United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EPA-450/3-80-028d
December 1 980
Air
Organic Chemical
Manufacturing
Volume 9: Selected
Processes
-------
EPA-450/3-80-028d
Organic Chemical Manufacturing
Volume 9: Selected Processes
Emission Standards and Engineering Division
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air, Noise, and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
December 1980
-------
REPORT 1
FORMALDEHYDE
R. J. Lovell
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.
D2R
-------
CONTENTS OF REPORT 1
Page
I. ABBREVIATIONS AND CONVERSION FACTORS 1-1
II. INDUSTRY DESCRIPTION II-l
A. Reason for Selection II-l
B. Formaldehyde Usage and Growth II-l
C. Domestic Producers II-3
D. References II-7
III. PROCESS DESCRIPTION III-l
A. Introduction III-l
B. Metallic-Silver-Catalyst Process III-2
C. Metal-Oxide-Catalyst Process III-5
D. Other Processes III-7
E. References III-9
IV. EMISSIONS IV-1
A. Formaldehyde from Methanol Process Using a Metallic IV-1
Silver Catalyst
B. Formaldehyde from Methanol Process Using a Metal IV-6
Oxide Catalyst
C. References IV-12
V. APPLICABLE CONTROL SYSTEMS V-l
A. Formaldehyde from Methanol Process Using a Metallic V-l
Silver Catalyst
B. Formaldehyde from Methanol Process Using a Metal V-5
Oxide Catalyst
C. References V-9
VI. IMPACT ANALYSIS VI-1
A. Environmental and Energy Impacts VI-1
B. Control Cost Impact VI-4
C. References VI-21
VII. SUMMARY VII-1
-------
l-v
APPENDICES OF REPORT 1
A. PHYSICAL PROPERTIES OF FORMALDEHYDE, METHANOL, AND PARAFORMALDEHYDE
B. AIR-DISPERSION PARAMETERS
C. FUGITIVE-EMISSION FACTORS
D. COST ESTIMATE DETAILS
E. EXISTING PLANT CONSIDERATIONS
F. LIST OF EPA INFORMATION SOURCES
-------
1-vii
TABLES OF REPORT 1
Number
II-l Formaldehyde Usage and Growth II-2
II-2 Formaldehyde Capacity II-4
IV-1 Uncontrolled Emissions from Metallic-Silver-Catalyst Process IV-3
IV-2 Model Plant Storage IV-3
IV-3 Emission Composition for Metallic-Silver-Catalyst Process IV-4
IV-4 Uncontrolled Emissions from Metal-Oxide-Catalyst Process IV-7
IV-5 Absorber Vent Gas Composition for Metal-Oxide-Catalyst Process IV-8
IV-6 Absorber Vent Emission Ratios IV-10
V-l Controlled Emissions for Metallic-Silver-Catalyst Process V-2
V-2 Controlled Emissions for Metal-Oxide-Catalyst Process V-6
VI-1 Environmental Impact of Controlled Model-Plant Formaldehyde VI-2
Production by Metallic-Silver-Catalyst Process
VI-2 Environmental Impact of Controlled Model-Plant Formaldehyde VI-3
Production by Metal-Oxide-Catalyst Process
VI-3 Cost Factors Used in Computing Annual Costs VI-5
VI-4 Control Device Cost Effectiveness for Metallic-Silver- VI-6
Catalyst Process
VI-5 Estimates of Emission Control and Reduction and Cost VI-19
Effectiveness for Formaldehyde Model Plant Using a Metal
Oxide Catalyst
VII-1 VOC Emission Summary for Model Plant VII-2
A-l Properties of Anhydrous Formaldehyde and Methanol A-l
A-2 Properties of Formaldehyde Solution (37 wt %) A-l
A-3 Properties of Paraformaldehyde A-2
B-l Air-Dispersion Parameters for Metallic-Silver-Catalyst B-l
Process Model Plant with a Capacity of 45 Gg/yr
B-2 Air-Dispersion Parameters for Metal-Oxide-Catalyst B-2
Process Model Plant with a Capacity of 45 Gg/yr
E-l Control Devices Currently Used by the Domestic Formaldehyde Industry E-4
-------
1-ix
FIGURES OF REPORT 1
Number Page
II-l Locations of Plants Manufacturing Formaldehyde II-6
III-l Flow Diagram for Metallic-Silver-Catalyst Process III-3
III-2 Flow Diagram for Metal-Oxide-Catalyst Process III-6
VI-1 Capital Cost for Thermal Oxidation VI-7
VI-2 Annual Cost for Thermal Oxidation VI-8
VI-3 Cost Effectiveness for Thermal Oxidation VI-9
VI-4 Installed Capital Cost vs Plant Capacity for Fractionator Vent VI-11
Emission Controls
VI-5 Net Annual Cost or Savings vs Plant Capacity for Fractionator Vent VI-12
Emission Controls
VI-6 Cost Effectiveness vs Plant Capacity for Fractionator Vent Emission VI-13
Controls
VI-7 Installed Capital Cost vs Plant Capacity for Formaldehyde Storage VI-15
Emission Control
VI-8 Net Annual Cost vs Plant Capacity for Formaldehyde Storage VI-16
Emission Control
VI-9 Cost Effectiveness vs Plant Capacity for Formaldehyde Storage VI-17
Emission Control
D-l Precision of Capital Cost Estimates D~2
-------
1-1
ABBREVIATIONS AND CONVERSION FACTORS
EPA policy is to express all measurements used in agency documents in metric
units. Listed below are the International System of Units (SI) abbreviations
and conversion factors for this report.
To Convert From
Pascal (Pa)
Joule (J)
Degree Celsius (°C)
Meter (m)
Cubic meter (m3)
Cubic meter (m3)
Cubic meter (m3)
Cubic meter/second
(m3/s)
Watt (W)
Meter (m)
Pascal (Pa)
Kilogram (kg)
Joule (J)
To
Atmosphere (760 mm Hg)
British thermal unit (Btu)
Degree Fahrenheit (°F)
Feet (ft)
Cubic feet (ft3)
Barrel (oil) (bbl)
Gallon (U.S. liquid) (gal)
Gallon (U.S. liquid)/min
(gpm)
Horsepower (electric) (hp)
Inch (in.)
Pound-force/inch2 (psi)
Pound-mass (Ib)
Watt-hour (Wh)
Multiply By
9.870 X 10"6
9.480 X 10"4
(°C X 9/5) + 32
3.28
3.531 X 101
6.290
2.643 X 102
1.585 X 104
1.340 X 10~3
3.937 X 101
1.450 X 10"4
2.205
2.778 X 10"4
Standard Conditions
68°F = 20°C
1 atmosphere = 101,325 Pascals
PREFIXES
Prefix
T
G
M
k
m
Symbol
tera
giga
mega
kilo
milli
micro
Multiplication
Factor
1012
109
106
103
10"3
. 10"6
Example
1 Tg = 1 X 1012 grams
1 Gg = 1 X 109 grams
1 Mg = 1 X 106 grams
1 km = 1 X 103 meters
1 raV = 1 X 10"3 volt
1 pg = 1 X 10"6 gram
-------
II-l
II. INDUSTRY DESCRIPTION
A. REASON FOR SELECTION
Formaldehyde production was selected for study because preliminary estimates
indicated that emissions of volatile organic compounds (VOC) from the industry
were high and because an increase in formaldehyde consumption was expected to
continue.
Pure, dry formaldehyde is a colorless gas characterized by a pungent odor. Its
stability in the gaseous state depends on its purity; however, even traces of
water will cause rapid polymerization. Therefore formaldehyde is manufactured
and transported only in solution or in the polymerized state. The domestic
production capacity of formaldehyde is reported traditionally on the basis of a
37 wt % solution, although it is manufactured and sold in different forms, e.g.,
37, 44, 50, 52, and 56 wt % solutions and as paraformaldehyde, a solid. When-
ever possible customers buy the high-concentration product in order to reduce
freight
dehyde.
freight charges. Appendix A gives the pertinent physical properties of formal-
B. FORMALDEHYDE USAGE AND GROWTH
The current production capacity of formaldehyde in the United States (based on
37 wt % solution) is 4066 Gg/yr, with the 1977 production being 2750 Gg/yr, or
68% of this capacity. ' Formaldehyde consumption is expected to increase at an
average annual rate of 4 to 5% during 1977—1982. ' At these rates production
will be 85% of current capacity by 1982.
The uses of formaldehyde and their expected growth rates are given in Table II-l.
The manufacture of adhesives constitutes 60% of the end use for the formaldehyde
produced. The major derivatives — urea-formaldehyde and phenol-formaldehyde
resins -- are used principally in the manufacture of particle board and plywood.
Thus the consumption pattern of formaldehyde depends largely on the construction
industry.
The manufacture of plastics accounts for approximately 10% of the formaldehyde
produced. Butanediol, a derivative of formaldehyde, is used in making polybutyl-
ene terephthalate (PBT). If the use of plastic in automobile production increases
-------
II-2
Table II-l. Formaldehyde Usage and Growth
End Use
Urea resins
Phenolic resins
Butanediol
Acetol resins
Pentaerythritol
Hexamethylenetetramine
Melamine resins
Urea formaldehyde concentrates
Chelating agents
4,4'-Methylenedianiline and 4,4'-
methylenediphenyl isocyanate
Textile treating applications
Pyridine chemicals
Trimethylolpropane
Nitroparaffin derivatives
Other
Production for
1977 (%)
25.4
24.3
7.7
7.0
6.0
4.5
4.2
3.6
3.6
2.6
1.8
1.3
1.3
0.4
6.3
Average Growth for
1977 — 1982 (%/yr)
0 to 3
4 to 5
12
9 to 10
1 to 3
2 to 3
7
3
5
12 to 15
-1 to +1
8
7
7
7
aSee ref. 1.
Growth rates are rounded to the nearest 1%.
-------
II-3
and PBT plastic is chosen as the principal material used, formaldehyde could have
a growth rate of more than 10% annually for a few years.
C. DOMESTIC PRODUCERS
Because most of the formaldehyde is manufactured and shipped as a solution con-
taining 50% or more water, the distance from the producing point to the consuming
point is minimized to reduce shipping costs. Therefore the industry is charac-
terized by a large number of relatively small plants. Since more than half the
formaldehyde used is in the manufacture of adhesives for wood products, the
producing plants are located predominantly in the south and northwest.
Sixteen producers were operating 55 formaldehyde plants at the end of 1977.
Table II-2 lists the producers, locations, capacities, and processes; Fig. II-l
shows the plant locations.
Formaldehyde has a tendency to polymerize on storage. When an inhibitor is added
to prevent excessive polymerization at lower storage temperatures, it is usually
methanol at 7 to 11% concentration. Most of the formaldehyde is sold uninhibited
and must be kept warm (above 54°C) to prevent polymerization.
All the formaldehyde produced in the United States is made from methanol either
by a combination oxidation-dehydrogenation process using a silver catalyst or by
catalytic oxidation in the vapor phase using a metal oxide catalyst. About half
the formaldehyde producers also produce methanol feedstock (Borden, Celanese, Du
Pont, Georgia-Pacific, Hercules, IMC, Monsanto, and Tenneco). Reichhold is the
only large producer that does not make its own methanol feedstock.
-------
II-4
Table I1-2. Formaldehyde Capacity3
Capacity (Gg/yr)
Location Silver, Metal Oxide
Key Producer Process Process6
Allied Chemical Corporation
1A South Point, OH 141
Borden, Inc.
2A Demopolis, AL 45
2B Diboll, TX 36
2C Fayetteville, NC 106
2D Geismar, LA 113
2E Louisville, KY 36
2F Sheboygan, WI 59
2G Fremont, CA 102
2H Kent, WA 36
21 LaGrande, OR 29
2J Missoula, MT 41
2K Springfield, OR 109
Celanese Chemical Company
3A Bishop, TX 680
3B Newark, NJ 53
3C Rock Hill, SC 53
Chembond Corporation
4A Springfield, OR 68
4B Winnfield, LA 32
Du Pont Company
5A Belle, WV 227
5B LaPorte, TX 145
5C Healing Springs, NC 100
5D Linden, NJ 73
5E Toledo, OH 122
GAF Corporation
6A Calvert City, KY 45
Georgia-Pacific Corporation
7A Albany, OR 54
7B Columbus, OH 45g
7C Coos Bay, OR 419
7D Crossett, AR 45 27
7E Russellville, SC 113g
7F Taylorsville, MS 54
7G Vienna, GA 45
7H Lufkin, TX 45
Gulf Oil Corporation
8A Vicksburg, MS 27
Hercules Inc.
9A Louisiana, MO 77
9B Wilmington, NC 45
Hooker Chemicals and Plastics Inc.
10A North Tonawanda, NY 61
-------
II-5
Table II-2. (Continued)
Location
Key
11A
11B
Producer
IMC Chemical Group, Inc.
Seiple, PA
Sterlington, LA
Capacity
Silver ,
Process
29
14
(Gg/yr)
Metal Oxide
Process
12A
12B
12C
12D
13A
14A
14B
14C
14D
14E
14F
14G
14H
ISA
15B
16A
Monsanto Corporation
Addyston, OH
Chocolate Bayou, TX
Eugene, OR
Springfield, MA
Pacific Resins & Chemicals, Inc.
Eugene, OR
Reichhold Chemicals, Inc.
Hampton, SC
Houston, TX
Kansas City, KA
Malvern, AR
Moncure, NC
Tacoma, WA
Tuscaloosa, AL
White City, OR
Tenneco Inc.
Fords, NJ
Garfield, NJ
Wright Chemical Corporation
Reigelwood, NC
Total process capacity
Number of plants
Percent of total industry capacity
Capacity of total industry
45
88
45
134
43
23
23
33
45
45
3040
35
74.8
54
50
55
22
113
39
1026
20
25.2
4066
See refs. 1 and 2. See Fig. II-l for plant locations. Because of space
limitations, symbols on map do not reflect precise locations. Based on 37 wt %
solution. Silver catalyst process. eMixed metal-oxide-catalyst process.
Capacity of Russellville, SC, plant, which came on-stream in 1975, is 113 Gg/yr
[see 1976 Directory of Chemical Producers. United States of America, p. 62 in
January to July Supplement, Chemical Information Services, Stanford Research
Institute, Menlo Park, CA (July 1977)]. gGeorgia-Pacific Corporation in 1979
stated their Columbus, Coos Bay, and Russellville capacities to be 32, 34, and
90 Gg/yr (per letter dated May 30, 1979, from V. J. Tretter, Georgia-Pacific, to
R. T. Walsh, EPA). In 1976 Rohm and Haas closed an 11-Gg/yr plant at Phila-
delphia, PA, and Union Carbide closed a 54-Gg/yr plant at Bound Brook, NJ.
-------
Fig. II-l. Locations of Plants Manufacturing Formaldehyde
(See Table II-2 for identification of plants)
-------
II-7
D- REFERENCES*
1. J. L. Blackford, "CEH Marketing Research Report on Formaldehyde," pp. 658.5031C--
658.5033E in Chemical Economics Handbook, Stanford Research Institute, Menlo Park,
CA (April 1977).
2. R. B. Morris et. al. , Engineering and Cost Study of Air Pollution Control for the
Petrochemical Industry, Volume 4; Formaldehyde Manufacture with the Silver
Catalyst Process, EPA-450/3-73-006-d, pp. FS-9 and 10, EPA, Research Triangle
Park, NC (March 1975).
3. B. F. Greek and W. F. Fallwell, "Gas-based Chemicals: Slow Growth Continues,"
Chemical and Engineering News 56, 10--14 (January 1978).
*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
III. PROCESS DESCRIPTION
A. INTRODUCTION
In the United States two major processes are used in the manufacture of formalde-
hyde from methanol: the metallic-silver-catalyst combination dehydrogenation-
oxidation process (silver process), which is used at 35 locations to produce 75%
of the formaldehyde manufactured, and the metal-oxide-catalyst oxidation process
(metal oxide process), which is used at 20 locations to produce 25% of the formal-
dehyde manufactured. The projected annual production growth for each process is
4 to 5%.1'2
Two gas-phase reactions are employed to form formaldehyde from methanol:
Reaction 1 — Dehydrogenation
CH OH * HCHO + H
w £,
(methanol) (formaldehyde) (hydrogen)
Reaction 2 — Oxidation
CH OH + ^O_ > HCHO + H-0
<3 £* £» £
(methanol) (oxygen) (formaldehyde) (water)
The silver process involves the dehydrogenation of methanol (reaction 1) followed
by oxidation of a portion of the hydrogen evolved to form water, or a combination
of dehydrogenation and oxidation of methanol (reactions 1 and 2). The metal oxide
process involves oxidation of methanol by reaction 2.
The major difference between the two processes is the amount of air mixed with
the methanol before conversion. Since air and methanol form explosive mixtures
at concentrations of approximately 6 to 37 vol % of methanol in air, the air/
methanol input must be controlled. The metallic-silver-catalyst dehydrogenation-
oxidation processes maintain an excess of methanol, and methanol concentrations
are above the explosive range, whereas the metal-oxide-catalyst oxidation proc-
esses use an excess of air so that the methanol concentrations are below the lower
limits of the explosive range. Off-gas from the metallic-silver-catalyst process
A
contains 18 to 20% hydrogen and less than 1% oxygen. Off-gas from the metal-oxide-
-------
III-2
catalyst process contains unreacted oxygen and no appreciable concentration of
hydrogen.
With the metallic-silver-catalyst process the excess methanol must be separated
to meet customer specifications. With the metal-oxide-catalyst process the large
excess of air requires a larger compressor and the process equipment must be larger
to handle the added air volume.
B. METALLIC-SILVER-CATALYST PROCESS
1. Process Description
The process flow diagram shown in Fig. III-l represents a typical continuous metallic'
silver-catalyst process.
The incoming air (Stream 1) is washed with caustic to remove traces of sulfur
dioxide, hydrogen sulfide, and other impurities that act as catalyst poisons.
The air is then compressed to 143 kPa and passed through a vaporizer column, where
it is heated and saturated with methanol vapor (Stream 2). The heated air and
methanol vapor must comprise more than 37 vol % methanol in order to be above the
high explosive limit of the methanol.
The mixture (Stream 3) then enters a battery of converters. Approximately 80% of
the methanol feed is reacted for a per-pass conversion ratio of 0.80. The converter
temperature is maintained at approximately 635°C (ref. 5) by heat generated by
the oxidation of a portion of the hydrogen evolved and/or by methanol oxidation.
The hot effluent gases (Stream 4) are quickly cooled to prevent decomposition of the
formaldehyde formed. Cooling is accomplished by heat interchange with the feed
mixture in the vaporizer and by then introducing the gas into the primary absorber.
The primary-absorber liquid is an aqueous solution of formaldehyde and methanol.
A portion of this liquid is withdrawn from the bottom of the absorber column and
recirculated to the top. The remainder (Stream 5) is pumped to the product frac-
tionation column. The uncondensed vapors and noncondensable gases (Stream 6) are
withdrawn from the top of the primary-absorber column and fed to a secondary absorber;
-------
METHA.UOI-
CAUSTIC WATER
AIR
^
—£00)
AIR
TOWER
(K)
SPEUT
CAUSTIC
STARTUP
PRIMARY
blOM POTEWT
DEMlUERALIZED
BOTTOMS
COOLER
WATER
REGEKJERAWT
RECYCLE
METHAUOL
SURGE
PRODUCT
FRACT lOUATlOW
\OU
i.(D)
PRODUCT
STQRAC.E
SALTS
Fig. III-l. Flow Diagram for Uncontrolled Model Plant Producing Formaldehyde from
Methanol with the Metallic-Silver-Catalyst Process
-------
III-4
The major portion of the uncondensed vapors is recovered in the secondary-absorber
column through contact with demineralized water, and the off-gas is vented (Vent A).
The weak formaldehyde/methanol solution (Stream 7) withdrawn from the bottom of
the secondary-absorber column is pumped to the primary-absorber column and used
as makeup solution.
The methanol-containing formaldehyde solution (Stream 5) is pumped to a fractiona-
tion column, where methanol is recovered. This vacuum distillation step yields an
overhead product of approximately 99% methanol for recycle to the reactor and a
bottom product of formaldehyde solution containing less than 1% methanol. The
methanol vapor from the top of the column is condensed and recycled to the vaporizer
(Stream 9). Uncondensed vapors are vented (Vent B). The formaldehyde solution
from the bottom of the fractionation column is pumped to product storage tanks.
When required by customer specifications the solution is treated for removal of
trace amounts of formic acid by being passed through an ion exchange system before
being stored. As a final step, water is added to provide a suitable concentration
for storage and shipping. Reported yields for the metallic-silver-catalyst process
range from 83 to 92%.1/3"5
All product storage tanks are heated to prevent polymer formation and precipita-
tion in storage. A series of tanks are used to blend and adjust the solution to
the desired formaldehyde and methanol concentrations before it is shipped to the
customer.
2. Process Variations
Vacuum distillation is described for the model-plant process step used for recovery
of the excess methanol in the product stream. However, the uncondensed gases and
vapors discharged from the vacuum producer must be vented or otherwise dealt with.
Many plants use pressure distillation equipment operated at increased temperature,
instead of vacuum fractionation, which dispenses with the need for the vent associ-
ated with the fractionator vacuum producer (Vent B, Fig. III-1) and therefore
eliminates the emission source. The quantity of dissolved gases in the feed to
the distillation column is very small.
Process development efforts have been directed toward reducing the excess methanol
in order to eliminate the fractionation step. The addition of water to the methanol
-------
III-5
to form a feed mixture containing 30 to 50 vol % water can produce a product con-
8
5
taining 7 to 8 wt % methanol. This concentration of methanol is suitable for
some markets.'
Many approaches have been taken by the various plants to improve efficiency for
heat utilization. In some plants heat from the reactor is used to heat the
methanol distillation reboiler or to generate steam to drive the fractionation-
column vacuum pump.
C. METAL-OXIDE-CATALYST PROCESS
1. Process Description
The process flow diagram shown in Fig. III-2 represents a typical continuous metal-
oxide-catalyst process. The catalyst system most often used is ferric molybdate.
Incoming air (Stream 1) may be washed with caustic solution in a packed tower to
remove dust and trace impurities. The air is then compressed, along with the
recycle gas (Stream 5) to 143 kPa. Recycle of a portion of the oxygen-lean vent
gases lowers the oxygen content of the air feed stream to below 10.9%. This re-
duces the explosion hazard of the feed mixture and increases the equipment output
by reducing the amount of excess air required to keep the methanol concentration
below the low explosive limit.
A portion of the air is passed through the vaporizer column, where it is saturated
with methanol (Stream 2). The methanol-saturated air is then mixed with the remain-
ing air and preheated by heat exchange with the product gas leaving the converter.
The feed gas mixture (Stream 3) then enters the converter. Conversion ratios of
97% are obtained. The converter, heated by the exothermic oxidation reaction, is
maintained at 345°C (ref. 5) by boiling heat transfer fluid in the reactor shell.
Steam is generated by condensation of the heat-transfer fluid vapors.
The product gas (Stream 4) is cooled by heat exchange with the feed gas mixture
and then quenched in the absorber column. The formaldehyde and methanol are re-
moved from the gas stream by absorption in the aqueous solution. The unabsorbed
gases and vapors exit at the top of the absorber column. A portion (60 to 80%)
of this gas is recycled (Stream 5) and the remaining gas is vented. The product
-------
H
CTl
FORMALDEHYDE
CAO6TIC
V/ASH
PRODUCT IOU PRODUCT
) - FURTIVE EM\-at>lOVJ*> - OVERAU-
Fig. III-2. Flow Diagram for Uncontrolled Model Plant Producing Formaldehyde from
Methanol with the Metal-Oxide-Catalyst Process
-------
III-7
solution drawn from the bottom of the absorber column contains approximately 0.8%
methanol and 0.005% formic acid. The solution is usually treated in an ion exchange
system to reduce the acidity and is then stored. As a final step water is added
to provide a suitable concentration for storage and shipping. Process yields of
91 to 93% are reported for the metal-oxide-catalyst process. '
The formaldehyde yield from the metal-oxide-catalyst process is higher than that
from the metallic-silver-catalyst process and the metal-oxide-catalyst process is
simpler, because methanol distillation is not required. The equipment costs for
the metal-oxide-catalyst process are greater because of the large volume gas
streams; also, because of a lower concentration of formaldehyde in the product
gas stream, the absorber column diameter is larger and the operating temperature
is lower than those used with the metallic-silver-catalyst process.
2. Process Variations
The industry makes use of various catalyst compositions and methods and extent of
heat recovery. Otherwise, the processes used are basically similar. Many older
plants do not recycle a portion of the absorber-column vent gas. For these plants
the vent gas volume and the ratio of volatile organic compounds emitted per unit
of product produced are increased.
The oxide catalyst is not susceptible to poisoning by traces of sulfides in the
air feed; thus many plants filter the incoming air rather than utilizing caustic
scrubbers.
D. OTHER PROCESSES
1. Formaldehyde by Partial Oxidation of Methane
Considerable research has been devoted to production of formaldehyde directly
from methane. The process is more complex and requires a higher capital invest-
ment than do processes utilizing methanol. Commercial attempts to produce formal-
dehyde from natural gas or methane has had limited success and the process cur-
rently is not used in the United States.
-------
III-8
2. Formaldehyde by Partial Oxidation of Light Hydrocarbons
Higher hydrocarbons, such as ethane, propane, or butane, may also be oxidized to
formaldehyde. This process was used originally by the Celanese Corporation at
their large Bishop, Texas, plant but has recently been replaced by the metallic-
silver-catalyst process using methanol as the feedstock. Because of the cost of
light hydrocarbons it is doubtful that any new facility in the United States will
again utilize this process.
-------
III-9
E. REFERENCES*
1. J. L. Blackford, "CEH Marketing Research Report on Formaldehyde," p. 658.5031D in
Chemical Economics Handbook. Stanford Research Institute, Henlo Park, CA
(April 1977).
2. "Chemical Profile on Formaldehyde," p. 9 in Chemical Marketing Reporter (Jan. 22,
1978).
3. J. F. Walker, Formaldehyde, 3d ed., p. 9, American Chemical Society Monograph
Series, Reinhold, New York (1974).
4. J. F. Walker, "Formaldehyde," p. 7799 in Kirk-Othmer Encyclopedia of Chemical
Technology, 2d ed., vol. 10, edited by A. Stenden et al., Interscience, New York,
1969.
5. G. E. Haddeland and G. K. Chang, Report No. 23. Formaldehyde, pp. 63—95,
A private report by the Process Economics Program, Stanford Research Institute,
Menlo Park, CA (February 1967).
6. R. B. Morris et al., Engineering and Cost Study of Air Pollution Control for the
Petrochemical Industry, Volume 5.- Formaldehyde Manufacture with the Mixed Oxide
Catalyst Process, EPA-450/3-73-006-e, pp. FS1--8, EPA, Research Triangle Park,
NC (March 1975).
*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 com-
pounds (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 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 estimated for the formaldehyde model plants are based on
the emissions reported in responses to EPA's requests for information from selected
companies, on EPA emission testing data, and on data obtained during visits to
Celanese and Borden formaldehyde production plants (see Appendix E). Also used
in sizing and design of the model plants were data from the EPA studies, SRI
reports, formaldehyde data compiled by J. F. Walker, and an understanding of
the process chemistry and yields.
A. FORMALDEHYDE FROM METHANOL PROCESS USING A METALLIC SILVER CATALYST
1. Model Plant
The model plant* for this study has a capacity of 45 Gg/yr, based on 8760 hr of
operation per year.** Although not an actual operating plant, it is typical of
many plants built recently. The plant utilizes the model metallic-silver-catalyst
process (Fig. III-l) and best fits today's formaldehyde manufacturing and engi-
neering technology for that process.
Typical raw-material, intermediate, and product-storage tank capacities were esti-
mated for the 45-Gg/yr model plant. Storage tank requirements are given in
Sect. IV.A.2.e. Estimates of potential fugitive sources, based on data from existing
*See p. 1-2 for a discussion of model plants.
**Process downtime is normally expected to range from 5 to 15%. If the hourly
rate remains constant, the annual production and annual VOC emissions will be
correspondingly reduced. Control devices will usually operate on the same cycle
as the process. From the standpoint of cost-effectiveness calculations, the
error introduced by assuming continuous operation is negligible.
-------
IV-2
plants, are given in Sect. IV.A.2.d. " Characteristics of the model plant that
are important in air-dispersion modeling are shown in Appendix B.
2. Sources and Emissions
Uncontrolled emission rates and sources for the metallic-silver-catalyst process
are summarized in Table IV-1 and are described below. The process emission rates
are based on emission data from existing plants. Potential storage, handling,
fugitive, and secondary emissions were calculated from characteristics of the
plant and from data on existing plants (see Table IV-2).
a. Absorber Vent — The absorber vent (Vent A, Fig. III-l) is the principal source
of emissions from the formaldehyde production plant. The volatile organic com-
pounds (VOC) in the vent gas include unreacted methanol, formaldehyde, methyl
formate, and methylal (see Table IV-3). Also included in the vent gas are hydrogen,
methane, carbon monoxide, carbon dioxide, nitrogen, water vapor, and a small amount
of unreacted oxygen. The composition of the absorber vent varies somewhat with
4
the catalyst's age and activity. The average emission during normal operation
is given in Table IV-1.
b. Product Fractionator Vent -- A product fractionator operated under vacuum is used
to separate and recover unreacted methanol from the product stream. A steam ejector
or vacuum pump is employed to produce the vacuum required. Emissions from this
source (Vent B) include methanol vapor, formaldehyde, water vapor, and a small
amount of inert gases (see Table IV-3).
c. Intermittent Air Emissions -- The formaldehyde plant is normally operated at design
conditions to achieve highest yields. It is shut down when product inventories
are filled. Since the metallic-silver-catalyst process operates above the upper
explosive limit of methanol, the plant startup procedure must be handled carefully.
Unstable conditions are often encountered and explosions can occur in the methanol
vaporizer and the reactor. Various startup procedures are used in the industry.
Usually during startup the output from the reactor is vented until stable operation
is achieved and an acceptable yield ratio is obtained. The flow is then switched
2
into the absorber, the total startup time is 1 to 2 hr. The reactor feed rate
varies as the startup proceeds. Initially the reactor produces mainly carbon
-------
IV-3
Table IV-1. Total VOC from Uncontrolled Emissions Produced by
Formaldehyde Metallic-Silver-Catalyst Process in a Model Plant
Emission Source
Absorber
Product fractionator
b
Intermittent
Methanol storage
Formaldehyde storage
Handling
Fugitive
Secondary
Total
Stream
Designation
(Fig.III-1)
A
B
C
D
D
F
H
K
Ratio
(g/kg)
4.73
1.58
0.11
0.08
0.03
0.01
1.08
0.01
7.63
Emissions
Rate
(kcr/hr)
24.5
8.2
0.28
0.43
0.2
0.1
5.7
0.03
39.44
g of total VOC per kg of 37% formaldehyde solution produced.
Average rate for entire year, based on 8 startups per year and flow of one-half the
normal rate during startup.
Table IV-2. Model Plant Storage
Content
Tank Size
(m3)
No. of Tanks
Required
Turnovers
Per Year
Bulk Liquid
Temperature
<°c>
Metallic-Silver-Catalyst Process
Methanol (feed) 190
Formaldehyde 190
Methanol (recycle) 4
2
4
1
68
45
25
54
38
Metal-Oxide-Catalyst Process
Methanol (feed)
Formaldehyde
190
190
2
4
65
45
25
54
Surge tank.
-------
IV-4
Table IV-3. Emission Composition for.
Model Metallic-Silver-Catalyst Process'
Emissions
Component
VOC
Formaldehyde
Methanol
Methyl formate
Methylal
Combustible gases
Hydrogen
Methane
Carbon monoxide
Other gases
Oxygen
Nitrogen
Carbon dioxide
Water vapor
Total
Absorber
Composition
(wt %)
0.75
(0.06)
(0.14)
(0.36)
(0.19)
2.38
(1.69)
(0.28)
(0.41)
93.96
(0.42)
(86.64)
(6.90)
2.91
100.00
Vent
b
Ratio
(g/kg)
4.73
(0.36)
(0.89)
(2.27)
(1.20)
14.97
(10.59)
(1.77)
(2.61)
590.23
(2.63)
(544.23)
(43.37)
18.25
628.17
Fractionator
Composition
(wt %)
83.16
'(17.49)
(65.67)
0.94
15.90
100.00
Vent
Ratiob
(g/kg)
1.58
(0.33)
(1.24)
0.02
0.3013
1.90
Emission rates are based on emission data from existing plants; see Appendix E.
g of emission per kg of 37 wt % formaldehyde solution produced.
-------
IV-5
dioxide and water vapor. As the temperature rises, the formaldehyde yield in-
creases, thereby enlarging the amount of VOC in the gas vented. The average annual
emission calculated from this source (see Table IV-1) is based on an average of
eight startups per year and a flow of one-half the normal rate during startup.
d. Fugitive Emissions — Process pumps, valves, and circulating process cooling water
are potential sources of fugitive emissions. The model plant is estimated to
have 13 pumps, 214 process valves, and 6 relief valves handling VOC. An esti-
mated 6.5 liters of cooling water per kg of product produced is circulated through
the cooling tower. Fugitive emission factors from Appendix C were applied to
determine the fugitive emissions shown in Table IV-1.
6. Storage and Handling Emissions -- Emissions result from the storage and handling
of methanol and formaldehyde. Sources for the model plant are shown in Fig. III-l
(Source D). Storage tank conditions for the model plant are given in Table IV-2.
The emissions in Table IV-1 are based on fixed-roof tanks, half full, and a diurnal
temperature variation of 11.1°C and were calculated based on the emission equations
from AP-42. However, calculated breathing losses were divided by 4 to account
for recent evidence indicating that the AP-42 breathing loss equation overpredicts
6
emissions.
Since uninhibited formaldehyde polymerizes at a low temperature, concentrated
solutions (over 30% HCHO) must be kept warm.7 Therefore the model-plant formal-
dehyde storage tanks are maintained at above 54°C. Since these tanks are tempera-
ture controlled, breathing losses are negligible and emissions given are based on
working losses only.
Emissions from the loading of formaldehyde solution into trucks and tank cars
were calculated with the equations from AP-42, with submerged-fill-pipe loading
o
assumed to be used. These emissions are also included in Table IV-1.
f - Secondary Emissions -- Secondary VOC emissions can result from the handling and
disposal of process-waste liquid streams. The potential sources (Source K) that
exist for the model plant are ion exchange system regeneration and blowdown water
from the cooling and air-wash towers. The calculated total secondary VOC emis-
sion from these sources is given in Table IV-1. Calculations are based on waste-
-------
IV-6
water data reported by industry and the assumptions that the ion exchange system
is operated for 10% of the production and that 5% of the VOC contained in the
wastewater evaporates before treatment.
B. FORMALDEHYDE FROM METHANOL PROCESS USING A METAL OXIDE CATALYST
1. Model Plant
The model plant for this study is based on the metal-oxide-catalyst process
utilizing vent gas recycling (Fig. III-2). This model process best fits today's
formaldehyde manufacturing and engineering technology for utilizing a metal oxide
catalyst. The model plant has a capacity of 45 Gg/yr based on 8760 hr of operation
per year.
Typical raw-material and product-storage tank capacities were estimated for the
45-Gg/yr model plant. Storage tank requirements are given in Sect. IV.B.2.e.
Estimates of potential fugitive sources, based on data from existing plants, are
given in Sect. IV.B.2.d. ~~ Characteristics of th<
in air-dispersion modeling are given in Appendix B.
given in Sect. IV.B.2.d. Characteristics of the model plant that are important
2. Sources and Emissions
Uncontrolled emission rates and sources for the metal-oxide-catalyst process are
summarized in Table IV-4 and are discussed below. Process emission rates are
based on emission data from existing plants. Potential storage, handling, fugitive,
and secondary emissions were calculated from model-plant characteristics and data
from existing plants.
Since the process operates below the explosive limit with an excess of air,
unstable conditions during startup are easily prevented. Venting of the reactor
during startup is not required.
a. Absorber Vent -- The product absorber vent (Vent A, Fig. III-2) is the main source
of emissions from the formaldehyde production plant. The VOC components in the
vent gas include methanol, formaldehyde, and dimethyl ether (Table IV-5). Also
included in the vent gas are carbon monoxide, carbon dioxide, nitrogen, oxygen,
and water vapor.
-------
IV-7
Table IV-4. Total VOC Uncontrolled Emissions for Formaldehyde
Produced by Metal-Oxide-Catalyst Process in a Model Plant
Source
Absorber
Methanol storage
Formaldehyde storage
Handling
Fugitive
Secondary
Total
Stream
Designation
(Fig. Ill- 1)
A
D
D
F
H
K
Ratio
(g/kg) a
3.15
0.08
0.03
0.01
0.74
0.05
4.06
Emissions
Rate
(kg/hr)
16.3
0.44
0.2
0.1
3.7
0.2
20.7
a
g of total VOC per kg of 37% formaldehyde solution produced
-------
IV-8
Table IV-5. Absorber Vent Gas Composition for
Model Metal-Oxide-Catalyst Processa
Component
VOC
Formaldehyde
Methanol
Dimethyl ether
Other gases
Oxygen
Nitrogen
Carbon dioxide
Carbon monoxide
Water vapor
Composition
(wt %)
0.27
(0.04)
(0.16)
(0.07)
95.45
(7.91)
(86.09)
(0.19)
(1.26)
4.28
100.00
Emissions
Ratio
(q/kcr)
3.15
(0.44)
(1.93)
(0.79)
1124.39
(93.20)
(1014.02)
(2.28)
(14.89)
50.37)
1177.91
Emission rates are based on emission data from existing plants;
see Appendix E.
g of emission per kg of 37 wt % formaldehyde solution produced.
-------
IV-9
The emission composition and flow rates are especially affected by the percent of
absorber gas recycled. By recycling a portion of the oxygen-lean vent gas, the
oxygen concentration in the reactor feed mixture can be reduced, making it possible
for the concentration of methanol to be increased without an explosive mixture
being formed. This reduces the volume of reaction gases and thus reduces equip-
ment size and horsepower required to drive the compressor. Recycling reduces the
9
emission rate and enhances the reactor equilibrium to produce a higher yield.
The effect of recycling on absorber vent emissions can be seen by the comparison
of absorber vent gas composition for recycling and nonrecycling operations given
in Table IV-6.
Other variables that affect the absorber vent emissions are catalyst formulation,
catalyst age, absorber temperature, and strength of formaldehyde produced. The
catalyst formulation can affect the overall process yield and thus the amount and
type of by-products or emissions produced. Also, product yields tend to decrease
as the catalyst ages. Lowering the absorber temperature increases its efficiency
and thus lowers the VOC emissions from the absorber. As the strength of the for-
maldehyde produced increases, the partial pressure due to the formaldehyde increases,
thus increasing the relative amount of formaldehyde in the vent emission.
The model-plant average absorber vent emission during normal operation is given
in Table IV-4. A recycle rate of 63 vol % for the product absorber vent gas was
calculated based on emissions data.
Fugitive Emissions -- Process pumps and valves are potential sources of fugitive
emissions. The model plant is estimated to have 8 pumps, 176 process valves, and
4 relief valves handling VOC. An estimated 6.5 liters of cooling water per kg of
product produced is circulated through the cooling tower. Fugitive emission factors
from Appendix C were applied to determine the fugitive emissions shown in Table IV-4.
Storage and Handling Emissions — Emissions result from the storage and handling
of methanol and formaldehyde. Sources for the model plant are shown in Fig. III-2
(Source D). Storage tank conditions for the model plant are given in Table IV-2.
The emissions in Table IV-4 are based on fixed-roof tanks, half full, and a diurnal
temperature variation of 11.1°C and were calculated based on the emission equations
-------
IV-10
Table IV-6. Absorber Vent Emission Ratios for Recycled Vent Gas vs
Nonrecycled Vent Gas for Model Metal-Oxide-Catalyst Process
Emission Ratio (g/kg)
Recycled Nonrecycled
Component Vent Gas Vent Gas
Total VOCb 3.2 12.7
Other
C 1174.8 5440.0
Total 1178.0 5452.7
ag of total VOC per kg of 37% formaldehyde solution produced.
blncludes formaldehyde, methanol, and dimethyl ether.
Includes nitrogen, oxygen, carbon dioxide, carbon monoxide, and water.
c
-------
IV-11
from AP-42. However, the breathing losses were divided by 4 to account for recent
evidence indicating that the AP-42 breathing-loss equation overpredicts emissions.
Since uninhibited formaldehyde polymerizes at low temperature, concentrated solu-
tions (over 30% HCHO) must be kept warm. Consequently the model-plant formalde-
hyde storage tanks are maintained above 54°C. Breathing losses are negligible
because the tank temperature is controlled. Emissions therefore are based on
only working losses.
Emissions from the loading of formaldehyde solution into trucks and tank cars
were calculi
Table IV-4.
Q
were calculated with the equations from AP-42. These emissions are included in
d. Secondary Emissions — Secondary VOC emissions can result from the handling and
disposal of process-waste liquid streams. The potential sources (Source K) that
exist for the model-plant are ion exchange system regeneration, blowdown water
from the cooling and air-wash towers, and emissions from the heat-transfer fluid
system vent. The total secondary VOC emission from these sources is given in
Table IV-4. Emissions were calculated based on wastewater data reported by industry
and the assumptions that the ion exchange system is operated for 100% of the produc-
tion and that 5% of the VOC contained in the wastewater evaporates before treatment.
-------
IV-12
C. REFERENCES*
1. J. F. Lawson, IT Enviroscience, Inc., Trip Report to Celanese Plant, Celanese
Chemical Company, Bishop, TX, July 26, 1977 (on file at EPA, ESED, Research
Triangle Park, NC).
2. J. F. Lawson, IT Enviroscience, Inc., Trip Report to Borden Plant, Borden, Inc.,
Fayetteville, NC, August 24, 1977 (on file at EPA, ESED, Research Triangle Park,
NC).
3. D. F. Dryden, Data Package for Formaldehyde Plant Fugitive Emissions Study, p. 2,
Walk, Haydel & Associates, Inc., New Orleans, LA (June 27, 1978).
4. R. B. Morris et. al., Engineering and Cost Study of Air Pollution Control for the
Petrochemical Industry, Volume 4; Formaldehyde Manufacture with the Silver
Catalyst Process, EPA-450/3-73-006-d, pp. FS-8—18, EPA, Research Triangle Park,
NC (March 1975).
5. C. C. Masser, "Storage of Petroleum Liquids," pp. 4.3-1--4.3-16 in Compilation
of Air Pollutant Emission Factors, 3d ed., Part A, AP-42, EPA, Research Triangle
Park, NC (April 1977).
6. E. C. Pulaski, TRW, letter to Richard Burr (EPA), May 30, 1979.
7. J. F. Walker, "Formaldehyde," p. 79 in Kirk-Othmer Encyclopedia of Chemical
Technology, 2d ed., vol. 10, edited by A. Standen e_t al., Interscience, New York,
1969.
8. C. C. Masser, "Transportation and Marketing of Petroleum Liquids," pp. 4.4-1—
4.4-6 in Compilation of Air Pollutant Emission Factors, 3d ed., Part A, AP-42,
EPA, Research Triangle Park, NC (April 1977).
9. C. W. Horner, "A Formaldehyde Process to Accommodate Rising Energy Costs,"
Chemical Engineering 84, 108—110 (July 4, 1977).
10. R. B. Morris e_t al., Engineering and Cost Study of Air Pollution Control for
the Petrochemical Industry. Volume 5: Formaldehyde Manufacture with the Mixed
Oxide Catalyst Process, EPA-450/3-73-006-e, p. FM-8, EPA, Research Triangle Park,
NC (March 1975).
*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. FORMALDEHYDE FROM METHANOL PROCESS USING A METALLIC SILVER CATALYST
1. Absorber Vent
Thermal oxidation can be used effectively to control the emissions from the absor-
ber vent. The vent stream is rich in hydrogen gas and contains other clean-burning
hydrocarbons. The heating value of the gas is high enough to self-sustain com-
bustion. If the gas is fired with a nominal amount of supplemental fuel to main-
tain stable combustion conditions, heat can be recovered from the flue gas. An
emission reduction efficiency of greater than 99% can be obtained with this system.
Thermal oxidizer systems and efficiencies are discussed in a separate EPA control
device evaluation report.
Control of model-plant absorber vent emissions (see Table V-l) is by a thermal
oxidizer coupled with a waste heat boiler to generate low-pressure steam. Sup-
plemental fuel (natural gas) and the vent gas are fired through separate burners.
The vent gas can also be effectively oxidized in a conventional steam generator
p
through use of a specially designed burner unit. When the vent gas is used as
supplemental fuel, its combustion in the boiler is essentially complete, with an
expected emission reduction efficiency of greater than 99%.
Flaring of the vent gases has been practiced at some plants in the past; flaring,
however, does not allow recovery of heat. A flare incorporating appropriate safety
features could be used for controlling the absorber vent emissions and startup
emissions if heat recovery is not to be considered. Flares and the use of emis-
sions as fuel are the subject of a separate EPA control device evaluation report.
Catalytic oxidation would not provide additional advantage over thermal oxidation,
since the gas mixture has a high heating value. However, if surplus heat avail-
ability negates the value of steam production, catalytic oxidation may be a viable
option.
2. Product Fractionator Vent
The model-plant fractionator vent emissions are controlled by recycle. A surface
condenser is used to condense the vapor issuing from the steam jet ejector. The
condensate, containing approximately 95% of the VOC, is returned to the secondary
-------
Table V-l. Total Controlled VOC Emissions for Model-Plant Formaldehyde
Production Using a Metallic Silver Catalyst
Stream
Designation
Source (Fig. III-l)
Absorber
Production fractionator
Methanol storage
Formaldehyde storage
Handling
Fugitive
A
B
D
D
F
H
Secondary K
Total (product fractionator emissions
Control Device
or Technique
Thermal oxidation
Recycle
Water scrubber
Internal-floating-roof
tank
Vent scrubber
Vapor recovery
Repair and maintenance
None
recycled)
Emission
Reductior
99
100
95
85
96
96
81
96
Emissions
a
i Ratio
(g/kg)
0.047
Negligible
0.079
0.012
0.003
0.001
0.21
0.006
0.28
Rate
(kg/hr)
0.24
Negligible
0.41
0.06
0.01
0.005
1.09
0.03
1.44
*g of total VOC per kg of 37% formaldehyde solution produced.
'Average rate for entire year, based on 8 startups per year.
