cumene Production Capacity, 1978 II~
IV-1 Total Uncontrolled VOC Emissions from the Model Plant for the IV-2
Cumene Manufacturing Process Using Solid Phosphoric Acid Catalyst
IV-2 Storage Tank Data for Model Plant Producing Cumene by Process IV-5
Using Solid Phosphoric Acid Catalyst
IV-3 Total Uncontrolled VOC Emissions from Model Plant for the Cumene IV-7
Manufacturing Process Using Aluminum Chloride Catalyst
IV-4 Storage Tank Data for Model Plant Producing Cumene by Process IV-10
Using Aluminum Chloride Catalyst
V-l VOC Controlled Emissions for Model Plant Producing Cumene by V-2
Process Using Solid Phosphoric Acid Catalyst
V-2 Storage Tank Data for Model Plant Producing Cumene by Process V-4
Using Solid Phosphoric Acid Catalyst
V-3 VOC Controlled Emissions for Model Plant Producing Cumene by V-6
Process Using Aluminum Chloride Catalyst
V-4 Storage Tank Data for Model Plant Producing Cumene by Process V-8
Using Aluminum Chloride Catalyst
VI-1 Environmental Impact of Controlled Model Plant Producing Cumene VI-2
by Process Using Solid Phosphoric Acid Catalyst
VI-2 Environmental Impact of Controlled Model Plant Producing Cumene VI-3
by Process Using Aluminum Chloride Catalyst
VII-1 Emission Summary for Model Plant Producing Cumene by Process VII-2
Using Solid Phosphoric Acid Catalyst
VII-2 Emission Summary for Model Plant Producing Cumene by Process VII-3
Using Aluminum Chloride Catalyst
A-l
A-l Physical Properties
B-l Air-Dispersion Parameters for Model Plant Producing B-l
Cumene by Process Using Solid Phosphoric Acid Catalyst and
with a Capacity of 227 Gg/yr
B-2 Air-Dispersion Parameters for Model Plant Producing B-2
Cumene by Process Using Aluminum Chloride Catalyst
D-l Emission Control Devices or Techniques Currently Used by Some D-2
Cumene Producers
-------
3-ix
FIGURES OF REPORT 3
Number Page
II-l Locations of Plants Manufacturing Cumene H-5
III-l Flow Diagram for Uncontrolled Model Plant Producing Cumene by III-2
Use of Solid Phosphoric Acid Catalyst
III-2 Flow Diagram for Uncontrolled Model Plant Producing Cumene by III-5
Use of Aluminum Chloride Catalyst
A-l Vapor Pressure vs Temperature A~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
(ms/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 IO1
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
1C9
106
103
io"3
io"6
Example
12
1 Tg = 1 X 10^ grams
1 Gg = 1 X IO9 grams
1 Mg = 1 X IO6 grams
1 km = 1 X IO3 meters
1 mV = 1 X IO"3 volt
1 (jg = 1 X IO"6 gram
-------
II-l
II. INDUSTRY DESCRIPTION
A. REASON FOR SELECTION1
Cumene production was selected for study because it is an aromatic chemical
that consumes benzene in its production; it is known that benzene causes harmful
health effects,-2 and the pattern of rapid industrial growth to high production
levels indicates that large quantities of benzene are being handled and consumed.
Benzene is present at relatively high levels in many of the process streams
during cumene manufacture,- so vents and other emission sources are likely to
discharge significant amounts of benzene vapors to the air unless appropriate
emission control techniques are used.
B. CUMENE USAGE AND GROWTH1'3*
Table II-l shows cumene production and growth rate. The predominant (99%) use
for cumene is in the manufacture of phenol and acetone by the cumene hydroperoxide
process. Small amounts of a-methylstyrene and acetophenone are also made from
cumene, usually as by-products from the cumene hydroperoxide process. In the
period from 1955 to 1975 the cumene hydroperoxide process grew to dominance as
the principal route used to manufacture phenol (and the co-product acetone).
In 1955, only 13% of the total domestic phenol and 8% of the domestic acetone
were manufactured from cumene. By 1975 these percentages had risen to 88% for
domestic phenol and 58% for domestic acetone.
Some cumene is sold on the open market to processors for conversion to phenol
and acetone, but a large share of the total cumene manufactured is further pro-
cessed to phenol and acetone by large, integrated producers that manufacture
cumene for use as an intermediate in their manufacturing complex. Because of
this large internal consumption of cumene by integrated producers, the data on
production of cumene shown in Table II-l are expected to contain some inaccuracies,
but these figures are the best numbers available. The current projected growth
rate of 4.4% is expected to continue through 1982.
*In order to minimize the revision time, the data used for the original draft
of this report have been retained. For our purposes the change in usage and
growth data is not believed to be significant.
-------
II-2
Table II-l. Cumene Production and Growth*
Year
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1982
Production Rate Growth Rate
(Gg/yr) (%/year)
72
77
80
97 15.3
99
133
175
196
249
301
406
514 24.3
611
765
899
972
1040
1209
1318 4.3
908
1197
1197
1257
1492 (est.) 4.4 (est.)
*Data for 1956 to 1976 from ref 1, p. 638.5030F;
data for 1977 through 1982 from ref 3.
-------
II-3
C. DOMESTIC PRODUCERS1'3—14
As of 1978, twelve producers of cumene in the United States were operating plants
at thirteen locations. Table II-2 lists the producers, plant capacity, and (where
known) the type of catalyst system used in the plant. Figure II-l shows the
locations of the 13 operating plants.
Marathon Oil Company has shut down their plant at Texas City, TX, which was
rated at a production capacity of 95.2 Gg/yr, and has indicated that they do
not intend to resume manufacture of cumene in this facility.4 Costal States
Petrochemical Company has converted their 64-Gg/yr cumene facility at Corpus
Christi, TX, to manufacture other products.12 The rated capacity of operating
plants in the United States is estimated at a total of 2193.6 Gg/yr (Table II-2).
The 1978 production was 1257 Gg (57% of capacity), and the estimate for 1982
production of 1492 Gg is only 68% of the rated capacity (Table II-l). Two new,
large plants have recently been started up: Shell's 317.5-Gg/yr plant at Deer
Park, TX (1977),3 and Georga Pacific's 340.1-Gg/yr plant at Houston, TX (1978).l
With these two new, large plants operating and with present and predicted operating
levels far below total plant capacity, it is expected that additional older,
smaller plants for manufacture of cumene will be shut down.
-------
II-4
Table II-2. Cumene Production Capacity, 1978
Company and Plant Location
Capacity
(Gg/yr)
Catalyst
System Type
Amoco Oil Co., Texas City, TX 13.6
Ashland Oil Co., Catlettsburg, KY 181.4b
Chevron Oil Co., El Segundo, CA 40.8a
Clark Oil Co., Blue Island, IL 54.43
Georgia Pacific Corp., Houston, TX 340.1C
Getty Oil Co., El Dorado, KS 61.2a
Gulf Oil Co., Philadelphia, PA 204.la
Gulf Oil Co., Port Arthur, TX 200.5b
Monsanto Chemical Co., Chocolate Bayou, TX 317.5
Shell Oil Co., Deer Park, TX 317.5a
Sun Petroleum Products Co., 104.3
Corpus Christi, TX
Texaco, Inc., Westville, NJ 68.Oa
Union Carbide Corp., Ponce, PR 290.2
Total 2193.6d
Unknown
Solid phosphoric acid
Unknown
Unknown
Solid phosphoric acia
Unknown
Solid phosphoric acia
Solid phosphoric acid
Solid phosphoric acid
Solid phosphoric acid
Solid phosphoric acid
Unknown
Aluminum chloride
From ref 3.
From individual company replies to EPA in response to their request for
information on cumene production.
"From ref 11.
Champlin Petroleum Co. is building a 181-Gg/yr plant at Corpus Christi, TX,
with completion scheduled for 1980; see ref 12.
-------
II-5
1. Amoco Oil Co., Texas City, TX
2. Ashland Oil Co., Catlettsburg, KY
3. Chevron Oil Co., El Segundo, CA
4. Clark Oil Co., Blue Island, IL
5. Georgia Pacific Corp., Houston, TX
6. Getty Oil Co., El Dorado, KS
7. Gulf Oil Co., Philadelphia, PA
8. Gulf Oil Co., Port Arthur, TX
9. Monsanto Chem. Co., Chocolate Bayou, TX
10. Shell Oil Co., Deer Park, TX
11. Sun Petroleum Products Co., Corpus Christi, TX
12. Texaco, Inc., Westville, NJ
13. Union Carbide Corp., Ponce, PR
Fig. II-l. Locations of Plants Manufacturing Cumene
-------
II-6
D. REFERENCES*
1. T. C. Gunn, "CEH Product Review on Cumene," pp. 638.5030A—638.5030N in
Chemical Economics Handbook, Stanford Research Institute, Menlo Park, CA
(March 1977).
2. "National Emission Standards for Hazardous Air Pollutants, Addition of Benzene
to List of Hazardous Air Pollutants," Federal Register 42 (110), 29332—29333
(Wednesday, June 8, 1977). —
3. "Chemical Profile on Cumene," in Chemical Marketing Reporter (June 12, 1978).
4. Albert 0. Learned, letter dated Sept. 11, 1978, to EPA from Marathon Oil Co.,
Texas City, TX, in response to EPA's request for information on the cumene
process.
5. J. R. Kampfhenkel, letter dated Sept. 12, 1978, to EPA from Sun Petroleum
Products Co., Corpus Christi, TX, in response to EPA's request for informaton
on the cumene process.
6. M. P. Zanotti, letter dated Sept. 19, 1978, to EPA from Gulf Oil Co.,
Port Arthur, TX, in response to EPA's request for information on the cumene
process.
7. F. D. Bess, letter dated Sept. 21, 1978, to EPA from Union Carbide Corp., South
Charleston, WV, in response to EPA's request for information on the cumene process.
8. Oliver J. Zandona, letter dated Sept. 25, 1978, to EPA from Ashland Petroleum
Co., Ashland, KY, in response to EPA's request for information on the cumene
process.
9. Michael A. Pierle, letter dated Oct. 23, 1978, to EPA from Monsanto Chemical
Co., St. Louis, MO, in response to EPA's request for information on the cumene
process.
10. Attachment II, Information on the Cumene Process, from Shell Oil Co., Deer
Park, TX, in response to EPA's request for information on the cumene process.
11. "Cumene Plant Operating at Georgia-Pacific Site," pp. 7, 49 in Chemical Marketing
Reporter (Feb. 12, 1979).
12. S. A. Al-Sayyari and Koon-Ling Ring, "Cumene," pp 638.5030A—638.5030Q in
Chemical Economics Handbook, Stanford Research Institute, Menlo Park, CA (March
1979).
13. J. B. Ellsworth, Georgia-Pacific Corp., letter dated Feb. 26, 1980, to J. R. Farmer,
EPA, with information on catalyst type used.
14. G. J. Wilson, Jr., Gulf Oil Co., U.S., letter dated Dec. 21, 1979, to J. R. Farmer,
EPA, with comments on draft Ci^ene report.
^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
INTRODUCTION1
In the United States at present all chemical-grade cumene is manufactured by
the alkylation of benzene with propylene. Benzene and propylene are reacted at
elevated temperatures and pressures in the presence of an acidic catalyst.
Excess benzene is used to minimize the formation of dialkylated or polyalkylated
benzenes. The catalysts used may be solid phosphoric acid (on a catalyst support,
such as alumina), aluminum chloride, or sulfuric acid. The reaction is exothermic.
Process yields are about 94%, based on the amount of benzene consumed, and
about 92%, based on the amount of propylene consumed. A simplified formula for
this reaction is as follows:
+ CH2=CH-CH3
(benzene) (propylene)
(cumene)
B.
CATALYSIS2
The selection of a catalyst system for the alkylation of benzene to cumene is
the most important choice that affects plant design, raw-material purity require-
ments, number of processing steps, material of construction constraints, emission
locations, and potential process emission quantities.
Solid Phosphoric Acid Catalyst3
Figure III-l is a typical flowsheet for the manufacture of cumene by the process
using phosphoric acid on a catalyst support (such as alumina). This is the
most favored catalyst system. Basic process patents are held by UOP, Inc.,
Institute Francais du Petrole, and Bayer A.G.1 Solid phosphoric acid is a
selective catalyst that promotes the alkylaLion of benzene with propylene in a
vapor-phase system that operates at about 205°C and 3.5 X 106 Pa.
Since the catalyst is selective, propylene feedstock for cumene manufacture
does not have to be thoroughly refined before use. Crude propylene streams (1)
from refinery crackers that are fractionated to about 70% propylene can be used
in this process without further purification. After the propylene is consumed,
-------
C>cx\y
D>S7/*,±. £ D
GUM c,\/ £
ff-e.Kf\/e&
TA.MK. (fj)
DtST/L.L^*.~r/O/V
Cur*E.rtm
QoT-row$
/?KC£ SV £&
TA**
-&
Fofi.
Use Ir*
A«. ruEi.
Fig. III-l. Flow Diagram for Uncontrolled Model Plant Producing Cumene by
Use of Solid Phosphoric Acid Catalyst
-------
Ill-3
the residual hydrocarbon stream (K3> (principally propane) can then be returned
to the refinery for use as feedstock or fuel gas. Higher boiling olefins such
as butylene should be removed from the propylene stream before they are used to
manufacture cumene.
The benzene (stream 4) used in this process does not have to be dried before it
is used, since the catalyst system requires small amounts of water vapor in the
reactor stream to activate the catalyst. The feed ratio is normally at least
4 moles of benzene (stream 4) per mole of propylene (stream I).4
Product purification is relatively simple with this catalyst, since no catalyst
removal processing is required. The propane (streams 9 and K ), the recycle
benzene (stream 3), and the cumene product (stream 12) can each be separated by
distillation. The residual bottoms (stream Kc) from the cumene distillation
b
column can be returned to the refinery for reforming or be used in the "gasoline
pool" or burned as fuel by inclusion in a fuel gas system.
The main process vent (A ) is associated with the depropanizer column and its
overhead receiver. Methane (or nitrogen) is used to blanket this system. A
pressure-control valve relieves excess pressure on this system by bleeding off
to the fuel gas system a mixture of methane (or nitrogen), propane, and accumu-
lated inert gases that are carried into the process with the crude propylene
(stream 1).
The second process vent (A ) is associated with the benzene recovery column.
This column is normally operated under pressure and is padded with methane (or
nitrogen). As pressure and receiver levels fluctuate, a pressure-control valve
relieves excess pressure on this system by bleeding off to the fuel gas system
a mixture of methane (or nitrogen), benzene vapor, and residual inert gases.
The third process vent (A ) is associated with the cumene distillation column.
This column is normally operated slightly above atmospheric pressure and is
padded with methane (or nitrogen) to protect the cumene from contact with the
air. As pressure and receiver levels fluctuate, a pressure-control valve relieves
excess pressure on this system by bleeding off a mixture of methane (or nitrogen)
and cumene vapor.
-------
III-4
Solid wastes (streams KI and KS respectively) are produced from the packed-bed
reactor and from the optional clay treatment vessels. These two solids streams
are not large, since they result from the periodic discharge of exhausted or
depleted bed solids, but the solids can contain some volatile organic compounds
(VOC). Purging and/or steam cleaning of the solids beds before the exhausted
solids are discharged would minimize the residual VOC they contain.
Contaminated wastewater streams (K2A/ K2B, and K2c respectively) exit from the
depropanizer column, the propane receiver tank, and the benzene receiver tank.
These wastewater streams will contain small quantities of dissolved VOC. The
principal contaminates will be benzene in streams K and K and propane in
Z.A /JC_
stream K The wastewater stream (principally K ) will also contain dissolved
£A
phosphoric acid and small quantities of dissolved or emulsified alkylbenzenes
such as cumene and diisopropylbenzene.
Propane is extracted from the crude product (stream 6) by the depropanizer column.
Some of the extracted propane is recycled (stream 9) to the reactor for cooling,
with the balance (stream K ) returned to the refinery for reuse.
The bottoms (stream K&) from the cumene distillation column contain principally
diisopropylbenzene, along with small amounts of other high-boiling materials.
This stream is returned to the refinery for reforming, for use in the refinery
gasoline pool, or for use as fuel. The overhead (stream 12) from this column is
the cumene product.
A purge stream (K4) of benzene is taken as a side stream from the recycle benzene
(stream 3) extracted from the crude product (stream 10) by the benzene recovery
column. The purge stream, which is sent back to the refinery for purification,
reforming, or use in the refinery gasoline pool, contains the small amount of
ethylbenzene and similar low boilers that were generated in the alkylation of
benzene with the crude propylene feed.
2. Aluminum Chloride Catalyst5
Figure III-2, pp. 1 and 2, is a typical flowsheet for cumene manufacture using
aluminum chloride as the alkylation catalyst. Aluminum chloride is a much more
active and much less selective alkylation catalyst than solid phosphoric acid.
Since aluminum chloride also functions as a transalkylation catalyst, diisopropyl-
-------
Fig. III-2. Flow Diagram for Uncontrolled Model Plant Producing Cumene by
Use of Aluminum Chloride Catalyst
(Page 1 of 2)
-------
H
n
o
3
ft
•0
(a
n>
K>
o
Ml
-------
III-7
benzene can be recycled back to the reaction system, where it reacts with excess
benzene to produce additional cumene. A simple equation for this transalkylation
reaction is as follows:
CH-(CH3)2
-CH-(CH3)2
(benzene) (mixed isomers of D.I.P.B.)
CH-(CH3)2
(cumene)
To prevent the generation of undesirable contaminating by-products, the propylene
used with this catalyst system must be purified to at least chemical grade (95%+
purity) and must contain no more than minute amounts of other olefins such as
ethylene and butylene. This propylene feedstock (stream 1) must also be dried
(stream 4) and treated to remove any residual organic sulfur compounds (stream 5).
Treatment of the propylene to remove residual water in fixed-bed dryers and
regeneration of the bed with heated methane generate a contaminated methane
(stream K ) that can be fed to the plant fuel gas manifold. The wastewater
(stream K ) generated by this process will contain traces of dissolved methane
and VOC.
Treatment of the propylene in a sulfur guard absorber will generate waste solid
(stream K ) in the form of spent absorbent. This waste solid will contain only
minor traces of VOC as propylene.
The benzene used in this process must be azeotropically dried (stream 7) to
remove dissolved water. The wastewater (stream K4) generated by the drying
step is saturated with dissolved benzene at about 2000 g/Mg of water. The azeo-
trope drying distillation generates a vent gas (stream AI) that is rich in benzene.
The aluminum chloride used as a catalyst in this process is received and handled
as a dry powder (stream 9). Benzene (stream 11) and diisopropylbenzene (stream 23)
are fed to a catalyst mix tank, where the aluminum chloride powder is added to
form the catalyst complex. This mixture is treated with hydrogen chloride gas
(stream 10) to activate the catalyst complex. The catalyst preparation operation
-------
III-8
generates a vent gas consisting of inert gases and hydrogen chloride gas saturated
with vapors of benzene and diisopropylbenzene. The scrubber is used to absorb
HC1 gas, and the residual vapors (stream AZ> are then vented.
The catalyst suspension (stream 13) and benzene (stream 12) are fed to the alkyla-
tion reactor as liquids, and the propylene (stream 5) is sparged into the bottom
of the reactor as a vapor. The alkylation reactor operates at about 90°C and
at relatively low pressure (about 150 kPa). The feed ratio to the alkylation
reactor is maintained at or above 4 moles of benzene per mole of propylene to
minimize formation of polyalkylated products and to permit transalkylation of
the recycle diisopropylbenzene to cumene. Since the alkylation reaction is
exothermic, heat is removed by jacket cooling water and/or by use of a reflux
condenser. A control valve after the reflux condenser maintains pressure in
the reactor system by discharging accumulated propane (stream 15) to the degassing
vessel as the reactor pressure rises above the setpoint.
The crude reaction mixture (stream 14) from the alkylation reactor is sent to
the degassing vessel, where dissolved low-boiling hydrocarbons (such as propane)
are released from solution.
The hydrocarbon vapor (stream 16) from the degassing vessel is sent to the caustic
gas scrubber, where a weak caustic solution (stream 18) is injected into the
scrubber system. The caustic solution (stream 20) is recycled over the scrubber
packing for absorption of residual hydrogen chloride out of the gas stream. A
side stream (21) of caustic solution is sent to the caustic wash tank.
The caustic washed hydrocarbon vapor (stream 22) is sent to the D.I.P.B. gas
scrubber, where it is contacted by recycled D.I.P.B. (stream 40). The D.I.P.B.
scrubber is used to extract residual unreacted propylene from the nonreactive
propane in the gas. After the vapor is scrubbed, the waste gas (stream K&) is
returned to the refinery, where it either is recycled to the olefins cracker
unit or is used as fuel gas. The D.I.P.B. liquid (stream 23) that contains the
absorbed propylene is sent to the catalyst mix tank.
The degassed product (stream 17) is sent to the acid wash tank, where it is
contacted with a weak acid solution (stream 24), which breaks down the catalyst
complex and dissolves the aluminum chloride in the water layer. The hydrocarbon
-------
III-9
portion of the catalyst complex blends with the rest of the hydrocarbon layer.
The water-hydrocarbon mixture (stream 25) is sent to the first decanter tank
for separation of the two layers. The wastewater (stream K ) from this decanter
tank contains some weak acid, dissolved aluminum chloride, and dissolved and
suspended residual hydrocarbons (principally benzene) as contained VOC.
The hydrocarbon layer (stream 26) from the first decanter tank enters the caustic
wash tank, where it is mixed with the dilute caustic (stream 21) from the caustic
gas scrubber. This dilute caustic layer extracts and neutralizes any residual
acid carried by the hydrocarbon layer. The mixed layers (stream 27) are sent
to the second decanter tank, where the hydrocarbon and aqueous layers settle
and separate. The wastewater (stream K0) from the second decanter tank contains
o
salt, traces of residual caustic, and some dissolved or suspended hydrocarbons
(principally benzene) as contained VOC.
The hydrocarbon layer (stream 28) from the second decanter tank enters the water
wash tank, where it is mixed with fresh process water. This fresh process water
extracts and removes any residual salt or other water soluble material from the
hydrocarbon layer. The mixed layers (stream 29) from the water wash tank are
sent to the third decanter tank, where the hydrocarbon and aqueous layers settle
and separate. The wastewater (stream K ) from the third decanter tank contains
traces of salt and some dissolved or suspended hydrocarbons (principally benzene)
as contained VOC.
The decanted hydrocarbon layer (stream 30) is stored in a washed-product receiver
tank. Traces of residual suspended water settle out in this receiver tank, and
the residual wastewater (stream K ) is periodically drained from the collection
sump of the receiver tank. This wastewater contains traces of salt and some
dissolved or suspended hydrocarbons (principally benzene) as contained VOC.
The entire wash-decanter system is tied together by one common vent-pad line
that furnishes nitrogen for blanketing this series of tanks. A pressure control
valve on the end of the vent-pad manifold periodically releases vent gas (stream A
as levels rise and fall in the various tanks of the wash-decanter system. The
vent gas is saturated with water vapor and hydrocarbon vapor (principally benzene)
as contained VOC.
-------
111-10
The crude product (stream 31) from the washed-product receiver tank is sent to
the benzene recovery column, where the excess benzene is stripped out of the
crude product. The recovered benzene (stream 33) is returned to the benzene
feed tank, and the crude cumene (stream 32) is stored in the crude cumene receiver
tank. Some residual water (stream K ) accumulates in the benzene-receiver-tank
collection sump and is periodically drained. This wastewater contains some
dissolved and/or suspended benzene as contained VOC. The vent line associated
with the benzene recovery column and with the benzene receiver tank releases
some vent gas (stream A ). This vent gas is principally inert gas saturated
with benzene vapor as the contained VOC.
The crude cumene (stream 32) is sent to the cumene distillation column for dis-
tillation of the cumene product (stream 35). The cumene distillation column
and the associated cumene receiver tank are operated above atmospheric pressure
and are blanketed with nitrogen (or methane) to protect the cumene from reaction
with oxygen from the air to form cumene hydroperoxide. The vent line associated
with the cumene distillation column and with the cumene receiver tank releases
some vent gas (stream A ). This vent gas is nitrogen (or methane) saturated
with cumene vapor as the contained VOC.
The crude D.I.P.B. (stream 34) is the bottoms stream from the cumene distillation.
column. This bottoms stream contains a small amount of cumene, along with mixed
isomers of diisopropylbenzene (D.I.P.B.) and a small amount of higher boiling
alkylbenzenes and miscellaneous tars. The crude D.I.P.B. stream is sent to the
D.I.P.B. stripping column, where it is stripped away from the residual higher
boiling alkylbenzenes and tars. This stripping column is normally operated
under vacuum because of the high boiling points of the D.I.P.B. isomers (about
200 to 210°C at atmospheric pressure). The vacuum system on the stripping column
does draw a vent stream (stream A,,) from the column condenser, and this vent
o
stream is air (or inert gas) saturated with cumene and D.I.P.B. vapors as the
contained VOC. Depending on the design and operation of the vacuum system for
the column, part or all of the vent gas (stream A&) could be discharged to the
atmosphere. The portion of the VOC in stream Afe that is not discharged directly
to the atmosphere would probably end up as a secondary waste stream that could
either be recovered for recycle or be sent to a waste disposal facility.
-------
III-ll
The distilled D.I.P.B. (stream 37) from the D.I.P.B. stripping column is recovered
and stored in the diisopropylbenzene storage tank.
The bottoms (stream 36) from the D.I.P.B. stripping column are stored in a bottoms
receiver tank and then sent to waste disposal for use as a fuel.
If excess distilled D.I.P.B. is accumulated from the cumene manufacturing
(alkylation) process, it (stream 39) can be added to the bottoms (stream 36) and
the combined waste stream (K ) be sent to waste disposal.
The recycle D.I.P.B. (stream 40) is sent to the D.I.P.B. gas scrubber, where it
is used to absorb residual propylene from the propane waste gas stream. This
recycle D.I.P.B. eventually returns to the alkylation reactor, where it is trans-
alkylated with excess benzene to generate additional cumene.
3. Other Catalysts
Other alkylation catalysts, such as concentrated sulfuric acid or anhydrous
hydrogen fluoride, can be used to catalyze the alkylation of benzene with propylene
to form cumene, but it is not known whether any of the present commercial producers
of cumene use any of these alternative catalysts. All the manufacturers of cumene
on which information on catalysts is known use either a solid phosphoric acid
catalyst or an aluminum chloride complex. If an alternate catalyst such as
concentrated sulfuric acid is used, the process and its characteristic emissions
would be similar to the process described for the aluminum chloride catalyst
system.
-------
111-12
C. REFERENCES*
1. T. C. Gunn, "CEH Product Review on Cumene," pp. 638.5030A—N in Chemical Economic;;
Handbook, Stanford Research Institute, Menlo Park, CA (March 1977).
2. Y. C. Yen, Phenol, Supplement A, pp. 19—41, Report No. 22A, A private report
by the Process Economics Program, Stanford Research Institute, Menlo Park, CA
(Sept. 1972).
3. D. J. Ward, "Cumene," pp. 543—546 in Kirk-Othmer Encyclopedia of Chemical
Technology, 2d ed., Vol. 6, edited by A. Standen et al., Wiley, New York, 1967.
4. R. H. Rosenwald, "Alkylation," pp. 58—61 in Kirk-Othmer Encyclopedia of Chemical
Technology, 3d ed. , Vol. 2, edited by M. Grayson e_t al. , Wiley-Interscience,
New York, 1978.
5. F. D. Bess, letter dated Sept. 21, 1978, to EPA from Union Carbide Corp., South
Charleston, WV, in response to EPA's request for information on the cumene 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 Sections 111 and 112 of the Clean Air Act since
there are associated health or welfare impacts other than those related to
oxone formation. It should be noted that although ethane is 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.
A. SOLID PHOSPHORIC ACID CATALYST PROCESS
1. Model Plant1—6
The model plant* for this study on the solid phosphoric acid catalyst process
for the manufacture of cumene has a production capacity of 227 Gg/yr based on
8760 hr/yr.** Actual capacities of the newer production plants using this catalyst
system vary from 181.4 to 317.5 Gg/yr. The flow diagram of the model plant
shown in Fig. III-l is typical of today's manufacturing and engineering technology.
The process shown is not necessarily identical to that used by any of the actual
operating plants, but the technology represented is close enough to be suitable
for emission control studies. Characteristics of the model plant important to
air dispersion are shown in Table B-l, Appendix B.
2. Sources and Emissions
Sources and emission rates for the solid phosphoric acid catalyst process are
summarized in Table IV-1.
*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.
-------
Table IV-1. Total Uncontrolled VOC Emissions from the Model Plant for the
Cumene Manufacturing Process Using Solid Phosphoric Acid Catalyst (227 Gg/yr)
Stream
Designation
Emission Source (Fig. III-l)
Cumene distillation A
system vent
Fugitive
Storage and handling
Secondary
Total
b
VOC Emissions Vent Gas VOC Emission Composition (wt %)
Ratio
(g/kg)c
0.03
0.24
0.27
0.008
0.55
Non-VOC
Rate Higher in Vent Gas
(kg/hr) C2 CB C4 Aliphatics Benzene Alkylbenzenes (wt %)
0.9 11. 7d 3.9d 6.1d 0 Trace 78.3 64.1
6.32
7.11
0.2
14.5
t
_
Uncontrolled emissions are emissions from the process for which no specific emission control devices (other than those M
necessary for economical operation) have been installed.
VOC emissions exclude methane, but include higher molecular weight organic compounds such as ethane, ethylene, propane,
propylene, butane, butenes, benzene, and various alkylbenzenes.
cg of emissions per kg of cumene produced.
The C2, C3/ and C^ indicated here are brought into the system with the crude methane used as an inert-gas blanket. If
pure methane or nitrogen were used as an inert-gas blanket, these emissions would not be present.
-------
IV-3
a. Cumene Distillation System Vent1—6 The cumene distillation system operates
slightly above atmospheric pressure to ensure that no air contacts the cumene
product. Cumene oxidizes easily to cumene hydroperoxide when contacted with
oxygen from the air, and the presence of cumene hydroperoxide (especially in a
cumene distillation system) could be very hazardous, since the vapor pressure
of cumene hydroperoxide is much lower (higher boiling point) than that of cumene
and could cause the cumene hydroperoxide to decompose violently if it accumulates
in the reboiler of the cumene distillation system.
The distillation system is pressurized with crude methane to maintain a minimum
pressure. As the pressure in the system fluctuates, a vent stream of crude
methane saturated with cumene vapors is periodically released through the pressure
control valve. The amount and composition given in Table IV-1 are intended to
represent typical emissions from a well-designed and -operated plant. If nitrogen
is used instead of crude methane for pressurization, the VOC emissions will be
less because the VOC from the crude methane will not be present. The VOC emis-
sions will be approximately the same as those shown for the model plant for the
aluminum chloride catalyst process (see Sect. IV-B-e) when nitrogen is used.
The crude methane stream is used to initially pressurize the cumene distillation
system and to maintain a minimum pressure on the system during operating. The
methane charged to the system is eventually vented (A , Fig. III-l) along with
•j
other hydrocarbon vapors. The crude methane is also used to purge the system
of liquid hydrocarbons during shutdowns and to drive out oxygen-containing air
before startups.
b. Fugitive Emissions Process pumps, piping flanges, and valves are potential
sources of fugitive emissions. The model plant is estimated to have 28 pumps
in light-liquid service, 200 process valves in light-liquid service, and 6 con-
trol valves (safety-relief valves) in vapor service. The factors in Appendix C
were used to determine the emission contribution of these equipment components.
For the model plant it is estimated that approximately 6.32 kg/hr as VOC is
lost to the atmosphere.
c. Storage and Handling Emissions7'8 Emissions result from the storage and handling
of raw materials, intermediates, and finished products. A list of the storage
-------
IV-4
tanks, the materials stored, and the assumed turnovers per year for the model
plant is given in Table IV-2. For material that is not produced or consumed
captively it is assumed that shipment is by rail car or by barge. The uncon-
trolled emissions were calculated based on fixed-roof tanks, painted white,
with conservation vents. Day-night temperature variations were assumed to
average 11.1°C. Emission equations from AP-42 were used with one modification.
The breathing losses were divided by 4 to account for recent evidence that the
AP-42 breathing-loss equation overpredicts emissions.
d- Secondary Emissions1—6 The principal sources of secondary VOC emissions are
the process wastewater streams from the depropanizer column, the propane receiver
tank, and the benzene receiver tank. It is assumed that these wastewater streams
are combined and sent through an oil skimmer tank for removal of any floating
layer of hydrocarbons and that the oily skimmings are returned to the process.
After the skimming step, it is assumed that the combined process wastewater
stream is sent to the plant wastewater system. This wastewater will still con-
tain dissolved hydrocarbons, such as propane (trace), benzene (up to 2000 ppm
of water), and assorted alkylbenzenes (up to 400 ppm of water). The total waste-
water flow is estimated to be 75 kg/hr for the model plant. The amount of benzene
and alkylbenzene in the wastewater is estimated to be approximately 0.2 kg/hr.
Extremely minor sources of secondary VOC emissions are the waste catalyst from
the multistage packed-bed reactor and the spent clay from the optional clay
treatment vessels. No estimate of the amount of VOC from these solid-waste
sources has been made.
B. ALUMINUM CHLORIDE CATALYST PROCESS
1. Model Plant1'9
The model plant for this study on the aluminum chloride catalyst process for
the manufacture of cumene has a production capacity of 227 Gg/yr based on
8760 hr/yr. The actual capacity of the one known cumene plant using an aluminum
chloride catalyst is 290 Gg/yr. There may be other plants (in the unknown-catalyst
category) that also use this catalyst system for the manufacture of cumene.
The flow diagram of the model plant shown in Fig. III-2 is typical of today's
manufacturing and engineering technology. The process is not necessarily identical
-------
IV-5
Table IV-2. Storage Tank Data for Model Plant Producing Cumene by
Process Using Solid Phosphoric Acid Catalyst
Contents
Benzene
Cumene bottoms
Finished cumene
Finished cumene
Cumene
Cumene
Total
Tank Size
(m3)*
8891
334
870
870
8891
8891
Bulk
Turnovers Temperature
per Year (°C)
20
77
150
150
14.8
14.8
20
20
20
20
20
20
Losses
(kg/hr)
6.08
0.009
0.111
0.111
0.399
0.399
7.11
*Fixed-roof tanks, painted white, with conservation vents; day-night
temperature variation averages 11.1°C.
-------
IV-6
to that used by any actual operating plant, but the technology represented is
close enough to be suitable for emission control studies. Characteristics of
the model plant important to air dispersion are shown in Table B-2, Appendix B.
2. Sources and Emissions
Sources and emission rates for the aluminum chloride catalyst process are sum-
marized in Table IV-3.
a. Benzene Azeotrope Drying Column Vent1'9 The vent (AI, Fig. Ill-2) from the
benzene azeotrope drying system discharges inert gas, water vapor, and benzene
vapor. This azeotrope distillation system operates above atmospheric pressure
and is blanketed by nitrogen (inert gas) to maintain column pressure and to
purge the column during shutdowns and startups. A pressure control valve is
used to maintain column pressure, and the discharge from this control valve
contains the VOC that is released. The composition and amount of this stream
are controlled by the vapor pressure of the benzene-water condensate and by the
amount of inert gas that must be vented. The amount and composition given in
Table IV-3 are intended to represent typical emissions from a well-designed and
operated process.
b. Catalyst Mix Tank Scrubber Vent1'9 The vent (AZ, Fig. III-2) from the catalyst
mix tank discharges a mixture of HCl gas and organic vapor consisting principally
of benzene and some diisopropylbenzene. Since HC1 gas is both toxic and corrosive,
this vent gas cannot be released directly to the atmosphere without treatment.
Normal treatment consists of scrubbing with water or an alkaline solution to
absorb and remove the HCl gas. Most of the organic vapors will also be condensed
and dissolved by the scrubber water used to remove the HCl. The residual vent
gas discharged by the scrubber will also carry some residual organic vapors
with it. The amount and composition given in Table IV-3 are intended to repre-
sent typical emissions from the vent of the scrubber in a well-designed and
operated process.
c. wash-Decanter System Vent1'9 The vent <&3, Fig. III-2, p. 2) from the wash-
decant system is shown as a common header with a nitrogen pad and a single relief-
valve outlet. Since the wash-decant system operates continuously with no significant
changes in liquid levels, the normal discharge from this vent is zero. Level
-------
Table IV-3. Total Uncontrolled3 VOC Emissions from Itodel Plant for the Cumene
Manufacturing Process Using Aluminum Chloride Catalyst (227 Gg/yr)
Stream
Designatior
Emission Source (Fig.III-2)
Benzene azeotrope drying
column
Catalyst mix tank scrubber
Wash -decanter system
Benzene recovery column
Cumene distillation system
D.I.P.B. stripping system
Fugitive
Storage and handling
Secondary
Total
Al
A2
A3
A4
A5
A6
VOC Emissions'3 Vent Gas VOC Emission
i Ratio
(g/kg)°
0.02
0.16
0.01
0.017
0.003
0.0009
0.51
0.97
0.23
1.92
Composition (wt %)
Rate
(kg/hr) C3 Benzene Alkylbenzene
0.54 100
4.0 Trace 99.4 0.6
0.3 78.4 21.6
0.43 100
0.07 100
0.02 100
13.3
25.1
6.0
49.8
Non-VOC
in Vent Gas
(wt %)
72
66
68
72
79
91
Uncontrolled emissions are emissions from the process for which no specific emission control devices (other
than those necessary for economic or safety reasons) have been installed.
bVOC emissions exclude methane, but include higher molecular weight organic compounds such as ethane, ethylene,
propane, propylene, butane, butenes, benzene, and various alkylbenzenes.
°g of emissions per kg of cumene produced.
-------
IV-8
fluctuations during startups and shutdowns can cause intermittent venting of
nitrogen gas contaminated with organic vapors, such as benzene, cumene, etc.
The amount and composition given in Table IV-3 are intended to represent the
average emissions generated by the periodic releases from this wash-decant system
in a well-designed and operated process.
d «»n«.ne Recovery Column Vent*"— The vent (A,. Fig. III-2. P- 2) from the benzene
recovery column discharges inert gas, water vapor, and benzene vapor. This
benzene recovery column operates above atmospheric pressure and is blanketed
with nitrogen (inert gas) to maintain column pressure and to purge the column
during startups and shutdowns. A pressure control valve is used to maintain
column pressure, and discharges from this control valve contain the VOC that is
released The composition and amount of VOC in this stream are controlled by
the vapor pressure of the benzene-water condensate and the amount of inert gas
that must be vented. The amount and composition given in Table IV-3 are intended
to represent typical emissions from a well-designed and operated process.
e Cumene Distillatior^^temJTent^ '»— The vent (A... Fig. III-2. p. 2) from the
cumene distillation system vent contains inert gas and cumene vapor. Thxs dis-
tillation system operates slightly above atmospheric pressure and is blanketed
with nitrogen to protect the cumene from oxidation to cumene hydroperoxide by-
atmospheric oxygen. A pressure control valve is used to maintain column pressure,
and discharges from this control valve contain the VOC that is released. The
composition and amount of VOC in this stream are controlled by the vapor pressure
of the cumene condensate and the amount of inert gas that must be vented. The
amount and composition given in Table IV-3 are intended to represent typical
emissions from a well-designed and operated process.
f
D ! P B stripping system contains inert gas (air) and diisopropylben2ene vapors.
This system operates at atmospheric pressure to strip off the diisopropylbensene
from the residual high-boiling impurities. The v.nt gas fro. this system contaln,,
minor counts of VOC in the for* of diisopropyibenzene vapors. The —* ^
D X P B is controlled by the -apor pressure of the diisopropyloensene condensate
and'the'a^ount of inert gas that Must be vented. The amount and compos^on
given in Table IV-3 are intended to represent typical emissions fr» a well-
designed and operated process.
-------
IV-9
g. Fugitive Emissions Process pumps, piping flanges, and valves are potential
sources of fugitive emissions. The model plant is estimated to have 56 pumps
in light-liquid service, 500 process valves in light-liquid service, and 10 control
valves (safety-relief valves) in vapor service. The factors in Appendix C were
used to determine the emission contribution of these equipment components. For
the model plant it is estimated that approximately 13.3 kg/hr as VOC is lost to
the atmosphere.
h. Storage and Handling Emissions7'8 Emissions result from the storage and handling
of raw materials, intermediates, and finished products. A list of the storage
tanks, the materials stored, and the assumed turnovers per year for the model
plant is given in Table IV-4. For material that is not produced or consumed
captively it is assumed that shipment is by rail car or by barge. The uncon-
trolled emissions were calculated based on fixed-roof tanks, painted white,
with conservation vents. Day-night temperature variations were assumed to average
11.1°C. Emission equations from AP-42 were used with one modification. The
breathing losses were divided by 4 to account for recent evidence that the AP-42
breathing-loss equation overpredicts emissions.
i. Secondary Emissions1'9 The principal sources of secondary emissions are the
various wastewater streams generated by the process. These wastewater streams
are-. K , wastewater from the gas driers,- K , wastewater from the decanter on
the benzene azeotrope drying columing,- K , catalyst mix tank scrubber wastewater;
K_, K0, and K0, wastewater streams from the decanters of the product wash steps ,-
78 9
and K and K , wastewater from the water collection sumps of the washed-product
receiver tank and the benzene receiver tank. It is assumed that all these waste-
water streams are collected, combined, and sent to a final oil skimmer sump for
collection of any residual oil layer. After the skimming step the combined
wastewater stream is sent through an underground sewer system to the plant waste-
water biooxidation treatment system, and the oil layer is returned to the washed-
product receiver tank. The combined wastewater stream will still contain dissolved
hydrocarbons, such as benzene (up to 2000 g/mg of water) and assorted alkylbenzenes
(up to 400 g/mg of water). The total wastewater flow is estimated to be about
2500 kg/hr for the model plant. This amount of benzene and alkylbenzene in the
wastewater is estimated to be approximately 6 kg/hr.
-------
IV-10
Table IV-4. Storage Tank Data for Model Plant Producing Cumene
by Process Using Aluminum Chloride Catalyst
Contents
Benzene
Benzene
Benzene
Benzene
Benzene
Mixture
Mixture
Cumene (crude)
Cumene (crude)
Cumene (finished)
Cumene (finished)
Cumene (finished)
Cumene (finished)
D.I.P.B. (crude)
D.I.P.B. (crude)
D.I.P.B. (finished)
D.I.P.B. (finished)
Heavy oil
D.I.P.B. (finished)
Total
Tank Size
(m3)*
8891
1800
1800
1800
1800
1800
1800
870
870
870
870
8891
8891
80
80
80
80
17.8
1422
Bulk
Turnovers Temperature
per Year (°C)
20
148
148
148
148
179
179
165
165
150
150
14.8
14.8
161
161
139
139
101
16.9
/
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
Losses
(kg/hr)
6.08
3.08
3.08
3.08
3.08
2.71
2.71
0.111
0.111
0.111
0.111
0.399
0.399
0.003
0.003
0.001
0.001
0.000
0.009
25.1
*Fixed-roof tanks, painted white, with conservation vents; day-night
temperature variation averages 11.1°C.
-------
IV-11
OTHER PROCESSES1
The literature describes other catalysis schemes that will promote the alkylation
of benzene with propylene to form cumene. These alternative catalyst systems
include the following: phosphoric acid—boron trifloride complex; aluminum
chloride—phosphoric acid complex; concentrated sulfuric acid; anhydrous hydro-
fluoric acid; boron-trifluoride-modified alumina; boron trifluoride complexed
with either water or sulfuric acid; alkane—sulfuric acid complex; silica-alumina,
with or without hydrogen chloride; zinc chloride on silica,- activated clay;
VOC13—(C2H5)2A1C1; rhenium chloride; and many others. The reaction schemes
using these various catalysts would be similar to that for either the solid
phosphoric acid catalyst process or the aluminum chloride catalyst process, and
their characteristic emissions would also be similar. Although other catalyst
systems have been described and patented, the two systems (solid phosphoric
acid and aluminum chloride) seem to dominate the industry, with the solid phosphoric
acid route being preferred by most producers.
-------
IV-12
D. REFERENCES*
1. Y. C. Yen, Phenol, Supplement A, pp. 19—41, A private report by the Process
Economics Program, Stanford Research Institute, Menlo Park, CA (September 1972).
2. J. R. Kampfhenkel, letter dated Sept. 12, 1978, to EPA from Sun Petroleum Products
Co., Corpus Christi, XX, in response to EPA's request for information on the
cumene process.
3. M. P. Zanotti, letter dated Sept. 19, 1978, to EPA from Gulf Oil Co., Port Arthur,
TX, in response to EPA's request for information the cumene process.
4. Oliver J. Zandona, letter dated Sept. 25, 1978, to EPA from Ashland Petroleum
Co., Ashland, KY, in response to EPA's request for information on the cumene
process.
5. Michael A. Pierle, letter dated Oct. 23, 1978, to EPA from Monsanto Chemical
Co., St. Louis, MO, in response to EPA's request for information on the cumene
process.
6. Attachment II, Information on the Cumene Process, from Shell Oil Co., Deer Park,
TX, in response to EPA's request for information on the cumene process.
7. 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).
8. E.G. Pulaski, TRW, letter dated May 30, 1979, to Richard Burr, EPA.
9. F. D. Bess, letter dated Sept. 21, 1978, to EPA from Union Carbide Corp., South
Charleston, WV, in response to EPA's request for information on the cumene
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.
-------
V-l
V. APPLICABLE CONTROL SYSTEMS
A. SOLID PHOSPHORIC ACID CATALYST PROCESS1—5
1. Cumene Distillation System Vent
The stream from the cumene distillation system vent (A , Fig. III-l) consists
principally of cumene vapors, together with some low-molecular-weight C2, C3,
and C4 hydrocarbons that are introduced with the crude methane used to blanket
the distillation system. Heating value of this vent stream (including the methane
used for blanketing) is approximately 0.13 GJ/hr for the model plant.
The control system evaluated for this vent stream is the installation of a piping
manifold to direct the vent gas, which contains VOC, to the plant emergency
flare system for destruction of the VOC by thermal oxidation. A VOC removal
efficiency of 95%* has been assumed when the flare is operating at less than
10% of design capacity.6 The controlled emission for this vent is shown in
Table V-l.
2. Fugitive Emission Sources
Controls for fugitive emissions from the synthetic organic chemicals manufac-
turing industry are discussed in a separate EPA report.7 Emissions from pumps
and valves can be controlled by an appropriate leak-detection system, along
with repair of leaky or defective equipment as needed. 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 major leaks are
detected and corrected as noted in Appendix C.