-------
V-3
absorber as a part of the makeup water. The remainder of the VOC is recovered by
returning the uncondensed gases to the air-compressor suction manifold, thus obtain-
ing essentially 100% reduction in emissions. The uncondensable portion of the
gases will be subsequently emitted from the absorber vent. These gases contain
essentially no VOC.
An alternate method of control could be achieved by directing the uncondensed
gases from the separator to the proposed absorber-vent thermal oxidizer. With
this option a total VOC reduction efficiency of 99.95% would be achieved, with
95% of the formaldehyde and methanol in the emission stream recovered.
A mechanical vacuum pump could be installed to replace the steam jet ejector.
The gases and vapors exhausted by the pump would be returned to the air-compres-
sor suction manifold, thus achieving total recycle.
A water scrubber could be used as an alternate control option, with a resultant
removal efficiency of greater than 95% based on average scrubber efficiencies
reported by scrubber manufacturers.* The discharge water can be used as makeup
for the product absorber, thus allowing the methanol and formaldehyde scrubbed
from the vent gas to be recovered.
A final control option would be to combine the emissions from the fractionator
vent with the absorber-vent emissions going to the proposed thermal oxidizer. An
overall reduction of 99% would be achieved.
3. Intermittent Air Emissions
The model plant is assumed to operate at one-half the normal production rate
during startup. Since a relatively small quantity of VOC is emitted during the
early startup phase, it is normally vented to the atmosphere until the emission
composition reaches a steady state above the flammable range. If a suitable
flare system is available at a nearby production unit on the plant site, the
startup emissions could possibly be routed to the existing flare. Suitable
safety precautions, such as purging with natural gas, would be required. The
capital cost of a flare system may be less than that for thermal oxidation; however,
the cost effectiveness of a flare is poor because heat recovery is precluded.
*Information contained in various catalogs and/or sales brochures.
-------
V-4
A multipurpose thermal oxidizer designed to handle the normal absorber-vent flow
and the converter startup flow would not be feasible because of the design condi-
tions and the infrequent occurrence of the startup emissions. At peak flow during
the startup period the heat produced by the quantity of combustible gases dis-
charged from the converter is approximately 3 times that of the heat capacity of
the proposed absorber-vent thermal oxidizer. The specific heat of the startup
mixture reaches a level of more than twice that of the normal absorber-vent gas.
The duct work from the reactor to the thermal oxidizer and the burner internals
would have to be designed to accommodate the hot gases that are discharged directly
from the reactor, or some arrangement would have to be included to cool the gas.
Explosive mixtures could inadvertently occur during startup, with the thermal
oxidizer acting as a potential source of ignition. During the latter phase of
the startup a source of ignition would not be needed, because the reactor tempera-
ture would be above the autoignition temperature of the mixture. The system there-
fore would have to be designed for the peak startup duty with turndown capability
for the normal vent flow. More sophisticated controls would be required to accom-
modate the varying composition and quantity of these gases during the startup
period, and explosion vents would have to be added to the system.
4. Fugitive Sources
4
Controls for fugitive sources are discussed in a separate EPA report. Fugitive
emissions from equipment, pumps, valves, and the process water cooling tower can
be controlled by an appropriate leak-detection system, plus repair and maintenance
as needed. Controlled fugitive emissions calculated with the factors given in
Appendix C are included in Table V-l. The factors are based on the assumption
that major leaks are detected and repaired.
5. Storage and Handling Sources
a. Methanol Storage -- Internal-floating-roof tanks* are commonly used for control
of storage-tank VOC emissions and are used in the model plant for methanol stor-
age control. The controlled methanol emissions given in Table V-l were calculated
by assuming that a contact-type internal floating roof with secondary seals will
reduce fixed-roof-tank emissions by 85%.
*Consist of internal floating covers or covered floating roofs as defined in API
25-19, 2d ed. (fixed-roof tanks with internal floating device to reduce vapor
loss).
-------
V-5
b. Formaldehyde Storage -- Formaldehyde storage emissions are controlled in the model
plant by a vent scrubber. A portion of the fresh feed water going to the second-
ary absorption tower is used as the scrubber medium. The water, after passing
through the vent scrubber, is used for product dilution or is returned to the
absorption tower as makeup. The tank emissions recovered are thus returned to
the process. A reduction efficiency of 96%, typical of an average scrubber system,
was used to calculate the controlled emissions given in Table V-l.
c. Other Tank Emissions -- The methanol recycle tank is small and has low emissions
and is uncontrolled in this model plant.
d. Handling — Emissions occurring during the loading of tank cars and tank trucks
are controlled by a vapor recovery system. The vapors displaced are returned to
the proposed formaldehyde tank-vent scrubber system. The controlled handling
emissions given in Table V-l were calculated on the assumption of 96% reduction
efficiency.
6. Secondary Emissions
Secondary emissions result from evaporation of VOC contained in aqueous effluent
from the plant. Control of secondary emissions is discussed in a separate EPA
report. No con
the model plant.
report. No control system has been identified for the secondary emissions from
7. Current Emission Control
The control devices being used by domestic formaldehyde producers are discussed
in Appendix E.
B. FORMALDEHYDE FROM METHANOL PROCESS USING A METAL OXIDE CATALYST
1. Absorber Vent
Thermal oxidation controls the absorber-vent emissions from the model plant. A
VOC reduction of greater than 99% can be achieved and was used to calculate the
controlled emission rate given in Table V-2. Heat from the oxidizer flue gases
can be recovered either by generating steam by a waste heat boiler or by preheat-
ing the incoming vent gas and combustion air with recuperative heat exchangers.
-------
Table V-2. Total Controlled VOC Emissions for Model-Plant
Formaldehyde Production Using a Metal Oxide Catalyst
Stream
Designation
Source (Fig. III-l)
Absorber
Methanol storage
Formaldehyde storage
Handling
Fugitive
Secondary
Total
A
D
D
F
H
K
Control Device
or Technique
Thermal oxidation
Covered floating- roof
tank
Vent scrubber
Vapor recovery
Repair and maintenance
None
Emission
Reduction
99
85
96
96
74
93
Emissions
Ratio
(g/kg)
0.032
0.014
0.003
0.001
0.19
0.047
0.287
Rate
(kg/hr)
0.16
0.07
0.01
0.005
0.98
0.24
1.47
of total VOC per kg of 37% formaldehyde solution produced.
-------
V-7
The vent gas is largely inert, thus requiring that supplemental fuel be fired
through a separate burner. Thermal oxidizer systems and efficiencies are dis-
cussed in a separate EPA control device evaluation report.
A larger portion of the heat can be recovered by generation of steam than by recu-
perative heating, making the steam generation option more attractive from an energy
standpoint. Since the metal-oxide-catalyst process itself produces an excess of
steam, recuperative heating may be attractive for plants having no use for the
steam. Recuperative heating greatly reduces the quantity of fuel required to
maintain combustion conditions.
One manufacturer reported using a water scrubber to control the absorber-vent
emissions. The performance of the scrubber, however, is hampered by the in-
soluble nature of the dimethyl ether contained in the vent stream. For the model
plant the overall VOC removal efficiency for this system would be 74%. However,
wastewater from the scrubber is discharged to wastewater treatment. Thus the
secondary emissions due to evaporation of the absorbed VOC would increase the
rate of secondary emissions by 350%.
2. Fugitive Sources
4
Controls for fugitive sources are discussed in a separate EPA report. Control
of emissions from pumps and valves can be attained by an appropriate leak detec-
tion system followed by repair maintenance. Controlled fugitive emissions have
been calculated with the factors given in Appendix C and are included in Table V-2.
The factors are based on the assumption that major leaks are detected and repaired.
3. Storage and Handling Sources
a. Methanol Storage -- Internal-floating-roof tanks are commonly used for control of
storage-tank VOC emissions and are used in the model plant for methanol storage
control. The controlled methanol emissions given in Table V-2 were calculated by
assuming that a contact-type internal floating roof with secondary seals will
reduce fixed-roof-tank emissions by 85%.
b. Formaldehyde Storage — A vent scrubber system is used to control the formalde-
hyde storage-tank vent emissions. A portion of the fresh feed water going to the
-------
V-8
secondary absorption tower is used as the scrubber medium. The scrubber discharge
water is returned to the absorption tower or is used for product dilution. The
tank emissions recovered are thus returned to the process. A reduction effi-
ciency of 96% was used to calculate the controlled emissions given in Table V-2.
c. Handling — Vapors displaced while tank cars and tank trucks are being loaded are
controlled by a vapor recovery system. A vent line is attached to the vessel
being filled and the vapors are returned to the proposed formaldehyde tank vent-
scrubber system. The controlled handling emissions given in Table V-2 were cal-
culated on the assumption of 96% reduction efficiency for the scrubber system.
4. Secondary Emissions
Sources of secondary emissions from a plant using the metal-oxide-catalyst process
are from evaporation of VOC contained in aqueous wastes going to wastewater treat-
ment and from the vent on the heat transfer system. No control system has been
identified for the secondary emissions from the model plant. Control of secondary
emissions is discussed in a separate EPA report.
5. Current Emission Control
The control devices being used by domestic formaldehyde producers are discussed
in Appendix E.
-------
V-9
C. REFERENCES*
1. J. W. Blackburn, IT Enviroscience, Inc., Control Device Evaluation. Thermal
Oxidation (July 1980) (EPA/ESED report, Research Triangle Park, NC).
2. J. F. Lawson, IT Enviroscience, Inc., Trip Report for Borden Plant, Fayetteville,
NC, Aug. 24, 1977 (on file at EPA, ESED, Research Triangle Park, NC).
3. 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).
4. D. G. Erikson and V. Kalcevic, IT Enviroscience, Inc., Fugitive Emissions
(September 1980) (EPA/ESED report. Research Triangle Park, NC).
5. W. T. Moody, TRW, Inc., letter dated Aug. 15, 1979, to D. A. Beck, EPA.
6. J. Cudahy and R. Standifer, IT Enviroscience, Inc., Secondary Emissions (June
1980) (EPA/ESED report, 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.
-------
VI-1
VI. IMPACT ANALYSIS
A. ENVIRONMENTAL AND ENERGY IMPACTS
1. Formaldehyde Model Plants
The environmental impact of reducing total VOC emissions by application of the
described control systems to the model plants (Table VI-1) would be 330 Mg/yr for
the metallic-silver-catalyst process and 170 Mg/yr for the metal-oxide-catalyst
process. By incorporating thermal oxidation with heat recovery 6892 MJ/hr of
energy is recovered by the metallic-silver-catalyst process and 7447 MJ/hr is
recovered by the metal-oxide-catalyst process. Deducting the energy required to
operate all emission controls, including thermal oxidizer auxiliary fuel, gives a
net energy gain of 5286 MJ/hr for the metallic-silver-catalyst process and a net
consumption of 2751 MJ/hr for the metal-oxide-catalyst process.
2. Metallic-Silver-Catalyst Process
Table VI-1 shows the environmental impact of reducing total VOC emissions by appli-
cation of the described control systems to the model plant. The addition of emis-
sion controls under option A will result in a reduction of 96%, or 330 Mg of VOC
emissions per year, and a net recovery of energy of 5286 MJ/hr.
A typical uncontrolled plant using the metallic-silver-catalyst process will
require about 198 kJ of energy per kg of formaldehyde solution produced. The
potential net energy savings is 1029 kJ/kg. Thus with heat recovery applied, a
typical plant could produce 831 kJ of excess energy per kg of formaldehyde solution
produced, which is equivalent to a production of 4269 MJ/hr for the model plant.
If heat recovery were not incorporated, the total energy consumption of the plant
with the emission controls applied would be 2569 MJ/hr.
3. Metallic-Oxide-Catalyst Process
Table VI-2 shows the environmental impact of reducing total VOC emissions by appli-
cation of the described emission control systems to the model plant. The controls
described will reduce total VOC emissions by 94%, or 171 Mg/yr.
Three types of thermal oxidation systems are described: oxidation with conven-
tional heat recovery (generation of steam), oxidation with recuperative heat
-------
Table VI-1. Environmental Impact of Controlled Model-Plant Formaldehyde
Production by Metallic-Silver-Catalyst Process
Source
Absorber
Product fractionator
Methanol storage
Formaldehyde storage
Handling
Fugitive
Secondary
Total (product
Stream
Designation
(Fig. III-l)
A
B
D
D
F
H
K
fractionator emissions
Control Device Total VOC Emission Reduction
or Technique
Thermal oxidation
Recycle process
Water scrubber*
Internal- floating- roof tank
Vent scrubber
Vapor recovery
Repair and maintenance
None
recycled)
(%)
99
100
95
85
96
96
81
96
(Mg/yr)
212.4
71.5
67.9
3.2
1.4
1.0
40.6
330.1
<
H
1
tvj
*Alternate system
-------
Table VI-2. Environmental Impact of Controlled Model-Plant Formaldehyde
Production by Metal-Oxide-Catalyst Process
Source
Absorber
Methanol storage
Formaldehyde storage
Handling
Fugitive
Secondary
Total
Stream
Designation
(Fig. III-l)
A
D
D
F
H
K
Control Device
or Technique
Thermal oxidation
Internal-floating-roof tank
Vent scrubber
Vapor recovery
Repair and maintenance
None
Total VOC Emission Reduction
(%)
99
85
96
96
74
94
(Mg/yr)
141.6
3.2
1.4
1.0
23.8
171.0
M
1
Ul
-------
recovery (preheating of the vent gas and combustion air), and oxidation without
heat recovery. With conventional heat recovery the net energy consumption for
all controls would be 2728 MJ/hr. With recuperative heat recovery the net energy
consumed would be 3878 MJ/hr and without heat recovery would be 10,156 MJ/hr.
A typical uncontrolled plant using the metal-oxide-catalyst process produces a
net excess of exportable energy in the form of low-pressure steam at the rate of
232 kJ per kg of formaldehyde solution produced. The emission controls described
for conventional heat recovery consume 531 kJ of energy per kg of formaldehyde
solution produced. The net energy consumed by the model plant would be 299 kJ
per kg of formaldehyde solution produced.
B. CONTROL COST IMPACT
Estimated costs and cost-effectiveness data for control of VOC emissions result-
ing from the production of formaldehyde are given in this section. Details of
the model plants (Figs. III-l and III-2) are given in Sects. Ill and IV. Cost
estimate calculations are included in Appendix D.
Capital cost estimates, based on December 1979 costs, represent the total invest-
ment 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 costs of formaldehyde produc-
tion lost during installation or startup, research and development, or land acquisi'
tion.
Bases for the annual cost estimates for the control alternatives include utilities,
operating labor, maintenance supplies and labor, recovery credits, capital charges,
and miscellaneous recurring costs such as taxes, insurance, and administrative
overhead. The cost factors used are itemized in Table VI-3. Recovery credits
2
are based on the raw-material market value for the material being recovered.
Annual costs are for a 1-year period beginning December 1979.
1. Metallic-Silver-Catalyst Process
a. Absorber Vent (Thermal Oxidizer) -- The cost of installing a thermal oxidizer
system with heat recovery to control VOC emissions from the model-plant absorber
-------
VI-5
Table VI-3. Cost Factors Used in Computing Annual Costs
Utilities
Cooling water
Electricity
Natural gas
Steam
Fixed costs
Maintenance labor plus materials, 6%
Capital recovery, 18% (10 yr life @ 12% int.)
Taxes, insurance, administration charges, 5%
Recovery credits*
Energy
Methanol
Formaldehyde (raw-material value)
$0.026/m ($0,10/M gal)
$8.33/GJ ($0.03/kWh)
$1.90/GJ ($2.00/M Btu)
$5.50/Mg ($2.50/M Btu)
29% of installed capital
$1.90/GJ ($2.00/M Btu)
$0.17/kg
$0.18/kg
*The values used for methanol and formaldehyde solution were taken from
Chemical Marketing Reporter, ref. 2.
-------
Table VI-4. Estimates of Emission Control and Reduction and Cost Effectiveness
for Formaldehyde Model Plant by Metallic-Silver-Catalyst Process
Emission Source
Product absorber vent
Product fractionation vent
Formaldehyde storage and handling
Total Installed
Capital Cost
Control (X 1000)
Thermal oxidation
No heat recovery $370
With heat recovery 540
Recycle condenser 47
Scrubber 40
Vent scrubber 49
Annual Operating Cost (X 1000) Total VOC Total VOC
... . ......... Emission Cost
Recovery Reduction Effectiveness
Fixed Cost Utilities Manpower Credit Net (Mg/yr) (per Hg)
$107 $43 $18 $ 0 $168 212.4 $ 791
157 43 .36 (132)* 104 212.4 489
14 0.5 b (11.8) 2.7 71.5 35
12 b 3.6 (11.7) 3.9 67.9 57
14 b 3.6 (0.4) 17.2 2.4 7,167
"values in parentheses in these columns represent savings.
Negligible.
H
1
cn
-------
VI-7
700
600
o
o
o
500 -
CJ
•O
U>
s
H
o
0)
Q
400 -
300 -
200
100
200
Plant Capacity (Gg/yr)
(1) Thermal oxidizer with conventional heat recovery, silver catalyst process
(2) Thermal oxidizer without heat recovery, silver catalyst process
(3) Thermal oxidizer with recuperative heat recovery, metal oxide process
(4) Thermal oxidizer with conventional heat recovery, metal oxide process
(5) Thermal oxidizer without heat recovery, metal oxide process
Fig. VI-1. Installed Capital Cost vs Plant Capacity
for Emission Control by Thermal Oxidation
-------
VI-8
500
o
o
o
4J
to
rH
Oj
g
-P
0
400 -
300 -
200 —
100 —
Plant Capacity (Gg/yr)
(1) Thermal oxidizer with conventional heat recovery, silver catalyst proc
(2) Thermal oxidizer without heat recovery, silver catalyst process
(3) Thermal oxidizer with recuperative heat recovery, metal oxide process
(4) Thermal oxidizer with conventional heat recovery, metal oxide process
(5) Thermal oxidizer without heat recovery, metal oxide process
Fig. VI-2. Net Annual Cost or Savings vs Plant Capacity
for Emission Control by Thermal Oxidation
-------
2500
I
•w-
M
to
V
c
gi
•rl
ti
OJ
IH
M
CO
8
2000 -
1500 _
1000 _
500 _
15
20
30
40
60
80
100
200
Plant Capacity (Gg/yr)
(1)
(2)
(3)
(4)
(5)
Thermal oxidizer with conventional heat recovery, silver catalyst process
Thermal oxidizer without heat recovery, silver catalyst process
Thermal oxidizer with recuperative heat recovery, metal oxide process
Thermal oxidizer with conventional heat recovery, metal oxide process
Thermal oxidizer without heat recovery, metal oxide process
Fig. VI-3. Cost Effectiveness vs Plant Capacity for
Emission Control by Thermal Oxidation
-------
VI-10
vent is estimated to be $540,000. If heat is not recovered, the cost of the system
would be $370,000 (see Table VI-4), based on installation of the equipment, piping,
and controls necessary for a complete and operating system. Since the vent gas
rate varies directly with production, a plant twice the size of the model plant
would have twice the emissions from this vent. Curves 1 and 2 of Fig. VT-1 were
plotted to show the variation of installed capital cost with plant capacity for
oxidation systems with and without heat recovery.
To determine the cost effectiveness of the thermal oxidation systems, estimates
were made of the direct operating cost, the capital recovery cost, and miscellan-
eous capital costs; for the system incorporating heat recovery a recovery credit
was calculated from the heating value of the vent gas. For the model plant re-
covering heat by conventional heat recovery the recovery credit is $132,000/yr,
resulting in a net annual cost of $104,000. Without heat recovery the net annual
cost would be $168,000 (see Table VI-4). The variation of net annual cost with plan
capacity is shown by curves 1 and 2 of Fig. VI-2 for both oxidation systems. The
variation of cost effectiveness with plant capacity is shown by curves 1 and 2 of
Fig. VI-3.
b. Fractionator Vent — The two options described in Sect. V for controlling the
fractionator-vent emissions are discussed below:
Recycle -- The emissions from the product fractionator vent are controlled by
recycling the vapors back to the process. The installed capital cost of the vacuum
jet condenser system for the model plant is estimated to be $47,000 (see Table VI-4)
The variation of the estimated installed cost of the recycling system with plant
capacity is shown by curve 1, Fig. VI-4. These estimates are based on installation
of a water-cooled condenser and drum separator and include the cost of all piping
and controls required for a complete and operating system. Recycling the fractiona*
tor emissions results in a net annual operating cost of $2500. Curve 1 of Fig. VI-5
shows the variation of net annual cost with plant capacity. The cost effectiveness
of the system results in a cost of $35 per Mg of VOC emission removed. The varia-
tion of cost effectiveness with plant capacity is given by curve 1 of Fig. VI-6.
Scrubber — A water scrubber unit is installed for control of emissions from the
product fractionation vent. Scrubber discharge water is used as makeup water for
the product absorber. The installed capital cost of the complete scrubber system
-------
VI-11
o
o
o
O
u
rt
4-1
•H
u
T3
(U
CO
H
u
Q)
Q
60
40 —
20
20
30 40 60 80
Plant Capacity (Gg/yr)
100
200
(1) Fractionator jet/condenser system, silver catalyst process
(2) Fractionator vent scrubber, silver catalyst process
Fig. VI-4. Installed Capital Cost vs Plant Capacity for
Fractionator Vent Emission Controls
-------
VI-12
10
o
o
o
C/J
CP
ti
•H
-u
en
o
U
c
c,
ft,
-p
0)
2
20
40 60 80
Plant Capacity (Gg/yr)
(1) Fractionator jet/condenser system, silver catalyst process
(2) Fractionator vent scrubbing, silver catalyst process
Fig. VI-5. Net Annual Cost vs Plant Capacity for
Fractionator Vent Emission Controls
-------
VI-13
300
w
ID
0)
£
•H
JJ
O
1)
M-l
IH
W
-P
(0
200
g 100'
u
100
I
I I
15
20
30 40 60 80 100
Plant Capacity (Gg/yr)
200
(1) Fractionator jet/condenser system, silver catalyst process
(2) Fractionator vent scrubber, silver catalyst process
Fig. VI-6. Cost Effectiveness vs Plant Capacity for
Fractionator Vent Emission Controls
-------
VI-14
for the model plant is estimated to be $40,000 (see Table VI-4). The variation
of the estimated cost of the scrubber system with plant capacity is shown by
curve 2 of Fig. VI-4. With recovery of the emissions by recycle of the scrubber
water the net annual cost is 3900. Curve 2 of Fig. VT-5 shows the variation of
net annual operating cost with plant capacity. The cost effectiveness results in
a cost of $57 per Mg of VOC removed (see curve 2 of Fig. VI-6).
c. Storage and Handling -- Storage and handling cost impacts for emissions control
resulting from the production of formaldehyde by the metallic-silver-catalyst
process are described below.-
Methanol -- Model-plant methanol storage emissions are controlled by the use of
floating-roof tanks. The installed capital cost, net annual cost, and cost-
effectiveness data for new internal-floating-roof tanks are discussed in a separate
report covering storage and handling.
Formaldehyde — Model-plant formaldehyde storage and handling emissions are con-
trolled by a vent scrubber system. The scrubber discharge water is used for pro-
duction dilution or is recycled to the product absorber. The installed capital
cost, net annual cost, and cost-effectiveness data for installation of a vent
scrubber system complete with vent manifold, piping, and controls are given in
Table VI-4. The variation of installed capital cost, net annual cost, and cost
effectiveness with plant capacity is shown by Fig. VI-7, Fig. VI-8, and Fig. VI-9,
respectively. For the model plant the cost effectiveness of the system is $7167
per Mg of VOC emissions removed.
e. Fugitive Sources -- A control system for fugitive sources is defined in Appendix C.
Fugitive emissions and their applicable controls for all the synthetic organic
4
chemicals manufacturing industry are discussed in a separate report.
f. Secondary Sources -- No control system has been identified for the secondary emis-
sions from the model plant. Secondary emissions and their control are covered by
a separate EPA report.
-------
VI-15
I I I ' ' '
40 50 60 80
Plant Capacity (Gg/yr)
100
'200
Fig. VI-7. Installed Capital Cost vs Plant Capacity for
Formadlehyde Storage Emission Control
-------
VI-16
25
o
o
o
Jj
to
5
4J
0)
20
15
10
I
I
_L
I I I i
15
20
30 40 50 60 80
Plant Capacity (Gg/yr)
100
Fig. VI-8. Net Annual Cost vs Plant Capacity for
Formaldehyde Storage Emission Control
-------
VI-17
14,000
tn
tn
0)
U
0)
M-l
U
•U
W
12,000
— 10,000
8,000
6,000
4,000
2,000
15
20
I
I
J I
30 40 50 60 70 8tf 100
Plant Capacity (Gg/yr)
200
Fig. VI-9. Cost Effectiveness vs Plant Capacity for
Formaldehyde Storage Emission Control
-------
VI-18
2. Metal-Oxide-Catalyst Process
a- Absorber Vent — The capital cost for installation of a thermal oxidizer system
with conventional steam-generation heat recovery to control VOC emissions from
the model-plant absorber vent is estimated to be $448,000. A system with recupera-
tive heat recovery would cost $459,000, and an oxidation system installed without
heat recovery would cost $350,000 (see Table VI-5). The variation of installed
capital cost with plant capacity for each type of oxidation system is shown by
curves 3, 4, and 5 of Fig. VI-1. To determine the cost effectiveness of each
oxidation system, estimates were made of the direct operating cost and of capital
recovery costs, and a capital recovery credit was calculated from the heating
value of the vent gas. The installed capital cost, net annual cost, and cost-
effectiveness data for installation of an oxidizer with conventional heat recovery,
with recuperative heat recovery, and without heat recovery are given in Table VI-5.
The variation of net annual cost for each system is shown by curves 3, 4, and 5
of Fig. VI-2. The variation of cost effectiveness for each oxidation system is
shown by curves 3, 4, and 5 of Fig. VI-3.
b. Storage and Handling — Storage and handling cost impacts for emission control
resulting from the production of formaldehyde from methanol by the metal-oxide-
catalyst process are discussed below:
Methanol -- Model-plant methanol storage emissions are controlled by the use of
floating-roof tanks. Installed capital cost, net annual cost, and cost-effective-
ness data for new internal-floating-roof tanks are discussed in a separate report
covering storage and handling.
Formaldehyde — Model-plant formaldehyde storage and handling emissions are con-
trolled by a vent scrubber system. The scrubber discharge water is used for product
dilution or is recycled to the product absorber. The installed capital cost, net
annual cost, and cost-effectiveness data for installation of a vent scrubber system
complete with vent manifold, piping, and controls are given in Table VI-5. The
variation of installed capital cost, net annual cost, and cost effectiveness with
plant capacity is shown by Fig. VI-7, Fig. VI-8, and Fig. VI-9, respectively.
-------
Table VI-5. Estimates of Emission Control and Reduction and Cost Effectiveness
for Formaldehyde Model Plant Using a Metal Oxide Catalyst
Emission Source
Product absorber vent
Formaldehyde storage and handling
Control
Thermal oxidation
Ho heat recovery
Steam generation
Recuperative heat
Vent scrubber
Total Installs
Capital Cost
(X 1000)
$350
448
459
49
j Annual Operating Cost (X
Fixed Cost
SlOl
130
133
14
Utilities
$135
135
13
a
Manpower
$16
36
32
3.6
1000)
Recovery
Credit
$0
(95)
0
(0.4)
Net
$254
206
178
17.2
Total VOC
Emission
Reduction
(Mg/yr)
142
142
142
2.4
Total VOC
Cost
Effectiveness
(per Mq)
$1,794
1,454
1,257
7,167
Negligible.
-------
VI-20
c. Fugitive Sources -- A control system for fugitive sources is defined in Appendix C.
4
A separate report covers fugitive emissions and their applicable controls for
all the synthetic organic chemicals manufacturing industry.
d. Secondary Sources — No control system has been identified for the secondary emis-
sions from the model plant. Secondary emissions and applicable controls are discus
sed in a separate report.
-------
VI-21
C. REFERENCES*
1. G. E. Haddeland and G. K. Chang, Report No. 23. Formaldehyde, pp. 104—106,
A private report by the Process Economics Program, Stanford Research Institute,
Menlo Park, CA (Feb. 1967).
2. "Current Prices of Chemicals and Related Materials," Chemical Marketing Reporter
214(5), 34, 35 (July 31, 1978).
3. D. G. Erikson, IT Enviroscience, Inc., Storage and Handling (September 1980)
(EPA/ESED report, Research Triangle Park, NC).
4. D. G. Erikson and V. Kalcevic, IT Enviroscience, Inc., Fugitive Emissions
(September 1980) (EPA/ESED report. Research Triangle Park, NC).
5. J. Cudahy and R. L. Standifer, IT Enviroscience, Inc., Secondary Emissions
(June 1980) (EPA/ESED report, 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.
-------
VII-1
VII. SUMMARY
Formaldehyde is produced in the United States from methanol by either a metallic-
silver-catalyst process or a metal-oxide-catalyst process. In 1977 the total
production from 55 operating plants was 2750 Gg. The formaldehyde production
capacity of these plants is reported to be 4066 Gg/yr. With the estimated average
annual consumption growth rate of 4 to 5% the production capacity is sufficient
to supply the demand through 1982.
Emission sources and control levels for the model plant are summarized in Table VII-1
Projecting these values for the entire domestic formaldehyde industry at full-capa-
city operation indicates that the total uncontrolled VOC emissions would be
3141 kg/hr. It is estimated that the total VOC emissions from the domestic formal-
dehyde industry in 1977 were of the order of 1153 kg/hr.
The predominant emission point is the product absorber vent. For the metallic-
silver-catalyst process the absorber-vent gas has a high heating value and can be
oxidized or be used as supplemental fuel in a steam generator. A VOC reduction
efficiency of greater than 99% results when the gas is burned. A thermal oxida-
tion system with a conventional steam-generating heat recovery boiler for the
silver catalyst model plant is estimated to cost $540,000. The recovery credit
for the steam generated would be $132,000/yr, for a net annual cost of $104,000.
The cost effectiveness of the system would be $489 per Mg of VOC removed.
The absorber-vent emissions from a plant using the metal-oxide-catalyst process
can also be controlled by thermal oxidation of the gas stream, which would result
in a total VOC reduction of greater than 99%. An oxidation system with a conven-
tional steam-generating heat recovery boiler would cost $448,000 for the model
plant. The recovery credit for the steam generated would be $95,000 per year.
However, since the vent gas contains large amounts of inert components, supple-
mental fuel must be used to fire the oxidizer. The net annual cost therefore
would be $206,000 per year, which results in a cost effectiveness of $1454 per Mg
of VOC removed.
lj. L. Blackford, "CEH Marketing Research Report on Formaldehyde," pp. 658.50310—
658.5033E in Chemical Economics Handbook, Stanford Research Institute, Menlo
Park, CA (April 1977).
-------
Table VII-1. VOC Emission Summary for Model Plant
Emission Source
Absorber vent
Fractionator vent
c
Intermittent
Storage and handling
Fugitive
Secondary
Total
Stream or Vent
Designation
(Figs. III-l, -2)
A
B
C
D.F
H
K
Emission Rate (kg/hr)
Metallic-Silver-Catalyst Process
Uncontrolled
24.5
8.2
0.28
0.73
5.7
0.03
39.44
Controlled
0.24
Negligible
0.28
0.075
1.09
0.03
1.71
Metal-Oxide-Catalyst Process
Uncontrolled
16.3
NAb
NA
0.74
3.7
0.2
20.94
Controlled
0.16
NA
NA
0.085
0.98
0.24
1.46
All emissions are based on 8760 hr of operation per year.
Not applicable to process.
CAverage rate for entire year, based on 8 startups per year.
H
H
tvj
-------
VII-3
For those plants using the metal-oxide-catalyst process that do not have a use
for the excess steam generated by the heat recovery boiler, an oxidation system
with recuperative heat recovery could be installed. This system, by preheating
vent gas and combustion air, greatly reduces the supplemental fuel required.
This system for the model plant would cost $459,000. Since this system does not
produce exportable energy, the net annual cost would be $178,000/yr, for a cost
effectiveness of $1257 per Mg of total VOC removed.
The metallic-silver-catalyst process incorporates a fractionator to separate the
excess methanol from the product. The emissions from the fractionator vent are
controlled in the model plant by recycling the vapor to the process, thus providing
essentially 100% VOC control. The cost of the model-plant recycle system is $47,000,
With a recovery credit of $11,800 taken for the value of the methanol and formal-
dehyde that are recovered, the net annual cost is $2700. The cost effectiveness
is $35 per Mg of VOC removed.
The model-plant methanol storage emissions are controlled by internal-floating-
roof tanks. Costs for internal floating roofs are given in a separate report
2
covering storage and handling. The model-plant formaldehyde-solution storage
and handling emissions are controlled by a vent scrubber system. The scrubber
cost for either process is $49,000. The net annual cost is $17,200/yr, for a
cost effectiveness of $7167 per Mg of VOC removed.
2D. G. Erikson, IT Enviroscience, Inc., Storage and Handling (September 1980)
(EPA/ESED report. Research Triangle Park, NC).
-------
A-l
APPENDIX A
Table A-l. Properties of Anhydrous Formaldehyde and Methanol
Methanol'
Formaldehyde
Synonyms
Molecular formula
Molecular weight
Physical state
Vapor pressure
Density
Boiling point
Melting point
Water solubility
Methyl alcohol, carbinol,
methyl hydroxide
32.04
Liquid
17.05 kPa at 25°C
0.7913 at 20°C/4°C
64.88C
-93.9°C
Infinite
Methanal, methyl
aldehyde
CH O
30.03
Colorless gas
259.67 KPa at 25°C
0.815 at 20°C/4°C
-21°C
-92 °C
Soluble
aFrom: J. Dorigan et al, "Formaldehyde," p. AIII-154 in Scoring of Organic
Air Pollutants. Chemistry, Production and Toxicity of Selected Synthetic
Organic Chemicals (Chemicals F-N), Rev 1, Appendix III, MTR-7248, MITRE Corp.,
McLean, VA (September 1976).
blbid., p AIII-12.
Table A-2. Properties of Formaldehyde Solution (37 wt %)
Synonyms
Molecular formula
Molecular weight
Physical state
Density
Boiling point
Water solubility
Formalin
CHnO
30.03
Clear liquid
1.113 g/ml at 18°C
99 °C
Soluble
-------
A-2
Table A-3. Properties of Paraformaldehyde
Molecular formula
Molecular weight
Physical state
Vapor specific gravity
Boiling point
Melting point
Water solubility
HO(CH_0) H
— a
(30.03) + 18
ri
White solid
0.815 at 20°C/4°C
Depolymerizes at 120 to 200°C
120 to 170°C
Soluble
n ranges from about 8 to 100.
-------
B-l
APPENDIX B
Table B-l. Air-Dispersion Parameters for
Metallic-Silver-Catalyst Process Model Plant
Source
with
Total
Emission
Bate
a Capacity of 45
Height
Diameter
(m)
Gg/yr
Discharge
Temperature
(K)
Flow Discharge
Rate Velocity
(m-Vsec) (m/sec)
Uncontrolled Emissions
Absorber vent
Fractionator vent
Startup vent
Methanol recycle tank
Methanol storage (2 tanks)
Formaldehyde storage (4 tanks)
Formaldehyde handling
b
Fugitive emissions
Secondary emissions
6.80
2.27
167.20
0.01
0.11
0.05
0.03
1.58
0.01
19.8
21.3
2.4
7.3
7.3
c
0.61
0.05
0.51
1.4
5.8
5.8
c
302
297
533
297
Ambient
327
327
Ambient
Ambient
3.18 3.32
0.02 2.00
5.59 8.37
Controlled Emissions
d
Absorber vent
Methanol storage (2 tanks)
Formaldehyde storage arid
handling (4 tanks)
Fugitive emissions
0.068
0.017
0.003
1307
30.0
7.3
7.3
0.61
5.8
5.8
533
Ambient
302
Ambient
9.60 10
Peak flow conditions during startup.
Fugitive emissions are distributed over a 50-m by 150-m area.
Surface of ground level waste water treatment system.
3
Thermal oxidizer system.
-------
Table B-2. Air-Dispersion Parameters for Metal-Oxide-Catalyst
Process Model Plant with a Capacity of 45 Gg/yr
Source
Absorber vent
Methanol storage (2 tanks)
Formaldehyde storage (4 tanks)
Formaldehyde handling
Fugitive emissions
Secondary emissions
Absorber vent
Methanol storage (2 tanks)
Formaldehyde storage and
handling (4 tanks)
Fugitive emissions
Total VOC
Emission
Rate Height Diameter
(g/sec) (m) (m)
Uncontrolled Emissions
4.54 19.8 0.76
0.12 7.3 5.8
0.05 7.3 5.8
0.03
1.06
0.07
Controlled Emissions
0.045 30.0 0.76
0.019 7.3 5.8
0.003 7.3 5.8
0.288
Discharge Discharge
Temperature e Velocity
(K) (m /sec) (m/sec)
302 4.99 3.34
Ambient
327
327
Ambient
Ambient
533 14.60 10
Ambient
302
Ambient
b
Fugitive emissions are distibuted over a 50-m by 150-m area.
Thermal oxidizer system.
Internal-floating-roof tanks.
vent scrubber.
Repaired and maintained.
-------
C-l
APPENDIX C
FUGITIVE-EMISSION FACTORS*
The Environmental Protection Agency recently completed an extensive testing
program that resulted in updated fugitive-emission factors for petroleum re-
fineries. Other preliminary test results suggest that fugitive emissions from
sources in chemical plants are comparable to fugitive emissions from correspond-
ing sources in petroleum refineries. Therefore the emission factors established
for refineries are used in this report to estimate fugitive emissions from
organic chemical manufacture. These factors are presented below.
Source
Uncontrolled
Emission Factor
(kg/hr)
Controlled
Emission Factor'
(kg/hr)
Pump seals ,
Light-liquid service
Heavy-liquid service
Pipeline valves
Gas/vapor service
Light-liquid service
Heavy-liquid service
Safety/relief valves
Gas/vapor service
Light-liquid service
Heavy-liquid service
Compressor seals
Flanges
Drains
0.12
0.02
0.021
0.010
0.0003
0.16
0.006
0.009
0.44
0.00026
0.032
0.03
0.02
0.002
0.003
0.0003
0.061
0.006
0.009
0.11
0.00026
0.019
Based on monthly inspection of selected equipment; no inspection of
heavy-liquid equipment, flanges, or light-liquid relief valves;
10,000 ppmv VOC concentration at source defines a leak; and 15 days
allowed for correction of leaks.
Light liquid means any liquid more volatile than kerosene.
*Radian Corp., Emission Factors and Frequency of Leak Occurrence for Fittings
in Refinery Process Units, EPA 600/2-79-044 (February 1979).
-------
D-l
APPENDIX D
COST ESTIMATE DETAILS
This appendix contains sample calculations showing how the costs presented in
this report were estimated.
The accuracy of an estimate is a function of the degree of data available when
the estimate was made. Figure D-l illustrates this relationship. The contin-
gency allowance indicated is included in the estimated costs to cover the
undefined scope of the project.
Capital costs given in this report are based on a screening study, as indicated
by Fig. D-l, based on general design criteria, block flowsheets, approximate
material balances, and data on general equipment requirements. These costs
have an accuracy range of +30% to -23%, depending on the reliability of the
data, and provide an acceptable basis to determine the most cost-effective
alternative within the limits of accuracy indicated.
A. THERMAL OXIDIZER METALLIC-SILVER-CATALYST MODEL PLANT
To determine the cost estimate for controlling the vent emissions from the
silver catalyst model-plant absorber vent, the emission flow details were taken
from Table IV-3:
628 g/kg X 45 Gg/yr -f 8760 hr/yr X 2.205 Ib/kg = 7114 Ib/hr.
The flow in mole/hr was calculated for each component; the total flow was
calculated to be 310 Ib-moles/hr, or 1855 scfm. The total lower heating value was
calculated to be 8.55 MM Btu/hr, or 77 Btu/scf.
The following control costs and cost-effective estimates given in Table D-l were
developed by making semilog extrapolations of the tables on pages B-22 and B-23
of the control device evaluation report for thermal oxidation.1
1J. W. Blackburn, IT Enviroscience, Inc., Control Device Evaluation. Thermal
Oxidation (July 1980) (EPA/ESED report, Research Triangle Park, NC).
-------
ESTIMATE. TYPE
USED BY ESTIMATOR
5CRCEkliUG(
(PRE-UM. EMC,. ^TUDV)
PHASE. IE
(PREUM. PROC. EUC^.)
PMA^E nr
(COMPLETE PROCESS
£fJCj. CESIQW^
•
•
•
•
•/• OP TOTAL.
PROBABL.E
CAP. COST)
E STI MAT E.D co ST
WlTH ALLOWANCE.
Fig. D-l. Precision of Capital Cost Estimates
-------
Table D-l.
Plant
Size
(Gg)
23
45
113
23
45
113
Rate
{scfm)
927
1855
4637
927
1855
4637
Installed
Capital
Cost
$327
370
430
$438
540
675
Fixed Cost
$ 95
107
125
$127
157
196
Annual
Utilities
$ 21
43
105
With
$21
43
105
Cost (XlOOO)
Manpower
No Heat
$18
18
18
Recovery
Credit
Recovery
$0
0
0
Net
$134
168
248
Operating
Cost
(per scfm)
$144
90
.53
VOC
Reduction
(Mg/yr)
106
212
531
Cost
Effectiveness
(per Mg)
$1,262
791
467
250-psi Steam Generator
$36
36
36
$66
132
329
$118
104
7
$127'
56
1.5
106
212
531
$1,111
489
13.
2
D
-------
D-4
THERMAL OXIDIZER—METAL-OXIDE-CATALYST MODEL PLANT
The emission flow details were taken from Table IV-5. Since the oxygen content
is significant and the thermal oxidation control device evaluation report is
based on no oxygen in the feed, the oxygen and the corresponding nitrogen will
be subtracted and assumed to be part of the air added for combustion.
The listed emissions of 1178 g/kg are therefore reduced by 93.2 g of oxygen and
93.2 (77/23) = 312 g of nitrogen. With zero oxygen and 1014 - 312 = 702 g of
nitrogen the total emissions for sizing the thermal oxidizers are 773 g/kg of
formaldehyde solution:
773 g/kg X 45 Gg/8760 X 2.205 Ib/kg = 8756 Ib/hr.
Component
Formaldehyde
Methanol
Dimethyl ether
CO
Totals
Feed Rate
(Ib/hr)
5.3
21.3
9.3
168.1
204
Heat Capacity „__
(Btu/lb)
7,410
8,896
12,358
4,347
5,267
I'Ul
(Btu/hr)
39,273
189,484
114,929
730,731
1,074,417
r Thermal Oxidizer Design
(Ib mole/hr)
0.17
0.67
0.21
6.00
7.05
Nitrogen with equivalent air subtracted
Oxygen 0
C02 25.4
284.03
0
0.58
291.66 X 359 =
104,706 scf/hr
1745 scfm
Therefore 1,074,417 -r 104,706 = 10.26 Btu/scf.
The control costs and cost-effectiveness estimates given in Table D-2 were
developed by making a semilog extrapolation of the tables on page B-16 of the
control device evaluation report for thermal oxidation.1
-------
Table D-2.
Annual Cost (X 1000)
Plant
Size
(Gg)
23
45
113
23
45
113
23
45
113
Rate
(scfiti)
872
1745
4362
872
1745
4362
872
1745
4362
Installed
Capital
Cost
$300
350
415
$370
459
575
$360
448
563
Fixed Cost
$87
101
120
$107
133
167
$140
130
164
Utilities
$67
135
338
70%
$ 6
13
32
$ 67
135
339
Net
Recovery Annual
Manpower Credit Cost .
No Heat Recovery
$18
18
18
Recuperative Heat Recovery
$32
32
32
250-psia Steam Generation
$36 $47
36 95
36 237
$172
254
476
$145
178
231
$160
206
302
Operating
Cost
(per scfra)
$192
145
109
$166
102
.53
$183
118
69
VOC
Reduction
(Mg/yr)
71
142
354
71
142
354
71
142
354
Cost
Effectiveness
(per Mg)
$2429
1794
1344
$2048
1257
652 0
Ul
$2260
1454
853
-------
D-6
FRACTIONATOR—JET CONDENSER SYSTEM METALLIC-SILVER-CATALYST PROCESS
The standard 56-ft2 jet condenser is adequate for the model plant. The capital
cost for the installed condenser with valving, fitting, piping, gauges, liquid
separator, level control, and pump is estimated to be $47,000.
Plant
Size
(Gq)
23
45
113
VOC
Reduction
(Mg/yr)
35.2
71.5
173.1
Installed
Capital
Cost
$47
47
49
Annual Cost (X 1000)
Fixed
Cost
$14
14
14
Cooling
Water
$0.5
0.5
0.7
Manpower
Negligible
Negligible
Negligible
Recovery
Credit
$6.0
11.8
29.8
Net
$ 8.
2.
(15.
- Cost
Effectiveness
(per Mg)
5 $241
7 35
4) t (89)
FRACTIONATOR VENT EMISSIONS SCRUBBER—METALLIC-SILVER-CATALYST PROCESS
The scrubber for the fractionator for the model plant is designed to handle
20 gpm of water and a gas flow of 10 scfm. The scrubber for the model plant is
made of type 304 stainless steel and is 8 in. in diameter and 10 ft high, with
8 ft of porcelain rings. There is no significant cost for utilities because
the water will also be used as feedwater for the product absorber.
Plant
Size
(Gg)
23
45
113
VOC
Reduction
(Mg/yr)
34.7
67.9
170.5
Installed
Capital
Cost
$38
40
52
Fixed Cost
$11
12
15
Annual Cost
Manpower
$3.6
3.6
3.6
(X 1000}
Recovery
Credit
$5.9
11.7
29.3
E
Net
$ 8.7
3.9
(10.7)
Cost
1 f f ectiveness
(per Mg)
$251
57
(63)
FORMALDEHYDE STORAGE SCRUBBER
All of the scrubber system for formaldehyde storage emissions is made of type
304 stainless steel. The scrubber for the model plant is 8 in. in diameter and
10 ft high, with 8 ft of porcelain rings. There is no significant cost for
utilities.
Plant.
Size
(Gg)
23
45
113
VOC
Reduction
(Mg/yr)
1.2
2.4
5.0
Installe
Capital
Cost
$46
49
61
Fixed Cost
$13
14
18
Annual Cost
Manpower
$3.6
3.6
3.6
(X 1000)
Recovery
Credit
$0.2
0.4
1.1
Net
$16.4
17.2
20.5
Cost
Effectiveness
(per Mg)
$13,667
7,167
3,417
-------
E-l
APPENDIX E
EXISTING PLANT CONSIDERATIONS
Table E-l lists process emission control devices reported to be in use by
industry. To gather information for the preparation of this report, two site
visits were made to manufacturers of formaldehyde. Trip reports have been
cleared by the companies concerned and are on file at EPA, ESED, in Research
Triangle Park, NC.1'2 Some of the pertinent information concerning process
emissions from these existing formaldehyde plants is presented in this appendix.
Pertinent process emission information was also obtained from the Chemical
Manufacturers Association and from a number of formaldehyde producers who
submitted comments in response to the draft of this reporti issued in February
1979.
A. PROCESS EMISSIONS FROM EXISTING PLANTS
1. Celanese, Bishop, TX1
The formaldehyde production facility consists of four metallic-silver-catalyst
process units. The process emissions are controlled by two incinerators. Heat
is recovered from the incinerator flue gases. The following composition is
reported to be typical for the process vent emissions:
Component Amount (%)
Hydrogen 20.57
Nitrogen + air 74.03
Methane 0.02
Methylal 0.19
Methyl formate 0.62
Methanol 0.06
Carbon monoxide 0.64
Carbon dioxide 3.87
Celanese has been averaging 1 or 2 startups a year per unit, and vents the
absorber emissions to the atmosphere during startup.
-------
E-2
2. Borden, Fayetteville, NC2
The formaldehyde production facility consists of three independent silver
catalyzed processes. The emissions from the product absorber consist of nitrogen,
hydrogen, carbon dioxide, carbon monoxide, and oxygen, along with small quantities
of VOC consisting of methyl formate, methylal, and methanol. Absorber vent
emissions are controlled by burning the gas in specially designed steam generators.
The system was developed by Borden and its design is proprietary. The plant
does not operate unless a boiler is in operation. During startup of a formaldehyde
unit the absorber vent gases are emitted to the atmosphere until stable operation
is achieved. These are normally from 4 to 12 startups per year. Startup
venting to the atmosphere lasts from 1 to 2 hr.
The fractionator vacuum system emission is discharged to the atmosphere. No
data were available on the composition or flow from the vent.
3. Monsanto Plastics & Resins Co.3
Monsanto objects to any consideration of the use of a flare during startup
because of the wide change in relative compositions of H2 and 02- They emphasize
the point that emissions are reduced by operating at one-half rate until the
startup procedure is completed.
Monsanto states that with their design of a vent condenser using refrigerated
water at 35°F to condense emissions from the product fractionator, they estimate
VOC emissions to be reduced by 80%. Methanol is thereby recovered in the
finished product, thus eliminating a subsequent waste disposal step.
4. Georgia-Pacific Corporation4
Georgia-Pacific states that new test data indicate that the emissions from the
absorber for their Lufkin metal-oxide-catalyst plant are much greater than they
had estimated for their permit application and much higher than indicated for
the metal oxide model plant in this report. The averages of five absorber
emission test results are reported as follows:
-------
E-3
Methanol Feed (wt %)
Dimethyl ether 1.81
Methanol 0.31
Formaldehyde 0.02
Total VOC 2.14
5. Reichhold Chemicals, Inc.5
Test data for the Reichhold Moncure metal-oxide plant indicate that their
absorber vent emissions are considerably less than the emissions listed for the
metal oxide model plant in this report. The averages of three absorber emission
test results are reported as follows:
Amount (ppm) Formaldehyde Produced (g/kg)
Formaldehyde 171 0.0037
Methanol 3380 0.0782
Dimethyl ether 1847 0.0614
Total VOC 5398 0.1433
B. TOTAL INDUSTRY EMISSIONS
Emissions from industry were estimated based on actual emission rates reported
by the individual plants. When the data reported were incomplete, the emission
rates used for the control measures reported by the plants (see Table E-l) were
based on model-plant data. For those plants not reporting data it was assumed
that control measures similar to those indicated in Table E-l would exist for
other plants operated by the same company. It was estimated that secondary
emissions were uncontrolled for all plants and that maintenance programs for
the control of fugitive emissions averaged half way between controlled and
uncontrolled.
Based on the above, total emissions from all plants during 1977 were approximately
10 Gg. The emissions from these plants would have been approximately 18 Gg
during 1977 if the emissions had been uncontrolled, or 1 Gg if all plants had
been controlled by the measures described for the model plants. It appears,
then, that the current level of control obtained by the industry is approximately
57% of that possible from application of model-plant controls.
-------
E-4
Table E-l. Control Devices Currently Used by
the Domestic Formaldehyde Industry
Company and Location
Allied
Irontown, OH
Bo r den
Fayetteville, NC
Geismar, LA
Springfield, OR
Celanese
Bishop, TX
Newark, NJ
Du Pont
Belle, WV
Healing Spring, NC
Lufkin, TX
LaPorte, TX
GAF
Calvert City, KY
Georgia-Pacific
Columbus, OH
Crossett, AR
Vienna, GA
Gulf
Vicksburg, MS
Hercules
Wilmington, NC
Hooker
North Tonawanda , NY
Monsanto
Addyston, OH
Chocolet Bayou, TX
Reichhold
Houston, TX
Kansas City, KA
Moncure, NC
Tuscaloosa, AL
Tenneco
Fords, NJ
Garfield, NJ
Type of
Process
Silver catalyst
Silver catalyst
Silver catalyst
Silver catalyst
Silver catalyst
Metal oxide
Silver catalyst
Silver catalyst
Metal oxide
Silver catalyst
Metal oxide
Metal oxide
Silver catalyst
Silver catalyst
Metal oxide
Silver catalyst
Silver catalyst
Silver catalyst
Silver catalyst
Metal oxide
Silver catalyst
Metal oxide
Silver catalyst
Metal oxide
Silver catalyst
For Absorber Vent
Thermal oxidizer (100%)
Steam boiler (100%)°
Steam boiler
Steam boiler (100%)
Thermal oxidizer (100%)C
Demister
Thermal oxidizer (99.8%)°
Thermal oxidizer
None
Flare
Demister
Demister
None
Demister
None
None
Steam boiler
Refrigerated
condenser (96.1%)C
None
None
None
None
Demister
Scrubber (94%)°
None
Control Device
For Fractionator Vent
D
NR
NR
D
NR
NA
NR
NR
NA
NR
NA
NA
None
D
NA
None
D
NR
Condenser
NA
Condenser
NA
D
NA
Scrubber
For Storage Tanji.
None
Conservation vent
None
Conservation vent
None
NR
None
Vent condenser
NR
Vent condenser
None
None
None
None
Conservation ven*
None
None
NR
None
Conservation veil'
None
Conservation vent
None
Scrubber
None
For those plants reporting information; see refs. 1—S.
D - distillation column used rather than vacuum fractionation.
NR - not reported.
NA - not applicable to metal-oxide-catalyst process.
Reported efficiency for control device on absorber vent.
-------
E-5
C. RETROFITTING CONTROLS
The primary difficulty associated with retrofitting may be in finding space to
fit the control device into the existing plant layout. Because of the costs
associated with this difficulty it may be appreciably more expensive to retrofit
emission control systems in existing plants than to install a control system
during construction of a new plant.
-------
E-6
C. REFERENCES*
1. J. F. Lawson, IT Enviroscience, Inc., Trip Report for Visit to Celanese
Chemicals Co.', Bishop, TX, July 26, 1977 (on file at EPA, ESED, Research
Triangle Park, NC).
2. J. F. Lawson, IT Enviroscience, Inc., Trip Report for Visit to Borden Chemical,
Inc., Fayetteville, NC, August 24, 1977 (on file at EPA, ESED, Research Triangle
Park, NC).
3. N. B. Galluzzo, Monsanto Resins Co., letter dated July 13, 1979, to R. J. Lovell,
IT Enviroscience.
4. V. J. Tretter, Jr., Georgia-Pacific Corp., letter dated May 30, 1979, to
R. T. Walsh, EPA.
5. P. S. Hewett, Reichhold Chemicals, Inc., letter dated July 21, 1978, to
R. J. Lovell, IT Enviroscience.
*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.
-------
F-l
APPENDIX F
LIST OF EPA INFORMATION SOURCES
W. B. Barton, EPA Questionnaire for Borden Inc., Fayetteville Plant, Aug. 29,
1973.
W. B. Barton, EPA Questionnaire for Borden Inc., Springfield Plant, Feb. 8, 1973.
J. S. Bellecci, Louisiana Air Control Commission Permit Application for Borden
Inc., Geismar Plant, Mar. 14, 1975.
J. W. Blackburn, IT Enviroscience, Inc., Control Device Evaluation. Thermal
Oxidation (July 1980) (EPA/ESED report, Research Triangle Park, NC).
W. R. Chalker, EPA Questionnaire for E. I. du Pont de Nemours & Co., Belle
Plant, Sept. 2, 1977.
J. Cudahy and R. Standifer, IT Enviroscience, Inc., Secondary Emissions (June
1980) (EPA/ESED report. Research Triangle Park, NC).
D. G. Erikson and V. Kalcevic, IT Enviroscience, Inc., Fugitive Emissions
(September 1980) (EPA/ESED report. Research Triangle Park, NC).
J. H. Frick, Texas Air Control Board 1975 Emissions Inventory Questionnaire
for Celanese Chemical Co., Bishop, TX Plant, Mar. 19, 1976.
J. H. Frick, EPA Questionnaire for Celanese Chemical Co., Bishop Plant,
Aug. 15, 1972.
C. R. Gerardy, EPA Questionnaire for Gulf Oil Corp., Vicksburg Plant,
Aug. 21, 1972.
P. S. Hewett, Reichhold Chemicals, Inc., letter to IT Enviroscience, Inc., July 21,
1978.
C. W. Horner, EPA Questionnaire for Reichhold Chemicals, Inc., Houston Plant,
Mar. 19, 1973.
C. W. Horner, EPA Questionnaire for Reichhold Chemicals, Inc., Moncure Plant,
Mar. 19, 1973.
F. Inzerillo, EPA Questionnaire for GAF Corp., Calvert City Plant, Apr. 9, 1973.
S. J. Jelich, EPA Questionnaire for Tenneco Inc., Fords Plant, Sept. 15, 1972.
R. H. Johnson, Texas Air Control Board 1975 Emissions Inventory Questionnaire
for E. l'. du Pont de Nemours & Company, LaPorte Plant, Mar. 19, 1976.
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).
G. D. Milian, EPA Questionnaire for Tenneco Chemicals, Inc., Garfield Plant,
Sept. 6, 1972.
-------
F-2
W. T. Moody, TRW, Inc., letter dated Aug. 15, 1979, to D. A. Beck, EPA.
H. E. Myers, EPA Questionnaire for Allied Chemical Corporation, South Point
Plant, Aug. 10, 1972.
G. Osterman, Ohio Air Pollution Control Permit Application for Monsanto Company,
Addyston Plant, Apr. 10, 1974.
F. T. Parkinson, EPA Questionnaire for Hercules Incorporated, Hanover Plant,
Sept. 7, 1972.
R. 0 Pfaff, Reichhold Chemicals, Inc., Moncore, NC, Emission Testing Report,
EMB Test No. 73-CHO-2, EPA, Research Triangle Park, NC (July 1973).
T. P. Shumaker, EPA Questionnaire for Reichhold Chemicals, Inc., Tuscaloosa
Plant, Sept. 14, 1972.
V. J. Tretter, Jr., EPA Questionnaire for Georgia-Pacific Corp., Columbus Plant,
Sept. 11, 1972.
V. J. Tretter, Jr., EPA Questionnaire for Georgia-Pacific Corporation, Crossett
Plant, Sept. 22, 1972.
V. J. Tretter, Jr., EPA Questionnaire for Georgia-Pacific Corporation, Vienna
Plant, Sept. 11, 1972.
V. J. Tretter, Jr., Georgia-Pacific Corporation, letter to EPA, July 19, 1978.
H. M. Walker, EPA Questionnaire for Monsanto Company, Chocolate Bayou Plant,
Aug. 9, 1972.
C. A. Williams, EPA Questionnaire for Reichhold Chemicals, Inc., Kansas City
Plant, Aug. 7, 1972.
-------
2-i
REPORT 2
Methanol
J. A. Key
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
November 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.
D43P
-------
2-iii
CONTENTS OF REPORT 2
I- ABBREVIATIONS AND CONVERSION FACTORS 1-1
II. INDUSTRY DESCRIPTION II-l
A. Reason for Selection II-l
B. Usage and Growth II-l
C. Domestic Producers II-3
D. References II-7
III. PROCESS DESCRIPTION III-l
A. Introduction III-l
B. Low-Pressure Process III-l
C. Process Variations III-5
D. References III-7
IV. EMISSIONS Iv-i
A. Low-Pressure Process IV-1
B. Process Variations IV-8
C. References IV-9
V. APPLICABLE CONTROL SYSTEMS V-l
A. Low-Pressure Process V-l
B. Process Variations V-4
C. References V-5
VI. IMPACT ANALYSIS VI-1
A. Environmental and Energy Impacts VI-1
B. Control Cost Impact VI-3
C. References VI-4
VII. SUMMARY VII-1
-------
2-v
APPENDICES OF REPORT 2
Page
A. PHYSICAL PROPERTIES OF METHANOL, DIMETHYL ETHER, AND A-l
METHYL FORMATE
B. AIR-DISPERSION PARAMETERS FOR MODEL PLANT B-l
C. FUGITIVE-EMISSION FACTORS C-l
D. EXISTING PLANT CONSIDERATIONS D-l
TABLES OF REPORT 2
Number Page
II-l Methanol Usage and Growth II-2
II-2 Methanol Capacity II-4
IV-1 Total Uncontrolled VOC Emissions for Methanol Model Plant IV-3
IV-2 Estimated Composition of Purge Gas from Model Plant IV-4
IV-3 Composition of Distillation Vent Gas from Model Plant IV-5
IV-4 Storage Tank Data for Methanol Model Plant IV-6
V-l VOC Controlled Emissions for Methanol Model Plant V-2
VI-1 Environmental Impact of Controlled Methanol Plant VI-2
VII-1 Emission Summary for Methanol Model Plant VII-2
A-l Physical Properties of Methanol A-l
A-2 Physical Properties of Dimethyl Ether A-2
A-3 Physical Properties of Methyl Formate A-3
B-l Air-Dispersion Parameters for Model Plant B-l
D-l Emission Control Devices or Techniques Currently Used D-3
FIGURES OF REPORT 2
II-l Locations of Plants Manufacturing Methanol II-5
III-l Flow Diagram for Uncontrolled Model Plant for Production of III-4
Methanol
-------
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 Kg)
British thermal unit (Btu)
Degree Fahrenheit (°F)
Feet (ft)
Cubic feet (ft3)
Barrel (oil) (bbl)
Gallon (U.S. liquid) (gal)
Gallon (U.S. liquid)/min
(gpm)
Horsepower (electric) (hp)
Inch (in.)
Pound-force/inch2 (psi)
Pound-mass (Ib)
Watt-hour (Wh)
Multiply By
9.870 X 10"6
9.480 X 10"4
(°C X 9/5) + 32
3.28
3.531 X 101
6.290
2.643 X 102
1.585 X 104
1.340 X 10"3
3.937 X 101
1.450 X 10"4
2.205
2.778 X 10"4
Standard Conditions
68°F = 20°C
1 atmosphere = 101,325 Pascals
PREFIXES
Prefix
T
G
M
k
m
M
Symbol
tera
giga
mega
kilo
milli
micro
Multiplication
Factor
1012
IO9
106
103
io"3
io"6
Example
1 Tg =
1 Gg =
1 Mg =
1 km =
1 raV =
1 Mg =
1 X 10 lz grams
1 X IO9 grams
1 X IO6 grams
1 X IO3 meters
1 X IO"3 volt
1 X IO"6 gram
-------
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
-------
CONTENTS
INTRODUCTION Vll
Product Report Page
1. FORMALDEHYDE 1-i
2. METHANOL 2-i
3. ETHYLENE 3-i
4. ETHYLENE OXIDE 4-i
5. VINYL ACETATE 5-i
6. ACETALDEHYDE 6-i
7. ETHANOLAMINES 7-i
8. ETHYLENE GLYCOL 8-i
9. GLYCOL ETHERS 9-i
-------
VIX
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: formaldehyde, methanol, ethylene, ethylene oxide, vinyl acetate,
acetaldehyde, ethanolamines, ethylene glycol, and glycol ethers. 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.
-------
II-l
II. INDUSTRY DESCRIPTION
A. REASON FOR SELECTION
Methanol was selected for consideration because it is produced in large amounts
and has a moderate volatility (see Appendix A for pertinent physical properties),
both of which contribute to potentially high emissions of volatile organic com-
pounds (VOC).
B. USAGE AND GROWTH
The end uses of methanol and the expected growth rate for each use are given in
Table II-l.* Formaldehyde production is the largest consumer (~40%) of methanol
and is expected to continue to be through 1981, when it will still account for
37% to 38% of the domestic consumption. The fastest growing use of methanol is
in the production of acetic acid, which by 1981 will account for 9 to 10% of
the domestic consumption and will make it the second-largest consumer of methanol.
Methanol is also used as a solvent, to a small extent as a fuel, and in the
production of numerous chemicals; only the largest consumers are shown in Table
II-l. Fuel use has been as an injection fluid in certain aircraft engines and
as automotive fuel in racing cars and boats. A new fuel use is in the production
of methyl tertiary butyl ether (MTBE), approved by the EPA in 1979 as a high-
octane component in gasoline. By 1983 this use could equal or exceed the use
1 2
of methanol to produce acetic acid. '
The domestic methanol production capacity for 1980 is reported to be about
2
4,310,000 Mg/yr and does not include a plant that has been mothballed by Du
Pont or a Valley Nitrogen plant that has been on stand-by since 1977. Two new
producers (Arco Chemical and Getty Refining and Marketing) are planning new
methanol capacities by 1983 in addition to expansions by existing producers,
and so there should be sufficient capacity to supply 1983 domestic demand if it
grows at the projected rate of 7% per year. Some methanol will be imported to
2
help meet the demand in times of shortages.
*In order to minimize the revision time, the 1976 data that were used for the
original draft of this report are retained. For our purposes the differences
are not believed to be significant.
-------
II-2
Table II-l. Methanol Usage and Growth*
End Use
Formaldehyde
Solvents
Chlorome thanes
Acetic acid
Methylamines
Methyl methacrylate
Dimethyl terephthalate
Glycol methyl ethers
Inhibitor for formaldehyde
Miscellaneous and fuel uses
Consumption
For 1976
(%)
42
8
8
5
5
5
4
1
1
21
Average Growth
For 1976—1981
(%/yr)
5
7
7
24
8
7
6
4
3
8
*See ref 1.
-------
II-3
C. DOMESTIC PRODUCERS
There were 8 producers operating 10 methanol plants in the United States in
2
1980. Table II-2 lists the producers, locations, and capacities; Fig. II-l
shows the plant locations.
Producing Companies
1. Air Products
About 45% of the methanol capacity is required to operate the methylamines faci-
lities at capacity.
2. Borden
Acetic acid and formaldehyde are produced from methanol at Geismar, LA, and
formaldehyde is produced at several other locations. The total estimated require-
ment for methanol is about 70% of capacity. Borden has announced that the
methanol plant will be modernized and expanded in 1980, with the existing high-
2
pressure process being replaced by ICI low-pressure synthesis technology.
3. Celanese
Formaldehyde production and acetic acid production require about 45% of the
total methanol capacity of two plants.
4. Du Pont
Formaldehyde, dimethyl terephthalate, methyl amines, and methyl methacrylate
are produced at several locations and consume about 75% of the methanol capacity
of the Beaumont, TX, plant. A new 600,000-Mg/yr methanol plant has been started
1 2
up at Deer Park, TX, and a 350,000-Mg/yr plant at Orange, TX, has been moth-
balled.3
5. Georgia-Pacific
Formaldehyde is produced from methanol at several locations and at capacity
requires about half of the methanol produced by their methanol plant when it is
run at capacity.
-------
II-4
Table II-2. Methanol Capacity
Plant
Air Products, Pensacola, FL
Allemania, Plaquemine, LA
Borden, Geismar, LA
Celanese, Bishop, TX
Celanese, Clear Lake, TX
Du Pont, Beaumont , TX
Du Pont, Deer Park, TX
Georgia-Pacific, Plaquemine, LA
Monsanto, Texas City, TX
Tenneco, Houston, TX
Total
Capacity as of 1980
(Mg/yr)
150,000
300,000
540,000
450,000
690,000
680,000
600,000
360,000
300,000
240,000
4,310,000
Process
Used
High pressure
High pressure
Low pressure
Low pressure
Low pressure
High pressure
Low pressure
Low pressure
Low pressure
High pressure
See ref 2.
Allemania will rebuild in 1981 the existing plant (formerly Hecofina) to
incorporate Lurgi low-pressure process with expanded capacity of 360,000
Mg/yr.
Q
Borden is to replace its existing high-pressure process with ICI low-pressure
technology by the end of 1980.
d
Tenneco will convert its plant in 1981 to Lurgi low-pressure process with a
capacity of 390,000 Mg/yr.
-------
II-5
1. Air Products, Pensacola, FL
2. Allemania, Plaquemine, LA
3. Borden, Geismar, LA
4. Celanese, Bishop, TX
5. Celanese, Clear Lake, TX
6. Du Pont, Beaumont, TX
7. Du Pont, Deer Park, TX
8. Georgia-Pacific, Plaquemine, LA
9. Monsanto, Texas City, TX
10. Tenneco, Houston, TX
Fig. II-l. Locations of Plants Manufacturing Methanol
-------
II-6
6. Allemania
The company will rebuild the existing plant (purchased from Hercofine in 1979)
2
to incorporate Lurgi low-pressure technology.
7. Monsanto
Formaldehyde is produced from raethanol at several locations, and acetic acid is
produced from methanol in a facility located in the same complex as the methanol
facility. The combined methanol requirements at capacity total about 80% of
the methanol production capacity.
8. Tenneco
Operation at capacity of two formaldehyde plants in New Jersey requires only
20% of the methanol capacity; the remainder is sold.
-------
II-7
D. REFERENCES
1. J. L. Blackford, "Methanol", pp. 674.5021AF and 674.5022A—674.5024L in Chemical
Economics Handbook, Stanford Research Institute, Menlo Park, CA (August, 1977).
2. A. D. Abshire et a_l. , "Methanol," pp 674.5021A—674.50211 and 674.5022A—
674.5026A in Chemical Economics Handbook, Stanford Research Institute, Menlo
Park, CA (June 1980).
3. D. W. Smith, E. I. du Pont de Nemours and Company, letter dated May 25, 1978,
to EPA with information on air emissions from the methanol plant at Beaumont, TX,
in response to EPA request.
4- "Manual of Current Indicators—Supplemental Data," p. 242 in Chemical Economics
Handbook, Stanford Research Institute, Menlo Park, CA (October 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.
-------
III-l
III. PROCESS DESCRIPTION
A. INTRODUCTION
Almost all methanol produced in the United States is made from natural gas.
The natural gas is steam reformed to produce synthesis gas, consisting of a
mixture of carbon monoxide, carbon dioxide, and hydrogen, which is then converted
into methanol. The large new methanol plant started up in 1980 by Du Pont at
Deer Park, TX, is supplied with synthesis gas produced by partial oxidation of
residual fuel oil. The off-gases from the partial oxidation of natural gas for
acetylene manufacture contain the right proportions of carbon monoxide, carbon
dioxide, and hydrogen for methanol synthesis and have been used for the production
of methanol. Other potential sources of synthesis gas are coal gasification,
pyrolysis of garbage, timber wastes, agricultural wastes, or municipal solid
wastes, and even a steel plant's basic oxygen furnace off-gases. None of these
sources are economically feasible at this time. —
The synthesis gas is converted to methanol by either a high-pressure (28 to
45 MPa) process or by one of the more recently developed low-pressure (5 to
10 MPa) processes. All new methanol capacity is based on low-pressure technology.
Although processes have been developed that employ medium pressures of 15 to
18 MPa, none are used in the United States. —
A small smount of methanol is obtained as a by-product from the oxidation of
butane to produce acetic acid and from destructive distillation of wood to produce
charcoal. When wood is carbonized by prolonged heating, condensable and nonconden-
sable volatiles are given off. The condensable portion - called pyroligneous
acid - contains methanol, acetic acid, and tars. Natural methanol and acetic acid
can be recovered by refining the pyroligeneous acid after the tars have been removed.
This process was discontinued in the United States during the early 1970s.
Methanol is regenerated in the production of polyester from dimethyl terephthalate
and is usually recycled to produce additional dimethyl terephthalate.
B. LOW-PRESSURE PROCESS
The steam reforming of natural gas (methane) to produce synthesis gas takes
place according to the following reactions:
-------
Ill-2
CH + HO *• 3 H + CO
(methane) (steam) (hydrogen) (carbon monoxide)
CH^ + 2H20 > 4H2 + CO
(methane) (steam) (hydrogen) (carbon dioxide)
The tubular reformer operates at 800 to 850°C and 1.7 to 2.1 MPa and is heated
with fuel gas as both reactions are endotherraic. A promoted nickel-based catalyst
is used and the steam to methane ratio is controlled to give a synthesis gas
that contains, in addition to the hydrogen and carbon oxides, only small amounts
of unreacted methane, plus the nitrogen and argon that were fed with the natural
3 6
gas. —
The hydrogen and carbon oxides in the synthesis gas are converted to methanol
by the exothermic reactions-.
2H2 + CO > CH OH
(hydrogen) (carbon monoxide) (metnanol)
3H + CO > CH OH + HO
(hydrogen) (carbon dioxide) (methanol) (water)
The reaction conditions are a temperature of 200 to 300°C and a pressure of 5
to 10 MPa. A very active copper-based synthesis catalyst is used that is easily
poisoned by sulfur compounds. The synthesis gas feed preferably contains less
than 1.0 ppm of sulfur, and if possible less than 0.1 ppm. When produced by
steam reforming of natural gas the synthesis gas contains an excess of hydrogen
over the stoichiometric amount needed; so carbon dioxide may be added to the
reformer feed or to the synthesis gas to provide the proper proportions. ' —
Efficient waste heat recovery from the reformer flue gases, from the synthesis
gas leaving the reformer, and from the product gases leaving the converter and
the use of purge gases and waste liquids as either fuel or as feeds to other
processes are very important factors affecting operating costs. Optimization
by a systems approach to the design of the operating parameters of all the sections
-------
III-3
of the methanol plant reforming, compression, synthesis, distillation, and
2
heat recovery reduces energy requirements. The overall plant may be self-
9
supporting with regard to steam during normal operation.
Figure III-l is a typical flow diagram for a low-pressure methanol process.
Natural gas (Stream 1) is desulfurized, generally by adsorption on activated
carbon, and then fed with steam and carbon dioxide (Streams 2 and 3) to a tubular
reformer fired with fuel gas. Heat is recovered from the synthesis gas leaving
the reformer (Stream 4) by producing steam in a heat recovery system. The
cooled synthesis gas (Stream 5) is compressed by the makeup gas compressor and
added to recycled synthesis gas (Stream 6) in the synthesis loop. The combined
synthesis gas is preheated to the reaction temperature by heat exchange with
the product gas (Stream 7) leaving the converter,- the preheated gas (Stream 8) then
enters the converter. A portion of the cold gas (Stream 9) from the synthesis
loop is injected into the converter at several locations to control the reaction
temperature. The product gas leaving the converter (Stream 7) is cooled by
heat exchange with the synthesis gas (Stream 6) and then further cooled by the
heat recovery and condensing system to condense methanol. The unreacted synthesis
gas and condensed methanol (Stream 10) go to the separator, where the crude
methanol (Stream 11) is removed from the unreacted gas and sent to the flash
tank. The synthesis gas from the separator is compressed by the recirculating
compressor for recycle after a portion has been purged (Stream 12) to remove
inert gases (Vent A) from the system. The pressure on the crude methanol is
reduced to near-atmospheric in the flash tank, where dissolved gases flash off
(Stream 13) and leave with the purge gas (Vent A). The degassed crude methanol
from the flash tank (Stream 14) goes to the crude storage tank. From storage
it is fed to the heads column, where the low-boiling impurities (mostly dimethyl
ether) are separated overhead (Stream 15) and sent either to another process or
to a boiler as fuel. The bottoms from the heads column (Stream 16) go to the
tails column, where purified methanol (Stream 17) is separated overhead and
sent to the check tanks where it is held until checked for meeting specifications.
High-boiling impurities (higher alcohols) are removed from the tails column as
a sidestream (Stream 18) and sent to a boiler for use as fuel. The bottoms
(Stream 19) from the tails column is the water separated from the crude methanol,
and, since it contains only 100 to 300 ppm organics (mostly methanol), it is
3 11
sent to a wastewater treatment system. —
-------
Fig. III-l. Flow Diagram for .Uncontrolled Model Plant for Production of Methanol
-------
III-5
The purge gas (Vent A) and the distillation section (Vent B) are sources of
process emissions,- in the sulfur removal section intermittent emissions can
result from the regeneration of the activated carbon with steam (Vent C). The
purge gas is usually burned, often in the reformer for its fuel value.—
Storage emission sources (Vents D through F) include crude storage, check tanks,
and product storage. Handling emission sources (G and H) are the loading of
methanol into railroad tank cars and into barges for shipment.
Fugitive emissions (I) occur when leaks develop in valves or in pump or compressor
seals. When process pressures are higher than the cooling-water pressure, VOC
can leak into the cooling water and escape as a fugitive emission from the cooling
tower.
Secondary emissions can occur when wastewater containing VOC is sent to a wastewater
treatment system or lagoon and the VOC are evaporated (J). Another source of
secondary emissions is the combustion of liquid and gas waste streams in a boiler
or process heater as fuel, where VOC are emitted with the flue gases (Vent K). J—ll
C. PROCESS VARIATIONS1'3—11
In the model plant* carbon dioxide is added to provide the correct stoichiometric
properties of carbon to hydrogen. One process variation involves the use of a
larger purge gas flow to remove the excess hydrogen from the synthesis loop.
Another process variation is the use of a shell and tube converter with boiling
water in the shell to control the reaction temperature and produce steam. Other
variations result from use of a different operating temperature of the converter
and a different arrangement of the purification section. The high-pressure processes
(28 to 45 MPa) were developed before the low-pressure processes were, and produce a
crude methanol with more impurities. The high pressures are necessary to obtain
commercially adequate reaction rates because the catalysts used are less active
but more resistant to sulfur poisoning than those used in the low-pressure processes.
The amount of water in the crude methanol is almost directly proportional to
the amount of carbon dioxide converted to methanol. The number of distillation
columns depends on the methanol specification that is to be met and can vary
from one to three with several different arrangements and operating pressures.
*See p 1-2 for a discussion of model plants.
-------
III-6
When partial oxidation of residual fuel oil is used to produce the synthesis
gas, the pressure of the synthesis gas is high enough for a makeup gas compressor
not to be required.
-------
III-7
REFERENCES*
J. L. Blackford, "Methanol," pp. 674.5021A—F and 674.5022A—674.5024L in
Chemical Economics Handbook, Stanford Research Institute, Menlo Park, CA
(August 1977)
B. Hedley, W. Powers, and R. B. Stobaugh, "Petrochemical Guide 15: Economics.
Methanol: How, Where, Who Future," Hydrocarbon Processing 49(9), 275—280
(1970). —
L. F. Hatch and S. Matar, "From Hydrocarbons to Petrochemicals...Part 6 Petro-
chemicals from Methane," Hydrocarbon Processing 56(10), 153—163 (1977).
G. E. Haddeland, Synthetic Methanol, Report No. 43, A private report by the
Process Economics Program, Stanford Research Institute, Menlo Park, CA
(October 1968).
G. E. Haddeland, Methanol Interim Report. Report No. 43A1, A private report by
the Process Economics Program, Stanford Research Institute, Menlo Park, CA
(July 1972)
H. F. Woodward, Jr., "Methanol," pp. 370—398 in Kirk-Othmer Encyclopedia of Chemical
Technology, Vol. 13, 2d ed.,edited by A. Standen et al., Wiley, New York, 1967.
"Methanol (ICI Low Pressure Process)," Hydrocarbon Processing 56(11), 182 (1977).
"Methanol (Lurgi Low Pressure Process)," Hydrocarbon Processing 5J>(11), 183
(1977). —
H. Miller and F. Marschner, "Lurgi Makes Low-Pressure Methanol," Hydrocarbon
Processing 49(9), 281—285 (1970).
B. Hedley, W. Powers, and R. B. Stobaugh, "Petrochemical Guide 15: Manufacture.
Methanol: How, Where, Who Future," Hydrocarbon Processing 49(6), 97—101 (1970).
D. D. Mehta and W. W. Pan, "Purify Methanol This Way," Hydrocarbon Processing
50(2), 115—120 (1971).
*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, partici-
pate 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. LOW-PRESSURE PROCESS
1. Model Plant*
The model plant for the low-pressure synthesis of methanol from natural gas
(Fig. III-l) has a capacity of 450,000 Mg/yr, based on 8760 hr** of operation
annually. The process and capacity are typical of those of recently built methanol
plants and of one plant started up in 1980. Characteristics of the model plant
important to air-dispersion modeling are shown in Table B-l in Appendix B.
2. Sources and Emissions
Emissions sources and rates for the low-pressure process are summarized in Table IV-1.
The process emissions estimated for the methanol model plant are based on information
12 3
given in reports of visits to Borden, Celanese, and Monsanto and in responses
to EPA's requests for information from selected companies, together with data
A
from a report published by Stanford Research Institute, and on an understanding
of the process chemistry and yields. The storage and handling emissions were
calculated based on physical properties. The fugitive emissions are based on
the petroleum refinery data referenced in Appendix C.
a. Purge Gas Vent The purge gas vent (Vent A, Fig. III-l) is the largest process
emission source. The vent gas contains the unreacted hydrogen, carbon monoxide.
*See p 1-2 for a discussion of model plants.
**Process downtime is normally expected to range from 5 to 15%. If the hourly
rate remains constant, the annual production and annual VOC emissions will be
correspondingly reduced. Control devices will usually operate on the same
cycle as the process. From the standpoint of cost-effectiveness calculations,
the error introduced by assuming continuous operation is negligible.
-------
IV-2
and carbon dioxide, the inert nitrogen and methane from the makeup synthesis
gas, and small amounts of the uncondensed methanol and water vapor remaining
after the crude methanol is condensed and separated. The composition of this
stream depends on the makeup synthesis gas composition, the conversion catalyst,
and the temperature and pressure in the converter. Flow and composition data
on this stream were not available; the estimated composition of the vent gas
from the model plant given in Table IV-2 is based on a material balance by
Haddeland for a process employing carbon dioxide addition to the natural gas
feed to the reformer. It was calculated by use of theoretical relationships
for a methane conversion of about 90%, and does not represent the emissions
from any specific plant or process. The estimate of the VOC emission rate
4
given in Table IV-1 is based on the same data.
b. Distillation Vent The vent gases from the heads column (Vent B, Fig. III-l)
are the noncondensables that are dissolved in the crude methanol fed to the
column and the VOC that are not condensed, i.e., methanol, dimethyl ether, and
methyl formate. The composition of the distillation vent gas from the model
plant, shown in Table IV-3, is based on the reported composition from the distilla-
tion area vent of a methanol plant with a low-pressure synthesis process designed
by ICI. The estimate of the VOC emission rate given in Table IV-1 is based on
4
the material balance by Haddeland, the composition shown in Table IV-3, and on
an estimate that 5% of the dimethyl ether is not condensed.
c. Sulfur Removal Vent The natural gas feed to the model plant (Stream 1, Fig. III-l)
normally contains no VOC and small amounts of sulfur compounds; as a result the
intermittent emission (Vent C) during regeneration of the sulfur removal section
2 5
contains no VOC, but only hydrogen sulfide, methane, and steam. ' This vent
can be a source of VOC emissions when the natural gas contains VOC that are
adsorbed in
generation.
adsorbed in the sulfur removal section and then desorbed during its steam re-
6,7
d. Storage and Handling Emissions Emissions result from the storage of crude and
purified methanol. Sources for the model plant are shown in Fig. III-l (sources
D through F). Storage tank parameters for the model plant are given in Table
IV-4. The uncontrolled storage emissions in Table IV-1 were calculated and are
based on fixed-roof tanks, half full, with an 11°C diurnal temperature variation.
-------
IV-3
Table IV-1.
Total Uncontrolled VOC Emissions for
Methanol Model Planta
Source
Purge gas vent
Distillation vent
Sulfur removal
Storage vents
Crude methanol
Check tanks
Product
Handling
Loading tank cars
Loading barges
Fugitive
Secondary
Wastewater treatment
Waste organic as fuel
Total
Vent
Designation
(Fig.III-1)
A
B
C
D
E
F
G
H
I
J
K
Ratio
(g /kg) b
1.1
0.4
c
0.0097
0.080
0.365
0.0486
0.0971
0.578
0.00044
0.000060
2.68
Emissions
Rate
(kg/hr)
56.5
20.5
c
0.50
4.1
18.8
2.49
4.99
29.7
0.023
0.0031
138
a . . .
control devices other than those necessary for economical operation.
g of emissions per kg of methanol produced.
CModel-plant emissions during the sulfur removal unit regeneration contain
no VOC.
Tanks for holding product until it is checked.
-------
IV-4
Table IV-2. Estimated Composition of Purge Gas from Model Plant'
Component
Methanol (VOC)
Methane
Hydrogen
Carbon monoxide
Nitrogen
Carbon dioxide
Water
Total
Composition
(wt %)
0.8
64.8
15.1
11.1
4.9
3.1
0.2
100
Emission Ratio
(9/kg)b
1.1
88.1
20.6
15.1
6.7
4.2
0.2
136
a
See ref 4.
g of emission per kg of methanol produced.
-------
IV-5
Table IV-3. Composition of Distillation Vent Gas
from Model Plant*
Component Composition (wt %)
Methanol 29.0
Dimethyl ether and methyl formate 12.1
Total VOC 41.1
Carbon dioxide 58.6
Water 0.3
Total 100
*See ref3.
-------
IV-6
Table IV-4. Storage Tank Data for Methanol Model Plant
Parameter
Contents
Number of tanks
Tank size (m )
Turnovers per year
Bulk temperature (°C)
Crude
Crude methanol
1
1890
6b
27
Tank
Check
Methanol
2
810
350
27
Product
Methanol
2
19,500
15
27
Tanks for holding product until it is checked.
This tank operates at approximately constant level, and the number of turnovers
indicated is an attempt to account for slight level variations.
-------
IV-7
Emission equations from AP-42 were used with one modification. The breathing
losses were divided by 4 to account for recent evidence indicating that the
8 9
AP-42 breathing-loss equation overpredicts emissions. '
Handling emissions result from the loading of methanol into railroad tank cars
and into barges for shipment (sources G and H). These emissions are shown in
Table IV-1 and were calculated with the equations from AP-42, based on submerged
loading of methanol at 27°C and on one third of the production being shipped in
tank cars and two thirds in barges.
e. Fugitive Emissions Process pumps, compressors, valves, and pressure-relief
devices are potential sources of fugitive emissions (source I). The model plant
is estimated to have 30 pumps, 1 compressor (with 2 seals), 1400 process valves,
1 ——3
and 40 pressure relief devices handling VOC. Half of the process valves
and pressure-relief devices are in gas/vapor service. Pumps, compressors, valves,
and pressure-relief valves not handling VOC (this includes the makeup gas compres-
sor) are not included in these estimates. The fugitive emission factors from
Appendix C were applied to these estimates, and the results are shown in Table IV-1
as fugitive emissions.
f. Secondary Emissions Secondary VOC emissions can result from the handling and
disposal of process waste streams. For the model plant two potential sources
are indicated on the flow diagram (sources J and K, Fig. III-l).
The secondary emissions from wastewater treatment (source J) were estimated by
. , 11
procedures that are discussed in a separate EPA report on secondary emissions.
The wastewater composition and flow rate were estimated based on data received
from methanol producers. ' A Henry's-law constant was then calculated for the
vapor-liquid system under consideration, and the emission rate was estimated by
comparison with informal
is shown in Table IV-1.
comparison with information given in existing literature. This emission rate
The boiler flue gas secondary VOC emissions, originating from waste organics
used as fuel, were calculated with the emission factors from AP-42 for distillate
oil combustion.12 The basis for estimating the amount of high-boiling impurities
and low-boiling impurities burned as fuel in a boiler for the model plant was a
-------
IV-8
crude methanol composition from a low-pressure process given by Killer and
Marschner. If the low-boiling impurities are sent to another process rather
than burned as fuel as in the model plant, this estimated amount would be lower
and therefore the calculated emissions would be lower. The secondary emissions
from burning waste organics as fuel (Vent K, Table IV-1) were based on burning
both low- and high-boiling impurities.
B. PROCESS VARIATIONS
It is reported that a high hydrogen—to—carbon monoxide ratio, which will result
when natural gas is steam reformed without carbon dioxide addition, suppresses
the undesirable side reactions but requires an increased purge gas flow to remove
the excess hydrogen. '14 Data were not available on the effect, if any, that
this has on the VOC emissions.
-------
IV-9
C. REFERENCES*
1. J. A. Key, IT Enviroscience, Inc., Trip Report for Visit to Borden, Geismar, LA,
Mar. 3, 1978 (on file at EPA, ESED, Research Triangle Park, NC).
2. J. A. Key, IT Enviroscience, Inc., Trip Report for Visit to Celanese, Bishop, TX,
Oct. 11, 1977 (on file at EPA, ESED, Research Triangle Park, NC).
3. J. A. Key, IT Enviroscience, Inc., Trip Report for Visit to Monsanto, Texas City.
TX, Dec. 13, 1977 (on file at EPA, ESED, Research Triangle Park, NC).
4. G. E. Haddeland, Methanol Interim Report, Report No. 43A1, A private report by
the Process Economics Program, Stanford Research Institute, Menlo Park, CA
(July 1972).
5. R. L. Duggan, Air Products and Chemicals Inc., letter dated May 11, 1978, to
EPA with information on air emissions from the methanol plant at Pensacola, FL,
in response to EPA request.
6. D. W. Smith, E. I. du Pont de Nemours and Company, letter dated May 25, 1978,
to EPA with information on air emissions from the methanol plant at Beaumont, TX,
in response to EPA request.
7. W. P. Anderson, Tenneco Chemicals, letter dated May 10, 1978, to EPA with informa-
tion on air emissions from the methanol plant at Pasadena, TX, in response to
EPA request.
8. C. C. Masser, "Storage of Petroleum Liquids," pp. 4.3-1—4.3-16 in Compilation of
Air Pollutant Emission Factors, 3d ed., Part A, AP-42, EPA, Research Triangle
Park, NC (April 1977).
9. E. C. Pulaski, TRW, Inc., letter dated May 30, 1979 to Richard Burr, EPA.
10. C. C. Masser, "Transportation and Marketing of Petroleum Liquids," o£. cit.,
pp. 4.4-1—4.4-6.
11. J. J. Cudahy and R. L. Standifer, IT Enviroscience, Inc., Secondary Emissions
(June 1980) (EPQ/ESED report, Research Triangle Park, NC).
12. T. Lahre, "Fuel Oil Combustion," pp. 1.3-1—1.3-5 in Compilation of Air Pollutant
Emission Factors, 3d ed., Part A, AP-42, EPA, Research Triangle Park, NC (April
1977).
13. H. Hiller and F. Marschner, "Lurgi Makes Low-Pressure Methanol," Hydrocarbon
Processing 49(9), 281—285 (1970).
14. D. D. Mehta and W. W. Pan, "Purify Methanol This Way," Hydrocarbon Processing
50(2), 115—120 (1971).
*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. LOW-PRESSURE PROCESS
1. Purge Gas Vent
Although the stream from the purge-gas vent is the largest uncontrolled process
emission source (Vent A, Fig. III-l) in the model plant, it is normally controlled
by being burned as fuel gas or by being transferred to another process to utilize
1 7
the methane, hydrogen, or carbon monoxide in it. — With the increasing cost
of energy, this is done primarily for economic reasons. The control option
selected for the purge gas in the model plant is the use of it as fuel gas in
the reformer, which is economically justified by recovery of its value as fuel.
The controlled VOC emissions in the reformer flue gas that originated in the
purge gas (see Table V-l) were calculated by applying the emission factors from
AP-42 for natural gas burned in ar
VOC emission reduction of 98.2%.*
o
AP-42 for natural gas burned in an industrial process boiler, and results in a
A flare is used by some plants to control the purge gas when for some reason it
cannot be used as fuel. ' The flare for a methanol plant would normally be
designed for process emergency venting conditions. When the purge gas is burned
9
in such a flare, the VOC destruction efficiency can be lower than 98%**.
2. Distillation Vent
The control option selected for the model-plant distillation vent is the flare
used to safely dispose of emergency releases in a methanol plant. ,A VOC reduction
of 98%** was used to calculate the controlled emissions from the flare that origi-
nated in the distillation vent, based on the estimate that the vent gases from
the distillation vent are greater than 10% of the maximum smokeless design flow
for the flare.9
*The destruction of the VOC entering the reformer is greater than 99.98%,
but after the VOC produced during combustion of the methane, hydrogen, etc.
in the purge gas (which are not considered to be VOC) is taken into account,
the net destruction efficiency is only 98.2%.
**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 Methanol Model Plant
Source
Purge gas vent
Distillation vent
Sulfur removal
Storage vents
Crude methanol
Check tanks
Product
Handling
Loading tank cars
Loading barges
Fugitive
Secondary
Wastewater treatment
Waste organic as fuel
Total
Vent
Des ignat ion
(Fig.III-1)
A
B
C
D
E
F
G
H
I
J
K
Control Device
or Technique
Used as fuel
Flare
None
Internal-floating-roof
tanks
Internal-floating-roof
tanks
Internal-floating-roof
tanks
Aqueous scrubber
Aqueous scrubber
Detection and correction
of major leaks
None
None
Total VOC
Emission
Reduction
98.2
98
85
85
85
99
99
80
VOC
Ratio
(g/kg)a
0.019
0.008
b
0.0014
0.012
0.055
Emissions
Rate
(kg/hr)
0.99
0.41
b
0.074
0.62
2.8
0.000486 0.0249
0.000971
0.116
0.00044
0.000060
0.213
0.0499
5.9
0.02?