3. Storage and Handling Sources
It is important to control the VOC emissions, particularly benzene, in the storage
and handling areas because of health and safety hazards. Options for control
of storage and handling emissions are covered in another EPA report.8 For the
model plant the VOC emissions from storage tanks containing benzene are controlled
*Flare efficiencies have not been satisfactorily documented except for specific
designs and operating conditions using specific fuels. Efficiencies cited
are for tentative comparison purposes.
-------
Table V-l. VOC Controlled Emissions for Model Plant Producing Cumene by Process Using
Solid Phosphoric Acid Catalyst
Stream
Designation
Emission Source (Fig- III-l)
Cumene distillation A3
system vent
Fugitive
Storage and handling
Benzene
Other
Secondary
Total
Control Device
or Technique
Plant flare
Detection and cor-
rection of major
leaks
Floating roofs
None
None
Total VOC
Emission
Reduction (%)
95C
71.4
85
0
0
VOC Controlled
Emissions
Ratio (g/kg)b Rate (kg/hr)
0.0015
0.070
0.035
0.040
0.008
0.155
0.05
1.8
0.912
1.029
0.20
3.99
From refs 1—5.
bg of emissions per kg of cumene produced.
C95% efficiency at less than 10% of flare design capacity.
<
-------
V-3
by using floating-roof tanks* in place of fixed-roof API tanks. The controlled
VOC emissions from storage tanks that contain benzene were calculated on the
assumption that a contact type of internal floating roof with secondary seals
will reduce fixed-roof-tank emissions by 85%9'10 and are listed in Table V-2
and summarized in Table V-l. No control has been identified for the tanks con-
taining cumene or by-products.
11
No
4. Secondary Sources
The control of secondary emissions is discussed in a separate EPA report
control system has been identified for the model plant.
B. ALUMINUM CHLORIDE CATALYST PROCESS12
1. Benzene Azeotrope Drying-Column Vent
The stream from the benzene azeotrope drying-column vent (h^, Fig. III-2) is
relatively small and consists largely of benzene vapor and inert gas. The heating
value of the vent vapor is approximately 0.02 GJ/hr for the model plant.
The control system evaluated for this vent stream is the installation of a piping
manifold to direct the VOC-containing gas to the plant emergency flare system
for destruction of the VOC by thermal oxidation. A VOC removal efficiency of
95% has been assumed when the flare is operating at less than 10% of design
capacity.6 The controlled emission for this vent is shown in Table V-3.
2. Catalyst Mix Tank Scrubber Vent
The stream from the vent (A Fig. III-2) scrubber on the catalyst mix tank is
the largest source of VOC process emission in the aluminum chloride catalyst
cumene model plant. The VOC in this vent stream consists largely of benzene
vapor. The heating value of the vent vapor is approximately 0.17 GJ/hr for the
model plant.
The control system evaluated for this vent stream is the installation of a piping
manifold to direct the VOC-containing gas to the plant emergency flare system
for destruction of the VOC by thermal oxidation. A VOC removal efficiency of
^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
Table V-2. Storage Tank Data for Model Plant Producing Cumene
by Process Using Solid Phosphoric Acid Catalyst
Contents
Benzene
Cumene bottoms
Finished cumene
Finished cumene
Cumene
Cumene
Total
Tank Size
(m3)a
8891
334
870
870
8891
8891
Roof
Style
Floating
Fixed
Fixed
Fixed
Fixed
Fixed
Turnovers
per Year
20
77
150
150
14.8
14.8
Bulk
Temp
(°c)
20
20
20
20
20
20
Losses ,
(kg/hr)°
0.912
0.009
0.111
0.111
0.399
0.399
1.94
a
Floating- or fixed-roof tanks painted white, with conservation vents on
fixed-roof tanks; day-night temperature variation averages 11.1°C.
From refs 9 and 10.
-------
V-5
95% has been assumed when a flare is operating at less than 10% of design capac-
ity.6 The controlled emission for this vent is shown in Table V-3.
3. Wash-Decanter System Vent
This stream from the wash-decanter system vent (A^, Fig. III-2, p. 2) consists
principally of benzene and alkylbenzene in an inorganic gas stream. The heating
value of the VOC in this vent gas is approximately 0.01 GJ/hr for the model
plant.
The control system evaluated for this minor vent stream is the installation of
a piping manifold to direct the VOC-containing gas to the plant emergency flare
system for destruction of the VOC by thermal oxidation. A VOC removal efficiency
of 95% has been assumed when the flare is operating at less than 10% of design
capacity.6 The controlled emisson for this vent is shown in Table V-3.
4. Benzene Recovery Column Vent
The stream from the benzene recovery column vent (A , Fig. III-2, p. 2) consists
principally of benzene in an inert-gas stream and is relatively small. The
heating value of the VOC in this vent gas is approximately 0.02 GJ/hr for the
model plant.
The control system evaluated for this vent stream is the installation of a piping
manifold to direct the VOC-containing gas to the plant emergency flare system
for destruction of the VOC by thermal oxidation. A VOC removal efficiency of
95% has been assumed when the flare is operating at less than 10% of design
capacity.6 The controlled emission for this vent is shown in Table V-3.
5. Cumene Distillation System Vent
The stream from the cumene distillation system vent (AS/ Fig. III-2, p. 2) con-
sists principally of cumene in an inert-gas stream. This vent stream contains
a very small amount of VOC, whose heating value is approximately 3 MJ/hr.
Since this VOC emission is so low, no emission control system was evaluated.
The emission from this vent i* shown in Table V-3.
-------
Table V-3. VOC Controlled Emissions for Model Plant Producing Cumene by Process Using
Aluminum Chloride Catalyst
Stream
Designation
Emission Source (Fig- IH-1)
Benzene azeotrope Al
drying column
Catalyst mix tank AZ
scrubber
Wash-decanter system A3
Benzene recovery A4
column
Cumene distillation A^
system
D.I.P.B. stripping A^
system
Fugitive
Storage and handling
J: Benzene
Other
Secondary
Total
Control Device
or Technique
Plant flare
Plant flare
Plant flare
Plant flare
None
None
Detection and cor-
rection of major
leaks
Floating roofs
None
None
Total VOC
Emission
Reduction (%)
95
95
95
95
0
0
71.5
85
0
0
VOC Controlled
Emissions
Ratio (g/kg) Rate (kg/hr)
0.001
0.008
0.0005
0.00085
0.003
0.0009
0.146
0.138
0.049
0.23
0.577
0.027
0.20
0.015
0.022
0.07
0.02
3.79
3.57
1.26
6.0
15.0
From ref 12.
3g of emissions per kg of cumene produced.
-------
V-7
6. D.I.P.B Stripping System Vent
This stream from the D.I.P.B. stripping system vent (A&, Fig. III-2, p. 2) con-
sists principally of D.I.P.B. vapors in an inert-gas stream. This vent stream
contains an extremely small amount of VOC, which has a heating value of approxi-
mately 0.8 MJ/hr.
Since this VOC emission is so low, no emission control system was evaluated.
The emission from this vent is shown in Table V-3.
7. Fugitive Emission Sources
Controls for fugitive emissions from the synthetic organic chemicals manufac-
turing industry are discussed in a separate EPA report.7 Emissions from pumps
and valves can be controlled by an appropriate leak-detection system, along
with repair of leaky or defective equipment as needed. Controlled fugitive
emissions calculated with factors given in Appendix C are included in Table V-3.
These factors are based on the assumption that major leaks are detected and
corrected as described in Appendix C.
8. Storage and Handling Sources
It is important to control the VOC emissions, particularly benzene, in the storage
and handling areas because of health and safety hazards. Options for control
of storage and handling emissions are covered in another EPA report.8 For the
model plant the VOC emissions from storage tanks containing benzene are controlled
by using floating-roof tanks in place of fixed-roof API tanks. The controlled
VOC emissions from storage tanks that contain benzene were calculated on the
assumption that a contact type of internal floating roof with secondary seals
will reduce fixed-roof-tank emissions by 85%9/1° and are listed in Table V-4
and summarized in Table V-3. No controls have been identified for tanks con-
taining cumene or by-products.
9. Secondary Sources
The control of secondary emissions is discussed in a separate EPA report.
No control system has been identified for the model plant.
-------
V-8
Table V-4. Storage Tank Data for Model Plant Producing Cumene
by Process Using Aluminum Chloride Catalyst
Contents
Benzene
Benzene
Benzene
Benzene
Benzene
Mixture
Mixture
Cumene (crude)
Cumene (crude)
Cumene (finished)
Cumene (finished)
Cumene (finished)
Cumene (finished)
D.I.P.B. (crude)
D.I.P-B. (crude)
D.I.P.B. (finished)
D.I.P.B. (finished)
Heavy oil
D.I.P.B. (finished)
Total
Tank Size
(m3)9
8891
1800
1800
1800
1800
1800
1800
870
870
870
870
8891
8891
80
80
80
80
17.8
1422
Roof
Style
Floating
Floating
Floating
Floating
Floating
Floating
Floating
Fixed
Fixed
Fixed
Fixed
Fixed
Fixed
Fixed
Fixed
Fixed
Fixed
Fixed
Fixed
Turnovers
per Year
20
148
148
148
148
179
179
165
165
150
150
14.8
14.8
161
161
139
139
101
16.9
Bulk
Temp
(°c)
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
Losses ,
(kg/hr)D
0.912
0.462
0.462
0.462
0.462
0.407
0.407
0.111
0.111
0.111
0.111
0.399
0.399
0.003
0.003
0.001
0.001
0.000
0.009
4.83
^Floating- or fixed-roof tanks, painted white, with conservation vents on
fixed-roof tanks,- day-night temperature variation averages 11.1°C.
DFrom refs 9 and 10.
-------
V-9
OTHER PROCESSES13
No attempt has been made to estimate VOC emissions, sources, or possible VOC
emission control techniques for other process routes or alternate catalyst sys-
tems that might be used to manufacture cumene. It is believed that the possible
alternate processes and catalyst systems will be similar in equipment character-
istics and process emissions to the two processes and catalyst systems described.
As far as is known, only the solid phosphoric catalyst process and the aluminum
chloride catalyst process are used commercially in the United States.
-------
V-10
D. REFERENCES*
1. J. R. Kampfhenkel, letter dated Sept. 12, 1978, to EPA from Sun Petroleum Products
Co., Corpus Christi, TX, in response to EPA's request for information on the
cumene process.
2. M. P. Zanotti, letter dated Sept. 19, 1978, to EPA from Gulf Oil Co., Port Arthur,
TX, in response to EPA's request for information the cumene process.
3. Oliver J. Zandona, letter dated Sept. 25, 1978, to EPA from Ashland Petroleum
Co., Ashland, KY, in response to EPA's request for information on the cumene
process.
4. Michael A. Pierle, letter dated Oct. 23, 1978, to EPA from Monsanto Chemical
Co., St. Louis, MO, in response to EPA's request for information on the cumene
process.
5. Attachment II, Information on the Cumene Process, from Shell Oil Co., Deer Park,
TX, in response to EPA's request for information on the cumene process.
6 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).
7. D. G. Erikson and V. Kalcevic, IT Enviroscience, Inc., Fugitive Emissions
(September 1980) (EPA/ESED report, Research Triangle Park, NC).
8. D. G. Erikson, IT Enviroscience, Inc., Storage and Handling (September 1980)
(EPA/ESED report, Research Triangle Park, NC).
9 C C Masser, "Storage of Petroleum Liquids," Sect. 4.3 in Compilation of Air
Pollutant Emission Factors, 3d ed., Part A, AP-42, EPA, Research Triangle Park,
NC (April 1977).
10. W. T. Moody, TRW, Inc., letter dated Aug. 15, 1959, to D. A. Beck, EPA.
11. J. J. Cudahy and R. L. Standifer, IT Enviroscience, Inc., Secondary Emissions
(June 1980) (EPA/ESED report. Research Triangle Park, NC).
12 F D Bess letter dated Sept. 21, 1978, to EPA from Union Carbide Corp., South
Charleston, WV, in response to EPA's request for information on the cumene process.
13. Y. C. Yen, Phenol. Supplement A, pp. 19—41, A private report by the Process
Economics Program, Stanford Research Institute, Menlo Park, CA (September 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 ?he 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
Tables VI-1 and VI-2 show the environmental impacts of reducing VOC emissions
from solid phosphoric catalyst cumene plants and aluminum chloride catalyst
cumene plants by application of the described control systems (Sect. V) to the
model plants. From an energy standpoint, typical uncontrolled model plants for
both processes will consume heat in the range of 4.6 to 7.0 MJ/kg of product
and will consume power in the range of 0.13 to 0.15 MJ/kg of product, while
releasing about 4.8 to 7.2 MJ/kg of product to the environment in the form of
low-temperature heat.1
1. Solid Phosphoric Acid Catalyst Process2—6
The emissions from the solid phosphoric acid model plant are discussed in Sect. IV,
and emission control techniques are discussed in Sect. V. It is estimated that
the current total domestic capacity for cumene manufacture by this process is
about 1750 Gg/yr. The environmental and energy impacts for control of emissions
from this process are as follows:
a. Cumene Distillation System Vent Emissions from the cumene distillation system
can be controlled by installing a piping manifold to direct the vent gas to the
plant emergency flare system. In the model plant, direction of this vent gas
to the plant flare would reduce VOC emissions from this source by about 7.5 Mg/yr.
Burning of the VOC in the plant emergency flare system would release about
0.13 GJ/hr of heat to the environment for the model plant.
b. Fugitive Emissions The control methods previously described for these emissions
are major leak detection and correction as described in Appendix C. Application
of these methods would result in a VOC emission reduction of 39.6 Mg/yr from
the model plant.
c. Storage and Handling7 The control method previously described for reduction
of VOC emissions from storage tanks consists of installing floating roofs on
the tanks that handle benzene or organic mixtures containing benzene. Applica-
tion of this method would reduce the VOC emissions from the model plant by about
45.3 Mg/yr.
-------
Table VI-1. Environmental Impact of Controlled Model Plant Producing Cumene by
Process Using Solid Phosphoric Acid Catalyst
Stream
Designation
Emission Source (Fig. III-l)
Cumene distillation system vent A3
Fugitive
Storage and handling
Benzene
Other
Secondary
Total
Control Device
or Technique
Plant flare
Detection and correction of
major leaks
Floating roofs
None
None
VOC Emission
(%)
95
71.4
85
0
0
Reduction
(Mg/yr)
7.5
39.6
45.3
0
0
92.4
M
I
-------
Table VI-2. Environmental Impact of Controlled Model Plant Producing Cumene by
Process Using Aluminum Chloride Catalyst
Emission Source
Benzene azeotrope drying
column vent
Catalyst mix tank
scrubber vent
Wash-decanter system vent
Benzene recovery column vent
Cumene distillation system vent
D.I.P.B. stripping system vent
Fugitive
Storage and handling
Benzene
Other
Secondary
Total
Stream
Designation
(Fiq. III-2)
Al
A2
A3
A4
A5
A6
Control Device
or Technique
Plant flare
Plant flare
Plant flare
Plant flare
None
None
Detection and correction of
major leaks
Floating roofs
None
None
VOC Emission
(%)
95
95
95
95
0
0
71.5
85
0
0
Reduction
(Mg/yr)
4.46
33.25
2.52
3.59
0
0
83.2
177
0
0
304
I
Ul
-------
VI-4
d- 1978 Industrial Emissions It has been estimated that the current industrial
capacity for manufacture of cumene by the solid phosphoric acid catalyst process
is 1750 Gg/yr. Using the figure of 57% for capacity utilization, this amounts
to a production level of 1000 Gg in 1978. It has been estimated that the actual
emissions from cumene manufacture by the solid phosphoric acid catalyst process
were 200 Mg in 1978 (assuming current control at 85% of the level to be achieved
by a controlled model plant). For the uncontrolled model plant at 227 Gg/yr the
emission level is 130 Mg/yr. For the controlled model plant at 227 Gg/yr the
low value for the emission level is 35 Mg/yr.
2. Aluminum Chloride Catalyst Process8
The emissions from the aluminum chloride catalyst process model plant are dis-
cussed in Sect. IV, and emission control techniques are discussed in Sect. V.
It is estimated that the current total domestic capacity for cumene manufacture
by this process or by closely allied equivalent processes is about 400 Gg/yr.
The environmental and energy impacts for control of emissions from this process
are as follows:
a- Benzene Azeotrope Drying-Column Vent The control method previously described
for reduction of VOC emissions from the benzene azeotrope drying column consists
of installing a piping manifold to deliver this vent gas to the plant emergency
flare system. Use of this method would reduce VOC emissions from the model
plant by about 4.46 Mg/yr. For the model plant burning of the VOC in the plant
emergency flare system would release about 0.02 GJ/hr as heat to the environ-
ment.
b. Catalyst Mix-Tank Scrubber Vent The control method previously described for
reduction of VOC emission from the catalyst mix-tank scrubber vent consists of
installing a piping manifold to deliver this vent gas to the plant emergency
flare. Use of this method would reduce VOC emissions from the model plant by
about 33.3 Mg/yr. For the model plant burning of the VOC in the plant emergency
flare system at 95% efficiency would release about 0.166 GJ/hr as heat to the
environment.
c. Wash-Decanter System Vent The control method previously described for reduction
of VOC emission from the wash-decanter system consists of installing a piping
-------
VI-5
manifold to deliver the vent gas to the plant emergency flare system. Use of
this method would reduce VOC emissions from the model plant by about 2.5 Mg/yr.
For the model plant burning of the VOC in the plant emergency flare system at
95% efficiency would release about 0.01 GJ/hr as heat to the environment.
d. Benzene Recovery Column Vent The control method previously described for reduc-
tion of VOC emission from the benzene recovery column consists of installing a
piping manifold to deliver the vent gas to the plant emergency flare system.
Use of this method would reduce the VOC emissions from the model plant by about
3.6 Mg/yr. For the model plant, burning of the VOC in the plant emergency flare
system at 95% efficiency would release about 0.019 GJ/hr as heat to the environment.
e. Cumene Distillation System Vent Because of the small amount of VOC emitted
from the cumene distillation system, no control technique for reduction of VOC
emissions was evaluated for normal operation.
f. D.I.P.B. Stripping System Vent Because of the small amount of VOC emitted
from the D.I.P.B. stripping system, no control technique for reduction of VOC
emissions was evaluated for normal operation.
b. Fugitive Emissions—-The control methods previously described for these emissions
are major leak detection and correction as described in Appendix C. Application
of these methods would result in a VOC emission reduction of 83.2 Mg/yr for the
model plant.
h. Storage and Handling7 The control method previously described for reduction
of VOC emissions from storage tanks consists of installing floating roofs on
tanks handling benzene or organic mixtures containing benzene. Application of
this method to the model plant would reduce emissions by about 177 Mg/yr.
i. 1978 Industrial Emissions It has been estimated that the current industrial
capacity for manufacture of cumene by the aluminum chloride catalyst process is
400 Gg/yr. Using the figure of 57% for capacity utilization, this amounts to a
production level of 230 Gg in 1978. For the uncontrolled model plant at 227 Gg/yr
the emission level is 440 Mg/yr. For the controlled model plant at 227 Gg/yr
the emission level is 130 Mg/yr. It has been estimated that the actual emissions
-------
VI-6
from cumene manufacture by the aluminum chloride catalyst process were 180 Mg
in 1978 (assuming current control at 85% of the level specified for control of
the model plant).
B. CONTROL COST IMPACT
Details of the model plants (Figs. III-l and III-2) are given in Sect. Ill and
control techniques are discussed in Sect. IV.
1. Solid Phosphoric Acid Catalyst Process
a. Cumene Distillation System Vent The VOC emissions from this vent are relatively
small. The only technique that seemed reasonable was to inject this vent gas
into the manifold leading to the plant emergency flare system. The cost impact
of connecting the cumene distillation system vent to the flare manifold is negli-
gible when a new plant is being designed. The cost of retrofitting this control
to an existing plant may be appreciably greater than the cost for a new installa-
tion if there is some distance between the source and the existing flare manifold.
b. Fugitive Emission Sources A control system for fugitive sources is defined in
Appendix C. A separate EPA report covers fugitive emissions and their applic-
able controls for the synthetic organic chemicals manufacturing industry.
a.
9
c. Storage and Handling Sources The use of floating roofs on tanks handling benzene
or mixtures containing benzene has been selected as the technique for reduction
of VOC emissions from the model plant. No economic evaluation or cost-benefit
analysis for floating-roof versus fixed-roof tanks has been prepared for this
report. The economics for floating-roof versus fixed- roof storage tanks is
covered in a separate EPA report.10
2. Aluminum Chloride Catalyst Process8
Process Vents The control technique that was selected for all the process
vents was to inject the vent gas into the manifold leading to the plant emergency
flare system. The cost impart of connecting these vents to the flare manifold
is negligible when a new plant is being designed. The cost of retrofitting
this control to an existing plant may be appreciably greater than the cost for
-------
VI-7
a new installation if there is some distance between the sources and the
existing flare manifold.
b- Cumene Distillation System Vent The VOC emissions from this vent during normal
operation are very small, and no control system was evaluated.
c- D.I.P.B. Stripping System Vent The VOC emissions from this vent during normal
operation are very small, and no control system was evaluated.
d- Fugitive Emission Sources A control system for fugitive emission sources is
defined in Appendix C. A separate EPA report covers fugitive emissions and
their applicable controls for the synthetic organic chemical manufacturing
industry.9
e- Storage and Handling Sources7'9 The use of floating roofs on tanks handling
benzene or mixtures containing benzene has been recommended as the technique
for reduction of VOC remissions from this model plant. No economic evaluation
or cost-benefit analysis for floating-roof versus fixed-roof tanks has been
prepared for this report. The economics for fixed-roof versus floating-roof
storage tanks are covered in a separate EPA report.10
-------
VI-8
C. REFERENCES*
1- Y. C. Yen, Phenol, Supplement A. pp. 19—41, A private report by the Process
Economics Program, Stanford Research Institute, Menlo Park, CA (September 1972).
2. J. R. Kampfhenkel, letter dated Sept. 12, 1978, to EPA from Sun Petroleum Products
Co., Corpus Chnsti, TX, in response to EPA's request for information on the
cumene process.
3. M. P. Zanotti, letter dated Sept. 19, 1978, to EPA from Gulf Oil Co., Port Arthur
TX, in response to EPA's request for information the cumene process.
4. Oliver J. Zandona, letter dated Sept. 25, 1978, to EPA from Ashland Petroleum
Co., Ashland, KY, in response to EPA's reuqest for information on the cumene
process.
5. Michael A. Pierle, letter dated Oct. 23, 1978, to EPA from Monsanto Chemical
Co., St. Louis, MO, in response to EPA's request for information on the cumene
process.
6. Attachment II, Information on the Cumene Process, from Shell Oil Co., Deer Park
TX, in response to EPA's request for information on the cumene process.
7. 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).
8. F^ D. Bess, letter dated Sept. 21, 1978, to EPA from Union Carbide Corp., South
Charleston, WV, in response to EPA's request for information on the cumene process.
9. D. G. Erikson and V. Kalcevic, IT Enviroscience, Inc., Fugitive Emissions
(September 1980) (EPA/ESED report. Research Triangle Park, NC).
10. D. G. Erikson, IT Enviroscience, Inc., Storage and Handling (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
Cumene is manufactured domestically by the alkylation of benzene with propylene.
The two processes of commercial significance use different catalysts and operating
conditions to promote the alkylation reaction. Domestic production of cumene
(including Puerto Rico) was estimated to be 1257 Gg in 1978, with an estimated
total plant capacity of 2194 Gg/yr, giving an industrial capacity utilization
rate of 57%. The principal domestic use of cumene is in the manufacture of
phenol, along with co-product acetone, by the cumene hydroperoxide process.
The estimated annual growth rate for cumene manufacture is 4.4%/yr.
Emission sources along with uncontrolled and controlled air emission rates for
the solid phosphoric acid catalyst model-plant process for cumene manufacture
are given in Table VII-1. The comparable sources and values for the aluminum
chloride catalyst model-plant process for cumene manufacture are given in
Table VII-2.
None of the process-generated VOC emissions from the solid phosphoric acid catalyst
process or from the aluminum chloride catalyst process are very large. The
technique that was evaluated for controlling these emissions would be to collect
them in a piping system and to inject the collected vent gases into the manifold
header leading to the plant emergency flare for thermal destruction. The largest
and most significant VOC emissions are released by storage tanks handling benzene.
The use of floating roofs on storage tanks handling benzene is the preferred
way to control these sources of VOC emissions.
The average level of control for VOC emissions from existing cumene manufacturing
plants is estimated to be at least 85% of the control level for the controlled
emission model plants. At this estimated level of control the 1978 total level
of VOC emissions is estimated to be about 380 Mg/yr.
The solid phosphoric acid process is preferred by most of the manufacturers of
cumene, since it can use a crude propylene stream from an adjacent refinery
cracker, as well as refined benzene from the same adjacent refinery. The solid
phosphoric acid catalyst is selective for alkylation of benzene with propylene,
with a minimum of other alkylbenzenes being generated. A fairly large purge
-------
VII-2
Table VII-1. Emission Summary for Model Plant Producing Cumene by
Process Using Solid Phosphoric Acid Catalyst (227 Gg/yr)
Designation VOC Emission Rate (kg/hr)
Emission Source (Fig. III-l) Uncontrolled Controlled
Cumene distillation system A, 0.9 0.05
Fugitive <6.3 1.8
Storage and handling 7.1 1.94
Secondary 0.2 0.2
Total 14.5 3.99
-------
VII-3
Table VII-2. Emission Summary for Model Plant Producing
Cumene by Process Using Aluminum Chloride Catalyst (227 Gg/yr)
Emission Source
Stream
Designation
(Fig. III-2)
VOC Emission Rate (kg/hr)
Uncontrolled Controlled
Benzene azeotrope drying
column
Catalyst mix tank scrubber
vent
Wash-decanter system
Benzene recovery column
Cumene distillation system
D.I.P.B. stripping column
Fugitive
Storage and handling
Secondary
Total
0.54
4.0
0.027
0.20
0.3
0.43
0.07
0.02
13.3
25.1
6.0
0.015
0.022
0.07
0.02
3.79
4.83
6.0
49.8
15.0
-------
VII-4
stream of recovered benzene is returned to the refinery to remove impurities
from the cumene plant recycle stream. The crude propane left over after the
propylene is extracted is also returned to the refinery. Because of the close
links to refinery operation, this solid phosphoric acid catalyst process is
economically attractive only when closely associated with an adjacent refinery.
It is estimated that the total cumene capacity by this route is about 1750 Gg/yr,
or about 80% of the total domestic cumene capacity.
The aluminum chloride catalyst process is preferred by a few manufacturers of
cumene, since it uses chemical-grade propylene (about 95% purity) and refined
benzene. Feedstock costs are higher for chemical-grade propylene than for a
crude refinery stream, but the amount of propane and other contaminants that
must be handled and rejected by this process is much lower than the amount of
those in the gas stream rejected by the solid phosphoric acid process. This
aluminum chloride process does not require close linkage to a refinery operation,
but can function as an independent plant. The by-product diisopropylbenzene
formed in this process can be recycled back to the reaction section for trans-
alkylation with excess benzene to form additional cumene, thereby increasing
yields. It is estimated that the total cumene capacity by this route is about
400 Gg/yr, or about 20% of the total domestic cumene capacity.
-------
A-l
APPENDIX A
Table A-l. Physical Properties*
Material
Propane
Propylene
Benzene
Ethylbenzene
Cumene
m-Diisopropylbenzene
g-Diisopropylbenzene
Molecular
Formula Weight
C3H8 44.1
C3H6 42 . 1
C6H6 78.1
C8H10 106.2
C9H12 120.2
C12H18 162.3
C^2 18 162.3
*Values abstracted from J. B. Maxwell, Data
New York City, 1955,
and from R. C. Weast
Boiling Freezing
Point Point
-44.5
-47.8
80.1
136.2
152.4
203
210
-189.9
-185.2
5.5
-94.9
-96.0
-63.0
-17
Book on Hydrocarbons ,
et al.,
Specific
Gravity,
20/4°C
of Liquid
0.508
0.522
0.878
0.867
0.866
0.856
0.857
Gross
Heat of
Combustion
(MJ/kg)
50.4
48.9
41.8
43.0
43.4
45.5
45.5
Van Nostrand,
Handbook of Chemistry and Physics,
The Chemical Rubber Co., Cleveland, 1964.
-------
A-2
10
.001
ISO* 2OCT 220" 2*O*C
Fig. A-l. Vapor Pressure vs Temperature
-------
Table B-l Air-Dispersion Parameters for Model Plant Producing Cumene by Process Using
Solid Phosphoric Acid Catalyst and with a Capacity of 227 Gg/yr
— ——————— —
Stream
Designation
tfi a III-l)
A3
A3
voc
Emission
Rate Height Diameter
(g/sec) (m) (m)
Uncontrolled
0.25 27 0.025
1.76
1.98
0.056
Controlled
0.013 73 Unknown
0.50
0.25
0.056
Total
Discharge Flow Discharge
Temperature Rate Velocity
(K) (m3/sec) (m/sec)
322 0.0008 1.6
322
293
298
Vv f*
1250 Variable Variable
322
293
298
Cumene distillation
system vent
Fugitive
Storage and handling
Secondary
Cumene distillation
system vent
a
Fugitive
Storage and handling
Secondary
fugitive emissions are distributed over an area of about 200 m by 300 m.
b . .
Minimum.
C1.2 minimum.
-------
B-2
Table B-2. Air-Dispersion Parameters for Model Plant Producing Cumene by
Process Using Aluminum Chloride Catalyst (Capacity, 227 Gg/yr)
Source
Benzene azeotrope drying
column
Catalyst mix tank scrubber
Wash-decanter system
Benzene recovery column
Cumene distillation system
D.l.P.B. stripping system
Fugitive*
Storage and handling
Secondary
Benzene azeotrope drying
column
Catalyst mix tank scrubber
Wash-decanter system
Benzene recovery column
Cumene distillation system
D.l.P.B. stripping system
Fugitive*
Storage and handling
Secondary
Stream
Designation
(Fig. III-2)
Al
A2
A3
A4
A5
A6
Al
A2
A3
A4
A5
A6
voc
Emission
Rate
(g/sec)
0.15
1.11
0.083
0.119
0.0194
0.006
3.69
6.97
1.667
0.0075
0.0556
0.0042
0.0061
0.0194
0.0056
1.053
1.34
1.667
Height
(m)
Uncontrolled
36
36
36
36
36
36
Controlled
73
73
73
73
36
36
Diameter
(m)
0.025
0.076
0.038
0.025
0.064
0.025
Unknown
Unknown
Unknown
Unknown
0.064
0.025
Discharge
Temperature
(K)
322
293
293
322
322
322
1250 min
1250 min
1250 min
1250 min
322
322
Total
Flow
Rate
(m3/sec)
0.00060
0.00449
0.00118
0.00053
0.0206
0.00030
Variable
Variable
Variable
Variable
0.0206
0.00030
Discharge
Velocity
(m/sec)
1.2
1.0
1.0
1.0
6.5
0.6
Variable
(1.2 min)
Variable
(1.2 min)
Variable
(1.2 min)
Variable
(1.2 min)
6.5
0.6
•Fugitive emissions are distributed over an area of about 200 m by 300 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
refineries. Other preliminary test results suggest that fugitive emissions
from sources in chemical plants are comparable to fugitive emissions from cor-
responding 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- source)
0.12
0.02
0.021
0.010
0.0003
0.16
0.006
0.009
0.44
0.00026
0.032
Controlled
a
Emission Factor
(kg/hr-source)
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 to correct 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, Radian Corporation, EPA 600/2-79-044 (February 1979)
-------
D-l
APPENDIX D
EXISTING PLANT CONSIDERATIONS
A. CHARACTERIZATION
Table D-l lists the emission control techniques reported in use by industry.
Sources of information in this appendix are letters in response to requests by
EPA for information on emissions from cumene plants.1—6
B. RETROFITTING CONTROLS
The primary difficulty with retrofitting the controls described in this report
is that the distances between the vents and the manifold to the emergency flare
may be so great that the cost of connecting the vents to the existing manifold
may be appreciably more than the cost of connecting the vents to the flare mani-
fold during construction of a new plant.
-------
Table D-l. Emission Control Devices or Techniques Currently Used by Some Cumene Producers'
Control Devices or Techniques Used
Stream By Ashland Oil
Emission Source Designation Company
Propane recovery A Vent to propane gas
system vent manifold
Benzene recovery A2 Vent to plant flare
system vent
Cumene distillation A Vent to plant flare
j
system vent
Benzene azeotrope A^
drying column
Catalyst mix tank A
scrubber
Wash-decanter system A^
Benzene recovery A
i ^
column
_••
Cumene distillation A
system
D.I.P.B. stripping A&
system
Fugit ive "* Unknown
Storage and handling Unknown
Secondary Small wastewater
(syrup) stream
with no oil layer?
sent to plant sewer
From refs 1 — 6.
bCxcess fuel gas over manifold capacity diverted to plant flara-
By Monsanto Chemical
By Gulf Oil Company Company By Shell Oil Company
Solid Phosphoric Acid Catalyst Process (Fig. III-l)
Vent to propane gas Vent to propane gas Vent to fuel gas
mani fold manifold mani f o Id
Vent to plant flare Vent to atmosphere Vent to fuel gas
through vent conden- manifold
ser
Vent to atmosphere Vent to atmosphere Vent to fuel gas
manifold
Aluminum Chloride Catalyst Process (Fig. III-2)
>,
Auxiliary Sources
Unknown Unknown Unknown
Unknown Floating roof on Unknown
benzene storage
tank
Waste water decanted Wastewater sent to Wastewater decanted
to remove oil general plant to remove oil
layer; then sent to chemical sewer layer; then sent to
plant biooxidation line for deep-well plant biooxidation
system injection system
-'*
By Sun Petroleum
Products Company
Vent to propane gas
manifold
Vent to plant flare
Vent to atmosphere
Unlcnown
Unknown
-
Tars steam-stripped
to flare; waste-
water decanted to
remove oil layer
and then sent to
plant biooxidation
system
By Union Carbide
Corporation
Through vent header
and collection pot
to atmosphere
Vent to atmosphere |
to
Vent through degas-
ser and ga.s wash
system to propane
gas manifold
Through vent header
and collection pot
to atmosphere0
Through vent header
and collection pot
to atmosphere
Through vent header
and collection pot
to atmosphere0
Unknown
Unknown
Wastewater decanted
layer; then sent to
plant wastewat*.r
treatment system
distillation column* operated under pressure with hiyh-p^sure shutdown Controls) manufacturer claims no venting of organics to vent header and collection pot under normal operating
conditions.
-------
D-3
C. REFERENCES*
1. J. R. Kampfhenkel, letter dated Sept, 12, 1978, to EPA from Sun Petroleum Products
Co., Corpus Christi, TX, in response to EPA's request for information on the
cumene process.
2. M. P. Zanotti, letter dated Sept. 19, 1978, to EPA from Gulf Oil Co., Port Arthur,
TX, in response to EPA's request for information on the cumene process.
3. Oliver J. Zandona, letter dated Sept. 25, 1978, to EPA from Ashland Petroleum
Co., Ashland, KY, in response to EPA's request for information on the cumene
process.
4. Michael A. Pierle, letter dated Oct. 23, 1978, to EPA from Monsanto Chemical
Co., St. Louis, MO, in response to EPA's request for information on the cumene
process.
5. Attachment II, Information on the Cumene Process, from Shell Oil Co., Deer Park,
TX, in response to EPA's request for information on the cumene process.
6. F. D. Bess, letter dated Sept. 21, 1978, to EPA from Union Carbide Corp., South
Charleston, WV, in response to EPA's request for information on the cumene 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.
-------
4-i
REPORT 4
TOLUENE DIISOCYANATE
David M. Pitts
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
March 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.
D41H
-------
4-iii
CONTENTS OF REPORT 4
Page
I. ABBREVIATIONS AND CONVERSION FACTORS 1-1
II. INDUSTRY DESCRIPTION II-l
A. Toluene Diisocyanate II-l
B. TDI Usage and Growth II-l
C. Domestic Producers II-2
D. References II-4
III. PROCESS DESCRIPTION III-l
A. Introduction III-l
B. Typical Process for the Production of TDI III-l
C. Process Variations III-5
D. Other Processes III-5
E. References III-7
IV. EMISSIONS IV-1
A. Typical Plant IV-1
B. Process Sources and Emissions IV-1
C. References IV-5
V. APPLICABLE CONTROL SYSTEMS V-l
A. Process Emission Controls for Typical Plants V-l
B. Industry Emissions V-l
C. Assessment V-3
D. References V-4
APPENDIX OF REPORT 4
Page
A. PHYSICAL PROPERTIES OF TOLUENE DIISOCYANATE A-l
-------
4-v
TABLES OF REPORT 4
Number
II-l
IV-1
IV-2
V-l
A-l
A-2
TDI Producers, Locations, and Capacities
Summary of Uncontrolled Process Emissions from Typical TDI
Process Plant
Estimated Typical Composition of Gas from the H2S04 Concentration
Unit (Vent B)
VOC Emissions from Controlled Process Sources in Typical TDI Plant
Physical Properties of 2,4-Toluene Diisocyanate
Physical Properties for Phosgene
II-3
IV-3
IV-3
V-2
A-l
A-2
FIGURES OF REPORT 4
Number
III-l Process Flow Diagram for Uncontrolled Integrated TDI Plant
Page
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)
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
Symbol
tera
giga
mega
kilo
milli
micro
Multiplication
Factor
1012
109
106
103
io"3
io"6
Example
1 Tg = 1 X IO12 grams
1 Gg = 1 X IO9 grams
1 Mg = 1 X IO6 grams
1 km = 1 X IO3 meters
1 mV = 1 X IO"3 volt
1 pg = 1 X IO"6 gram
-------
II-l
II. INDUSTRY DESCRIPTION
A. TOLUENE DIISOCYANATE
Toluene diisocyanate (TDI) production was selected for study because preliminary
estimates indicated that emissions of volatile organic compounds (VOC) and the
potential toxicity of the chlorinated hydrocarbon raw materials were relatively
high.
TDI is the most important diisocyanate for the production of polyurethane materials.
The bulk of commercially used TDI is a mixture of 80 parts of the 2,4-isomer and
20 parts of the 2,6-isomer. Pertinent physical properties of TDI are given in
Appendix A. A 65:35 mixture of the 2,4- and 2,6-TDI isomers is also available
commerically, as is the pure 2,4-isomer. They are not, however, widely used.
TDI is produced by the phosgenation of toluene diamine, which is manufactured by
the reduction of dinitrotoluene , which in turn is produced by the nitration of
toluene. Either nitration-grade toluene or highly refined toluene (99.95+%) is
used as the basic feed stock by most TDI manufacturers.
B. TDI USAGE AND GROWTH
The total domestic consumption of TDI in 1977 was 265 Gg, with the following
breakdown in usage: foams, 185 Gg; coatings, 12.7 Gg,- elastomers 5.9 Gg; other
uses, 3.2 Gg; exports, 58.2 Gg. The total consumption of TDI in 1982 is estimated
to be 300 to 322 Gg, which represents an estimated annual growth rate of 2.6 to
Overall demand for flexible foams is expected to increase only modestly, with
the major growth in uses for bedding and underpadding. The demand for TDI for
uses in rigid foams for insulation in refrigerators and freezers is not expected
to grow because of the increasing use of polymeric isocyanates in this applica-
tion.
It is a matter of speculation as to whether the use of fluorocarbons in flexible
foam production will be banned and what the effects of such a ban would be on
the flexible foam industry and therefore on the demand for TDI. More TDI might
be required if fluorocarbons are not used in the foam manufacture, but this more
expensive foam may have a decreased market demand.
-------
II-2
The consumption of TDI for use in commercial and industrial coating systems is
projected to grow at a rate of 5—7% per year.
The use of TDI for elastomers and similar products is projected to grow at an
annual rate of 3 to 5%. Other uses include foundry core binders, fabric coatings,
adhesives and sealants, injection-molding resins, millable gums, and fibers.
Exports of TDI are not expected to increase above 1977 levels and may even de-
cline slightly as output of the large new Bayer plant at Brunebuettel, Federal
Republic of Germany, continues to be used in export markets.1
DOMESTIC PRODUCERS
There are eight major producers of TDI in the United States at ten plants.
Table II-l lists the producers, plant locations, and overall annual capacities
as of January 1978 for each company.1 In the latter part of 1978 the 25-Gg/yr
Union Carbide facility at Institute, West Virginia, was shut down,2 making the
total TDI annual production capacity at 340.5 Gg at the end of 1978. Normally
plants operate at 80 to 85% of nameplate capacity and additional capacity may
be required before 1982.
-------
II-3
Table II-l. TDI Producers, Locations, and Capacities
Company
Plant Location
Annual
Capacity
Mobay Chemical Corp.
Olin Corp.
BASF Wyandotte
Dow Chemical USA
Allied Chemical Corp.
Du Pont
Union Carbide
Rubicon Chemicals
Total
Cedar Bayou, TX
New Martinsville, WV
Ashtabula, OH
Lake Charles, LA
Giesmar, LA
Freeport, TX
Moundsville, WV
Deepwater Point, NJ
Institute
Geismar, LA
See ref 1.
bThis plant was closed in the latter part of 1978.
-------
II-4
D. REFERENCES*
1- H. E. Frey and Andrew J. Wolfe, "Diisocyanates and Polyisocyanates,"
pp. 666.5021A—666.5023B in Chemical Economics Handbook, Stanford Research
Institute, Menlo Park, CA (September 1978).
2. Telephone conversation, March 1979, between John Bresland, Allied Chemical,
and David M. Pitts, IT Enviroscience, Inc.
3. "Chemical Profile on TDI," in Chemical Marketing Reporter, Feb. 14, 1977.
*When a reference number is used at the end of a paragraph or on a heading,
it usually refers to the entire paragraph or material under the heading.
When, however, an additional reference is required for only a certain portion
of the paragraph or captioned material, the earlier reference number may not
apply to that particular portion.
-------
III-l
HI. PROCESS DESCRIPTION
A. INTRODUCTION
The manufacture of commercial toluene diisocyanate is based on the phosgenation
of primary amines. As stated previously most commercial TDI plants are integrated
with the production of the intermediates dinitrotoluene (DNT), toluene diamine (TDA)
j u 1—3
and phosgene.
B. TYPICAL PROCESS FOR THE PRODUCTION OF TDI
TDI is produced by the following chemical reactions:
Reaction 1:
+ 2HNO,
(toluene) (nitric
acid)
Reaction 2:
+ 6H
N°2
(d i u i t i o-
toluene)
(hydio-
gen)
Reaction 3:
CH.
H2S°4
(sulfuric
acid)
Catalyst
+ 2COC1,
Heat
NH2
(toluene
diamine)
(phosgene)
(2,4-dinitro-
toluene)
(water)
NH2
( 2,4-toluene
diamine)
NCO
NCO
(2,4-toluene
diisocyanate)
(water)
+ 4HC1
(hydrogen
chloride)
-------
III-2
The nitration product (Reaction 1) typically contains 80% 2,4-dinitrotoluene
isomers and 20% 2,6-dinitrotoluene isomer. Other isomers (2,3- and 3,4-dinitro-
toluene) and some unreacted toluene and nitrotoluene may be present in small
amounts. To simplify presentation the formula is shown as the 2,4-isomer only.
The phosgenation reaction (Reaction 3) is carried out using either monochloro-
or o-dichlorobenzene as a solvent. Approximately 0.7 Ib of toluene and 1.3 Ib
of phosgene are consumed for each pound of distilled 80:20 TDI produced. Hydro-
chloric acid is the only useful by-product produced, about 0.8 Ib per pound of
TDI.3
The typical TDI plant operates continuously and is integrated with the production
of DNT and TDA. An integrated facility may use natural gas and chlorine as raw
materials and make its own hydrogen and phosgene for use in the reduction and
phosgenation reactions respectively. — This report, however, does not include
hydrogen and phosgene production as part of the typical process.
The process flow diagram shown in Fig. III-l represents a typical continuous
process for the production of TDI using toluene, nitric acid, hydrogen, and
phosgene as raw materials. '
As indicated by Fig. III-l, the first step of the TDI process is nitration.
Nitration-grade toluene (stream 1) is reacted with nitric acid (stream 2) to
form DNT (stream 3). The reaction is carried out at ~49 to 66°C in cooled
reactors, which vent inert gases (stream C) and some VOC through a water scrubber.
The reaction is catalyzed by sulfuric acid. The spent sulfuric acid (-70%) is
phase separated from the reaction mixture and concentrated to ~93% in a direct-
contact evaporator, which uses the combustion gases from a natural gas burner.
The concentrated H2SO solution is recycled to the reactor. The vent from the
sulfuric acid concentrator (stream B) represents a potential VOC emission.
The DNT from the nitration reactor is washed in a wash tank and then reacted
with hydrogen (stream 4) in catalytic reduction reactors to form crude TDA
(stream 5). Excess hydrogen is taken overhead from the reactors, along with
some water of reaction. The water of reaction is removed from the hydrogen and
1,2
the hydrogen is recycled to the reactors.
-------
TOI
PUK.\F>CAT>OkJ
DtSTIU-ATlOU
TO
LAKJOFIU-
HCI
Fig. III-l. Process Flow Diagram for Uncontrolled Integrated TDI Plant
-------
III-4
The solid catalyst (palladium on carbon) is separated from the crude IDA in a
filter that is vented to the atmosphere (stream D). The vent represents a poten-
124
tial VOC emission. The filtered catalyst is recycled to the reduction reactors. ' '
The filtered TDA (stream 6) is dried by distillation. The dried TDA (stream 7)
is sent to vacuum distillation columns to remove lights, which are condensed and
124
burned in a liquid incinerator. ' ' The vacuum jet associated with this distilla-
tion is normally vented through a condenser and represents a potential VOC emission
2
(stream E).
The purified TDA (stream 8) is reacted with phosgene (stream 9) in the presence
of o-dichlorobenzene solvent (stream 10) to form crude TDI (stream 11). Phosgene
is condensed out of the by-product HCl, which goes overhead from the reactor. The
condensed phosgene is recycled to the reactor. The HCl that goes overhead from the
condenser (stream 12) may contain trace amounts of phosgene and is therefore sent
to the phosgene absorber. The crude TDI mixture from the phosgenation reactor
is sent to a distillation column for removal of phosgene. The phosgene overhead
(stream 13) from this distillation is combined with the HCl and trace-phosgene
stream (stream 12) from the reactor condenser and sent to a column that absorbs
phosgene with the dichlorobenzene solvent (stream 14). The solvent is then stripped
of phosgene in a distillation column and recycled to the absorber. The phosgene
is condensed and recycled to the phosgenation reactor. The HCl overhead from the
phosgene absorber and from the stripper condenser is absorbed with water in the
HCl absorber. Aqueous HCl is sent to by-product storage from the bottom of the
HCl absorber.1'2'4—7
The TDI-dichlorobenzene solvent mixture (stream 15) from the phosgene removal
distillation column is sent to a vacuum distillation column to recover the dichloro-
benzene solvent overhead, which is recycled to the phosgenation reactor. The
crude TDI (stream 16) from the bottom of the solvent recovery distillation column
is vaporized by vacuum flash distillation to separate TDI from any polymeric
isocyanates that might have been formed. The TDI taken overhead from the flash
is condensed (stream 17) and sent to a vacuum distillation column that takes
purified TDI product overhead, which is condensed (stream 18) and sent to product
storage. The bottoms from the TDI purification distillation are recycled to the
TDI vaporizer (flash distillation). The vacuum jet condensates from the solvent
-------
III-5
recovery distillation, from the flash distillation, and from the TDI purification
distillation are sent to wastewater treatment. The bottoms from the TDI vaporizer
(stream 22) are sent to a vacuum distillation column, which separates the polymeric
isocyanate residue from any comparatively low boiling compounds that might be
contained in the residue. The residue from the bottom of this separation column
is sent to landfill. The vacuum jet condensate from this distillation is also
124 7
sent to wastewater treatment. ' ' —
The residue separation vacuum jet vent (G) and the vacuum jet vents (F) asso-
ciated with the solvent recovery distillation, the TDI flash distillation, and
the TDI purification distillation and the HCl absorber vent (H) represent
247
potential sources of VOC emissions. ' —
C. PROCESS VARIATIONS
The available data indicate the potential for significant process variations to
exist among the different manufacturers with respect to the type of equipment
used and the sequence of operations for a given process step. Major process
differences reflect differences in raw materials. In some cases dinitrotoluene
is purchased, obviating the requirement for toluene nitration and thus elimi-
nating the H~SO concentration unit. At least one manufacturer (Olin) pur-
chases toluene diamine, thus eliminating the TDA reaction step. It is known
that at least one manufacturer (Allied) makes phosgene as part of the integrated
TDI facility.
Very limited data indicate differences in the TDI recovery, purification, and
residue recovery steps although no significant details are available. All TDI
recovery and purification steps, however, should require vacuum distillation
and/or evaporation steps, which would give rise to similar types of VOC
. . 1,2,4 7
emissions. —
D. OTHER PROCESSES
Mitsui Toatsu Chemicals, Inc., in Japan has developed a TDI process based on
dinitrotoluene carbonylation. In this process, dinitrotoluene is catalytically
carbonylated in the presence of an alcohol to give diurethane intermediate, which
is then thermally decomposed to TDI. The absence of a phosgenation step is the
-------
III-6
principal difference between the Japanese process and the current commercial
process. Mitsui has announced plans to build a 50-Gg/yr TDI plant in Japan using
this process, to be completed in I960.3 This report, however, covers only the
present commerical process.
-------
III-7
E. REFERENCES*
1. Yen-Chen Yen, Isocyanates Supplement B, Report No. IB, Process Economics
Program, Stanford Research Institute, Menlo Park, CA (November 1973).
2. David M. Pitts, IT Enviroscience, Inc., Trip Report on Site Visit to Allied
Chemical Corp., Morristown, NJ, Mar. 15, 1978 (on file at EPA, ESED, Research
Triangle Park, NC).
3. H. E. Frey and Andrew J. Wolfe, "Diisocyanates and Polyisocyanates,"
pp. 666-5021A—666-5023B in Chemical Economics Handbook, Stanford Research
Institute, Menlo Park, CA (September 1978).
4. T. R. Kovacevich, BASF Wyandotte Corp., letter dated May 31, 1978, regarding
toluene diisocyanate process at the Geismar plant, in response to EPA's
request for information on emissions data from TOO production facilities.
5. Donald W. Smith, E. I. du Pont de Nemours & Co., letter dated May 17, 1978,
regarding toluene diisocyanate process at Chambers Works, in response to
EPA's request for information on emissions data from TOO production facilities.
6. Lee P. Hughes, Mobay Chemical Corp., letter dated May 3, 1978, regarding the
toluene diisocynate process at Cedar Bayou plant, in response to EPA's request
for information on emissions data from TOO production facilities.
*When a reference number is used at the end of a paragraph or on a heading, it
usually refers to the entire paragraph or material under the heading. When,
however, an additional reference is required for only a certain portion of the
paragraph or captioned material, the earlier reference number may not apply to
that particular portion.
-------
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, par-
ticipate 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. TYPICAL PLANT
The capacity of the typical integrated plant for the production of TDI developed
for this study is 45 Gg/yr, based on 8760* hr of production annually. Although
not an actual operating facility, the size of the plant is typical of most present
industrial operating units using the typical process described in Sect. III.
B. PROCESS SOURCES AND EMISSIONS
As indicated in Section III, there are nine potential sources of process emissions
(labeled B-H in Fig. III-l) in the manufacture of TDI by the typical process
considered in this report. Uncontrolled process emissions have been calculated
for the most part from estimated and measured data on controlled emissions and
estimated control efficiencies provided by the Allied Chemical Corporation and
from process and emission data from other sources. — These estimated uncon-
trolled process emissions are summarized in Table IV-1.
Storage and handling, fugitive, and secondary emissions are not considered in
the abbreviated report but they are covered for the entire synthetic organic
manufacturing industry by separate EPA documents. For convenience, sources of
^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
storage emissions are labeled A and potential sources of secondary emissions are
labeled S in Fig. III-l.
As indicated by Table IV-1, the most significant uncontrolled VOC emission from
TDI manufacturing (vent B in Table III-l) results from the H2S04 concentration
unit. This unit uses hot combustion gases to evaporate water from the spent H2S04
solution coming from the nitration reactors. The estimated uncontrolled composition
given in Table IV-2 for this vent stream was calculated from data on controlled
emissions and estimated control efficiencies provided by the Allied Chemical
Corporation.
Vent C represents the emissions from the nitration reactors and contains inert
gases (mostly air), SO , NO , and small amounts of nitroaromatic compounds.
£ X
Vent D represents the emissions from the TDA reactors via the catalyst separation
unit and contains air and small amounts of organic amines. '
Vent E represents the emissions from the vacuum jet associated with the distilla-
tion to remove low-boiling organic amines from the TDA. The air that is discharged
through the vacuum-jet hot well carries some of these light organic amines with
it.1'3
Vents F represent the emissions from the vacuum-jet hot wells associated with the
dichlorobenzene solvent recovery distillation, the TDI flash distillation, and
the TDI purification distillation. These vents taken together represent the
second most significant uncontrolled VOC emission from the typical TDI plant
according to estimated data from Allied. No detailed composition data on these
streams are available although it has been estimated, based on other industry data,
that the major VOC component of the combined uncontrolled emission is phosgene
(~99%) and that the remainder of the VOC is dichlorobenzene.
Vent G represents the emissions from the vacuum-jet hot well associated with the
residue separation distillation. This emission contains mostly air and trace
amounts of chlorinated hydrocarbons, which can be formed from the phosgenation
1,3 6
reaction. —
-------
IV-3
Table IV-1. Summary of Uncontrolled Process Emissions from
Typical TDI Process Plant
C/-\i i v- CG
H SO concentrator
2 4
Nitration reactor (s)
TDA -reaction via
Stream
Designation
(Fig.III-1)
B
C
D
VOC
Ratio
(g/kg)*
5.0
0.025
Emissions
Rate
(kg/hr)
25.90
0.13
0.0005 0.0026
catalyst filtration
TDA lights removal
distillation
Solvent recovery, flash,
and product purification
distillations
0.0033
4.6
-5
*g of VOC per kg of TDI produced.
0.017
23.8
-5
Residue separation
HC1 absorber
Total process emissions
G
H
1.1 X 10
4.6 X 10~?
9.63
5.7 X 10
2.4 X 10~6
49.8
Table IV-2. Estimated Typical Composition of
Gas from the H Concentration Unit (Vent B)
Component
Composition (wt
Combustion products
and HO vapor
so2
NO
H2S04
Nitroaromatics
Total
99.68
0.005
0.06
0.18
0.075
100
-------
IV-4
Vent H represents the emissions from the HCl absorber and contains small amounts of
phosgene in the CO- and water vapor discharged from the acid recovery system. ' —
It should be noted that phosgene represents a large percentage (~99%) of the
estimated uncontrolled VOC emissions associated with the solvent recovery and
TDI product distillations. Because of its toxicity, emissions of phosgene must
be controlled to extremely low levels.
-------
IV-5
C. REFERENCES*
1. David M. Pitts, IT Enviroscience, Inc., Trip Report on Site Visit to Allied Chemical
Corp., Morristown, NJ, Mar. 15, 1978 (on file at EPA, ESED, Research Triangle
Park, NC).
2. H. E. Frey and Andrew J. Wolfe, "Diisocyanates and Polyisocyanates,"
pp. 666-5021A—666-5023B in Chemical Economics Handbook, Stanford Research
Institute, Menlo Park, CA (September 1978).
3. T. R. Kovacevich, BASF Wyandotte Corp., letter dated May 31, 1978, regarding
toluene diisocyanate process at the Geismar plant, in response to EPA's
request for information on emissions data from TDI production facilities.
4. Donald W. Smith, E. I. du Pont de Nemours & Co., letter dated May 17, 1978,
regarding toluene diisocyanate process at Chembers Works, in response to
EPA's request for information of emissions data from TDI production facilities.
5. Lee P. Hughes, Mobay Chemical Corp., letter dated May 3, 1978, regarding toluene
diisocyanate process at Cedar Bayou plant, in response to EPA's request for
information on emissions data from TDI production facilities.
6. J. C. Ketchum, Union Carbide Corp., letter dated May 16, 1978, regarding toluene
diisocyanate process at the Institute plant, in response to EPA's request for
information on emissions data from TOO production facilities.
*When a reference number is used at the end of a paragraph or on a heading, it
usually refers to the entire paragraph or material under the heading. When,
however, an additional reference is required for only a certain portion of the
paragraph or captioned material, the earlier reference number may not apply to
that particular portion.
-------
V-l
V. APPLICABLE CONTROL SYSTEMS
A. PROCESS EMISSION CONTROLS FOR TYPICAL PLANTS
Table V-l shows the control devices, estimated VOC reduction efficiencies, and
resulting emissions for each of the vent streams shown in Fig. III-l and discussed
in Sect. IV. The control devices used and the estimated reduction efficiencies
represent nonconfidential data obtained from one company. Based on limited
O C,
information from other sources, — the data given in Table V-l are felt to be
representative of the TDI industry in general. The cost and cost effectiveness
for these applications have not been determined.
With respect to the sulfuric acid concentrator and the nitration reactor vents
(B and C) the primary function of the wet scrubber control devices is to remove
H»SO . These devices, however, have been estimated to be ~60 to 80% efficient
for removing VOC because of the nature of the nitro-organic compounds being scrub-
bed.1
With respect to the necessary control of phosgene emissions from vents F and H,
all data indicate the use of dilute caustic and/or water (hydrolysis) scrubbing.
The caustic scrubber or hydrolysis column is normally estimated to have >99%
removal efficiency for phosgene.1—6 In the case of TDI manufacture it is esti-
mated that >98% of the other relatively high boiling VOC would be removed by
condensation in the scrubbing device.1 (Note: virtually 100% control of phos-
gene emissions may be required in order to protect workers from toxic concen-
trations in the vicinity of these vents.)
B. INDUSTRY EMISSIONS
From the data reported in Table V-l the overall process emission ratio has been
calculated to be 2.056 g of VOC per kg of TDI produced for the typical plant.
This is believed to be typical of the TDI plants operating today. Storage and
handling, secondary, and fugitive emissions ara not included in the ratio. Compari-
son of the data in Table IV-1 with those in Table V-l indicates that the TDI
industry is -78.6% controlled overall with respect to process emissions of VOC
and that the major process emission results from the H2S04 concentrator. From
the data in Table V-l and the estimated 1978 total TDI production of 280 Gg, the
process emissions of VOC from the TDI industry have been estimated to be 0.576 Gg,
-------
Table V-l. VOC Emissions from Controlled Process Sources in Typical TDI Plant
Estimated
VOC _ .
^ . . Emissions
Source
H SO concentrator
Nitration reactors
btream IMIH^-LOU
Designation Control Device Reduction Ratio
(Fig.III-1) or Technique (%) (g/kg)*
B Wet venturi scrubber 60 2.0
for removal of
H2S°4
C Water scrubber (spray 60 0.01
tower) for removal of
V°4
Rate
(kg/hr)
10.36
0.052
TDA reaction via
catalyst filtration
TDA lights removal
distillation
Solvent recovery, flash,
and product purification
distillations
Residue separation
HC1 absorber
H
Wet venturi scrubber for 80
particulate removal
Water-cooled surface con- 97
densers for removal of
organic amines
Dilute caustic scrubber or ^99
hydrolysis column for
phosgene removal
Water-cooled surface con- 97
densers
Packed water scrubber 98
(hydrolysis column for
control of trace phosgene)
1.0 X 10 4 5.1 X 10~4
0.046
9.2 X 10
-9
0.24
-7 -6
3.3 X 10 1.71 X 10
4.8 X 10
*g of VOC per kg of TDI produced.
-------
V-3
not including secondary, fugitive, or storage and handling emissions. When danger
exists for operator exposure to highly toxic phosgene, extra precautions are
required. Therefore fugitive emissions are expected to be significantly below
the normal VOC fugitive emission rate for the synthetic organic chemicals manu-
facturing industry.
C. ASSESSMENT
As indicated in Tables IV-1 and V-l, the major emissions from the TDI process
result from the H2SO concentration step. A separate EPA report specifically
covers the emissions resulting from H2SOA concentration units.
-------
V-4
C. REFERENCES*
1. David M. Pitts, IT Enviroscience, Inc., Trip Report on Site Visit to Allied Chemical
Corp., Morristown, NJ, Mar. 15, 1978 (on file at EPA, ESED, Research Triangle
Park, NC).
2. H. E. Frey and Andrew J. Wolfe, "Diisocyanates and Polyisocyanates,"
pp. 666-5021A—666-5023B in Chemical Economics Handbook, Stanford Research
Institute, Menlo Park, CA (September 1978).
3. T. R. Kovacevich, BASF Wayndotte Corp., letter dated May 31, 1978, regarding
toluene diisocyanate process at the Geismar plant, in response to EPA's
request for information of emissions data from TDI production facilities.
4. Donald W. Smith, E. I. du Pont de Nemours & Co., letter dated May 17, 1978,
regarding toluene diisocyanate process at Chember Works, in response to
EPA's request for information on emissions data from TOO production facilities.
5. Lee P. Hughes, Mobay Chemical Corp., letter dated May 3, 1978, regarding toluene
diisocyanate process at the Cedar Bayou plant, in response to EPA's request
for information on emissions data from TDI production facilities.
6. J. C. Ketchum, Union Carbide Corp., letter dated May 16, 1978, regarding toluene
diisocyanate process at the Institute plant, in response to EPA's request for
information on emissions data from TDI production facilities.
7. J. A. Key, IT Enviroscience, Inc., Waste Sulfuric Acid Treatment for Acid Recovery
(EPA, ESED, Research Triangle Park, NC).
*When a reference number is used at the end of a paragraph or on a heading,
it usually refers to the entire paragraph or material under the heading.
When, however, an additional reference is required for only a certain portion
of the paragraph or captioned material, the earlier reference number may not
apply to that particular portion.
-------
A-l
APPENDIX A
Table A-l. Physical Properties of 2,4-Toluene Diisocyanate^
Synonyms TDI, isocyanic acid, methyl phenylene ester
Molecular formula C H N O
9622
Molecular weight 174.16
Physical state Liquid
Specific gravity 1.22 at 20°C/4°C
Vapor pressure <0.01 mm Hg at 20°C
Boiling point 238.3°C
Melting point 19.5 - 21.5°C
Water solubility Reacts with HO to produce CO
*From: J. Dorigan et al., "Toluene Diisocyanate," p. AIV-214 in
Scoring of Organic Air Pollutants. Chemistry, Production arid_
Toxicity of Selected Synthetic Organic Chemicals (Chemicals A—C) ,
Rev. 1, Appendix IV, MTR-7248, MITRE Corp., McLean, VA (September
1976).
-------
A-2
Table A-2. Physical Properties for Phosgene*
Synonyms
Molecular formula
Molecular weight
Physical state
Specific gravity
Vapor pressure
Boiling point
Melting point
Water solubility
Safety hazard
Carbonoxychloride, carbonylchloride, CG
ccl2o
98.92
Gas or volatile liquid
1.392 at 19°C/4°C
1428 mm Hg at 25 °C
7.56°C =
Decomposes in H^O
Disaster hazard; highly dangerous; toxic fumes
*From: J. Dorigan e_t al. , "Phosgene, p. AIV-42 in Scoring of Organic Air
Pollutants. Chemistry, Production and Toxicity of Selected Synthetic Organic
Chemicals (Chemicals O-Z, Rev. 1, Appendix IV, MTR-7248, MITRE Corp.) McLean,
VA (September, 1976).
-------
5-i
REPORT 5
CRUDE TEREPHTHALIC ACID, DIMETHYL TEREPHTHALATE, AND
PURIFIED THERPHTHALIC ACID
S. W. Dylewski
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
January 1981
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.
D54K
-------
5-iii
CONTENTS OF REPORT 5
I.
II.
ABBREVIATIONS AND CONVERSION FACTORS
III
INDUSTRY DESCRIPTION
A. Reason for Selection
B. Usage and Growth
C. Domestic Producers
D. References
PROCESS DESCRIPTION
A. Introduction
B. Air-Oxidation Process for C-TPA
C. Process Variation
D. DMT by Esterification of C-TPA
E. Purified TPA from C-TPA
F. References
IV. EMISSIONS
A. Crude Terephthalic Acid Process
B. C-TPA Process Variation
C. DMT by Esterification of C-TPA
D. Process Variation
E. Purified TPA from C-TPA
F. References
V. APPLICABLE CONTROL SYSTEMS
A. Crude Terephthalic Acid Process
B. C-TPA Process Variation
C. Current Emission Control Used in C-TPA Production
D. DMT by Esterification of C-TPA
E. Current Emission Control Used in DMT Production
F. DMT Process Variation
G. Purified TPA from C-TPA
H. References
Page
1-1
II-l
II-l
II-l
II-3
II-7
III-l
III-l
III-l
III-5
III-5
III-8
III-ll
IV-1
IV-1
IV-7
IV-7
IV-9
IV-9
IV-11
V-l
V-l
V-4
V-4
V-4
V-7
V-7
V-7
V-8
-------
5-v
CONTENTS (Continued)
Page
VI. IMPACT ANALYSIS VI-1
A. Environmental and Energy Impacts VI-1
B. Control Cost Impact VI-5
C. References VI-11
VII. SUMMARY VII-1
APPENDICES OF REPORT 5
A. PHYSICAL PROPERTIES A~1
B. AIR-DISPERSION PARAMETERS B_!
C. FUGITIVE-EMISSION FACTORS C-i
D. COST ESTIMATING PROCEDURE D_!
E. EXISTING PLANT CONSIDERATIONS E_!
-------
5-vii
TABLES OF REPORT 5
Number
II-1 DMT and P-TPA Usage and Growth II-2
II-2 DMT and P-TPA Capacity II-4
IV-1 Uncontrolled Emissions from C-TPA Model Plant IV-3
IV-2 Composition of Reactor Vent Gas IV-4
IV-3 C-TPA Model Plant Storage Tank Data IV-6
IV-4 Uncontrolled Emissions from Typical DMT Plant IV-8
IV-5 Emissions from P-TPA Typical Plant IV-10
V-l Controlled Emissions from C-TPA Model Plant V-3
V-2 Controlled Emissions from Typical DMT Plant V-5
VI-1 Environmental Impact of Controlled C-TPA Model Plant VI-2
VI-2 Environmental Impact of Controlled DMT Typical Plant VI-4
VI-3 Factors Used in Computing Annual Costs VI-6
VI-4 Cost Effectveness of Carbon Adsorption in C-TPA Model Plant VI-10
VII-1 Emission Summary for DMT Typical Plant VII-2
VII-2 Emission Summary for C-TPA Model Plant VII-3
-------
5-ix
FIGURES OF REPORT 5
Number Pa9e
II-l Locations of Plants Manufacturing DMT and P-TPA H-5
III-l Crude Terephthalic Acid Process III-2
III-2 DMT by Esterification of C-TPA III-6
III-3 P-TPA by Purification of C-TPA III-9
VI-1 Capital Cost of Carbon Adsorption VI~7
VI-2 Net Annual Cost of Carbon Adsorption VI~9
-------
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
Symbol
tera
giga
mega
kilo
milli
micro
Multiplication
Factor
1012
10°
106
103
10
Example
12
1 Tg = 1 X 10** grams
1 Gg = 1 X 109 grams
1 Mg = 1 X 106 grams
1 km = 1 X 103 meters
1 mV = 1 X 10"3 volt
1 |jg = 1 X 10"6 gram
-------
II-l
II. INDUSTRY DESCRIPTION
A. REASON FOR SELECTION
Production of terephthalic acid (TPA) and dimethyl terephthalate (DMT) was
selected for study because of the large amounts produced and because of the
significant emissions of VOC projected from their manufacture. The DMT study
has been abbreviated because industry data indicate the emissions from the DMT
process to be much lower than were previously estimated. The future DMT processes
are expected to be based on esterification of crude TPA, which is the process
generating the lowest emissions. Appendix A lists pertinent physical properties
of the chemicals of significance that are involved.
B. USAGE AND GROWTH
Dimethyl terephthalate and purified terephthalic acid (P-TPA) are alternative
raw materials for the manufacture of polyester products, where 1.17 g of DMT is
equivalent to 1 g of P-TPA. When DMT is used, methanol is recovered and recycled
to the DMT process. Table II-l shows the end uses of DMT and P-TPA, the percentage
of consumption by each end use, and the growth rate for each use from 1976 to
1981.l
The predominant use is in the manufacture of polyethylene terephthalate (poly-
ester) fibers, with small percentages going to polyester films, polybutylene
terephthalate resins, exports, and other uses. Polyethylene terephthalate
(PET) barrier resins for carbonated beverage bottles accounted for about 0.2%
in 1976; however, it is the fastest growing end use and is projected to reach
3.5 to 4% of the total demand in 1981.1
The 1978 domestic annual capacity is reported to be 1997 Gg of DMT and 1314 Gg
of P-TPA. Production was reported to be about 61% of capacity during 1978.
Based on a projected growth rate of 6.5 to 9.0% for both products the capacity
utilization will reach 78 to 86% by 1982.1—8
P-TPA capacity was recently expanded by 53% when Amoco Chemicals dedicated its
new plant in Cooper River, SC, in late 1978. However, there have been no recent
increases of DMT production capacity. There actually may be some shifting in
1 9
capacity from DMT to P-TPA. '
-------
II-2
Table II-l. Dimethyl Terephthalate and Purified Terephthatic Acid
Usage and Growth*
End Use
Polyester fibers
Polyester films
Polybutylene terephthalate resins
PET barier resins
Miscellaneous
Exports and Inventory Building
ConsumDtion
DMT
84.2
8.3
1.4
0.2
0.4
5.5
(%) for 1976
P-TPA
89.2
3.8
0
0.2
1.2
5.6
Average growth
Average Growth
1976 — 81
(%/yr)
5.5 — 7.5
8 — 10
14.5 — 19.0
84 — 92
4 . 5 — 8 . 5
Not available
rate 6 . 5 — 9 . 0
*See ref 1.
-------
II-3
C. DOMESTIC PRODUCERS
As of 1978 there were three active domestic producers of DMT in five locations
and one domestic producer of P-TPA in two locations. Table II-2 lists the pro-
1 12
ducers, locations, and capacities. Figure II-l shows the plant locations.
Late in the writing of this report the Chemical Marketing Reporter published an
estimated capacity of domestic producers that is not significantly different
from that in Table 11-2.
I. American Hoechst
1 4
The plant is based on the Hercules/Imhausen-Witten (Hercules) process ' for
DMT, which proceeds from p_-xylene via a methyl p_-toluate intermediate rather
than through a TPA intermediate. The facility was shut down in mid-1978 and
may not be reopened.5 Through a lease arrangement Hereofina is using the plant
2
facilities and supplying Hoechst with DMT.
2. E. I. du Pont de Nemours
Both operating plants produce DMT by air oxidation of p_-xylene to crude TPA
(C-TPA) by the Amoco process, followed by esterification of C-TPA to DMT by the
Tennessee Eastman process.4 The DMT produced is used captively in fiber produc-
tion. Following expansion by the addition of a second train at its Wilmington, NC,
location, the company curtailed its formerly large purchases of DMT. A
126-Gg/yr DMT plant at Gibbstown, NJ, was shut down indefinitely in 1974; the
1 2
plant has been sold and will be dismantled.
3. Eastman Kodak (Tennessee Eastman Division and Carolina Eastman Division)
Both plants use Eastman processes to produce C-TPA and DMT and use the DMT
-i o (L
captively in their fibers and films plants.
4. Hereofina (Joint Venture of Hercules and American Petrofina)
The Hercules process is used to produce DMT for the merchant market. Some TPA
is produced by hydrolysis of DMT.1 Construction was halted in 1975 on a DMT
plant in Eastover, SC, which was scheduled to have a capacity of 363 Gg/yr.
. 1,9
This plant is being redesigned and may be converted to TPA production.
Hercules is also experimenting with a new process for production of TPA and is
modifying part of its Wilmington, NC, plant to include the new technology. A
68-Gg/yr DMT plant at Burlington, NJ, was shut down indefinitely in 1974. The
-------
II-4
Table II-2. Dimethyl Terephthalate and
Purified Terephthalic Acid Capacity
Plant
Capacity
as of 1978
(Gg/yr)
DMT
P-TPA
American Hoechst Corp., Spartanburg, SC
a
E. I. du Pont de Nemours and Co., Inc.
Cape Fear (Wilmington), NC
Old Hickory, TN
Eastman Kodak Co.
Columbia, SC
Kingsport, TN
Q
Hereofina, Wilmington, NC
c
Standard Oil (Indiana) - Amoco Chemicals"
Cooper River, SC
Decatur, AL
73
567
250°
226
281C
600
a,b
454
860J
1997 1314
See refs 1,4.
DShut down in mid-1978; see refs 2,5.
'See ref 2.
See ref 6.
"See refs 1, 9, and 11.
See ref 7.
3See ref 12.
Started up late 1978; see ref 3.
"""See ref 8.
-------
II-5
! American Hoechst, Spartanburg, SC
2'. Du Pont, Cape Fear, NC
3. DU Pont, Old Hickory, TN
4 Eastman Kodak, Columbia, SC
5
6
7
Q'.
Eastman Kodak, Kingsport, TN
Hereof ina, Wilmington, NC
Standard Oil-Amoco, Cooper River, SC
Standard Oil-Amoco, Decatur, Al
T ations of Plants Manufacturing
Fig. n-1. Locations of Terephthalic Acid
Dimethyl Terephthalate
-------
II-6
plant has since been sold and will be dismantled. 1 Hercofina has a captive
supply of xylenes.
5. Standard Oil (Indiana) (Amoco Chemicals; subsidiary)
Crude TPA is produced by oxidation of g-xylene in an acetic acid medium in the
presence of a manganese acetate or cobalt acetate catalyst and an inorganic
bromide. ' C-TPA is purified to pure TPA (P-TPA) for the merchant market.1
Amoco recently increased the P-TPA capacity over 50% by dedicating a new 454-Gg/yr
plant at Cooper River, near Charleston, SC. Amoco is also a producer of raw
material p_-xylene. Amoco has shut down a 91-Gg/yr DMT plant at Decatur, AL,
and a 68-Gg/yr DMT plant at Joliet, IL. Also, a 45-Gg/yr TPA plant in Joliet,
IL, was converted to isophthalic acid production several years ago.
-------
II-7
D. REFERENCES*
1. J. L. Blackford, "Dimethyl Terephthalate and Terephthalic Acid," pp. 695.4021A—
695.4023H in Chemical Economics Handbook, Stanford Research Institute, Menlo
Park, CA (July 1977).
2. E. M. Klapproth, "Xylene Isomers," p. 300.7404K in Chemical Economics Handbook,
Stanford Research Institute, Menlo Park, CA (December 1978).
3. "Checkoff," Chemical and Engineering News 15(15) (Jan. 22, 1979).
4. D. F. Durocher et al., p. 4 in Screening Study to Determine Need for Standards
of Performance for New Sources of Dimethyl Terephthalate and Terephthalic Acid
Manufacturing, EPA Contract No. 68-02-1316, Task Order No. 18 (July 1976).
5. R. T. Monaghan, Hoechst Fibers Industries, letter dated Aug. 14, 1978, in
response to EPA request for information on emissions from DMT/TPA production
facilities.
6. J. C. Edwards, Tennessee Eastman Company, letter dated Aug. 31, 1978, in response
to EPA request for information on emissions from DMT/TPA production facilities.
7. S. W. Dylewski, IT Enviroscience, Inc., Trip Report for Visit to Hereofina,
Wilmington, NC, Nov. 17, 18, 1977 (on file at EPA, ESED, Research Triangle
Park, NC).
8. S. W. Dylewski, IT Enviroscience, Inc., Trip Report for Visit to Amoco Chemicals
Corporation, Decatur, AL, Oct. 31, Nov. 1, 1977 (on file at EPA, ESED, Research
Triangle Park, NC).
9. "Chementator," Chemical Engineering 84(15), 51 (1977).
10. S. W. Dylewski, IT Enviroscience, Inc., Trip Report for Visit to Carolina Eastman
Company, Columbia, SC, Dec. 6, 7, 1977 (on file at EPA, ESED, Research Triangle
Park, NC).
11. "CPI News Briefs," Chemical Engineering 86(8), 70 (1979).
12. H. M. Brennan, Amoco Chemicals Corp., letter dated Aug. 16, 1978, to D. J. Mangum,
EPA, Office of Air Quality Planning and Standards, Research Triangle Park, NC.
13. "Chemical Profile on DMT-TPA," p. 9 in Chemical Marketing Reporter (Apr. 30, 1979)
*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 DMT or TPA used to make polyester must be of very high purity. Crude
terephthalic acid (C-TPA) that was formerly made by nitric acid oxidation of
p--xylene contained impurities that were unacceptable to the polyester industry.
The methanol esterification process for dimethyl terephthalate (DMT) provides a
means of removing these impurities from C-TPA and produces a product of accep-
2
table quality.
C-TPA made by air oxidation of p_-xylene is of higher quality than that made by
nitric acid oxidation but still requires purification for use in polyester fibers.
This can be done by esterification with methanol,3 as discussed above, or by
1,2,4
hydrogenation and crystallization from water.
Another commercial route for producing DMT of polyester fiber quality is by air
oxidation of a mixture of E-xylene and methyl toluate to toluic acid and mono-
methyl terephthalate, respectively, and subsequent methanol esterification.
The methyl toluate that is formed in esterification is recycled to oxidation,
and the DMT is recovered and purified by distillation.
This report is primarily concerned with the air-oxidation process for C-TPA,
the methanol esterification process for DMT, and the hydrogenation and crystalli-
zation process for purified terephthalic acid (P-TPA). The process for oxidation
of a mixture of p_-xylene and methyl toluate as practiced by Hercofina is not
likely to be selected for new construction. The nitric acid oxidation process
for C-TPA is no longer practiced domestically1'6 and is not further considered
in this report.
B. AIR-OXIDATION PROCESS FOR C-TPA
The model continuous process for the manufacture of C-TPA is shown in Fig. III-l -
The oxidation and product recovery portion is essentially as is practiced by
Amoco Chemicals, whereas the recovery and recycle of acetic acid and recovery
of methyl acetate are essentially as practiced by Carolina Eastman.
-------
FU6,lT\VE
I UQU\D
StPAEATOR.
METHYL.
ACETATE
STORAGE
1
1 '
200'C ^-
/Cx
C-TPA, TO
PUEjFICA.TIOkl
•
j
J
£>
. _1
TT— ^
j.^ - —
-JcTp'
•S> 4^
\.
r-"1
*-
J
,
<<
^
•^
^
<^>
1
J
-*-
^
ME
THYU
ACETATE
STIU_
"-r:
..
/^
^*s
WAS1
rewATEp. Az
t)Tli_L.
1
/-TV Te\w4.TEJi
\^J W ^A -
12.ECOVER.ED
EOTRO
~
i
RESIDUE
STVUl-
I
NJ
WATER
Fig. III-l. Crude Terephthalic Acid Process
UQ.UID
YDROOJ
IUCIUE.RA.TOR.
WASTE WATE.E
-------
III-3
Chemistry
HOAC +
(acetic
acid
solvent)
CH3 \ / CH3
(p_-xylene)
cat
0 0
H sf~^ "
HO-C-C V-C-OH
(terephthalic acid)
(See footnote*)
2H2°
(water)
+ C02 + H20
(carbon (carbon (water)
monoxide) dioxide)
Products of partial oxidation of p_-xylene, such as E-toluic acid and p_-formyl
benzoic acid, are formed, with some of them appearing as impurities in TPA.
Methyl acetate is also formed in significant amounts in the reaction.
2. Oxidation of p_-Xylene
E-Xylene (stream 1), fresh acetic acid (stream 2), a catalyst system (stream 3),
such as manganese or cobalt acetate and sodium bromide,7 and recovered acetic
acid (stream 4) are combined to comprise the liquid stream entering the reactor
(stream 5). Air (stream 6), compressed to reaction pressure (about 2000 kPa),
is fed to the reactor. The temperature of the exothermic reaction is maintained
at about 200°C by controlling the pressure at which the reaction mixture is
permitted to boil and form the vapor stream (stream 7) leaving the reactor.
Inert gases, excess oxygen, CO, C02, and volatile organic compounds (VOC)
(stream 8) leave the gas/liquid separator and are sent to the high-pressure
absorber. This stream is scrubbed with water under pressure, resulting in a
gas stream (stream 9) with reduced VOC content. Part of the discharge from the
high-pressure absorber is dried and is used as a source of inert gas
-------
III-4
return of condensed VOC and water. The partially oxidized impurities are more
soluble in acetic acid and tend to remain in solution while TPA crystallizes
from the liquor. The inert gas that was dissolved and entrained in the liquid
under pressure is released when the pressure is relieved and is subsequently
vented to the atmosphere along with the contained VOC (vent B). The slurry
(stream 11) from the crystallizers is sent to solid-liquid separators, where
the TPA is recovered as a wet cake (stream 14). The mother liquor (stream 12)
from the solid-liquid separators is sent to the distillation section, while the
vent gas (stream 13) is discharged to the atmosphere (vent B).
4. Drying, Handling, and Storage
The wet cake (stream 14) from solid-liquid separation is sent to dryers, where
with the use of heat and IG the moisture, predominantly acetic acid, is removed,
leaving the product, C-TPA, as a dry flowable solid (stream 19).
The hot, VOC-laden IG is cooled to condense and recover VOC (stream 18). The
cooled IG (stream 16) is vented to the atmosphere (vent B). The condensate
(stream 18) is sent to the azeotrope still for recovery of acetic acid. IG is
used to convey the product (stream 19) to storage silos. The transporting gas
(stream 21) is vented from the silos to dust collectors (bag-type), where its
particulate loading is reduced. It is then discharged to the atmosphere (vent D).
The solids (S) from the bag filter can be forwarded to purification or be disposed
of by incineration.
5. Distillation and Recovery
The mother liquor (stream 12) from solid-liquid separation flows to the residue
still, where acetic acid, methyl acetate, and water are recovered overhead
(stream 26). The bottoms (stream L) from the still contain the products of
partial oxidation, tars, catalyst residue, and some acetic acid and are sent to
a liquid-waste incinerator for destruction. The overhead (stream 26) from the
still and the streams (25) from the high-pressure scrubber and the product dryer
are processed in the azeotrope still to remove water as an overhead stream ancl
produce a bottoms acetic acid stream (stream 4) essentially free of water.
n-Propyl acetate, used as an azeotroping agent to facilitate the separation,
enters the azeotrope still through stream 27. The vapors from the still con-
taining water, n-propyl acetate, and methyl acetate are condensed and decanted.
-------
III-5
The aqueous phase (stream 28) is forwarded to the wastewater still, whereas the
organic phase (stream 27), mainly n-propyl acetate, is returned to the azeotrope
still. The aqueous phase (stream 28) contains saturation amounts of n-propyl
acetate and methyl acetate, which are stripped from the aqueous phase in the
wastewater still. Part of the bottoms product is used as process water in absorp-
tion and the remainder (N) is sent to wastewater treatment. A purge stream of
the organic phase (stream 30) is sent to the methyl acetate still, where methyl
acetate and saturation amounts of water are recovered as an overhead product
(stream 31) and disposed of as a fuel (discharge M). n-Propyl acetate, obtained
as the bottoms product (stream 32), is returned to the azeotrope still. A small
amount of inert gas, which is used for blanketing and instrument purging, is
emitted to the atmosphere (vent C). Process losses of n-propyl acetate are
made up from storage (stream 33).
C. PROCESS VARIATION
In the model plant, acetic acid, used as a reaction solvent, is supplied as a
raw material to replace losses of acetic acid as oxidation products and to emissions.
A variation practiced by Carolina Eastman is the use of acetaldehyde as a source
of acetic acid for the replacements of losses. Carolina Eastman uses a somewhat
different catalyst system including bromine in the form of HBr, whereas others
use a bromine salt. Otherwise, the processes are very similar.
The process used by Hercofina is different from the model plant in the majority
of its processing steps. Air is used for oxidation of the £>-xylene as in the
model plant; however, in the Hercofina process the oxidation is conducted in an
excess of methyl toluate solvent, where methyl toluate is oxidized to mono-
methyl terephthalate. Monomethyl terephthalate, which is the main oxidation
product, is esterified subsequently to DMT; terephthalic acid is not recovered
as a product or intermediate in this process.
D. DMT BY ESTERIFICATION OF C-TPA
The purpose of the typical process as shown in Fig. III-2 is to convert the tereph-
thalic acid contained in C-TPA to a form that will permit its separation from
the impurities. This process is representative of current DMT technology.
Terephthalic acid is converted to the dimethyl ester and is separated by frac-
tional distillation.
-------
O
a
er
w
ft
n>
O
Jl)
rr
i—
O
3
O
l-h
O
•t)
"1
• ©
7^
3
R I
9-III
-------
III-7
1. Chemistry
00 0
>
HO-C- ) -C-OH + 2CH OH - > H.C-O-C- f -C-CH0 + 2H,0
\ — / J •j \ / 3 2
(terephthalic acid) (methanol) (dimethylterephthalate) (water)
2. Esterification
C-TPA (stream 1) is sent by mechanically assisted gravity feed from storage
silos to slurry mix tanks, where it is mixed with methanol (stream 2) to form a
slurry (stream 3) that is adequate for pumping to the continuous reactor. The
esterification reaction consumes methanol and terephthalic acid and forms
dimethyl terephthalate and water. A liquid purge stream (stream 4) is drawn
from the reactor and is sent to the sludge evaporator and stripper for the removal
and disposal of nonvolatile waste (discharge N) . The volatile portion (stream 5)
of the purge stream is returned to the process.
3 . Methanol Recovery Still
The liquid stream (stream 6) from the reactor contains excess methanol; water,
dimethyl ether, and other low boilers formed in the reactor,- methyl p_-toluate
and methyl £-formyl benzoate that were formed in the reactor from impurities in
C-TPA; and dimethyl terephthalate. Water formed in esterification is removed
as the o-xylene — water azeotrope and after decantation is sent (stream 9) to
the methanol flash still for recovery of the methanol that it contains. Makeup
amounts of o-xylene are supplied by stream 7. Recovered methanol (stream 8),
which contains lower boiling materials, is forwarded for further purification.
The crude DMT (stream 10) is forwarded to DMT purification for further frac-
tionation.
4. DMT Purification
By successive vacuum fractionation any o-xylene and light ends (stream 11) are
recovered for reycle, methyl p_-toluate and Denzoate (MPTB) are recovered for
sale or disposal as a burnable waste (discharge 0), and methyl g-formyl benzoate
and other materials are recovered as burnable wastes (discharge P). Finally,
DMT in high purity is recovered as a finished product (stream 12) and is sent
to storage. Higher boiling materials, including terephthalic acid (stream 13),
are recycled to the reactor.
-------
III-8
5. Methanol Purification
The aqueous layer (stream 9) from methanol recovery and recycled methanol
(stream 14) returned from polyester processors are sent to the methanol flash
still, where methanol and saturation amounts of o-xylene and any low boilers
are taken overhead (stream 15). The bottoms (discharge R), essentially water,
is sent to wastewater treatment. The methanol-rich overhead streams (streams 8
and 15) are sent to the low-boiler still, where dimethyl ether, other low boilers,
and any noncondensable gases are removed and forwarded for use as fuel (stream E).
The purified methanol (stream 16) leaves the bottom of the still and is returned
to the slurry tanks, along with any fresh methanol (stream 17) needed to satisfy
the methanol requirement. Scrap DMT (off-grade, etc.) is recycled (stream 18)
to crude DMT storage.
E. PURIFIED TPA FROM C-TPA
The purpose of the typical process shown in Fig. III-3 is to purify C-TPA to
make a terephthalic acid of quality acceptable for polyester fiber production.
This is done by hydrogenation in an aqueous medium to convert the impurities,
for example, p_-formyl benzoic acid, to a water-soluble form such as p_-toluic
acid and by crystallization to yield a product [purified TPA (P-TPA)] of very
o
high quality.
i Feed Slurry Preparation
C-TPA (stream 1) is sent by mechanically assisted gravity feed, along with hot
water (stream 2), to feed slurry tanks. The gases trapped in the C-TPA granules
are released to the atmosphere (vent A). The slurry of required consistency
(stream 3) is sent to the dissolver, where, with the application of pressure to
maintain a liquid phase, the temperature is raised to about 250°C to put the
terephthalic acid in solution in the water (stream 4).
2. Reaction
Hydrogen (stream 5) in the amount of abou . 0.004 g per g of C-TPA, which includes
a significant excess of the stoichiometrie requirement, is fed to the reactor.
The primary impurity, p_-formyl benzoic acid, is removed by converting it to the
8
more water-soluble £-toluic acid by the following reaction-.
-------
Fig. III-3. P-TPA by Purification of C-TPA
-------
111-10
00 0
'I catalvst
c-OH + 2H y > Hc- c-OH
(g-formyl benzole acid) (hydrogen) (p_-toluic acid) (water)
3. Crystallization
The discharge (stream 6) from the reactor is fed to crystallizers in series
wherein the temperature is lowered in stages to permit adequate crystal growth
during crystallization. Heat is removed from the crystallizing mass by allowing
the water to boil under controlled pressure in each crystallizer . Since tereph-
thalic acid exerts a vapor pressure of about 13 Pa at 100°C (see Appendix A),
some TPA is emitted in the vapor form along with water vapor and the excess
hydrogen (vent B). When vapors of terephthalic acid are cooled in the atmosphere,
they sublime to form solid particles that settle to the ground. The slurry of
terephthalic acid in water (stream 7) is sent from the crystallizers to the
atmospheric centrifuge feed tank, where the last stage of cooling and crystalli-
zation occurs. This is again accompanied by some discharge of water and tereph-
thalic acid vapors to the atmosphere (vent C).
4. Centrifuging, Drying, and Storage
The slurry (stream 8) of terephthalic acid in water is fed to centrifuges, where
the mother liquor, containing the undesired impurities in solution, is removed
(discharge W3) and sent to wastewater treatment. The wet cake, still in the
centrifuge, is washed with hot water to displace any remaining mother liquor.
The resultant wash liquor (stream 9), which is low in impurities, is forwarded
to the feed slurry tanks as part of the water (stream 2) required in the reactor;
the balance is made up by fresh process water (stream 12).
The wet cake (stream 10) leaves the centrifuges and is sent to the dryer, where
with the application of heat and a small amount of inert gas (IG) the moisture
content of the terephthalic acid is reduced to the desired level. The moisture
that is removed from the cake along with ' ne IG is discharged to the atmosphere
through a header that also vents the above-mentioned centrifuges (vent D).
Inert gas is used to convey the dried P-TPA (stream 11) to product storage.
The transport gas leaving the silos is discharged to the atmosphere (vent E).
-------
III-ll
F. REFERENCES*
1. J. L. Blackford, "Dimethyl Terephthalate and Terephthalic Acid," pp. 695.4021A—
695.4023H in Chemical Economics Handbook, Stanford Research Institute, Menlo Park,
CA (July 1977).
2. B. V. Vora et al., "The Technology and Economics of Polyester Intermediates,"
Chemical Engineering Progress 73(8), 74—80 (August 1977).
3. S. W. Dylewski, IT Enviroscience, Inc., Trip Report for Visit to Carolina Eastman
Company, Columbia, SC, Dec. 6, 7, 1977 (data on file at EPA, ESED, Research
Triangle Park, NC).
4. S. W. Dylewski, IT Enviroscience, Inc., Trip Report for Visit to Amoco Chemicals
Corporation, Decatur, AL, Oct. 31, Nov. 1, 1977 (data on file at EPA, ESED,
Research Triangle Park, NC).
5. S. W. Dylewski, IT Enviroscience, Inc., Trip Report for Visit to Hereofina,
Wilmington, NC, Nov. 17, 18, 1977 (data on file at EPA, ESED, Research Triangle
Park, NC).
6. D. F. Durocher e_t al., p. 4 in Screening Study to Determine Need for Standards
of Performance for New Sources of Dimethyl Terephthalate and Terephthalic Acid
Manufacturing, EPA Contract No. 68-02-1316, Task Order No. 18 (July 1976).
7. L. M. Elkin, "Terephthalic Acid and Dimethyl Terephthalate," pp. 49—55 in
Report No. 9, A private report by the Process Economics Program, Stanford
Research Institute, Menlo Park, CA (February 1966).
8. AMOCO, Standard Oil Co. (Indiana), Terephthalic Acid and Purified Terephthalic
Acid Processes [16-105-P(l-75)] (unpublished report).
^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.
Process emissions from the model plants are based on emission data included in
trip reports, responses to EPA letters requesting information from sites not
visited, and the GCA technology reports.1—7 Literature sources, such as
the SRI Chemical Economics Handbook and the Kirk-Othmer Encyclopedia of Chemical
Technology, were utilized to gain a better understanding of process unit opera-
tions and process chemistry.
A. CRUDE TEREPHTHALIC ACID PROCESS
1.. Model Plant*
The model plant (Fig. III-l) has a crude terephthalic acid (C-TPA) capacity of
230 Gg/yr based on operating 8760 hr/yr.** A number of existing production units
are of this size, but the older units are smaller.