0.0031
10.9
ag of emission per kg of methanol produced.
Model-plant emissions during the sulfur removal unit regeneration contain no VOC.
Tanks for holding product until it is checked.
-------
V-3
Another option used for control of the VOC in the distillation vent gases is to
send them to the fuel gas system , either with or without a compressor, depending
on the pressures involved. Usually this approach is justified by the economics
of recovering the fuel value of the vent gases. The VOC reduction efficiency
may be slightly better than that of a flare, but the difference in the controlled
VOC emissions is minor.
3. Sulfur Removal Vent
No control option has been identified for this vent (Vent C) in the model plant
because no VOC are emitted. Plants that do emit VOC during the regeneration of
their sulfur removal system do not report any control devices, likely because
the small amount of VOC and the sulfur compounds present make this a difficult
147
control problem. ' '
4. Storage and Handling Emissions
The emissions from the model-plant storage tanks are controlled by use of internal-
floating-roof tanks.* Options for control of storage and handling emissions
are covered in another EPA report.
The VOC emissions from loading tank cars and barges are controlled by aqueous
scrubbers in the model plant. Aqueous scrubbers are used in similar applica-
tions to control emissions from methanol storage tanks, with the scrubber effluent
sent to the crude methanol storage tank so that the methanol scrubbed from the
2
vent gases is recovered.
The controlled storage emissions given in Table V-l were calculated on the assump-
tion that a contact type of internal floating roof with secondary seals will
reduce fixed-roof-tank emissions by 85%.11/12 A VOC (methanol) removal efficiency
of 99%, was used to calculate the controlled emissions from loading tank cars
and barges for the model plant (see Table V-l). Calculation of removal efficiencies
for once-through absorbers is discussed in Control Device Evaluation. Gas
13
Absorption.
*Consist of internal floating covers or covered floating roofs as defined in
API 25-19, 2d ed., 1976 (fixed-roof tanks with internal floating device to
reduce vapor loss).
-------
V-4
5. Fugitive Emissions
Controls for fugitive emissions from the synthetic organic chemicals manufac-
14
turing industry will be discussed in another EPA report. Emissions from pumps
and valves can be controlled by an appropriate inspection system and repair 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.
6. Secondary Emissions
a. Wastewater Treatment Calculations based on estimated wastewater flow rates
and compositions for the model plant indicate that the VOC emissions from the
wastewater treatment (source J) are relatively small. No control system has
been identified for the model plant. Control of secondary emissions are dis-
cussed in another EPA report.
b. Waste Organic as Fuel Estimates of the VOC emissions in the flue gases of the
model plant reformer that originate in the waste organic streams used as fuel
(source K) indicate they are very small. No control system has been identified
for the model plant.
B. PROCESS VARIATIONS
The applicable controls for the high-pressure processes or for the processes
where carbon dioxide is not added are the same as those for the low-pressure
model plant.
-------
V-5
C. REFERENCES*
1. R. L. Duggan, Air Products and Chemicals Inc., letter dated May 11, 1978, to
EPA with information on air emissions from the methanol plant at Pensacola, FL,
in response to EPA request.
2. J. A. Key, IT Enviroscience, Inc., Trip Report for Visit to Borden, Geismar, LA,
Mar. 3, 1978 (on file at EPA, ESED, Research Triangle Park, NC).
3. J. A. Key, IT Enviroscience, Inc., Trip Report for Visit to Celanese, Bishop, TX,
Oct. 11, 1977 (on file at EPA, ESED, Research Triangle Park, NC).
4. D. W. Smith, E. I. du Pont de Nemours and Company, letter dated May 25, 1978,
to EPA with information on air emissions from the methanol plant at Beaumont, TX,
in response to EPA request.
5. J. A. Key, IT Enviroscience, Inc., Trip Report for Visit to Monsanto, Texas City, TX,
Dec. 13, 1977 (on file at EPA, ESED, Research Triangle Park, NC).
6. D. A. Copeland, Rohm and Haas Texas Incorporated, letter dated May 19, 1978, to
EPA with information on air emissions from the methanol plant at Deer Park, TX,
in response to EPA request.
7. W. P. Anderson, Tenneco Chemicals, letter dated May 10, 1978, to EPA with infor-
mation on air emissions from the methanol plant at Pasadena, TX, in response to
EPA request.
8. T. Lahre, "Natural Gas Combustion," pp. 1.41—1.4-3 in Compilation of Air
Pollutant Emission Factors, 3d ed., Part A, AP-42, EPA, Research Triangle
Park, NC (May 1974).
9. 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).
10. D. G. Erikson, IT Enviroscience, Inc., Storage and Handling (September 1980)
(EPA/ESED report. Research Triangle Park, NC)
11. 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).
12. W. T. Moody, TRW, Inc., letter dated Aug. 15, 1959, to D. A. Beck, EPA.
13. R. L. Standifer, IT Enviroscience, Inc., Control Device Evaluation. Gas
Adsorption (October 1980) (EPA/ESED report, Research Triangle Park, NC).
-------
V-6
14. D. G. Erikson and V. Kalcevic, IT Enviroscience, Inc., Fugitive Emissions
(September 1980) (EPA/ESED report, Research Triangle Park, NC).
15. J. J. Cudahy and R. L. Standifer, IT Enviroscience, Inc., Secondary Emissions
(June 1980) (EPA/ESED report. Research Triangle Park, NC).
*Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of
that paragraph, that reference number is indicated on the material involved.
When the reference appears on a heading, it refers to all the text covered by
that heading.
-------
VI-1
VI. IMPACT ANALYSIS
A. ENVIRONMENTAL AND ENERGY IMPACTS
1. Low Pressure Process
Table VI-1 shows the environmental impact of reducing the total VOC emissions
by application of the described control systems (Sect. V) to the model plant
described in Sects. Ill and IV. Use of these control devices or techniques
results in the reduction of total VOC emissions by about 1100 Mg/yr for the
model plant, and in the controlled emissions from the model plant being about
95 Mg/yr.
a. Purge Gas Vent The use of purge gas as fuel reduces the model-plant VOC emissions
by an estimated 486 Mg/yr and also reduces the natural gas needed as fuel, an
increasingly important economic factor as the cost of natural gas increases.
All domestic methanol producers normally use purge gas as fuel or as feed to
another process (see Appendix D).
b. Distillation Vent Sending the distillation vent gases to a flare reduces the
model plant VOC emissions by 176 Mg/yr. A flare is needed to safely dispose of
emergency releases from other parts of the methanol process.
c- Other Emissions (Storage, Handling, and Fugitive) These sources are controlled
in the model plant by internal-floating-roof storage tanks, aqueous scrubbers,
and repair of leaking components for fugitive emissions. Application of these
controls results in a VOC emission reduction of 448 Mg/yr for the model plant.
Internal-floating-roof tanks for emission control neither consume energy nor
have adverse environmental or energy impacts. The electrical energy and process
water required for the aqueous scrubbers are negligible. The scrubbing water
is returned to process for recovery of the methanol in it.
2. Process Variations
The environmental and energy inpacts of controlling the high-pressure processes
and the processes where carbon dioxide is not added are similar to the impacts
described for the low-pressure model plant.
-------
Table VI-1. Environmental Impact of Controlled Methanol Model Plant
Emission Source
Purge gas vent
Distillation vent
Sulfur removal
Vent
Designation
(Fig.III-1)
A
B
C
Control Device
or Technique
Used as fuel
Flare
None
VOC Emission Reduction
(%) (Mg/yr)
98.2 486
98 176
Storage vents
Crude methanol
Check tanks*
Product
Handling
Loading tank cars
Loading barges
Fugitive
Secondary
Wastewater treatment
Waste organic as fuel
Total
D
E
F
G
H
I
J
K
Internal-floating-roof tanks 85
Internal-floating-roof tanks 85
Internal-floating-roof tanks 85
Aqueous scrubber 99
Aqueous scrubber 99
Detection and repair of major 80
leaks
None
None
3.7
31
140
22
43
208
<
H
I
1100
*Tanks for holding product until it is checked
-------
VI-3
3. 1979 Industry Emissions
The total VOC emissions from the domestic methanol industry in 1979 are estimated
to be 3000 Mg, and includes the estimated emissions from the process, fugitive,
secondary, and storage and handling sources (see Appendix D).
B. CONTROL COST IMPACT
1. Process Vents
Use of the purge gas as fuel or as feed to another process is necessary if the
plant is to be competitive economically; so there is no cost impact to providing
this control. A flare is necessary in a methanol plant for safe disposal of
emergency releases of flammable gases, and the cost impact of connecting the
distillation vent to it or to an equivalent control is negligible when a new
plant is being designed. The cost of retrofitting these controls to an existing
plant may be appreciably greater than their cost for a new installation if there
is some distance between the source and the existing control. No control has
been identified for the sulfur removal vent.
2. Storage and Handling Sources
The control system for storage sources is the use of internal-floating-roof
tanks. Aqueous scrubbers are used to control the model-plant methanol handling
emissions from loading tank cars and barges. Another EPA report covers storage
and handling emissions and their applicable controls for all the synthetic organic
chemicals manufacturing industry.
3. Fugitive Sources
EPA fugitive emissions and their applicable controls are discussed in a separate
EPA report.
4. Secondary Sources
No control system has been identified for controlling the secondary emissions
from wastewater treatment or from burning the waste organic streams as fuel.
Another EPA report covers secondary emissions and their applicable controls.
-------
VI-4
C. REFERENCES*
1. D. G. Erikson, IT Enviroscience, Inc., Storage and Handling (September 1980)
(EPA/ESED report. Research Triangle Park, NC).
2. D. G. Erikson and V. Kalcevic, IT Enviroscience, Inc., Fugitive Emissions
(September 1980) (EPA/ESED report, Research Triangle Park, NC).
3. J. J. Cudahy and R. L. Standifer, IT Enviroscience, Inc., Secondary Emissions
(September 1980) (EPA/ESED report, Research Triangle Park, NC).
*Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of
that paragraph, that reference number is indicated on the material involved.
When the reference appears on a heading, it refers to all the text covered by
that heading.
-------
VII-1
VII. SUMMARY
Methanol is produced from synthesis gas, a mixture of hydrogen and carbon oxides,
usually obtained by steam reforming of natural gas. The domestic production
capacity of methanol for 1980 was estimated to be 4,310,000 Mg. Formaldehyde
production consumes about 40% of the methanol produced. The estimated methanol
consumption annual growth rate is 7%. When current new constructon and expan-
sions are completed, capacity will be sufficient to satisfy domestic require-
ments beyond 1983.
Emission sources and uncontrolled and controlled emission rates for the methanol
model plant are given in Table VII-1. The major uncontrolled emission source
is the purge gas vent; normally the purge gas is used as fuel or as feed to
another process for economic reasons, therefore actual controlled emissions
from the source are small. The emissions from the distillation vent can be
flared in the plant flare system designed for safe disposal of emergency releases
of flammable gases. VOC emissions from the sulfur removal system vent during
regeneration are minor or nonexistent, depending on the amount of VOC in the
natural gas used as feed.
The model-plant methanol storage emissions are controlled by internal-floating-
roof tanks, and the emissions from loading tank cars and barges by aqueous scrub-
bers. Potential secondary emissions are minor. The total methanol industry
VOC emissions are estimated at 3000 Mg in 1979, with most of the uncontrolled
VOC emissions coming from fugitive, storage, and handling emissions.
A. D. Abshire, et al., "Methanol," pp. 674.5021A—I and 674.5022A—674.5026A
in Chemical Economics Handbook, Stanford Research Institute, Menlo Park, CA
(June 1980).
-------
VII-2
Table VII-1 Emission Summary for Methanol Model Plant
(450,000 Mg/yr)
Emission
Purge gas vent
Distillation vent
Sulfur removal
Storage vents
Crude methanol
Check tanks
Product
Handling
Loading tank cars
Loading barges
Fugitive
Secondary
Wastewater treatment
Waste organics as fuel
Total
Vent
Designation
(Fig.III-1)
A
B
C
D
E
F
G
H
I
J
K
VOC Emission Rate (kg/hr)
Uncontrolled
56.5
20.5
a
0.50
4.1
18.8
2.49
4.99
29.7
0.023
0.0031
138
Controlled
0.99
0.41
a
0.074
0.62
2.8
0.0249
0.0499
5.9
0.023
0.0031
10.9
Model-plant emissions during the sulfur removal unit regeneration contain no VOC.
Tanks for holding product until it is checked.
-------
A-l
APPENDIX A
Table A-l. 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
CH .O
4
32.04
Liquid
127.9 mm Hg at 25°C
1.11
64.8°C at 760 mm Hg
-93.9°C
0.7913 g/ml at 20°C/4°C
Soluble
*From: J. Dorigan et al., "Methyl Alcohol," p. AIII-154 in Scoring of
Organic Air Pollutants. Chemistry, Production and Toxicity of Selected
Synthetic Organic Chemicals (Chemicals F-N), Rev. 1, Appendix IIIf
MTR-7248, MITRE Corp., McLean, VA (September 1976).
-------
A-2
Table A-2. Physical Properies of
Dimethyl Ether*
Synonyms Methyl ether, methyl
oxide, methoxymeth-
ane
Molecular formula C2H60
Molecular weight 46.07
Physical state Gas
Vapor pressure 4551.3 mm Hg at 25°C
Vapor specific gravity 1.59
Boiling point -23.7°C at 760 mm Hg
Melting point -138.5°C
Density 0.661 g/ml
Water solubility 74,000 mg/liter of H20
*From: J. Dorigan et al., "Dimethyl Ether," p AII-144
in Scoring of Organic Air Pollutants. Chemistry,
Production and Toxicity of Selected Synthetic Or-
ganic Chemicals (Chemicals D—E), Rev 1, Appendix II
MTR-7248, MITRE Corp., McLean, VA (September 1976).
D18P(1)
-------
A-3
Table A-3. Physical Properties of
Methyl Formate*
Synonyms Methyl ester of formic
acid, methyl methano-
ate
Molecular formula C2H402
Molecular weight 60.05
Physical state Liquid
Vapor pressure 602.5 mm Hg at 25°C
Vapor specific gravity 2.07
Boiling point 32.0°C at 760 mm Hg
Melting point -99°C
Density 0.975 g/ml at 20°C/4°C
Water solubility Soluble
*From: J. Dorigan et al., "Methyl Formate," p AIII-194
in Scoring of Organic Air Pollutants. Chemistry,
Production and Toxicity of Selected Synthetic Organic
Chemicals (Chemicals F—N), Rev 1, Appendix III,
MTR-7248, MITRE Corp., McLean, VA (September 1976).
D18P(2)
-------
B-l
APPENDIX B
Table B-l. Air-Dispersion Parameters for
Model Plant with a Capacity of 450,000 Mg/yr
Source
purge gas vent
pistillation vent
Storage vents
Crude methane 1
Check tanXs (2)
product (2)
Dandling
Loading tank cars
Loading barges
Fugitive
Secondary
Wastewater treatment
Waste organic as fuel
b
purge gas vent
Flare (distillation vent)
Storage
Crude methanol
Check tanks (2)
product (2)
Aqueous scrubber
(loading emissions)
Fugitive
vcc
Emission
Rate
(g/sec)
15.7
5.7
0.14
0.57 (each)
2-6 (each)
0.69
1.4
8.2
0.0063
0.00086
0.28
0.11
0.021
0.086 (each)
0-39 (each)
0.021
1.G4
Discharge Flow Discharge
Height Diameter Temperature Rate Velocity
(nO (m)
Uncontrolled Emissions
20 0.6
20 0.1
12.2 14
9.8 10.3
14.6 41.2
4 0.5
2 0.5
1 30
30 16
Controlled Emissions
60 0.5
12.2 14
9.8 10.3
14.6 41.2
5 0.3
IK) (m3/sec) (m/sec)
320 6.3 22
320 0.0092 1.2
300
300
300
300
300
300
450 90 4.5
1250
300
300
300
300 0.063 0.9
3pugi.tive emission:"! are distributed over an area of Z50 m X 400 m.
used as fuel gas.
-------
C-l
APPENDIX C
FUGITIVE-EMISSION FACTORS*
The Environmental Protection Agency recently completed an extensive testing
program that resulted in updated fugitive-emission factors for petroleum re-
fineries. Other preliminary test results suggest that fugitive emissions from
sources in chemical plants are comparable to fugitive emissions from correspond-
ing sources in petroleum refineries. Therefore the emission factors established
for refineries are used in this report to estimate fugitive emissions from
organic chemical manufacture. These factors are presented below.
Source
Uncontrolled
Emission Factor
(kg/hr)
Controlled
Emission Factor*
(kg/hr)
Pump seals ,
Light-liquid service
Heavy- liquid service
Pipeline valves
Gas/vapor service
Light-liquid service
Heavy-liquid service
Safety/relief valves
Gas/vapor service
Light-liquid service
Heavy-liquid service
Compressor seals
Flanges
Drains
0.12
0.02
0.021
0.010
0.0003
0.16
0.006
0.009
0.44
0.00026
0.032
0.03
0.02
0.002
0.003
0.0003
0.061
0.006
0.009
0.11
0.00026
0.019
Based on monthly inspection of selected equipment; no inspection of
heavy-liquid equipment, flanges, or light-liquid relief valves;
10,000 ppmv VOC concentration at source defines a leak; and 15 days
allowed for correction of leaks.
Light liquid means any liquid more volatile than kerosene.
*Radian Corp., Emission Factors and Frequency of Leak Occurrence for Fittings
in Refinery Process Units, EPA 600/2-79-044 (February 1979).
-------
D-l
APPENDIX D
EXISTING PLANT CONSIDERATIONS
A. 1979 INDUSTRY EMISSIONS
The total VOC emissions from the domestic methanol industry in 1979 are estimated
to be 3000 Hg, and includes the estimated emissions from the process, fugitive,
secondary, and storage and handling sources. This estimate is based on an
estimated 1979 level of production of 3,400,000 Mg, calculated by applying the
estimated future 7% growth per year (see Sect. II) to the reported production
for 1977 and from the emission ratios from Tables IV-1 and V-l, together with
an estimate of the percentage of production associated with controlled and with
uncontrolled emissions in each category. These estimates are based on engineering
1 7
judgement, on data from individual methanol producers, — on state and local
emission control agencies, and on the open literature. The following individual
estimated projections were made:
Source 1979 VOC Emissions (Mg/yr)
Process 250
Storage and handling 1700
Fugitive 1200
Secondary 2
Total (rounded) 3000
The sources of the largest amounts of VOC emissions are storage and handling
and fugitive emissions. Since the retrofit of controls to these sources, that
is, internal-floating-roof tanks, vent scrubbers, or refrigerated vent condensers
for tanks and inspection of equipment for fugitive emissions, is often cost
effective, many producers are working to do so. Refrigerated vent condensers
are reportedly more economical to retrofit to existing tanks and only slightly
8 9
less efficient than internal floating roofs or scrubbers. —
Data comparing the uncontrolled VOC emissions from a Rohm and Haas methanol
plant using a high-pressure process with those estimated for the model plant
low-pressure plant were included in the draft version of this report but have
been removed because this plant no longer exists.
-------
D-2
B. EXISTING PLANT CHARACTERIZATION
Table D-l lists emission control devices reported in use by industry. To
gather information for the preparation of this report three site visits were
made to producers of methanol. Trip reports have been cleared by the companies
235
concerned and are on file at EPA, ESED, Research Triangle Park, NC. ' ' Other
sources of information in this appendix are letters in response to requests by
EPA for information on emissions from methanol plants. ' ' '
1. Air Products, Pensacola, FL
The following data on process emissions were supplied:
a. Distillation Condenser Vent This is a vent of noncondensibles form the vent
condenser on the methanol distillation column. Most of this stream is recovered
and sent to the boilers. Excess gas that cannot be handled by the vent compressor
is vented to the atmosphere.
Flow 0.86 acfm
Temperature 90°F
Methanol emissions 1.2 Ib/hr 4.81 tons/yr
b. Synthesis Vent—This vent is used infrequently and emits methanol synthesis gas
just upstream of the converter loop. It is used only during startups and
shutdowns.
Flow
Temperature
Emissions
H2
CO
CH
248 acfm
80°F
58.4 Ib/hr
68.5 Ib/hr
41.5 Ib/yr
234 tons/yr
275 tons/yr
166 tons/yr
c. Organic Sulfur Removal Unit Regeneration Vent (Intermittent) This vent is
used intermittently end the vent gases are comprised of steam and traces of
sulfur. High-pressure steam is used to regenerate the catalyst beds in organic
sulfur removal twice per week.
-------
Table D-l. Emission Control Devices or Techniques Currently Used by Some Methanol Producers
Control Devices or Techniques Used
Sourer
rur-jt; g-^s vent
Dintil Lntion vent
Stor.-ig^ vents
r™*"--".
H.-ind 1 inr)
''sec ref 1-
bSoe ref 2.
C-;o-. r-f 3.
lsce ref 4.
csoo rof 5; this plant
By Air Products3
Transferred to another
process; flash gas
scrubbed and sent to
boiler
Compressed and sent to
boiler; excess vented
to atmosphere
Hone
Hone
Hot reported
is no longer in existence
By Borden
Transferred to another
process or to reformer
fuel gas
Transferred to another
process; pure column
vented to atmosphere
Aqueous scrubber
Aqueous scrubber
None
(ref 10) .
By Celanese
To fuel gas with
flare alternate
To fuel gas
Floating roof
Floating roof or
to flare
To another
process
By Du Pont By Monsanto6
To fuel gas most of the To fuel gas or to
time reformer
Compressed and sent to Flare
fuel gash
Compressed by ejectors to None
fuel gas; motive gas is
purge gas
None Internal floating
roof
None None
By Rohm & Haas By Tenneco'
To fuel gas with flare To fuel gas
alternate
Flare Compressed
and sent to
fuel gas
Flare Not reported
Internal floating None
roof
Not applicable Hot reported
(used on site)
See rcf 6.
:3cne vunts go directly to the atmosphere. Monsanto reports that they arc on a program to install internal floating roofs and to reduce storage vent emissions to the atmosphere (see ref
-------
D-4
Flow 823 acfm
Temperature 450°F
H2S emissions 1.47 Ib/hr 0.18 tons/yr
2 11
2. Borden, Geismar, LA '
The methanol "B" plant uses Chemico high-pressure technology and was started up
in 1967. Since the site visit, Borden has been replacing this process with ICI
low-pressure technology and is expecting to start up the revamped and expanded
plant by the end of 1980.
3. Celanese, Bishop, TX
The methanol synthesis uses Lurgi low-pressure technology and was started up
in 1976.
4
4. Du Pont, Beaumont, TX
The following data on process emissions were supplied:
a. Natural Gas Desulfurizer Regeneration Vent This vent removes sulfur compounds
from activated carbon catalyst during regeneration of catalyst with steam (340
hr/yr).
Flow 825 Ib/hr
Composition, wt%
H2S 0.1
CH4 10.9
C2H6' C3H8' C4H10 °-4
H20 88.6
Temperature, °C 100
b. Reactor Vent This vent removes inert gases from the reactor system during
periods when there is no CO in the feed (approximately 35 days per year).
-------
D-5
Flow 15,000 Ib/hr
Composition, wt %
H2 36.0
CO 35.0
C02 15.0
CH4 10.5
N
2
dimethyl ether
CH OH, HO and 0.5
Temperature, °C 40 to 50
The above data were determined from composition analysis of purge stream and
estimated flow to the atmosphere. During periods when CO is being fed to the
reformer, this stream is burned as fuel in the reformer without being vented to
the atmosphere.
c Reformer Process Vents (2) - These vents discharge process gas from the reformer
during startup and shutdown operations. There is no discharge during normal
operation.
Flow 128,600 Ib/hr
Composition, wt %
H2 16.3
CH4 2.3
C02 39.6
CO 41.8
Temperature, °C 120
The above data were determined by gas analysis.
d Dehydrator Column Vent - This vent emits noncondensable gases from the condenser.
Flow 4 Ib/hr
Composition, wt %
CH3OH 85
N2 15
Temperature, °C 66
-------
D-6
The above data were determined by calculations based on operating conditions.
e. Splitter Column Vent This vent emits low-boiling organics during plant startup,
shutdown, or process upsets.
Flow 5850 Ib/hr
Composition, wt %
Dimethyl ether 62
Ch OH 37
Methyl formate, 1
methylal, acetone
Temperature, °C 55
The above data were determined by calculation based on operating conditions and
flow measurements. Emissions are normally compressed and burned in the reformer.
Only during startup, shutdown, or compressor shutdown are there any emissions
to atmosphere.
5. Tenneco, Pasadena, TX
The following data on process emissions were supplied:
a. Compressor Interstage Knockout Vent This vent provides liquid-vapor separation
during compression of synthesis gas.
Flow 202 scfm
Composition, %
H20 83
CO 17
MeOH 0.001
Temperature, °F 87
b. Desulfurizer Carbon-Bed Vent (Regeneration) Activated carbon adsorbes sulfur
compounds from natural gas prior to reforming. Carbon must be regenerated with
steam periodically to restore activity.
-------
D-7
Intermittent (during regeneration only)
Total emissions, tons/yr
H20 580
CH4 130
Ethane 16
Propane 24
i-Butane 16
n-butane 21
Temperature, °F 420 max.
-------
D-8
C. REFERENCES*
1. R. L. Duggan, Air Products and Chemicals Inc., letter dated May 11, 1978, to
EPA with information on air emissions from the methanol plant at Pensacola, FL,
in response to EPA request.
2. J. A. Key, IT Enviroscience, Inc., Trip Report for Visit to Borden, Geismar, LA,
Mar. 3, 1978 (on file at EPA, ESED, Research Triangle Park, NC).
3. J. A. Key, IT Enviroscience, Inc., Trip Report for Visit to Celanese, Bishop, TX,
Oct. 11, 1977 (on file at EPA, ESED, Research Triangle Park, NC).
4. D. W. Smith, E. I. du Pont de Nemours and Company, letter dated May 25, 1978,
to EPA with information on air emissions from the methanol plant at Beaumont,
TX, in response to EPA request.
5. J. A. Key, IT Enviroscience, Inc., Trip Report for Visit to Monsanto, Texas
City, TX, Dec. 13, 1977 (on file at EPA, ESED, Research Triangle Park, NC).
6. D. A. Copeland, Rohm and Haas Texas Incorporated, letter dated May 19, 1978, to
EPA with information on air emissions from the methanol plant at Deer Park, TX,
in response to EPA request.
7. W. P. Anderson, Tenneco Chemicals, letter dated May 10, 1978, to EPA with
information on air emissions from the methanol plant at Pasadena, TX, in
response to EPA request.
8. K. D. Dastur, E. I. du Pont de Nemours and Company, letter dated Sept. 27,
1979, to D. R. Patrick, EPA, with comments on draft Methanol report.
9. N. B. Galluzzo, Monsanto Plastics & Resins Co., letter dated Oct. 18, 1979, to
D. R. Patrick, EPA, with comments on draft Methanol report.
10. Rohm and Haas Company, letter dated Sept. 18, 1979, to D. R. Patrick, EPA,
with comments on draft Methanol report.
11. A. D. Abshire et al., "Methanol," pp 674.5021A—674.50211 and 674.5022A—
674.5026A in Chemical Economics Handbook, Stanford Research Institute, Menlo
Park, CA (June 1980).
*Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of
that paragraph, that reference number is indicated on the material involved.
When the reference appears on a heading, it refers to all the text covered by
that heading.
-------
3-i
REPORT 3
ETHYLENE
R. L. Standifer
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.
D17D
-------
3-iii
CONTENTS OP REPORT 3
Page
I. ABBREVIATIONS AND CONVERSION FACTORS 1-1
II. INDUSTRY DESCRIPTION II-I
A. Ethylene (Olefins) II-l
B. Ethylene Usage and Growth II-l
C. Domestic Producers II-4
D. References II-8
III. PROCESS DESCRIPTION III-l
A. Introduction III-l
B. Naphtha/Gas-Oil Process III-3
C. Process Variations III-ll
D. New Processes 111-19
E. Foreign Processes 111-20
F. References 111-21
IV. EMISSIONS IV-1
A. Current Pyrolysis Processes IV-1
B. Other Processes IV-23
C. References IV-24
V. APPLICABLE CONTROL SYSTEMS V-l
A. Current Pyrolysis Processes V-l
B. Other Processes V-17
C. References V-18
VI. IMPACT ANALYSIS VI-1
A. Control Cost Impact VI-1
B. Environmental and Energy Impacts VI-12
C. Reference VI-17
SUMMARY VII-1
-------
3-V"
APPENDICES OF REPORT 3
Page
A. PHYSICAL PROPERTIES OF FEEDSTOCKS AND PRODUCTS A-l
B. AIR-DISPERSION PARAMETERS (MODELS III AND VII) B-l
C. FUGITIVE-EMISSION FACTORS C-l
D. COST ESTIMATE DETAILS D-l
E. INTERMITTENT-EMISSION SAMPLE CALCULATIONS E-l
F. SALT-DOME STORAGE-EMISSION SAMPLE CALCULATIONS F-l
G. LIST OF EPA INFORMATION SOURCES G-l
H. EXISTING PLANT CONSIDERATIONS H-l
-------
3-vii
TABLES OF REPORT 3
Table No. Page
II-l Feed Requirements and Co-Product Yields from Various II-2
Feedstocks
II-2 Ethylene Usage and Growth Rate II-3
II-3 Producers and Plant Locations II-5
II-4 New Plants—Locations, Feedstocks, Capacities II-7
III-l Estimated Material Losses Due to Compressor Outages 111-18
IV-1 Capacity Data for Plants Using Various Feedstocks IV-2
IV-2 Production from Various Feedstock Combinations IV-3
IV-3 Benzene and VOC Uncontrolled Emissions for Model- IV-5
Plants I-II
IV-4 Benzene and VOC Uncontrolled Emissions for Model- IV-6
Plants III-IV
IV-5 Benzene and VOC Uncontrolled Emissions for Model- IV-7
Plants V-VI
IV-6 Benzene and VOC Uncontrolled Emissions for Model- IV-8
Plants VII-VIII
IV-7 Benzene and VOC Uncontrolled Emissions for Model- IV-9
Plants IX-X
IV-8 Assumed Charge Gas Composition from Model Plants IV-16
IV-9 Atmospheric Storage Tank Conditions IV-20
IV-10 Wastewater Parameters IV-22
V-l Benzene and VOC Controlled Emissions for Model-Plant I V-2
V-2 Benzene and VOC Controlled Emissions for Model-Plant II V-3
V-3 Benzene and VOC Controlled Emissions for Model-Plant III V-4
V-4 Benzene and VOC Controlled Emissions for Model-Plant IV V-5
V-5 Benzene and VOC Controlled Emissions for Model-Plant V V-6
V-6 Benzene and VOC Controlled Emissions for Model-Plant VI V-7
-------
3-ix
Tables (continued)
Table No. Page
V-7 Benzene and VOC Controlled Emissions for Model-Plant VII V-8
V-8 Benzene and VOC Controlled Emissions for Model-Plant VIII V-9
V-9 Benzene and VOC Controlled Emissions for Model-Plant IX V-10
V-10 Benzene and VOC Controlled Emissions for Model-Plant X V-ll
VI-1 Cost Factors Used in Computing Annual Costs VI-2
VI-2 Cost and Cost Effectiveness for Model-Plant Flares VI-3
VI-3 Environmental Impact of Controlled Model-Plant III VI-13
VI-4 Environmental Impact of Controlled Model-Plant VIII VI-14
VII-1 Emission Summary for Model-Plant III VII-2
VII-2 Emission Summary for Model-Plant VIII VII-2
VII-3 Composite Model-Plant Emissions Summary VII-3
VII-4 Estimated Emissions for the Industry VII-5
VII-5 Cost Effectiveness Ratios for Model-Plants III & VIII VII-5
-------
3-xi
FIGURES OF REPORT 3
Figure No. Page
II-l Locations of Plants Manufacturing Ethylene II-6
III-l Process Flow Diagram for Plant with Naphtha/Gas-Oil Feed III-4
III-2 Process Flow Diagram for Process with Ethane/Propane Feed 111-12
IV-1 Total Uncontrolled VOC Emissions vs Plant Capacity IV-10
IV-2 Uncontrolled Benzene Emissions vs Plant Capacity IV-12
V-l Total Controlled VOC Emissions vs Plant Capacity V-12
V-2 Total Controlled Benzene Emissions vs Plant Capacity V-14
VI-1 Capital Cost of Flare Options vs Plant Capacity for VI-5
Intermittent Emission Control Model-Plants I-X
VI-2 Cost Effectiveness of Flare Options vs Plant Capacity VI-6
for Control of Intermittent VOC Emissions
VI-3 Cost Effectiveness of Flare Options vs Plant Capacity VI-8
for Control of Intermittent Benzene Emissions
VI-4 Capital Cost of Flares vs Plant Capacity VI-10
for Model-Plants III Through V (50:50 E/P)
-------
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 nun 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
1
1
1
1
1
Tg =
Gg =
Mg =
km =
mV =
ug =
1
1
1
1
1
1
X
X
X
X
X
X
IO12
IO9
IO6
IO3
io"3
grams
grams
grams
meters
volt
10~6 gram
-------
II-l
II. INDUSTRY DESCRIPTION
A. ETHYLENE (OLEFINS)
The production of ethylene, a basic building block for a large segment of the
organic chemical industry, was selected for study because (1) it is produced in
greater quantity than any other organic chemical, (2) the total estimated emis-
sions of VOC are relatively high, (3) significant expansion in ethylene produc-
tion capacity is expected, and (4) some processes emit significant quantities
4
of benzene, which was listed as a hazardous pollutant by EPA in the Federal
Register on June 8, 1977.
Most ethylene is produced by the pyrolysis or thermal cracking of natural-gas
concentrates (primarily ethane and propane) or the heavier petroleum liquids
(primarily naphthas and atmospheric gas oils).
Propylene and 1,3-butadiene, formed with ethylene during pyrolysis, are generally
recovered as co-products. Therefore, the processes producing ethylene are also
termed "olefins" processes. Pyrolysis gasoline, another significant co-product,
is a complex mixture of the C,. and heavier compounds formed during pyrolysis,
with benzene as a major component. When liquid petroleum feedstocks, such as
naphthas and gas oils, are cracked, significant quantities of pyrolysis fuel oil
, , 2,5
are also produced.
Emissions include all components present in the cracked gas (see Appendix A for
pertinent physical properties of feedstocks and primary products). The cracked-
gas composition and product yield structure are strongly dependent on the feed-
stocks used and on the thermal cracking conditions. Table II-l shows typical
feed requirements and co-product yields for plants annually producing 454 Gg (1
billion Ib) of ethylene from several of the more common feedstocks.
B. ETHYLENE USAGE AND GROWTH
Table 11-2 shows the major ethylene uses and expected growth rate. The predomi-
nant uses are for the manufacture of polyethylene, styrene, ethylene oxide, and
vinyl chloride.
-------
Table II-l. Typical Feed Requirements and Co-Products Produced for
Plants Producing 453.5 Gg/yr (1 Billion Ib/yr) of Ethylene from Various Feedstocks'
Co-Product Yields (Gg/yr) for 453.5-Gg/yr Ethylene Plants
Feedstock
Ethane (E)
Propane (P)
50:50 E/P
Naphthas (N)
Atmosphere,
gas oils (G)
50:50 N/G
Feed
Requirements
(Gg/yr)
551.7
1030.6
791.1
1439.2
1774.3
1606.8
Based on averages of ranges
Hydrogen
and Methane
66.1
295.2
180.7
225.8
211.7
218.8
given in ref 1.
Propylene
11.2
165.1
88.1
205.5
269.2
237.3
Total C4
Mixture
13.2
45.5
29.4
137.6
169.8
153.7
Butadiene in
C4 Mixture
9.2
30.4
19.8
66.8
82.8
74.8
Pyro lysis
Gasoline
7.3
61.8
34.5
307.7
324.0
315.9
Benzene in
Pyro lysis
Gasoline
3.4
26.8
15.1
83.4
96.8
90.1
Fuel Oil
0.4
9.4
4.9
109.0
346.1
227.6
H
H
1
t\J
-------
II-3
Table II-2. Ethylene Usage and Growth Rate'
End Use
Low-density polyethylene
High-density polyethylene
Ethylene oxide
Vinyl chloride
Ethylbenzene , styrene
Ethyl alcohol
Aliphatic alochols
Acetaldehyde
Vinyl acetate
Ethyl chloride
Alpha olefins
Other
Percent of
Production
(1979)
27.2
17.5
17.8
11.9
8.0
2.7
2.4
2.3
2.2
1.0
2.1
4.9
100.0
Projected Average
Annual Percent Growth
(1979 — 1984)
4.0 — 4.5
6.0—^.5
2.0 — 3.0
6.0 — 7.0
2.5 — 3.5
1.5 — 2.5
5.0 — 6.0
(-5.0) — (-6.0)
2.5 — 3.5
(-7.0)— (-8.0)
6.5 — 7.5
Avg. 4.0 — 4.5
See ref 1.
-------
II-4
Domestic ethylene production in 1979 was 13,200 Gg. The estimated production
capacity at year's end was 16,500 Gg. Four new units and major expansions planned
for 1980 and 1981 will increase the capacity to 18,600 Gg by the end of 1981.
With a predicted annual growth rate of 4 to 4.5% the industry is expected to be
operating at about 75% of capacity by the end of 1981. '
C. DOMESTIC PRODUCERS
As of 1979 twenty-five manufacturers were producing ethylene in the U.S. and
Puerto Rico, at 36 locations. Table II-3 lists the producers and the feedstocks
used and Fig. II-l shows the plant locations.
The bulk of the increase in capacity is taking place on the Texas side of the
Gulf Coast. Table II-4 lists the producers, plant locations, and expected feed-
stocks for the new plants or significant expansions expected to be completed in
1980 and 1981.
With the increasing scarcity and higher cost of natural-gas liquids, almost all
new capacity planned after 1978 will use heavy-liquid feedstocks, primarily
naphthas and atmospheric gas oils. ' ' By 1981, 42 to 47% of U.S. ethylene
production is expected to be from heavy-liquid feeds, in contrast to 1969, when
90% was from natural-gas concentrates and refinery off-gases. The shift to
heavy- liquid feedstocks will result in the oil companies providing most of the
expansion in ethylene capacity because petroleum refineries are generally the
sources of these feedstocks, and large quantities of petroleum type of co-products,
especially gasoline and fuel oil, are produced (see Table II-l for comparative
yield structures).
Although very little expansion in ethylene capacity based on natural-gas concen-
trates is expected and even though a few of the older and less fuel-efficient
plants may close between 1979 and 1981, the production of ethylene from natural-
gas liquids is not expected to decrease appreciably during this period.
-------
II-5
Table II-3. Ethylene Producers as of 1979 and
Their Plant Locations, Feedstocks, and Capacities'
Producing Company
Plant Location
Feedstock
Ethylene Production
Capacity (69/yr)
Allied Chemical/BASF Wyandotte/Borg-Warner
Atlantic Richfield (Arco)
Atlantic Richfield (Arco)
Chenplex
Cities Service
Continental Oil
Dow Chemical
Dow chemical
E. i. du Pont de Nemours
Eastman Kodak
El Paso Products/Rexene Polyolefins
Exxon
Exxon
B.F. Goodrich
Gulf Oil
Gulf Oil
Mobil
Monsanto
Monsanto
National Distillers and Chemical
Internorth, Inc.
Olin Corp.
Phillips Petroleum
Shell oil
Shell oil
Standard Oil (Indiana)
Sun Company
Sunolin Chemical
Texaco
Texaco (Jefferson Chemical)
Union Carbide
Union Carbide
Dnion Carbide
Union Carbide
Union Carbide
U.S. Steel Corp.
Total
Geismar, LA
Channelview, TX
Wilmington, CA
Clinton, IO
Lake Charles, LA
Lake Charles, LA
Freeport, TX
Plaquemine, LA
Orange, TX
Longview, TX
Odessa, TX
Baton Rouge, LA
Baytown, TX
Calvert City, KY
Cedar Bayou, TX
Port Arthur, TX
Beaumont, TX
Alvin, TX
Texas City, TX
Tuscola, IL
East Morris, IL
Brandenburg, KY
Sweeny, TX
Deer Park, TX
Dorco, LA
Alvin, TX
Corpus Christi,TX
Claymont, DL
Port Arthur, TX
Port Nechfls, TX
Seadrift, TX
Taft, LA
Texas City, TX
Torrance, CA
Penuelas, PR
Houston, TX
Ethane, propane • 336
Naphtha, gas oil 1,179
Refinery off-gas, ethane, propane 45
Ethane, propane 238
Ethane, propane, butane .385
Ethane 295
Ethane, propane 1,020
Ethane, propane 544
Ethane, propane 374
Ethane, propane 576
Ethane, propane, butane 236
Ethane, propane, naphtha, gas oil 616
Gas.oil, naphtha 590
Propane 159
Ethane, propane, naphtha, gas oil 730
Refinery off-gas 567
Naphtha, gas liquids, refinery off-gas 408
Field condensate naphtha, gas oil 295
Field condensate 45
Ethane 181
Ethane, propane, butane 408
Ethane 45
Ethane, propane, butane 971
Ethane, propane, gaa oil 1,270
Ethane, propane, heavy liquids 635
Naphtha, gas liquids 907
Refinery gas 9
Refinery gas, ethane, propane 102
Naphtha, propane, butane 454
Ethane, propane, refinery gas 238
Ethane, propane, crude oil 544
Naphtha, ethane, propane 526
Ethane, propane 590
Ethane, propane 75
Naphtha 454
Refinery gas, ethane, propane 227
16,474
See raf 1.
-------
Fig. II-l. Locations of Plants Manufacturing Ethylene (ref 1.
(See Table II-3 for specific plant and location designations.)
Rico
-------
II-7
Table II-4. New Ethylene Plants (Expected to Be Completed Between 1980-
1981), Their Locations, Feedstocks, and Capacitiesa
Producing Company
Dow
Corpus Christ i
Petrochemicals
Conoco /Monsanto
Shell
Plant Location
Plaquemine, LA
Corpus Christi, TX
Chocolate Bayou, TX
Norco , LA
Feedstock
Naphtha
Naphtha
Naphtha
Gas oil
Ethylene Productic
Capacity (Gg/yr)
454
544
680b
680
See ref 1. •• .
Includes capacity of Monsanto's existing plant, which will be integrated into
new plant.
-------
II-8
D. REFERENCES*
1. S. Q. Cogswell, A. C. Gaessler, and T. A. Gibson, "Ethylene," pp 300.5200A—300.52051
in Chemical Economics Handbook,Stanford Research Institute, Menlo Park, CA (August
1980).
2. T. B. Baba and J. R. Kennedy, "Ethylene and Its Coproducts: The New Economics,"
Chemical Engineering 83(1), 116--128 (1976).
3. A. J. Cahill, "Ethylene—Past, Present, Near-Term Future," Chemical Engineering
Progress 73(7), 26--3S (1977).
4. Texas Air Control Board 1975 Emission Inventory Questionnaires.
5. A. D. Little, Inc., Environmental Considerations of Selected Energy Conserving
Manufacturing Process Options: Volume VI, "Olefins Industry Report," PB 264 272
(EPA-600/7-76-034f) U.S. Dept. of Commerce (December 1976) (available from the
National Technical Information Service, Springfield, VA).
6. "U.S. Ethylene Capacity.- Too Much of a Good Thing?" Chemical Engineering 85(7),
80, 81 (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 para-
graph, 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
Almost all commercial ethylene is produced by pyrolysis of natural-gas concentrates
and petroleum fractions. Although significant amounts of ethylene were once
extracted from by-product petroleum refinery streams (40% of the U.S. production
in 1956), only about 2% of the current ethylene production is derived from this
source. Most of the plants that are extracting ethylene from refinery streams
1 2
also produce ethylene by pyrolysis. '
Several alternative pyrolysis processes, primarily utilizing feedstocks not
currently in common use, are either being commercially attempted on a limited
scale or are in the development stage with expectations of limited commercial
application between 1980 and 1985. Although these processes are all expected to
be commercially proven within 5 years, wide application will depend on demon-
strated favorable process economics. No significant impact on total olefins
production is anticipated from these developmental processes for at least 10
2
years.
The primary difference between the domestic and foreign olefins industries has
been in the feedstocks used for pyrolysis. In Japan and Europe natural-gas
liquids have historically been scarce and naphtha has been the predominant feed-
2
stock.
The pyrolysis reaction mechanisms by which ethylene and co-products are formed
are very complex, particularly for the heavier feedstocks. The simplest example
2
of ethylene formation is the following free-radical sequence for ethane:
Initiation C,H, -» (CHj* + (CH,)*
2 o j -j
Propagation
-------
III-2
processes, pyrolysis is accomplished noncatalytically, inside radiantly heated
tubes. Optimum ethylene yields require a short residence time at pyrolytic
temperatures, followed by rapid quenching. Additional process operations are
required for removing water and undesirable impurities and for separating the
product fractions. '
Although all pyrolysis processes are alike in these basic requirements, there are
many specific process variations. The 35 domestic plant locations have at least
60 separate process units, which have been built over a time span of 40 years. '
Generally each unit was custom-designed to satisfy specific requirements and was
based on the prevailing technology, economic conditions, and regulatory require-
ments .
The most notable variations in relatively recent pyrolysis processes stem from
variations in feed composition. Since product yield structure and process
requirements depend strongly on feed composition, pyrolysis processes are
designed for specific feedstocks or combinations and generally cannot be operated
2 4
efficiently if there is much deviation from the design feed ranges. ' Although
a variety of specific feedstocks and combinations are used, the most prevalent
pyrolysis processes can be classified as those using natural-gas concentrates and
refinery off-gases, composed primarily of ethane, propane, and butane (E/P pro-
cesses); and those using heavy petroleum-based liquids, primarily naphthas and
2 3
atmospheric gas oils (N/G processes). '
Among the newer commercial processes, variations are much less significant. Most
new processes (1978 and later) have the following common characteristics:
1. Heavy-liquid feedstocks (primarily naphthas and atmospheric gas oils) --
As is shown in Table II-4, all significant new capacity projected for 1980
or later will use naphthas and gas oils.
2. High capacity — Most new units will produce from 450 to 680 Gg of ethylene
per year. Most smaller capacity increases will involve the expansion of
existing facilities (see Table II-4).
-------
III-3
3. Separation of product fractions by low-temperature distillation -- Although
low-temperature distillation has been the most common separation method for
many years, several older processes use selective absorption or adsorption.