Typical raw-material, in-process, product, and waste by-product storage-tank
capacities are estimated for the 230-Gg/yr plant. The storage-tank parameters
are given in Sect. IV.A.Z.e, and estimates of potential fugitive emission sources
are given in Sect. IV.A.Z.f. Characteristics of the model plant that are important
in air-dispersion modeling are given in Tfole B-l in Appendix B.
"~*See~~pTl-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 corres
Tondingly reduced. Control devices will usually operate on the same cycle as
the process. From the standpoint of cost-effectiveness calculations, the erro.
introduced by assuming continuous operation is negligible.
-------
IV-2
2. Sources and Emissions
Emission sources and quantities for the C-TPA process are summarized in Table IV-1.
a. Reactor Vent The reactor vent gas (A, Fig. III-l) contains nitrogen (from air
oxidation); unreacted oxygen; unreacted p_-xylene; acetic acid (reaction solvent);
carbon monoxide, carbon dioxide, and methyl acetate resulting from oxidation of
jD-xylene and acetic acid that are not recovered by the high-pressure absorber,-
and water, some of which results from oxidation and some from evaporation during
absorption with water in the high-pressure absorber. Table IV-2 gives the composi-
tion of this stream based on consideration of data from several sources.
The quantity of VOC emitted at vent A can be higher if the absorber is operated
at a lower pressure than that in the model plant. The quantity can also vary
with the temperature of the exiting vent gases.
b. Crystallization, Separation, and Drying Vent The gases vented from the crys-
tallization of terephthalic acid and the separation of the crystallized solids
from the solvent by centrifugation or by filtration are the noncondensable gases
that are released during crystallization and the VOC vapors that are carried by
those gases. These vent gases and the C-TPA dryer vent gas are combined and
released to the atmosphere (B, Fig. III-l). Different methods employed in this
processing section can result in less noncondensable gases and less accompanyiriq
VOC being emitted from this vent. However, the VOC emission from the reactor
vent may be commensurately increased. ' ' '
c. Distillation and Recovery Vent (C, Fig. III-l) The gases vented from the dis-
tillation section are the small amount of gases dissolved in the feed stream to
distillation, the inert gas used in inert blanketing, in instrument purging,
and in pressure control, and the VOC vapors that are carried by the noncondens-
able gases. The quantity of this discharge is normally small.
d. Product Transfer Vent The gas vented (D Fig. III-l) from the bag filters on
the product storage tanks (silos) is dry, reaction-generated, inert gas con-
taining the VOC that were not absorbed in the high-pressure absorber. The vented
gas stream contains a small quantity of TPA particulate that is not removed by
347
the bag filters. ' '
-------
IV-3
Table IV-1. Uncontrolled VOC Emissions from
Crude Terephthalic Acid Model Plant
Emissions
Emission Source
Stream
Designation
(Fig.III-1)
Ratio (g/kg)
VOC CO
separation, and drying vent
Distillation and recovery
vent
Product transfer vent
Storage and handling
D
1.14
1.78
Rate (kg/hr)
VOC CO
Reactor vent
Crystallization,
A
B
14.6
1.9
17 383.3
49.9
446
29.9
46.7
g of emission per kg of product produced.
Based on 8760 hr of operation per year.
CStream contains 0.7 g of TPA particulatesAg; not included.
VOC and CO emissions originated in reactor off-gas used for transfer.
53
Raw material storage
Other storage
Fugitive
Secondary
Incinerator
Wastewater treatment
Total
F,G,I
H,J
K
L
M
N
0.112
0.006
0.58
0.00482
0.00123
<0.004
20.13 19
2.94
0.17
15.26
0.126
0.0323
<0.1
528.4 499
-------
IV-4
Table IV-2. Composition of Model-Plant
Reactor Vent Gas (Vent A)a
Component
Nitrogen
Oxygen
C°2
CO
p_-Xylene
Acetic acid
Methyl acetate
Water
Composition
(wt %)
94.71
2.58
0.91
0.81
0.29
0.03
0.38
0.29
100.00
Emission
Ratio (g/kg)b
1985
54
19
17
6
0.6
8
6
2095.6
aSee refs 3, 4, 7, and 8.
g of emission per kg of C-TPA produced.
-------
e.
IV-5
Storage and Handling Emissions Emissions result from storage of p_-xylene,
acetic acid, and n-propyl acetate. The emission from p_-xylene storage occurs
only during filling of the tanks since they are maintained at a constant tempera-
ture. Sources for the model plant are shown in Fig. III-l (F through J). Storage-
tank parameters for the model plant are given in Table IV-3. The calculated
emissions in Table IV-1 are based on fixed-roof tanks, half full, an 11°C diurnal
9
temperature variation, and the use of the emission equations from AP-42. How-
ever, breathing losses were divided by 4 to account for recent evidence
indicating that the AP-42 breathing-loss equation overestimates emissions.
There are no VOC handling emissions since the product, C-TPA, is transferred in
the solid form and by-product waste methyl acetate is transported by pipeline
to incinerators.
f. Fugitive Emissions Pumps, compressors, valves, and pressure relief devices on
VOC-containing streams are potential sources of fugitive emissions (K in Fig. III-l)
The model plant is estimated to have 50 pumps, 900 process valves, and 40 pressure
relief devices in VOC service. The fugitive emission factors from Appendix C
were applied to these estimates, and the totals are shown in Table IV-1.
q. Secondary Emissions Secondary emissions can result from the handling and dis-
posal of process waste-liquid streams. Three potential sources (L, M, and N)
are indicated in Fig. III-l for the model plant. The secondary emissions from
burning still residues (L) and methyl acetate waste (M) were calculated with
the emission factors from AP-42 for residue oil and distillate oil, respectively.
The still residues also contain some bromine compounds and inorganic solids.
Care must be exercised upon incineration to avoid the release of free bromine
and particulates to the atmosphere.
The secondary emissions from wastewater treatment (source P) were estimated by
procedures that are discussed in a separate EPA report on secondary emissions.
An estmate of wastewater composition and flow rate was made, based on industry
data.4 A Henry's-law constant was then calculated for the vapor-liquid system
and the emission rate was estimated by the estimating approaches given in the
literature.
-------
IV-6
Table IV-3. Crude Terephthalic Acid Model-Plant Storage-Tank Data
Purpose
Raw material
In-process
Raw material
Mother liquor
Raw material
Catalyst mix
Burner feed
Product
Content
p-Xylene
£-Xylene
Acetic acid
Acetic acid
Propyl acetate
Acetic acid
Methyl acetate
C-TPA
Quantity
2
1
1
1
1
1
1
4
Size
-(n3)
5770
1000
660
1200
114
455
114
4600
Turnovers/
yr
15.9
2°
15.9
2°
12
2C
2C
21
Temperature
(°C)a
42b
42b
25
40
25
40
25
25
Average bulk temperature.
Controlled temperature.
"These tanks operate at essentially constant level, and the turnovers represent
shutdown events.
-------
IV-7
B. C-TPA PROCESS VARIATION
In the Carolina Eastman process, where acetaldehyde is used to make up acetic
acid losses, the VOC emissions are very similar to those associated with the
model process with the exception that in the acetaldehyde process a small amount
4
of methyl bromide is also emitted.
A discussion regarding the Hereofina process is presented in Sect. IV-D.
C. DMT BY ESTERIFICATION OF C-TPA
1. Typical Plant
The typical plant (Fig. III-2) for dimethyl terephthalate (DMT) production has
a capacity of 269 Gg/yr (1.17 X C-TPA capacity) based on operating 8760 hr/yr.
4 7,13
Some existing production units are of this size; other units are smaller.
New construction will likely be of the capacity of the typical plant.
2. Sources and Emissions
Uncontrolled VOC emission quantities from process, storage, fugitive, and secondary
sources in DMT production are summarized in Table IV-4 and are discussed below.
The discharge locations are shown in Fig. III-2.
Slurry Mix Tank Vent The gases present in the voids of the crude terephthalic
acid (C-TPA) bulk solid are displaced by and saturated with methanol during
4 6
slurry preparation. The gas/vapor mixture is released at vent A.
b. Reactor Sludge Transfer Vent Some of the impurities and the catalyst contained
in C-TPA are discharged from the crude reaction stream after evaporation and
stripping of the catalyst. This discharge is accompanied by some DMT parti-
4 6
culate emission at vent B. —
c Vacuum Jet Condenser Vent Air in-leakages occurring during vacuum distillation,
-—— 4 e
along with some VOC, are discharged at vent D. —
d. Methanol Flash Still Vent Inert gases that originate in recycled methanol
(returned from polymer plant) and that are introduced for blanketting, along
4 6
with some VOC, are discharged at vent F. —
-------
IV-8
Table IV-4. Uncontrolled VOC Emissions from Typical
Dimethyl Terephthalate Plant
Emission Source
Slurry mix tank vent
Reactor sludge transfer vent
Vacuum jet condenser vent
Methanol flash still vent
Storage vents
Crude DMT
Methanol
DMT
Other storage
Fugitive
Secondary
Process boiler
Incinerator
Wastewater
Stream
Designation
(Fig.III-2)
A
B
D
F
C
G,H
L
I K
M
E
N — P
QrR
Emissions
a
Ratio
(g/kg)
1.0
c
0.34
0.02
0.09
0.13
e
0.03
0.66
0.0018
NSd
NS
2.27
Rate
(kg/hr)
30.72
c
10.44
0.61
2.80
3.99
e
0.92
20.43
0.06
NS
NS
69.9
g of emission per kg of product.
DBased on 8760 hr of operation per year.
"Particulate emission of 0.038 g/kg and 1.17 kg/hr.
Not significant.
6Particulate emission of 0.18 g/kg and 5.53 kg/hr.
-------
IV-9
e.
Storage and Handling Emissions Emissions result from storage of recycled and
fresh methanol, from o-xylene, and from certain in-process tanks containing
VOC. Location of storage tank vents C, G—L are shown in Fig. III-2. The
quantities shown in Table IV-4 are representative of the emissions reported by
• j «. 4 6
industry. —
f. Fugitive Emissions Pumps, compressors, process valves, and pressure relief
devices on VOC-containing streams are potential sources of fugitive emissions
(M, Fig. III-2). The typical DMT plant is estimated to have 89 pumps, 1100 process
valves, and 16 pressure relief devices. The fugitive emission factors from
Appendix C were applied to these estimates, and the totals are shown in Table IV-4.
g. Secondary Emissions Emissions can result from the handling and disposal of
gaseous and liquid process wastes. Stream E, containing dimethyl ether and
other vapors, is sent by pipeline to a process boiler, where it is used as a
fuel. The emission from this source is very small, as are the emissions from
incineration of waste streams N, 0, and P. The wastewater streams (Q,R) going
to wastewater treatment are small and the emission from their disposal is also
expected to be small.
D. PROCESS VARIATION
In the process used by Hereofina, where air is used to oxidize a mixture of
p_-xylene and methyl p-toluate, the light ends that are formed and the p_-xylene
that does not react are carried on a stream of nitrogen and other gases. The
uncontrolled VOC from the oxidation reactor can be as much as 124 g/kg of DMT and
from the esterification can be 68 g/kg of DMT; other emissions from this process
14
are similar to those for the typical plant for production of DMT.
E. PURIFIED TPA FROM C-TPA
The purification of C-TPA by hydrogenation in an aqueous medium does not involve
the handling or generation of VOC; therefore no VOC are emitted to the atmosphere.
During the venting of excess hydrogen and water vapor at elevated temperatures,
vaporized TPA is emitted; however, the TPA vapors sublime on contact with the
atmosphere and fall to the ground. Since this emission is not considered to be
VOC, no further treatment of this process will be addressed. The particulate
emissions for P-TPA of 230-Gg/yr capacity are shown in Table IV-5.
-------
IV-10
Table IV-5. Emission from Purified Terephthalic Acid Typical Plant
a
Emission Source
Feed slurry tank vent
Crystallizer vent
Atmospheric centrifuge feed tank
Dryer vent
Silo dust collector vent
Stream
Designation
(Fig.III-3)
A
B
C
D
E
. . b
Emissions
Q
Ratio
(g/kg)
0.088
0.098
0.023
0.0012
0.0017
d
Rate
(kg/hr)
5.08e
5.69e
1.32e
0.07e
0.10
See ref 7.
Emissions shown are TPA particulates. No VOC present in processing steps.
/-i
g of emission per kg of product.
Based on 8760-hr/yr operation.
'Emission quantities following water scrubber.
e
-------
IV-11
F. REFERENCES*
1. Amoco, Standard Oil Co. (Indiana), Terephthalic Acid and Purified Terephthalic
Process [16-105-P(l-75>] (unpublished report).
2. L. M. Elkin, Terephthalic Acid and Dimethyl Terephthalate, pp 49--5S in
Report No. 9, A private report by the Process Economics Program, Stanford
Research Institute, Menlo Park, CA (February 1966).
3 S. W. Dylewski, IT Enviroscience, Inc., Trip Report for Visit to Amoco Chemicals
Corporation, Decatur, AL, Oct. 31. Nov. 1, 1977 (on file at EPA, ESED,
Research Triangle Park, NC).
4 S W Dylewski, IT Enviroscience, Inc., Trip Report for Visit to Carolina Eastman
Company, Columbia, SC, Dec. 6, 7, 1977 (on file at EPA, ESED, Research
Triangle Park, NC).
5. J. C. Edwards, Tennessee Eastman Company, letter dated Aug. 31, 1978, in response
to EPA request for information on emissions from TPA/DMT production facilities.
6. D. W. Smith, E. I. du Pont de Nemours and Co., letter dated Oct. 29, 1978, in
response to EPA's request for information on emissions from DMT/TPA production
facilities.
7. D. F. Durocher et al., p. 4 in Screening Study to Determine Need for Standards
of Performance for New Sources of Dimethyl Terephthalate and Terephthalic Acid
Manufacturing, EPA Contract No. 68-02-1316, Task Order No. 18 (July 1976).
8. L. M. Elkin, "Terephthalic Acid and Dimethyl Terephthalate,11 pp. 49—55 in Report
No. 9, A private report by the Process Economics Program, Stanford Research
Institute, Menlo Park, CA (February 1966).
9,. C. C. Masser, "Storage of Petroleum Liquids," pp. 4.3-1—4.3-17 in Compilation
of Air Pollutant Emission Factors, AP-42, 3d ed., Part A (August 1977).
10. T. Lehre, "Fuel Oil Combustion," pp. 1.3-1—1.3-5 in Compilation of Air Pollutant
Emission Factors, AP-42, 3d ed., Part A (August 1977).
11. 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 48,
1975, sponsored by AIChE and EPA Technology Transfer.
12. D. Mackay and P. J. Leinonen, "Rate of Evaporation of Low-Solubility Contaminants
from Water Bodies to Atmosphere," Environmental Science and Technology 9(13),
1178—1180 (December 1975).
-------
IV-12
13. Telephone conversation of S. W. Dylewski, IT Enviroscience, Inc., with D. W. Smith,
E. I. du Pont de Nemours, May 14, 1979.
14. S. W. Dylewski, IT Enviroscience, Inc., Trip Report for Visit to Hereofina,
Wilmington, NC, Nov. 17, 18, 1977 (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. CRUDE TEREPHTHALIC ACID PROCESS
1. Reactor Vent and Product Transfer Vent
There is demonstrated performance of carbon adsorption of VOC from a gas stream
similar to the reactor vent gas and product transfer vent gas (A and D, respec-
tively, Fig. III-l).1 It is estimated that the vent stream from the model plant
will perform similarly in carbon adsorption and effect a VOC emission reduction
of 97% or greater. It should be noted that the CO emissions will not be reduced
by carbon adsorption.
The reactor vent gas passes through one of the carbon beds, where the VOC are
adsorbed, and is then released to the atmosphere. When the first carbon bed
approaches breakthrough, the feed gas is routed to another carbon bed. At this
point regeneration of the first bed by steam stripping is started. The VOC-laden
stripping steam is then condensed and decanted. The p_-xylene layer is returned
to the reactor section, and the aqueous layer is forwarded to distillation for
recovery of the water-soluble VOC. When essentially all the VOC are stripped
from the first bed, a purge stream of VOC-depleted effluent from the second bed
is forced by a blower through the first bed to purge the remaining VOC and to
cool the bed for adsorption.
An alternative to the carbon adsorption system employed in the controlled model
plant is a thermal oxidizer. With a properly designed system operating at 1100°C
for efficient CO destruction, a reduction of 99% or greater in VOC and in CO
can be achieved. Because of the high percentage of nitrogen present in the
vent gas, 176 GJ of supplemental fuel per hour is needed to achieve the desired
temperature. Although 133 GJ/hr of energy as steam can be recovered, the energy
requirement balance of the plant needs to be considered. Thermal oxidation is
2
covered by a separate EPA report.
A reduction in emissions from vents A and D can be achieved by a change in
the high-pressure absorber in the model plant by providing a compound system
rather than the usual multistage system wherein the liquor from the lower
portion is largely recycled and the upper portion is irrigated by once-through
water, as is practiced by Carolina Eastman.3 This modification could reduce VOC
emissions from vents A and D by 36%.
-------
V-2
2.
Crystallization, Separation, and Drying Vent and Distillation and Recovery Vent
The emissions from the crystallization, separation, and drying vent (B, Fig. III-l)
and from the distillation and recovery vent (C, Fig. III-l) can be piped to a
header; the combined streams can be controlled by compressing the vent gas with
a blower, combining it with stream 9 (Fig. III-l), and sending the combined
stream to the carbon adsorption system. The VOC emission reduction is estimated
to be 97% or greater (B, C, Table V-l).
An alternative to the carbon adsorber for vent streams B and C is the use of
aqueous absorbers. The absorption of acetic acid from stream B will reduce
4
emissions by 98%, and the absorption of methyl acetate from stream C will reduce
emissions by 96%.
Storage and Handling Emissions
The emissions from p_-xylene storage tanks (F, Fig. III-l) are not large at the
storage temperature of 42°C and a vapor pressure of 3 kPa; therefore no controls
are indicated. The industry does, however, use conservation vents to minimize
losses. The emissions from acetic acid storage tanks (G, H, Fig. III-l) are
controlled by being vented through aqueous absorbers, as is done in industry.4'5
The resultant aqueous solution is returned to the process. The emissions from
methyl acetate storage (J) are also controlled by an aqueous absorber. An alter-
native to the use of an aqueous absorber is to collect, compress, and send the
emissions from vents G, H, and J to the above carbon adsorber. Handling of
the product, a solid with a high melting point (see Appendix A), does not result
in VOC emissions. Options for control of storage and handling emissions are
covered in a separate EPA report.
Fugitive Emissions
Controls for fugitive emissions from the synthetic organic chemical manufacturing
fj
industry are discussed in a separate EPA document. Controlled fugitive emissions
(K) calculated with factors given in Appendix C are included in Table V-l; these
factors are based on the assumption that major leaks are detected by an appropriate
leak-detection system and corrected.
-------
Table V-l. Controlled Emissions from Crude Terephthalic Acid Model Plantc
Stream
Designation
Emission Source (Fig. III-l)
Reactor vent A
Crystallization, B
separation, and drying
vent
Distillation and C
recovery vent
Product transfer D
c,d
vent
Storage vents
p_-Xylene F
Acetic acid and G,H,J
methyl acetate
Propyl acetate I
Fugitive K
Secondary
Incineration L/M
Wastewater N
treatment
Emission
Control Device Reduction
or Technique (%)
Carbon adsorber 97+
Carbon adsorber 97+
Carbon adsorber 97+
Carbon adsorber 97+
None
Aqueous absorber 98
None
Detection and correction 71
of major leaks
None
None
Ratio
VOC
0.44
0.057
0.034
0.053
0.11
Emissions
(g/kg)b Rate (kg/hr)
CO VOC
17 11.5
1.50
0.89
2 1.41
2.81
CO
446
53
0.0001 0.003
0.001
0.169
0.006
<0.004
0.874
0.036
4.42
0.158
<0.100
19 22.8
499
3A11 emissions are based on 8760 hr of operation per year.
bg of emission per kg of product.
CVOC and CO emissions originated in reactor off-gas used for transfer.
dStream also contains 0.7 g of TPA particulates/kg; not included.
-------
V-4
5. Secondary Emissions
Secondary VOC emissions resulting from burning the still residues and methyl
acetate waste (L and M, respectively, Fig. III-l) are estimated to be very small.
No control has been identified for the model plant. Still residues (L) containing
bromine and inorganic solids will probably require either prior removal or post-
incineration emission-control devices to control bromine and particulate emissions
to the atmosphere. Calculations based on estimated wastewater flow rates and
compositions for the model plant indicate that the emissions from wastewater
treatment of these wastes are relatively small. No control system has been
identified for the model plant.
B. C-TPA PROCESS VARIATION
In the Carolina Eastman process, where acetaldehyde is used to make up acetic
acid losses, a small amount of methyl bromide is present in the emissions and
it is not certain how stable this chemical is in carbon adsorption nor how
effectively it can be removed and recovered.
C. CURRENT EMISSION CONTROL USED IN C-TPA PRODUCTION
The control devices and techniques in current use by the terephthalic acid pro-
ducers are discussed in Appendix E.
D, DMT BY ESTERIFICATION OF C-TPA
1. Slurry Mix Tank Vent
The VOC emission from the slurry mix tank vent (A, Fig. III-2) can be controlled
by passing the vent gas through an o-xylene absorber. o-Xylene has a higher
boiling point than methanol and is a solvent for methanol (see Appendix A).
Based on industry experience and supported by engineering data, the VOC emission
reduction is estimated to be 96% or greater (A, Table V-2).
2. Reactor Sludge Transfer Vent
The DMT particulate emission from the reactor sludge transfer vent (B, Fig. III-2)
can be essentially completely controlled by an o-xylene absorber,- however, some
VOC emission is created by the vaporization of o-xylene into the carrier gas
(B, Table V-2).
-------
Table V-2. Controlled VOC Emissions from Dimethyl Terephthalate Typical Plant
Emission Source
Slurry mix tank vent
Reactor sludge transfer
vent
Vacuum jet condenser
vent
Methanol flash still
vent
Storage vents
Crude DMT
Methanol
DMT
Other storage
Fugitive
Secondary
Stream
Designation
(Fig. III-2)
A
B
D
F
C
G,H
L
I — K
M
E
N — P
Q,R
Emission
Control Device Reduction
or Technique (%)
o-Xylene absorber 96
o-Xylene scrubber c
Refrigerated condenser 81
None
o-Xylene absorber 99
Water absorber 90
Methanol scrubber c
Conservation vent
Detection and correction 73
of major leaks
None
None
None
VOC
Ratio (g/kg)
0.04
o.oiid
0.065
0.02
0.0009
0.013
0.196
0.03
0.175
0.0018
NSf
NS
0.54
Emissions
3 Rate (kg/hr)b
1.23
0.34d
1.98
0.61
0.028 f
Ul
0.40
5.84
0.92
5.45
0.06
NS
NS
16.8
ag of emission per kg of product.
Based on 3760 hr of operation per year.
CParticuLate reduction is essentially 100°
Some o-xylene is vaporized.
SMethanol vaporized during scrubbing.
Not significant.
-------
V-6
3. Vacuum-Jet Condenser Vent
The VOC emission from the vacuum-jet condenser vent (D, Fig. III-2) is mainly
o-xylene, which can be reduced by 81% with the use of a refrigerated condenser
(D, Table V-2).
4. Methanol Flash-Still Vent
No control has been indicated for this source (F, Table V-2). The gases from
this vent can be controlled by combining it with vent A, which is controlled by
an o-xylene absorber.
5. Storage Emissions
Emission of VOC from crude DMT storage, at an elevated temperature, is controlled
by an o-xylene absorber and results in a reduction of 99% or greater (C, Table V-2).
Methanol storage vents are controlled by a water absorber and results in a reduction
of 90% or greater (G and H, Table V-2).
The particulate emisson from DMT storage is essentially completely controlled
by a methanol scrubber. There is, however, some vaporization of methanol into
the carrier gas (L, Fig. III-2). If o-xylene were considered by industry to be
a feasible medium for scrubbing DMT particulate, with a lower vapor pressure
less vapors would be emitted with the carrier gas. The remaining storage tanks
with minor emissions are equippped with conservation vents but are otherwise
uncontrolled (I-K, Fig. III-2). Options for control of storage emissions are
6
covered in a recent EPA report.
6. Fugitive Emissions
Controls for fugitive emissions from the synthetic organic chemical manufacturing
industry are discussed in a separate EPA document.7 Controlled fugitive emissions
calculated with factors given in Appendix C are included in Table V-l; these
factors are based on the assumption that mrjor leaks are detected by an appropriate
leak-detection system and corrected.
7. Secondary Emissions
No controls are identified for emissions occurring when discharges E, N—P (Fig. HI-2
are burned. No controls are identified for emission from wastewater treatment
of discharges Q and R (Fig. III-2).
-------
V-7
E. CURRENT EMISSION CONTROL USED IN DMT PRODUCTION
The control devices and techniques in current use by the dimethyl terephthalate
acid producers are discussed in Appendix E.
F. DMT PROCESS VARIATION
In the process used by Hereofina, where air is used to oxidize a mixture of
£-xylene and methyl E-toluate, an aqueous absorber is used to control the emission
of methanol and low boilers that are carried by inert gases from methanol recovery.
About 99% of the methanol emission is reduced. No data are available on reduction
of low-boiler emissions.1
G. PURIFIED TPA FROM C-TPA
Water scrubbers are used to control the particulate emissions from the purified
4
TPA process.
-------
V-8
H. REFERENCES*
1. S. W. Dylewski, IT Enviroscience, Inc., Trip Report for Visit to Hereofina,
Wilmington, NC, Nov. 17, 18, 1977 (on file at EPA, ESED, Research Triangle
Park, NC).
2. J. Blackburn, IT Enviroscience, Inc., Thermal Oxidation (July 1980) (EPA/ESED
report, Research Triangle Park, NC).
3. S. W. Dylewski, IT Enviroscience, Inc., Trip Report for Visit to Carolina Eastman
Company, Columbia, SC, Dec. 6, 7, 1977 (on file at EPA, ESED, Research Triangle
Park, NC).
4. S. W. Dylewski, IT Enviroscience, Inc., Trip Report for Visit to Amoco Chemicals
Corporation, Decatur. AL, Oct. 31, Nov. 1, 1977 (on file at EPA, ESED, Research
Triangle Park, NC).
5. D. W. Smith, E. I. du Pont de Nemours and Co., letter dated Oct. 20, 1978, in
response to EPA's request for information on emissions from TPA/DMT production
facilities.
6. D. G. Erikson, IT Enviroscience, Inc., Storage and Handling (September 1980)
(EPA/ESED report, Research Triangle Park, NC).
7. D. G. Erikson and V. Kalcevic, IT Enviroscience, Inc., Fugitive Emissions
(September 1980) (EPA/ESED report, Research Triangle Park, NC).
8. D. F. Durocher et al., p. 4 in Screening Study to Determine Need for
Standards of Performance for New Sources of Dimethyl Terephthalate and
Terephthalic Acid Manufacturing, EPA Contract No. 68-02-1316, Task Order No. 18
(July 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. ENVIRONMENTAL AND ENERGY IMPACTS
1. Crude Terephthalic Acid Model Process
Table VI-1 shows the environmental impact of reducing the VOC emissions by appli-
cation of the indicated controls to the several sources from the model plant.
The total reduction is indicated to be 4429 Mg/yr for the model plant.
a. Process Vents The carbon adsorber used for control of emissions from the reactor
vent (A), the crystallization, separation, and drying vent (B), the distillation
and recovery vent (C), and the product transfer vent (D) reduces the VOC emissions
by 4333 Mg/yr. The carbon adsorber uses steam and cooling water during regeneration
and power for the blowers, instruments, and lighting. The aqueous stream containing
VOC that is recovered requires additional steam and cooling water in the recovery
steps. The total energy required in the form of steam and power to recover the
VOC as indicated is 16 GJ/hr.
b. Other Emissions (Storage. Fugitive, and Secondary) Control methods described
for these sources for the model plant are aqueous absorbers for some storage
vents and correction of leaks for fugitive emissions. Application of these
systems results in a VOC emission reduction of 96 Mg/yr for the model plant.
The electrical energy and the process water required for the aqueous absorber
are negligible.
2. C-TPA Process Variations
The environmental and energy impacts of controlling the emissions from processes
using acetaldehyde to make up acetic acid losses are similar to the impacts
described for the model plant except for the possible need for a small amount
of caustic to neutralize by-products of me thyl bromide hydrolysis.
3. 1979 C-TPA Industry Emissions
The total VOC emissions from process, storage, fugitive, and secondary sources
during the domestic production of crude terephthalic acid in 1979 are estimated
to be 28.6 Gg. This is based on an estimated 1979 level of production of 1655 Gg
-------
Table VI-1- Environmental Impact of Controlled Crude Terephthalic Acid Model Plant
. '
Reactor vent
Crystallization, separation, and
drying vent
Distillation and recovery vent
Product transfer vent
Storage
Fugitive
Secondary
Total
Vent
Designation
(Fia. III-D
B
C
D
F I
G,H,J
K
L,M,N
VOC Emission Reduction
Control Device
or Technique (%)
Carbon adsorber 97
Carbon adsorber 9.
Carbon adsorber 97
Carbon adsorber 97
None
Aqueous absorber 98
Detection and correction of 71
major leaks
None
(Mg/yr)
3257
423
255
398
1
95
<
H
1
4429
*Basis is 8760 hr of operation per year.
-------
VI-3
of C-TPA required to produce the P-TPA and DMT (Hereofina not included). The
demand for these products is calculated by applying the estimated 7.75% annual
growth rate to the reported production for 1978 (see Sect. II), the estimated
emission ratios (see Tables IV-1 and V-l), and the level of control practiced
in the industry (see Table E-l). The process emissions are estimated to be 14%
controlled, storage emissions to be 5% controlled, fugitive emissions to be 50%
controlled, and secondary sources to be uncontrolled.
4. DMT by Typical Process for Esterification of C-TPA
Table VI-2 shows the environmental impact of reducing the VOC emissions by appli-
cation of the indicated controls to the several sources from the typical plant.
The total reduction is indicated to be 465 Mg/yr for the typical plant.
a.
Process Vents The o-xylene absorbers used for control of emissions from the
slurry mix tank vent (A), the reactor sludge transfer vent (B), and the refrige-
rated condenser used for control of emissions from the vacuum-jet condenser
vent (D) reduce the VOC emissions by 330 Mg/yr. The energy impact of these
emission control devices will not be significant.
b. Other Emissions (Storage, Fugitive, and Secondary) Control methods described
for these sources for the typical plant are an o-xylene absorber for crude DMT
storage, a water absorber for methanol storage, and a methanol scrubber for
control of particulates from DMT storage. Control of fugitive emissions is by
adequate methods of leak detection and maintenance. Application of these systems
results in a VOC emission reduction of 136 Mg/yr.
5. 1979 Industry Emissions from DMT via C-TPA (70% of DMT Production)
The total VOC emissions from process, storage, fugitive, and secondary sources
from the production of DMT from C-TPA domestically in 1979 are estimated to be
1.26 Gg. This is based on an estimated 1979 level of production of 926 Gg,
which is calculated by applying the estimated 7.75% annual growth rate to the
reported production for 1978 (see Sect. II),1 the estimated emission ratios
(see Tables IV-4, V-2), and the level of control practiced in the industry (see
Table E-2). The process emissions are estimated to be 49% controlled, storage
emissions to be 28% controlled, fugitive emissions to be 50% controlled, and
secondary sources to be uncontrolled.
-------
Table VI-2. Environmental Impact of Controlled Dimethyl Terephthalate Typical Plant
Emission Source
Vent
Designation
(Fig. III-2)
Control Device
or Technique
VOC Emission Reduction
Slurry mix tank vent
Reactor sludge transfer vent
Vacuum jet condenser vent
Methanol flash still vent
Storage vents
Crude DMT
Methanol
DMTb
Other storage
Fugitive
Secondary
A
B
D
F
C
G,H
L
I—K
M
E, N—P, Q, R
o-Xylene absorber
o-Xylene scrubber
Refrigerated condenser
None
o-Xylene absorber
Water absorber
Methanol scrubber
None
Detection and correction of
major leaks
None
96
81
99
90
73
258.33
(2.98)C
74.11
24.28
31.45
(51.16)'
131
465.03
<
H
I
aBasis is 8760 hr of operation per year.
DMT particulate emitted.
GVaporized o-xylene emitted.
vaporized methanol emitted.
-------
VI-5
6. 1979 Industry Emissions from DMT via Hereofina Process (30% of DMT Production)
The total VOC emissions from process sources from the production of DMT via the
Hereofina process in 1979 are estimated to be 6.0 Gg. This is based on an
estimated 1979 level of production of 394 Gg, which is the same percent of capa-
city operation as the estimate for the entire industry.
B. CONTROL COST IMPACT
1. Crude Terephthalic Acid Process
This section gives estimated costs and cost-effectiveness data for control of
VOC emissions from crude terephthalic acid production. Details of the model
plant (Fig. III-l) are given in Sects. Ill and IV. Cost estimates were deter-
mined by using the control device evaluatii
procedure used is described in Appendix D.
2
mined by using the control device evaluation report for carbon adsorption. The
Capital cost estimates represent the total investment required for purchase and
installation of all equipment and material needed for a complete emission control
system performing as defined for a new plant at a typical location. These estimates
do not include the cost of crude terephthalic acid production lost during installa-
tion or startup, research and development, or land acquisition.
The bases for the annual cost estimates for the control alternatives include
utilities, waste disposal, operating labor, maintenance supplies and labor,
recovery credits, capital charges, and miscellaneous recurring costs such as
taxes, insurance, and administrative overhead. The cost factors used are
itemized in Table VI-3. Recovery credits are based on the raw-material value
or the fuel value of the materials being recovered. Annual costs are end-of-year
costs for 1979.
a. Process Vents The estimated installed capital cost of a carbon adsorption
system designed to reduce the VOC emissions from the process vents by 97% or
greater is $1,100,000 (see Appendix D).
The process-vent gas rate varies directly with the production rate; Fig. VI-1
was plotted to show the variation of installed capital cost of a carbon adsorp-
tion system versus plant capacity.
-------
VI-6
Table VI-3. Factors Used in Computing Annual Costs
Carbon loading
Steam for regeneration
Granular activated carbon replacement every 5 yr
Utilities
Steam
Electricity
Cooling water
Fixed costs
Maintenance labor plus materials, 6%
Capital recovery, 18% (10 yr life @ 12% interest)
Taxes, insurance, administration charges, 5%
Recovery credits
Acetic acid
p-Xylene
Methyl acetate
15 kg VOC/100 kg carbon
0.6 kg/kg of carbon
$2.57/kga
$2.37/GJ ($2.50/10 Ib)
$8.33/GJ ($0.03/kWh)
$0.026/m3 ($0.10/103 gal)
29% installed capital
$0.42/kg
$0.44/kgb
$0.0083/kgC
aif it became necessary to replace the carbon every 2
$24,314/yr.
bSee ref 3.
CBased on fuel equivalent value of $1.90/GJ.
yr, the cost would increase
-------
Installed Capital (X $1000) (December 1979)
"3
H-
id
H
I
W 3
W U)
H- ft
O P)
8-8
M H-
ft
tr &
0 Q
(D 0
3 <
in
>
en H
0 0)
'TJ ft
ft
H- O
0 p>
3 T)
n
H-
ft
H
-------
VI-?
To determine the cost effectiveness of a carbon adsorption system, estimates
were made of the direct operating cost, of those related to miscellaneous capital,
and of capital recovery cost. The recovery credits for acetic acid and p_-xylene
were based on current market prices; credit for methyl acetate was based on
its fuel value. The net savings for a carbon adsorption system was calculated
to be $27/Mg of VOC emission reduced ($117,300 per year), as shown by Table VI-4.
The variation in savings versus plant capacity is shown in Fig. VI-2.
The cost effectiveness for control by thermal oxidation was not completed for
this study. Thermal oxidation is not practiced in the industry. Thermal oxi-
dation does not have the potential for resource recovery that is displayed by
carbon adsorption.
b. Storage Sources The control of vents from acetic acid and methyl acetate storage
is by use of aqueous absorbers. A separate 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.
A separate report cove
manufacturing industry.
A separate report covers fugitive emissions for all the synthetic organic chemicals
d. Secondary Sources No control system has been identified for controlling the
secondary emissions from incinerator or wastewater treatment. A separate EPA
report covers secondary emissions and their applicable controls for all the
synthetic organic chemicals manufacturing industry.
2. Dimethyl Terephthalate Process
The DMT process emissions are relatively minor and are controlled primarily by
xylene absorbers. The cost and cost effectiveness of these absorbers have not
been developed for this report.
-------
50
o
o
o
X!
tn
Cn
C
•H
Ul
4J
(fl
8
I
300
350
400
200
300 400
Plant Capacity (Gg/yr)
500
600
Fig. VI-2,
Net Annual Cost (Savings) vs Plant Capacity for
Emission Control by Carbon Adsorption
-------
VI-10
Table VI-4. Cost-Effectiveness Estimate for
Control of Model-Plant
C-TPA Process Emissions by Carbon Adsorption
Total installed capital cost $1,100,000
Annual costs 408,000
Recovery credits
Acetic acid (77,800)
p-Xylene (789,630)
Methyl acetate3 (19,520)
Net annual savings ($117,300)
Total VOC emission reduction 4333 Mg/yr
Cost effectiveness (savings) ($27/yr)
aFuel equivalent value of $1.90/GJ, or $0.0083/kg.
-------
VI-11
C. REFERENCES*
1 J L. Blackford, "Dimethyl Terephthalate and Terephthalic Acid," pp. 695.4021A—
695.4023H in Chemical Economics Handbook, Stanford Research Institute, Menlo
Park, CA (July 1977).
2 H. S. Basdekis, IT Enviroscience, Inc., Control Device Evaluation. Carbon
Adsorption (January 1981) (EPA/ESED report. Research Triangle Park, NC) .
3. "Current Prices of Chemicals and Related Materials." Chemical Marketing Reporter
215(16), 46, 57 (1979).
4. D. Erikson, IT Enviroscience, Inc., Storage and Handling (September 1980)
(EPA/ESED report. Research Triangle Park, NC)
5. D. Erikson and V. Kalcevic, IT Enviroscience, Inc., Fugitive Emissions
(September 1980) (EPA/ESED report, Research Triangle Park, NC).
6. J. Cudahy and R. 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.
-------
VII-1
VII. SUMMARY
Dimethyl terephthalate (DMT) is produced by the esterification purification of
crude terephthalic acid (C-TPA).1 Purified terephthalic acid (P-TPA) is produced
by the hydrogenation purification of C-TPA.2 C-TPA, in turn, is produced by the
air oxidation of E-xylene in the presence of acetic acid. DMT is also produced
by the air oxidation of a mixture of E-xylene and methyl p_-toluate followed by
esterification.
The annual growth rate of DMT and P-TPA is estimated to be 6.5 to 9.0%, and
production is expected to reach an average of 78 to 86% of capacity for both
products by 1982.
Emission sources and uncontrolled and controlled VOC emission rates for the DMT
process are given in Table VII-1; there are no VOC emissions from the P-TPA
process; the VOC emission sources and rates for C-TPA, the intermediate product,
are given in Table VII-2. The current emissions projected for the domestic
DMT/P-TPA industry based on the estimated degree of control existing in 1979
are 1.26 Gg of VOC from DMT via C-TPA, no VOC emissions from P-TPA via C-TPA,
28.6 Gg of VOC from C-TPA, and 6.0 Gg of VOC from DMT via the Herofina process.
Control devices for process vents on operating plants include a carbon adsorber
in C-TPA production and an o-xylene absorber and refrigerated condenser in DMT
production. An emission reduction of 97% or greater may be realized in a carbon
adsorber. The installed capital cost of a carbon adsorption system is $1,100,000.
The energy requirement for regeneration of the carbon bed and for recovery of
the VOC is 16 GJ/hr.
For the carbon adsorption system the cost effectiveness is a net savings of
$27/Mg of VOC reduction.
1B V Vora et al., "The Technology and Economics of Polyester Intermediates,"
Chemical Engineering Progress 73(8), 74—80 (August 1977).
'AMOCO Standard Oil Co. (Indian), Terephthalic Acid and Purified Terephthalic
Acid Processes [16-105-P(l-75)] (unpublished report). ^
ST T Blackford "Dimethyl Terephthalate and Terephthalic Acid," pp. 695.4021A
695 4023H in ^^E^icl^n^^^ Stanford Research Institute, Menlo Park,
CA (July 1977).
-------
VII-2
Table VII-1. Emission Summary for Typical Plant Producing
Dimethyl Terephthalate via C-TPA
(Capacity: 269 Gg/yr)
Emission
Slurry mix tank vents
Reactor sludge transfer
vents
Vacuum- jet condenser
vent
Methanol flash still
vent
Storage vents
Crude DMT
Methanol
DMT
Other storage
Fugitive
Secondary
Process boiler
Incinerator
Wastewater treatment
Vent
Designation
(Fig. III-l)
A
B
D
F
C
G,H
L
I-K
M
E
N-P
Q,R
VOC Emission Rate (kg/hr)
Uncontrolled
30.72
b
10.44
0.61
2.80
3.99
d
0.92
20.43
0.06
NSf
NS
69.9
Controlled
1.23
0.34C
1.98
0.61
0.028
0.40
5.84e
0.92
5.45
0.06
NS
NS
16.8
Based on 8760-hr/yr operation.
Particulate emission of 1.17 kg/hr.
CEmission resulting from vaporization of o-xylene scrubbing liquid.
Particulate emission of 5.53 kg/hr.
eEmission resulting from vaporization of methanol scrubbing liquid.
Not significant.
-------
VII-3
Table VII-2.
Emission Summary for Model Plant Producing
Crude Terephthalic Acid
(Capacity: 230 Gg/yr)
— —
Emission
Reactor vent
Crystallization, separa-
tion, and drying vent
Distillation and recovery
vent
Product transfer vent
Storage vents
p_-Xylene
Acetic acid and methyl
acetate
Propyl acetate
Fugitive
Secondary
Incinerator
Wastewater treatment
.
Vent
Designation
(Fiq. III-D
B
C
D
G,H,J
M
N
VOC Emission
Uncontrolled
383.3
49.9
29.9
46 7
*±U • '
2.81
n 1 7
\j . j. /
0.13
15.26
0.126
0.0323
528.4
Rate (kg/hr)*
Controlled
11.5
1.5
0.89
1.41
2.81
0.003
0.036
4.42
0.126
0.0323
22.8
Based on 8760-hr/yr operation.
-------
VII-4
The emission reduction for the o-xylene absorber on process emissions is 96%
and for the refrigerated condenser is 81%. The DMT process emissions are small,
and therefore cost and cost effectiveness of these controls have not been developed
for this report.
-------
A-l
APPENDIX A
Table A-l. Physical Properties of Acetaldehyde*
Synonyms
Molecular formula
Molecular weight
Physical state
Vapor pressure
Vapor specific gravity
Boiling point
Melting point
Density
Water solubility
Acetic aldehyde, ethyl aldehyde
C2H4°
44.05
Liquid
123,060 Pa at 25°C
1.52
20.8°C
0.7834 at 18°C/4°C
Infinite (hot H2
-------
A-2
Table A-2. Physical Properties of Acetic Acid*
Synonyms
Molecular formula
Molecular weight
Physical state
Vapor pressure
Vapor specific gravity
Boiling point
Melting point
Density
Water solubility
Methyl carboxylic acid, ethylic
acid, glacial acetic acid,
ethanoic acid, vinegar acid
C2H4°2
60.05
Liquid
1520 Pa to 20°C
2.07
117.9°C
16.6°C
1.0492 at 20°C/4°C
Infinite
*From: J. Dorigan et al., "Acetic Acid," p. AI-16 in Scoring
of Organic Air Pollutants. Chemistry, Production and Toxicity
of Selected Synthetic Organic Chemicals (Chemicals A-C),
MTR-7248, Appendix II, Rev. 1, MITRE Corp., McLean, VA
(September 1976).
-------
A-3
Table A-3. Physical Properties of Methanol*
Synonyms
Molecular formula
Molecular weight
Physical state
Vapor pressure
Vapor specific gravity
Boiling point
Melting point
Density
Water solubility
Methyl alcohol, carbinol, methyl
hydroxide
CH40
32.04
Liquid
17,050 Pa at 25°C
1.10
64.8°C
-93.9°C
0.7913 at 20°C/4°C
Infinite
*From: J. Dorigan et al., "Methanol," p. AIII-154 in Scoring _
of Organic Air Pollutants. Chemistry, Production and Toxicity
of Selected Synthetic Organic Chemicals (Chemicals F-N),
MTR-7248, Appendix II, Rev. 1, MITRE Corp., McLean, VA
(September 1976).
-------
A-4
Table A-4. Physical Properties of Methyl Acetate*
Synonyms
Molecular formula
Molecular weight
Physical state
Vapor pressure
Vapor specific gravity
Boiling point
Melting point
Density
Water solubility
Acetic acid, methyl ester
C3H6°2
74.08
Liquid
28,330 Pa at 25°C
2.55
57.8°C
-98.1°C
0.9330 at 20°C/4°C
Very soluble
*From: J. Dorigan e_t al_. , "Methyl Acetate," p. AIII-148 in Scoring
of Organic Air Pollutants. Chemistry, Production and Toxicity
of Selected Synthetic Organic Chemicals (Chemicals F-N),
MTR-7248, Appendix II, Rev. 1, MITRE Corp., McLean, VA
(September 1976).
-------
A-5
Table A-5. Physical Properties of Terephthalic Acid*
Synonyms
Molecular formula
Molecular weight
Physical state
Vapor pressure
Vapor specific gravity
Boiling point
Melting point
Density
Water solubility
p-Phthalic acid, TPA, benzene-
~~ p-dicarboxylic acid
C8H6°4
166.14
Solid
Negligible
Sublimes
>300°C sublimes without melting
1.51
Insoluble
*From- J. Dorigan et al. , "Terephthalic Acid," p. AIV-174 inScoring,
of Organic Air Pollutants. Chemistry, Production and Toxicity
of Selected Synthetic Organic Chemicals (Chemicals O-Z),
MTR-7248, Appendix II, Rev. 1, MITRE Corp., McLean, VA
(Spetember 1976).
-------
A-6
Table A-6. Physical Properties of Terephthalic Acid, Dimethyl
Ester*
Synonyms
Molecular formula
Molecular weight
Physical state
Vapor pressure
Vapor specific gravity
Boiling point
Melting point
Density
Water solubility
Dimethylterephthalate, DMT, 1,4-
benzene dicarboxylic acid, di-
methyl ester, dimethyl 1,4-ben-
zene carboxylate dimethyl ester
C10H10°4
194.19
Solid
133.3 Pa at 100°C
6.70
Sublimes
141.0 to 141.8°C
1.194 at 20°C/4°C
Slightly (hot)
*From: J. Dorigan e_t al_., "Dimethyl Terephthalates," p. AII-162
in Scoring of Organic Air Pollutants. Chemistry, Production
and Toxicity of Selected Synthetic Organic Chemicals (Chemicals
P-E) , MTR-7248, Appendix II, Rev. 1, MITRE Corp., Mclean, VA
(September 1976).
-------
A-7
Table A-7. Physical Properties of o-Xylene*
Synonym
Molecular formula
Molecular weight
Physical state
Vapor pressure
Vapor specific gravity
Boiling point
Melting point
Density
Water solubility
o_-Xylol
C8H10
106.2
Liquid
1,333 Pa at 32.1°C
3.66
144.4°C
-25°C
0.880 at 20°C/4°C
Insoluble
*From: J. Dorigan e_t a_l. , "o-Xylene," p. AIV-298 in Scoring
of Organic Air Pollutants. Chemistry, Production and
Toxicity of Selected Synthetic Organic Chemicals (Chemicals
Q-Z), MTR-7248, Appendix II, Rev. 1, MITRE Corp., McLean, VA
(September 1976).
-------
A-8
Table A-8. Physical Properties of p-Xylene*
Synonym
Molecular formula
Molecular weight
Physical state
Vapor pressure
Vapor specific gravity
Boiling point
Melting point
Density
Water solubility
£-Xylol
C8H10
106.2
Liquid
1,333 Pa at 27.3°C
3.66
138.5°C
13.2°C
0.8611 at 20°C/4°C
Insoluble
*From: J. Dorigan e_t a^. , "p_-Xylene", p. AIV-300 in Scoring
of Organic Air Pollutants. Chemistry, Production and
Toxicity of Selected Synthetic Organic Chemicals (Chemicals
0-Z), MTR-7248, Appendix II, Rev. 1, MITRE Corp., Mclean, VA
(September 1976).
-------
B-l
APPENDIX B
Table B-l. Air-Dispersion Parameters for
Crude Terephthalic Acid Model Plant with a Capacity of 230 Gg/yr
Reactor vent
Crystallization ,
separation, and
drying vent
Distillation and
recovery vent
Product transfer vent
Storage vents
p_-Xylene
Acetic acid
Methyl acetate
Propyl acetate
Fugitive *
Secondary
Incinerator
Wastewater
Carbon adsorber vent
Storage vents
g-Xylene
Acetic acid
Methyl acetate
Propyl acetate
Fugitive *
Secondary
Incinerator
Wastewater
voc
Emission
Rate
(g/sec)
106.5
13.9
8.3
13.0
0.75
0.060
0.029
~- 0.012
4.24
0.044
0.028
4.25
0.78
0.0002
0.001
0.008
1.25
0.044
0.028
Discharge
Height Diameter Temperature
(m) (m) (K)
Uncontrolled Emissions
20 0.76 311
20 0.1 311
20 0.05 311
30 0.46 311
12.2
9.8
7.3
7.3
30 1.58 1250
Controlled Emissions
30 1.22 316
12.2
20
20
7.3
30 1.58 1250
Flow Discharge
Rate Velocity
(m3/sec) (m/sec)
14.4 31.7
0.045 5.7
0.0014 0.71
1.76 10.6
27.9 14.2
14.4 12.3
27.9 14.2
•Fugitive emissions are distributed over an area of 100 m X 200 m.
-------
B-2
Table B-2. Air-Dispersion Parameters for Typical Plant for
Dimethyl Terephthalate with a Capacity of 269 Gg/yr
Source
Slurry mix tank vent
Reactor sludge
transfer vent
Vacuum-jet condenser
vent
Methanol flash still
vent
Storage vents
Crude DMT
Methanol
DMT
Other storage
Fugitive
Fecondary
Process boiler
Incinerator
Wastewater
Slurry mix tank vent
Rpactor sludge
'-Tansfer vent
Vacuum -jet condenser
vent
"ethanoJ flash still
vent
Storage vents
Crude DMT
Methanol
DMT
Other storage
Fugitive
Secondary
Process boiler
Incinerator
Wastewater
VOC
Emission
Rate
(g/sec)
8.53
a
2.90
0.17
0.78
1.11
0.26
5.71
0.017
Nil
Nil
0.34
0.094
0.54
0.17
0.0078
0.11
1.62
0.26
1.5
0.017
Nil
Nil
Discharge
Height Diameter Temperature
<">) (m) (K)
Uncontrolled Emissions
20 0.05 311
20 0.10 311
20 0.05 311
20
9.8
7.3
Controlled Emissions
20 0.05 311
20 , 0.05 311
20 0.10 293
20 0.05 311
Flow Discharge
Bate Velocity
(m /sec) (m/sec)
0.006 3.3
0.035 1.5
0.0002 0.1
0.006 3.0
0.0002 O.t
0.035 . 1.3
0.0002 0,1
Particulate emissions only.
"Fugitive emissions are distributed over an area of 150 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 Factor"
(kg/hr)
Pump seals ,
Light-liguid 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
b
3Based 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 ESTIMATING PROCEDURES
CRUDE TEREPHTHALIC ACID PROCESS—CARBON ADSORPTION (CA) SYSTEM COST ESTIMATE
As shown by Table IV-1 the total VOC flow from vents A, B, C, and D equals
509 8 kg/hr. The total gas flow fro, these vents is estimated to be 72,900 kg/hr,
which equals 34,000 scfm. The average VOC molecular weight is approximately
84 the estimated VOC concentration is 2360 pPmv, and the estimated loading
capacity is 15 Ib of VOC/100 Ib of carbon. From Fig. II-l of the control
device evaluation report for carbon adsorption,1 3.7 Ib of carbon are required/
1000 scf of waste gas. The total carbon requirement is therefore
37lb C 34.000 (60) scf 3_hL_ - gl^5QO_lb_of_C -
1000 scf hr cycle cycle
From Fig. IV-1 of the carbon adsorption report the December 1979 installed
capital for a 34,000-^scfm CA system is $750,000. To adjust the cost for stain-
less steel requirements a 1.5 adjustment was applied to the installed cost
except for the initial carbon. The total installed cost preliminary estimate
is $1 100 000 including the carbon cost for three beds. Figure IV-3 in the
carbon adsorption report indicates the annual cost to be $12/scfm, or $408,000.
The annual cost adjustments for fixed costs associated with the added capital
for stainless steel construction, the added utilities for product recovery
separation, and the equivalent raw-material recovery credits are included in
Table D-l.
S Basdekis and C S. Parmele, IT Enviroscience , Inc., Control Device
Adsorption (January 1981) (EPA/ESED report, Research
Triangle Park, NC) .
-------
D-2
Table D-l. Carbon Adsorption Control Cost Summary
Total installed capital
a
Annual cost
Fixed cost for extra capital
Utilities for recovery distillation
Recovery credits
Acetic acid
p-Xylene
Methyl acetate
Net annual cost
VOC emission reduction
Cost effectiveness (savings)
per Mg of VOC reduced)
Model Plant
230 Gg/yr
(34,000 scfm)
$1,100,000
$408,000
101,500
260,150b
(77,800)
(789,630)
(19,520)
($117,300)
4333 Mg/yr
($27.07)
Capacity
Model Plant
350 Gg/yr
(51,740 scfm)
$1,300,000
$595,000
119,000
395,900
(118,390)
(1,201,610)
(29,700)
($239,800)
6593 Mg/yr
($36.37)
Model Plant
450 Gg/yr
(66,520 scfm)
$1,600,000
$732,000
145,000
509,000
(152 ;22D)
(1,544,930)
(38,190)
($349,340)
8478 Mg/yr
$41.21)
aFrom Fig. IV-3 of the carbon adsorption report.
°Azeotropic distillation: steam $232,140; cooling water $28,010.
-------
E-l
APPENDIX E
EXISTING PLANT CONSIDERATIONS
Tables E-l and E-21-5 list process control devices reported to be in use by
industry. To gather information for the preparation of this report three site
visits were made to manufacturers of terephthalic acid and dimethyl terephthalate.
Trip reports have been cleared by the companies concerned and are on file at
EPA, ESED, in Research Triangle Park, NC.1'3'4 Some of the pertinent information
concerning process emissions from these existing terephthalic acid and dimethyl
terephthalate plants is presented in this appendix.
A. PROCESS EMISSIONS FROM EXISTING PLANTS
4
1 Hereofina Hanover Plant, Wilmington, NC
Hercofina manufactures dimethyl terephthalate by the Hercules-Imhausen-Witten
process. In this process E-xylene and recycled methyl toluate are oxidized
with air to form toluric acid, monomethyl terephthalate, and terephthalic acid.
This mixture is esterified to produce dimethyl terephthalate. p_-Xylene and
pure methanol are received by barge. Reclaimed methanol is received by tank
car and tank truck.
The main emission from the process, oxidizer off-gas, is due to the large
amount of nitrogen present as a result of air oxidation. The emission is
controlled by a carbon adsorption system (see Table E-3). Several processing
steps involve the use of methanol. The emissions from the processing steps are
collected by the methanol recovery header and fed to the methanol recovery
absorber for emission control before being released to the atmosphere (see
Table E-3). A portion of the DMT produced is converted to a solid form by
being passed through a flaker. Emissions from the flaker are discharged to the
atmosphere (see Table E-3).
Water is a by-product of the oxidation of £-xylene and the esterification of
toluic and terephthalic acids. After appropriate decantation and stripping,
the wastewater (containing soluble, nonstrippable organics) is sent to the
thermal oxidizer for disposal (see Table E-4). Residues formed throughout the
process resulting from oxidation and distillation are discharged and disposed
of by incineration at the site (see Table E-4).
-------
Table E-l• Emission Control Devices Currently Used by Terephthalic Acid Producers
Control Devices in Use
Source
Reactor vent
Crystallization,
separation, drying
vent
Distillation and
recovery vent
Product transfer vent
Storage vents
p-Xylene
Acetic acid
a
By Amoco
None6
Aqueous
absorber
None
Bag filter
Conservation
vent
Aqueous
absorber
By Du Pont,
Cape Fear'3
e
None
Aqueous
absorber
Aqueous
absorber
Bag filter
Conservation
vent
Aqueous
absorber
By Du Pont,
Hickoryb
e
None
Aqueous
absorber
Aqueous
absorber
Bag filter
Double-seal
floating roof
Aqueous
absorber
By Eastman, By
Carolina0 Hercofinad
None6 ' Carbon
adsorber
q
Aqueous NA
absorber
Aqueous NA
absorber
Bag filter NA
Conservation Conservatioi
vent vent
NA NA
See ref 1.
DSee ref 2.
"See ref 3.
3See ref 4.
2High-pressure absorber is considered to be a part of basic process.
fA small side stream is passed through a carbon adsorber for organic removal.
9Not applicable.
-------
i'able 5-2. Emission Control Devices Currently Used by Dimethyl Terephthalate Producers
Control Devices in Use
Source
Slurry mix tank
vent
Reactor sludge
transfer vent
Crude DMT tank vent
Methanol recovery
still, low-boiler
still vents
Methanol flash still
Storage
Methanol
o-Xylene
MPTB, MFB
DMT
MFB, MPTB waste
Sludge waste
Wastewaters
By Eastman,
Carolina
None
Water scrubber
Xylene absorber
Burned as fuel
None
Conservation
vent
Conservation
vent
Conservation
vent
Methanol
absorber
Incinerator
Incinerator
Wastewater
treatment
By Eastman,
Tennessee
Conservation
vent
Water scrubber
Conservation
vent
Burned as fuel
Conservation
vent
Conservation
vent
Conservation
vent
Conservation
vent
Methanol
absorber
Incinerator
Incinerator
Wastewater
treatment
By Du Pont,
c
Cape Fear
Hydrocarbon
scrubber
and vent
condenser
Hydrocarbon
scrubber
Hydrocarbon
scrubber
Burned as fuel
Conservation
vent
Water absorber
Conservation
vent
ND
Hydrocarbon vortex
scrubber
ND
Boiler
Wastewater
treatment
By Du Pont, By
Hickory Hercofina
Hydrocarbon NA
scrubber
and vent
condenser
Hydrocarbon NA
scrubber
Hydrocarbon NA
scrubber
Burned as fuel Carbon
adsorber
M
1
Conservation Water <-o
vent absorber
Double seal Floating
floating-roof, roof
conservation
vent
Conservation NA
vent
Bag filter NA
Hydrocarbon Methanol
scrubber scrubber
ND NA
Incinerator ND
Wastewater Incinerator
treatment
Sae ref 3.
3See ref 5.
'See ref 2.
3See ref 4.
"Not applicable.
fNo data.
-------
Table E-3. Direct Emissions (Hereofina Hanover Plant)
Pollutant Flow
(lb/1000 Ib of Product)
Emission Source
Emergency reactor pressure
relief
Oxidizer off-gas
Xylene-water decanter and
storage
Process tank vents
Emergency relief
Methanol recovery header
Vacuum jet barometric
tank
Emergency relief
DMT crystallization melt
tank, emergency
DMT flaker
DMT dust vent
Vent to:
Atmosphere
Carbon adsorption
column
Condenser
Xylene vent scrubber
Atmosphere
Methanol recovery
absorber
Atmosphere
Atmosphere
Atmosphere
Atmosphere
DMT dust collector
Pollutant
£-xy lene /oxidate
p_-xylene
Light VOC
p-xylene
Aromatic methyl
esters and
xylene
Methanol or
wastewater
Methanol
VOC
Aromatic
methyl esters
Methanol
Light ends
DMT dust
Before BCD3
None
28
60
No data
36
None
68
No data
None
None
0.2
5.02
After ECDa
None
0.93
12.06
No data
None
1
No data
None
None
0.2
0.03
M
I
aEmissions control device.
-------
E-5
Table E-4. Secondary Emissions (Hercofina Hanover Plant)
Pollutant Flow
(lb/1000 Ib of Product)
Potential
Emission Source
Wastewater
DMT finishina
Discharged to Pollutant
Incinerator HOAC
Formic acid
Formaldehyde
Methanol
Disposal No data
Pollutant
Rate
12
4
4
1
No data
Stream
Rate
281
No data
still residue
-------
E-6
Table E-5 lists the information received on emission control devices.
Carolina Eastman Company, Columbia, SC
Carolina Eastman Company at Columbia, SC, manufactures terephthalic acid (TPA)
and uses it as a raw material in the manufacture of dimethyl terephthalate
(DMT). The facilities at this site also include processes for converting DMT
to polyester products. The TPA processing steps are conducted in multiple
units, including six oxidizers, that are operated interchangeably depending on
product demand and maintenance needs. The DMT process is a single-train design.
These facilities were put into operation starting in Novermber 1976.
p_-Xylene, acetaldehyde, and fresh methanol are received by tank car. Recycle
methanol is transferred by pipeline from the polymer plant, where it is released
from DMT by transesterification with a glycol. TPA is conveyed by low-oxygen
gas from the TPA plant to the DMT plant. DMT is transferred by pipeline to the
polymer plant.
The TPA process used at the Columbia plant is the cobalt bromide—catalyzed air
oxidation of p_-xylene in the presence of acetic acid. The main discharge from
the process is the result of using air for oxidation. The reactor off-gas is
passed through a water absorber for recovery and control of emissions. A small
amount of low-oxygen gas is produced by passing the scrubbed gas through a
carbon adsorption bed. Some of the scrubbed gas is used to convey the product
to storage and is discharged to the atmosphere after it is passed through a bag
filter. The remainder is discharged directly to the atmosphere (see Table E-6).
The emission from the distillation and recovery of low boilers is sent to an
absorber for emission control before being discharged to the atmosphere (see
Table E-6). Small amounts of VOC emission are released during water-removal
distillation and during filtration and drying.
Emissions can result from the handling and disposal of wastewater from the
diltillation system. This stream may contain small amounts of methyl acetate,
n-propyl acetate, and acetic acid.
-------
Table E-5. Emission Control Devices (Hereofina Hanover Plant)
Emission
Control Device
Carbon adsorber
Solvent scrubber
Thermal oxidizer
w/heat recovery
Chilled solvent
scrubber
Dust collector
Control
Pollutant Efficiency (%)
£-xylene, 97, 80
light VOC
£-xylene , 97
other VOC
Acetic acid M.OO
Formic acid
Formaldehyde
Methanol
Methanol 99
Light ends
DMT dust 99
Control
Agent
Active
carbon
Xylene
Fire
Chilled
Solvent
Bag filter
Cost ($/M Ib
Size Capital
Two 9.5-ft diam, 1.05/M lba
X 22 ft long
1750 -gal tank 0.2ia
750-ft conden-
ser
140 MM Btu/hr 3.97°
6000-ft2 0.6ia
condenser
2
483-ft
cooler
8000 cfm 0.47°
of Product)
Annual
Operating
0.55b
O.llb
0.88d
0.32b
0.166
1972 basis.
31977 basis.
:1973 basis.
1975 basis.
not known.
M
I
-------
Table E-6. Direct Emissions from TPA Manufacture (Carolina Eastman)
Emission Source
Reactor off-gas
Low-boiler distillation
Decanter vent
Solids transport vent
Filter vent
Emission
Control Device
Solvent absorber
Solvent absorber
Atmosphere
Dust collector
Solvent absorber
Pollutant Flow
(lb/1000 Ib of Product)
Pollutant
MeOAc , p-xylene , MeBr ,
acetalydehyde , methanol
CO
MeOAc
Propylacetate
TPA particulate
VOC
CO
Acetic acid
a
Before BCD
NDb
ND
ND
0.0038
NDb
NDb
ND
NDb
After ECDa
4.26
11.2
0.035
0.0038
b
0.38°
1.72C
0.0017
w
i
00
Emission control device.
No data.
-------
E-9
The process for the manufacture of DMT at the Carolina Eastman Columbia plant
is the direct esterification of crude terephthalic acid with methanol. Emissions
from the slurry mix tank vents are caused by filling losses from batch prepara-
tion. The discharges from the slurry feed tank vents are also caused by filling
losses from batch feed preparation. The discharge from the jet seal pot is
caused by air in-leakage during vacuum, distillation. The discharges from the
sludge hoods are caused by evaporation losses during transfer of reactor sludge
from the stripper into containers used for transferring material to sales or to
landfill (see Table E-7).
The esterification process results in the formation of low-boiling materials
such as dimethyl ether and methyl acetate, which are handled by pipeline and
are disposed of in a fired boiler. It is estimated that the destruction is
essentially 100%. The by-products MPTB and MFB are also disposed of by burning.
Wastewater containing unknown amounts of VOC is discharged to the wastewater
system, and emissions can result from the handling and disposal of these materials
The sources of this water are water formed in esterification, steam from jets,
and water from the scrubber on the sludge discharge hood.
Amoco Chemicals Corp., Decatur, AL
The process used at Decatur for the manufacture of terephthalic acid is the
continuous air oxidation of p_-xylene in acetic acid solution. The first TA
plant at Decatur was built in 1966, and the last unit was completed ten years
later. p-Xylene is received by barge, and makeup acetic acid is received by
tank car. The PTA product is a solid and is shipped by rail car.
The main discharge from the process is due to the large amount of nitrogen
present as the result of air oxidation. The emission is passed through a
high-pressure water absorber before being released to the atmosphere (see
Table E-8).
VOC discharges result from the venting of dissolved inert gases present in the
liquid leaving the reactor under reactor pressure (see Table E-8). A minor
discharge results from miscellaneous process vents controlled by a low-pressure
absorber (see Table E-8).
-------
Table E-7. Direct Emissions from Dimethyl Terephthalate Process (Carolina Eastman)
Pollutant Flow
(lb/1000 Ib of Product)
Emission
Emission Source Control Device Pollutant Before BCD After BCD
TA slurry mix tanks, vents Atmosphere MeOH, o-xylene 0.0090 0.0090
TA slurry feed tanks, vents Atmosphere MeOH, o-xylene 0.0186 0.0186
Vent from sludge recovery Contact condenser o-Xylene, others 0.0913 0.0084
Decanter Atmosphere o-Xylene 7 x 15 ^ 7X 10 ^
Jet seal pot vent Atmosphere o-Xylene 3 X 10 3 X 10
Product transfer Solvent absorber DMT .0.171
MeOH 0.0113 0.146
Sludge hood vent Scrubber Particulate 12 g/m^ 2.97 mg/m^
Sludge hood vent Scrubber Particulate 15 ?/m 1'25 mg/m
-------
Taijle E-8. Direct Emissions TPA (Amoco, Decatur, AL)
Pollutant Flow
(lb/1000 Ib of Product)
Emission ouui.<-.c
Emergency vent
Nitrogen vent
Crystallizer
Process vent
Dehydration
tower
3No emission control
No data.
Emission
Control Device
a
Atmosphere
High pressure
absorber
Atmosphere
Low pressure
absorber
Atmosphere
device.
Pollutant
Acetic acid
p_-xylene
Acetic acid
Methyl acetate
£-xylene
Acetic acid
Acetic acid
Methyl acetate
Acetic acid
Methyl acetate
p-xylene
Before BCD
0.04
v>
N.D.
2.1
0.23
N.D.
Trace
4.1
Trace
After ECD
0.04
0.72
9.0
2.95
2.1
0.01
Trace
Trace
4.1
Trace
ft]
I
-------
E-12
The inorganic portions of the catalyst, the by-products and residues formed in
the reaction and distillation sections, and the unrecoverable portions of the
product are carried through the process in the liquid phase and are ultimately
discharged as the residue from the residue still. This stream contains some
acetic acid, which is disposed of in a rotary kiln incinerator.
2
4. Du Pont and Co., Cape Fear, NC, and Old Hickory, TN
The emission factor data presented here represent annual averages for the
combined TPA and DMT processes at each location.
TPA (Ib of VOC/CWT) DMT (Ib of VOC/cwt)
High-pressure absorber vents 2.03
Atmospheric absorbers 0.027
Silo bag filters 0.21
TPA process incinerators Neg.
Methanol column vents 0.037
Vacuum-jet condenser vents 0.035
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-13
C.
REFERENCES*
1. S. W. Dylewski, IT Enviroscience, Inc., Trip Report for Visit to Amoco Chemical
Corp., Decatur, AL, Oct. 31, 1977 (on file at EPA, ESED, Research Triangle Park,
NC).
2. D. W. Smith, E. I. du Pont de Nemours and Co., letter dated Oct. 20, 1978,
in response to EPA's request for information on emissions from TPA/DMT
production facilities.
3. S. W. Dylewski, IT Enviroscience, Inc., Trip Report for Visit to Carolina
Eastman Co., Columbia, SC, Dec. 6, 7, 1977 (on file at EPA, ESED, Research
Triangle Park, NC).
4. S. W. Dylewski, IT Enviroscience, Inc., Trip Report for Visit to Hereofina Co.,
Wilmington, NC, Nov. 17, 1977 (on file at EPA, ESED, Research Triangle Park,
NC).
5. J. C. Edwards, Tennessee Eastman Co., letter dated Aug. 31, 1978, in response
to EPAs request for information on TPA/DMT production facilities.
6. L. M. Elkin, Terephthalic Acid and Dimethyl Terephthalate, pp 49—55 in
Report No. 9, A private report by the Process Economics Program, Stanford,
Research Institute, Menlo Park, CA (February 1966).
^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.
-------
6-i
REPORT 6
PHENOL/ACETONE
C. W. Stuewe
F. D. Hobbs
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
February 1981
This report contains certain information whirh 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.
D106A
-------
6-iii
CONTENTS OF REPORT 6
I. ABBREVIATIONS AND CONVERSION FACTORS 1-1
II. INDUSTRY DESCRIPTION II-1
A. Reasons for Selection II-l
B. Acetone Usage and Growth II-l
C. Phenol Usage and Growth II-3
D. Domestic Producers II-3
E. References 11-12
III. PROCESS DESCRIPTION III-l
A. Introduction III-l
B. Cumene Peroxidation Processes III-l
C. Other Commercial Phenol Processes 111-12
D. Other Commercial Acetone Processes 111-13
E. References 111-14
IV. EMISSIONS IV-1
A. Process via Allied Technology IV-1
B. Process by Hercules Technology IV-6
C. References IV-12
APPLICABLE CONTROL SYSTEMS V-l
A. Process via Allied Technology V-l
B. Process by Hercules Technology V-4
C. References V-8
VI. IMPACT ANALYSIS VI-1
A. Environmental and Energy Impacts VI-1
B. Cost Control Impact VI-5
C. References VI-8
VII. SUMMARY VII-1
-------
G-V
APPENDICES OF REPORT 6
A. PHYSICAL PROPERTIES OF ACETONE, CUMENE, AND PHENOL A-l
B. FUGITIVE-EMISSION FACTORS B-l
C. EXISTING PLANT CONSIDERATIONS C-l
D. COST ESTIMATE PROCEDURE FOR PROCESS EMISSION CONTROL D-l
WITH CARBON ADSORPTION
-------
6-vii
TABLES OF REPORT 6
Number
II-l Acetone Usage and Growth II-2
II-2 Phenol Usage and Growth II-4
II-3 Acetone Capacity II-5
II-4 Phenol Capacity I]:~7
III-l Phenol Plants Using Allied and Hercules Licensed Process III-2
Technology
IV-1 Total Uncontrolled VOC Emissions from a Model Plant Using Allied IV-2
Technology
IV-2 Estimated Composition of Oxidation Vent Gas from Model Plant IV-4
Using Allied Technology
IV-3 Storage Requirements for 200,000-Mg/yr Model Plant Using Allied IV-5
Technology
IV-4 Total Uncontrolled VOC Emissions from a Model Plant Using IV-8
Hercules Technology
IV-5 Estimated Composition of Oxidation Vent Gas from Model Plant IV-9
Using Hercules Technology
IV-6 Storage Requirements for 200,000-Mg/yr Model Plant Using IV-11
Hercules Technology
V-l Estimates of Controlled VOC Emissions from a Model Plant Based V-2
on Allied Technology
. - 2 Estimates of Controlled VOC Emissions from a Model Plant Based V-5
on Hercules Technology
VI-1 Environmental Impact of Controlled Model Plant Using Allied VI-2
Technology
VI-2 Environmental Impact of Controlled Model Plant Using Hercules VI-4
Technology
VI-3 Summary of Costs and Cost Effectiveness for Carbon Adsorption VI-7
Applied to Allied and Hercules Model Plants
VII-1 Emission Summary for the Model Plant Using Allied Technology VII-2
VII-2 Emission Summary for the Model Plant Using Hercules Technology VII-3
A-l
A-l Properties of Acetone
A- 2
A-2 Properties of Cumene
A— 3
A-3 Properties of Phenol
C-l Control Devices and Techniques Reported by Existing Plants
c~2
-------
6-ix
FIGURES OF REPORT 6
Number p-^2e_
II-l Locations of Plants Manufacturing Acetone II-6
II-2 Locations of Plants Manufacturing Phenol II-Q
III-l Flow Diagram for Phenol/Acetone from Cumene Using Allied III-4
Technology
III-2 Flow Diagram for Phenol/Acetone from Cumene Using Hercules III-8
Technology
-------
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 I0"e
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
Symbol
tera
giga
mega
kilo
milli
micro
Multiplication
Factor
1012
ior
106
103
io"3
io"6
Example
1 Tg = 1 X
grams
1 Gg = 1 X IO9 grams
1 Mg = 1 X IO6 grams
1 km = 1 X IO3 meters
1 mV = 1 X IO"3 volt
1 |jg = 1 X IO"6 gram
-------
II-l
II. INDUSTRY DESCRIPTION
A. REASON FOR SELECTION
Production of acetone and phenol was selected for study because their manufacture
results in significant emissions of volatile organic compounds (VOC). A major
portion of both acetone and phenol domestic production is based on the cumene
peroxidation process. As of 1978, 67% of domestic acetone production was based
on this process, with most of the remainder being derived from isopropyl
alcohol.1 A small amount of listed capacity is derived as by-product of other
products, including 2-naphthol, hydroquinone, and propylene oxide.2 As of 1978,
94% of the listed domestic synthetic phenol capacity was based on the cumene
peroxidation process, with the remaining synthetic phenol capacity being based on
the benzene sulfonation process and the toluene oxidation process.3 A small
amount (less than 2% of the total domestic production in 1974) of phenol, called
t\
natural phenol, is recovered from coal tar and petroleum streams.
VOC emissions from the cumene peroxidation process include acetone, cumene,
phenol, acetaldehyde, and a-methylstyrene. VOC emissions from the isopropyl
alcohol process include acetone and isopropyl alcohol. Acetone constitutes the
major VOC in emissions from both the cumene peroxidation process and the iso-
propyl alcohol process because of the volatility of that VOC (see Appendix A for
pertinent physical properties).
Although the isopropyl alcohol process is included in the above discussion for
completeness, the subject of this report is the cumene peroxidation route to
phenol and acetone. In the following sections processes other than cumene per-
oxidation are described only briefly, and discussions of emissions, emission
controls, and control impacts are exclusively devoted to the cumene peroxidation
process.
B. ACETONE USAGE AND GROWTH
Table II-l shows the acetone end products, the percentages of total consumption,
and the projected growth rates. The largest single consumption of acetone is in
production of methyl methacrylate, which is converted to acrylic sheet. The next
largest acetone consumer is methyl isobutyl ketone production, but this use is
declining because of environmental legislation restrictions on the use of methyl
-------
II-2
Table II-l. Acetone Usage and Growth
Average Annual
End Use 1977 Production (%) Growth (%) 1977-1982
Methyl methacrylate 25 7.0-8.0
Methyl isobutyl ketone 9 (2.5-3.5)
Bisphenol A 6 10.0-11.0
Methacrylic acid and higher 5 7.0-8.0
methacrylate
Methyl isobutyl carbinol 2 0.0-2.0
Aldol chemicals 9 2.5-3.5
Solvent uses 22 3.0-3.5
Miscellaneous 22 3.0-3.5
See ref 1.
-------
11-3
isobutyl ketone as a solvent. Consumption for production of bisphenol A, which
is used for epoxy and polycarbonate resins manufacture, is expected to increase
rapidly.4
Domestic acetone capacity in 1978 was reported3 to be about 1326 Gg/yr, with
reported5 1978 production utilizing about 71% of that capacity. Production would
reach 83—86% of current capacity by 1982 based on the projected4 4 to 5% annual
growth rate.
C. PHENOL USAGE AND GROWTH
Table II-2 shows the phenol end products, the percentages of total consumption,
and the projected growth rates.
The largest consumer of phenol is phenolic resins, which are used as adhesives.
The second largest use of phenol is an intermediate for bisphenol A, which is
used in the manufacture of epoxy resins. Large amounts of phenol are used to
manufacture cyclohexanone, which is converted to caprolactam through a series of
reactions. Caprolactam is used in the production of nylon fibers.6
Domestic phenol capacity in 1978, including natural phenol, was reported3 to be
1624 Gg/yr, with 1978 production7 utilizing about 77% of that capacity. Production
would reach about 92% of current capacity by 1982 based on the projected8 4.5%
annual growth rate.
D. DOMESTIC PRODUCERS
As of the end of 1977 there were 15 producers of acetone, as listed in Table II-3,
at the plant locations shown in Fig. II-l and 12 producers of synthetic phenol,
listed in Table II-4, at the plant locations shown in Fig. II-2. Six producers
separate natural phenol from coal tar and petroleum; they are listed in Table II-4
but are not shown in Fig. II-2. Following are brief descriptions of those com-
panics producing acetone and phenol.
1, Allied
Acetone and phenol are produced by cumene peroxidation. Phenol is used in the pro-
duction of adipic acid and cyclohexanone for caprolactam. Some phenol is sold.
-------
II-4
Table II-2. Phenol Usage and Growth0
End Use
Phenolic resins
Bisphenol A
Caprolactam
Nony Ipheno 1
Salicylic acid
Dodecylphenol
Adipic acid
Miscellaneous
1977 Production (%)
44
17
15
2
1
I
1
19
Average Annual
Growth (%) 1977-1982
3.5-4.5
10.0-11.0
5.0-5.5
4.0-5.0
2.5-4.5
1.5-2.5
1.5-2.5
4.0-5.0
See ref 6.
-------
II-5
Table II-3. Acetone Capacity
Plant
Location
1978
Capacity
(Mg, X 103)
Process
Allied
American Cyanamid
Clark Oil
Dow
Eastmen Kodak
Exxon
Georgia-Pacific
Getty Oil
Goodyear
Monsanto
Oxirane
Shell
Standard Oil
Tin ior Carbide
United States Steel
Total
a.
Frankford, PA
Willow Island, WV
Blue Island, IL
Oyster Creek, TX 4
Kingsport, TN
Bayway, NJ
Plaquemine, LA
El Dorado, KS
Bayport, TX
Chocolate Bayou, TX
Bayport, TX
Deer Park, TX
Deer Park, TX
Dominquez, CA
Richmond, CA
Bound Brook, NJ
Institute and
South Charleston, WV
Penuelas, PR
Haverhill, OH
163
5
24
127
36
63
71
25
5
136
18
136
181
45
15
50
77
59
90
1326
Cumene peroxidation
2-Naphtol by-product
Cumene peroxidation
Cumene peroxidation
Isopropyl alcohol
Isopropyl alcohol
Cumene peroxidation
Cumene peroxidation
Hydroquinone by-product
Cumene peroxidation
Propylene oxide by-product
Cumene perioxidation
Isopropyl alcohol
Isopropyl alcohol
Cumene peroxidation
Cumene peroxidation
Isopropyl alcohol
Cumene peroxidation
Cumene peroxidation
See ref 2.
-------
II-6
1. Allied Chemical Corp., Frankford, PA
2. American Cyanamid Co., Willow Island, WV
3. Clark. Oil s Refining Corp., Blue Island, IL
4. Dow Chemical Co., Oyster Creek, TX
5. Eastman Kodak Co., Kingsport, TN
6. Exxon Corp. , Bayway, NJ
7. Georgia-Pacific Corp., Plaquemine, LA
8. Getty Oil Co., El Dorado, KS
9. Goodyear Tire & Rubber Co., Bayport, TX
10. Monsanto Co., Chocolate Bayou, TX
11. Oxirane Corp., Bayport, TX
12. Shell Chemical Co., Deer Park, TX
13. Shell Chemical Co., Dominquez, CA
14. Standard Oil Co. of CA, Richmond, CA
15. Union Carbide Corp., Bound B ook, NJ
16. Union Carbide Corp., Institute and South Charleston, WV
17. Union Carbide Corp., Penuelas, PR
18. United States Steel Corp., Haverhill, OH
Fig. II-l. Locations of Plants Manufacturing Acetone
-------
Table II-4. Phenol Capacity
Plant
Location
1978 Capacity (Mg)
(X 103)
Process
Allied Chemical
Clark
Dow
Fallik Chemical
Ferro
Georgia-Pacific
Getty Oil
Kalama
Koppers
Merichem
Monsanto
Reichhold
Shell
Standard Oil
Stimson
Union Carbide
U.S. Steel Corp.
Total
Frankford, PA
Blue Island, IL
Oyster Creek, TX
Tuscaloosa, AL
Santa Fe Springs, CA
Plaquemine, LA
El Dorado, KS
Kalama, WA
Follansbee, WV
Houston, TX
Chocolate Bayou, TX
Tuscaloosa, AL
Deer Park, TX
Richmond, CA
Anacortes, CA
Bound Brook, NJ
Penuelas, PR
Clairton, PA
Haverhill, OH
272
40
211
b
c
118
43
34
b
b
227
70d
227
25
b
82
100
c
148*
1,624
Cumene peroxidation
Cumene peroxidation
Cumene peroxidation
Unknown (natural phenol)
Coal tar and petroleum
Cumene peroxidation
Cumene peroxidation
Toluene oxidation
Coal tar
Petroleum
Cumene peroxidation
Benzene sulfonation
Cumene peroxidation
Cumene peroxidation
Petroleum
Cumene peroxidation
Cumene peroxidation
Coal tar
Cumene peroxidation
See ref 3.
bThese four plants combined have a natural-phenol capacity of about 27 X 10 Mg/yr, which is included in total capacity.
-i
"Not available.
^Closed; placed on standby in March 1978.
eCapacity recently increased by 195 X 10 Mg/yr.
-------
1. Allied Chemical Corp., Frankford, PA
2. Clark Oil & Refining Corp., Blue Island, IL
3. Dow Chemical Co., Oyster Creek, TX
4. Georgia-Pacific Corp., Plaquemine, LA
5. Getty Oil Co., El Dorado, KS
6. Kalama Chemical Co., Kalama, WA
7. Monsanto Co., Chocolate Bayou, TX
8. Reichhold Chemicals, Inc., Tuscaloosa, AL
9. Shell Chemical Co., Deer Park, TX
10. Standard Oil Co. of CA, Richmond, CA
11. Union Carbide Corp., Bound Brook, NJ
12. Union Carbide Corp., Penuelas, PR
13. United States Steel Corp., Haverhill, OH
Fig. II-2. Locations of Plants Manufacturing Phenol
-------
II-9
2. American Cyanamid
Acetone is produced as a by-product of 2-naphthol.2
3. Clark Oil
Acetone and phenol are produced by cumene peroxidation. Phenol is mainly sold,
but some is used in production of phenolic resins.6
4. Dow
Acetone and phenol are produced by the cumene peroxidation process.6
5. Eastman Kodak
Acetone is produced from isopropyl alcohol.2
6. Exxon
Acetone is produced from isopropyl alcohol.2
/ Fallek
Natural phenol is recovered from an unreported feed stock.3
8. Ferro
Natural phenol is recovered from coal tar and petroleum streams.3
°. General Electric
General Electric, which is not listed in the tables of producers as a current
producer, plans to build a cumene-based phenol/acetone plant with 100-Gg/yr
acetone capacity and 181-Gg/yr phenol capacity at Mount Vernon, IN, beginning in
I960.8—10
10. Georgia-Pacific
Acetone and phenol are produced by cumene peroxidation. About 50% of the phenol
is sold and the remainder is consumed in t'.ie production of phenolic resins.6
ji. Getty Oil
Acetone and phenol are produced by cumene peroxidation.2'3
-------
11-10
12. Goodyear
Acetone is produced as a by-product of hydroquinone.2
13. Gulf Oil
Gulf Oil, which is not listed in the tables as a current producer, plans to have
a plant completed in 1981s'9 that will have an acetone capacity of 136 Gg/yr and
a phenol capacity of 227 Gg/yr.
14. Kalama
Phenol is produced by toluene oxidation.3 It was reported8 that the capacity
would be expanded by 9 Gg/yr in 1978.
15. Koppers
Natural phenol is separated from coal tar.3
16. Merichem
Natural phenol is separated from petroleum.3
17. Monsanto
Acetone and phenol are produced by cumene peroxidation. The phenol is used as an
intermediate for manufacture of a number of different chemicals and also is
sold.*
IS. Oxirane
Acetone is produced as a by-product of propylene oxide.2
19. Reichold
Phenol was produced by benzene sulfonation to produce phenolic resins, penta-
chlorophenol, and miscellaneous chemicals, as well as for sale.6 This capacity
was placed on standby in March 1978.
?• , Shell
Acetone and phenol are produced by cumene peroxidation at the Deer Park, TX,
plant.3 Acetone is produced from isopropyl alcohol at Deer Park, TX, and
Dominguez, CA. A new acetone plant with 136-Gg/yr capacity is due to be com-
pleted at Wood River, IL, in 1979.9 Acetone was produced by oxidation of iso-
-------
11-11
propyl alcohol at the Norco, LA, plant, but that process has been permanently
shut down. All acetone produced from isopropyl alcohol by Shell is now produced
by the dehydrogenation process.11
21. Standard Oil
Acetone and phenol are produced by cumene peroxidation. Phenol is used for the
manufacture of alkylphenols. Some of the phenol is sold.
6
22. Stimson Lumber
Natural phenol is separated from petroleum.3
23. Union Carbide
Acetone and phenol are produced by cumene peroxidation at the Bound Brook, NJ,
and Penuelas, PR, plants.3 Acetone also is produced from isopropyl alcohol at
Institute and South Charleston, WV.2
24 United States Steel
Natural phenol is separated from coal tar at Clairton, PA. Acetone and phenol
are produced by cumene peroxidation at Haverhill, OH. Phenol capacity was in-
creased by 90,000 Mg/yr in 1979.1X
-------
11-12
E. REFERENCES*
I. S. A. Cogswell, "Acetone," p 604.5032A in Chemical Economics Handbook, Stanford
Research Institute, Menlo Park, CA (July 1978).
2. "Chemical Information Services," pp 419 and 420 in 1979 Directory of Chemical
Producers, United States of America, SRI International, Menlo Park, CA (1979).
3. Ibid., p 807.
4. S. A. Cogswell, "Acetone," pp 604.5031C-D in Chemical Economics Handbook,
Stanford Research Institute, Menlo Park, CA (July 1978).
5. "Acetone," p 228 in Chemical Economics Handbook, Manual of Current Indicators
Supplemental Data, Chemical Information Services, Stanford Research Institute,
Menlo Park, CA (October 1979).
6. S. A. Cogswell, "Phenol," pp 686.5021A—686.5023J in Chemical Economics Handbook,
Stanford Research Institute, Menlo Park, CA (October 1978).
7. "Phenol," p 274 in Chemical Economics Handbook, Manual of Current Indicators
Supplemental Data, Chemical Information Services, Stanford Research Institute,
Menlo Park, CA (October 1979).
8. "Chemical Profile on Phenol," p 9 in Chemical Marketing Reporter (Feb. 6, 1978).
9. "Chemical Profile on Acetone," p 9 in Chemical Marketing Reporter (Nov. 21,
1977).
10. "Chemical Information Services," 1979 Directory of Chemical Producers Supple-
ment II, SRI International, Menlo Park, CA.
ii. J. Beale, Chemical Manufacturers Association, letter dated Nov. 14, 1980, to
Robert E. Rosensteel, 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 DESCRIPTION
A. INTRODUCTION
In the United States 97%1/2 of the phenol is manufactured by the peroxidation
of cumene followed by cleavage of the resulting cumene hydroperoxide (CHP).
The two basic reactions of the cumene route to phenol and acetone are as
follows:
1. C6H5CH(CH3)2 + 02 - * C6H5COOH(CH3)2
(cumene) (air) (cumene hydroperoxide)
2. C6H5COOH(CH3)2 > CeHSOE + CH3COCH3
(cumene hydroperoxide) (phenol) (acetone)
In the peroxidation reaction, as practiced commercially, relatively pure3
(•v99.8%) cumene manufactured on-site or shipped to the site is reacted with
oxygen in air in an autocatalytic4 liquid-phase reaction to form CHP. The re-
action is exothermic (about 1000 kJ/kg of cumene4). Impurities in the cumene
result in increased by-product formation, such as acetaldehyde, methyl ethyl
ketone, and propionaldehyde . These by-products are usually3 vented.
In the second reaction the CHP product of the peroxidation reaction is cleaved
to phenol and acetone in the presence of dilute sulfuric acid. The acid promotes
this exothermic (about 2700 kJ/kg of phenol) decomposition reaction,4 which is
extremely fast and temperature dependent. After cleavage, the acid in the cleav-
age product is neutralized and the products and by-products are separated in a
series of distillation columns. In addition to the products phenol and acetone,
a-methyl styrene and acetophenone are recovered as by-products by some producers.
B. CUMENE PEROXIDATION PROCESSES
At the present time about 47% of the installed phenol capacity using the cumene
route is based on process technology licensed by Allied Chemical. The remaining
capacity uses processing technology licensed by Hercules (see Table III-l).
The major differences between the Allied and Hercules processes involve the
-------
Table III-l. Phenol Plants Using Allied and Hercules Licensed Process Technology
Plant
Allied Chemical
Clark Oil and
Refining
Dow Chemical
a
Getty Oil Co.
Union Carbide
Location
_ n
Frankfort, PA
Blue Island, IL
Oyster Creek, TX
El Dorado, KS
Bound Brook, NJ
Penuelas, PR
Formerly Skelly Oil Co, ?any.
Hercules Technology
1978 Capacity
(Gq)
272
40
211
43
82
100
748
Plant
Georgia Pacific
Monsanto
Shell Oil
Standard Oil of
California
U. S. Steel Corp.
Total
Location
Plaquemine , LA
Chocolate Bayou, TX
Deer Park, TX
Richmond, CA
Haverhill, OH
1978 Capacity
(Gg)
118
227
227
25
b
236
833
H
I
1979 capacity.
-------
III-3
operating conditions of the peroxidation reaction and the method of neutraliza-
tion of the acid in the cleavage product. These differences affect the plant
design primarily in the peroxidation and cleavage-product neutralization steps,
in the location of process emission points, and in the potential quantity of
process emissions.
1. Allied Process
Figure III-l is a typical flowsheet for the manufacture of phenol and acetone
by the Allied process. Cumene (1)* manufactured on-site or shipped to the site
and recycle cumene (2) are combined and fed with air to the multiple-reactor
system connected in series. The Allied process operates at relatively low tem-
peratures and pressures (compared with those used in the Hercules process) and
uses no catalyst or alkaline buffer in the oxidation step.5 Cooling is re-
quired for this exothermic reaction step. Substantial quantities of cumene (5)
are carried out of the reactors with the spent air, which contains about 5 vol
% 02- Part of the cumene is recovered and recycled from a refrigerated vent
system operated at about 5°C and atmospheric pressure.
The reaction product (6), containing primarily cumene and CHP, is flashed in
the CHP concentration column under vacuum to remove most of the cumene, which
is recycled. The concentrated CHP (8) flows through the CHP concentrate tank
to the cleavage reactor. The cleavage product (10) is neutralized in ion-
exchange columns and fed through the crude-product surge tank to a multicolumn
distillation system.
The distillation system shown on Fig. III-l is illustrative of the Allied
process5 and recovers, in addition to phenol and acetone, by-products a-methyl
styrene and acetophenone. In the crude-acetone column acetone and lower boil-
ing impurities such as acetaldehyde and formaldehyde are distilled overhead.
This product (12) flows through the crude-acetone surge tank to the acetone
finishing column, where the acetone is distilled overhead to product quality.
Acetone product (14) is accumulated in the acetone day tanks and stored in the
acetone storage tank for subsequent loading.
*Such numbers in parentheses refer to the streams shown on Figs. III-l and
III-2,- capital letters refer to emission sources.
-------
?
ACETOUS
FWOOOCT
TO LOACH U3
-TO CRUDE.
CUMEkJE
,e.cov£«y
COV.UMU
Fig. III-l. Flow Diagram for Phenol/Acetone from Cumene Using Allied Technology
Page 1 of 2
-------
TAUK
Fig. III-l. (Continued)
Page 2 of 2
-------
III-6
Bottoms (13) from the crude-acetone column are distilled to remove cumene (16),
which, after being washed with dilute caustic to convert phenol to an aqueous
phenate solution for removal, is recycled.