4. Large-equipment and single-train operation -- Most newer olefins processes
are composed of a single process train and minimize the use of multiple,
parallel, major items of equipment. This trend has been most notable with
the increased use of large centrifugal compressors and larger fractionation
towers.
NAPHTHA/GAS-OIL PROCESS
The process flow diagram shown in Fig. III-l represents a recent process for
O fi *J
naphtha and/or atmospheric gas-oil feedstocks, ' ' the projected feedstocks for
almost all new ethylene capacity. Processes using natural-gas liquids (E/P
processes) are generally similar but less complex, with fewer and much smaller
quantities of co-products obtained (see Sect. III.C).
Naphtha and/or gas oil (Stream 1), diluted with steam, is fed in parallel to a
number of gas- or oil-fired tubular pyrolysis furnaces. The fuel gas and oil
(Stream 2) for these furnaces are supplied from gas and oil fractions removed
from the cracked gas in subsequent separation steps. In cracking naphtha and/or
gas oil, the ratio of steam to feed must be high for optimum yields and minimal
formation of coke in the reactor tubes. Ethane and propane, which are present in
the cracked gas and are separated in subsequent distillation steps (Streams 35
and 38), are combined and recycled (Stream 3) through a separate cracking fur-
nace. The resulting cracked gas is combined with the cracked gas from the naphtha/
gas-oil furnaces (Stream 5). The flue gas from the pyrolysis furnaces is vented
(Vent A). During operation, coke accumulates on the inside walls of the reactor
coils, and each furnace must be periodically taken out of service for removal of
the accumulated coke. Normally one furnace is out of service for decoking at all
times. Decoking is accomplished by passing steam and air through the coil while
the furnace is maintained at an elevated temperature, effectively burning the
carbon out of the coil. While a furnace is being decoked, the exhaust is diverted
(Stream 7) to an emission control device (Vent B) whose primary function is to
reduce particulate emissions. The collected particles are removed as a slurry
-------
H
H
I
Fig. III-l. Process Flow Diagram for Naphtha and/or Gas-Oil Feeds
-------
H
H
M
I
in
Fig. III-l. (Continued)
-------
III-6
(Stream 8). The cracked gas (Stream 4) leaving the pyrolysis furnaces is rapidly
7
cooled (quenched) to 250--300°C by passing it through transfer-line exchangers,
which terminate pyrolysis and simultaneously generate steam. The streams from
the transfer-line exchangers (Stream 5) are combined and further quenched by the
injection of recycled pyrolysis fuel oil from the gasoline fractionator
(Stream 6).
The remaining operations shown on Fig. III-l are required for separation of the
various product fractions formed in the cracking of gas oil and/or naphtha, for
removal of acid gases (primarily H S and CO ) and water, and for hydrogenation of
acetylene compounds to olefins or paraffins.
The quenched cracked gas (Stream 9) passes to the gasoline fractionator, where
pyrolysis fuel oil is separated. Most of the fuel oil passes through water-
cooled heat exchangers and is recycled (Stream 6) to the preceding oil-quenching
operation. The surplus fuel oil (Stream 10), equivalent to the quantity initi-
ally present in the cracked gas, passes first to the fuel oil stripper, where
light fractions are removed, and then to fuel oil storage. The light fractions
(Stream 11) removed in the fuel oil stripper are recycled to the gasoline frac-
tionator. The gasoline fractionator temperatures are well above the vaporization
temperature of water, and the contained water remains as superheated steam, with
the overhead stream containing the lighter cracked-gas components.
The overhead stream from the gasoline fractionator (Stream 12) passes to the
quench tower, where the temperature is further reduced, condensing most of the
water and part of the C_ and heavier compounds. The condensed organic phase
(Stream 13) is stripped of the lighter components in the gasoline stripper and is
passed to raw pyrolysis gasoline intermediate storage (Stream 14). Most of the
water phase, which is saturated with organics, is separated in the quench tower
(Stream 15), passed through water-cooled heat exchangers (Stream 16), and then
recycled to the quench tower to provide the necessary quench cooling. The sur-
plus water (Stream 17), approximately equivalent to the quantity of steam
7
injected with the pyrolysis furnace feed, passes to the dilution steam generator,
where it is vaporized and recycled as steam to the pyrolysis furnaces. Slowdown
from the recycle steam generator is removed as a wastewater stream (Stream 18).
-------
III-7
The overhead stream from the quench tower (Stream 19) passes to a centrifugal
charge-gas compressor (first three stages), where it is compressed to approxi-
mately 1.5 MPa. Water (Stream 20) and organic fractions (Stream 21) condensed
during compression and cooling are recycled to the quench tower and gasoline
stripper.
Lubricating oil (seal oil) discharged from the charge-gas compressor is stripped
of volatile organics in a separator pot before the oil is recirculated. The
organic vapor is vented to atmosphere (Vent G). Similar separator pots separate
volatile organics from lubricating oil from ethylene and propylene refrigeration
compressors (Streams 48 and 49).
Following compression, acid gas (H S, CO ) is removed by absorption in diethanol-
amine (DBA) or other similar solvents in the amine wash tower followed by a
caustic wash step. The amine stripper strips the acid gas (Stream 22) from the
saturated DEA and the DEA (Stream 23) is recycled to the amine wash tower. Very
little blowdown from the DEA cycle is required.
The waste caustic solution, blowdown from the DEA cycle, and wastewater from the
caustic wash tower are neutralized, stripped of acid gas, and removed as liquid-
waste streams (Streams 24 and 25). The acid gas stripped from the DEA and
caustic waste (Stream 22) passes to an emission control device (Vent D), primarily
to control H S emissions.
Following acid gas removal, the remaining process gas stream (Stream 26) is
further compressed to approximately 3.5—4 MPa (Stages 4 and 5) and is then
passed through drying traps that contain a desiccant, where the water content is
reduced to the low level necessary to prevent ice or hydrate formation in the
low-temperature distillation operations. The drying traps are operated on a
cyclic basis, with periodic regeneration necessary to remove accumulated water
from the desiccant. The desiccant is regenerated with heated fuel gas (not
shown), and the effluent gas is routed to the fuel system. Fouling of the desic-
cant by polymer formation necessitates periodic replacement, which results in the
generation of a solid waste (Stream 27); however, with normal desiccant service
life (possibly several years) this waste source is relatively minor.
-------
III-8
With the exception of three catalytic hydrogenation operations, the remaining
process steps involve a series of fractionations in which the various product
fractions are successively separated.
The demethanizer separates a mixture of hydrogen and methane from the C? and
heavier components of the process gas (Stream 28). The demethanizer overhead
stream (hydrogen and methane) is further separated into hydrogen-rich and methane-
rich streams (Streams 29 and 30) in the low-temperature chilling section. The
methane-rich stream is used primarily for furnace fuel. Hydrogen is required in
the catalytic hydrogenation operations.
The de-ethanizer separated the C components (ethylene, ethane, and acetylene)
(Stream 31) from the C, and heavier components (Stream 32). Following catalytic
hydrogenation of acetylene to ethylene by the acetylene converter (Stream 33),
the ethylene-ethane split is made by the ethylene fractionator. The overhead
from the ethylene fractionator (Stream 34) is removed as the purified ethylene
product, and the ethane fraction (Stream 35) is recycled to the ethane/propane
cracking furnace.
The de-ethanizer bottoms (C and heavier compounds) (Stream 32) pass to the
depropanizer, where a C -C. split is made. The depropanizer overhead stream
(primarily propylene and propane) (Stream 36) passes to a catalytic hydrogenation
reactor (C converter), where traces of propadiene and methyl acetylene are
hydrogenated. Following hydrogenation, the C fraction passes to the propylene
fractionator, where propylene is removed overhead as a purified product
(Stream 37). The propane (Stream 38) is recycled to the ethane/propane pyrolysis
furnace.
The C and heavier components (Stream 39) from the depropanizer pass to the
debutanizer, where a C -Cj. split is made. The overhead C stream (Stream 40) is
removed as feed to a separate butadiene process.
The stream containing C,. and heavier compounds from the debutanizer (Stream 41)
is combined with the bottoms fraction from the gasoline stripper (Stream 14) as
raw pyrolysis gasoline. The combined stream (Stream 42) is hydrogenated in the
gasoline treatment section. Following the stripping of lights (Stream 43), which
-------
III-9
are recycled to the cracked-gas compressor, the C and heavier compounds
(Stream 44) are transferred to storage as treated pyrolysis gasoline. This
stream contains benzene and other aromatics formed by pyrolysis.
To meet the low-temperature requirements of most of the fractionation columns,
liquid ethylene and propylene are used as refrigerants. A significant part of
the process equipment is included in the refrigeration cycles. These cycles
consist of centrifugal propylene and ethylene compressors, a complex of flash
tanks, condensers, and heat exchangers, all of which are necessary to attain the
required low temperatures and to efficiently utilize the refrigeration. As the
only normal process emissions from the refrigeration cycles are from the ethylene
and propylene compressor lubricating oil vents (Streams 48 and 49), the refrigera-
tion cycles are not included in the flowsheet.
The three catalytic hydrogenation reactors for acetylene, C compounds, and
pyrolysis gasoline all require periodic regeneration of the catalyst to remove
contaminants. The catalyst is generally regenerated every four to six months.
At the start of regeneration, as superheated steam (Stream 45) is passed through
a reactor, a mixture of steam and hydrocarbons leaving the reactor (Stream 46) is
passed to the quench tower (arrow not shown). After sufficient time has elapsed
for stripping of malodorous organics (approximately 48 hr), the exhaust is
directed to an atmospheric vent (Vent F) and a steam-air mixture is passed
through the catalyst to remove residual carbon. This operation continues for an
additional 24 to 48 hr. The presence of air during this phase of the regenera-
tion prevents the vented vapor from being returned to the process.
Emissions from the gasoline hydrogenation reactor heater (Vent A) and the cata-
lyst regeneration steam superheater (Vent A) are composed of flue gas formed from
the combustion of gaseous fuel.
The process described in this section is characterized by very low VOC emissions
from process vents during normal operation, with only one minor benzene emission
source (charge-gas compressor lubricating oil vent, Stream 47, Vent G).
Most process emissons of VOC occur during abnormal conditions, such as schedule
startups and shutdowns, process upsets, and emergency situations. Emissions
-------
111-10
result from the activation of pressure-relief devices, the intentional venting of
off-specification materials, and the depressurization and purging of equipment in
preparation for maintenance. The greatest quantity of intermittent emissions
results from outages of the refrigeration and charge-gas compressors. Although
compressor outages are relatively infrequent and of short-term duration, the
resultant high rates cause significant quantities of VOC emissions, including
benzene. With the exception of the demethanizer relief valves, which may vent
directly to the atmosphere (not shown), all pressure-relief devices and con-
trolled emergency vents are routed through the main process vent (Vent E) to an
emission-control device. Emissions resulting from activation of demethanizer
relief valves are infrequent and are composed primarily of hydrogen and methane.
Fugitive emissions can contain all components present in the cracked gas, includ-
ing benzene. The extreme variation in composition throughout the process pro-
duces widely varying compositions of fugitive emissions. As with most organic
chemical processes, leaks into cooling water can occur, allowing volatile organic
compounds (VOC) to escape.
Storage-tank emission sources (labeled C on Fig. III-l) include naphtha, gas oil,
pyrolysis fuel oil, and pyrolysis gasoline. Primary storage of ethylene is in
pressurized underground salt domes. Emissions that occur when dissolved VOC is
stripped from salt brine displaced from the storage domes are vented (Vent H).
Since feedstock and products are transferred by pipeline, handling emissions are
not significant.
The five potential sources of secondary emissions (labeled K on Fig. III-l) are
(1) blowdown from the dilution steam generator, (2) spent caustic, (3) wash water
from the caustic wash tower, (4) coke generated from pyrolysis furnace and trans-
fer line exchanger decoking, and (5) spent desiccant from the process gas dryers.
The coke is composed of uncombined carbon containing organics with very high
molecular weights, and secondary emissions are very low. Secondary emissions
from spent desiccant are insignificant since the organics remaining after steam
purging of the traps have high molecular weights and desiccant replacement is
infrequent (approximately evary 4 or 5 years).
-------
III-ll
C. PROCESS VARIATIONS
With more than 60 domestic process units, constructed over a time span of approxi-
mately 40 years ' and frequently designed to satisfy individual requirements,
variations in commercial ethylene processes are numerous. The most significant
variations and their effects on VOC emissions are summarized below:
1 Feedstock Composition
Most new ethylene production will use naphthas and atmospheric gas oils as feed-
stocks, but most current ethylene production is derived from ethane and propane.
Although various feedstock combinations are employed, most processes use either
heavy liquids (N/G processes) or gas concentrates (E/P processes). Butane and
naphtha are supplemental feeds in some processes that use primarily ethane and
propane feedstocks. Heavy-liquid processes are generally designed for either
naphtha or gas oil or for a combination of the two, with butane as a relatively
minor supplemental feed.
238
a Ethane-Propane Feed (E/P Processes) — The E/P process shown in Fig. III-2 ' '
is similar to the N/G process (Fig. III-l) but is less complex. All process
steps and stream designations shown on Fig. III-2 are included in Fig. III-l with
identical stream and emission source designations. Stream and emission sources
designated in Fig. III-l for the N/G process that do not occur in the E/P process
(Fig. III-2) are omitted. The E/P process simplifications are as follows:
1. A gasoline fractionator and fuel oil stripper are not included because
essentially no pyrolysis fuel oil is produced (see Table II-l).
2. The cracked gas is not partially quenched by oil. It is entirely quenched
in transfer line exchangers and a quench tower.
3. Since the sulfur content of ethane and propane obtained from natural-gas
concentrates generally is very low, an MEA tower may not be included. The
smaller quantities of acid gas (H S, CO ) may be removed by caustic scrubbing
£• £*
alone. H^S or SO emissions, a significant problem with heavier feedstocks
(particularly gas oils), are not generally a significant problem with E/P
processes.
-------
H
H
M
I-1
NJ
Fig. III-2. Flow Diagram for Ethane-Propane Feed Process
-------
<$
LOW -T EMPti. A ATuft C
CMiu-iKia
SC.CT1OKJ
1
J«?>
JC
T0h
V
4 1
.MtRATieXi
«.Ul»t*-
I
M
U)
Fig. III-2. (Continued)
-------
111-14
o o
4. A pyrolysis gasoline treatment section is not usually required. '
In the simplest case, in which ethane is the sole feedstock, the process may be
further simplified by elimination of the depropanizer, propylene fractionator,
? Q
and debutanizer. ' The small quantities of co-products formed are not separated
but are recycled to the pyrolysis furnaces or are burned as fuel.
b. Effects of Feedstock Variations -- Increasingly heavy feedstocks (progressing
from ethane to gas oil) result in the following general trends in process char-
acteristics :
1. More raw material is required and larger quantities of co-products are
formed per unit of ethylene produced, resulting in increased process capac-
ity requirements (see Table II-l).
Q
2. The process becomes increasingly complex, with more process steps required.
3. Shorter residence time, higher heat flux, and higher steam dilution ratios
2
during pyrolysis are required.
4. The quantity of coke formed during pyrolysis is greater and decoking is more
frequent.
2
5. The quantity of process wastewater is increased.
6. Operating problems caused by heavy-residue formation are more severe, result-
2
ing in more frequent maintenance.
2
7. The sulfur content of feedstocks is generally much higher.
Although the process variations described do not significantly affect direct
process emissions of VOC during normal operation, intermittent, fugitive, and
secondary VOC emissions are potentially greater with the heavier feedstocks
because of the greater quantities of materials processed, the increased process
complexity, and the greater quantities of wastewater generated. With higher
concentrations of benzene in the cracked gas, intermittent, fugitive and secon-
dary emissions of benzene are significantly greater with the heavier feedstocks.
-------
111-15
2
Storage tank emissions are generally less from E/P processes. Atmospheric
tanks, used for naphtha, gas oil, and pyrolysis fuel oil storage when heavy-
liquid feeds are used, are not used and pyrolysis gasoline storage requirements
are less. Ethane and propane are generally received by pipeline and are either
fed directly to the pyrolysis furnaces or stored in pressure vessels. Vapor
vented from pressurized storage vessels is also introduced as pyrolysis furnace
feed.
Cracked-Gas Quenching and Quench-Water Treatment
Variations in the methods of quenching the cracked gas and in the cooling, hand-
ling, and disposal of effluent water from quenching operations result in signifi-
cant variations in VOC emissions. The most significant variations, prevalent
primarily in older processes, can result in relatively large quantities of VOC
and benzene emissions. Effluent quench water is potentially a significant VOC
and benzene emission source. The water is saturated with the organic compounds
present in the cracked gas at the quench-tower operating pressure (approximately
200 kPa), and will release organic vapor if the pressure is reduced as it is
discharged. Because benzene is a relatively high-boiling cracked-gas component,
the proportion of benzene in the organics emitted from quench water is signifi-
cantly higher than the benzene concentration in the cracked gas.
g
Very significant VOC emissions from quench water have been reported for older
processes in which (1) the hot quench water was discharged directly to vented,
atmospheric, settling basins, where organic residues were separated; and (2) the
effluent water from the settling basins was then passed through cooling towers
before being recycled to the quench tower, effectively stripping and venting most
of the remaining organics. In one case combined VOC emissions of more than*
400 kg/hr from both sources were initially reported. The reported emissions in
this case were subsequently reduced by approximately 90% with the installation of
a vacuum stripping tower. In the revised process the quench-tower effluent water
passes through the stripping tower before it is discharged to the settling basins,
and the stripped vapor is recycled to the quench tower or is incinerated.
There are no significant emissions from recycled quench water in the process
shown in Fig. III-l because (1) phase separation is attained in the quench tower
base, thereby eliminating venting of the contained vapor, and (2) the process
water is cooled in water-cooled heat exchangers. When heat exchangers are used,
-------
111-16
the cooling water that subsequently passes through cooling towers does not con-
tact process organics.
In most older processes the excess quench water effluent, which results from the
addition of dilution steam in the pyrolysis furnaces, is not recycled as dilution
steam but is removed as a wastewater stream. The quantity of wastewater result-
ing from steam dilution is potentially greater when heavy-liquid feedstocks are
used because higher ratios of steam to feed are necessary.
Some older processes do not utilize transfer line exchangers for primary quench-
3 9
ing. ' The primary purpose of transfer line exchangers is to improve process
thermal efficiency. Emissions are only indirectly affected. When transfer line
exchangers are not used, two variations are: a two-stage quench system composed
of a primary oil-quench, followed by a water-quench step; or quenching solely
with water. With a total water quench the greater amounts of water used may
result in increased emissions from the discharged quench water.
3. Compressors
In contrast to recent processes that primarily utilize high-capacity, single-
train, centrifugal compressors (typically equipped with high-efficiency, oil-
purged mechanical seals), many older processes use larger numbers of lower
capacity compressors operated in parallel. As a result of the larger number of
compressors and the use of either reciprocating compressors or older centrifugal
compressors with less efficient seals, compressor fugitive emissions are typi-
cally greater than those from the more recent processes.
Intermittent process emissions caused by compressor outages are typically much
greater with single compressor trains than with parallel dual or multiple trains.
A compressor shutdown will result in the venting of process material, generally
to a flare.
The quantity of material vented depends on the venting rate and the time required
to re-establish stable operation. When a compressor outage occurs, the venting
rate is much less with dual compressor trains and the return to stable operation
is much more rapid. The emergency shutdown of one of two parallel compressors
does not interrupt the other machine, and product purity at a production rate of
50% of capacity can be maintained.
-------
111-17
The specific compressor that trips (e.g., charge gas, propylene, ethylene) affects
the quantity of material vented. Generally charge-gas compressor outages are
much more frequent than refrigeration compressor outages.
The estimated material losses caused by compressor outages for single- and dual-
train plants experienced by one producer are shown in Table III-l. Losses
were estimated for the first 5 years of operation, from plants producing 450-Gg/yr
ethylene from naphtha/gas-oil feeds. Losses resulting from compressor outages
tended to decrease during the first 5 years of operation, because operating
problems were eliminated and compressor reliability improved, becoming relatively
constant after the fifth year. Atmospheric emissions resulting from these losses
depend on the efficiency of the emission-control devices.
Raw Material and Product Transfer
Although pipeline transfer of all raw materials and products is widely used,
shipment of heavier products (i.e., propylene, crude butadiene, pyrolysis fuel
oil, pyrolysis gasoline) by tank car, tank truck, and barge is also common. When
methods other than pipeline transfer are used, potential emissions are signifi-
cantly greater. Other methods for transferring ethylene are relatively rare and
are generally used only for relatively small quantities of ethylene.
Integration of Related Processes
The ethylene (olefins) process shown in Fig. III-l does not include the separa-
tion of 1,3-butadiene from the C fraction nor the separation of C_ compounds,
benzene, toluene, and xylene contained in pyrolysis gasoline. Although generally
performed in separate process units, in some cases these operations are included
as integral operations within the olefins units.
Separation of Other Product Fractions
Acetylene may be removed as a separate product by an absorption-stripping opera-
tion instead of eliminating it by catalytic hydrogenation. Other variations
include separating pyrolysis fuel oil into two or more fractions, separating the
Cg fraction from pyrolysis gasoline, and removing propane as a product instead of
recycling to pyrolysis.
-------
111-18
Table III-l. Estimated Material Losses Due to Compressor Outages'
(453.5 Gg/yr Ethylene, N/G Feed)
Year of Operation
Outages and Material Losses
Annual trips and checks
Material lost per trip, Mg
Annual compressor maintenance
Material lost per occurrence, Mg
Total material lost, Mg
Cost at raw-material value ($1000)
1
10
816
3
571
9878
1180
2
Single
7
816
2
571
6857
820
3
4
5
Compressor Trains^
6
816
1
571
5469
650
Dual Compressor
Annual trips and checks
Material lost per trip, Mg
Annual compressor maintenance
Material lost per occurrence, Mg
Total material lost, Mg
Cost at raw-material value ($1000)
20
61
6
0
1224
150
14
61
4
0
857
100
12
61
2
0
735
90
5
816
1
571
4653
560
Trains
10
61
2
0
612
70
4
816
1
571
3837
460
8
61
2
0
490
60
See ref 10.
-------
111-19
7- Refrigerants and Refrigeration Cycles
The type of refrigerants used varies, as do the number of levels of refrigeration
and the refrigeration cycle equipment configurations. Alternative refrigerants
include propane, methane, and ethane. The refrigerants are almost invariably
pyrolysis gas components.
8. Miscellaneous Process Variations
Other process variations that have relatively minor effects on emissions are
1. separation of products by selective absorption or adsorption instead of
low-temperature fractionation,
2. variation in the order of some process steps, including charge-gas compres-
sion, acid-gas removal, water removal, acetylene hydrogenation, and product
fraction separations,
3. variations in fractionation tower conditions (i.e., pressure, temperature,
reflux ratios),
4. use of other processes or solvents for the removal of acid gases.
D. NEW PROCESSES
Pyrolysis processes primarily using feedstocks not currently in common use for
the production of olefins are being commercially attempted on a limited scale or
are in the development stage, with expectations of limited commercial application
between 1980 and 1985.2 Wider application will depend on demonstrated favorable
process economics, and no significant impact on total olefins production is
anticipated for at least 10 years. Such processes are conventional pyrolysis of
vacuum gas oil (Exxon), autothermic pyrolysis of crude oil (Union Carbide/
Kureha), fluid-bed pyrolysis of petroleum residues (AIST), and fluid-bed pyrolysis
of coal (Garret Corp.).
The preliminary nature of these processes makes specific emission data unavail-
able. Variations from current processes are primarily in the pyrolysis steps,
with generally similar separation and purification steps. Following the general
-------
111-20
trends of heavier feedstocks, VOC emissions are probably slightly higher than
2
from current atmospheric gas-oil processes.
E. FOREIGN PROCESSES
In most foreign commercial ethylene processes, liquid petroleum feedstocks are
used primarily. Processes are generally similar to domestic naphtha/gas-oil
2
processes. The alternative processes <
both domestic and foreign developments.
2
processes. The alternative processes discussed in the preceding section include
-------
111-21
F. REFERENCES*
1. E. M. Carlson and M. G. Erskine, "Ethylene," pp. 648.5051--648.5055H in
Chemical Economics Handbook, Stanford Research Institute, Menlo Park, CA
(February 1975); see also S. L. Soder and R. E. Davenport, ibid.,
pp. 648.5051A-648.5055Y (January 1978).
2. A. D. Little, Inc., Environmental Considerations of Selected Energy Conserving
Manufacturing Process Options: Volume VI. "Olefins Industry Report," PB 264 272
(EPA-600/7-76-034f), U.S. Dept. of Commerce (December 1976) (available from the
National Technical Information Service, Springfield, VA).
3. S. Takaoka, "Ethylene," Report No. 29, Process Economics Program, Stanford
Research Institute, Menlo Park, CA (August 1967).
4. "Ethylene: The End of an Era," Chemical Engineering 84(7), 63—65, 1977.
5. R. L. Standifer, IT Enviroscience, Trip Report for Arco Chemical Co., Channel-
view, TX, Aug. 16-17, 1977 (on file at EPA, ESED, Research Triangle Park, NC).
6. R. L. Standifer, IT Enviroscience, Trip Report for Gulf Oil Chemicals Co., Cedar
Bayou Olefins Plant Cedar Bayou, TX, Sept. 13-14, 1977 (on file at EPA, ESED,
Research Triangle Park, NC).
7. Lumus Co., Process for the Pyrolysis of Hydrocarbons, British Patent 1,047,905
(Nov. 9, 1966).
8. T. Baba and J. Kennedy, "Ethylene and Its Coproducts: The New Economics,"
Chemical Engineering 83(1), 116--128 (1976).
9. Texas Air Control Board, 1975 Emission Inventory Questionnaires.
10. R. p. Paveletic, H. C. Skinner, and D. Stewart, "Why Dual Ethylene Unit Com-
pressors?" Hydrocarbon Processing 55(10), 135--138 (1976).
11. W. E. Nelson, "Compressor Seal Fundamentals," Hydrocarbon Processing 56(12),
91—95 (1977). —
^Usually, when a reference is located at the end of a paragraph, it refers to the
entire paragraph. If another reference relates to certain portions of that
paragraph, that reference number is indicated on the material involved. When the
reference'appears on a heading, it refers to all the text covered by that head-
ing.
-------
IV-1
IV. EMISSIONS
A.
CURRENT PYROLYSIS PROCESSES
1. Model Plants
Current operating ethylene plants and those scheduled to go on-stream between
1978 and 1981 encompass a wide range of production capacities and feedstock
combinations. Tables IV-1 and IV-2 provide breakdowns as to capacities and feed-
stocks by production plants. Many plant sites have more than one ethylene
unit. The total capacity shown (20,456 Gg/yr) and the total number of process
units (67) include both those currently in operation and those projected to start
up within the next 5 years. As a number of older units are expected to shut down
during the period, the actual total capacity at any specific time will be some-
what less. As shown by Table IV-1, the most prevalent processes in terms of
total production capacity and number of operating units include ethane/propane
plants with ethylene capacities of 160 to 340 Gg/yr and naphtha and/or gas-oil
plants with capacities of 340 to 610 Gg/yr. Most units expected to go on stream
within the next 2 years will have naphtha/gas-oil flexibility and will be capable
of producing at least 450 Gg of ethylene per year.
The following ten model plants will be considered based on 8760 hr of operation
annually:
Model No. Feed Composition
I 100% ethane
II 100% propane
III 50% ethane, 50% propane
IV 50% ethane, 50% propane
V 50% ethane, 50% propane
VI 100% naphtha
VII 100% gas oil
VIII 50% naphtha, 50% gas oil
IX 50% naphtha, 50% gas oil
X 50% naphtha, 50% gas oil
Ethylene Production Capacity*
[Gg/yr (lb/vr)]
226.8 (500 X 106)
226.8 (500 X 106)
226.8 (500 X 106)
158.7 (350 X 106)
340.1 (750 X 106)
544.2 (1200 X 106)
544.2 (1200 X 106)
544.2 (1200 X 106)
680.3 (1500 X 106)
408.2 (900 X 106*
*Process downtime is normally expected to range from 5 to 15%. If the hourly
rate remains constant, the annual production and the 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 gegligible.
-------
IV-2
Table IV-1. Capacity Data for Plants Using Various Feedstocks
Feedstock
Ethane, propane, butane
Naphtha, gas oil, field
condensate
Mixed gas liquid and
heavy liquid
Number
of Plants
5
30
4
2
14
2
2
4
0
Ethylene
Plant
<160
160-340
>340
<340
340-610
>610
<340
340-610
>610
Capacity
Total
381
6952
1753
426
7166
1451
526
1723
0
(Gg/yr)
Average
77
232
438
213
512
726
263
431
0
-------
IV-3
Table IV-2. Ethylene Production from Various Feedstock Combinations
Feedstocks Number
Ethane
Propane
Ethane, propane
Ethane, propane, butane
Refinery off-gas
Refinery off-gas, ethane, propane
Ethane, propane, naphtha
Naphtha
Naphtha, gas oil
Gas oil
Field condensate
Naphtha, field condensate, raffinate
Naphtha, ethane, propane, refinery off-gas
Butane
of Plants
3
1
25
5
3
5
5
4
8
3
1
2
1
1
Total Ethylene
Capacity (Gg/yr)
521
159
5642
1818
50
839
1841
1447
4580
1950
340
680
408
181
67
20,456
-------
IV-4
Estimated emissions for Models VI through X (naphtha and/or gas oil) are based on
the process shown in Fig. III-l and described in Sect. III-B. Naphtha/gas-oil
flexibility is assumed (i.e., process equipment and storage tanks for Models VI
through VIII are the same and variations in emissions are caused by variations in
feed composition only).
Emissions for Models I through V (ethane and/or propane) are based on the process
shown in Fig. III-2, with storage tank capacities sized for specific feed com-
position.
For Model I (ethane feed) it was assumed that propylene, C4 compounds, and pyrol-
ysis gasoline are separated as co-products. This simplifying assumption permits
emissions from processes using mixed ethane/propane feeds to be estimated for the
entire range of ethane/propane ratios from the single-feed component models.
Processes with ethane as the sole feed component and in which co-products are not
separated are relatively few and account for only a small fraction of total
ethylene production.
Criteria for process, storage, fugitive, and secondary emissions for the models
are discussed in the corresponding emissions sections: process emissions,
Sects. 2.a. and b.; fugitive emissions, Sect. 2.c; storage emissions, Sects. 2.d
and e; and secondary emissions, Sect. 2.f. Atmospheric dispersion parameters for
Model-Plants III and VIII are given in Appendix B.
2. Sources and Emissions
Benzene and VOC emission rates and ratios (emissions/ethylene production) for the
ten models are summarized in Tables IV-3 through IV-7 and are shown graphically
in Figs. IV-1 and IV-2. Estimated VOC emissions do not include methane (methane
and hydrogen are significant components of intermittent and fugitive emissions).
Because intermittent emissions (Vent E) predominate and are significantly less
with dual refrigeration and charge-gas compressor trains, estimates of inter-
mittent emissions are included for both single- and dual-train processes.
a. Normal Process Emissions — Process emissions of VOC and benzene occurring during
normal operation are very low for all models. Characteristics of the emissions
-------
IV-5
Table IV-3. Benzene and Total VDC Uncontrolled Emissions for
Model-Plants I and II
Emission Ratio
Source , .
_.___,._.__ (g/Mg)
Source
Compressor lube-oil vents
Single compressor trains
Dual compressor trains
Other normal process
emissions
Intermittent emissions
Single compressor trains
Dual compressor trains
Storage tanks
Salt-dome storage
Fugitive
Secondary
Compressor lube-oil vents
Single compressor trains
Dual compressor trains
Other normal process
emissions
Intermittent emissions
Single compressor trains
Dual compressor trains
Storage tanks
Salt-dome storage
Fugitive
Secondary
UGH j_yuai_.Li_iii
(Fig. III-2) Benzene
b
Model-Plant I
G
0.0348
0.0696
A,B,D,F
E
29.8
3.8
C 3.4
H
8.3
K 8.9
Model-Plant II
G
0.187
0.374
A,B,D,F
E
235.8
30.1
C 27.7
H
43.0
K 8.9
Total VOC
13.9
27.8
27.6
6910
1506
24
274
3608
23
13.9
27.8
51.5
8460
1704
195
274
3295
23
Emission Rate (kg/hr)
Benzene
0.0009
0.0018
0.77
0.099
0.088
0.21
0.23
0.00483
0.00966
6.1
0.78
0.71
1.11
0.23
Total VOC
0.36
0.72
0.71
179.0
39.0
0.61
7.1
93.4
0.59
0.36
0.72
1.33
219.0
44.1
5.0
7.1
85.3
0.59
ag of benzene or total VOC per Mg of ethylene produced.
Feed, ethane; ethylene capacity, 226.8 Gg/yr.
CFeed, propane; ethylene capacity, 226.8 Gg/yr.
-------
IV-6
Table IV-4. Benzene and Total VOC Uncontrolled Emissions for
Model-Plants III and IV
]
Source
Compressor lube-oil vents
Single compressor trains
Dual compressor trains
Other normal process
emissions
Intermittent emissions
Single compressor trains
Dual compressor trains
Storage tanks
Salt -dome storage
Fugitive
Secondary
Compressor lube-oil vents
Single compressor trains
Dual compressor trains
Other normal process
emissions
Intermittent emissions
Single compressor trains
Dual compressor trains
Storage tanks
Salt -dome storage
Fugitive
Secondary
g of benzene or total VOC per
Feed, 50% ethane-50% propane;
°Feed, 50% ethane-50% propane;
Emission Ratio
^ Source (q/Mg) Emission Rate (kg/hr)
Designation -• - -
(Fig. III-2) Benzene
Model Plant II Ib
G
0.111
0.222
A,B,D,F
E
116.3
14.9
C 15.8
H
25.6
K 8.9
Model Plant IVC
G
0.159
0.318
A,B,D,F
E
116.3
14.9
C 15.9
H
36.7
K 8.9
Mg of ethylene produced.
ethylene capacity, 226.8
ethylene capacity, 158.7
Total VOC Benzene
13.9 0.00288
27.8 0.00576
39.6
7822 3.0
1621 0.39
111 0.41
274
3453 0.66
23 0.23
19.9 0.00288
39.7 0.00576
39.6
7822 2.1
1621 0.27
112 0.29
276
4935 0.66
23 0.16
Gg/yr.
Gg/yr.
Total VOC
0.36
0.72
1.025
202.5
42.0
2.9
7.1
89.4
0.59
0.36
0.72
0.72
141.7
29.4
2.0
5.0
89.4
0.42
-------
IV-7
Table IV-5. Benzene and Total VOC Uncontrolled Emissions for
Model-Plants V and VI
a
Emission Ratio
Source . . .
..__.,...,„ (g/Mg)
Source
Compressor lube-oil vents
Single compressor trains
Dual compressor trains
Other normal process
emissions
Intermittent emissions
Single compressor trains
Dual compressor trains
Storage tanks
Salt-dome storage
Fugitive
Secondary
Compressor lube-oil vents
Single compressor trains
Dual compressor trains
Other normal process
emissions
Intermittent emissions
Single compressor trains
Dual compressor trains
Storage tanks
Salt-dome storage
Fugitive
Secondary
uca j.yiicii.jLuii —^— — — —
(Fig. III-2) Benzene
Model Plant V13
G
0.0740
0.1480
A,B,D,F
E
116.4
14.9
C 15.6
H
17.1
K 8.9
Model Plant VIC
G
0.193
0.386
A,B,D,F
E
733.2
93.6
c 143.2
H
44.5
K 15.5
Total VOC
0.25
18.50
39.6
7822
1621
110
276
2303
23
5.79
11.60
72.0
11,690
2,114
1,067
274
1,450
40
Emission
Benzene
0.00288
0.00576
4.5
0.58
0.61
0.66
0.34
0.0120
0.0240
45.6
5.8
8.9
2.8
0.96
Rate (kg/hr)
Total VOC
0.36
0.72
1.54
303.7
63.0
4.3
10.7
89.4
0.90
0.36
0.72
4.47
726.0
131.3
66.3
17.0
90.1
2.47
g of benzene or total VOC per Mg of ethylene
bFeed, 50% ethane-50% propane; ethylene capacity, 340.1 Gg/yr.
°Feed, naphtha; ethylene capacity, 544.2 Gg/yr.
-------
IV-8
Table IV-6. Benzene and Total VOC Uncontrolled Emissions for
Model-Plants VII and VIII
. a
Emission Ratio
_ source. (g/Mg)
Source
Compressor lube-oil vents
Single compressor trains
Dual compressor trains
Other normal process
emissions
Intermittent emissions
Single compressor trains
Dual compressor trains
Storage tanks
Salt-dome storage
Fugitive
Secondary
Compressor lube-oil vents
Single compressor trains
Dual compressor trains
Other normal process
emissions
Intermittent emissions
Single compressor trains
Dual compressor trains
Storage tanks
Salt-dome storage
Fugitive
Secondary
Lies iyndu j.uu
(Fig. III-2) Benzene
Model-Plant VII
G
0.206
0.412
A,B,D,F
E
851.4
108.7
C 151.6
H
47.4
Total VOC
b
5.79
11.60
88.8
12,968
2,278
1,068
274
1,466
K 28.0 73
Model-Plant VIIIC
G
0.200
0.400
A,B,D,F
E
791.9
101.1
C 147
H
45.9
K 21.7
5.79
11.60
80.4
12,314
2,195
1,068
274
1,460
56
Emission
Benzene
0.0128
0.0256
52.9
6.8
9.4
2.9
1.74
0.0124
0.0248
49.2
6.3
9.2
2.85
1.35
Rate (kg/hr
Total VOC
0.36
0.72
5.51
805.6
141.5
66.3
17.0
91.1
4.52
0.36
0.72
4.99
765.0
136.3
66.3
17.0
90.7
3.49
g of benzene or total VOC per Mg of ethylene produced.
DFeed, gas oil; ethylene capacity, 544.2 Gg/yr.
'Feed, 50% naphtha-50% gas oil; ethylene capacity, 544.2 Gg/yr.
-------
IV-9
Table IV-7. Benzene and Total VOC Uncontrolled Emissions for
Model-Plants IX and X
Source
Compressor lube-oil vents
Single compressor trains
Dual compressor trains
Other normal process
emissions
Intermittent emissions
Single compressor trains
Dual compressor trains
Storage tanks
Salt-dome storage
Fugitive
Secondary
Compressor lube-oil vents
Single compressor trains
Dual compressor trains
Other normal process
emissions
Intermittent emissions
Single compressor trains
Dual compressor trains
Storage tanks
Salt-dome storage
fugitive
Secondary
Emission
Source , ,
(<3/Mg)
a
Ratio
(Fig. III-2) Benzene Total VOC
Model-Plant IX
G
0.160
0.320
A,B,D,F
E
791.9
101.1
C 145.8
H
36.8
K 21.7
Model-Plant Xc
G
0.266
0.532
A,B,D,F
E
791.9
101.1
C 148.5
H
61.2
K 21.7
4.64
9.28
80.4
12,314
2,195
1,055
274
1,168
56
7.73
15.46
80.4
12,313
2,194
1,076
275
1,946
56
Emission
Benzene
0.0124
0.0248
61.5
7.9
11.3
2.85
1.69
0.0124
0.0248
36.9
4.7
6.9
2.85
1.01
Rate (kg/hr!
Total VOC
0.36
0.72
6.24
956.2
170.5
82.0
21.3
90.7
4.36
0.36
0,72
3.75
573.7
102.2
50.1
12.8
90.7
2.61
9 of benzene or total VOC
Feed, 50% naphtha-50% gas
•^
"Feed, 50% naphtha-50% gas
per Mg of ethylene produced.
oil; ethylene capacity, 680.3 Gg/yr.
oil; ethylene capacity, 408.2 Gg/yr.
-------
IV-10
1000
100
10
100
200
300
400
500
600
700
Fig. IV-1. Total Uncontrolled VOC Emissions vs Plant
Capacity for Model Plants I Through X
-------
IV-11
Legend for Fig. IV-1.
Curve
la
Ib
Ic
Id
le
If
2a
2b
2c
2d
2e
2f
3
4a
4b
4c
5
6
7
8
9
10
Emission Source
intermittent (single compressor trains)
Intermittent (single compressor trains)
Intermittent (single compressor trains)
Intermittent (dual compressor trains)
Intermittent (dual compressor trains)
Intermittent (dual compressor trains)
Intermittent (single compressor trains)
Intermittent (single compressor trains)
Intermittent (single compressor trains)
Intermittent (dual compressor trains)
Intermittent (dual compressor trains)
Intermittent (dual compressor trains)
Secondary emissions (below scale)
Secondary emissions
Secondary emissions
Secondary emissions
Fugitive emissions
Fugitive emissions
Storage tanks
Storage tanks
Normal process emissions (single compressor
trains)
Normal process emissions (single compressor.
trains)
Feedstock
50:50 E/P
Ethane
Propane
50:50 E/P
Ethane
Propane
50:50 N/G
Naphtha
Gas oil
50:50 N/G
Naphtha
Gas oil
50:50 E/P
50:50 N/G
Naphtha
Gas oil
50:50 E/P
50:50 N/G
50:50 E/P
50:50 N/G
50:50 E/P
50:50 N/G
-------
IV-12
100.0
0.1
100
20O
300
400
500
600
700
Ethylene Capacity (Gg/yr)
Fia IV-2 Total Uncontrolled Benzene Emissions vs
Plant'capacity for Model Plants I Through X
-------
IV-13
Legend for Fig. IV-2.
Curve
la
Ib
Ic
Id
le
If
2a
2b
2c
2d
2e
2f
3
4a
4b
4c
5a
5b
5c
6
7
8
Emission Source
Intermittent (single compressor train)
Intermittent (single compressor train)
Intermittent (single compressor train)
Intermittent (dual compressor trains)
Intermittent (dual compressor train)
Intermittent (dual compressor train)
Intermittent (single compressor train)
Intermittent (single compressor train)
Intermittent (single compressor train)
Intermittent (dual compressor train)
Intermittent (dual compressor train)
Intermittent (dual compressor train)
Secondary emissions
Secondary emissions
Secondary emissions
Secondary emissions
Fugitive emissions
Fugitive emissions
Fugitive emissions
Fugitive emissions
Storage tanks
Storage tanks
Feedstock
50:50 E/P
Ethane
Propane
50:50 E/P
Ethane
Propane
50: 50 N/G
Naphtha
Gas oil
50:50 N/G
Naphtha
Gas oil
50:50 E/P
50:50 N/G
Naphtha
Gas oil
50:50 E/P
Ethane
Propane
50:50 N/G
50:50 E/P
50:50 N/G
-------
IV-14
from the sources identified in Figs. III-l and III-2 are summarized in the fol-
lowing items (i)--(v).
(i) Flue gas (Vent A) -- Emissions are composed of combustion products generated
from the combustion of primarily gaseous fuels (primarily H and CH ) sepa-
rated from the cracked gas. The naphtha and/or gas-oil feedstock processes
(Models VI through X) also utilize fuel oil produced from the process. VOC
concentrations are relatively low, with no significant benzene concentra-
tion.
(ii) Pyrolysis furnace decoking (Vent B) -- Emissions are composed primarily of
air, steam, CO , CO, and particles of unburned carbon, with no significant
concentrations of VOC or benzene. The primary purpose of emission control
devices is for particulate control.
(iii) Acid gas removal (Vent D) — Emissions are composed primarily of H S, SO ,
and CO.. Emission controls either
convert H S to SO before venting.
and CO.. Emission controls either remove and recover H-S as sulfur or
(iv) Hydrogenation catalyst regeneration (Vent F) — Emissions are infrequent,
occurring only during catalyst regeneration (4- to 6-month intervals). VOC
emissions are very low, with no significant concentrations of benzene.
(v) Compressor lubricating oil vents (Vent G) -- Although VOC concentrations are
significant, the flow rates are relatively low. The charge-gas-compressor
lubricating oil vent (Stream 47) is the only normal process vent emitting
benzene.
Estimated normal process VOC emissions for Models VI through X (naphtha and/or
gas-oil feedstocks) were developed directly from data received from ethylene
4 5
manufacturers. ' Because the reported data were for processes using both naph-
tha and gas oil, the normal process emissions for Models VI through VIII were
assumed to be identical; the actual differences are minor. Estimated VOC emis-
sions for Models I through V (ethane and/or propane) were developed from the same
data based on the following criteria:
-------
IV-15
(i) The quantity of flue gas vented is proportional to the total quantity of
feedstocks consumed. Feedstock requirements and yield structures for
Models I through V were based on the ethane and propane data in Table II-l.
(ii) VOC concentration in the flue gas is identical for all models.
(iii) Total VOC emissions from compressor lubricating oil vents are the same for
all models.
Normal process emissions of benzene (charge-gas compressor lube-oil vent only)
are based on the following criteria:
(i) Benzene concentrations in the charge-gas compressor lube-oil vents are the
same as compressor inlet concentrations (see Table IV-8).
(ii) Compressor inlet concentrations of benzene are as follows (from Table IV-8)
Model
I
II
III-V
VI
VII
VIII-X
Feed Composition
Ethane
Propane
50:50 E/P
Naphtha
Gas oil
50:50 N/G
Benzene Concentration (wt %)
0.44
2.28
1.36
5.66
6.03
5.85
b. Intermittent Emissions (Vent E) -- Most process emissions from all ethylene model
plants occur as intermittent emissions. Intermittent emissions result from the
activation of pressure-relief devices, the depressurization and purging of equip-
ment in preparation for maintenance, and the intentional venting of off-specification
products generated during abnormal conditions. Most intermittent emissions are
caused by compressor outages, which primarily result in the venting of pyrolysis
gas (process compressor charge gas, Stream 19).
With the exception of emissions from demethanizer relief valves all intermittent
emissions are vented through the main process vent (Vent E). The demethanizer
relief valves, which release primarily hydrogen and methane, are vented sepa-
-------
Table IV-8. Typical Charge-Gas Compositions and Rates for Plants Producing
453.5 Gg/yr (1 billion Ib/yr) of Ethylene from Various Feedstocks
Feedstock.