The bottoms (17) from the cumene recovery column contain primarily phenol, AMS,
acetophenone, and other organics (heavy ends) with higher boiling points than
phenol and are fed to the crude-AMS column. The crude-AMS column overhead stream
(18) is washed with caustic to convert phenol to an aqueous phenate stream for
removal, flows (19) through the crude-AMS storage tank to the AMS refining column,
is distilled overhead (21) from the AMS refining column, and is then stored in
the AMS product tanks. Bottoms (22) from the AMS refining column, containing
higher boiling hydrocarbons, are purged to on-site fuel uses.
Crude phenol (20) from the bottom of the crude-AMS column flows to the phenol
refining column, where phenol is distilled overhead (23) to the phenol-product
day tanks. The product is stored in the phenol storage tank for subsequent
loading.
Bottoms (24) from the phenol refining column are fed to the heavy-ends column,
where primarily acetophenone with impurities such as AMS and some dimethyIphenyl
carbinol is distilled overhead (26) and the higher boiling ends such as para-
alpha-cumylphenol, dimers of AMS, and tars exit (25) from the bottom of the
column. This tarry product is stored in the tars tank and sold or used as heavy
fuel oil.
Acetophenone is separated as the bottoms product (28) of the acetophenone column
and stored in the acetophenone tank for loading. The overhead stream (27) from
the acetophenone column is recycled to recover the AMS content and to remove
the phenol impurity.
The main process vent (A) is associated wi ,-h the spent-air stream from the air
oxidation reaction. Nitrogen and unused oxygen, which are vented at approxi-
mately atmospheric pressure, carry out a mixture of hydrocarbons, predominantly
cumene.
-------
III-7
The second process vent (B) is associated with the vacuum jet on the accumulator
of the CHP concentration column. Inert gases, primarily nitrogen, dissolved in
the oxidation reaction product (6) are stripped and vented along with cumene,
primarily.
The third process vent (C) is associated with the accumulator on the crude-
acetone column. Low-boiling hydrocarbons such as acetaldehyde and formaldehyde
formed during the two reaction steps are vented, along with some acetone.
The fourth process vent (D) is associated with the acetone finishing column.
The VOC in the vent stream is acetone.
The final process vent (E) is associated collectively with the vacuum jets from
the remaining six distillation columns in the distillation system. Unreacted
ethylbenzene and toluene introduced with the cumene feed, as well as the other
VOC products and by-products, are vented. Contaminated wastewater streams (K)
result (1) from dilute caustic washes of recycle cumene to remove acidic and
phenolic components, which may cause degradation of the product or inhibit the
reaction rate in the peroxidation step, (2) from the caustic regeneration of
the ion-exchange columns, (3) from wash of the crude-AMS recycle to remove the
phenol contaminant as phenate before it is distilled, and (4) from the bottoms
from the acetone refining column.
Hercules Process
Figure III-2 is a typical flowsheet for the manufacture of phenol and acetone
by the Hercules process.
Cumene from storage (1) and recycle cumene (2) are combined and then fed with
air (4) to the multiple-reactor system connected in series.6'7 Additionally,
an aqueous Na2C03 solution (3) is fed to the reactor system to promote the peroxi-
dation reaction.7 This oxidation step is operated at about 95°C and 6.5 X 105 Pa
(ref 8). The spent air (5) exiting from the reactors contains about 5 vol % oxygen.
Cumene vaporized and flushed from the reactors with the spent air provides cooling
for this reaction step. Most of the cumene is recovered and recycled from a
refrigerated vent system8 operated at about 5°C and 5.9 X 105 Pa.
-------
4^&
CUMEUE
^e^.
HtO
FEED /
CLEAVAGE.
PRODUCT
CRUDE
PHOOUCT
.. TO VAEAW- •
C.OL.UMVJ
CRUDE CRyoe
PHEMOL / ACE-TOU6. >
COUUMNJ T>WK
Fig. IH-2. Flow Diagram for Phenol/Acetone from Cumene Using Hercules Technology
Page 1 of 2
-------
M
M
COL.UMU
Fig. III-2. (Continued)
Page 2 of 2
-------
111-10
The oxidation reaction product (6) flows into a separator to remove the spent
carbonate solution7 and then is washed with water to remove remaining carbonate
and other soluble components. The separation and wash steps are operated at
close to atmospheric pressure; as a result the reaction product is degassed
before it is concentrated. The degassed product (8) is concentrated in a column
operated under vacuum to minimize thermal decomposition of the CHP to dimethyl-
phenylcarbinol (DMPC). The recovered cumene (9) is recycled and the concentrate
(10) is transferred through a surge tank to an agitated9 reactor. Sulfuric
acid, diluted to 5 to 10% with acetone,10 is added to catalyze the decomposi-
tion of CHP to phenol and acetone. The heat of reaction is removed by acetone
being vaporized at the controlled operating pressure and temperature.
Excess acid in the cleaved mixture (11) is neutralized with sodium hydroxide
solution. The neutralized stream (12) flows through the crude-product surge
tank to an 8-column distillation train to produce product-grade phenol, acetone,
and AMS.
In actual practice the operating conditions and the separation sequence of the
distillation system vary from plant to plant, depending on the product mix,
impurities, and mass-transfer operation preferences. The separation sequence
shown in Fig. III-2 is believed to be similar to those used in practice.
The crude product is separated in the first distillation column into a crude
acetone fraction (13) and a crude phenol stream (14). The crude acetone is
combined with recycled HC (25) from the phenol topping column and fed to the
light-ends column to strip low-boiling HC impurities, such as acetaldehyde and
formaldehyde, which are vented. The bottoms stream (16) from the light-ends
column is fed to the acetone finishing column, which is operated under vacuum.
The acetone product (18) is taken overhead to the acetone day tanks and subse-
quently to acetone product storage and loading. The bottoms stream (17) is
washed with dilute sodium hydroxide and dec mted to remove any phenolic impuri-
ties as the phenates.
The washed stream (19) flows through a surge tank to the AMS topping column.
A light-oil fraction (20), consisting of unreacted ethyl benzene and toluene
introduced with the cumene raw material and other impurities (e.g., mesityloxide)
-------
III-ll
is removed overhead and used on-site for its fuel value. An impure-cumene
stream (21) is removed and recycled, and AMS product (22) is transferred to
storage.
The crude-phenol stream (14) from the crude phenol/acetone column and the bottoms
(28) from the phenol finishing column are fed to the heavy-ends column and
distilled under vacuum to separate tars (23) from the impure-phenol stream (24).
Hydrocarbons in the tar stream (e.g., cumyl phenols, AMS dimers, acetophenone,
DMPC, and phenate9) are used as heavy fuel oil for their fuel content4 (about
37 MJ/kg).
The impure phenol (24) is fed to the phenol topping column to remove hydrocar-
bons such as cumene and AMS, which remained with the crude phenol stream (14),
and AMS formed by dehydration of the DMPC component in the heavy-ends-column
feed stream. The phenolic stream (26) is then fed to a dehydrating column,
where water is removed overhead as a phenol/water azeotrope.
The dried-phenol stream (27) is distilled under vacuum in the phenol finishing
column to separate product-quality phenol (29) from higher boiling components,
which are recycled (28).
The main process vent (A) is associated with the spent air stream from the per-
oxidation reaction following the refrigerated condenser system. Nitrogen, unused
oxygen, and a mixture of HC, predominantly cumene, are vented.
Three process vent points (B, C, and D) are associated with the oxidate washer,
CHP concentrator, and CHP cleavage reactor. Vents B and C emit cumene primarily,
with vent gases desorbed from the oxidation reaction product as the operating
pressure is decreased. Vent D emits acetone from the refrigerated condenser on
the cleavage reactor.
Another process vent (E) is associated with the accumulator on the light-ends
column. Low-boiling hydrocarbons (e.g., acetaldehyde) formed during the two
reaction steps are vented, along with some acetone.
-------
111-12
Another process vent (F) is associated with the accumulator on the acetone
finishing column; the VOC is acetone.
The final process vent (G) is associated collectively with the other five dis-
tillation columns and emits a mixture of hydrocarbons.
Contaminated wastewater streams (K) result (1) from separation of the spent
carbonate and oxidate wash solution, (2) from dilute caustic washes to neutral-
ize excess cleavage acid and to remove phenolic impurities in the crude-AMS
stream, and (3) from water removed in the phenol dehydrating column.
3. Process Variations
There are many possible variations in operating conditions and procedures that
will influence the types and quantities of emissions. One example is that the
excess oxygen in the spent air can be varied and will directly affect the quan-
tity of spent air and thus the VOC emission rate from the main process vent
(A).
Another variation that could greatly reduce the emissions from vent A would be
the use of oxygen instead of air in the oxidation step, thereby greatly reduc-
ing the inert-gas venting. However, the use of oxygen would increase the explo-
sion hazard and is reportedly10 not economical. None of the old or newer
plants for which detailed process data were secured5—8 use oxygen instead of
air. Both the Allied and Hercules process technologies are based on the use of
air in the cumene oxidation step.
Another process variation is the hydrogenation of the crude-AMS stream to
produce cumene for recycle rather than to produce an AMS product for sale.
This variation would result in a higher yield of phenol and acetone from the
cumene raw material and change the emission points and emissions associated
with AMS product distillation and storage.
C, OTHER COMMERCIAL PHENOL PROCESSES
The only commercial route to phenol in the United States today other than
cumene peroxidation is by toluene oxidation. About 2% of the synthetic phenol
is produced by the toluene process. In this process toluene is oxidized, by
-------
111-13
air in the liquid phase at elevated temperature and pressure (160°C/ 5 X 105 Pa)
in the presence of cobalt acetate catalyst, to benzoic acid. Following separa-
tion, the benzoic acid is catalytically converted to phenol in a liquid-phase
oxidative decarboxylation reaction with air at elevated temperature (240°C) and
atmospheric pressure.2
The only plant producing phenol by benzene sulfonation was reportedly closed
and put on standby as of March 1978. This process involves reacting benzene
and concentrated sulfuric acid to form benzene sulfonic acid, which is then
reacted with sodium sulfite to form sodium benzene sulfonate. The sulfonate is
fused with sodium hydroxide to form sodium phenate, which is acidified with
sulfur dioxide in the presence of sulfuric acid to form phenol.2
D. OTHER COMMERCIAL ACETONE PROCESSES
The only commercial process used in the United States other than the cumene
peroxidation route that produces and separates acetone as a product is based on
catalytic dehydrogenation of isopropyl alcohol (IPA). In this process IPA is
catalytically dehydrogenated to acetone in a vapor-phase reaction at 400 to
500°C.
-------
111-14
E. REFERENCES*
1. "No Switch from Cumene, Say Phenol Manufacturers," Chemical Engineering 86(8),
64 (Apr. 9, 1979). —
2. S. A. Cogswell, "Phenol," pp 686.5021A— 686.5023J, in Chemical Economics Hand--
book, Stanford Research Institute, Menlo Park, CA (October 1978).
3. Yen-chen Yen, Report No. 22, Phenol, A private report by the Process Economics
Program, Stanford Research Institute, Menlo Park, CA (April 1967).
4. P. R. Pujado, J. R.. Salazar, and C. V. Berger, "Cheapest Route to Phenol,"
Hydro-carbon Processing 55(3), 91—96 (1976).
5. C. W. Stuewe, IT Enviroscience, Trip Report for Visit to Allied Chemical
Corp., Philadelphia, PA, Mar. 16, 1978 (on file at EPA, ESED, Research Triangle
Park, NC).
6. C. W. Stuewe, IT Enviroscience, Trip Report for Visit to Monsanto Chemical
Intermediates Co., Alvin, TX, July 28, 1977 (on file at EPA, ESED, Research
Triangle Park, NC)..
7. C. W. Stuewe, IT Enviroscience, Trip Report for Visit to Georgia Pacific Corp.,
Plaquemine, LA, Aug. 2, 1977 (on file at EPA, ESED, Research Triangle Park, NC)
8. Shell Oil Co./Shell Chemical Co., Deer Park, TX, Texas Air Control Board Permit
Application for Phenol-2 as revised May 9, 1975.
9. Yen-chen Yen, Report No. 22A. Phenol Supplement A, A private report by the
Process Economics Program, Stanford Research Institute, Menlo Park, CA
(September 1972).
30. J. L. Delaney, and T. W. Hughes, Monsanto Research Corp., Source Assessment
Manufacture of Acetone and Phenol from Cumene, EPA-600/2-79-019d, (May 1979).
^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.
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 are photochemically unreactive. However, many
photochemically unreactive organic chemicals are of concern and may not be
exempt from 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. PROCESS VIA ALLIED TECHNOLOGY
1. Model Plant*
The model plant for the synthesis of phenol and acetone from cumene using
Allied Chemical licensed technology has a phenol capacity of 200,000 Mg/yr and
an acetone capacity of 120,000 Mg/yr based on 8760 hr** of operation annually.
These capacities are typical of recently built or announced plants manufacturing
phenol and acetone from cumene. In addition, 10,500 Mg of AMS and 3,750 Mg of
acetophenone are recovered annually as by-products. The process shown in
Fig. III-l is believed to be typical of actual processes using Allied techno-
logy,- however, not all plants recover the AMS and acetophenone by-products.
2. Sources and Emissions
Uncontrolled emission sources and rates are summarized in Table IV-1 and are
further described below.
Cumene Oxidation—Spent air vented (A, Fig. III-l) from the cumene oxidation
reactors following the refrigerated condenser system is the most significant
*See page 1-2 for a discussion of model plants.
**Process downtime is normally expected to range from 5 to 15%'
the error introduced by assuming continuous operation is negligible.
-------
IV-2
Table IV-1. Total Uncontrolled VOC Emissions from a Model
Plant Using Allied Technology3
Source
Cumene oxidation
CHP concentration
Crude-acetone (light-ends) column
Acetone finishing column
Other distillation column
Storage vents
Handling
Fugitive
Wastewater treatment
Incineration of tars
Total
Vent
Designation
(Fig. III-l)
A
B
C
D
E
H
I
J
K
L
VOC
Ratio
(g/kg)b
20.630
1.825
0.300
0.648
0.060
0.663
0.250
1.654
0.018
0.006
26.054
Emission
Rate
(kg/hr)
471.00
41.67
6.85
14.79
1.37
15.14
5.71
37.76
0.41
0.13
594.83
Uncontrolled emissions are emissions from a process for which there are no
control devices other than those necessary for economical operation.
g of emissions per kg of phenol produced.
-------
IV-3
c.
source of VOC emitted from the process. The estimated composition of the
uncontrolled vent gas, shown in Table IV-2, is based on reported1" composi-
tions after the use of carbon adsorption for emission control with a reported
VOC removal efficiency of 92%.
CHP Concentration—The uncontrolled emission from this vacuum distillation
step is considered to be the vent stream immediately before the Det after-
condenser. The stream consists primarily of cumene and spent air previously
held in solution in the oxidation reaction product plus water vapor from the
steam jet. The estimate of the uncontrolled emissions is based on the re-
ported1 controlled emissions and an estimated control efficiency of 98%.
rrnjg-Acetone. Acetone Finishing, and othej^istillation^olumM---Estimates of
the emissions from these columns are based on reported rates.1'3'4 Light
hydrocarbons, such as acetaldehyde, generated in the process are vented from
the overheads accumulator on the crude-acetone column (vent C, Fig. III-l).
Acetone and inert gases are vented from a refrigerated condenser system on the
acetone finishing column (vent D, Fig. III-l) - Emissions from the other dxstU-
lation columns (vent E, Fig. III-l) are associated with the vacuum jets on the
columns and consist of various hydrocarbons, including predominantly cumene,
AMS, and ethyIbenzene/toluene.
storage and Handling Emissions—Emissions result from feed, intermediate-
product, and final-product storage tanks. Sources are described in Table IV-3
and shown as vent H on Fig. III-l- Storage tank data were calculated by use of
equations from AP-42* based on fixed-roof tanks, half full, with a dxurnal
temperature variation of ll'C. However, breathing losses were divided by 4 to
account for recent evidence indicating that the AP-42 breathing-loss equate
overpredicts emissions.« Handling emissions result from the loading (vent I,
Fig III-D of acetone and phenol into tank cars and tank trucks for shipment.
Handling emissions are shown in Table IV-1 and were calculated with the equa-
tions from AP-42,* based on submerged loading of tank cars and tank trucks,
with phenol at 49°C and all other products at 27°c. Emissions from the loadxng
of AMS and acetophenone are insignificant (44 X 10 « g/kg and 2 X 10 g/kg
respectively). Acetone accounts for two-thirds of the total VOC in the storage
emissions and for over 95% of the VOC in the handling emissions.
-------
IV-4
Table IV-2. Estimated Composition of Oxidation Vent Gas
from Model Plant Using Allied Technology
Component
Cumene
Other VOC
Total VOC
Spent air (O2, N
Composition
(wt %)
0.92
0.23
1.15
98.85
-------
IV-5
Table IV-3. Storage Requirements for 200,000-Mg/yr
Model Plant Using Allied Technology
Stored Material
Cumene
Cumene feed/recycle
Cumene /CHP
Crude product
Crude acetone
Acetone (day)
Acetone product
Crude AMS
AMS product
Phenol (day)
Phenol product
Tars
Acetophenone
Number
of Tanks
1
1
1
1
1
2
1
1
2
2
I
1
1
Tank
Size
(M gal)
3000
1000
1000
300
100
150
300
20
100
150
3000
10
20
Turnovers
Per Year
23
6a
6a
6a
63
133
133
6a
15
163
16
204
48
Bulk Liquid
Temperature
80
80
160
110
80
80
80
80
80
120
120
200
85
Surge tanks with nearly constant level.
-------
IV-6
e- Fugitive Emissions Process pumps, process valves, and pressure-relief valves
are potential sources (J) of fugitive emissions. The model plant is estimated
to have 148 pumps, 998 process valves, and 54 relief valves, based on data
supplied by a producer.7 The fugitive emission factors from Appendix B were
applied to these estimates; the results are shown in Table IV-1.
f- Secondary Emissions Emissions can result from handling and disposal of pro-
cess waste streams. For the model plant, sources of wastewater and tars or
residuals (K,L) are indicated on Fig. III-l.
Estimates of the secondary emissions from wastewater treatment are based on
reported1'3 flows and organic contents of phenolic and nonphenolic wastewater.
Emissions from wastewater will be discussed in an EPA report8 on secondary
emissions.
The cumene process forms substantial quantities of tarry products that can be
used as fuel or can be disposed of by incineration.9—ll The venting of flue
gas from combustion of these waste products results in secondary emissions of
VOC. Emissions from such sources are characteristically low. An emissions
estimate was based on AP-42,12 with the tars assumed to be similar to residual
oil in industrial and commercial boiler service.
B, PROCESS BY HERCULES TECHNOLOGY
1. Model Plant
The model plant for the synthesis of phenol and acetone from cumene using
Hercules licensed technology has the same product capacity as the model plant
representing Allied technology; however, the by-product mix is different. The
capacities are respectively 200,000, 120,000, and 10,500 Mg/year for phenol,
acetone, and AMS based on 8760 hr of operation annually. Acetophenone is not
recovered as a by-product but remains with ,-he waste tars from the process.
This capacity is typical of recently built or announced plants manufacturing
phenol and acetone from cumene. The process depicted in Fig. III-2 is believed
to be typical of actual processes using Hercules technology.
-------
IV-7
2. Sources and Emissions
Uncontrolled emission sources and emission rates are summarized in Table IV-4
and described in greater detail below.
a. Cumene Oxidation The largest source of VOC emitted from this process is the
spent air vented from the cumene oxidation reactors (vent A, Fig. III-2) follow-
ing the refrigerated condenser system. The composition of the uncontrolled
vent gas, shown in Table IV-5, is based on reported3 data. It should be noted
that the order-of-magnitude difference in uncontrolled emissions from this step
as shown in Tables IV-1 and IV-4 for Allied and Hercules technology respectively
is due to the comparatively high operating pressure for the refrigerated con-
denser system in the Hercules process.
b. Oxidate Wash and Separation Estimates of this source (vent B, Fig. III-2) are
for the vent stream following partial recovery of VOC using a water-cooled
condenser. The estimate is based on the estimated release of inert gases from
the oxidate stream as the system pressure is reduced. The vent emission con-
sists primarily of cumene and spent air.
c. CHP Concentration The uncontrolled emissions from this vacuum distillation
step is the vent stream from the accumulator immediately before the jet after-
condenser (Vent C, Fig. III-2). The stream consists primarily of cumene, spent
air, and water vapor from the steam jet. The uncontrolled emissions estimate
is based on an estimate of the solubility of inert gases in the oxidate stream
prior to distillation.
d. CHP Cleavage This source of uncontrolled emissions (vent D, Fig. III-2) is
determined at a point immediately following the refrigerated condenser. The
emitted VOC is primarily acetone. The emission estimate is based on reported3
data.
e. Light-Ends, Acetone Finishing, and Other Distillation Columns Estimates of
the emissions from the various distillation columns are based on reported
rates.1'3'4'7 The light-ends source (vent E, Fig. III-2) consists of light
hydrocarbons such as acetaldehyde that are generated in the process. These
light hydrocarbons are purged from the process, along with acetone, from the
-------
IV-8
Table IV-4. Total Uncontrolled VOC Emissions from a Model
Plant Using Hercules Technology3
Source
Cumene oxidation
Oxidate wash/separation
CHP concentration
CHP cleavage
Light-ends column
Acetone finishing column
Other distillation column
Storage vents
Handling
Fugitive
Secondary
Wastewater treatment
Incineration of tars
Total
Vent
Designation
(Fig. III-1)
A
B
C
D
E
F
G
H
I
J
K
L
VOC
Ratio
(g/kg)b
2.314
0.078
1.217
0.473
0.300
0.648
0.060
0.660
0.249
1.654
0.027
0.008
7.688
Emission
Rate
(kg/hr)
52.83
1.79
27.78
10.80
6.85
14.79
1.37
15.06
5.70
37.76
0.62
0.17
175.52
Uncontrolled emissions are emissions from a process for which there are no
control devices other than those necessary for economical operation.
g of emissions per kg of phenol produced.
-------
IV-9
Table IV-5. Estimated Composition of Oxidation Vent Gas
from Model Plant Using Hercules Technologya
Composition
_ Component _ ________ _ (wt *)
Cumene 0.12
Other VOC 0.01
Total VOC 0.13
Spent air (O , N2, C02) 99.78
H0 0.09
Total 100.00
See ref 3.
-------
IV-10
overhead accumulator on the column. The vent from the refrigerated condenser
on the acetone finishing column (vent F, Fig. III-2) emits acetone and inert
gases. Vents from the other distillation columns (vents G, Fig. III-2) are
associated with the accumulators, vacuum jets, and condensers on the columns
and contain various hydrocarbons, including, predominantly, cumene, AMS, and
ethyl benzene.
f. Storage and Handling Emissions Emissions result from feed, intermediate-product,
and final-product storage tanks. Sources are shown as vents H in Fig. III-2
and are further described in Table IV-6. Equations from AP-425 were used for
calculating storage-tank data based on fixed-roof tanks, operated half full,
and experiencing a diurnal temperature variation of 11°C. The resulting
breathing-loss data were divided by 4 to account for recent evidence indicating
that the AP-42 breathing-loss equation overpredicts emissions.6 Loading acetone
and phenol into tank cars and tank trucks for shipment results in handling-
emission sources (vent I, Fig. III-2). These emissions are shown in Table IV-4
and were calculated with the equations from AP-42,5 based on submerged loading
in tank cars and tank trucks, with phenol at 49°C and all other products at
27°C. Acetone accounts for about two-thirds of the total VOC in the storage
emissions and over 95% of the VOC in the handling emissions. Emissions from
loading AMS are insignificant (44 X 10 6 g/kg of phenol produced).
g. Fugitive Emissions The estimate and bases are the same as those used for the
Allied technology model plant discussed in Sect. IV-A-2e. The sources are
identified as vent J in Fig. III-2.
h. Secondary Emissions Sources of emissions (vents K and L, Fig. III-2) are
wastewater and tars or residuals. The bases and discussion in Sect. IV-A-2f
also apply to the Hercules model plant. In addition to phenolic and non-
phenolic wastewater streams similar to those in the Allied technology, the
Hercules process generates a spent aqueous Na2C03 stream containing VOC, pri-
marily cumene. The secondary-emission estimate in Table IV-4 includes an
estimate for this added source.
-------
IV-11
Table IV-6. Storage Requirements for 200,000-Mg/yr
Model Plant Using Hercules Technology
Stored Material
Cumene
Cumene feed /re cycle
Cumene/CHP
Crude product
Crude acetone
Acetone (day)
Acetone product
Crude AMS
AMS product
Phenol (day)
Phenol product
Tars
Number
of Tanks
I
I
1
1
1
2
1
1
2
2
1
1
Tank
Size
(M gal)
3000
1000
1000
300
100
150
300
20
100
150
3000
20
Turnovers
Per Year
23
6a
63
6a
6a
133
133
6a
15
163
16
165
Bulk Liquid
Temperature
80
80
160
110
80
80
80
80
80
120
120
200
Surge tanks with nearly constant level.
-------
IV-12
C. REFERENCES*
1. C. W. Stuewe, IT Enviroscience, Trip Report for visit to Allied Chemical Corpo-
ration, Philadelphia, PA, March 16, 1978 (on file at EPA, ESED, Research
Triangle Park, NC) (June 1979).
2. M. H. Siemens, Dow Chemical Company, Freeport, TX, Texas Air Control Board
Emissions Inventory Questionnaire for 1975.
3. Shell Oil Company/Shell Chemical Company, Deer Park, TX, Texas Air Control
Board Permit Application for phenol-2 as revised May 9, 1975.
4. C. W. Stuewe, IT Enviroscience, Trip Report for Visit to Georgia Pacific Corpo-
ration, Plaquemine, LA, August 2, 1977 (on file at EPA, ESED, Research Triangle
Park, NC) (July 1979).
5. C. C. Masser, "Storage of Petroleum Liquids," pp. 4.3-1—4.3-16 in Compilation
of Air Pollution Emission Factors, 3d ed., Part A, AP-42 (April 1977).
6. Letter dated May 30, 1979, from E. C. Pulaski, TRW, Inc., to Richard Burr, EPA,
Research Triangle Park, NC.
7. C. W. Stuewe, IT Enviroscience, Trip Report for Visit to Monsanto Chemical
Intermediates Co., Alvin, TX, July 28, 1977 (on file at EPA, ESED, Research
Triangle Park, NC) (July 1979).
8. J. J. Cudahy and R. L. Standifer, IT Enviroscience, Secondary Emissions (June
1980) (EPA/ESED report, Research Triangle Park, NC).
9. P. R. Pujado, J. R. Salazar, and C. V. Berger, "Cheapest Route to Phenol,"
Hydrocarbon Processing 55(3), 91—96 (1976).
10. Yen-Chen Yen, Report No. 22. Phenol, p 12, A private report by the Process
Economics Program, Stanford Research Institute, Menlo Park, CA (April 1967).
11. Yen-chen Yen, Report No. 22A. Phenol Supplement A, p 65, A private report by
the Process Economics Program, Stanford Research Institute, Menlo Park, CA
(September 1972).
12. T. Lahre, "Fuel Oil Combustion," Table 1.3-1 in Compilation of Air Pollution
Emission Factors, 3d ed., Part A, AP-42 (August 1977).
*Usually, when a reference is located at thr 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 VIA ALLIED TECHNOLOGY
Applicable control systems and emission estimates are summarized in Table V-l
and discussed below.
1. Cumene Oxidation
In the Allied process 88% of the uncontrolled process emissions come from
vent A (Fig. III-l). The control option selected for the model-plant cumene
oxidation vent is carbon adsorption. With good design and operation the VOC
content in the vent from the carbon adsorption unit is estimated to fall in the
range of 50 to 100 ppm , with 0.3 kg of steam/kg of carbon used for regenera-
tion. The resulting VOC emission reduction is 97.5% at an expected 70 ppm .
At 100 ppm , VOC emission reduction would be 96.4%. The design is based on a
0.91-m-deep bed, a superficial velocity of 0.51 m/s, and an estimated loading
capacity of 11 Ib of VOC/100 Ib of carbon (see the report1 on carbon adsorp-
tion) . Potential alternative controls for this emission source include the use
of other adsorbents (e.g., resins).2
The regeneration cycle operation can have a significant effect on the VOC
content in the vent. With operation at a regeneration steam ratio of 1 kg of
steam/kg of carbon it is estimated that the VOC content in the vent would fall
in the range of 5 to 20 ppm . At an expected 12 ppm the VOC emission reduc-
tion would be 99.6%; at 20 ppm it would be 99.3%.
2. CHP Concentration
The primary VOC in vent B (Fig. III-l) is cumene. Condensation at 4.4°C and
atmospheric pressure was selected as the control option. Vacuum conditions on
this distillation column are maintained by use of a steam-jet and condenser
system. Use of the refrigerated condenser after the jet condenser is partic-
ularly effective, due to both the overall -ncreased system pressure and the
reduced temperature, in decreasing the VOC in the vent. The overall VOC reduc-
tion is estimated to be greater than 98%. An EPA report3 will cover condensa-
tion as a control option.
-------
Table V-l. Estimates of Controlled VOC Emissions from a Model
Plant Based on Allied Technology
Source
Cumene oxidation
CHP concentration
Light-ends column
Acetone finishing column
Other distillation columns
Storage and handling
Fugitive
Secondary
Wastewater treatment
Incineration of tars and
residuals
Total
Designation
(Fig. III-l)
A
B
C
D
E
H, I
J
K
L
Control Device or Technique
Carbon adsorption
Refrigerated condenser
Combustion in boiler
Vent scrubber
No controls identified
Vent scrubber on acetone
emitting vents
Detection and correction of
major leaks
None
None
Total VOC
Emission
Reduction
97.5
.98
'VIOO
96
76
71
VOC
Ratio
(g/kg) a
0.523
0.036
VL X 10~
0.026
0.060
0.222
0.478
0.018
0.006
1.369
Emission
Rate
(kg/hr)
11.94
0.83
4 -3
^2 X 10
0.59
1.37
5.07
10.92
0.41
0.13
31.26
g of emission per kg of phenol produced.
""Regeneration with 0.3 Ib of steam/lb of carbon.
I
NJ
-------
V-3
3. Light-Ends Vent
This vent stream is rich in acetone, aldehydes, and other combustible hydro-
carbons. The control option selected for the light-ends vent is combustion in
an existing boiler or incinerator. Based on emission factors from AP-424 the
VOC reduction is estimated to be almost 100%. Installation of an incinerator
solely for the purpose of controlling this source would not be justifiable;
therefore this control method is applicable only if an existing combustion
chamber can be used. This vent stream is flammable, and safe handling prac-
tices should be considered in the design and operation of the collection and
transport system.
Another option used for control of the VOC in this vent stream is aqueous
scrubbing.5 It is estimated that a VOC reduction of 96 to 98% could easily be
obtained since the major VOC constituents are highly soluble in water. A
potential disadvantage of aqueous scrubbing is that part of the VOC removed may
be emitted as secondary emissions during wastewater treatment. Treatment of
the scrubbing liquor in an acetone recovery system before it is sent to waste-
water treatment would result in recovery of other light hydrocarbons and defeat
the purpose of the light-hydrocarbon (light ends) stripping in the crude-
acetone column.
-l. Acetone Finishing Column
The VOC in this vent should be relatively pure acetone and thus recoverable.
Aqueous scrubbing of the acetone finishing column vent was selected as the
control option. A slightly reduced pressure in this column is maintained with
a steam-jet and condenser system to enhance separation efficiencies. The
scrubber would be applied to the vent from the jet after-condenser. It is
estimated that the overall VOC emission reduction would be 96%. A future EPA
report6 will discuss the use of absorption as a control option.
An alternative control option could be chilled condensation. It is estimated
that the overall VOC emission reduction would be only about 40% based on the
physical properties of the vent stream at condensation conditions of 4.4°C and
atmospheric pressure.
-------
V-4
5. Other Distillation Columns
The VOC in the emissions from the other distillation columns contain phenol,
cumene, AMS, and other hydrocarbons. Since the emission level is relatively
low, no control options were identified for the model plant.
6. Storage and Handling
The major component of the VOC in the vents from storage and from handling,
particularly, is acetone. The control option selected for the model-plant
storage and handling sources is aqueous scrubbing on the acetone emitting
vents. These vents include acetone loading, acetone day tanks, acetone product
tank, crude-product tank, and crude-acetone tank. A conservative estimate of
96% VOC removal efficiency was used to calculate the reduction of VOC in these
vents, resulting in an overall VOC emission reduction of 76%.
Floating-roof tanks have been reported as a control option on acetone
tanks.5'7'8 The controlled storage emissions on the acetone tanks were calcu-
lated by assuming that a contact type of internal floating roof with secondary
seals will reduce fixed-roof-tank emissions by 85%.9 With this control on only
the acetone emitting tanks the overall reduction will be 45%. Another EPA
report10 covers control options for storage and handling.
"i. Fugitive
Controls for fugitive emissions from the synthetic organic chemicals manufac-
turing industry will be discussed in a future EPA document.11 Emissions from
pumps and valves can be controlled by appropriate leak-detection systems,
repairs, and maintenance as required. Controlled fugitive emissions calculated
with the factors given in Appendix B are included in Table V-l; these factors
are based on the assumption that major leaks are detected and corrected.
8. Secondary Emissions
No additional control systems for secondary emissions have been identified for
the model plant. An EPA report12 discusses control of secondary emissions.
B. PROCESS BY HERCULES TECHNOLOGY
A summary of applicable control systems and emission estimates is given in
Table V-2 and discussed below.
-------
Table V-2. Estimates of Controlled VOC Emissions from a Model
Plant Based on Hercules Technology
Vent
Designation
Source (Fig. III-3)
Cumene oxidation
Oxidate wash separation
CHP concentration
CHP cleavage
Light-ends column
Acetone finishing column
Other distillation columns
Storage and handlirg
Fugitive
Secondary
Wastewater treatment
Incineration of tars and
residuals
A
B
C
D
E
F
G
H, I
J
K
L
Total VOC
Emission
Reduction
Control Device or Technique (%)
Carbon adsorption 77 . 4
Refrigerated condenser 86
Refrigerated condenser 98
Vent scrubber 96
Combustion in boiler M.OO
Vent scrubber 96
No controls identified
Vent scrubber on acetone 76
emitting vents
Detection and correction of 71
major leaks
None
None
VOC Emission
Ratio
(g/kg) a
0.523
0.011
0.024
0.019
~1 X 10~4
0.026
0.060
0.218
0.478
0.027
0.008
Rate
(kg/hr)
11.94
0.25
0.56
0.43
^2 X 40~
0.59
1.37
4.98
10.92
0.62
0.17
Total
1.394
31.83
-------
V-6
1. Cumene Oxidation
In the Hercules process 45% of the uncontrolled process emissions emanate from
the cumene oxidation vent source (vent A, Fig. III-2). For the model plant,
carbon adsorption was selected as the control option for the cumene oxidation
vent. With proper design and operation the VOC content in the vent from the
carbon adsorption unit should be within a range of 50 and 100 ppm , with 0.3 kg
of steam/kg of carbon used for regeneration. The resulting VOC emission reduc-
tion is 77.4% at the expected 70 ppm . In the Hercules process the vent stream
exiting from the refrigerated condenser at 4 to 5°C and 5.9 X 105 Pa can be
cross-exchanged with the hot vent stream from the reactors both to recover heat
and, more importantly, to decrease the relative humidity of the water vapor in
the gas stream. At high relative-saturation pressures, water vapor will com-
pete with the organic vapors for the carbon's adsorptive capacity.13 The
system design is based on a 0.91-m-deep bed, a superficial velocity of
0.51 m/s, and a loading factor calculated by the method given in an EPA report1
on carbon adsorption.
At a regeneration steam ratio of 1 kg of stream/kg of carbon the VOC content in
the vent is estimated to fall between 5 am
the VOC emission reduction would be 96.1%.
the vent is estimated to fall between 5 and 20 ppm . At an expected 12 ppm
2. Oxidate Wash/Separation
This relatively small source of VOC (vent B, Fig. III-2) consists primarily of
cumene with inert gases. The control option selected for this source is con-
densation by use of a refrigerated coolant. The estimate of controlled emis-
sions is based on physical properties for the estimated stream composition at
the condensing conditions of 4.4°C and atmospheric pressure. The estimated
emission reduction is 86%. A future EPA report3 will cover condensation as a
control device.
3. CHP Concentration
The control-option selection and discussion in Sect. V-A-2 for the Allied proc-
ess is directly applicable to this vent in the Hercules process.
-------
V-7
4. CHP Cleavage
The VOC in this vent stream is primarily acetone. Economical operation requires
partial condensation of acetone vapor by use of a refrigerated coolant as part
of the process. The control option selected for the CHP cleavage vent is
aqueous scrubbing and would be applied to the vent from the refrigerated con-
denser. It is estimated that the overall VOC emission reduction would be 96%.
Another EPA report6 will further discuss absorption as a control option.
5. Light-Ends Column Vent
Although the distillation sequence differs for the Allied and Hercules model
plants, the light-ends vent stream is similar for the two processes. The same
control-option selection and discussion given in Sect. V-A-3 for the Allied
process are applicable to this source (vent E, Fig. III-2) for the Hercules
process.
6. Acetone Finishing Column
The control-option selection and discussion in Sect. V-A-4 for the Allied
process are applicable for this source (vent F, Fig. III-2) for the Hercules
process.
7. Other Distillation Columns
Since the emission level is relatively low, no control options were identified
for this source (vent G, Fig. III-2) for the Hercules model plant.
8. Storage and Handling
The control-option selection and discussion in Sect. V-A-6 for the Allied model
plant are applicable for these sources (vents H and I, Fig. III-2) for the
Hercules model plant.
9. Fugitive
This source (vent J, Fig. III-2) can be cor.rolled in the manner that is dis-
cussed for the Allied model plant in Sect. V-A-7.
10. Secondary Emissions
No additional controls were identified for this source (vents K and L,
Fig. III-2).
-------
V-8
C. REFERENCES*
1. H. S. Basdekis and C. S. Parmele, IT Enviroscience, Control Device Evaluation.
Carbon Adsorption (January 1981) (EPA/ESED report, Research Triangle Park, NC).
2. J. Beale, CMA, letter dated Nov. 14, 1980, to Robert E. Rosensteel, EPA, ESED,
Research Triangle Park, NC.
3. D. G. Erikson, IT Enviroscience, Inc., Control Device Evaluation. Condensation
(December 1980) (EPA/ESED report, Research Triangle Park, NC).
4. W. M. Vatavuk, "Petroleum Industry," Table 9.1-1 in Compilation of Air Pollution
Emission Factors, 3d ed., Part B, AP-42 (August 1977).
5. C. W. Stuewe, IT Enviroscience, Trip Report for Visit to Monsanto Chemical
Intermediates Co., Alvin, TX, July 28, 1977 (on file at EPA, ESED, Research
Triangle Park, NC) (July 1979).
6. R. L. Standifer, IT Enviroscience, Control Device Evaluation. Gas Absorption
(October 1980) (EPA/ESED report, Research Triangle Park, NC).
7. C. W. Stuewe, IT Enviroscience, Trip Report for Visit to Allied Chemical Corpo-
ration, Philadelphia, PA, March 16, 1973 (on file at EPA, ESED, Research
Triangle Park, NC) (June 1979).
8. Shell Oil Company/Shell Chemical Company, Deer Park, TX, Texas Air Control
Board Permit Application for phenol-2 as revised May 9, 1975.
9. W. T. Moody, TRW, Inc., letter dated Aug. 15, 1979, to David A Beck, EPA, ESED,
Research Triangle Park, NC.
10. D. G. Erikson, IT Enviroscience, Storage and Handling (September 1980) (EPA/
ESED report, Research Triangle Park, NC).
11. D. G. Erikson and V. Kalcevic, IT Enviroscience, Fugitive Emissions (September
1980) (EPA/ESED report, Research Triangle Park, NC).
12. J. J. Cudahy and R. L. Standifer, IT Enviroscience, Secondary Emissions (June
1980) (EPA/ESED report, Research Triangle Park, NC).
13. C. S. Parmele, W. L. OConnell, and H. S. Basdekis, "Vapor-Phase Adsorption Cuts
Pollution, Recovers Solvents," Chemical Engineering 86(28), 62 (December 1979).
*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. Process by Allied Technology
Table VI-1 gives 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 an estimated reduction of total VOC emissions by 94.8%, or about
4940 Mg/yr for the model plant, resulting in controlled emissions from the
model plant of about 270 Mg/yr.
a.
Cumene Oxidation Vent The adsorption with carbon of VOC from the spent air
from the oxidation reactors reduces the model-plant VOC emissions by 4021 Mg/yr.
Adsorbed VOC is recovered and then recycled as process feed. The major energy
impact results from the required regeneration steam, which is estimated to be
equivalent to about 6000 MJ/Mg of VOC removed.
All Other Process Vents The control of vent sources B, C, and D by the control
options shown in Table VI-1 reduces the model-plant VOC emissions by 541 Mg/yr.
The energy for these controls is impacted by the required refrigeration for the
condenser coolant and the energy required to either recover the acetone from
the scrubber effluent or destroy the acetone in a biological wastewater treat-
ment system. These energy requirements are offset by the potential heat re-
covery from combustion of the light ends in an existing boiler. The overall
energy impact is estimated to be a net credit of about 86 MJ/hr. The impact
ratio is estimated to be a credit of about 1380 MJ/Mg of VOC removed.
Nonprocess Emissions {Storage, Handling, and Fugitive) Storage and handling
emissions from the model plant are partly controlled by aqueous scrubbing of
the acetone emitting tanks and the acetone xoading vents. Application of this
control results in a VOC emission reduction of 138 Mg/yr for the model plant.
Fugitive emissions are controlled by the repair of leaking components. VOC
emissions reduction by control of fugitive emissions is estimated to be
235 Mg/yr. A separate EPA report1 covers energy requirements for the control
of storage and handling emissions.
-------
VI-2
Table VI-1. Environmental Impact of Controlled Model
Plant Using Allied Technology
Emission Source
Cumene oxidation
CHP concentration
Crude-acetone (light-ends)
column
Acetone finishing column
Other distillation column
Storage and handling vents
Fugitive
Secondary
Wastewater treatment
Incineration of tars
Total
Vent
Designation
(Fig. III-l)
A
B
C
D
E
H, I
J
K
L
Control Device or Technique
Carbon adsorption
Refrigerated condenser
Combustion in existing
boiler
Vent scrubber
None
Vent scrubber on acetone
emitting vents
Detection and correction
of major leaks
None
None
VOC Emission
Reduction
(%) (Mg/yr)
97.5 4021
98 358
^100 60
96 124
76 138
71 235
4936
-------
VI-3
2. Process by Hercules Technology
Table VI-2 summarizes the environmental impact of reducing the total VOC emis-
sions by application of the described control systems (Sect. V) to the model
plant described in Sects. Ill and IV. Use of these control devices results in
an estimated reduction in total VOC emissions by 82%, or about 1260 Mg/yr, and
results in controlled emissions from the model plant of about 275 Mg/yr.
a. Cumene Oxidation Vent Application of carbon adsorption to the spent air from
the oxidation reactors reduces model-plant VOC emissions by 358 Mg/yr. The
adsorbed VOC is recovered and then recycled to the process. The main energy
impact results from the steam required for regeneration of the carbon. The
energy equivalent of the steam is estimated to be about 11 GJ/Mg of VOC removed.
b. All Other Process Vents Control of vent sources B—F by the control options
shown in Table VI-2 reduces the model-plant VOC emissions by 529 Mg/yr. Energy
for these controls is impacted by the required refrigeration for the condenser
coolants and the energy required to either recover the acetone from the scrubber
effluent or destroy the acetone in a biological wastewater treatment system.
These energy requirements are partly offset by the potential heat recovery from
combustion of the light ends in an existing boiler. The overall energy impact
is estimated to be a net requirement of about 4 MJ/hr. The impact ratio is
estimated to be about 66 MJ/Mg of VOC removed.
c. Nonprocess Emissions (Storage, Handling, and Fugitive) Emissions from the
model-plant storage and handling are partly controlled by aqueous scrubbing of
the acetone emitting tanks and the acetone loading vents. The estimated VOC
emission reduction for the model plant through application of this control is
138 Mg/yr. Fugitive emissions are controlled by the repair of leaking com-
ponents, with an estimated VOC emission reduction of 235 Mg/yr. A separate EPA
report1 covers energy requirements for the control of storage and handling
emissions.
3. 1980 Industry Emissions
The total VOC emissions from the domestic production of phenol/acetone by the
cumene process are estimated at 4030 Mg and include estimated emissions from
the process, fugitive, secondary, and storage and handling sources. This
-------
VI-4
Table VI-2. Environmental Impact of Controlled Model
Plant Using Hercules Technology
Emission Source
Cumene oxidation
Oxidate wash/separation
CHP concentration
CHP cleavage
Light-ends column
Acetone finishing column
Other distillation column
Storage and handling vents
Fugitive
Vent
Designation
(Fig. III-2)
A
B
C
D
E
F
G
H, I
J
VOC Emission
Reduction
Control Device or Technique
Carbon adsorption
Refrigerated condenser
Refrigerated condenser
Vent Scrubber
Combustion in boiler
Vent scrubber
None
Vent scrubber on
acetone emitting vents
Detection and correction
of major leaks
(%)
77.4
86
98
96
VLOO
96
76
71
(Mg/yr)
358
13
238
91
60
124
138
235
Secondary
Wastewater treatment
Incineration of tars
Total
K
L
None
None
1257
-------
VI-5
estimate is based on a projected 1980 level of production of 1,320,000 Mg. The
estimated emissions were determined by applying the emission ratios from Tables
IV-1, IV-4, V-l, and V-2. Process emissions are estimated to be 92% con-
trolled, storage and handling emissions to be 67% controlled, and fugitive
emissions to be uncontrolled. Emissions from secondary sources are believed to
be negligible. Controls reported by producers are summarized in Appendix C.
B. COST CONTROL IMPACT
The cost control impact described below relates to the production of phenol/
acetone by the cumene process by Allied and Hercules technology. Details of
the model plants (Figs. III-l and 2) are given in Sects. Ill and IV.
Capital cost estimates represent the total investment required for purchase and
installation of all equipment and material needed for a complete 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, research and development, or land acquisition. If retrofitting is
considered for these controls, it should be recognized that a primary diffi-
culty in retrofitting may be in finding space to fit the control system into
the existing plant layout. Because of these associated costs the cost of
retrofitting emission control systems in existing plants may be appreciably
greater than that for a new installation.