Ethane
Propane
50:50 E/P
Naphtha
Atmospheric
gas oil
50:50 N/G
% of Feed
Recycled
40.0
15.0
23.7
10.0
10.0
10.0
Total Pyrolysis
Gas Rate from
Furnaces (Gg/yr)
772.4
1185.2
978.8
1583.1
1951.7
1767.4
Compressor
Charge- Gas
(Gg/yr)
772.0
1175.8
973.9
1474.1
1605.6
1539.9
Charge Gas Composition (wt %)
Hydrogen and
Methane
8.6
25.1
16.9
15.3
13.2
14.3
Ethylene
58.7
38.6
48.7
30.8
28.1
29.4
Propylene
1.5
14.0
7.8
13.9
16.8
15.4
Ethane and
Propane
28.6
13.1
20.8
9.8
11.1
10.4
Total C4's
1.7
3.9
'2.8
9.3
10.6
10.0
Contained
Butadiene
1.2
2.6
1.9
4.5
5.2
4.8
Pyrolysis
Gasoline
0.9
5. 3
3.1
20.9
20.2
20.6
Contained
Benzene
0.44
2.28
1.36
5.66
6.03
5.85
f
-------
IV-17
rately. Estimates of intermittent emissions for the model plants are based on
the following criteria:
(i) Emissions of process compressor charge gas (Streams 50—52) are based on the
5-year average material losses caused by compressor outages, shown in
Table III-l.6 Table III-l was developed for a plant with naphtha--gas-oil
flexibility, producing 453.5 Gg of ethylene per year. The quantity of
charge-gas emissions for all models is assumed to be proportional to the
quantity of feedstocks consumed less the fuel oil produced (Table II-l).
Fuel oil initially present in the pyrolysis gas (N/G processes only) is
separated before compression and is usually not vented when a compressor
outage occurs. Because Table III-l applies to a process with naphtha/gas-
oil flexibility, average values for naphtha and gas oil from Table II-l were
assumed for all naphtha and/or gas-oil models (Models VI-X). Although
ethylene production/feed ratios are generally lower for gas oil than for
naphtha, charge-gas compressor inlet rates (after pyrolysis fuel oil
removal) are approximately the same at the same ethylene production rates.
(ii) The assumed compositions and rates of charge gas and the corresponding
compositions of charge-gas emissions for the model plants, based primarily
on the feed requirements and product yield data presented in Table II-l and
on typical recycle ratios for the various feed materials, are given in
Table IV-8.
(iii) Charge-gas emissions from Table III-l account for 95% of total intermittent
emissions occurring with single compressor train operation (Table III-l
conditions only). For the model plants charge-gas emissions are assumed to
be proportional to the respective charge-gas production rates, and other
intermittent emissions are assumed to be proportional to the respective
ethylene production rates and to contain no significant quantities of hydro-
gen, methane, or benzene. Other intermittent emissions are the same for
both single and dual trains for all models. (The methods for calculating
intermittent emissions are detailed in the sample calculations in Appendix F.)
Fugitive Emissions — Process pumps, valves, compressors, continuous process
analyzers, and process samples are potential sources of fugitive emissions. The
-------
IV-18
factors used to establish the emission rates for pumps, valves, and compressors
are shown in Appendix C. The number of emission sources is based on the fol-
lowing data for a naphtha/gas-oil feedstock process:
Type of Source Number
Pumps 165
Compressors 8
Valves 4150
Relief valves 65
Process samples 100/day
(Notes: Emissions from refrigeration and charge-gas compressor lube-oil vents
are included in normal process emissons; fugitive emissions from relief valves are
vented through the main process vent and are controlled with intermittent
process emissions.)
Estimated emissions for the ten model plants are based on the following criteria:
(i) Total fugitive emissions (including H and CH ) for all models are the same.
(ii) 50% of fugitive emissions contain 100% nonmethane VOC.
(iii) 50% of fugitive emissions contain 85% nonmethane VOC.
(iv) 50% of fugitive emissions contain no benzene.
(v) The average benzene concentrations for the remaining 50% fugitive emissions
are as follows (from Table IV-8):
Model Benzene Concentration (%)
I 0.44
II 2.28
III-V 1.36
VI 5.66
VII 6.03
VIII-X 5.85
-------
IV-19
(vi) The collection of process samples in sample containers results in fugitive
emissions of 0.085 m /sample.
-3 3
(vii) Sample flow rates to continuous process analyzers are 2.8 X 10 m /min for
each stream sampled. Twenty streams are continuously sampled in all model
plants.
d. Storage Tanks — Naphtha and gas-oil feedstocks (Models VI through X only)
pyrolysis gasoline, and pyrolysis fuel oil (Models VI through X only) are the
only feedstocks or products stored in atmospheric storage tanks (see Table IV-9).
Ethylene is stored in underground salt domes. Gas liquid feedstocks (Models I
through V) and other products (i.e., propylene, C compounds) are stored in
pressurized storage tanks, with no significant emissions resulting. All storage
tank emission sources are shown on the flow diagram in Fig. III-2 (Vent C).
Equations from AP-42 were used for the emission calculations. However, breathing
losses were divided by 4 to account for recent evidence indicating that the AP-42
7
breathing-loss equation overpredicts emissions.
Storage tanks for naphtha and gas-oil feedstocks (Models VI through X only) are
sized to provide a 3-day supply. Feedstocks are normally consumed at the same
rates at which they are received by pipeline, with storage tanks providing surge
capacity for short-term differences between receiving and consuming rates.
Pyrolysis gasoline storage tanks for all models have a 14-day capacity, and
pyrolysis fuel oil storage tank capacity (Models VI through X) is 3 days. These
4 5
storage capacities are consistent with data received from ethylene manufacturers. '
Tanks are sized to conform with the yield structures shown in Table II-l. To
provide naphtha/gas-oil flexibility, tanks are sized for the feedstock composition
requiring the greatest capacity.
e. Salt-Dome Storage (Vent H) -- Emissions may result from the venting of ethylene
absorbed in salt brine, which is displaced from the salt domes as ethylene is
placed in storage. The estimated maximum uncontrolled emissions resulting from
salt-dome storage that are given in Tables IV-1 through IV-7 are based on the
following criteria:
1. Emissions are proportional to throughput and are independent of the total
amount of ethylene in storage.
-------
Table IV-9. Atmospheric Storage Tank Conditions
Tank Size ~{m )
Contents
Naphtha/gas oil
Raw pyrolysis gasoline
Treated pyrolysis gasoline
Pyrolysis gasoline (models
c
Light pyrolysis fuel oil
c
Heavy pyrolysis fuel oil
Emissions from fixed-roof
naphtha only.
No. of Model
Tanks I II III IV V
2
1
2
I — V) 1 283 1105 695 486 1043
1
1
tanks containing gas oil were insignificant.
Turnovers Tank Level
VI-VIII
10,670
20,920
10,460
1,139
2,152
IX
13,340
26,150
13,070
1,424
2,690
Floating-roof- tank
X Per
8,000 0
15,690 0
7,850 26
26
854 0
1,614 O
Year (% Of Height
b
Constant 80%
Constant 80%
10-80%
10-80%
b
Constant 80%
b
Constant 80%
emissions were determined for H
<
0
CEmissions from fixed-roof tanks were insignificant; floating-roof-tank emissions were not determined.
-------
IV-21
£
2. The annual throughput for all models is equal to 15 days' production.
3. The discharged brine is saturated with ethylene at the storage dome con-
ditions .
Sample calculations are given in Appendix G.
The estimated emissions resulting from salt-dome storage, given in Tables IV-1
through IV-7, represent the maximum emissions that would occur from this source
only if brine saturated with ethylene is discharged. Because only the brine
adjacent to the ethylene-brine interface probably approaches saturation, the
actual emissions are estimated to be much less if a substantial brine level is
maintained at all times and if the brine is discharged from the bottom of the
i 8"10
dome only.
Secondary Emissions -- Secondary emissions of VOC and benzene can result pri-
marily from the handling and disposal of process wastewater. For the model
plants four potential secondary sources are indicated on the flow diagram,
Fig. III-2 (Source K). The solid wastes from the process (coke, spent desiccant)
do not present a significant emission potential. Coke, removed from the pyrol-
ysis furnace coils and from transfer line exchangers, is primarily free carbon
and polymer-like organics with very low vapor pressures. Desiccant replacement
is infrequent (every 3 to 5 years), and most residual organics with significant
vapor pressures are removed by steam purging before desiccant replacement.
Solid-waste disposal is normally by landfill.
No actual plant data on emissions from process wastewater were available. For
the model plants emission estimates are based on the criteria given in
Table IV-10 (see refs. 11 and 12).
Actual VOC and benzene emissions from wastewater may vary significantly from the
estimated emissions for the models. Additional plant and experimental data are
needed for a more accurate assessment. Secondary emissions and controls for all
areas of the synthetic organic chemical industry are covered in a separate
,. 12
report.
-------
IV-2 2
Table IV-10. Wastewater Parameters
Model No.
and
Feed
I— V, E/P
VI , naphtha
VII, gas oil
VIII--X,
50-50 N/G
Wastewater/
Ethylene
Ratio (m^/Gg)
889
1528
2804
2166
Parameters
Avg . Cone .
of Organic
Compounds (ppm by wt)
104
104
104
104
Emission
Ratio
(kg/kg)
0.25
0.25
0.25
0.25
Benzene
Concentration in
a
Emitted VOC (%)
38.5
38.5
38.5
38.5
m of wastewater per Gg of ethylene product; see ref 11.
b
See ref. 4.
kg of VOC emitted per kg of organics in wastewater. Results of air stripping experiments
(see ref 12) indicate that from 0 to 50% of contained organics are vented during transfer
and biological treatment of wastewater, depending primarily on activity coefficient of
organic components. An average value of 25% was assumed.
d
Benzene concentration for all model plants is based on the composition of vapor vented
from the quench water from an ethane/propane feedstock process.
-------
IV-23
B- OTHER PROCESSES
Emission data for the developmental processes described in Sect. III-D are not
currently available.
-------
IV-24
C. REFERENCES*
1. E. M. Carlson and M. G. Erskine, "Ethylene," pp. 648.5051--648.5055H in Chemical
Economics Handbook, Stanford Research Institute, Menlo Park, CA (February 1975);
see also S. L. Soder and R. E. Davenport, ibid., pp. 648.5051A--648.5055Y
(January 1978).
2. T. Baba and J. Kennedy, "Ethylene and Its Coproducts.- The New Economics,"
Chemical Engineering 83(1), 116--128 (1976).
3. S. Takaoka, "Ethylene," Report No. 29, Process Economics Program, Stanford
Research Institute, Menlo Park, CA (August 1967).
4. R. L. Standifer, IT Enviroscience, Trip Report for Gulf Oil Chemicals Co.,
Cedar Bayou Olefins Plant, Cedar Bayou, TX, Sept. 13-14, 1977 (data on file at
EPA, ESED, Research Triangle Park, NC).
5. R. L. Standifer, IT Enviroscience, Trip Report for Arco Chemical Co.,
Channelview, TX, Aug. 16-17, 1977 (data on file at EPA, ESED, Research Triangle
Park, NC).
6. R. P. Paveletic, H. C. Skinner, and D. Stewart, "Why Dual Ethylene Unit Com-
pressors?" Hydrocarbon Processing 55(10), 135--138 (1976).
7. E. C. Pulaski, TRW Inc., letter dated May 30, 1979, to Richard Burr, EPA.
8. J. P. Walsh, Exxon Chemical Co., letter dated Feb. 26, 1979, to EPA with informa-
tion on ethylene process at Baton Rouge, LA, in response to EPA request for
comments on ethylene draft report.
9. R. J. Feldman, C. E. Lummus Co., letter dated Mar. 22, 1979, to EPA responding to
EPA request for comments on ethylene draft report.
10. C. A. Gosoline, MCA, letter dated Jan. 25, 1979, to EPA responding to EPA request
for comments on ethylene draftreport.
11. A. D. Little, Inc., Environmental Considerations of Selected Energy Conserving
Manufacturing Process Options: Volume VI. "Olefins Industry Report," PB 264 272
(EPA-600/7-76-034f), U.S. Dept. of Commerce (December 1976) (available from the
National Technical Information Service, Springfield, VA).
12. J. J. Cudahy and R. L. Standifer, IT Enviroscience, Secondary Emissions (June
1980) (EAP/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 head-
ing.
-------
V-l
V. APPLICABLE CONTROL SYSTEMS
A. CURRENT PYROLYSIS PROCESSES
Estimates of total controlled VOC and benzene emission rates and ratios (emis-
sions/ ethylene production) for the ten model plants, with single and dual com-
pressor trains and with various control options are summarized in Tables V-l*
through V-10 and are shown graphically in Figs. V-l and V-2.
1. Normal Process Emissions
The compressor lubricating-oil vents are relatively small sources of VOC and
benzene emissions and no specific controls have been specified for the model
plants. These emissions can be effectively controlled by being recycled to the
charge-gas compressor suction; however, because of the relatively small quantities
of emissions occurring, the cost of the required piping may not be justifiable.
No additional controls for normal process emissions are indicated.
2. Intermittent Emissions
Intermittent process emissions can be effectively controlled by flares. Because
these emissions are relatively infrequent, of short-term duration, and occur at
extremely high and variable rates, other control methods are not generally
applicable. Estimates of controlled intermittent emissions are included for both
single and dual compressor train processes. Dual compressor trains are considered
a process variation. Retrofitting of processes with single compressor trains to
dual trains is not considered to be feasible.
Elevated flares that utilize steam injection to provide smokeless emissions are
specified for the model plants. Other types of flares, primarily ground flares,
and other methods of improving combustion (e.g., air injection, water spray) are
less commonly used in the ethylene industry. One ethylene manufacturer who
uses a ground flare with water spray for this purpose indicates that it has
significant advantages. Flare efficiencies have not been satisfactorily documented
except for specific designs and operating conditions using specific fuels. The
*The internal-floating-roof tanks referred to in Tables V-l—V-10 consist of
internal floating covers or covered floating roofs as defined in API 25-19,
2d ed., 1976 (fixed-roof tanks with internal floating device to reduce vapor
loss).
-------
Table V-l. Benzene and Total VOC Controlled Emissions for Model-Plant
Source
Designation
Source (Fig. III-2)
Compressor lube-oil vents G
Single compressor trains
Dual compressor trains
Other normal process emissions A,B,D,F
Intermittent process emissions E
Single compressor trains
At 98* flare efficiency0
At 90\ flare efficiency0
Dual compressor trains
At 98* flare efficiency0
At 90% flare efficiency0
Storage tanks C
Salt-dome storage H
Fugitive
Secondary K
VOC
Control Device Emission
or Technique Reduction (*)
None
None
Flares
98.0
90.0
98.0
90.0
Floating-roof tanks 85.0
Maintain adequate brine 100.0
inventory
Detection and correction of 80.8
leaks, mechanical seals
None
Emission Ratio
(g/Mg)
Benzene
0.0348
0.0696
0.6
3.0
0.08
0.38
0.5
1.58
8.9
Total VOC
13.9
27.8
.27.6
138
691
30
151
3.6
0
690
23
Emission
Benzene
0.0009
0.0018
0.015
0.077
0.002
0.010
0.013
0.041
0.23
Rate (kg/hr)
Total VOC
0.36
0.72
0.71
3.6
17.9
0.8
3.9
0.09
0
17.9
0.59
Feed, ethane; ethylene capacity, 226.7 Gg/yr.
bg of benzene or total VOC per Mg of ethylene produced.
cFlare efficiencies have not been satisfactorily documented except for specific designs and operating conditions using specific fuels.
are for tentative comparison purposes.
Efficiencies used
-------
Table V-2. Benzene and Total VOC Controlled Emissions for Model-Plant II*
Source
Designation
Source (Fig. IH-2)
Compressor lube-oil vents G
Single compressor trains
Dual compressor trains
Other normal process emissions A,B,D,F
Intermittent process emissions E
Single compressor trains
At 98% flare efficiency
At 90% flare efficiency5
Dual compressor trains
At 98% flare efficiency
At 90% flare efficiency0
Storage tanks C
Salt-dome storage H
Fugitive
Secondary K
VOC
Control Device Emission
or Technique Reduction (%)
None
None
Flares
98.0
90.0
98.0
90.0
Floating-roof tanks 85.0
Maintain adequate brine 100.0
inventory
Detection and .correction of 80.9
leaks, mechanical seals
Hone
Emission Ratio
-------
Table V-3. Benzene and Total VOC Controlled Emissions for Model-Plant III'
Source
Designation
Source (Fig. III-2)
Compressor lube-oil vents G
Single compressor trains
Dual compressor trains
Other normal process emissions A,B,D,F
Intermittent process emissions E
Single compressor trains
At 98* flare efficiency0
At 90% flare efficiency0
Dual compressor trains
At 981 flare efficiency
At 90* flare efficiency0
Storage tanks C
Salt -do me storage H
Fugitive
Secondary K
Control Device
or Technique
None
None
Flares
Floating-roof tanks
Maintain adequate brine .
inventory
Detection and correction of
leaks, mechanical seals
None
VOC
Emission
Reduction (%)
98.0
90.0
98.0
90.0
85.0
100.0
80.9
Emission Ratio
-------
Table V-4. Benzene and Total VOC Controlled Emissions for Model-Plant IVC
Source
Designation
Source (Fig. III-2)
Compressor lube-oil vents G
Single compressor trains
Dual compressor trains
Other normal process emissions A,B,D,F
Intermittent process emissions E
Single compressor trains
At 98% flare efficiency
At 90% flare efficiency
Dual compressor trains
At 98% flare efficiency0
At 90% flare efficiency0
Storage tanks C
Salt-dome storage H
Fugitive
Secondary K
Control Device
or Technique
None
Nona
Flares
Floating-roof tanks
Maintain adequate brine
inventory
Detection and correction of
leaks, mechanical seals
None
VOC
Emission
Reduction (%)
98.0
90.0
98.0
90.0
85.0
100.0
80.9
Emission Ratio
(g/Mg)
Benzene
0.159
0.318
2.3
11.6
0.30
1.5
2.39
7.0
8.9
Total VOC
19.0
39.7
39.6
156
782
32
162
16.8
0
943
23
Emission
Benzene
0.00298
0.00576
0.04
0.21
0.005
0.027
0.044
0.127
0.16
Rate (kq/hr)
Total VOC
0.36
0.72
0.72
2.8
14.2
0.6
2.9
0.30
0
17.1
0.42
8Feed, ethane; ethylene capacity, 226.7 Gg/yr.
bg of benzene or total VOC per Mg of ethylene produced.
CFlare efficiencies have not been satisfactorily documented except for specific designs and operating conditions using specific fuels. Efficiencies
are for tentative comparison purposes.
used
-------
Table V-5. Benzene and Total VDC Controlled Emissions for Model-Plant
Source
Designation
Source (fid. III-2)
Compressor lube-oil vents G
Single compressor trains
Dual compressor trains
Other normal process emissions A,B,D,F
Intermittent process emissions E
Single compressor trains
At 98% flare efficiency0
At 90* flare efficiency0
Dual compressor trains
At 98* flare efficiency0
At 90* flare efficiency"
Storage tanks c
Salt-dome storage H
Fugitive
Secondary K
Control Device
or Technique
None
None
Flares
Floating-roof tanks
Maintain adequate brine
inventory
Detection and correction of
leaks, mechanical seals
None
VOC
Emission
Reduction (*)
98.0
90.0
98.0
90.0
85.0
100.0
80.9
Emission Ratio
(g/Mg)
Benzene
0.0740
0.1480
2.3
11.6
0.30
1.5
2.34
3.3
8.8
Total VOC
9.25
18.50
39.6
156
782
32
162
16.5
0
440
23
Emission
Benzene
0.00288
0.00576
0.09
0.45
0.012
0.058
0.092
0.127
0.34
Rate (kg/hr)
Total VOC
0.36
0.72
1.54
6.1
38.4
1.3
6.3
0.65 ethane; ethylene capacity, 226.7 Gg/yr.
bg of benzene or total VOC per Mg of ethylene produced.
°Flare efficiencies have not been satisfactorily documented except for specific designs and operating conditions using specific fuels. Efficiencies used
are for tentative comparison purposes.
-------
Table V-6. Benzene and Total VOC Controlled Emissions for Model-Plant VI
Source
Designation
Source (Fig. 111-2}
Compressor lube-oil vents G
Single compressor trains
Dual compressor trains
Other normal process emissions A,B,D,F
Intermittent process emissions E
Single compressor trains
At 98% flare efficiency0
At 901 flare efficiency0
Dual compressor trains
At 98% flare efficiency0
At 90% flare efficiency0
Storage tanks C
Salt -dome storage H
Fugitive
Secondary K
Control Device
or Technique
None
None
Flares
Floating-roof tanks
Maintain adequate brine
inventory
Detection and correction of
leaks, mechanical seals
None
VOC
Emission
Reduction (%)
98.0
90.0
98. 0
90.0
85.0
100.0
80.9
Emission Ratio
(g/Mg)
Benzene
0.193
0.386
14.7
73.3
1.9
9.4
21.5
8.5
15.5
Total VOC
5.79
11.60
.72.0
234
1169
42
211
160.0
0
277
40
Emission
Benzene
0.0120
0.0240
0.91
4.56
0.12
0.58
1.34
0.53
0.96
Rate (kg/hr)
Total VOC
0.36
0.72
4.47
14.5
72.6
2.6
13.1
9.9 f
0 ~J
17.2
2.47
aFeed, ethane; ethylene capacity, 226.7 Gg/yr.
bg of benzene or total VOC per Mg of ethylene produced.
cFlare efficiencies have not been satisfactorily documented except for specific designs and operating conditions using specific fuels. Efficiencies
are for tentative comparison purposes.
used
-------
Table V-7. Benzene and Total VOC Controlled Emissions for Itodel-Plant VIIC
Source
Designation
Source (Fig. III-2)
Compressor lube-oil vents G
Single compressor trains
Dual compressor trains
Other normal process emissions A,B,D,F
Intermittent process emissions E
Single compressor trains
At 98% flare efficiency0
At 90% flare efficiency0
Dual compressor trains
At 98* flare efficiency
At 90% flare efficiency0
Storage tanks c
Salt-done storage H
Fugitive
Secondary K
Control Device
or Technique
None
None
Flares
Floating-roof tanks
Maintain adequate brine
inventory
Detection and correction of
leaks, mechanical seals
None
VOC
Emission
Reduction (%)
98.0
90.0
98.0
90.0
85.0
100.0
80.9
Emission Ratio
(g/Mg)
Benzene
0.206
0.412
17.0
85.1
2.2
10.9
22.7
9.1
28.0
Total VOC
5.79
11.60
88.8
259
1297
46
228
160
0
281
73
Emission
Benzene
0.0128
0.0256
1.06
5.29
0.14
0.68
1.41
0.56
1.74
Rate (kg/hr)
Total VOC
0.36
0.72
5.51
16.1
80.6
2.8
14.2
9.9 *?
0 °°
17.4
4.52
3Feed, ethane; ethylene capacity, 226.7 Gg/yr.
bg of benzene or total VOC per Mg of ethylene produced.
CFlare efficiencies have not been satisfactorily documented except for specific designs and operating conditions using specific fuels. Efficiencies used
are for tentative comparison purposes.
-------
Table V-8. Benzene and Total VOC Controlled Emissions for Model-Plant VIII
Source
Designation
Source (Fig. III-2)
Compressor lube-oil vents G
Single compressor trains
Dual compressor trains
Other normal process emissions A,B,D,F
Intermittent process emissions E
Single compressor trains
At 98% flare efficiency0
At 90% flare efficiency0
Dual compressor trains
At 98% flare efficiency
At 90% flare efficiency0
Storage tanks C
Salt-dome storage H
Fugitive
Secondary K
Control Device
or Technique
None
None
Flares
Floating-roof tanks
Maintain adequate brine
inventory
Detection and correction of
leaks, mechanical seals
None
VOC
Emission
Reduction (%)
98.0
90.0
98.0
90.0
85.0
100.0
80.9
Emission Ratio
(g/Mg)
Benzene
0.200
0.400
15.8
79.2
2.0
10.1
22.1.
8.8
21.7
Total VOC
5.79
11.60
.80.4
246
1231
44
220
160.2
0
279
56
Emission
Benzene
0.0124
0.0248
0.93
4.92
0.13
0.63
1.38
0.55
1.35
Rate (kg/hr)
Total VOC
0.36
0.72
4.99
15.3
76.5
2.7
13.6
9.9
0
17.3
3.49
<
1
VD
aFeed, ethane; ethylene capacity. 226.7 Gg/yr.
bg of benzene or total VOC per Hg of ethylene produced.
°Flare efficiencies have not been satisfactorily documented except for specific designs and operating conditions using specific fuels. Efficiencies used
are for tentative comparison purposes.
-------
Table V-9. Benzene and Total VOC Controlled Emissions for Model-Plant IXC
Source
Designation
Source (Fig. III-2)
Compressor lube-oil vents G
Single compressor trains
Dual compressor trains
Other normal process emissions A,B,D,F
Intermittent process emissions E
Single compressor trains
At 98% flare efficiency6
At 90% flare efficiency0
Dual compressor tfains
At 98% flare efficiency
At 90% flare efficiency0
Storage tanks C
Salt -dome storage H
Fugitive
Secondary K
Control Device
or Technique
None
None
Flares
Floating-roof tanks
Maintain adequate brine
inventory
Detection and correction of
leaks, mechanical seals
Hone
VOC
Emission
Reduction (%)
98.0
90.0
98.0
90.0
85.0
100.0
80.9
Emission
(g/M<
Benzene
0,160
0.320
15.8
79.2
2.0
10.1
21.9.
7.0
21.7
Ratio*1
J)
Total VOC
4.64
9.28
80.4
246
1231
44
220
158
0
223
56
Emission
Benzene
0.0124
0.0248
1.2
6.2
0.16
0.79
1.69
0.55
1.69
Rate (kn/hr)
Total VOC
O.K
0.72
6.24
19.1
95.6
3.4
17.1
12.3
0
17.3
4.36
I
O
aFeed, ethane; ethylene capacity, 226.7 Gg/yr.
bg of benzene or total VOC per Hg of ethylene produced.
cFlare efficiencies have not been satisfactorily documented except for specific designs and operating conditions using specific fuels. Efficiencies
are for tentative comparison purposes.
used
-------
Table V-10. Benzene and Total VOC Controlled Emissions for Model-Plant X£
Source
Designation
Source (Fig. III-2)
Compressor lube-oil vents G
Single compressor trains
Dual compressor trains
Other normal process emissions A,B,D,F
Intermittent process emissions E
Single compressor trains
At 98% flare efficiency
At 90% flare efficiency0
Dual compressor trains
At 98% flare efficiency
At 90% flare efficiency0
Storage tanks c
Salt-dome storage H
Fugitive
Secondary *
VOC
Control Device Emission
or Technique Reduction (%)
None
None
Flares
98.0
90.0
98.0
90.0
Floating-roof tanks 85.0
Maintain adequate brine 100.0
inventory
Detection and correction of BO. 9
leaks, mechanical seals
None
Emission Ratio
(g/Mg)
Benzene
0.266
0.532
15.8
79.2
2.0
10.1
22.3
11.7
21.7
Total VOC
7.73
15.46
80.4
246
1231
44
219
161
0
372
56
Emission
Benzene
0.0124
0.0248
0.74
3.69
0.09
0.47
1.04
0.55
1.01
Rate (kg/hr)
Total VOC
0.36
0.72
3.75
11.5
57.4
2.0
10.2
7.5
° E
17.3
2.61
aFeed. ethane; ethylene capacity, 226.7 Gg/yr.
bg of benzene or total VOC per Mg of ethylene produced.
cFlare efficiencies have not been satisfactorily documented except for specific designs and operating conditions using specific fuels. Efficiencies
are for tentative comparison purposes.
used
-------
V-12
100
o
•H
0]
to
10
-------
V-13
Legend for Fig. V-l
Curve
Emission Source
Feedstock
Control Option
la Intermittent (single compressor trains)
lb Intermittent (single compressor trains)
lc Intermittent (single compressor trains)
Id Intermittent (single compressor trains)
le Intermittent (single compressor trains)
If Intermittent (single compressor trains)
2a Intermittent (single compressor trains)
2b Intermittent (single compressor trains)
2c Intermittent (single compressor trains)
2d Intermittent (single compressor trains)
2& Intermittent (single compressor trains)
2f Intermittent (single compressor trains)
3 Secondary emissions
4 Secondary emissions
5 Fugitive emissions
6 Fugitive emissions
7 Storage tanks
8 Storage tanks
9 Normal process emissions (single
compressor trains)
10 Normal process emissions (single
compressor trains)
50:50 E/P
Ethane
Propane
50:50 E/P
Ethane
Propane
50:50 N/G
Naphtha
Gas oil
50:50 N/G
Naphtha
Gas oil
50:50 E/P
50:50 N/G
50:50 E/P
N/G (all
ratios)
50:50 E/P
50:50 N/G
50:50 E/P
50:50 N/G
Flare (98% removal)
Flare (98% removal)
Flare (98% removal)
Flare (90% removal)
Flare (90% removal)
Flare (90% removal)
Flare (98% removal)
Flare (98% removal)
Flare (98% removal)
Flare (90% removal)
Flare (90% removal)
Flare (90% removal)
None
None
Miscellaneous
Miscellaneous
Floating-roof tanks
Floating-roof tanks
None
None
-------
V-14
10.0
D1
(0
§
-rl
U)
CO
•H
I
0)
0)
N
c
•o
0)
o
VI
id
-p
o
1.0 _
o.i _
0.01
100 200 300 400 -500
Ethylene Capacity (Gg/yr)
600
700
Fig. V-2. Total Controlled Benzene Emissions vs
Plant Capacity for Model Plants I Through X
-------
V-15
Legend for Fig. V-2.
Curve
Emission Source
Feedstock
Control Option
la Intermittent (single compressor trains) 50:50 E/P
lb Intermittent (single compressor trains) Ethane
lc Intermittent (single compressor trains) Propane
Id Intermittent (dual compressor trains) 50:50 E/P
(off scale)
le Intermittent (single compressor trains) 50:50 E/P
If Intermittent (single compressor trains) Ethane
lg Intermittent (single compressor trains) Propane
lh Intermittent (dual compressor trains) 50:50 E/P
2a Intermittent (single compressor trains) 50:50 N/G
2b Intermittent (single compressor trains) Naphtha
2c Intermittent (single compressor trains) Gas oil
2d Intermittent (dual compressor trains) 50:50 N/G
2e Intermittent (single compressor trains) 50:50 N/G
2f Intermittent (single comrpessor trains) Naphtha
2g Intermittent (single compressor trains) Gas oil
2h Intermittent (dual compressor trains) 50:50 N/G
3 Secondary emissions 50:50 E/P
4a Secondary emissions 50:50 N/G
4b Secondary emissions Naphtha
4c Secondary emissions Gas oil
5 Fugitive emissions 50:50 E/P
6 Fugitive emissions 50:50 N/G
7 Storage tanks 50:50 E/P
8 Storage tanks 50:50 N/G
9 Normal process emissions (below scale) 50:50 E/P
10 Normal process emissions (below scale) 50:50 N/G
Flare (98% efficiency)
Flare (98% efficiency)
Flare (98% efficiency)
Flare (98% efficiency)
Flare (90% efficiency)
Flare (90% efficiency)
Flare (90% efficiency)
Flare (90% efficiency)
Flare (98% efficiency)
Flare (98% efficiency)
Flare (98% efficiency)
Flare (98% efficiency)
Flare (90% efficiency)
Flare (90% efficiency)
Flare (90% efficiency)
Flare (90% efficiency)
None
None
None
None
Miscellaneous
Miscellaneous
Floating-roof tanks
Floating-roof tanks
None
None
-------
V-16
efficiencies used (90 and 98%) are for tentative comparison purposes. A detailed
discussion of flares is presented in a separate EPA report.
3. Fugitive Sources
Fugitive emissions and controls for the entire synthetic organic chemical industry
7
are covered in a separate report. Controlled fugitive emissions from valves,
pumps, and compressors are based on the factors given in Appendix C and are
included in Tables V-l through V-10. These factors are based on the assumption
that major leaks will be detected and repaired.
2
One ethylene manufacturer reports the use of tandem seals (with the space between
seals vented to pump suction) on all pumps in organic service to reduce fugitive
losses. Data on the resulting reduction in emissions and the cost effectiveness
are not currently available.
Emissions from process samples primarily result from purging of the sample lines
and containers. Emissions can be effectively controlled by piping sample purge
gas to the suction of the charge-gas compressor or to an existing combustion
chamber. Continuous sample streams from process analyzers can be controlled with
a similar collection system. A combined removal efficiency (process samples and
continuous analyzers) of 95% is considered to be attainable.
4. Storage Tanks
Q
Storage guidelines for all producers are discussed in a separate EPA report.
Control of storage losses with floating-roof tanks is considered for naphtha and
pyrolysis gasoline only. Emissions from other materials stored at atmospheric
pressure (i.e., gas oil, pyrolysis fuel oil) are extremely low and floating-roof
tanks are not indicated. For processes with naphtha/gas-oil flexibility
(Models VI—X), tanks primarily used for gas oil storage are equipped with float-
ing roofs to permit alternate use for naphtha storage. Excluding ethylene stor-
age (stored in salt domes) other feedstocks and products (i.e., gas liquid
feedstocks, propylene, C compounds) are stored in pressurized storage tanks,
with no significant emissions.
-------
V-17
Storage tank emissions listed in Tables V-l through V-10 are based on the assump-
tion that a contact type of internal floating roof* with secondary seals will
9 10
reduce fixed-roof-tank emissions by 85%. '
5. Salt-Dome Storage
Emissions of ethylene vented from brine (displaced from the salt domes as ethylene
is stored) are believed to be effectively controlled by maintaining a wide separa-
tion between the brine-ethylene interface and the brine discharge piping, with
11—14
the brine being discharged from the bottom of the cavities. Although no
actual emission data are available, it is estimated that emissions are negligible
if a "buffer zone" of at least 25—30 ft is maintained. An alternative control
method would be to flare the vapor released by the discharged brine.
5. Secondary Emissions
Actual emissions from process wastewater may vary significantly from the emis-
sions estimated for the model plants. The recycling of excess quench tower water
as process steam (Stream 17), considered as a basic process feature in the flow
sheets for the model plants, is an effective method of minimizing the quantity of
quench water discharged. Specific data as to the concentrations of VOC and
benzene in discharged quench water and in the blowdown from recycle steam gener-
ators are needed to estimate the reduction in VOC and benzene emissions that can
be accomplished by recycling generated steam. No controls for secondary emissions
are shown for the model plants.
B. OTHER PROCESSES
Data are not currently available for determination of the emission control
requirements for the development processes described in Sect. III.D.
*Consist of internal floating covers or covered floating roofs as defined in API
25-19, 2d ed., 1976 (fixed-roof tanks with internal floating device to reduce
vapor loss).
-------
V-18
C. REFERENCES*
1. R. L. Standifer, IT Enviroscience, Trip Report for Arco Chemical Co.,
Channelview, TX, Aug. 16-17, 1977 (on file at EPA, ESED, Research Triangle Park,
NC).
2. R. L. Standifer, IT Enviroscience, Trip Report for Gulf Oil Chemicals Co.,
Cedar Bayou Olefins Plant, Cedar Bayou, TX, Sept. 13-14, 1977 (on file at EPA,
ESED, Research Triangle Park, NC).
3. Responses to EPA request for information on emissions from ethylene manufacturers
(see Appendix H).
4. Responses to EPA Questionnaires, Air Pollution Control and Cost Study of the
Petrochemical Industry, OMB Approval No. 158 S 72019 (see Appendix H).
5. W. H. Lauderback, "Unique Flare System Retards Smoke," pp. 127, 128 Hydrocarbon
Processing (January 1972).
6. V. Kalcevic, IT Enviroscience, Control Device Evaluation. Flares and the
Use of Emissions as Fuels (in preparation for EPA, ESED.) (Research Triangle
Park, NC) (August 1980).
7. D. G. Erikson and V. Kalcevic, IT Enviroscience, Fugitive Emissions (September
1980) (EPA/ESED report, Research Triangle Park, NC)
8. D. G. Erikson, IT Enviroscience, Storage and Handling (September 1980) (EPA/ESED
report, Research Triangle Park, NC).
9. C. C. Mosser, "Storage of Petroleum Liquids," Sect. 4.3 in Compilation of Air
Pollutant Emission Factors, 3d ed., part A, AP-42, EPA (April 1977).
10. W. T. Moody, TRW Inc., letter dated Aug. 15, 1979, to D. A. Beck, EPA.
11. J. P. Walsh, Exxon Chemical Co., letter dated Feb. 26, 1979, to EPA with infor-
mation on ethylene process at Baton Rouge, LA, in response to EPA request for
comments on ethylene draft report.
12. R. J. Feldman, C.E. Lummus Co., letter dated Mar. 22, 1979, to EPA responding to
EPA request for comments on ethylene draft report.
13. C. A. Gosoline, CMA, letter dated Jan. 25, 1979, to EPA responding to EPA request
for comments on ethylene draft report.
14. J. A. Mullins, Shell Oil Co., letter dated Jan. 3, 1979, to EPA responding to EPA
request for comments on ethylene draft 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 head-
ing.
-------
VI-1
VI. IMPACT ANALYSIS
A. CONTROL COST IMPACT
This section presents estimated costs and cost-effectiveness data for control
of intermittent emissions of benzene and total VOC by the use of steam-assisted
elevated flares. Details of the model plants are given in Sects. Ill and IV.
The capital and annual costs presented for the process emission controls were
obtained from the control device evaluation report for flares and the use of
emissions as fuel. The procedures
systems are detailed in Appendix D.
emissions as fuel. The procedures used to develop the costs for the control
Capital cost estimates represent the total investment required to purchase and
install a complete flare system as defined in the control device evaluation
report. Specific features of ethylene plant flare systems that are required
because these systems must handle liquids and vapors released at low temperatures
(e.g., vaporizers, additional knockout drums, materials suitable for low tempera-
tures) may increase these costs significantly. These items are considered to
be site-specific.
The bases for the annual cost estimates for the elevated flare systems include
utilities, operating labor, maintenance supplies and labor, capital recovery
charges, and miscellaneous recurring costs such as taxes, insurance, and
administrative overhead. Annual costs are for a 1-year period beginning in
December 1979. The cost factors used to compute annual costs are given in
Table VI-1.
Current Pyrolysis Processes
intermittent Emissions (Vent E)--Intermittent emissions, resulting from the
activation of pressure relief devices, the depressurization and purging of
equipment in preparation for maintenance, and the intentional venting of off-
specification products generated during abnormal conditions, are controlled by
elevated flares, with steam injection used to obtain smokeless emissions.
Estimates of emission reductions, capital costs, total operating costs, and
cost-effectiveness ratios for the ten model plants, with single and dual com-
pressor trains, are summarized in Table VI-2. The relationships between
-------
VI-2
Table VI-1. Annual Cost Parameters
Operating factor
Operating labor
Fixed costs
Maintenance labor plus
materials, 6%
Capital recovery, 18%
Taxes, insurances,
administration charges, 5%
Utilities
Process water
Electric power
Steam
Natural gas
Heat recovery credits
(equivalent to natural gas)
8760 hr/yrc
$15/man-hr
29% of installed capital cost
$0.07/m ($0.25/thousand gal)
$8.33/GJ ($0.03/kWh)
$5.50/Mg ($2.50/thousand Ib or
million Btu)
$1.90/GJ ($2.00/thousand ft3 or
million Btu)
$1.90/GJ ($2.00/million Btu)
Process downtime is normally expected to range from 5 to 15%. If the hourly
rate remains constant, the annual production and annual VOC emissions will be
correspondingly reduced. Control devices will usually operate on the same
cycle as the process. From the standpoint of cost-effectiveness calculations,
the error introduced by assuming continuous operation is negligible.
Based on 10-year life and 12% interest.
-------
Table Vl-2. Cost and Cost-Effectiveness Summary for Model-Plant Flares
Model
Plant
I
I
II
II
III
III
IV
IV
V
V
VI
VI
VII
VII
VIII
VIII
IX
IX
X
X
Compressor
Configuration
Single
Dual
Single
Dual
Single
Dual
Single
Dual
Single
Dual
Single
Dual
Single
Dual
Single
Dual
Single
Dual
Single
Dual
Maxima
Snokeless
Flaring Rate
(M U>/hr)
46.9
46.9
71.4
71.4
59.1
59.1
41.4
41.4
68.7
88.7
214.7
214.7
233.9
233.9
224.3
224.3
280.4
280.4
168.2
168.2
Installed
. Capital.
Cost
(MS)
175
175
239
239
201
201
154
154
282
282
572
572
617
617
595
595
731
731
453
453
Average VOC
Discharge
Rate
)
9,405
70,594
1,626
12,290
2,769
20,550
3,026
22,740
2,697
19,790
563
4.142
525.6
3,820 <
543.7 V
3.972 W
540.1
3,930
544.5
3,997
-------
VI-4
production capacity and capital costs, total operating costs, and cost-effec-
tiveness ratios are shown in Figs. VI-1 through VI-4.
Flare capital and operating costs for the specific models were obtained direct-
ly from the appropriate graphs and equations in the control device evaluation
report on flares and the use of emissions as fuels.
for each model were based on the following criteria:
report on flares and the use of emissions as fuels. The flare requirements
1. The flare is sized to smokelessly combust the maximum quantities of inter-
mittent emissions vented during normal startup and shutdown operations.
The maximum flare nonsmokeless capacity available for severe upset situa-
tions is much greater than the smokeless capacity. The relationship
between smokeless capacity and maximum capacity at a pressure drop of
4.47 kPa (18 in. HO) is given in Fig. IV-1 of the cited control device
evaluation report.
2. The maximum quantity of intermittent emissions anticipated during normal
startup and shutdown of a naphtha/gas-oil ethylene plant with a capacity
of 589.7 Gg/yr (1300 million Ib/yr) of ethylene is 110.2 Mg/hr
(243,000 lb/hr).2
3. The maximum quantities of intermittent emissions anticipated for all model
plants are proportional to the respective compressor charge-gas rates (see
Table IV-8). Sample calculations of estimates of capital costs, operating
costs, and cost-effectiveness ratios are given in Appendix D.
b. Normal Process Emissions—The compressor lubricating-oil vents (vent G) are the
only normal process sources of benzene and VOC emissions for which controls
were considered. (The control of SO , H S, and particulate emissions is not
included in this report.) Emissions from the lubricating-oil vents can be
effectively controlled by routing the vents to the charge-gas compressor suc-
tion. However, because the lubricating-oil vents are normally relatively minor
sources of emissions and the cost of control is very site-specific, controls
for the lubricating-oil vents are not specified for the model plants.
-------
VI-5
r-
en
o
o>
Q
o
o
o
X
8
-p
•H
u
H
100
100
200 300 400 500
Ethylene Capacity (Gg/yr)
600
700
Fig. VI-1. Elevated Flare System Capital Costs
-------
VI-6
300
:§
o
X.
"' a
r-t
ps-
i *w
•P
• 0)
Cn
.5
0)
to
200 —
80 —
60 _
50 —
40
30
100 200 300 400 500
Ethylene Capacity (Gg/yr)
600
700
Fig. VI-2. Elevated-Flare-System Gross Annual Operating Cost
-------
VI-7
Legend for Fig. Vl-2,
Curve
1
2
3
4
5
6
7
8
9
10
11
Feedstock
Propane
Propane
50:50 E/P
50:50 E/P
Ethane
Ethane
Gas oil
50:50 N/G
50:50 N/G
Naphtha
Naphtha
Process Configuration
Single compressor trains
Dual compressor trains
Single compressor trains
Dual compressor trains
Single compressor trains
Dual compressor trains
Single compressor trains
Single compressor trains
Dual compressor trains
Single compressor trains
Dual compessor trains
Emission Source
Intermittent
Intermittent
Intermittent
Intermittent
Intermittent
Intermittent
Intermittent
Intermittent
Intermittent
Intermittent
Intermittent
Control
Flare
Flare
Flare
Flare
Flare
Flare
Flare
Flare
Flare
Flare
Flare
-------
VI-8
O
03
H
t,
"4-1
O
(0
to
•H
tJ
.8
<4-l
U
0
O
220
210
200
190
14-1
O
I
£ 180
170
160
50
40
30
100 200 300 400 500
Ethylene Capacity (Gg/yr)
600
700
Fig. VI-3. Cost Effectiveness of Flares vs Plant Capacity for
Control of Intermittent VOC Emissions from Model Plants I Through X
-------
VI-9
Legend for Fig. VI-3.
Curve
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Feedstock
Propane
Propane
50:50 E/P
Ethane
50:50 E/P
Ethane
Propane
Propane
50:50 E/P
Ethane
50:50 E/P
Ethane
N/G (all ratios)
N/G (all ratios)
N/G (all ratios)
N/G (all ratios)
Compressor Trains
Dual
Dual
Dual
Dual
Dual
Dual
Single
Single
Single
Single
Single
Single
Dual
Dual
Single
Single
Flare Efficiency (*)
90
98
90
90
98
98
90
98
90
90
98
98
90
98
90
98
-------
VI-10
100,000
10,000
0)
G
0)
N
c
en
E
w
(1)
0)
in
0)
tJ 1,000
0)
w
•4J
w
8
100
1
I
,20,22
21,19
17
1445
18,15
I
i
I
100
200 300 400 500
Ethylene Capacity (Gg/yr)
600
700
Fig. VI-4. Cost Effectiveness of Flares vs Plant Capacity for
Control of Benzene Emissions from Model Plants I Through X
-------
VI-11
Legend for Pig. VI-4.
Curve
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Feedstock
50:50 E/P
Ethane
Propane
50:50 E/P
Ethane
Propane
50:50 E/P
Ethane
Propane
50:50 E/P
Ethane
Propane
50:50 N/G
Naphtha
Gas oil
50:50 N/G
Naphtha
Gas oil
50:50 N/G
Naphtha
Gas oil
50:50 N/G
Naphtha
Gas oil
Compressor Trains
Single
Single
Single
Single
Single
Single
Dual
Dual
Dual
Dual
Dual
Dual
S ing le
Single
Single
Single
Single
Single
Dual
Dual
Dual
Dual
Dual
Dual
Flare Efficiency (%)
98
98
98
90
90
90
98
98
98
90
90
90
98
98
98
90
90
90
98
98
96
90
90
90
-------
VI-12
c. Storage Tank Emissions—Emissions of VOC and benzene from pyrolysis gasoline
storage (all model plants) and emissions of VOC from naphtha storage
(models VI—X only) are controlled by the use of floating-roof tanks.