Bases for the annual cost estimates for the control alternatives include utili-
ties, raw materials, maintenance supplies and labor, recovery credits, capital
charges, and miscellaneous recurring costs such as taxes, insurance, and admin-
istrative overhead. (Incremental operating labor costs are assumed to be
minimal and therefore are not included.) Emission recovery credits are based
on the raw-material value of the material recovered.2 Annual costs are for a
1-year period beginning mid-1979.
1. Cumene Oxidation Emissions
The major source of emissions from the production of phenol/acetone by the
cumene process for both Allied and Hercules technology is the spent air from
the cumene oxidation reaction. These emissions are controlled by a carbon
adsorption system. The cost estimate for the control system is based on a
separate EPA report on carbon adsorption as a control option.3 As applied to
-------
VI-6
the Hercules model plant the carbon adsorption system does not require a vent
stream blower. Capital and operating costs were adjusted to reflect this
change. The costs and cost effectiveness are summarized in Table VI-3 at two
regeneration steam ratios: 0.3 and 1.0 kg of steam/kg of carbon. The VOC
emission reduction given in Tables VI-1 and VI-2 are based on a regeneration
steam ratio of 0.3 kg of steam/kg of carbon. The VOC emission reduction benefit
resulting from use of the higher steam ratio is discussed in Sects. V-A-1 and
V-B-1.
2. Other Process Emissions
Emissions from other process vents are controlled as shown in Tables VI-1 and
VI-2 by condensation, combustion, and absorption (vent scrubbing). Condensation
and absorption are covered in separate EPA reports.4'5 The predominant cost
involved in the use of an existing boiler or incinerator would be installation
of the piping necessary to transfer the vent stream to the combustion device.
As the cost of the required piping will depend primarily on the distance of the
phenol/acetone plant from the combustion device, which can vary greatly, the
cost impact was not determined. Another EPA report6 covers the use of emissions
as fuel.
3. Storage and Handling Sources
The control method for storage and handling is aqueous scrubbing of the acetone
emitting vents. Another EPA report1 covers applicable controls for storage and
handling emissions.
4. Fugitive Sources
A future EPA document7 will cover fugitive emissions and applicable controls.
5. Secondary Sources
No control system has been identified for controlling the secondary emissions
from wastewater treatment or from the dispjSal of residues by incineration. An.
EPA document8 covers secondary emissions for the synthetic organic chemicals
manufacturing industry.
-------
Table VI-3. Summary of Costs and Cost Effectiveness for Carbon Adsorption
Applied to Allied and Hercules Model Plants
. •
Technology
Allied
Hercules
a
Savings.
. — • ~ —
Regeneration
Steam Ratio
(kg of steam/kg of carbon
0.3
1.0
0.3
1.0
. —
Installed
) Capital
$574,000
574,000
$517,000
517,000
Annual
$259,000
434,000
$177,000
200,000
Costs
Annual
Recovery
Credit
$1,443,000
1,469,000
$120,000
147,000
Net
Annual
($1,184, 000) a
( l,035,000)a
$57,000
53,000
Cost
Effectiveness
(per Mg removed)
($294)a
( 252)a
$156
119
<
M
1
-------
VI-8
C.
1.
2.
3.
4.
5.
6.
7.
8.
REFERENCES*
D. G. Erikson, IT Enviroscience, storage and Handling (September 1980) (EPA/
ESED report, Research Triangle Park, NC) .
-Current Prices of Chemicals and Related Materials," Chemi^lJiaJ±etin3
Reporter, May 28, 1979.
D o Erikson and V. Kalcevic, It wlro^Unc.. fHHLSHS (September
I960) (EPA/ESED report, Research Triangle Park, NC) .
j a cudahy and R. L. Standifer, IT I»iro«i««. SecondarvJ.isslons (June
1980) (EPA/ESED report. Research Triangle Park, NC).
*Usuany, »hen a reference i
the entire paragraph. If another ,re«^" th material involved.
paragraph, that reference number is indicated on the n
the reference appears on a heading, it rer.rs
heading.
-------
VII-1
VII. SUMMARY
Phenol and acetone are co-products of the cumene peroxidation process, which
accounts for about 97%1/2 of the phenol manufactured in the United States. As
of 1978 the cumene process also accounted for about 67% of the domestic acetone
production.3 At projected annual growth rates of 4.5%4 for phenol and 4 to 5%5
for acetone, production will reach about 87% and 77 to 81% of current capacity
by 1982 for phenol and acetone respectively.
Two process variations of the basic cumene peroxidation route are practiced
commercially. About 47% of the current capacity utilizes a process based on
Allied Chemical licensed technology. The remaining capacity is based on Her-
cules licensed technology.
Emission sources and uncontrolled and controlled emission rates for the model
plants based on the two processes are given in Tables VII-1 and VII-2. The
most significant process emission sources of both processes are the cumene
oxidation reaction vents, which are controlled in the model plants by carbon
adsorption.
Storage and handling emissions are predominantly acetone. These emissions for
the model plants are controlled by aqueous scrubbing. Potential secondary
emissions are minor. The total industry VOC emissions from processes based on
cumene peroxidation were estimated in this study to be 4030 Mg in 1980, with
most of the uncontrolled VOC emissions coming from fugitive sources.
lnNo Switch from Cumene, Say Phenol Manufacturers," Chemical Engineering 86(8),
64 (Apr. 9, 1979). —
2S. A. Cogswell, "Phenol," pp 686.5021A—686.5023J in Chemical Economics Hand-
book, Stanford Research Institute, Menlo Park, CA (October 1978).
3S. A. Cogswell, "Acetone," p 604.5032A in Jhemical Economics Handbook, Stanford
Research Institute, Menlo Park, CA (July 1978).
4"Chemical Profile in Phenol," p. 9 in Chemical Marketing Reporter, Feb. 6,
1978.
5S. A. Cogswell, "Acetone," pp 604.5031 C—D in Chemical Economics Handbook,
Stanford Research Institute, Menlo Park, CA (July 1978).
-------
VII-2
Table VII-1. Emission Summary for the Model
Plant Using Allied Technology
Emission Source
Cumene oxidation
CHP concentration
Crude-acetone (light-ends)
column
Acetone finishing column
Other distillation column
Storage and handling vents
Fugitive
Secondary
Wastewater treatment
Incineration of tars
Total
Vent
Designation
(Fig. III-l)
A
B
C
D
E
H, I
J
K
L
VOC Emission
Uncontrolled
471.0
41.7
6.8
14.8
1.4
20.8
37.8
0.41
0.13
594.8
Rate (kg Air)
Controlled
11.9
0.83
0.002
0.59
1.4
5.1
10.9
0.41
0.13
31.3
-------
VII-3
Table VII-2. Emission Summary for the Model
Plant Using Hercules Technology
Emission Source
Cumene oxidation
Oxidate wash/separation
CHP concentration
CHP cleavage
Light-ends column
Acetone finishing column
Other distillation column
Storage and handling vents
Fugitive
Secondary
Wastewater treatment
Incineration of tars
Total
Vent
Designation
(Fig. III-2)
A
B
C
D
E
F
G
H, I
J
K
L
VOC Emission
Uncontrolled
52.8
1.8
27.8
10.8
6.8
14.8
1.4
20.8
37.8
0.62
0.17
175.6
Rate (kg/hr)
Controlled
11.9
0.25
0.56
0.43
0.003
0.59
1.4
5.0
10.9
0.62
0.17
31.8
-------
A-l
APPENDIX A
Table A-l. Properties of Acetone*
Synonyms
Molecular formula
Molecular weight
Physical state
Vapor pressure
Vapor specific gravity
Boiling point
Melting point
Density
Water solubility
2-Propanone, dimethyl ketone, B-ketone
propane, methyl ketone, pyroacetic
ether
C3H6°
58.08
Liquid
400 mm at 39.5°C
2.0
56.2°C at 760 mm
-95.35°C
0.7972 g/ml at 150C/4°C
Infinite
*From: J. Dorigan et aL., "Acetone," p. AI-20 in Scoring of Organic Air Pol-
lutants. 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-2. Properties of Cumene*
Synonyms
Molecular formula
Molecular weight
Physical state
Vapor pressure
Vapor specific gravity
Boiling point
Melting point
Density
Water solubility
Isopropyl benzene, 2-phenyl propane, cunvol
SH11
120.21
Liquid
6.56 at 25°C
4.1
152 °C
-96°C
0.864 g/ml at 20°C/4°C
Insoluble
*From: J. Dorigan £t _al., "Cumene," p. AI-306 in Scoring of Organic Mr Pol-
lutants. Chemistry, Production and Toxicity of Selected Organic Chemicals
(Chemicals A-C), MTR - 7248, Rev. 1, Appendix I, Metre Corp., McLean, VA
(September 1976) .
-------
B-l
APPENDIX B
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
Q.00'03
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).
-------
C-l
APPENDIX C
EXISTING PLANT CONSIDERATIONS
Data reported on control devices and techniques used in existing Allied,
Georgia Pacific, Monsanto, and Shell phenol/acetone plants are summarized in
Table C-l and discussed below.
1. Cumene Oxidation Vent
Allied reported1 the use of a carbon adsorption system with an overall hydro-
carbon removal efficiency of 92%. The carbon adsorber follows a refrigerated
condenser. Georgia Pacific and Shell also reported2'3 the use of carbon adsorp-
tion following a refrigerated condenser. Georgia Pacific reported2 a control
efficiency of 99% for their carbon adsorption unit including condensation, and
Shell reported3 design data from which a control efficiency of 83.4% was calcu-
lated for the carbon adsorption step. Monsanto reported4'5 that they use a
refrigerated condenser as the control device with a 90% control efficiency at 4
to 5°C and 85 psia.
2. Oxidate Wash/Separation Vent
The oxidate wash/separation vent is not applicable to plants using the process
based on Allied technology. Georgia Pacific and Monsanto, who use Hercules
technology, reported2'4 emission control by condensation. Georgia Pacific
reported2 a control efficiency of 84%.
XC. W. Stuewe, IT Enviroscience, Trip Report for Visit to Allied Chemical
Corp., Philadelphis, PA, Mar. 16, 1978 (on file at EPA, ESED, Research
Triangle Park, NC).
2C. W. Stuewe, IT Enviroscience, Trip Report for Visit to Georgia Pacific
Corporation, Plaquemine, LA, Aug. 2, 1977 (on file at EPA, ESED, Research
Triangle Park, NC).
3Shell Oil Co./Shell Chemical Co., Deer Park, TX, Texas Air Control Board
Permit Application for phenol-2 as revised May 9, 1975.
4C. W. Stuewe, IT Enviroscience, Trip Repon for Visit to Monsanto
Chemical Intermediates Co., Alvin, TX, July 28, 1977 (on file at EPA, ESED,
Research Triangle Park, NC)
5Texas Air Control Board, Permits 1985 and 1986 issued to Monsanto Co.,
Chocolate Bayou Plant, Alvin, TX, for phenol/acetone manufacture.
-------
Table C-l. Control Devices and Techniques Reported by Existing Plants
Control Device or Technique Used by
Emission Source
Cumene oxidation
Oxidate wash/separa-
tion
CHP concentration
CHP cleavage and
neutralization
Light-ends column
Acetone finishing
column
Other distillation
columns
Storage
Allied
Carbon adsorption
Not applicable
Chilled-brine
condenser
No vent
Condensation
Water scrubber
Water scrubber for
cumene recovery
distillation; con-
densation on AMS,
phenol, and aceto-
phenone columns
IFRS on 2 acetone
and 2 cumene tanks;
vent scrubber on
other acetone tanks
Georgia-Pacific
Monsanto
Shell
Carbon adsorption
Condensation
Condensation
Condensation on
cleavage and vent
water scrubber on
neutralization
Handling
Vent scrubber on
acetone loading
Pressurized refrigerated
condensation
Condens ation
Condensation
Vent condenser for
cleavage and for
neutralization
Carbon adsorption
Water scrubber for
cleavage
Incineration-existing Water scrubber
boiler
Condensation Condensation
Incineration in
existing boiler
on AMS and heavy-
ends columns
Water scrubber on
acetone day tanks
and cleavage prod-
uct tank; vent
condensers on
acetone storage and
light- and heavy-
oil tanks
Control on acetone
by unnamed device
Condensation on crude
acetone/phenol, heavy-
ends, and phenol puri-
fication columns
IFR for crude AMS and
most acetone tanks
Incineration in existing i
fire box
Water scrubber
Incineration in existing
fire box for crude
acetone columns
FR on acetone tanks;
vent scrubber on
phenol, heavy-ends,
and light HC tanks;
refrigerated conden-
sation on cumene/CHP
and on crude-product
tanks
Vent scrubber on acetone
loading
o
1 ee ref I bSee ref 2. CSee ref 4. dsee ref 3. Internal floating-roof tank. Floating-roof tank.
-------
C-3
3. CHP Concentration Vent
The CHP concentration column is operated under vacuum. Allied reported1 using
chilled-brine condensation to control the vent. Georgia Pacific reported2 a
control efficiency of 95% using condensation. Monsanto also reported4 using
condensation to control the vent.
4. CHP Cleavage and Neutralization Vents
Georgia Pacific reported2 a cleavage ejector condenser for the cleavage vent
with a control efficiency of 93% and a water scrubber on the neutralization
vent. Condensation on both the cleavage and the neutralization vents was
reported by Monsanto.4 Shell reported3 use of refrigerated condensation
followed by water scrubbing of the cleavage reactor vent. Based on the data
supplied a control efficiency of 96% was calculated for the Shell scrubber.
5. Light-Ends Column Vent
Allied reported1 the use of condensation to control the light-ends column vent.
Georgia Pacific and Shell reported2'3 that they incinerated the vent stream by
using it as part of the fuel for existing fire boxes. Control by aqueous
scrubbing of the vent and eventual disposal of the wastewater by underground
injection was reported4 by Monsanto.
6. Acetone Finishing Column Vent
Georgia Pacific and Monsanto reported2'4 the use of condensation for control of
the acetone finishing column vent. Allied and Shell use aqueous scrubbing for
control of this vent.1'3 The Shell scrubber follows refrigerated condensation,
and based on the data reported,3 a control efficiency of 95% was calculated for
the water scrubber.
7. Other Distillation Column Vents
Allied, Georgia Pacific, Shell, and Monsanto reported1—4 varying control
techniques for selected distillation column vents. The control techniques
reported were aqueous scrubbing, condensation, and incineration.
8. Storage
Allied, Shell, and Monsanto reported1'3'4 the use of floating-roof tanks for
storage of acetone. Floating-roof tanks were also reported by Allied1 for
-------
C-4
cumene storage and by Monsanto4 for crude-AMS storage. Aqueous scrubbing of
acetone tank vents was reported1—3 by Allied, Georgia Pacific, and Shell. The
use of condensation on selected tanks was reported2'3 by Georgia Pacific and
Shell. Shell also reported3 using aqueous scrubbing, of tank vents containing
phenol, for hydrocarbon and odor control.
9. Handling
Allied and Shell reported1'3 using aqueous scrubbing to control acetone-loading
vents. Georgia Pacific reported2 a control efficiency of 70% on the acetone-
loading vents.
-------
D-l
APPENDIX D
COST ESTIMATE PROCEDURE FOR PROCESS EMISSION CONTROL WITH CARBON ADSORPTION
A. EMISSION TO CARBON ADSORPTION
From cumene oxidation vent of a model plant using Allied technology:
377 kg „ 2.2 Ib hr Ib-mole 359 ft3 f
Cumene — * — — X - X X
Other VOC MJ^2 at 58.1 Ib/lb-mole = 21 scfm
hr
Total VOC 62
Spent air 40580 kg/hr at 28.5 Ib/lb-mole = 18,740 scfm
Total waste gas to carbon adsorption = 18,800 scfm
377 kg „ 2.2 Ib hr _ _
Cumene - X —- X - - 13.8
Other VOC 94. g = 3.5 Ib/min
hr -
Total VOC 17.3 Ib/min
VOC content - 17.3 Ib/min = 0 92 ib of VOC/1000 scf
voc content 18 8 x 1000 scf/min
B. TOTAL INSTALLED CAPITAL
From Fig. IV- 1 of the control device evaluation report for carbon adsorption,1
the December 1979 installed capital cost of a carbon adsorption system for
18,800 scfm of waste gas is $574,000
C. CARBON REQUIREMENT
For a VOC content of 0.92 lb/1000 scf and an estimated loading capacity of
11 Ib of VOC/100 Ib of carbon the carbon requirement shown in Fig. II-l of the
carbon adsorption report1 is 8 Ib of carbon/1000 scf.
H S Basdekis and C. S. Parmele, IT Enviroscience, Control Device Evaluation
Carbon Adsorption (January 1981) (EPA/ESED report, Research Triangle Park, NC).
-------
D-2
D NET ANNUAL COST f carbon
' - °<
for 1
for recovered VOC, or $259,000.
From Table VI-1 of this report the VOC adsorbed is
4021 Mg x 2205 lb _ 8/866,000 lb/yr.
of 90* of the VOC adsorbed and a raw-material value
Using an estimated recovery of 90^0 of the v
of $0 181/lb of VOC the recovery credit is as follows-.
yr
g x $0^181 = $1/443,000/yr.
The net annual cost is
$259,000-$1,443,000 = -$1,184,000, or a savings.
The cost effectiveness is
-$l_L184,000/yr _ _$294/Mg removed.
4021 Mg/yr
-------
7-i
REPORT 7
LINEAR ALKYLBENZENE
C. A. Peterson, Jr.
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
February 1981
This report contain, certain information which
Economics Handbook,
reside with Stanford
D18
-------
CONTENTS OF REPORT 7
Page
I- ABBREVIATIONS AND CONVERSION FACTORS
1-1
II. INDUSTRY DESCRIPTION II~1
A. Reason for Selection II-l
B. Usage and Growth II-l
C. Domestic Producers II-l
D. References
III. PROCESS DESCRIPTION III-l
A. Introduction III-l
B. Olefin Process " III-l
C. LAB Chlorination Process III-8
D. References 111-15
IV. EMISSIONS IV~1
A. LAB Olefin Process IV~1
B. LAB Chlorination Process IV~5
C. References IV-1
V. APPLICABLE CONTROL SYSTEMS V~I
A. LAB Olefin Process v~1
B. LAB Chlorination Process v~4
V— 8
C. References
VI. IMPACT ANALYSIS VI"1
A. Environmental Impacts VI-1
B. Other Impacts VI-1
-------
7-v
APPENDICES OF REPORT 7
Page
A PHYSICAL PROPERTIES OF ORGANIC RAW MATERIALS, END PRODUCTS A-l
AND BY-PRODUCTS FOR THE LINEAR ALKYLBENZENE PROCESSES
B. EXISTING PLANT CONSIDERATIONS B_l
C. LIST OF EPA INFORMATION SOURCES C"1
-------
Number
IV-3
7-vii
TABLES OF REPORT 7
Page
II-l Linear Alkylbenzene Usage and Growth II-2
II-2 Linear Alkylbenzene Capacity I]:~3
IV-1 LAB Olefin Model-Plant Storage IV~3
IV-2 Benzene and Total VOC Uncontrolled Emissions, IV'4
LAB Olefin Process
LAB Chlorination Model-Plant Storage IV~6
IV-4 Benzene and Total VOC Uncontrolled Emissions, IV-"7
LAB Chlorination Process
V-l Benzene and Total VOC Controlled Emissions, LAB V-3
Olefin Process
V-2 Benzene and Total VOC Controlled Emissions, LAB Chlorination V-5
Process
VI-1 Environmental Impact, LAB Olefin, Controlled VI-2
VI-2 Environmental Impact, LAB Chlorination, Controlled VI-3
A-l
A-l Physical Properties of Benzene
A-2 Physical Properties of n-Paraffins A~
A-3 Physical Properties of Linear Alkylbenzene A~3
A-4 Physical Properties of LAB By-Products A~4
B-l Control Devices and Techniques Used in LAB Olefin Process B-2
B-2 Control Devices and Techniques Used in LAB Chlorination B-3
B-3 Estimated Emissions from Monsanto LAB Plant B~4
B-4 Estimated Emissions from Union Carbide LAB Plant B-5
B-5 Estimated Emissions from Conoco LAB Plant B~7
-------
7-ix
FIGURES OF REPORT 7
Number Page
II-l Location of Plants Manufacturing LAB II-4
III-l Flow Diagram for LAB Olefin Process III-3
III-2 Flow Diagram for LAB Chlorination Process 111-10
-------
1-1
ABBREVIATIONS AND CONVERSION FACTORS
EPA policy is to express all measurements used in ager.cy 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"e
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
10
12
io6
io3
io~3
io"6
Example
1 Tg = 1 X IO12 grams
1 Gg = 1 X IO9 grams
1 Mg = 1 X IO6 grams
1 km = 1 X IO3 meters
1 mV = 1 X IO"3 volt
1 ug = 1 X 10~6 gram
-------
II-l
II. INDUSTRY DESCRIPTION
A. REASON FOR SELECTION
Linear alkylbenzene (LAB) production was selected for consideration because
preliminary estimates indicated that emissions of volatile organic compounds
(VOC) are relatively high and that the predominant manufacturing process emits
significant quantities of benzene, which was listed as a hazardous pollutant by
the EPA in the Federal Register on June 8, 1977. This report has been changed to
an abbreviated format because the data received during its preparation indicate
that benzene emissions from a new LAB plant can be satisfactorily controlled and
because of the low vapor pressures of all the other VOC used in LAB manufacture.
LAB is a viscous liquid with low vapor pressure at ambient conditions. It is
normally processed at elevated temperatures, where the viscosity is lower and the
vapor pressure is higher. Benzene, the predominant emission, is a volatile
liquid at ambient conditions but is emitted as a gas. (See Appendix A for
pertinent physical properties.)
B. USAGE AND GROWTH
Table II-l (refs. 1—3) shows LAB usage and growth rate. The predominant end use
for LAB is in the manufacture of linear alkyl sulfonate for use in synthetic
detergent formulations.
The domestic LAB nameplate production capacity for 1979 was reported to be
1 2
304,000 Mg, with 93% of this capacity being utilized. ' Actual production plant
capacities vary with product mix and operating conditions. With the planned new
LAB capacity announced by Conoco for 1982 there should be sufficient capacity to
supply domestic demand through 1994 if it grows 2% annually as projected.
C. DOMESTIC PRODUCERS1'2
As of 1980 there were four domestic producers of LAB. Table II-2 (refs. 1,2)
lists the producers, the plant locations, rnd the processes being used; the plant
locations are shown in Fig. II-l. Approximately 36% of the 304,000-Mg/yr
domestic capacity is based on the olefin conversion process wherein n-paraffin
feedstock is dehydrogenated to mono-olefins before alkylation with benzene to
LAB. The rest of the domestic capacity uses the chlorination process, wherein
the n-paraffin feedstock is chlorinated to mono-chloroparaffin before alkylation
-------
II-2
Table ll-l. Linear Alkylbenzene (LAB) Usage and Growth
Year
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
Production
(Gg/yr)
218
218
253b
240
251
249°
238°
226d
242
22 46
245
239d
239d
284
Growth
(%/yr)
3.3
-5.2
4.5
-0.5
-4.7
-5.0
7.0
-7.1
9.1
-2.4
18.8
See refs 1-3.
temporary production spurt caused by a fire in the Shell Nederland Chemie NV
wax cracking plant at Pernis, The Netherland.
CExport shipments to Europe dropped when new Spanish LAB plant became operational.
dTight supplies of raw materials, both chlorine and benzene in 1973 and
n-paraffin in 1977 and 1978, limited production.
Recession.
-------
II-3
Table II-2. Linear Alkylbenzene (LAB) Capacity
-" -
Company
b,c
Conoco, Inc.
d
Monsanto Co .
d
Union Carbide Corp.
b
Whit co Chemical Corp.
Location
Baltimore, MD
Alvin, TX
Institute, WV
Carson , CA
1980
Capacity
(Gg/yr)
109
109
66
20
Process Type
Paraffin chlori-
nation
Olefin (paraffin
dehydrogenation)
Paraffin chlori-
nation
Paraffin chlori-
nation
aSee refs 1,2.
bPart of the LAB produced is converted to LAS in an adjoining sulfonation
facility; the rest of the LAB is sold to other companies for conversion to
LAS
c
:Conoco has announced that it will build a new 68-Gg/yr LAB plant at Lake
Charles, LA, with completion expected in 1982.
dAll the LAB produced by these manufacturers is sold to other companies for
conversion to LAS.
-------
(1) Monsanto Co., Alvin, TX
(2) Conoco Chemicals Div., Baltimore, MD
(3) Union Carbide Corp., Institute, WV
(4) Witco Chemical Corp., Carson, CA
Fig.
II-l.
Locations of Plants Manufacturing LAB
-------
II-5
with benzene to LAB. Data are not availabale on the comparative economics of
these two production routes for the manufacture of LAB.
Prior to 1966 the principal alkylate used for manufacture of synthetic detergents
was a branched-chain material produced by the alkylation of propylene tetramer
with benzene. Sulfonation of this alkylate produced a cheap and effective alkyl-
benzene sulfonate (ABS) detergent used in most of the synthetic detergent formula-
tion. Since ABS is resistent to biodegradation, governmental regulations forced
the detergent industry to switch to LAB as the alkylate material for sulfonation
to detergent alkylate. Linear alkylate sulfonate (LAS) produced from LAB is much
more biodegradable in natural water systems than the branched-chain alkylate
sulfonate (ABS) it replaced.
The manufacture of LAS-based synthetic detergents based on the use of LAB is a
mature industry with small growth potential. Newer detergents are coming on the
market. These new synthetic detergents are based on linear paraffin sulfonates
and the nonionic, ethoxylated mixed linear alcohols. LAB is expected to continue
its dominant role in synthetic-detergent manufacture, but the newer detergent
materials are taking over the growth portion of the detergent market.
-------
II-6
D. REFERENCES*
1 L^yyi^S sssr^B'i-Si^-ss?^:
(September 1980).
2 • RshJ:ass-ss^
(January 1979).
5022H in
CA
3.
•« 1 orated at the end of a paragraph, it refers to LUC.
Usually, when a reference is located at tne certain portions of that
entire paragraph. If anotherKreferen^"eated on the material involved. When the
paragraph, that reference number is indicated on the ma ^
reference appears on a heading, it refers to all tne
heading.
-------
III-l
III. PROCESS DESCRIPTION
A. INTRODUCTION
Two major processes are used to manufacture linear alkylbenzene (LAB) in the
United States. Approximately 64% is manufactured by three companies using the
paraffin chlorination process, and approximately 36% is manufactured by one com-
pany using the olefin (paraffin dehydrogenation) process (see Table II-2). The
projected growth rate for the domestic total LAB market is only 2% per year.
The only significant foreign process not normally used in the United States uses
as feedstock the linear alpha olefins produced by Shell's wax cracking process
(Shell Nederland Chemie NV, Pernis, The Netherlands). These linear alpha olefins
are alkylated with benzene at several locations to produce a linear alkylbenzene
(LAB), but the LAB from linear alpha olefins produces a detergent with a slightly
different balance of detergent properties. When n-paraffins were in short supply
during the late 1970s, linear alpha olefins were used as raw material for LAB in
1 2
the United States. '
B. OLEFIN PROCESS
. . 1--5
1. Basic Reactions and Process Description
LAB is produced from n-paraffins (CIQ to C^ mixtures) and benzene in a two-step
sequence of reactions. In the first step n-paraffins are dehydrogenated to
n-olefins by passing hot, vaporized paraffins through a catalyst bed, where
hydrogen is split off from the paraffin molecule, leaving an olefinic double
bond.1'3 A simple illustration of this reaction is
[R and R represent groups of various chain lengths, from a minimum of
hydrogen to cnH2n+1 (alkyl).]
The resultant olefin mixture contains some alpha olefins (10 to 30%), as well as
a mixture of internal olefins, unreacted paraffins, some diolefins, and lower
molecular weight "cracked" materials. Space velocities are high and residence
times are low through the catalyst bed to minimize the amount of isomerization,
-------
III-2
polymerization, coking, and chain scission that can occur. The exit gas mixture
is quenched by contact with a cold liquid stream to minimize thermally promoted
side reactions after the vapor exits from the catalyst bed.
Reaction conditions are selected to achieve an economic balance between the
amount of unreacted paraffin left in the olefin mixture and the amount of mate-
rial degraded to low-molecular-weight oils and residual coke.
Gas separated from the reaction product consists of hydrogen and low-molecular-
weight hydrocarbons such as methane, ethane, ethylene, propane, etc. This gas
can be used as fuel in the process burners, can be piped to an auxiliary process
that uses hydrogen, or can be vented to a flare stack. The most common practice
is to use the mixed gas stream as a process fuel.
The process flow diagram shown in Fig. III-l was developed from literature
sources to illustrate the olefin process. Some variations from the flowsheet in
Fig. III-l exist in current industrial practice, but it is accurate enough to be
useful for air emission evaluations.
In the second reaction step benzene is reacted with the olefin stream from the
first reaction step in the presence of an alkylation catalyst to form the linear
alkylbenzene. A simple illustration of this reaction is
R CH=CHRQ + > R CH -CHR
1 2 L I) 1
The benzene is dried by azeotropic distillation to remove all traces of water
before the above reaction occurs. In the alkylation reactor the benzene, olefin,
and alkylation catalyst are blended intimately and held at reaction conditions
long enough for the alkylation reaction to go to completion. Hydrogen fluoride
is the catalyst of choice for alkylation of benzene with linear olefins, since
-------
MI H f flVKK.
T/utfC
Fig. III-l. Flow Diagram
for LAB olefin Process Model Plant with Uncontrolled Emissions
-------
III-4
yields are higher with hydrogen fluoride than with either sulfuric acid or alumi-
num chloride. A large excess of benzene is used in the reaction mixture to mini-
mize the formation of polyalkylated benzenes.
After alkylation, a settler is used to separate the liquid hydrogen fluoride from
the hydrocarbon product stream. The hydrogen fluoride layer is then recycled to
the alkylation reactor along with fresh makeup hydrogen fluoride.
The hydrocarbon layer is fed to a series of four distillation columns for separa-
tion and recovery of the various components. The benzene is stripped off first
and returned to the benzene feed storage tank. The vent from the benzene strip-
ping column does contain some hydrogen fluoride vapor, as well as some volatile
organic chemicals (VOC), predominantly benzene.
A lime-water scrubber system is used to remove hydrogen fluoride from the vent
gases, since hydrogen fluoride vapor is both toxic and corrosive. Some VOC is
also condensed and absorbed in this scrubber system.
The second distillation column removes unreacted paraffin from the product for
recycle to the paraffin feed tank.
The third distillation column recovers a by-product from the main product stream.
This by-product is stored and sold.
The fourth distillation column recovers and purifies the main LAB product from
the plant, which is stored and/or sold. The bottoms residue from this last
distillation column is stored and sold separately as a heavy by-product.
2. Main Process Vents
There are six main process vents as described below:
a Combustion Vent - The combustion gas vent from the catalytic furnace discharges
the products of combustion generated by burning plant fuel gas or natural gas in
the furnace combustion chamber. Since the oxygen (air) intake to the combustxon
chamber is well above stochiometric levels needed for combustion (2 to 3 tunes
theoretical) and since combustion chamber temperatures run above 900°C, combus-
tion is complete and emissions do not contain measurable quantities of VOC.
-------
III-5
b. Benzene Azeotrope Column Vent A -- The vent after the condenser on the benzene
azeotrope column does contain significant levels of benzene vapor. The amount of
benzene emitted here is influenced by the amount of noncondensables (inert gases
and air) venting from the column and by the operating temperature and design of
the reflux condenser.
c. Hydrogen Fluoride Scrubber Vent A,, — This vent is the discharge vent from the
hydrogen fluoride scrubber. The amount of VOC emitted here is influenced by the
inert gases and air venting from this scrubber system, along with the operating
temperature of the scrubber fluid, the solubility of the VOC in the scrubber
fluid, and the purge rate of the scrubber fluid. The vent gases from the hydro-
gen fluoride scrubber go to a flare, which acts as an emission control device.
Paraffin stripping column vent A -- The paraffin stripping column operates under
a vacuum of 24 kPa absolute, and the column is vented through a steam jet to the
atmosphere. Any VOC that exit from the vacuum line after the vent condenser
would be discharged to the atmosphere. Air or inert gases that enter the column
and exit through the vacuum line would sweep VOC with the noncondensables.
Operating temperature and design of the vent condenser influence the amount of
VOC emitted. The reboiler furnace on the column emits direct combustion products
to the atmosphere. Fuel for this furnace is plant fuel gas or natural gas.
Lights stripping column vent A -- The lights stripping column operates under a
vacuum of 13.3 kPa absolute, and the column is vented through a steam jet to the
atmosphere. Any VOC that exit from the vacuum line after the vent condenser
would be discharged to the atmosphere. Air or inert gases that enter the column
and exit through the vacuum line would sweep VOC with the noncondensables.
Operating temperature and design of the vent condenser influence the amount of
VOC emitted. The reboiler furnace on the column emits combustion products from
the direct combustion of plant fuel gas or nacural gas.
LAB product column vent A,. -- This vent operates under a vacuum of 1.3 kPa
absolute, with a two-stage steam jet with intercondenser used as the vacuum
source. The discharge from the primary jet is condensed and discharged as waste-
water, and the secondary jet discharges to the atmosphere. Any VOC that exit
from the vacuum line after the vent condenser would be condensed with the jet
-------
III-6
condensate or be vented to the atmosphere. (It is estimated that almost all the
VOC in the vacuum line would be condensed and discharged as a wastewater contam-
inant, probably as an oily film on the water.) Air or inert gases that enter the
column and exit through the vacuum line would sweep VOC with the noncondensables.
Operating temperature and design of the vent condenser influence the amount of
VOC emitted. The reboiler furnace on the column emits combustion products
directly to the atmosphere. Fuel for this furnace is plant fuel gas or natural
gas.
Other Emission Sources
Fugitive leaks throughout the process can emit benzene, paraffin, olefin, LAB,
by-products, or hydrogen fluoride. Corrosion can occur in the alkylation section
wherever moisture from air or water lines contact streams containing hydrogen
fluoride. Benzene distillation columns operate above atmospheric pressure.
Pressure in the process side of the reflux condenser may be higher than the pres-
sure in the cooling-water side of the reflux condenser. Any leaks in heat ex-
changers where the pressure of the organic side is higher than the pressure on
the water side would permit the cooling water to be contaminated with VOC. This
VOC would eventually be released into the atmosphere from the cooling tower
system.
Storage and handling emission sources (labeled C on Fig. III-l) include benzene,
paraffin, olefin, LAB, and by-products.
There are five potential sources of secondary emission (labeled K on Fig. III-l):
the hydrogen-hydrocarbon gas from the compressor on the denydrogenation (paraffin
to olefin) system, the wastewater from the benzene azeotrope column receiver, the
wastewater from the hydrogen fluoride scrubber system filter, the wet solids from
the hydrogen fluoride scrubber system filter, and the wastewater from the LAB
column jet condenser. The hydrogen-hydrocarbon gas from the dehydrogenation
system is a satisfactory fuel in the direc.-fired furnaces of the catalytic
furnace and the direct-fired reboilers of the three columns, replacing natural
gas as fuel for these units. Since this process gas burns cleanly and complete-
ly, no VOC is emitted when the gas is used as fuel. The wastewater from the
benzene azeotrope column receiver is saturated with benzene. The amount of waste-
water from the azeotrope column receiver is fixed by the amount of water in the
-------
III-7
benzene raw material purchased for use in this plant. The wastewater from the
hydrogen fluoride scrubber system filter normally contains a mixture of VOC, pre-
dominantly benzene with some paraffin, olefin, LAB, etc. The solids from the HF
scrubber system filter are discharged at the rate of about 9000 g/Mg of product
(dry basis). Washing the filter cake with fresh water will transfer most of the
VOC to the wastewater stream. The wastewater from the LAB jet condenser con-
tains very low levels of VOC. This wastewater stream is added to the other plant
wastewater streams.
4. Process Variations
There are many possible variations of the paraffin dehydrogenation step. (Exist-
ing plant considerations are given in Appendix B.) Various catalysts can be used
to accelerate dehydrogenation, and one version (thermal) can dehydrogenate paraf-
fins without a catalyst. Reaction times and temperatures vary, depending on the
catalyst used. The reaction technique and type of catalyst can also change the
amount of paraffin to olefin conversion and the amount of side reactions that
occur. If a large amount of low-molecular-weight by-products is formed, these
impurities may have to be stripped before the output stream is sent to alkyla-
3—5
tion.
Alkylation can be catalyzed by various catalyst systems, such as hydrogen fluor-
ide, sulfuric acid, and aluminum chloride. Reaction conditions and process ves-
sel design can also influence the rate of alkylation and the amount of side reac-
tions. For olefin alkylation with benzene, hydrogen fluoride is the catalyst of
choice, since yields are higher and side reactions are lower than with other
catalysts.2 Excess benzene (usually 3 to 5 times theoretical quantity) is used
2 ,4
to minimize the formation of polyalkylbenzene.
Alkylation catalyst selection, in turn, dictates the type of system used for
catalyst removal. Hydrogen fluoride can be separated from the product stream by
settling and decantation; hydrogen fluoride is too expensive to be discarded,
recovery and recycling are necessary. The hydrocarbon layer is saturated with
dissolved hydrogen fluoride, which must be removed by a distillation opera-
2—4
tion.
-------
III-8
Product cleanup is necessary after the alkylation step. Multiple distillation
will separate the various hydrocarbon fractions. Benzene is normally removed
first in a benzene stripping column. Residual hydrogen fluoride vapors are
emitted during the benzene distillation, and will sweep some benzene vapor with
them as they exit from the benzene stripping system. A hydrogen fluoride
scrubber system must be used to remove hydrogen fluoride vapors from the vent
stream, since hydrogen fluoride is too toxic and corrosive to be vented to the
atmosphere.2 After benzene is removed, vacuum stripping distillation is used to
remove residual paraffin for recycle. A second vacuum distillation at lower
absolute pressure is used to remove a by-product fraction. A third distillation
at even lower absolute pressure is used to separate the main LAB product stream
from a "heavies" by-product fraction. Various distillation schemes and various
designs of distillation towers can be used to accomplish this separation of the
alkylate hydrocarbon into various recycle, by-product, and product fractions.
Some VOC will be emitted by the vent lines or vacuum systems used on each distil-
lation column.
C. LAB CHLORINATION PROCESS
. . 1,6 — 8
1. Basic Reactions and Process Description
LAB is produced from n-paraffins (CIQ to C^ mixtures) and benzene in a two-step
sequence of reactions. In the first step, dry n-paraffins are reacted with
gaseous chlorine to form n-chloroparaffins and by-product HC1. Ultraviolet light
is used to promote the reaction. A simple illustration of this reaction is
VCH2~R2 + C12 * Rl"(|H"R2 + HC1 + hSat
Cl
[R and R2 represent groups of various chain lengths, from a minimum of hydrogen
to^ H 2 (alkyl)]. An excess of n-paraffin is used in this reaction step to
minimized formation of chloroparaffin with more than one chlorine attached to
a single paraffin chain. Reactants and equipment are kept "dry" to minimize the
corrosive attack of wet hydrogen chloride on metallic equipment.
in the second reaction step dry benzene is reacted with the crude chloroparaffin
mixture in the presence of aluminum chloride catalyst to form linear alkylbenzene
(LAB). A simple illustration of this reaction is
-------
III-9
HC1 •* heat
"2 ^ >J 1 I 2
Cl
An excess of benzene is used in this reaction step to minimize the formation of
polyalkylbenzenes. In addition to by-product hydrogen chloride, other degrada-
tion products and by-products are formed. Some of these by-products and degrada-
tion products are olefins, short-chain paraffins, short-chain alkylbenzenes, poly-
alkylbenzenes, and miscellaneous "tars."
The process flow diagram shown in Fig. III-2 was developed from open literature
sources to illustrate the paraffin chlorination process for the manufacture of
linear alkylbenzene (LAB).
After alkylation the catalyst sludge is separated from the crude LAB by settling.
The catalyst sludge is then hydrolyzed with water to separate the water-soluble
aluminum chloride from the organic materials in the sludge. The organic materi-
als recovered after hydrolysis are a complex mixture of benzene, LAB, and various
degradation products or tars. Since the tars content is high, this stream of
organic materials is collected and used for fuel or is sold.
The crude LAB is washed with alkaline water to neutralize it and is then sepa-
rated from the alkaline wash by decanting. The crude LAB is washed again with
water and is then separated from the water layer by another decanting operation.
The water layers from the hydrolysis and washing steps are sent to the plant
wastewater treatment facility.
After the washing step, the crude LAB is sent through a series of distillation
columns to separate the crude LAB mixture -nto its various components.
The first distillation column operates at atmospheric pressure and strips resid-
ual benzene out of the crude LAB mixture. This recovered benzene is returned to
the benzene feed tank.
-------
©
A*>K £
i«. III-2. Flow
Diagram for LAB Chlorination Process Model Plant with Uncontrolled Emissions
-------
III-ll
The second distillation column operates under vacuum and strips residual n-paraf-
fin out of the crude LAB mixture. The recovered n-paraffin is returned to the
n-paraffin feed tank.
The third distillation column operates under vacuum and strips "light oil" (a
low-molecular-weight mixture of alkylbenzene and tars) out of the crude LAB mix-
ture. This recovered light oil is either sold as a lubricating oil basestock or
is burnt as fuel.
The fourth distillation column operates under vacuum and separates the LAB prod-
ucts from the bottoms or residual high-boiling materials. The overhead LAB prod-
uct is stored for final treatment. The bottoms are collected and sold as deter-
gent base stock for use in the manufacture of motor oil additives.
The overhead LAB product is passed through a treatment system for removal of
residual impurities and colored materials. After this final treatment the
finished LAB is shipped to detergent manufacturers for conversion to linear alkyl-
benzene sulfonate (LAS) and incorporation into finished detergent formulations.
r o
2. Main Process Vents
The main process vents from the chlorination process are as described below:
a. Paraffin Drying Column Vent -- This vent is normally interconnected with storage
tank vents in a connected vent system so that direct discharge from the column to
the atmosphere does not occur.
b. Benzene Azeotrope Column -- The quantity of benzene in this column vent will vary
depending on the wetness of the benzene feed to the azeotrope column and on the
design of the azeotrope column condenser.
c. HC1 Absorber System -- The hydrogen chloride gas out of the VOC absorber system
carries some VOC with it, and the acid absorber is normally operated to minimize
the quantity of VOC dissolved in the aqueous hydrochloric acid. The quantity
will vary, depending on the temperature of the fluid in the VOC absorber and the
vapor pressure of the mixed absorber fluid. Some of the VOC could be absorbed in
the aqueous hydrochloric acid and then be removed from the acid stream.
-------
111-12
e.
Post-Alkylation Treatment Vents -- These vents include vents from catalyst
settling, catalyst hydrolysis, catalyst hydrolysis decanting, product neutrali-
zation, product neutralization decanting, product washing, and product washing
decanting. The seven vents are tied together with one common vent line that is
padded with nitrogen. A conservation vent on this nitrogen-padded common vent
line does discharge some VOC to the atmosphere.
Benzene Stripping Column Vent - This vent can contain significant amounts of
benzene vapor. The amount of benzene vented here is influenced by the amount of
noncondensables (inert gases) venting from the column, and by the operating tem-
perature and design of the reflex condenser.
f. Vacuum Pump (or Steam Jet Vent) on Paraffin Stripping Column -- This vent dis-
charges the vapors from the column to the atmosphere. The discharge amount is
influenced by the air in-leakage into the column and by the design and operating
temperature of the reflux condenser.
g. "Liaht Oil" Stripping Column Vacuum Pump (or Steam Jet Vent) - The amount of VOC
contained in this vent stream varies, depending on design and operating condi-
tions.
h. LAB Product Column Vacuum Pump (or Steam Jet Vent) -- This vent discharges the
vapors from the column to the atmosphere. The amount of VOC discharged is in-
fluenced by the air in-leakage into the column and by the design and operating
temperature of the column reflux condenser.
3. Other Emission Sources
Fugitive leaks throughout the process can emit benzene, paraffin, chloroparaffin,
LAB, by-products, chlorine, or hydrogen chloride. Corrosion can occur in the
chlorination and alkylation sections wherever moisture from air or process
streams contact a process stream containinc chlorine or hydrogen chloride. In
some production plants benzene distillation columns operate above atmospheric
pressure Pressure in the reflux condenser may be higher than the pressure in
the water cooling the condcuser. Leaks in heat exchangers where the pressure on
the organic side is higher than the pressure on the water side would permxt con-
tamination of the cooling water with VOC. The VOC would eventually be released
into the atmosphere from the cooling tower system.
-------
111-13
Storage and handling emission sources (labeled C on Fig.III-2) include benzene,
r o
paraffin, chloroparaffin, LAB, and by-products.
There are five potential sources of secondary emissions (labeled K on Fig. 111-2}-.
the VOC-contaminated wastewater discharged from the n-paraffin drying column, the
VOC-contaminated wastewater discharged from the benzene azeotrope column, the
VOC-contaminated wastewater discharged from the catalyst hydrolysis decanter
tank, the VOC-contaminated wastewater discharged from the neutralization decanter
tank, and the VOC-contaminated wastewater discharged from the wash decanter tank.
If steam jets with aftercondensers were used as vacuum pumps on the four vacuum
distillation columns, the condensate from these steam jets would be contaminated
with VOC and would constitute additional sources of potential secondary emis-
sions. These various sources of wastewater will all carry dissolved and sus-
pended VOC. They can be combined and sent to a plant wastewater treatment
facility, but some of the VOC will escape to the air in the treatment plants.
4. Process Variations
There are many possible minor variations of the LAB chlorination process. (See
Appendix B for existing plant considerations.) Various reaction conditions, the
concentration, the use of pure versus impure chlorine, the reactor design, the
use of reaction accelerators (such as ultraviolet lamps), and the techniques used
to absorb the heat of reaction all influence the performance of a facility for
LAB via chlorination. Additional factors that affect plant performance is the
technique used to remove VOC from the exit stream of hydrogen chloride gas and
the technique used to convert the by-product hydrogen chloride to salable or
useful forms or to otherwise dispose of this acid gas.
Production of LAB via chloroparaffin can be handled in several ways. Chloroparaf-
fins can be refined or separated from unreacted n-paraffins before alkylation, or
the crude reaction mixture can be alkylated before the refining steps are taken.
Another possible reaction route involves t,ie conversion of chloroparaffins to
olefins in a separate reaction step before the olefins are alkylated with ben-
zene. The olefins could be refined or purified before alkylation. The commer-
cial practice normally is the one-step approach, in which crude chloroparaffins
are alkylated with excess benzene in the presence of aluminum chloride complex as
the alkylation catalyst. Refining and separation then take place after alkyla-
-------
111-14
tion. The techniques used to remove the heat of reaction and those used to
remove VOC from the by-product hydrogen chloride gas also affect process results.