Installed capital cost, net annual cost, and cost-effectiveness data for retro-
fitting the Model-Plant fixed-roof tanks and the corresponding incremental
costs of new internal-floating-roof tanks (based on the capital cost of new
internal-floating-roof tanks minus capital cost of new fixed-roof tanks) are
3
given in a separate EPA report.
d. Salt-Dome Storage Emissions—Emissions of VOC, which occur when ethylene
absorbed in brine is released, are controlled by maintaining a wide separation
between the brine-ethylene interface and the brine discharge piping, thus
preventing absorption of ethylene in the brine that is discharged. No specific
costs are involved in this control technique.
e. Fugitive Sources—Control emission factors for fugitive sources are described
in Appendix C. A separate EPA report covers fugitive emissions and their
applicable controls for the synthetic organic chemicals manufacturing indus-
try. Capital requirements for controls for process samples and analyzer vents
(see Sect. V-3) have not been determined. Cost estimates and cost-effective-
ness ratios are not included.
f. Secondary Sources—No control systems are defined for secondary emissions from
the model plants. Secondary sources and their controls are discussed in a
separate EPA report.
2. Other Processes
No data are available for determining the cost of any control devices required
to control emissions from the alternative processes described in Sect. III-D.
B. ENVIRONMENTAL AND ENERGY IMPACTS
1. Current Pyrolysis Processes
Tables VI-3 and VI-4 show the environmental impacts of reducing benzene and VOC
emissions by the application of the described control systems to Model-
Plant III (50:50 E/P feed, 226.8-Gg/yr ethylene) and to Model-Plant VIII
-------
Table VI-3. Environmental Impact of Controlled Model-Plant III
(50:50 E/P Feed; 226.8-Gg/yr Ethylene) with Single and Dual Compressor Trains
Source
Normal process emissions
Intermittent process emissions
Single compressor trains
Dual compressor trains
Storage tanks
Salt-dome storage
Stream
Designation Control Device
(Fig. Ill -3) or Technique
A,B,D,F,G None
E Elevated flare
C Internal floating roofs
H Maintain adequate brine
Emission
Reduction
(%)
98a
90a
98a
90a
85
100
Emission
Reduction (Mg/yr)
Benzene
25.8
23.7
3.3
3.1
3.0
Total VOC
1738
1596.
361
331
21.5
62.1
Fugitive
inventory
Detection and correction of
major leaks
81
Secondary K None
Total with single compressor trains and 98% flare efficiency
Total with single compressor trains and 90% flare efficiency
Total with dual compressor grains and 98% flare efficiency
Total with dual compressor trains and 90% flare efficiency
4.7
33.5
31.4
11.0
10.8
w
633
2455
2313
1078
1047
aFlare efficiencies have not been satisfactorily documented except for specific designs and operating conditions
using specific fuels. Efficiencies used are for tentative comparison purposes.
-------
Table VI-4. Environmental Impact of Controlled Model-Plant VIII
(50:50 N/G Feed; 544.2-Gg/yr Ethylene) with single and Dual Compressor Trains
Stream
Designation
Source (Fig. III-2)
Normal process emissions A,B,D,F,G
Intermittent process emissions E
Single compressor trains
Dual compressor trains
Storage tanks C
Salt-dome storage H
Fugitive
Secondary K
Emission
Control Device Reduction
or Technique (%)
None
Elevated flare
a
90
903
Internal floating-roofs 85
Maintain adequate brine 100
inventory
Detection and correction of 81
major leaks
None
Total with single compressor trains and 98% flare efficiency
Total with single compressor trains and 90% flare efficiency
Total with dual compressor trains and 98% flare
Total with dual compressor trains and 90% flare
efficiency
efficiency
Emission
Reduction (Mg/yr)
Benzene Total VOC
455 6923
420 6387
58.4 1216
-54.0 1120
70.3 494
149
20.6 646
546 8212
511 7676
149 2505
145 2409
aFlare efficiencies have not been satisfactorily documented except for specific designs and operating conditions
using specific fuels. Efficiencies used are for tentative comparison purposes.
-------
VI-15
(50:50 N/G feed, 544.2-Gg/yr ethylene) with both single and dual compressor
trains. Comparable information for other combinations of plant capacities,
feedstocks, and process variations can be determined from the information
presented in previous sections and from the appendices. Total energy consump-
tion for a typical ethylene plant (excluding fuel value of products) is about
40 MJ per kg of ethylene produced.
Elevated Flare System—Because flare efficiencies have not been satisfactorily
documented except for specific designs and operating conditions using specific
fuels, the environmental impacts of flare systems with efficiencies of both 98%
and 90% are included for tentative comparison purposes. With a flare effi-
ciency of 98%, intermittent emissions of VOC from Model-Plant III are reduced
by 1738 Mg/yr from plants with single compressor trains and by 361 Mg/yr from
plants with dual compressor trains, with corresponding reductions in benzene
emissions of 25.8 Mg/yr and 3.3 Mg/yr respectively. At a flare efficiency of
90%, VOC emissions are reduced by 1596 Mg/yr with single compressor trains and
by 331 Mg/yr with dual trains. The corresponding reductions in benzene emis-
sions are 23.7 Mg/yr and 3.1 Mg/yr, respectively.
The flare system energy requirements for Model-Plant III with single compressor
trains are 3570 GJ/yr, which includes steam usage and the fuel gas required for
the pilots and for purging.
At a flare efficiency of 98%, intermittent emissions of VOC from Model-
plant VIII are reduced by 6923 Mg/yr with single compressor trains and by
1216 Mg/yr with dual trains. The corresponding reductions in benzene emissions
are 455 Mg/yr and 58.4 Mg/yr, respectively. With a flare efficiency of 90%,
VOC emissions are reduced by 6387 Mg/yr with single compressor trains and by
1120 Mg/yr with dual trains. The corresponding reductions in benzene emissions
are 420 Mg/yr and 54.0 Mg/yr, respectively. Energy requirements for the flare
system for Model-Plant VIII with single compressor trains are 18,500 GJ/yr.
controls for Other Emission Sources (Storage Tanks, Salt-Dome Storage, Fugitive)
Control methods for these sources are floating-roof storage tanks, leak correc-
tions for fugitive sources, and maintenance of adequate brine levels for salt-
-------
VI-16
dome storage emissions. Application of these systems or methods result in VOC
emission reductions of 717 Mg/yr and benzene emission reductions of 7.7 Mg/yr
for Model-Plant III; in Model-Plant VIII the VOC emission reduction is 1289 Mg/
yr and the benzene emission reduction is 91 Mg/yr. These control methods do
not consume energy and have no adverse environmental or energy impacts.
No additional controls are proposed for secondary emissions.
2. Other Processes
Emission control systems for the developmental processes described in
Sect. III-D have not been described.
-------
VI-17
C. REFERENCES*
1. V. Kalcevic, IT Enviroscience, Control Device Evaluation Flares and the
Use of Emissions as Fuels (in preparation for EPA/ESED, Reasearch Triangle
Park, NC).
2. R. L. Standifer, IT Enviroscience, Trip Report for Arco Chemical Company,
Channelview, TX, Aug. 16-17, 1977 (on file at EPA/ESED, Research Triangle Park,
NC).
3. D. G. Erikson, IT Enviroscience, Storage and Handling (September 1980) (EPA/
ESED report, Research Triangle Park, NC).
4. D. G. Erikson and V. Kalcevic, IT Enviroscience, Fugitive Emissions (September
I960) (EPA/ESED report, Research Triangle Park, NC).
5. J. J. Cudahy and R. L. Standifer, IT Enviroscience, Secondary Emissions (June
1980) (EPA/ESED report, Research Triangle Park, NC).
6- A. D. Little, Inc., Environmental Considerations of Selected Energy Conserving
Manufacturing Process Options: Volume VI, "Olefins Industry Report," PB264 272
(EPA-600/7-76-034f), U.S. Dept. of Commerce (December 1976) (available from the
National Information Service, Springfield, VA).
*Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of that
paragraph, that reference number is indicated on the material involved. When
the reference appears on a heading, it refers to all the text covered by that
heading.
-------
VI I-1
VII. SUMMARY
Ethylene is produced in the United States predominately by the pyrolysis of
natural-gas concentrates and refinery off-gas (primarily ethane and propane)
and by the pyrolysis of petroleum liquids (primarily naphthas and atmospheric
gas oils). Although ethylene produced from natural-gas concentrates and refinery
off-gas accounted for about 70% of total ethylene in 1976 and production from
these feedstocks is not expected to decrease significantly before 1982, almost
all new capacity after 1979 will use the heavier petroleum liquid feedstocks.
The annual growth rate in ethylene production is estimated to be 4 to 4.5%
through 1984; however, the development of new-production capacity is expected
to stay well ahead of demand, with projected production at only 75% of capacity
by 1981.1
Emission sources and control levels for Model-Plant III (feed, 50:50 ethane-
propane,- ethylene capacity, 226.8 Gg/yr) and Model-Plant VIII (feed, 50:50
naphtha—gas oil; ethylene capacity, 544.2 Gg/yr) are summarized in Tables
VII-1 and VII-2 and are based on the use of single compressor trains with
elevated flare systems controlling intermittent process emissions.
Table VII-3 gives a composite emission summary for all models based on the
weighted averages of emissions from individual models. Emissions from indivi-
dual models were weighted according to estimates of the actual industry dis-
tribution of feedstocks and process configurations. Because flare efficiencies
have not been satisfactorily documented except for specific designs and operating
conditions, using specific fuels, emission estimates based on flare efficiencies
of both 98% and 90% are included for tentative comparison purposes.
The current emissions projected for the domestic ethylene industry based on the
estimated degree of control existing in 1980, with an average flare efficiency
of 98%, are 1170 Mg/yr for benzene and 29,200 Mg/yr for total VOC. With an
average flare efficiency of 90% the corresponding emissions are 1500 Mg/yr for
S. A. Cogswell, A. C. Gaessler, and T. A. Gibson, "CEH Marketing Research
Report on Ethylene," pp. 300.5200H--300.52051 in Chemical Economics Handbook,
Stanford Research Institute, Menlo Park, CA (August 1980).
-------
VI I-2
Table VII-1. Emission Summary for Model-Plant III
(Feed, 50:50 E/P; Ethylene Capacity, 226.8 Gg/yr) with Single Compressor Trains
Emission Rate (kg/hr)
Uncontrolled
Emission Source
Normal process emissions
Intermittent process emissions
Flare efficiency 98%
Flare efficiency 90%
Storage and handling
Fugitive
Secondary
Total - 98% flare efficiency
Total - 90% flare efficiency
Table VII-2. Emission
(Feed, 50:50 N/G; Ethylene Capacity,
Benzene
0.0029
3.0
3.0
0.41
0.66
0.23
4.3
4.3
Summary for
544.2 Gg/yr)
VOC
1.4
202.5
202.5
10.0
89.4
0.6
303.9
303.9
Controlled
Benzene
0.0029
0.06
0.30
0.06
0.13
0.23
0.48
0.72
Model- Plant VIII
with Single Compressor
Emission
Uncontrolled
Emission Source
Normal process emissions
Intermittent process emissions
Flare efficiency 98%
Flare efficiency 90%
Storage and handling
Fugitive
Secondary
Total - 98% flare efficiency
Total - 90% flare efficiency
Benzene
0.0124
49.2
49.2
9.2
2.9
1.4
62.7
62.7
VOC
5.4
765.0
765.0
83.2
90.7
3.5
942.4
942.4
Rate (kg/hr)
VOC
1.4
4.1
20.3
0.44
17.1
0.6
23.6
39.8
Trains
Controlled
Benzene
0.0124
0.98
4.92
1.38
0.55
1.35
4.3
8.2
VOC
5.4
15.3
76.5
9.9
17.3
3.5
51.4
112.6
-------
VII-3
Table VII-3. Composite Model-Plant Emission Summary'
Emission Ratio (g/Mg)
Uncontrolled
Emission Source
Normal process emissions
Intermittent emissions
Storage
Fugitive
Secondary
Total
Benzene
0.2
325.9
66.4
31.3
14.0
437.8
VOC
69
7,247
755
2,677
36
10,784
Controlled
Benzene
0.2
6.5°
d
32.6
10.0
6.0
14.0
36. 7°
rl
62. 8a
VOC
69
145C
d
725
72
572
36
834°
d
1414
aBased on the weighted averages of the emissions from individual model plants.
Emissions from individual models were weighted according to the following
criteria (estimated to approximate actual industry distribution):
Feedstock distribution
Ethane/propane: 60% (37% ethane, 23% propane)
Naphtha/gas oil: 40%
Distribution of compressor configurations
E/P processes: 50% with single trains, 50% with dual or multiple trains
N/G processes: 90% with single trains, 10% with dual or multiple trains
g of benzene or total VOC per Mg of ethylene produced.
°With a flare efficiency of 98%.
With a flare efficiency of 90%.
-------
VII-4
benzene and 36,400 Mg/yr for total VOC. These emission estimates are based on
engineering judgment and data from individual ethylene producers, state emis-
sion control agencies, and the open literature. Individual estimated projec-
tions are shown in Table VII-4.
The predominant emission sources are intermittent process emissions and fugi-
tive emissions from both the gas-concentrate and petroleum-liquid feedstock
processes and storage tank emissions from petroleum-liquid feedstock processes.
Intermittent process emissions can be effectively controlled by flares.
Because these emissions are relatively infrequent, are of short-term duration,
and occur at extremely high and variable rates, other control methods are not
generally applicable. The installed cost of a steam-assisted elevated flare
system for Model-Plant III is $200,000 and is $595,000 for Model-Plant VIII,
based on information on general flare costs presented in a separate EPA
report.2 Specific features of ethylene plant flare systems may increase these
costs significantly.
The corresponding cost-effectiveness ratios of steam-assisted elevated flare
systems for Model-Plants III and VIII, with single and dual compressor trains
and with flare efficiencies of 93% and 90%, are summarized in Table VII-5.
Emissions of benzene and VOC from atmospheric storage tanks can be effectively
controlled by using internal-floating-roof tanks. The emission reduction would
be 85%. Emissions resulting from the displacement of brine from salt-dome
storage can be controlled by maintaining adequate separation of the ethylene-
brine interface from the brine inlet line.
2V. Kalcevic, IT Enviroscience, Control Device Evaluation. Flares and the Use
Of Emissions as Fuels (in preparation for EPA, ESED, Research Triangle Park,
NC).
-------
VII-5
Table VII-4. Estimated Emissions for the Industry
Source
Normal process
Intermittent (98% flare efficiency)
Intermittent (90% flare efficiency)
Storage
Fugitive
Secondary
Total (98% flare efficiency)
Total (90% flare efficiency)
1980 Emissions
Benzene
2
530
855
294
153
194
1173
1498
(Mg/yr)
Total VOC
960
11,800
19,000
2,880
13,040
500
29,180
36,380
Table VII-5. Cost Effectiveness Ratios for Model-Plants III and VIII
Model
Plant
III
III
III
III
VIII
VIII
VIII
VIII
Compressor
Configuration
Single trains
Single trains
Dual trains
Dual trains
Single trains
Single trains
Dual trains
Dual trains
Flare
Efficiency (%)
98
90
98
90
98
90
98
90
Cost Effectiveness (?/Mg)
Benzene
2,540
2,770
18,900
20,600
500
540
3,650
3,970
Total VOC
38
41
175
191
32
35
169
184
-------
APPENDIX A
Table A-l. Physical Properties of Feedstocks and Products'
Molecular
Formula
Feedstocks
Ethane C2H&
Propane ^3^8
n-Butane C4H10
Naphtha
Gas oil
Products
Ethylene C2H4
Propylene C3H6
Butadiene ^"4^6
Pyrolysis gasoline
Light fuel oil
Heavy fuel oil
Molecular
Weight
30.07
44.09
58.12
28.05
42.08
54.09
Liquid Vapor
Density Sp. Grav.
0.546 at -88°C 1.049
0.585 at -44.5°C 1.562
0.579 at 20°C
0.694 at 35°C
0.873 at 35°C
0.566 at -102°C 0.975
0.609 at -47°C 1.498
0.621 at 20°C
0.804 at 41°C
0.972 at 54°C
1.01 at 74°C
Boiling Vapor
Point (°C) Pressure
-88.6
-42.1
-0.5
91C/d 53.8 kPa at 35°Ce
320 3.03 Pa at 35°Ce
-103.7
-47.7
-4.4
43C'd 58.6 kPa at 41°C6
205C/d 80.0 Pa at 54°C6
278C/d 65.5 Pa at 74°C6
Unless otherwise noted, all values are from N. A. Lange, "Physical Properties of Organic Compounds," pp. 366—703
in Handbook of Chemistry, 8th ed., Handbook Publishers, Sandusky, OH, 1952.
Properties of these materials vary. Values shown are typical examples.
Initial boiling point.
(R. L. Standifer, IT Enviroscience, Inc., Trip Report for Arco Chemical Co., Chanhelview, Texas, August 16, 17,
1977 (on file at EPA/ESED, Research Triangle Park, NC).
eR. L. Standifer, IT Enviroscience, Inc., Trip Report for Gulf Oil Chemicals Co., Cedar Bayou Olefins Plant,
Cedar Bayou, Texas, Sept. 13, 14, 1977 (on file at EPA/ESED, Research Triangle Park, NC).
-------
APPENDIX B
Table B-l. Air-Dispersion Parameters for Model-Plant III
(226.8-Gg/yr Ethylene, 50:50 Ethane/Propane Feed), Single Compressor Trains
Source
Ethane /propane pyrolysis
furnace flue gas
Catalyst regeneration heater
flue gas
Catalyst regeneration off -gas --
acetylene conv.
Compressor lube-oil vents
(uncontrolled)
Charge gas
Propylene
Ethylene
Main vent (uncontrolled)
Salt-dome emissions (uncontrolled)
Flares
98% efficiency
90% efficiency
Storage tanks (uncontrolled)
Pyrolysis gasoline
Storage tanks (controlled)
Pyrolysis gasoline
Fugitive emissions (uncontrolled)
Fugitive emissions (controlled)
Secondary emissions (uncontrolled)
No. of
Units
6
1
1
1
1
1
1
1
1
1
1
1
Emission
Rate (g/sec)
VOC
0.02
0.004
0.05
0.075
0.075
56
2.0
1.1
5.6
0.8
0.12
24.8
4.8
0.16
Benzene
0.0008
0.8
0.017
0.083
0.11
0.017
0.18
0.035
0.06
Height
(m)
27.5
15.2
18.3
18.3
18.3
18.3
76.2
a
95
95
9.8
9.8
Dia.
(m)
1.0
0.4
0.4
0.3
0.3
0.3
0.73
a
0.73
0.73
9.5
9.5
Discharge Flow Discharge
Temp. Rate Velocity
(°C) (m /sec) (m/sec)
170 12.3 15
704 1.7 15
510 7.0 60
w
60 *,
60
60
38 64 (max) 152 (max)
Ambient a a
b b b
38b 63.6 (max)b 152 (max)3
38b 63.6 (max)b 152 (max)3
Vented from brine storage pond.
Conditions before combustion.
-------
Table B-2. Air-Dispersion Parameters for Model-Plant VIII
(544.2-Gg/yr Ethylene, 50:50 Naphtha/Gas-Oil Feed), Single Compressor Trains
Source
Naphtha /gas -oil pyrolysis
furnace flue gas
Ethane /propane pyrolysis
furnace flue gas
Gasoline hydrogenation heater
flue gas
Catalyst regeneration heater
flue gas
Catalyst regeneration off-gas
Acetylene converter
Gasoline treatment
C_ converter
Amine stripper vent
Compressor lube-oil vents
(uncontrolled)
Charge gas
Propylene
Ethylene
Main vent (uncontrolled)
Salt-dome emissions (uncontrolled)
No. of
Units
6
1
1
1
1
1
1
1
1
1
1
1
1
Emission
Rate (g/sec)
VOC Benzene
0.1
0.1
0.04
0.01
0.05 0.0008
0.025
0.025
213 13.7
4.7
Height
(m)
40
27.5
39.6
15.2
18.3
18.3
15.2
130
18.3
18.3
18.3
76.2
b
Dia.
On)
2.3
4.4
1.2
0.6
0.6
0.6
0.3
0.1
0.3
0.3
0.3
1.5
b
Discharge
Temp.
171
171
399
704
510
510
510
Ambient
60
60
60
38
Ambient
Flow
Bate
(m /sec)
59
24
10
4
17
16
4
0.065
269 (max)
b
Discharge
Velocity
(m/sec)
14.1
15.9
9.0
14.8
59.1
56.7
51.2
8.3
152 (max)
b
-------
Table B-2. (Continued)
No. of
Source Units
Flares
98% efficiency 1
90% efficiency 1
Storage tanks (uncontrolled)
Naphtha/ gas-oil
(naphtha emissions only) 2
Raw pyrolysis gas 1
Treated pyrolysis gas 2
Storage tanks (controlled)
Naphtha/ gas oil
(naphtha emissions only) 2
Raw pyrolysis gas 1
Treated pyrolysis gas 2
Fugitive emissions (uncontrolled)
Fugitive emissions (controlled)
Secondary emissions (uncontrolled)
Emission
Rate (g/sec)
VOC
4.3
21.3
0.6
4.8
6.7
0.3
0.8
0.6
25.2
4.8
1.0
Benzene
0.3
1.4
0.5
1.0
0.14
0.12
0.79
0.15
0.38
Discharge Flow Discharge
Height Dia. Temp. Rate Velocity
(m) (m) (bC) (ni /sec) (m/sec)
C C C
130 1.5 38° 269C 152C
130 1.5 38° 269° 152C
12.2 33.4 Ambient
14.6 42.7 Ambient
to
12.2 33.0 Ambient u
12.2 33.4 Ambient
14.6 42.7 Ambient
12.2 33.0 Ambient
aVented at top of main-vent flare stack.
Vented from brine storage pond.
n
Conditions before combustion.
-------
C-l
APPENDIX C
FUGITIVE-EMISSION FACTORS*
The Environmental Protection Agency recently completed an extensive testing
program that resulted in updated fugitive-emission factors for petroleum re-
fineries. Other preliminary test results suggest that fugitive emissions from
sources in chemical plants are comparable to fugitive emissions from correspond-
ing sources in petroleum refineries. Therefore the emission factors established
for refineries are used in this report to estimate fugitive emissions from
organic chemical manufacture. These factors are presented below.
Source
Pump seals k
Light-liquid service
Heavy-liquid service
Pipeline valves
Gas/vapor service
Light-liquid service
Heavy-liquid service
Safety/relief valves
Gas/vapor service
Light-liquid service
Heavy-liquid service
Compressor seals
Flanges
Drains
Uncontrolled
Emission Factor
(kg/hr)
0.12
0.02
0.021
0.010
0.0003
0.16
0.006
0.009
0.44
0.00026
0.032
Controlled
Emission Factor
(kg/hr)
0.03
0.02
0.002
0 . 003
o.od'bs
0.061
0.006
0.009
0.11
0.00026
0.019
aBased on monthly inspection of selected equipment; no inspection of
heavy-liquid equipment, flanges, or light-liquid relief valves,-
10,000 ppmv VOC concentration at source defines a leak; and 15 days
allowed for correction of leaks.
bLight liquid means any liquid more volatile than kerosene.
*Radian Corp., Emission Factors and Frequency of Leak Occurrence for Fittings
in Refinery Process Units, EPA 600/2-79-044 (February 1979).
-------
D-l
APPENDIX D
COST ESTIMATE DETAILS
This appendix contains sample calculations showing how the costs presented in
this report were estimated.
The accuracy of an estimate is a function of the degree of data available when
the estimate was made. Figure D-l illustrates this relationship. The contin-
gency allowance indicated is included in the estimated costs to cover the unde-
fined scope of the project.
Capital costs given in this report are based on a screening study, as indicated
by Fig. D-l, based on general design criteria, block flowsheets, approximate
material balances, and data on general equipment requirements. These costs have
an accuracy range of +30% to -23%, depending on the reliability of the data, and
provide an acceptable basis to determine the most cost-effective alternative
within the limits of accuracy indicated.
STEAM-ASSISTED ELEVATED FLARE CONTROLLING INTERMITTENT EMISSIONS FROM MODEL-
PLANT I (ETHANE FEED, 226.8 Gg/yr) WITH SINGLE COMPRESSOR TRAINS
Installed Capital Cost
The model plant flares are sized to smokelessly combust the maximum emissions
vented during normal startup and shutdown operations. For a plant with naphtha
and gas-oil flexibility and producing 589.6-Gg/yr (1.3 billion Ib/yr) ethylene
the required smokeless capacity was reported to be 243,000 Ib/hr. For the model
plants the required smokeless capacities are based on this value but are estimated
to be proportional to the respective compressor charge-gas rates, as given in
Table IV-8 for plants producing 453.5 Gg/yr (1 billion Ib/yr) of ethylene from
various feedstocks. For Model-Plant I (ethane feed, 226.8 Gg/yr) the required
smokeless capacity is determined as follows:
= 46'900 lb/hr-
aModel-Plant I ethylene capacity, Gg/yr.
Ethylene capacity of plant with required flare smokeless capacity of
243,000 lb/hr.
cCharge-gas rate for plant producing 453.5 Gg/yr from ethane.
Charge-gas rate for plant producing 453.5 Gg/yr from 50:50 N/G feed.
-------
ESTIMATE. TYPE
USED BY ESTIMATOR
5CRE.EKJlk»Gi
(PRELIM. E»JG|. STUDY)
PHAjq. DESiqu)
•
•
•
•
\\\
\
MlU. PROS.
cow
\\
\
COST
WITH ALLOWANCE.
. MA*.
\ covr
O > 2 3 4 -fcO -4o -to O 20 4O fcO
APPRO*. COST
EW(^R. 4 E'ST.
(•/- OF TOTAU
CAP. COST")
, - PROBABLE.
ACTUW- PROJECT
COST («M
D
10 ZO
TO /WC.LUDE,
Fig. D-l. Precision of Capital Cost Estimates
-------
D-3
2
From Fig. V-l of the flare report the installed capital cost of the required
elevated flare system is estimated to be $175,000.
Gross Annual Operating Cost
From Table VI-1 of this report the total fixed costs, including capital recovery,
are 29% of the installed capital cost:
$175,000 X 0.29 = $50,800/yr.
2
From Fig. IV- 1 of the flare report the required flare tip diameter is determined
as 16 in., and from IV-4 of the flare report the corresponding natural gas used
for the pilots is 60 scfh and for purging is 155 scfh. From Table VI-1 the cost
of gas is $2.00 per thousand ft3:
(60 + 155) X 8760 X = $3770/yr.
From Sect. IV-A-1 of the flare report it is estimated that 0.3 Ib of steam is
required per pound of emission; from Table VI-1 the cost of steam is $2.50/
thousand Ib. From Table IV-3 of this report the average emission from Model-
Plant I with single compressor trains is 179 kg/hr (395 lb/hr):
2 50
0.3 X 395 X 8760 X = $2600/yr.
The annual cost summary is as follows:
Fixed $50,800
Natural gas 3,800
Steam 2,600
Total $57,200
Cost Effectiveness
Cost effectiveness is the gross annual operating cost $57,200 divided by the
annual VOC or benzene destroyed at 98% or 90% efficiency.* From Tables IV-3 and
V-l of this report the total VOC reduction of intermittent emissions is 1540 Mg/
yr
*Flare efficiencies have not been satisfactorily documented except for specific
designs and operating conditions using specific fuels. Efficiencies used are for
tentative comparison purposes.
-------
D-4
at a flare efficiency of 98% and is 1410 Mg/yr at a efficiency of 90%. The total
benzene destroyed is 6.6 Mg/yr at 98% flare efficiency and is 6.1 Mg/yr at 90% flare
efficiency:
= $37.1/Mg of VOC destroyed (98% flare efficiency).
$5141Q° = $40-5/M9 of voc Destroyed (90% flare efficiency).
= $8600/Mg of benzene destroyed (98% efficiency).
6.6
,200 _
' = $9400/Mg of benzene destroyed (90% efficiency).
6.1
-------
D-5
B. REFERENCES*
1. R. L. Standifer, IT Enviroscience, Trip Report for Arco Chemical Co.,
Channelview, TX, Aug. 16--17, 1977 (on file at EPA/ESED, Research Triangle Park,
NC>.
2. V. Kalcevic, IT Enviroscience, Control Device Evaluation. Flares and the Use of
Emissions as Fuels (in preparation for EPA/ESED, Research Triangle Park, NC).
*Usually, when a reference is located at the end of a paragraph, it refers to the
entire paragraph. If another reference relates to certain portions of that
paragraph, that reference number is indicated on the material involved. When the
reference appears on a heading, it refers to all the text covered by that head-
ing.
-------
E-l
APPENDIX E
INTERMITTENT-EMISSION SAMPLE CALCULATIONS
A. BASE CASE: 50:50 N/G, 453.6-Gg/yr Ethylene
1. Charge-Gas Emissions, Single Compressor Trains
From Table III-l the 5-year average material loss caused by compressor outages,
including hydrogen and methane (total charge gas emissions), for a plant with N/G
flexibility (50:50 N/G assumed), single compressor trains, producing 453.6 Gg/yr
ethylene is
9878 + 6857 + 5469 + 4653 + 3837 = ^ ? Mg/yr = ^ fcg/hr>.
The uncontrolled (nonmethane) VOC emissions are
700.8 X 0.868b = 608.3-kg/hr VOC.
The uncontrolled benzene emissions are
700.8 X 0.0585C = 41.0-kg/hr benzene.
2. Charge-Gas Emissions, Dual Compressor Trains
From Table III-l for the same plant with dual compressor trains the 5-year average
material loss caused by compressor outages (total charge gas emissions) is
1224 + 857 + 735 + 612 + 490 -«,,«/ oo c i /u a
= 783.6 Mg/yr =89.5 kg/hr.
5
The uncontrolled (nonmethane) VOC emissions are
89.5 kg/hr X 0.868 = 77.7 kg/hr.
The uncontrolled benzene emissions are
89.5 kg/hr X 0.0585 = 5.2 kg/hr.
Compressor outage material losses (total charge-gas emissions) for the model plants
are based on these values as shown by the sample calculations for Model-Plant I.
86.8% nonmethane VOC in charge gas.
f»
5.85% benzene in charge gas.
-------
E-2
3. Miscellaneous (Other) Intermittent Emissions (Single and Dual Trains)
For the base case (Table III-l conditions) charge-gas emissions account for 95%
of the total intermittent emissions (including hydrogen and methane) for plants
with single compressor trains. The remaining 5% intermittent emissions (based on
single compressor trains) contain no significant quantities of hydrogen, methane,
or benzene. The following miscellaneous intermittent emissions are assumed to be
the same for single or dual compressor trains:
700.8 X j^|| = 36.9-kg/hr VOC.a
4. Total Intermittent (Nonmethane) VOC Emissions
For single compressor trains the emissions are
608.3 + 36.9 = 645.2 kg/hr.
For dual compressor trains the emissions are
77.7 + 36.9 = 114.6 kg/hr.
B MODEL-PLANT I (ETHANE FEED, 226.8-Gg/yr ETHYLENE) SINGLE COMPRESSOR TRAINS
1. Charge-Gas Emissions
Estimates of charge-gas emissions for the model plants are based on the estimates
developed for the Table III-l conditions and are assumed to be proportional to
the respective quantities of compressor charge gas produced. The charge-gas
quantities and compositions used are given in Table IV-8 and are based on the
feed requirements and yield structures given in Table II-l:
(551.7 Gg of ethane X l.4)b - 0.4 Gq of fuel oil produced e
453.5 Gg of ethylene produced of ethy?ene prod2ced.
For Table III-l conditions (N/G flexibility) and for the 50:50 N/G model plants
(VIII—X), a charge gasrethylene ratio of 3.395 was estimated based on averages
of the naphtha and gas oil value.
Miscellaneous intermittent emissions for the model plants are based on this value,
as shown by the sample calculations for Model-Plant I.
Based on a recycle ratio of 0.4 for ethane (Table IV-8).
-------
E-3
Then the total charge-gas emissions (including hydrogen and methane) for
Model-Plant I with single compressor trains are
700.8a X ^^- X i^- = 175.7 kg/hr.
453.5C 3.395e
The uncontrolled nonmethane VOC emissions are
175.7 X 0.914f = 160.6 kg/hr.
The uncontrolled benzene emissions are
175.7 X 0.0044g = 0.77 kg/hr.
2. Miscellaneous (Other) Intermittent Emissions
Estimates of miscellaneous intermittent emissions for the model plants are based
on the estimates developed for the Table III-l conditions and are assumed to be
proportional to ethylene production. For Model-Plant I the miscellaneous inter-
mittent emissions are
??fi fib
36.9 X = 18.5 kg/hr.
453.6C
3- Total Emissions
The total uncontrolled intermittent (nonmethane) VOC emissions for Model-Plant I
with single compressor trains are as follows:
160.6 + 18.5 = 179.0 kg/hr.
aTotal charge-gas emissions for Table III-l conditions, single trains.
Ethylene production, Model-Plant I.
°Ethylene production, Table III-l conditions.
Charge gas:ethylene ratio for ethane feed.
eCharge gas:ethylene ratio for 50:50 N/G feed (Table III-l conditions).
91.4% nonmethane VOC in charge gas.
"o.44% benzene in charge gas.
-------
E-4
4. Controlled Intermittent Emissions
The nonmethane VOC emissions with 98% flare efficiency are
178.6 X 0.02b = 3.6 kg/hr.
The benzene emissions are
0.77 X 0.02b = 0.015 kg/hr.
The main-vent flare emissions with 90% flare efficiency are
178.6 X 0.10 = 17.9 kg/hr for VOC
and
0.77 X 0.10 = 0.077 kg/hr for benzene.
-------
F-l
APPENDIX F
SALT-DOME STORAGE-EMISSION SAMPLE CALCULATIONS
A. MODEL VIII -- 50:50 N/G FEED, 544.2-Gg/yr ETHYLENE BASIS
Annual throughput, 15 days' ethylene production.
Storage dome conditions, 20°C, 100-atm pressure.
Brine conditions, saturated sodium chloride brine.
I m Throughput of Ethylene
(544.2 Gg/yrX) = 22.4 Gg/yr.
2. Volume of Brine Displaced
Specific volume of ethylene at 20°C, 100 atm:
V°
V=
p M.W.
where
f = Compressibility factor
= 0.360,
M.W. = molecular weight,
V = the volume given in Table 160 of ref . 1
= 22,240,
_ (0.360)(22,240)
(100)(28)
=2.86 cc/g = 2860 m3/Gg.
The volume of brine displaced = (22. 4) (2860) = 64,064 m .
1J. H. Perry, Chemical Engineer's Handbook, 3d ed., pp. 205 208, McGraw-Hill,
New York, 1950.
-------
F-2
3. Solubility of Ethylene in Brine
The solubility of ethylene in HO at 20°C, 100 atm, was calculated as follows:
Pa
Xa H '
where
x = the mole fraction of a in the liquid
dt,
100 = 9.8 X 10"3 (ref. 2),
10,200
P = the partial pressure of component a in the vapor,
3.
H = Henry's law constant (ref. 2).
w , x& . ^ = (9.8 X 10'3) -|ff[ = 15.24 X 10"3 g of C^/g of Kf.
s
The solubility of ethylene in saturated NaCl brine was calculated as follows:
Solubility of NaCl at 20°C =6.1 " ^1 (ref- 2) '
S
K = - log T^ (ref. 3),
where
S = solubility of gas in pure HO (g/1000 g of H
S = solubility of gas in solution (g/1000 g of H
K = 0.134 at 20°C (ref. 4),
_ g equiv of salt.
C ~ 1000 g of H20
2ibid., pp. 673—675.
3N. A. Lange, Handbook of Chemistry, 8th ed., 1952, p. 289.
4Winkler, "Landolt-Bbrnstein Physicalisch-Chemische Tabellen."
-------
F-3
Then
n ,,. 1 _ 15.24
°-134 = n log~r-
and
s = 2.32 X 10"3 g of C^/g of H20.
4. Ethylene Emission
, /-, Mn\ / ^ 9 of C HA
(64064 in ) (:L-|a) (2.32 X 10 - oTHo)= 148-6 Mg/yr'
\m/\ g°2/
The average VOC emission rate = 17.0 kg/hr.
3_ Maximum Emission Rate
Assuming that an average of 15 days' production is moved in and put of storage,
the maximum emission rate is
17.0 kg/hr X = 410 kg/hr.
-------
APPENDIX G
LIST OF EPA INFORMATION SOURCES*
1. R. B. Ruston, EPA questionnaire for Allied Chemical Corporation, Geismar Complex,
Geismar, Louisiana.
2. G. Delodder, EPA questionnaire for Union Carbide Corporation, Texas City Plant,
Texas City, Texas.
3. W. R. Chalker, EPA questionnaire for E. I. du Pont de Nemours and Company, Sabine
River Works, Orange, Texas.
4. H. McNair, EPA questionnaire for Dow Chemical Company, Plant A and Plant B,
Freeport, Texas.
5. D. G. Pringle, EPA questionnaire for Texas Eastman Company, Longview, Texas.
5. C. B. Brantley, EPA questionnaire for Gulf Oil Company, Port Arthur Refinery,
Port Arthur, Texas.
7. H. J.LaBorde, EPA questionnaire for Northern Petrochemical Company, Joliet Plant,
Morris, Illinois.
8. R. L. Maycock, EPA questionnaire for Shell Chemical Company, Houston Chemical
Plant, Deer Park, Texas.
g. H. M. Walker, EPA questionnaire for Monsanto Company, Chocoloate Bayou Plant,
Alvin, Texas.
0. D- W. Smith, letter to EPA from E. I. du Pont de Nemours & Company, February 3, 1978.
l. J. P. Walsh, letter to EPA from Exxon Chemical Co., February 10, 1978.
!2. R- J- Brenner, letter to EPA from Mobil Chemical Co., January 26, 1978.
J3. L. A. McReynolds, letter to EPA from Phillips Petroleum Co., January 27, 1978.
X4. A. G. Smith, letter to EPA from Shell Oil Company, February 22, 1978.
*Sources 1—9 were part of a data-gathering program in the preparation of Survey
Report on Atmospheric Emissions from the Petrochemical Industry, Vol. II, by
J. W. Pervier et al., EPA-450/3-73-005-b (April 1974).
-------
H-l
APPENDIX H
EXISTING PLANT CONSIDERATIONS
EXISTING PLANT CHARACTERIZATION
1 ——8
Table H-l lists emission control devices reported to be in use by industry.
To gather information for the preparation of this report two site visits were
made to manufacturers of ethylene. Trip reports have been cleared with the
companies concerned and are on file at EPA, ESED, in Durham, NC; EPA also has
received letters in response to requests for information on air emissions from
ethylene plants and in response to requests for comments on the draft version of
this report. Some of the pertinent information concerning process emissions from
existing ethylene plants is presented in this appendix.
Arco Chemical Co., Channelview, TX (Lyondell Plant)
Two nearly identical olefin units, designed and engineered by the Luraus Company,
are located at the Lyondell plant. At the time of the site visit (August 1977)
Olefin unit No. 1 had been in operation for less than one year and construction
of Olefin unit No. 2 was nearing completion. Each unit has an annual design
capacity of 1.3 billion Ib of ethylene. Feedstocks for both units are naphtha
and/or gas oil.
The more significant VOC emission control devices include two flares and their
associated equipment and a brine degassing system associated with product storage
in underground salt domes. Possibly of greater significance in the control of
VOC emissions are some of the internal features of the process. Most notable
among these are (1) the recycle of most wastewater to the process, (2) the
generation and effective recycle of steam from process wastewater, (3) the use of
high-capacity centrifugal compressors designed for low emission levels, (4) the
transfer of all products by pipeline,minimizing transfer losses, and (5) the
routing of most relief valves and process vents to the flare system. Table H-2
summarizes Arco's estimates of controlled VOC emissions from the No. 1 olefin
unit. Not included are estimates of uncontrolled intermittent emissions, VOC
emissions discharged from the flare system, or fugitive or secondary emissions.
The steam-assisted elevated flare for the No. 1 olefin unit is designed to smoke-
lessly burn up to 243,000 Ib/hr of hydrocarbons, the maximum anticipated flaring
rate during normal startup and shutdown. The maximum design capacity of the
-------
Table H-l. Control Devices Currently Used by the United States Ethylene industry
Control Devices and Techniques for Various Emission Sources
Company Location
Arco* Channelview, TX
Gulf0 Cedar Bayou, TX
1591 (E/P) unit
1592 (N/G) unit
Mobild Beaumont, TX
Texaco6 Port Neches, TX
Texas-Eastman Longview, TX
Absorption plant
Cryogenic plant
Du Pont Orancie, TX
Phillips1* Svieeny, TX
Shell1 Deer Park, TX
Exxon3 Baton Rouga, LA
"see ref 1.
^Total emissions only 0.6 Ib/hr.
c
See ref 2.
See ref 8.
See ref 3.
See ref 4.
9See ref 5.
hsee ref 6.
See ref 7.
Intermittent
Elevated flare
Elevated flare
Elevated flare
Elevated flare
Elevated flare
Horizontal flare
Horizontal flare
Elevated flare
Elevated flare
Elevated flare
Elevated flare
Lube-Oil Vents Atmospheric Storage Tanks Salt-Dome Storage
High-efficiency seals Floating-roof tanks
Hone Floating-roof tanks
High-efficiency sealsb Floating-roof tanks
Not reported Not reported
Hone Not reported
Mot reported *>*• reported
Mot reported Not reported
Hot reported Floating-roof tanks
Hot reported Not reported
Hot reported Not reported
70% controlled by flaring Not reported
Flare
Storage contracted
Not reported
Not reported
Hot reported
Not reported
Hot reported
Not reported
Not reported ffi
C H storage contracted, M
C H controlled with
25-230 ft brine
"buffer" zone
3See ref 2.
-------
H-3
Table H-2. Estimated Emissions from Arco-Lyondell Plant
Olefin Unit No. 1 (Naphtha/Gas Oil Feed)3
Average VOC Emissions
Source (Ib/hr)
Naphtha storage 15
Pyrolysis gasoline storage 27
Other storage tanks 0.2
Lube-oil vents 0.6
asee ref 1.
-------
H-4
flare (not smokeless) is 2.3 million Ib/hr of hydrocarbons. All process relief
valves except those relieving the demethanizer column are vented to the flare
system. Emissions from the demethanizer relief valves would be primarily hydrogen
and methane.
2
2. Gulf Oil Chemicals Co., Cedar Bayou, TX
Two olefin units are located at this plant. The older of the two units (unit 1591),
which has been in operation since 1963, has a rated ethylene capacity of 400 million
Ib/yr. Feedstocks for this unit are ethane and propane. The newer unit (unit 1592),
designed and engineered by the Lumus Company, was started in 1976 and has a rated
ethylene production capacity of 1.2 billion Ib/yr. Feedstocks are naphtha and/or
gas oil. The primary emission control devices for both units are steam-assisted
elevated flares. Estimated emissions from both units are given in Tables H-3 and
H-4. Not included are estimates of uncontrolled intermittent emissions, VOC
emissions discharged from the flare system, or fugitive or secondary emissions.
9
3. Exxon Chemical Co., Baton Rough, LA
Exxon's Baton Rouge chemical plant (BRCP) has an ethylene production capacity of
695.3 Gg/yr. The BRCP is an older plant that has gone through many modifications,
with new equipment incorporated with older, existing equipment. At the BRCP,
sidestreams are sent to other units and are converted to products such as butadiene,
isobutylene, n-butylene, and isoprene. A detailed list of products is provided
in Table H-5. A simplified description of the BRCP is that it consists of two
single compressor train units. However, the ethylene complex is further broken
down into three gas oil cracking sections, two ethane/propane gas cracking sections,
and two purification sections.
Table H-6 presents a comparison between Exxon's estimates of uncontrolled inter-
mittent and lubricating-oil vent emissions and emission ratios for the BRCP with
those for model-plant VII. The experience at Exxon's BRCP has been that upset
emissions resulting from compressor outages do not decrease after the first
5 years of operation.
4. Texaco, Port Neches, TX
Texaco operates two older ethane/propane feedstock units at Port Neches.
Following are general comments concerning emission sources and controls for these
older units:
-------
H-5
Table H-3. Estimated Emissions from Gulf Oil Chemicals Co-r
Cedar Bayou, TX, 1591 Olefin Unit (Ethlene/Propane Feed)a .
Average VOC Emissions
Source (Ib/hr)
Storage tanks 21
Propylene compressor 18
Cracked-gas compressor "24
Flue gas 3
asee ref 2.
Table H-4. Estimated Emissions from Gulf Oil Chemicals Co.,
Cedar Bayou, TX, 1592 Olefin Unit (Naphtha/Gas Oil Feed)
Average VOC Emissions
Source (Ib/hr)
Flue gas 12
Naphtha storage 0.9
Pyrolysis gasoline storage 31
Other storage tanks 0.2
Lube-oil vents 0.6
aSee ref 2.
-------
H-6
Table H-5. Exxon Baton Rouge Chemical Plant Olefin
Unit Products3
Major Products
Indirect Products
Tar
Aromatic high boilers
Low-pressure distillate
Heartcut distillate
Heavy naphtha
Hydrogen
Tail gas
Ethylene
Propylene
Isobutylene
Butadiene
Isoprene
Other by-product streams
aSee ref 3.
Isopropanol (from propylene)
Petroleum resins
Chlorobutyl rubber
Other rubbers
-------
H-7
Table H-6. Comparison of Benzene and Total VOC
Uncontrolled Emissions from Selected Sources from the
Exxon BRCP with Model-Plant-VII Emissions*1
__ Source
Lube oil vents
Exxon BRCP - precontrolled
Exxon BRCP - after large
volume vents controlled
Model Plant VII , single trains
Intermittent emissions
Exxon BRCP
Model Plant VII, single trains
Emission
Benzene
2.33
0.655
0.206
792
Ratio (g/Mg)b
Total VOC
280
27.6
5.79
14,100
12,300
Emission
Benzene
0.145
0.041
0.0128
49.2
Rate (kg/hr) C
Total VOC
17.4
1.71
0.36
876. ld
765.0
See ref 3.
g of benzene or VOC per Mg of ethylene produced.
=Exxon BRCP emissions prorated to model-plant-VII ethylene capacity.
97% of Exxon BRCP intermittent emissions are flared.
-------
H-8
a. Compressor lubricating-oil vents no buffer gas in seals and the seal oil
is routed to an oil/water collection system, which, in turn, is vented to
the atmosphere through a carbon filter.
b. Furnace stack and decoking gases vented to the atmosphere.
c. Compressor outage emissions vented to the flare.
d. Relief valves the quench tower relief valve is vented to the flare; the
recovery section relief valves are vented to the atmosphere; however, during
normal operation the pressure control on the towers is maintained by venturi
to the flare system.
e. Analyzers are vented to the atmosphere.