Again, the techniques used to convert the by-product hydrogen chloride to useful
or salable forms or to otherwise dispose of this acid gas also influence plant
performance.
Distillation techniques are used for separation of the various fractions in the
crude alkylate product. The distillation columns can vary in design and oper-
ating technique, but high temperatures and low pressures are needed for effective
6--8
separation into useful fractions, recycle materials, and by-products.
As an alternate to the use of a treatment system, some manufacturers react the
distilled (overhead) LAB with sulfuric acid and caustic solutions to remove the
fi — — 8
residual impurities and colored materials.
-------
111-15
D. REFERENCES*
1. R. F. Modler et al
Chemical Economics
al., "Normal Paraffins (C -C )," pp. 683.5022D—683.5022H in
cs Handbook, Stanford Research Institute, Menlo Park, CA
(September 1980).
2. R. G. Hoy, "Olefins, Higher," pp. 321--326 in Kirk-Othmer Enclyclopedia of
Chemical Technology, 2d ed. , vol. 14, edited by A. Standen et al., Interscience,
New York, 1967.
3. C. A. Peterson, IT Enviroscience, Trip Report for Visit to Monsanto Chemical
Company, Inc., Alvin, TX, Nov. 8, 1977 (on file at EPA, ESED, Research Triangle
Park, NC).
4. R. H. Rosenwald, "Alkylation," pp. 890, 891 in Encylcopedia of Chemical
Technology, 2d ed., vol. 1, edited by A. Standen e_t al., Interscience, New York,
1963.
5. W. L. Nelson, "Petroleum Refinery Processes," pp. 18--31 in Kirk-Othmer
Encyclopedia of Chemical Technology, 2d ed., vol. IS, edited by A. Standen e_t
al., Interscience, New York, 1968.
6. C. A. Peterson, IT Enviroscience, Trip Report for Visit to Union Carbide
Corp., Institute, WV, Dec. 8, 1977 (on file at EPA, ESED, Research Triangle Park,
NC) .
7. Letter dated Feb. 17, 1978, from D. J. Lorine, Conoco Chemicals, Continental Oil
Company, Inc., to D. R. Goodwin, EPA, Research Triangle Park, NC.
8. Letter dated Feb. 6, 1978, from E. A. Vistica, Witco Chemical, to D. R. Goodwin,
EPA, 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 head-
ing.
-------
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, par-
ticipate 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 emissions calculated for the LAB olefin process model plant* are
based on information received from Monsanto, the only operator of the LAB olefin
process in the United States. The process emissions calculated for the LAB
chlorination process model plant are based on information received from Union
Carbide, Conoco, and Witco, the three companies that operate chlorination process
plants in the United States, and on data received from MCA, the State of Maryland
Environmental Health Administration, and on data from EPA from a testing program.
The emission quantities reported vary widely from plant to plant.
A. LAB OLEFIN PROCESS1'2
1. Model Plant*
The model plant for the LAB olefin process has a capacity of 90 Gg/yr based on
8760 hr of operation per year.** Though this is not an actual operating plant,
it is similar to the one existing plant in the United States. The model LAB
olefin process, shown in Fig. III-l, reasonably conforms with current technology.
A single process train is typical for today's manufacturing and engineering
technology. The model process dehydrogenates n-paraffins to n-olefins and then
reacts the n-olefins with benzene, with hydrogen fluoride used as the catalyst,
to produce LAB.
*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 corre-
spondingly 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
Typical raw material, intermediate, by-product, and product storage-tank capaci-
ties are estimated for a 90-Gg/yr plant. The storage-tank requirements are given
in Table IV-1.
2. Sources and Emissions
All estimated process emission rates and sources for the LAB olefin process are
summarized in Table IV-2.
a. Benzene Azeotrope Column Vent -- This column vent releases some benzene into the
atmosphere. All benzene used in the process passes through the azeotrope column
for removal of traces of water from the benzene. Since benzene freezes at 5.5°C
(42°F), the column condenser must be operated above this temperature. At normal
condenser temperatures of about 27°C (80°F), benzene has a vapor pressure of
13.7 kPa, and some benzene is normally lost out of the column vents.
b. Hydrogen Fluoride Scrubber Vent -- The largest process vent is the hydrogen
fluoride scrubber vent. This scrubber receives vent gas from the alkylator and
the benzene stripping column. In the uncontrolled model plant these process ve'nt
streams contain significant quantities of benzene and other VOC, as well as
hydrogen fluoride vapor and system nitrogen purge gas. The hydrogen fluoride
scrubber removes hydrogen fluoride from the vent stream by reacting it with
alkaline (calcium hydroxide) scrubber water. Benzene and other VOC condense in
this scrubber water. The scrubber normally operates at 32 to 38°C. The nitrogen
is purged through the alkylation system at a flow rate of about 1.7 m /hr to
prevent the backflow of water vapor into any of the system components. This flow
of purge gas sweeps volatile benzene vapor out of the scrubber vent at an esti-
mated rate of about 0.11 kg/hr. This is the largest process loss of benzene to
the atmosphere. The increased use of nitrogen purge gas during startups or
shutdowns, as well as process upsets, can drastically increase this normal loss
rate by a factor of 5 to 10.
c. Vacuum Refining Column Vents -- The three product refining columns that operate
under vacuum discharge the exhaust gases from their vacuum pump (steam jets)
vents directly to the atmosphere. Since these columns operate at high head
temperatures, the main column condensers must operate hot to prevent cooling of
the reflux that is returned to the top of the columns. Auxiliary vent condensers
-------
IV-3
Table IV-1. LAB Olefin Model-Plant Storage (Organics Only)
— — — —
Contents
n-Paraffin (11) S (bulk)
n-Paraffin (12) a (bulk)
n-Paraffin (13) a (bulk)
n-Paraffin (feed)
n-Paraffin (feed)
n-Olefin (feed)
n-Olefin (feed)
Benzene (bulk)
Benzene (feed)
Benzene (dry feed)
By-product (receiver)
By-product (bulk)
Heavies (receiver)
Heavies (bulk)
LAB (receiver)
LAB (receiver)
LAB (11 )a (bulk)
LAB (12 )a (bulk)
LAB (13) a (bulk)
LAB (11) a (bulk)
LAB (12 )a (bulk)
LAB (13) a (bulk)
Tank Size
(m3)
3200
3200
3200
213
213
213
213
3200
213
213
18
213
18
334
213
213
334
334
334
3200
3200
3200
Turnovers
Per Year
10
10
10
230
230
230
230
13
200
200
150
13
255
14
250
250
15
15
15
11
11
11
Molecular
Weight
164
175
186
175
175
173
173
78
78
78
118
118
420
420
243
243
236
243
261
236
243
261
Bulk Liquid
Temperature
(°C)
32
32
32
32
32
32
32
27
27
27
38
38
43
43
43
43
43
43
43
43
43
43
aAverage chain length.
-------
Table IV-2. Benzene and Total VOC Uncontrolled Emissions for 90-Gg/yr
Model Plant Using the LAB Olefin Process
Source
Benzene azeotrope column vent
Hydrogen fluoride scrubber column
vent
Paraffin stripping column vent
By-product stripping column vent
LAB product column vent
Stream
Designation
(Fiq. III-l)
Al
A2
A3
A4
A5
Emission Ratio (g/Mg)
Benzene Total VOC
3.7 3.7
11 11
88
1
0.0014
Emission Rate (kg/hr)
Benzene Total VOC
0.038 0.038
0.11 0.11
0.9
0.01
0.000014
of emission per Mg of LAB produced.
f
-------
IV-5
have been provided on these column vacuum lines to prevent flooding of the vacuum
pumps with hot vapors. Column air leakage and vapor pressure in the vent con-
denser determine the amount of VOC in the vacuum pump vents. Process upsets,
startups, and shutdowns do not have much impact on the VOC emissions from these
vents.
d. Other Emissions -- Storage, fugitive, and secondary emissions for the entire
synthetic organic chemicals manufacturing industry are covered by separate EPA
, . 10—12
documents.
3—9
B. LAB CHLORINATION PROCESS
1. Model Plant
The model plant for this study has a capacity of 90 Gg/yr based on 8760 hr of
operation per year. Although the model plant is not in actual operation, it is
similar in most design features to the three existing plants in the United
States. The model plant is sized midway between the two largest LAB chlorination
process domestic plants. The model LAB chlorination process, shown in
Fig. III-2, is a reasonable concept of current technology. A single process
train is typical for today's manufacturing and engineering technology. The model
process chlorinates n-paraffins to monochlorinated n-paraffins and then reacts
the crude chloroparaffin with benzene in the presence of aluminum chloride
catalyst to form the crude, linear alkylbenzene (LAB) products. Product separa-
tion, distillation, and final purification steps are used for separating and
refining the final LAB products and for removal of the by-products and recycle
materials.
Typical raw material, intermediate, by-product, and product storage-tank capac-
ities are estimated for a 90-Gg/yr plant. The storage-tank requirements are
given in Table IV-3.
2. Sources and Emissions
Estimated process emission rates and sources for the LAB chlorination process are
summarized in Table IV-4.
-------
IV-6
Table IV-3. LAB Chlorination Model-Plant Storage (Organic Only)
Contents
n-Paraffin (II)3 (bulk)
n-Paraffin (12) a (balk)
n-Paraffin (13)a (bulk)
n-Paraffin (feed)
n-Paraffin (feed)
n-Paraffin (dry feed)
n-Paraffin (dry feed)
Crude chloroparaf fin (feed)
Crude chloroparaf fin (feed)
Benzene (bulk)
Benzene (feed)
Benzene (dry feed)
Waste oil (receiver)
Waste oil (bulk)
By-product (receiver)
By-product (bulk)
Heavies (receiver)
Heavies (bulk)
LAB (receiver)
LAB (receiver)
LAB (11 )a (bulk)
LAB (12) a (bulk)
LAB (13) a (bulk)
LAB (11) S (bulk)
LAB (12 )a (bulk)
LAB (13) a (bulk)
Tank
Size
(m3)
3200
3200
3200
213
213
640
640
640
640
3200
870
870
18
213
40
640
80
1420
213
213
334
334
334
3200
3200
3200
Turnovers
Per Year
10
10
10
230
230
225
225
225
225
13
240
240
150
12
200
12
260
15
250
250
15
15
15
11
11
11
Molecular
Weight
164
175
186
175
175
175
175
210
210
78
78
78
118
118
118
118
420
420
243
243
236
243
261
236
243
261
Bulk Liquid
Temperature
(°C)
32
32
32
32
32
32
32
32
32
27
27
27
32
32
38
38
43
43
43
43
43
43
43
43
43
43
Average chain length.
-------
Table IV-4. Benzene and Total VOC Uncontrolled Emissions from LAB
Chlorination Process Used in 90-Gg/yr Model Plant
Source
Paraffin drying column vent
Benzene azeotrope column vent
Hydrochloric acid absorber vent
Atmospheric wash decanter vents
Benzene stripping column vent
b
Vacuum refining column vents
Stream
Designation
(Fig.III-2)
Al
A2
A3
A4
A5
A6
a
Emission Ratio (g/Mg)
Benzene
3.7
250
12.3
3.7
Total VOC
2.8
3.7
250
12.4
3.7
92
Emission Rate (kg/hr)
Benzene
0
2
0
0
.038
.6
.126
.038
Total VOC
0.029
0.038
2.6
0.127
0.038
0.95
ag of emissions per Mg of LAB produced.
bAssumed use of refrigerated vent condensers to minimize venting of VOC vapors through the vacuum pumps on
the vacuum refining column vents.
-------
IV-8
a- n-Paraffin Drying Column Vent -- The n-paraffin drying column operates under
vacuum to keep the still bottoms temperature below the n-paraffin decomposition
range. The primary reflux condenser operates at high head temperature to prevent
subcooled reflux from being returned to the top of the column. An auxiliary vent
condenser is provided to prevent flooding of the vacuum pump with hot vapors.
Column air leakage and vapor pressure in the vent condenser determine the amount
of VOC in the vacuum pump vent. Normal leakage rates were assumed to permit
calculation of estimated emissions. Process upsets, startups, and shutdowns do
not have much impact on the VOC emissions from this vent.
b- Benzene Azeotrope Column Vent -- This column vent releases some benzene into the
atmosphere. All benzene used in the process passes through the azeotrope column
for removal of traces of water from the benzene. Since benzene freezes at 5.5°C
(42°F), the column condenser must be operated above this temperature. At normal
condenser temperatures of about 27°C (80°F) benzene has a vapor pressure of
13.7 kPa, and some benzene is normally lost out of the column vent.
c- Hydrogen Chloride Absorber Vent -- This vent is the largest process vent for the
LAB chlorination process. All of the vent gas from the paraffin chlorinators and
the alkylation reactors is directed first through a volatile organic absorber
system and then through the acid absorber before being discharged to the atmo-
sphere. The amount of VOC in the hydrogen chloride gas going to the acid ab-
sorber is regulated by the performance of the volatile organic absorber system.
All the crude chloroparaffin is used as the absorption fluid in the volatile
organic absorber. The principal VOC that escape from the organic absorption
system is benzene, with some traces of n-paraffin and paraffin degradation prod-
ucts. The acid absorption system operates as an adiabatic absorber, with the
heat of solution of the hydrogen chloride in water raising the temperature of the
acid solution to the boiling point to prevent absorption of VOC in the acid.
(Absorption of VOC in the acid by-product would contaminate the acid with dis-
solved organic material, and its removal would be necessary if the acid were
sold.) The nitrogen purge gas charged to the alkylator escapes through the vent
from the hydrogen chloride absorber, carrying with it the residual VOC that
escapes from the volatile organic absorber system. (Nitrogen is purged through
3
the alkylator at a flow rate of about 1.7 m /hr to prevent backflow of water
vapor into any of the alkylator system components.) Variations in inert-gas
-------
IV-9
content in the chlorine gas used for chlorination also influence this gas flow.
The increased use of nitrogen purge gas during startups or shutdowns, as well as
process upsets, can drastically increase this normal loss rate by a factor of 5
to 10. Benzene emissions from this vent as reported by industry vary from about
5 g/Mg of LAB to over 10,000 g/Mg (see Appendix B).
d. Atmospheric Wash and Decanter Vents -- The series of process vessels used for
settling the catalyst slurry, hydrolyzing the spent catalyst, neutralizing the
organic product stream, washing the organic product stream, and decanting the
various oil layers from the various hydrolysis, neutralization, and wash water
streams are all vented to the atmosphere through conservation vents. Since these
vessels normally operate at constant liquid levels, the only VOC losses are
breathing losses. Startups, shutdowns, and process upsets could drastically
increase this loss by vapor space displacement due to changes in liquid levels.
e. Benzene Stripping Column Vent -- The benzene stripping column operates at atmo-
spheric pressure, and the vent line from the condenser reflux receiver vents to
the atmosphere. Since benzene has a significant vapor pressure (24.3 kPa) at the
column condenser temperature of about 40°C, some benzene vapors are lost to the
atmosphere at this point.
f. Vacuum Refining Column Vents -- The three product refining columns that operate
under vacuum discharge the exhaust gases from their vacuum pump vents directly to
the atmosphere. Since these columns operate at high head temperatures, the main
column condensers must operate hot to prevent subcooled reflux from being
returned to the top of the columns. Auxiliary vent condensers have been provided
on these column vacuum lines to prevent flooding of the vacuum pumps with hot
vapors. Column air leakage and vapor pressure in the vent condenser determine
the amount of VOC in the vacuum pump vents. Process upsets, startups, and shut-
downs do not have much impact on the VOC emissions from these vents.
g. Other Sources -- Storage, fugitive, and secondary emissions for the entire syn-
thetic organic chemicals manufacturing industry are covered by separate EPA docu-
10--12
ments.
-------
IV-10
0. REFERENCESA
1. C. A. Peterson, IT Enviroscience, Trip Report for Visit to Monsanto Industrial
Chemicals Co., Alvin, TX, Nov. 8, 1977 (on file at EPA, ESED, Research Triangle
Park, NC.).
2. Letter dated May 31, 1979, from J. H. Craddock, Manager, Product Safety, Monsanto
Industrial Chemicals Co., St. Louis, MO, to D. R. Patrick, EPA, with comments on
draft LAB report.
3. C. A. Peterson, IT Enviroscience, Trip Report for Visit to Union Carbide
Corp., Institute, WV, Dec. 8, 1977 (on file at EPA, ESED, Research Triangle Park,
NC).
4. Letter dated May 16, 1979, from R. L. Foster, Union Carbide Corp., South
Charleston, WV, to D. R. Patrick, EPA, with comments on draft LAB report.
5. Letter dated Feb. 6, 1978, from E. A. Vistica, Vice President, Witco Chemical
Corporation, Wilmington, CA, to D. R. Godwin, Director, ESED Division, EPA.
6. Letter dated Feb. 17, 1978, from D. J. Lorine, Chief Engineer, Conoco Chemicals
Div., to D. R. Godwin, Director, ESED Division, EPA.
7. Letter dated Apr. 26, 1979, from R. A. Oliver, Public Health Engineer, State of.
Maryland Environmental Health Administration, Baltimore, MD, to D. R. Patrick
ESED, EPA, with comments on draft LAB report.
8. Chemical Manufacturers Association, Review Comments on Draft Linear Alkylbenzene
Product Report (nd).
9. Letter dated Nov. 3, 1978, from J. L. Shumaker, ESED, EPA, to C. A. Peterson, IT
Enviroscience, with preliminary results on the LAB test.
10. D. G. Erikson, IT Enviroscience, Storage and Handling (September 1980) (EPA/ESED
report, Research Triangle Park, NC).
11. D. G. Erikson and V. Kalcevic, IT Enviroscience, Fugitive Emissions (September
1980) (EPA/ESED report, Research Triangle Park, NC).
12. J. Cudahy and R. Standifer, IT Enviroscience, Secondary Emissions (June 1980)
(EPA/ESED report, Research Triangle Park, NC).
^Usually, when a reference 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 tent covered by that head-
ing.
-------
V-l
V. APPLICABLE CONTROL SYSTEMS
A. LAB OLEFIN PROCESS
1. Hydrogen Fluoride Scrubber Vent
The main process vent for the LAB olefin process is the vent from the hydrogen
fluoride scrubber column. The vent gas from this column is rich in benzene
vapor, releasing approximately 11 g of benzene per Mg of LAB product. An emis-
sion control system is the destruction of the hydrogen fluoride scrubber vent
vapors by combustion.
Since a flare stack system would have to be installed for control of emergency
emissions due to process upsets and malfunctions, this hydrogen fluoride scrubber
vent emission stream could be directed to the flare stack for destruction by
combustion. Properly designed flare tips with steam or air injection and con-
tinuous pilot lights can assure combustion of flammable vapors at removal effi-
ciencies of 90%* or better. This control method is used by industry to control
the emissions from a hydrogen fluoride vent (see Appendix B).
Another possible control technique is to use the emissions from the hydrogen
fluoride scrubber vent and the benzene azeotrope column vent as fuel by piping
the vent gases into the plant fuel gas header where one is used. The VOC de-
struction efficiency for this technique can be greater than 99.98%. ' This
control method is used by industry on other processes, and the incremental cost
for connecting the vent to the fuel gas header should be negligible when done as
a new plant is being designed.
n-Paraffin Stripping Column Vent
This secondary process vent for the LAB olefin process is rich in n-paraffin
vapor, releasing approximately 88 g of n-paraffin per Mg of LAB product. For a
90-Gg/yr plant the yearly emissions from this source would be approximately
7.9 Mg.
*Flare efficiencies have not been satisfactorily documented except for specific
designs and operating conditions using specific fuels. Efficiencies cited are
for tentative comparison purposes.
-------
V-2
a- Reduction of Air In-Leakage -- The vent emissions calculated for this n-paraffin
stripping column are based on the assumption of an air in-leakage rate typical of
a normally assembled and maintained vacuum distillation column with no special
precautions or techniques used to achieve better than average column tightness.
Special testing, maintenance, and assembly techniques can be used to reduce this
air in-leakage rate and the resultant VOC emissions.
b- Condensation of Jet Exhaust -- The uncontrolled model plant shows a single-stage
steam jet as the vacuum pump on the n-paraffin stripping column, with the stearn
and entrained vapors discharging directly to the atmosphere. Addition of a
surface condenser to condense and subcool the steam and entrained vapors to about
38°C would remove at least 92% of the n-paraffin vapors from this vent stream.'
The condensate would contain the condensed n-paraffin, and this condensate stream
could be sent to the plant wastewater skimmer system for separation and recovery
of the n-paraffin organic layer. The controlled emissions from this surface
condenser are shown in Table V-l.
3. Other Process Vents
The emissions from the other process vents (lights stripping column and LAB
product column) are already low enough to warrant no further effort to reduce VOC
from these process vents.
4. Fugitive Sources
Controls for fugitive sources are discussed in another EPA report covering fugi-
4
tive emissions from the synthetic organic chemicals manufacturing industry.
5. Storage and Handling Sources
Control of benzene and other VOC storage emissions for the synthetic organic
chemicals industry is covered in another EPA report.
6. Secondary Sources
Controls for secondary emissions from the synthetic organic chemicals industry
are discussed in another EPA report.
-------
Table V-l. Benzene and Total VOC Controlled Emission for
90-Gg/yr Model Plant Using the LAB Olefin Process
Emissions
Benzene azeotrope column
vent
Hydrogen fluoride scrubber
column vent
Paraffin stripping column
vent
Lights stripping column
vent
LAB product column xant
Stream
Designation
(Fig. III-l)
Al
A2
A3
A4
A5
Control Device
or Technique
Used as fuel
Used as fuel
(Alt 1)
Flare (Alt 2)
Surface condenser
b
None
b
None
Total VOC
Emission
Reduction (%)
99.98
99.98
90
92
Ratio (g/Mg) a
Benzene Total VOC
0.00074 0.00074
0.0022 0.0022
1.1 1.1
7.0
1 0
0 0014
Rate (kg/hr)
Benzene Total VOC
0.0000076 0.0000076
0.000023 0.000023
0.011 0.011
0.072
0.01
0.000014
ag of benzene or total VOC per Mg of LAB produced.
DJet exhaust surface condenser recommended for suppression of steam plume.
-------
B. LAB CHLORINATION PROCESS
1. Hydrochloric Acid Absorber Vent
The main process vent for the LAB chlorination process is the vent from the
hydrochloric acid absorber. The vent gas from this absorber column is rich in
benzene vapor, releasing approximately 250 g of benzene per Mg of LAB product.
The only control technique reported for this vent is the operation of the hydro-
chloric acid absorber so that benzene goes with the aqueous acid, followed by
removal of the benzene from the acid by an oil-water separator and activated
carbon. No data were given on the removal efficiency achieved when this tech-
nique is used. Emission of benzene at one plant was reported as 50 g per Mg of
LAB produced (see Appendix B).
Another control technique for this vent is to scrub the vent gases with caustic
and pipe the neutralized vent gases into a plant fuel-gas header if one is used.
The VOC destruction efficiency for this control technique can be greater than
1 2
99.98%. ' This method is used to control alkylation vent gases from the manu-
facture of ethylbenzene. An alternative is to pipe the neutralized vent gases to
a flare, a technique that is used by industry for other processes. The incre-
mental cost for using either of these techniques in a new plant is negligible.
Another possible control technique is the use of carbon adsorption. In order to
use carbon adsorption, the exhaust gas stream must be scrubbed with caustic to
remove acid and water-soluble organics. Benzene is likely the only VOC remain-
ing. Two or more carbon beds are needed since the exhaust stream passes through
one bed while the other bed is being regenerated with steam. The steam conden-
sate is decanted to separate the benzene for recycle to the process, and the
benzene-saturated aqueous layer is sent to waste disposal. This control techni-
que has not been demonstrated on this vent stream, but based on engineering
experience with similar applications it is believed that a carbon adsorption
system can be designed and operated at a sustained removal efficiency of greater
than 99%.8
A removal efficiency of 99.98% for use of the vent gases as fuel has been used to
project the controlled hydiochloric acid absorber vent emissions from the model
plant (Table V-2).
-------
Table V-2. Benzene and Total VOC Controlled Emissions for
90-Gg/yr Model Plant Using LAB Chlorination Process
Stream
Designatior
Source (Fig. Ill- 3
Paraffin azeotrope A^
column vent
Benzene azeotrope A
column vent
Hydrochloric acid absorber A
vent
Atmospheric wash A^
decanter vents
Benzene stripping A^
column vent
Vacuum refining A
column vents
Emissions
Total VOC ^ ^ . , ,„ . a _ . ,,/•,,
„ „ . . Ratio (g/Mg) Rate (kg/hr)
i Control Device Emission
.) or Technique Reduction (%) Benzene Total VOC Benzene Total VOC
Used as fuel 99.98 0.00056 0.0000058
Used as fuel 99.98 0.00074 , 0.00074 0.0000076 0.0000076
Used as fuel 99.98 0.05 0.05 0.00051 0.00051
Used as fuel 99.98 0.0025 0.0025 0.000025 0.000025
Used as fuel 99.98 0.00074 0.00074 0.0000076 0.0000076
Used as fuel 99.98 0.018 0.00019
of benzene or total VOC per Mg of LAB produced.
-------
V-6
2. n-Paraffin Azeotrope Column Vent
A vacuum vent condenser is incorporated in the design for the n-paraffin azeo-
trope column vent to prevent flooding of the vacuum pump with hot n-paraffin
vapors. The VOC remaining that are emitted from the vacuum pump discharge of the
model plant are controlled by being piped to the plant fuel-gas header for use as
1 2
fuel. A VOC destruction efficiency of 99.98% ' was used to calculate the con-
trolled emissions that originate in this vent, as was done for all process vents
in the model plant (see Table V-2). An alternative control technique could
consist in piping the emissions to the emergency flare or to the carbon adsorber
if one of those techniques is used for controlling the hydrochloric acid absorber
vent.
3. Benzene Azeotrope Column Vent
The emission control selected for this vent for the model plant is the use of the
1 2
vent gases as fuel and a VOC destruction efficiency of 99.98% ' was used to
calculate the model-plant controlled emissions from this vent.
4. Atmospheric Wash-Decanter Vents
These series of wash-decant process vessels are tied together by one common vent
line, padded with nitrogen, and terminated with a conservation vent. The emis-
sions from this vent are breathing losses that are controlled in the model plant
1 2
by using them as fuel. A VOC destruction efficiency of 99.98% ' was used to
calculate the model plant controlled emissions from this vent.
5. Benzene Stripping Column Vent
The emission control selected for this vent for the model plant is the use of the
1 2
vent gases as fuel. A VOC destruction efficiency of 99.98% ' was used to calcu-
late the model-plant controlled emissions from this vent.
6 . Vacuum Column Vents
The n-paraffin stripping column vent is rich in n-paraffin vapor, releasing
approximately 88 g of n-paraffin per Mg of LAB product for a 90 Gg/yr plant. The
emissions from the n-paraffin stripping column vent and from the other vacuum
columns are controlled in the model plant by using them as fuel. A VOC destruc-
tion efficiency of 99.98%1;2 was used to calculate the model-plant controlled
emissions from this vent.
-------
V-7
7. Fugitive Sources
Controls for fugitive sources are discussed in another EPA report covering fugi-
tive emissions from the entire synthetic organic chemicals manufacturing indus-
try.
8. Storage and Handling Sources
Control of benzene and other VOC storage emissions for the entire synthetic
organic chemicals industry is covered in a separate EPA report. Information on
LAB manufacturing locations indicates that benzene is stored in both fixed-roof
and floating-roof API style tanks. A floating roof is commonly used to control
storage-tank emissions for VOC in the vapor pressure range of benzene. The vapor
pressures of all the other organic raw materials, intermediates, and finished
products or by-products are low. The vent lines on these storage tanks could be
interconnected and the final output vent sent to some control device or system if
it were cost effective.
9. Secondary Sources
Control of secondary emissions is discussed in a separate EPA report.
-------
V-8
C. REFERENCES*
1. V. Kalcevic, IT Enviroscience, Control Device Evaluation. Flares and the Use of
Emissions as Fuels (in preparation for EPA, ESED, Research Triangle Park, NC).
2. T. Lahre, "Natural Gas Combustion," pp. 1.41—1.4-3 in Compilation of Air Pollu-
tant Emission Factors, 3d ed., Part A, AP-42, EPA, Research Triangle Park, NC
(May 1974).
3. D. G. Erikson, IT Enviroscience, Control Device Evaluation. Condensation
(December 1980) (EPA/ESED report, Research Triangle Park, NC).
4. D. G. Erikson and V. Kalcevic, IT Enviroscience, Fugitive Emissions (September
1980) (EPA/ESED report, Research Triangle Park, NC).
5. D. G. Erikson, IT Enviroscience, Storage and Handling (September 1980) (EPA/ESED
report, Research Triangle Park, NC).
6. J. J. Cudahy and R. L. Standifer, IT Enviroscience, Secondary Emissions
(September 1980) (EPA/ESED report, Research Triangle Park, NC).
7. J. A. Key and F. D. Hobbs, IT Enviroscience, Ethylbenzene and Styrene (September
1980) (EPA/ESED report, Research Triangle Park, NC).
8. H. S. Basdekis and C. S. Parmele, IT Enviroscience, Control Device Evaluation.
Carbon Adsorption (January 1981) (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 rebates 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 IMPACT
Tables VI-L and VI-2 show the environmental impact of reducing VOC emissions by
application of the described control devices or techniques (Sect. V) to new
plants producing 90 Gg/yr of LAB by the model olefin process and by the model
chlorination process respectively. The environmental impacts of controlling VOC
emissions from storage and handling, fugitive, and secondary sources are not
included in the estimates in Tables VI-1 and VI-2 but are believed to be similar
to those from other processes in the synthetic organic chemicals manufacturing
industry.
Based on a projected estimate of 290 Gg of LAB produced in 1980 and on a current
removal efficiency of approximately 10%, a very rough estimate of emissions from
the LAB industry in 1980 is 1200 Mg of benzene and 1400 Mg of total VOC. This
estimate includes process, storage and handling, fugitive, and secondary sources.
If planned retrofitting of emission controls has been completed, the estimate may
be high (see Appendix B), depending on the reduction efficiency actually
achieved.
B. OTHER IMPACTS
Energy and control cost impacts have not been determined for the control tech-
niques selected in Sect. V. The impacts are believed to be negligible when the
techniques are applied during the design of a new plant.
-------
Table VI-1. Environmental Impact of Controlled LAB Olefin 90-Gg/yr Model Plant
Stream
Designation
Source (Fig. III-l)
Benzene azeotrope column A^
vent
HF scrubber column vent AZ
Paraffin stripping column A^
vent
Lights stripping column A4
vent
LAB product column vent A^
Control Device Emission
or Technique Reduction (%)
Used as fuel 99.98
Used as fuel 99.98
Flare 90
Surface condenser 92
None
None
Emission Reduction (Mg/yr)
Benzene Total VOC
0.33 0.33
0.99 0.99
0.89 0.89
7.3
H
1
M
-------
Table VI-2. Environmental Impact of Controlled LAB Chlorination 90-Gg/yr Model Plant
Source
Paraffin azeotrope
vent
Benzene azeotrope
vent
Hydrochloric acid
vent
Stream
Designation
(Fig. IH-2)
column A
column A
absorber A
Atmospheric wash decanter A
vents
Benzene stripping
vent
column A
Vacuum refining column A
Control Device
or Technique
Used as
Used as
Used as
Used as
Used as
Used as
fuel
fuel
fuel
fuel
fuel
fuel
Emission
Reduction (%)
99.
99.
99.
99.
99.
99.
98
98
98
98
98
98
Emission Reduction (Mg/yr)
Benzene Total VOC
0.25
0.33 0.33
22.5 22.5
1.1 1.1
0.33 0.33
<
H
8.3 u
vents
-------
A-l
APPENDIX A
Table A-l. 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
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 20°C/4°C
1.79 g/liter
2.28
From: J. Dorigan et^ al_ . , "Benzene," p. AI-102 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) .
-------
A-2
Table A-2. Typical Physical Properties of n-Paraffins*
Low-Range Mid-Range High-Range
Value Value Value
Chain distribution (%)
Below C
cio
Cll
C12
C13
C14
C15
C16
Average molecular weight
Specific gravity at 60°F
Flash point (°F) (Pensky-Martin)
Melting range (°C)
Viscosity at 60°F (cs)
Distillation range (°F)
Initial boiling point
50%
90%
End point
Abstracted from Conoco Normal Paraffins
<2
16
38
40
6
1
161
0.745
155
-22 to -25
1.78
360
386
408
446
, Conoco Chemicals
<2
1 <1
16 20
51 47
32 23
1 8
2
189 186
0.756 0.767
210 210
-11 to -13 -3 to 0
2.80
435 435
453
468
482 558
Division, Continental
Oil Co., Houston, TX (nd).
-------
A-3
Table A-3. Typical Physical Properties of Linear Alkylbenzenes"
Low-Range Mid-Range
Value Value
Chain distribution (%)
Below C
cio
Cll
C12
C13
C14
C15
C16
Average molecular weight
2-Phenyl isomer (%)
Specific gravity at 60°F
Viscosity at 100°F (cs)
Bromine number
Flash point (°F) (Pensky-Martin)
Distillation range (°F)
Initial boiling point
5%
50%
95%
End point
*
Abstracted from: Conoco Nalkylene 500
Continental Oil Co., Houston, TX (nd) ,
& A230) , Monsanto Industrial Chemicals
Information. UCANE Alkylate 12 Linear
<2
18
32
37
10
>2
238
20 — 40
0.866
4.3
0.003
280-290
536
546
555
578
586
Detergent
<0.5
10
28
39
15
7
<0.5
244
20 — 30
0.865
4.7
0.01
290-300
543
553
563
593
603
Alky late, Conoco Chemicals
High-Range
Value
<0.5
1.5
15
47
34
<3
262
20 — 30
0.865
5.9
0.01
295-305
577
588
597
615
621
Division,
Product Data Sheet (on Alkylate A215, A225,
Co., St.
11 and 12
Louis, MO (December 1976)
,- Product
Alkylbenzene, Union Carbide Corp.,
New York (nd) .
-------
A-4
Table A-4. Typical Physical Properties of LAB By-Products'
Average molecular weight 40°
n Rft? O-888 °'891
Specific gravity at 60°F 0.88J
Flash point («F) (Pensky-Martin) 415-430 380 380
-70 -60
Pour point (°F)
Viscosity
at 60°F 125
at 100°F 14
14
at 122°F
Distillation range (°F)
. . 626 680
Initial boiling point
642 730
5%
682 800
50%
714 975
95%
o LMR (
thvlensoos^n^o
Chemicals Division, Continental Oil Co., Houston, TX (nd) .
__ _ .
Molecular Ratio), Conoco N-B-D (Distilled Total.
DPA (Diphenylalkangl, product bulletins, Conoco
-------
B-l
APPENDIX B
EXISTING PLANT CONSIDERATIONS
1 2--4
Tables B-l and B-2 list the emission control devices and techniques reported
to be in use by the LAB industry. To gather information for this report, two
site visits were made to manufacturers of LAB. Trip reports have been cleared by
1 ?
the companies concerned and are on file at ESED in Durham, NC. ' Some of the
pertinent information concerning process emissions from these existing LAB plants
is presented in this appendix. Other information is from letters to EPA from the
other two companies that produce LAB, in response to requests for information on
3 4
process emissions from their plants. ' Also included is information received
C -I
with comments on the draft Linear Alkylbenzene Report.
A. CONTROLS AT EXISTING PLANTS
1. Monsanto, Alvin, TX1'5
Monsanto is the only operator of the LAB olefin process in the United States and
uses a process developed by Monsanto using refining and reaction principles
originally developed in the petroleum refining industry. See Table B-l for the
emission control devices and techniques used by Monsanto. No measurements of
emissions were reported; however, Monsanto believes its process should not
require additional controls. See Table B-3 for Monsanto's estimates of actual
emission ratios for its process.
2. Union Carbide, Institute, WV '6
The Union Carbide plant uses the paraffin chlorination process for production of
LAB. See Table B-2 for the emission control devices and techniques used by Union
Carbide. Table B-4 gives the emissions reported by Union Carbide. In the Union
Carbide plant the HC1 gas stream from the alkylation reaction is scrubbed with
all the crude chlorinated paraffin to remove benzene and then is sent to lime-
stone "pits," where the HC1 is neutralized. Union Carbide and EPA have sampled
this gas stream and analyzed it for organic content. The reported presence of
relatively large quantities of compounds that cannot be reasonably accounted for
and the inability to calculate a material balance from the data are reasons for
doubting the results of the EPA study. Union Carbide reports that VOC losses
with the wastewater from their LAB process are from 3 to 5 kg of VOC per Mg of
LAB produced. This wastewater goes to their plant wastewater system. Their
-------
B-2
Table B-l. Control Devices and Techniques
Currently Used in the LAB Olefin Process
Source
Stream
Designation
(Fig. III-l)
Emission
Control Devices
and Techniques
Benzene azeotrope
control vent-
Hydrogen fluoride
scrubber column
vent
Paraffin stripping
column vent
Lights stripping
column vent
A,
None
Vent gases sent to flare for
combustion
Vent condenser used to minimize
VOC to vacuum jet; no condenser
used on jet exhaust
Vent condenser used to minimize
VOC to vacuum jet
LAB product column A
vent
Vent condenser used to minimize
VOC to vacuum jets; surface
condenser used as intercondenser
between second-stage jet and
final steam jet
Storage and handling C
emissions
Fugitive emissions
Secondary emissions K
Used by Monsanto ; see ref 1,
Refrigerated vent condenser used
to reduce emissions from fixed-
roof benzene storage tank used
to feed process
Mechanical single and double seals
used on centrifugal pumps;
quality of maintenance on valves,
etc., not known, but plant ap-
peared to be clean and neat;
special precautions used during
plant shutdowns and turnarounds
Plant wastewater streams combined,
put through an enclosed skimming
tank to remove floating organics;
then skimmed, filtered wastewater
fed to a deep-well injection
syscem for disposal; filtered
solids are then sent to land
fill; organic skimmings are re-
covered and returned to the
process or are burned as fuel
-------
B-3
Table B-2. Control Devices and Techniques Currently Used in the
LAB Chlorination Process
Source
Paraffin drying
Stream
Designation
(Fig. III-2)
\
Control Devices
Union Carbide
None
and Technologies Used
Conoco
None
By c
Whitco
Sent to heater
for
column vent
Benzene azeotrope
column vent
Hydrochloric acid
absorber vent
Atmospheric wash
decanter vents
Benzene stripping
column vent
Vacuum refining
column vents
Storage and handling
emissions
Fugitive emissions
A
A,
Secondary emissions
K
None
None
None
None
None
Insulation of
benzene stor-
age tanks
Single and double
mechanical
seals used on
pumps handling
VOC
Skimmer used to
remove floating
VOC; wastewater
sent to plant
wastewater
system
oxidation
None Sent to heater for
oxidation
None Removed by oil/
water separator
and activated
carbon adsorp-
tion
None Sent to heater for
oxidation
None Sent to heater for
oxidation
None Surface condensers
used to condense
jet exhaust; re-
sidual exhaust
sent to heater
for oxidation
Not re- Not reported
ported
Not re- Not reported
ported
Not re- Wastewater scrub-
ported bed with air to
remove benzene
and other VOC;
air from scrub-
ber sent to
heater for
oxidation
See ref 2.
3See ref 3.
See ref 4.
-------
B-4
Table B-3. Estimated Emissions from Monsanto LAB Plant'
Emission Ratio (g/Mg)
Actual
Potential
Source
Benzene drying vent
HF scrubber vent
Paraffin stripper vent
Lights stripper vent
LAB prod, column vent
Storage and handling
Fugitive emissions
Secondary emissions
Total emissions
Benzene
3.7
11.0
602
4.9
71
692.6
VOC
3.7
11.0
88.0
1
0.0014
668
20.3
142.0
933
Existing
Benzene
3.7
1.1
141.1
4.9
16
166.8
VOC
3.7
1.1
88
1
0.0014
203.1
20.3
33
350.2
See ref 5.
g of emission per Mg of LAB produced.
-------
B-5
Table B-4. Emissions from Union Carbide LAB Plant
Source
Emission Ratio (g/Mg)'
Catalyst tank vent
Water scrubber on sludge destruction
decanter vent
Wash-decantation vent
Stills
Benzene storage
72
2
0.026
3
3.3
See refs 2 and 6.
g of emissions per Mg of LAB produced.
-------
B-6
plant differs from the model plant by having a vent on the catalyst mix tank.
This vent is needed to discharge the nitrogen that is used to force the catalyst
from the storage bins into the tank containing benzene. The tank is agitated and
operates at atmospheric pressure.
3. Conoco, Baltimore, MD
The Conoco plant was at one time estimated to be emitting more than 3 tons of
benzene per day. It employs the paraffin chlorination process but differs from
the model plant in that (1) Conoco uses a molecular sieve for drying the feed
benzene, (2) the HCl absorbers are not operated to minimize absorption of ben-
zene, (3) the A1C13 catalyst sludge recovery is vented separately, and (4) the
sprung oil settler is vented separately. The estimated emissions from the Conoco
LAB plant are given in Table B-5. Conoco reported plans for retrofit emission
control to several of these sources by late 1978. In ref. 7 it is reported that
the need for a vent on the Aid catalyst sludge recovery system was eliminated
by installation of a static mixer, which cut the emission from that vent to zero.
Conoco also reported that a paraffin absorber was proposed as a control on
several of the vents shown in Table B-5 and an estimate of 5 ppm of benzene in
the exit stream from the absorber. A spray tower was reported as the proposed
control on the hydrochloric acid absorber vent and a surface aftercondenser was
reported as the proposed control on the vacuum refining column vents, with a
projection that the benzene in the vent gases from these control devices is
minimal.
4
4. Witco Chemical, Carson, CA
Witco reports that almost all of their benzene-containing vent gas streams are
burned in their heater. The HCl absorber is operated so that the organics go
with the muriatic acid; 18 Ib of benzene and 36 Ib of n-paraffin are removed from
the muriatic acid by an oil-water separator and activated carbon. Approximately
40 gpm of wastewater containing 400 to 600 mg/liter of benzene is scrubbed with
250 to 300 scfh of air, which is then sen*, to the heater for oxidation of the
benzene.
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
-------
B-7
Table B-5. Estimated Emissions from Conoco LAB Plant
Source
Benzene Emission Ratio (g/Mg)
Hydrochloric acid absorber vent
Atmospheric wash decanter vents
Benzene stripping column vent
Vacuum refining column vent
AlCl catalyst sludge recovery vent
Sprung oil settling vent
Storage and handling
Fugitive emissions
Secondary emissions
Total
62.5 (841)
175 (7428)°
526
1073
3949
131 (3898)C
1472d
No information
62 4d
8012.5
See refs 3 and 7.
g of benzene per Mg of LAB produced.
CNumbers in paretheses are emission ratios calculated for flow at upset conditions
by assuming that the concentration of benzene in the vent gases does not change.
See ref 3.
Emission ratios from ref 7, which states "the figures ... have been adjusted
to a production rate similar to that of the model plant." No explanation is
given for why ratios are expected to change with production rate. All other
ratios were calculated from data in Conoco letter (ref 3) .
-------
B-E
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. Connecting existing vents to existing flares
or fuel header systems can require a significant smount of piping. Pressure
considerations are more of a problem with existing equipment, which may not
operate properly if back-pressure is increased with the addition of emission
control equipment.
-------
B-9
APPENDIX B REFERENCES*
1. C. A. Peterson, IT Enviroscience, Trip Report for Visit to Monsanto Chemicals Co.,
Alvin, TX, Nov. 8, 1977 (on file at EPA, ESED, Research Triangle Park, NC).
2. C. A. Peterson, IT Enviroscience, Trip Report for Visit to Union Carbide Corp.,
Institute, WV, Dec. 8, 1977 (on file at EPA, ESED, Research Triangle Park, NC).
3. Letter from D. J. Lorine, Chief Engineer, Conoco Chemical Division, to D. R.
Godwin, Director, ESED Division, EPA, Feb. 17, 1978.
4. Letter from E. A. Vistica, Vice President, Witco Chemical Corp., Wilmington, CA,
to D. R. Godwin, Director, ESED Division, EPA, Feb. 6, 1978.
5. Letter from J. H. Craddock, Manager, Product Safety, Monsanto Industrial Chemicals
Co., St. Louis, MO, to D. R. Patrick, EPA, May 31, 1979, with comments on draft
LAB report.
6. Letter from R. L. Foster, Union Carbide Corp., South Charleston, WV, to D. R.
Patrick, ESED, EPA, May 16, 1979, with comments on draft LAB report.
7. Letter from R. A. Oliver, Public Health Engineer, State of Maryland Environmental
Health Administration, Baltimore, MD, to D. R. Patrick, ESED, EPA, Apr. 26, 1979,
with comments on draft LAB report.
*Usually, when a reference is located at the end of a paragraph, it refers to the
entire paragraph. If another reference rentes 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.
-------
C-l
APPENDIX C
LIST OF EPA INFORMATION SOURCES
Letter from E. A. Vistica, Witco Chemical Corp., Wilmington, CA, to D. R. Godwin,
EPA, ESED Division, Feb. 6, 1978.
Letter from D. J. Lorine, Conoco Chemicals Division, to D. R. Godwin, EPA, ESED
Division, Feb. 17, 1978.
Harry M. Walker, Texas Air Control Board 1975 Emission Inventory Questionnaire
for Monsanto Chemical Co., Chocolate Bayou, LA, Plant.
-------
TECHNICAL REPORT DATA
/Please read Immicnons on the reverse before completing)
tEPORT NO. 2
EPA-450/3-80-028b
ITL = AND SUBTITLE
Organic Chemical Manufacturing
Volume 7: Selected Processes
.UTHOR(S)
. D. Hobbs C. W. Stuewe S. W. Dylewski
i. M. Pitts C. A. Peterson
ERFORMING ORGANIZATION NAME AND ADDRESS
IT Enviroscience, Inc.
9041 Executive Park Drive
Suite 226
Knoxville, Tennessee 37923
SPONSORING AGENCY NAME AND AQDRESS
DAA for Air Quality Planning and Standards
Office of Air, Noise, and Radiation
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
3. RECIPIENT'S ACCESSION NO.
5. REPORT DATE
December 1980
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO.
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-02-2577
13. IYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
EPA/200/04
SUPPLEMENTARY NOTES
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.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
D'STRiBUTiON STATEMENT
Jnlimited Distribution
b. IDENTIFIERS/OPEN ENDED TERMS
19 SECURITY CLASS fThis Report)
Unclassified
20 SECURITY CLASS (This page I
Unclassified
c. COSATI Held/Group
13B
21. NO. OF PAGES
398
22. PRICE
: Forrr-. 2220-1 (Rev. 4-77)
PREVIOUS EDITION 'S OBSOLETE
------- |