B. RETROFITTING CONTROLS
The primary difficulty associated with retrofitting may be in finding space to
fit the control device into the existing plant layout. Because of the costs
associated with this difficulty it may be appreciably more expensive to retrofit
emission control systems in existing plants than to install a control system
during construction of a new plant.
-------
H-9
C. REFERENCES*
1. R. L. Standifer, IT Enviroscience, Trip Report for Arco Chemical Co.,
Channelview, TX, Aug. 16, 17, 1977 (on file at EPA/ESED, Research Triangle Park,
NC).
2. R. L. Standifer, IT Enviroscience, Trip Report for Gulf Oil Chemicals Co.,
Cedar Bayou Oletius Plant, Cedar Bayou, TX, Sept. 13, 14, 1977 (on file at
EPA/ESED, Research Triangle Park, NC).
3. J. F. Cooper, Texaco, Inc., letter dated Feb. 2, 1979, to EPA with information on
ethylene processes at Port Arthur and Port Naches, TX, in response to EPA request
for comments on ethylene draft report.
4. G. Prendergast, Texas Eastman Co., letter dated Feb. 21, 1978, to EPA with infor-
mation on ethylene processes at Longview, TX, in response to EPA request for
information.
5. D. W. Smith, Dupont, letter dated Feb. 3, 1978, to EPA with information on
ethylene processes at Orange, TX, in response to EPA request for information.
6. L. A. McReynolds, Phillips Petroleum Co., letter dated Jan. 27, 1978, to EPA with
information on ethylene processes at Sweeny, TX, in response to EPA request for
information.
7. A. G. Smith, Shell Oil Co., letter dated Feb. 22, 1978, to EPA with information
on ethylene processes at Deer Park, TX, in response to EPA request for informa-
tion.
8. P. B. Mullin, Mobil Chemical Co., letter dated Jan. 26, 1978, to EPA with infor-
mation on ethylene processes at Beaumont, TX, in response to EPA request for
information.
9. J. P. Walsh, Exxon Chemical Co., letter dated Feb. 26, 1979, to EPA with informa-
tion on ethylene processes at Baton Rouge, LA, in response to EPA request for
comments on Ethylene draft 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 head-
ing.
-------
4-i
REPORT 4
ETHYLENE OXIDE
V. Kalcevic
J. F. Lawson
IT Enviroscience
9041 Executive Park Drive
Knoxville, Tennessee 37923
Prepared for
Emission Standards and Engineering Division
Office of Air Quality Planning and Standards
ENVIRONMENTAL PROTECTION AGENCY
Research Triangle Park, North Carolina
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
D25B
-------
4-iii
CONTENTS OF REPORT 4
I- ABBREVIATIONS AND CONVERSION FACTORS I_1
II. INDUSTRY DESCRIPTION II_1
A. Reason for Selection II-l
B. EO Usage and Growth II-l
C. Domestic Producers II-l
D. References jj_7
III- PROCESS DESCRIPTIONS III-l
A. Introduction III-l
B. Oxidation of Ethylene III-l
C. References III-7
IV. EMISSIONS jy.j
A. Air-Oxidation Process iv-l
B. Oxygen-Oxidation Process IV-5
C. References IV-13
V. APPLICABLE CONTROL SYSTEMS v_x
A. Air-Oxidation Process v-1
B. Oxygen-Oxidation Process v_5
C. References V-10
VI. IMPACT ANALYSIS VI-1
A. Environmental and Energy Impacts VI-1
B. Control Cost Impact VI-5
C. Reference VI-16
VII. SUMMARY VII-1
-------
4-v
APPENDICES FOR REPORT 4
Page
A. PHYSICAL PROPERTIES OF ETHYLENE AND ETHYLENE OXIDE A-l
B. AIR-DISPERSION PARAMETERS B-l
C. FUGITIVE-EMISSION FACTORS C-l
D. COST ESTIMATE DETAILS D'l
E. EXISTING PLANT CONSIDERATIONS E-l
-------
4-vii
TABLES OF REPORT 4
Number Page
II-l Ethylene Oxide Usage and Growth II-2
11-2 Ethylene Oxide Capacity 11-3
IV-1 Uncontrolled Emissions for Air-Oxidation Process IV-2
IV-2 Main Process Vent Compositions, Air-Oxidation Process IV-4
IV-3 Stripper Purge Vent Composition, Air-Oxidation Process IV-4
IV-4 Model Plant Storage, Air-Oxidation Process IV-6.
IV-5 Uncontrolled Emissions for Oxygen-Oxidation Process IV-8
IV-6 CO Purge Vent Composition, Oxygen-Oxidation Process IV-8
IV-8 Stripper Purge Vent Composition, Oxygen-Oxidation Process IV-11
IV-7 Argon Purge Vent Composition, Oxygen-Oxidation Process IV-9
IV-9 Model Plant Storage, Oxygen-Oxidation Process IV-11
V-l Controlled Emissions for Air-Oxidation Process V-2
V-2 Controlled Emissions for Oxygen-Oxidation Process V-6
VI-1 Emission Reduction by Control Devices for Air-Oxidation VI-2
Process
VI-2 Emission Reduction by Control Devices for Oxygen-Oxidation VI-4
Process
Vl-3 Cost Factors Used in Computing Annual Costs VI-6
VI-4 Emission Control Cost Estimates for Model Plants Using VI-7
Oxygen-Oxidation or Air-Oxidation Process
Vli-1 Emission Summary, Air-Oxidation Process VII-2
Vll-2 Emission Summary, Oxygen-Oxidation Process VII-2
A-l Properties of Ethylene A~1
A-2 Properties of Ethylene Oxide A~2
B-l Air-Dispersion Parameters for Air-Oxidation Model Plant B-l
B-2 Air-Dispersion Parameters for Oxygen-Oxidation Model Plant B-2
-------
4-ix
TABLES (continued)
Number Page
D-l Emission Flow Data D-3
D-2 Catalytic Oxidizer Control Cost Data for Air-Oxidation D-5
Process
D-3 Thermal Oxidizer Control Cost Data for Air-Oxidation
D-6
Process
D-4 Stripper Purge Vent Control Cost Data for Air-Oxidation D-7
Process
D-5 Stripper Purge Vent Control Cost for Oxygen-Oxidation D-8
Process
E-l Emission Control Devices Used by Some Domestic
Ethylene Oxide Procedures E-2
E-2 Emission Data for Process Vents E-5
-------
4-xi
FIGURES OF REPORT 4
Number
II-l Location of Plants Manufacturing Ethylene Oxide II-4
III-l Flow Diagram for Air-Oxidation Process III-3
III-2 Flow Diagram for Oxygen-Oxidation Process III-5
VI-1 Capital Cost for Emission Control, Air-Oxidation Process VI-8
VI-2 Annual Cost for Emission Control, Air-Oxidaton Process VI-9
VI-3 Cost Effectiveness for Emission Control, Air-Oxidation Process VI-10
VI-4 Capital Cost for Emission Controls for Oxygen-Oxidation VI-13
Process
VI-5 Annual Cost for Emission Controls Oxygen-Oxidation Process VI-14
VI-6 Cost Effectiveness for Emission Control, Oxygen-Oxidation VI-15
Process
D-l Precision of Capital Cost Estimates D_2
-------
1-1
ABBREVIATIONS AND CONVERSION FACTORS
EPA policy is to express all measurements used in agency documents in metric
units. Listed below are the International System of Units (SI) abbreviations
and conversion factors for this report.
To Convert From
Pascal (Pa)
Joule (J)
Degree Celsius (°C)
Meter (m)
Cubic meter (m3)
Cubic meter (m3)
Cubic meter (m3)
Cubic meter/second
(m3/s)
Watt (W)
Meter (m)
Pascal (Pa)
Kilogram (kg)
Joule (J)
To
Atmosphere (760 mm Hg)
British thermal unit (Btu)
Degree Fahrenheit (°F)
Feet (ft)
Cubic feet (ft3)
Barrel (oil) (bbl)
Gallon (U.S. liquid) (gal)
Gallon (U.S. liquid)/min
(gpm)
Horsepower (electric) (hp)
Inch (in.)
Pound-force/inch2 (psi)
Pound-mass (Ib)
Watt-hour (Wh)
Multiply By
9.870 X 10"6
9.480 X 10"4
(°C X 9/5) + 32
3.28
3.531 X 101
6.290
2.643 X 102
1.585 X 104
1.340 X 10"3
3.937 X 101
1.450 X 10"4
2.205
2.778 X 10"4
Standard Conditions
68°F = 20°C
1 atmosphere = 101,325 Pascals
PREFIXES
Prefix
T
G
M
k
m
Symbol
tera
giga
mega
kilo
milli
micro
Multiplication
Factor
1012
109
106
103
10'3
10"6
Example
1
1
1
1
1
1
Tg =
Gg =
Mg =
km =
mV =
ug =
1
1
1
1
1
1
X
X
X
X
X
X
10
10
10
10
10
10
12 grams
9
6
3
"
M
grams
grams
meters
3 volt
6 gram
-------
II-l
II. INDUSTRY DESCRIPTION
A. REASON FOR SELECTION
Ethylene oxide (EO) production was selected as a product for study because of
the indication by preliminary data of relatively high total emissions of volatile
organic compounds (VOC), of the suspicion of harmful health effects caused by
1 2
EO, and of the expected industry growth.
Although EO generally is handled as a liquid, it is a gas at ambient conditions
(see Appendix A for pertinent physical properties of EO). The largest process
emission is unreacted ethylene, the organic raw material for EO production.
B. EO USAGE AND GROWTH
Table II-l shows EO end products and their expected growth rates. The predomi-
nant end use is the production of ethylene glycol, from which antifreeze and
polyethylene terephthalate fibers and films are made.
The domestic annual production capacity for EO on July 1, 1979, was estimated
to be 2783 Gg and the 1978 production was 82% of this capacity. Historically,
the industry has operated at approximately 87% of nameplate capacity. Industry
capacity is projected to be 3366 Gg by the end of 1981, and consumption by 1983
2
is expected to be about 2900 Gg.
A substitute feedstock for producing EO is reported to be under development by
Union Carbide. No information is <
will not be covered in this report.
Union Carbide. No information is available on this development; therefore it
DOMESTIC PRODUCERS
As of July 1, 1978, eleven producers of EO in the United States were operating
16 plants at 14 locations. Table II-2 lists the producers and the processes
being used; Fig. II-l shows the plant locations. Approximately 63% of the present
domestic capacity is produced by air oxidation of ethylene and 37% by oxygen
oxidation of ethylene.
-------
II-2
Table II-l. Ethylene Oxide Usage and Growth
End Use
Ethylene glycol
Diethylene glycol
Triethylene glycol
Polyethylene glycol
Glycol ethers
Ethanolamines
Nonionic surface
active agents
Other
Total
Production for
1978
(%)
58.7
4.6
2.5
2.7
7.8
6.3
12.0
5.4
100
Average Growth
for 1978 — 1983
(%/yr)
5.0 — 6.0
4.0 — 5.0
4.0 — 5.0
4.5 — 5.5
4.5 — 5.5
4.5 — 5.5
4.0 — 5.0
4.5 — 5.5
3Ref 2.
-------
II-3
Table II-2. Ethylene Oxide Capacity
Capacity (Gg)
Company
BASF Wyandotte
Calcasieu
Celanese
Dow
Eastman
ICI
Northern Petrochemical
Olin
PPG
Shell
SunOlin
Texaco
Union Carbide
Total
Location
Geismar, LA
Lake Charles , LA
Clear Lake City, TX
Freeport , TX
Plaquemine, LA
Longview, TX
Bayport, TX
Joliet, IL
Br and enbur g , KY
Beaumont, TX
Geismar , LA
Claymont, DE
Port Neches, TX
Penuelas, PR
Seadrift, TX
Taft, LA
July 1, 1979
155
193
118
204
88
104
50
68
318
45
215
268
417
540
2783
Year-End 1981
218
102°
193
118
204
88
227
104
50
68
318
45
315
290
477
578
3366
Process
Oxygen
Oxygen
Oxygen
Air
Air
Oxygen
Oxygen
Oxygen
Oxygen
Oxygen
Oxygen
Oxygen
Air
Air
Air
Air
See ref 2.
b
Air is listed for plant believed to be using air as the oxidant feed and oxygen is
listed for plants believed to be using purified oxygen as the oxidant feed.
c
Plant was shut down in early 1978 after an explosion and fire; was to have resumed
operation in late 1979.
-------
II-4
1. BASF Wyandotte Corp., Geismar, LA
2. Calcasieu Chemical Corp., Lake
Charles, LA
3. Celanese Chemical Co., Clear Lake
City, TX
4. Dow Chemical Co., Freeport,
5. Dow Chemical Co.,
6. Eastman Kodak Co.
7. Id, Bayport, TX
8. Northern Petrochemical Co., Joliet, IL
TX
Plaquemine, LA
, Longview, TX
9. Olin Corp., Brandenburg, KY
10. PPG Industries, Inc.
Beaumont, TX
11. Shell Chemical Co., Geismar, LA
12. Sunolin Chemical Co., Claymont, DB
13. Texaco, Port Neches, TX
14. Union Carbide Corp,, Ponce, PR
15. Union Carbide Corp., Seadrift, TX
16. Union Carbide Corp., Taft, LA
Fig. II-l. Locations of Plants Manufacturing Ethylene Oxide
-------
II-5
The choice of process (air vs oxygen) is based on the cost of ethylene, on the
cost of energy, and on other considerations. The air-oxidation process is more
adaptable to large units, and results in a lower total investment. The oxygen-
oxidation process results in a high ethylene yield and is adaptable to any unit
4
size.
Producing Companies
1. BASF Wyandotte Corp.
This facility has two production trains, one built in 1957 and one in 1967,
2
with a combined capacity of 155 Gg/yr. The plant built in 1957 was the first
commercial use of the oxygen-oxidation process.
2. Calcasieu Chemical Corp.
The Lake Charles, LA, facility expanded its annual capacity from 74,800 Mg to
104,000 Mg in 1976. A fire in early 1978 shut down this facility; it was to
have resumed operation in late 1979.
3. Celanese Chemical Co.
Much of the EO from the 193-Gg/yr plant is used to produce ethylene glycol for
2 7
captive use. The plant, built in 1967, uses a process licensed from Shell.
4. Dow Chemical Co.
Dow's combined production from two separate plants is 322 Gg/yr, most of which
2
is used captively to produce glycols, glycol ethers, and ethanolamines.
5. Texas Eastman Co.
The major portion of the EO from the 88-Gg/yr capacity is used to produce mono-,
2
di-, and triethylene glycols and glycol ethers.
6. Northern Petrochemical Co.
a
Some of the EO produced is marketed; the remainder is converted to glycols.
7. Olin Corp.
The 50-Gg/yr plant produces EO that is used captively to produce glycols, glycol
2
ethers, ethanolamine, and ethoxylated phenol.
-------
II-6
8. PPG Industries
PPG expanded the capacity at their Beaumont, TX, plant from 30 to 68 Gg/yr in
1977. The EO is used captively to produce glycols, with a very small amount
2
going to glycol ethers.
9. Shell Chemical Co.
2
Shell's plant in Geismar, LA, can produce 318 Gg of EO per year.
10. SunOlin Cemical Co.
2
The plant can produce 45 Gg of EO per year; all the EO is marketed.
11. Texaco
The EO produced by this 215-Gg/yr plant is used for producing glycols (mono-,
di-, tri-, and polyethylene), glycol ethers, ethanolamines, ethoxylated phenols,
2
and mixed linear alcohols.
12. Union Carbide Corp.
Union Carbide, the largest producer of EO, has three plants, with a total capa-
2
city of 1,225 Gg/yr. Their Taft, LA, facility is adjacent to an ethylene oxide-
g
glycol handling and shipping complex, reported to be the world's largest.
-------
II-7
D. REFERENCES*
1. "Ethylene Oxide Comes Under Increasing Suspicion of Harmful Health Effects,"
Chemical Engineering 83(14), 64 (1977).
2. S. A. Cogswell, "Ethylene Oxide," pp. 654.5031A—654.5033F, in Chemical Economics
Handbook, Stanford Research Institute, Menlo Park, CA (January 1980).
3. S. C. Johnson, "U.S. EO/EG Past, Present, and Future," Hydrocarbon Processing
83(6), 109--113 (1976).
4. J. F. Lawson, IT Enviroscience, Trip Report for Visit to Union Carbide
Corp., South Charleston, WV, Dec. 7, 1977 (on file at EPA, ESED, Research
Triangle Park, NC).
5. J. F. Lawson, IT Enviroscience, Trip Report for Visit to BASF Wyandotte
Corp., Geismar, LA. July 11, 1977 (on file at EPA, ESED, Research Triangle
Park, NC).
6. I. Kiquchi, T. Kumazawa, and T. Nakai, "For EO: Air and Oxygen Equal,"
Hydrocarbon Processing 55(3), 69—72 (1976).
7. J. F. Lawson, IT Enviroscience, Trip Report for Visit to Celanese Chemical
Co., Clear Lake, TX, June 21 and 22, 1977 (on file at EPA, ESED, Research
Triangle Park, NC).
8. J. Starkey, Northern Petrochemical Co., letter dated May 2, 1979, to David Patrick,
EPA.
9. "Ethylene Oxide Plant Goes Onstream," Chemical Engineering 83(14), 69 (1977).
*Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of
that paragraph, that reference number is indicated on the material involved.
When the reference appears on a heading, it refers to all the text covered by
that heading.
-------
III-l
III. PROCESS DESCRIPTIONS
A. INTRODUCTION
In the United States two major processes are used in the manufacture of ethylene
oxide (EO); 37% is manufactured at eight locations by oxygen oxidation of
ethylene, and 63% at six locations by air oxidation of ethylene. Both
processes are projected to continue to grow.
Another process — the chlorohydrin process -- was the main method of manu-
facture of EO until 1957 but is no longer used. In 1972 Dow Chemical converted
the remainder of its chlorohydrin capacity to the production of propylene
oxide.
B. OXIDATION OF ETHYLENE
In commercial processes the direct oxidation of ethylene to ethylene oxide is
carried out in the vapor phase, with either air or oxygen used as the oxidant
and with a silver catalyst. Oxidation takes place according to the reaction
2
0
(ethylene) (oxygen) (EO)
A second reaction is
CH2=CH2 + 302 * 2C02 + 2H20
(ethylene) (oxygen) (carbon dioxide) (water)
Under optimum conditions the reactor variables are controlled to give less than
30% per-pass ethylene conversion, which with a fresh catalyst gives a selec-
tivity in the range of 70%, with most of the remainder being converted to CO
and water. A higher per-pass ethylene conversion or catalyst aging results in
a selectivity shift from EO to C02- Catalyst aging is sensitive to the condi-
tions that it is subjected to and catalyst life can run from a very short time
up to maybe 5 years. One producer changes the catalyst when selectivity drops
2
to approximately 60%.
-------
Ill-2
1. Air Oxidation of Ethylene
Figure III-l is a typical flow diagram for a continuous air-oxidation process.
Ethylene and air (Streams 1 and 2) are added to a recycle stream (Stream 3),
which feeds one or more primary reactors operated in parallel. The fresh air:
ethylene feed ratio, usually about 10:1, is varied with the recycle gas to en-
sure an optimum oxygen:ethylene ratio. Oxidation takes place over a silver
catalyst packed in tubes. The reactor is surrounded by a heat transfer fluid
to control the temperature,- the reaction temperature and pressure are main-
tained at 220 to 280°C and 1 to 3 MPa. The unreacted ethylene is separated
from the reaction products and recycled through the reactor until consumed.
The effluent from the primary reactor (Stream 4) is cooled by the recycle
stream from the main absorber (Stream 3) to about 38°C in a shell-and-tube heat
4
exchanger. It is then compressed before entering the main absorber. It
4
passes up the main absorber countercurrent to cold water, in which the EO,
along with some of the carbon dioxide from the stream, dissolves. The water
solution is removed from the base of the main absorber (Stream 5).
Unabsorbed gas passing overhead from the main absorber is split into two un-
equal portions. The larger portion (Stream 3) recycles through the reactor
effluent cooler and joins the fresh reactor feed. The smaller portion
(Stream 6) is passed through a heat exchanger to raise its temperature and then
enters the purge reactor (secondary ethylene conversion reactor). The effluent
from the purge reactor (Stream 7) is cooled by the incoming feed to the purge
reactor and enters the purge absorber, where ethylene oxide is removed from the
stream with water, as in the main absorber. The overhead gas (Vent A) is
4 5
vented from the purge absorber. ' There can be more than one stage of purge
reaction, depending on the economics of the value of ethylene recovered versus
cost.
The dilute water solutions containing ethylene oxide, carbon dioxide, and other
4
VOC from both absorbers are combined (Stream 8). The mixture is fed to the
top of the desorber, where the crude EO (Stream 9) is distilled off the top and
4
compressed for further refining. A stripper removes carbon dioxide and inert
gases overhead (Vent B), and the EO, stripped of carbon dioxide (Stream 10), is
fed to the midsection of the refiner, where it is distilled overhead to
-------
C - FUGITIVE EMI^>IOkl-OVERA,U-
- SECOtODARV EMlS^jlOKJ
Fig. III-l. Flow Diagram for Uncontrolled Model Plant for Production of Ethylene Oxide by
Continuous Air-Oxidation Process
-------
III-4
99.5 mole % EO. ' The product (Stream 11) is stored under a nitrogen atmo-
4
sphere. The secondary reaction of ethylene that produces CO represents not
only a loss of ethylene but also a release of more than 13 times as much energy
as the primary product reaction, i.e., 50.4 vs 3.7 MJ/kg of ethylene. The
large difference in energy released by oxidation when the reaction shifts
4
toward CQ production is illustrated by the following:
'o
Selectivity ratio (ethylene to EO) 70% 60% 50%
Total heat, MJ/kg of ethylene converted 17.8 22.5 27
A 227,000-Mg/yr ethylene oxide plant with 70% EO selectivity would release
400 GJ of heat per hour. If the selectivity should be decreased to 50% as the
result of improper control or catalyst activity, the heat release would be more
4
than doubled, or 900 GJ/hr.
Oxygen Oxidation of Ethylene
Figure III-2 is a typical flow diagram for a continuous oxygen-oxidation
process, which differs slightly from the air-oxidation process. A higher con-
centration of ethylene allows more ethylene to be converted per pass without
exceeding the 30% ethylene conversion favorable for optimum selectivity for EO
formation.
The oxygen-oxidation process also allows recirculation of the unabsorbed gas
through the reactor to achieve a higher ultimate conversion. The higher
ultimate ethylene conversion eliminates the need for the purge-reactor absorber
system required by the air-oxidation process.
As shown by the oxygen-oxidation flow diagram, Fig. III-2, the purge reactor
and purge absorber of the air-oxidation process are replaced by a CO absorber
and reactivator. EO, along with some CO , is dissolved in the water stream
leaving the base of the main absorber (Stream 6) and is fed to the top of the
desorber. The desorption, stripping, and refining steps are similar to those
of the aqueous effluent from the main.absorber of the air-oxidation process
(Stream 8, Fig. III-l).4'5
-------
ETKVUEKJC.
o
PURQE
VEWT
VJATEK
SPEV4T I
CATALYST f(£)
PURGES.
MAIM
REACTOR
MAIM
ABSORBER
ABSORBER
REACT1VATOR
WATER
Fig. III-2. Flow Diagram for Uncontrolled Model Plant for Production of Ethylene Oxide by
Continuous Oxygen-Oxidation Process
-------
III-6
Part of the unabsorbed gas overhead from the main absorber of the oxygen-oxida-
tion process (Stream 7) passes through a CO absorber before being recycled
back to the reactor feed (Stream 8). Carbon dioxide must be removed to main-
tain favorable catalyst activity and favorable conversion to EO. The CO
absorbent, usually potassium carbonate (Stream 9), is heated by the bottoms
from the reactivator and then fed to the top of the reactivator, where it is
4 5
stripped of carbon dioxide and recycled to the CO absorber. '
Small amounts of gaseous impurities in the feed, such as argon, must be removed
5 7
since they will accumulate in the closed system. ' Some of the recycle gas
stream is purged through the argon purge vent (Vent B). This emission flow
rate is automatically regulated by the argon concentration.
3. Process Variation
Some producers captively react EO to glycol or to other products in adjoining
facilities without purification and isolation of EO. In these cases some of
the purge vents shown on Figs. III-l and 2 would exist in the integrated
facility. Also, Figs. III-l and 2 are general process schemes and do not
2
illustrate all the process and operation variations that are practiced.
a. Air-Oxidation Process — The model plant, Fig. III-l, has a two-stage reaction
system, which consists of a main reactor followed in series by a purge reactor.
Large plants may increase the number of stages by incorporating additional
purge reactors in series. Additional stages increase the freedom to optimize
Q
the reaction conditions of each stage. This results in an improved EO yield
by higher ethylene reaction selectivity to EO and in reduction of the ethylene
Q
lost in the process vent gas.
b. Oxygen-Oxidation Process -- Yield improvements have been achieved by intro-
ducing methane into the reactor feed stream. The addition of methane to the
ethylene/ oxygen mixture serves to narrow the flammability limits of the inlet
gas, thereby allowing greater feed ratio flexibility. ' ' The argon purge
Q
vent gas produced (Vent B) then is suitable for boiler fuel.
Operating conditions, such as pressure, of the CO absorber system can affect
the relative quantity and composition of the material vented through Vents A
and B.2
-------
III-7
C. REFERENCES*
1. J. L. Blackford, "Ethylene Oxide," pp. 654.5032C, D, E, in Chemical Economics
Handbook, Stanford Research Institute, Menlo Park, CA (September 1976).
2. J. Starkey, Northern Petroleum Co., letter dated May 2, 1979, to EPA with
information on EO process.
3. I. Kiguchi, T. Kumazawa, and T. Nakai, "For EO: Air and Oxygen Equal,"
Hydrocarbon Processing 55(3) 69--72 (1976).
4. H. C. Schultze, "Ethylene Oxide," p. 523 in Kirk-Othmer Encyclopedia of Chemical
Technology, 3d ed., Vol. 8, 2d ed., edited by A. Standen et al., Wiley, New
York, 1967.
5. M. Cans and B. J. Ozero, "For EO: Air or Oxygen?" Hydrocarbon Processing
55(3), 73—77 (1976).
6- J. F. Lawson, IT Enviroscience, Trip Report for Visit to BASF Wyandotte
Corp., Geismar, LA. July 11, 1977 (on file at EPA, ESED, Research Triangle
Park, NC).
7- J. F. Lawson, IT Enviroscience, Trip Report for Visit to Celanese Chemical
Co., Clear Lake City. TX, June 21 and 22, 1977 (on file at EPA, ESED, Research
Triangle Park, NC).
8. J. F. Lawson, IT Enviroscience, Trip Report for Visit to Union Carbide Corp.,
South Charleston, WV, Dec. 7, 1977 (on file at EPA, ESED, Research Triangle
Park, NC).
9- D. E. Field et al., Houdry Division of Air Products, Inc., Engineering and Cost
Study of Air~?ollution Control for the Petrochemical Industry. Vol. 6,
Ethylene Oxide Manufacture by Direct Oxidation of Ethylene, EPA-450/3-73-006F,
Research Triangle Park, NC (June 1975).
*Usually, when a reference is located at the end of a paragraph, it refers to
the entire paragraph. If another reference relates to certain portions of that
paragraph, that reference number is indicated on the material involved. When
the reference appears on a heading, it refers to all the text covered by that
heading.
-------
IV-1
IV. EMISSIONS
Emissions in this report are usually identified in terms of volatile organic
compounds (VOC). VOC are currently considered by the EPA to be those of a
large group of organic chemicals, most of which, when emitted to the atmo-
sphere, participate in photochemical reactions producing ozone. A relatively
small number of organic chemicals have low or negligible photochemical reactiv-
ity. However, many of these organic chemicals are of concern and may be sub-
ject to regulation by EPA under Sections 111 or 112 of the Clean Air Act since
there are associated health or welfare impacts other than those related to
ozone formation. It should be noted that, although ethane 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. AIR-OXIDATION PROCESS
1. Model Plant
The model plant* for this study has an EO capacity of 227 Mg/yr, typical for
the industry; this capacity is based on 8760 hr of operation annually.** The
model air-oxidation process shown in Fig. III-l reflects today's manufacturing
and engineering technology. Single process trains with multiple, parallel,
main reactors and with at least one purge reaction stage are typical. Charac-
teristics of the model plant important to air dispersion are shown in Table B-l
in Appendix B.
2. Sources and Emissions
Sources and emission rates for the air-oxidation process are summarized in
Table IV-1. They are meant to represent the typical emissions for the model
process; actual emissions of a producer could vary widely for an individual
source but the overall total plant emissions from various plants are probably
relatively consistent.
*See p. 1-2 for a discussion of model plants.
**Process downtime is normally expected to range from 5 to 15%. If the hourly
production rate remains constant, the annual production and annual VOC emis-
sions will be correspondingly reduced. Control devices will normally operate
on the same cycle as the process. Therefore, from the standpoint of cost ef-
fectiveness calculations, the error introduced by assuming continuous operation
is negligible.
-------
IV-2
Table IV-1. Total Uncontrolled VOC Emissions from Model Plant for
Air-Oxidation Process
Source
Main process vent
Stripper purge vent
Fugitive
Storage and handling
Secondary
Total
Stream
Designation
(Fig.III-1)
A
B
C
D
E
Emissions
Ratio
(9/kg)c
59.6
4.0
0.42
2.6
0.0116
66.6
Rate
(kg/hr)
1544
104
10.8
68
0.3
1727
Uncontrolled emissions are emissions from process for which there are no
control devices other than those necessary for economical operation.
Emissions include ethylene oxide, ethylene, and ethane.
Q
g of emissions per kg of ethylene oxide produced.
-------
IV-3
a. Main Process Vent — The main process vent (Vent A, Fig. III-l) is the largest
process emission source. It contains the nitrogen and unreacted oxygen from
the air fed to the reactor, ethane and unreacted ethylene from the ethylene
feed, and product EO and by-product CO . The composition of this stream, given
in Table IV-2, for the model plant depends on the catalyst and the reactor
conditions, ethylene feed purity, number of purge reaction stages, and absorber
operating conditions.
During startup of a reactor the air feed rate is brought up slowly in correla-
tion with the ethylene feed. Emission ratios during startup are essentially
2
the same as those for normal operation. Process upsets, such as the loss of
the stripper feed compressor, can cause a sharp emission increase. When an
upset occurs, the ethylene feed is cut back, which reduces the VOC level
3t€
4
exhausted from the reactor. The vent can also be directed to an emergency
flare.
b. Stripper Purge Vent — The overhead stream from the stripper column is vented
through the stripper purge vent (Vent B, Fig. III-l). The stream is composed
of the inert gases and ethylene that become dissolved in the main and purge
absorber waters during the recovery of EO from the reaction gases. Normally,
any EO that this stream may contain is scrubbed out with water and returned to
the process. Table IV-3 gives the composition of this stream for the model
plant after EO has been scrubbed out. The emissions from the vent are not
2
affected by process startups or shutdowns. Since the emission is a function
of gas solubilities in the circulating water, the water rate used has an effect
on the emission.
c. Fugitive Emissions -- Process pumps, compressors, and valves are potential
sources of fugitive emissions (Source C, Fig. III-l). The model plant is
estimated to have 10 pumps, 2 compressors, and 400 valves handling VOC. The
factors in Appendix C were used to determine the emission contribution of these
equipment components. The process water from the desorber bottoms is also a
potential source of fugitive emissions. The water is recycled through an
atmospheric cooling tower, with the excess discharged to wastewater treatment.
For the model plant 2500 kg of glycols and aldehydes is recycled per hour in
this stream. It is estimated that 0.15%, or 3.8 kg/hr, is lost as VOC to the
atmosphere.
-------
IV-4
Table IV-2. Gas Composition of Main Process Vent in
Air-Oxidation-Process Model Plant
Component
Ethylene oxide
Ethylene
Ethane
Total VOC
Nitrogen
Oxygen
Carbon dioxide
Water
Total
Composition (wt %)
0.02
0.80
0.09
0.91
80.23
3.07
15.65
0.15
100
Emission Ratio (g/kg)
1
52
6
59
5253
201
1025
10
6548
.0
.6
.0
.6
See refs. 1—3.
g of emission per kg of ethylene oxide produced.
Table IV-3. Gas Composition of Stripper Purge Vent in
Air-Oxidation-Process Model Plant
Componont
Ethylene
Nitrogen
Oxygen
Carbon dioxide
Total
Composition (wt %)
3.30
13.1
1.7
81.9
100
. b
Emission Ratio
4.0
16
2
100
122
(gAg)
See refs. 1—3.
°g of emission per kg of ethylene oxide produced.
-------
IV-5
d. Storage and Handling Emissions — Emissions result from the storage and
handling of EO, Source D, Fig. III-l. For the model plant EO is stored at
10°C under a nitrogen pad in pressure tanks. For production that is not used
captively, shipment is by tank car. Storage tank conditions for the model
plant are given in Table IV-4. The uncontrolled storage emissions in
Table IV-1 were calculated by assuming that the day tanks are vapor balanced
with the storage tanks and that an equivalent amount of vapor, saturated with
EO at 10°C, is displaced from the system for each volume of EO produced. Tank
car loading losses were calculated similarly, but it was estimated that only
20% of EO production is shipped and that the average handling temperature is
16°C.
Storage and handling practices of an individual plant can vary widely from the
model-plant conditions; for instance, some producers store at refrigerated
temperatures and near-ambient pressures, whereas others store at ambient
temperatures and elevated pressures. Also, the amount shipped, if any, will
depend on the individual producer's situation.
e. Secondary Emissions — Secondary VOC emissions can result from the handling and
disposal of process waste streams. For the model plant two potential sources
are indicated on the flow diagram (Sources E, Fig. III-l): the heavy ends from
the refiner column and the spent catalyst from the reactors.
The refiner heavy ends for the model plant are estimated to be 115-kg/hr total
organic containing 6-kg/hr VOC. The VOC emitted to the atmosphere is
estimated to be 5%. The spent catalyst is purged before removal, is changed
infrequently, and is reclaimed off-site. The potential for emissions from this
source is slight.
B. OXYGEN-OXIDATION PROCESS
1. Model Plant
The model plant for this study has an EO capacity of 136 Gg/yr, a typical
industrial capacity; this value is based on 8760 hr of operation annually. The
model oxygen-oxidation process in Fig. III-2 reflects today's manufacturing and
engineering technology. Single process trains with multiple parallel reactors
-------
IV-6
Table IV-4. Storage Tank Data for
Air-Oxidation-Process Model Plant
Parameter
Contents
No. of tanks required
3
Tank size (m )
Turnovers per year
Bulk temperature (°C)
Day
EO
2
225
550
10
Tank
Storage
EO
6
470
89
10
-------
IV-7
are typical. Characteristics of the model plant important to air dispersion
are shown in Table B-2 in Appendix B.
2. Sources and Emissions
Sources and emission rates for the oxygen-oxidation process are summarized in
Table IV-5. They are meant to represent the typical emissions for the model
process; actual emissions of a plant could vary widely for an individual
source, but the overall total plant emissions from various plants are probably
relatively consistent.
a. CO Purge Vent -- The overhead stream from the reactivator column is vented
through the CO purge vent (Vent A, Fig. III-2). The column reactivates the
CO absorbent medium for recycle by stripping it of carbon dioxide. The vent
4Ci
emissions consist of most of the by-product C02 formed in the reactors,
together with some of the ethane from the ethylene feed. The composition of
this stream, given in Table IV-6, for the model plant depends mainly on the C02
absorbent medium used and the C02 absorber operating conditions. The emissions
from this source normally are not affected by process startups or shutdowns.
b. Argon Purge Vent — A discard stream from the reaction recycle gases is vented
through the argon purge vent (Vent B, Fig. III-2). The stream contains most of
the argon and nitrogen that enter with the oxygen feed and the ethane that
enters with the ethylene feed. The composition of this stream (see Table IV-7)
for the model plant depends on the argon level that can be tolerated in the
reaction system. Argon has a low specific heat compared to other gases, and
too high a concentration can affect the reactor temperature control. The
quantity vented is directly related to the composition of the oxygen feed. For
the process variation in which methane is added to the reactor feed stream the
emissions from the vent will also contain methane.
Process upsets and shutdowns normally do not affect this emission source, nor
do startups, provided that the oxygen feed composition is established before
startup.
c. Stripper Purge Vent -- The overhead stream from the stripper column is vented
through the stripper purge vent (Vent C, Fig. III-2), and is composed of the
-------
IV-8
Table IV-5. Total Uncontrolled& VOC Emissions from Model Plant for
Oxygen-Oxidation Process
Source
CO purge vent
Argon purge vent
Stripper purge vent
Fugitive
Storage and handling
Secondary
Stream
Designation
(Fig.III-2)
A
B
C
D
E
F
Emissions
Ratio
(g/kg) c
4.0
10.9
2.8
0.55
2.6
0.013
20.9
Rate
(kg/hr)
62
170
44
8.7
41
0.2
326
Uncontrolled emissions are emissions from process for which there are no
control devices other than those necessary for economical operation.
Emissions include ethylene oxide, ethylene, and ethane.
^
~g of emissions per kg of ethylene oxide produced.
Table IV-6. Gas Composition of CO Purge Vent in
Oxygen-Oxidation-Process Model Plant
Component
Ethylene
Ethane
Total VOC
Oxygen
Carbon dioxide
Water
Total
Composition (wt %)
0.23
0.05
0.28
0.01
46.16
53.55
100
Emission Ratio (g/kg)
3.3
0.66
4.0
0.10
666
773
1443
See refs. 3—5.
3g of emissions per kg of ethylene oxide produced.
-------
IV-9
Table IV-7. Gas Composition of Argon Purge Vent in
Oxygen-Oxidation-Process Model Plant
Component
Ethylene
Ethane
Total VOC
Nitrogen
Oxygen
Carbon dioxide
Argon
Water
Total
Composition
(wt %)
21.3
__!_•!
26.4
16.6
6.4
31.2
18.4
1.0
100
b
Emission Ratio
(g/kg)
8.8
2.1
10.9
7.7
2.6
12.9
7.7
0.42
42
See refs 3 and 4.
Dg of emission per kg of ethylene oxide produced.
-------
IV-10
inert gases and ethylene that become dissolved in the main absorber water dur-
ing the recovery of EO from the reaction gases. Normally, any EO that this
stream may contain is scrubbed out with water and returned to the process. The
composition of this stream after the EO is scrubbed is given in Table IV-8.
The emission is not affected by process startups or shutdowns. Since the
emission is a function of gas solubilities in the circulating water, the water
rate used has an effect on the emission.
d. Fugitive Emissions -- Process pumps, compressors, and valves are potential
sources of fugitive emissions (Source D, Fig. III-2). The model plant is
estimated to have 10 pumps, 2 compressors, and 400 valves handling VOC. The
factors in Appendix C were used to determine the emission contribution of these
equipment components. The process water from the desorber bottoms is also a
potential source of fugitive emissions. The water is recycled through an
atmospheric cooling tower, with the excess discharged to wastewater treatment.
For the model plant 1100 kg of glycols and aldehydes is recycled in this stream
per hour. It is estimated that 0.15%, or 1.7 kg/hr, is lost as VOC to the
atmosphere.
e. Storage and Handling Emissions -- Emissions result from the storage and
handling of EO, Source D, Fig. III-2. For the model plant EO is stored at 10°C
under a nitrogen pad in pressure tanks. For production that is not used
captively, shipment is by tank car. Storage tank conditions for the model
plant are given in Table IV-9. The uncontrolled storage emissions in
Table IV-6 were calculated by assuming that the day tanks are vapor balanced
with the storage tanks and that an equivalent amount of vapor, saturated with
EO at 10°C, is displaced from the system for each volume of EO produced. Tank
car loading losses were calculated similarly, but it was assumed that only 20%
of EO production is shipped and that the average handling temperature is 16°C.
Storage and handling practices of an individual plant can vary widely from the
model-plant conditions; for instance, some producers store at refrigerated
temperatures and near-ambient pressures, whereas others store at ambient tempera-
ture and elevated pressures. Also, the amount shipped, if any, will depend on
the individual producer's situation.
-------
IV-11
Table IV-8. Gas Composition of Stripper Purge Vent in
Oxygen-Oxidation-Process Model Plant
Component
Ethylene
Ethane
Total VOC
Nitrogen
Oxygen
Carbon dioxide
Water
Total
Composition
(wt %)
34.2
0.2
34.4
12.2
8.2
16.6
28.6
100
. b
Emission Ratio
(g/kg)
2.8
0.0.19.
2.8
1.0
0.67
1.4
2.3
8.2
See ref 4.
3g of emission per kg of ethylene oxide produced.
Table IV-9. Storage Tank Data for
Oxygen-Oxidation-Process Model Plant
Tank
Parameter
Contents
No. of tanks required
Tank size (m )
Turnovers per year
Bulk temperature (°C)
Day
EO
3
150
330
10
Storage
EO
5
470
64
10
-------
IV-12
The refiner heavy ends for the model plant are estimated to be 70-kg/hr total
4 5
organic containing 4-kg/hr VOC. ' The VOC emitted to the atmosphere is esti-
mated to be 5%. The spent catalyst is purged before removal, is changed
infrequently, and is reclaimed off-site. The potential for emissions from this
source is slight.
-------
IV-13
C. REFERENCES*
1. M. Cans and B. J. Ozero, "For EO: Air or Oxygen?" Hydrocarbon Processing
55(3), 73--77 (1976).
2. J. F. Lawson, IT Enviroscience, Trip Report for Visit to Union Carbide
Corp., South Charleston, WV, Dec. 7, 1977 (on file at EPA, ESED, Research
Triangle Park, NC).
3. D. E. Field e_t al., Houdry Division of Air Products, Inc., Engineering and Cost
Study of Air Pollution Control for the Petrochemical Industry. Volume 6:
"Ethylene Oxide Manufacture by Direct Oxidation of Ethylene," EPA-450/3-73-006F,
Research Triangle Park, NC (June 1975).
4. J. F. Lawson, IT Enviroscience, Trip Report for Visit to BASF Wyandotte
Corp., Geismar, LA, July 11, 1977 (on file at EPA, ESED, Research Triangle
Park, NC).
5. J. F. Lawson, IT Enviroscience, Trip Report for Visit to Celanese Chemical
Co., Clear Lake City, TX, June 21 and 22, 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. AIR-OXIDATION PROCESS
1 - Main Process Vent
The stream from the main process vent is the largest process emission source
(Vent A, Fig. III-l) in the model plant. Control by catalytic oxidation and by
thermal oxidation will be evaluated for this source. Although only catalytic
oxidation is currently being used for this source, thermal oxidation has the
potential for higher efficiency operation, and is used on similar waste streams
in other chemical production processes.
a. Catalytic Oxidation -- A catalytic oxidation unit normally consists of a pre-
heater, a catalyst bed, a heat recovery unit, and a stack, together with the
necessary controls, blowers, and supplemental fuel supply. In a modern
ethylene oxide plant, a catalytic oxidation unit can play a key role in the
energy engineering of the plant. It can be an integral part of the system that
recovers energy from the process off-gases by use of turbines for compression
of the process air feed.
Modern catalytic oxidation units designed for this service can reduce the
average effluent ethylene concentration to less than 500 ppm, and probably to
as low as 200 ppm. Ethane is more difficult to burn than ethylene, because it
has a higher ignition temperature. The ethane effluent concentration will be
in the range of twice that of ethylene. The difference in their burning
characteristics requires that the ethylene-to-ethane ratio in the feed be con-
trolled to ensure ethane ignition. This limits the amount of ethane that can
be burned so as not to exceed the temperature limitation of the catalyst. It
has been stated that the next generation of catalytic oxidation units being
developed may have the potential to reduce effluent concentrations to the range
of 60 ppm.
For the model plant, based on the present modern technology, VOC removal
efficiency of 95% is used for the main process vent. The controlled emission
is shown in Table V-l.
-------
Table V-l. VOC Controlled Emissions for Air-Oxidation-Process Model Plant
Source
Main process vent
Stripper purge vent
Fugitive
Storage and handling
Secondary
Total
Stream
Designation Control Device
(Fig.III-1) or Technique
A Catalytic oxidation
Thermal oxidation
B Compression and recycle
Catalytic oxidation
Thermal oxidation
C Detection and correc-
tion of major leaks
D Aqueous scrubber
E None
Total VOC
Emission
Reduction (%)
95
99
97.3
95
99
50
99.5
VOC Emissions
Ratio (g/kg)
3.0
0.6
0.11
0.20
0.040
0.21
0.013
0.012
3.3
Rate (kg/hr)
77
15
2.8
5.2
1.0
5.4
0.34
0.3
85.8
f
M
g of emission per kg of ethylene oxide produced.
DBased on catalytic oxidation for the main process vent and compression and recycle for the stripper purge vent.
-------
V-3
b. Thermal Oxidation -- A direct-fired thermal oxidation unit for the main process
vent would have a combustion chamber that provides sufficient residence tiae
for complete combustion, a supplementary fuel burner that provides auxiliary
heat to raise the fume temperatures sufficiently for complete combustion, and
provision for good mixing of the fuaes vith the combustion gases, together with
the necessary controls, blower, and stack. A heat recovery system to preheat
the feed or generate steam could also be incorporated.
Based on similar incineration applications it is concluded that a properly
designed and operated incinerator for this service will achieve a VOC removal
efficiency of greater than 99%. An incineration teraerature of 870°C and a
retention time of 0.5 sec is specified to ensure complete combustion of the
waste VOC. While it is possible that greater than 99% VOC removal efficiency
could be obtained at lower temperatures, it cannot be predicted dependably.
For the model plant a main process vent removal efficiency of 99% is projected.
The controlled emission is shown in Table V-l.
2. Stripper Purge Vent
Two methods of handling the stream from the stripper purge vent {Vent B,
Fig. III-l) will be evaluated: compression and recycle to the process and
oxidation in the main-process-vent control. Of these two controls, only the
recycle option is in present use,- another control that has been practiced is
burning the stream in a boiler fire box,
a. Compression and Recycle -- In this option the stream is compressed and returned
to the purge reactor. When the carbon dioxide contained in the stream is rein-
jected into the reaction cycle, it can cause impairment of the ethylene selec-
tivity and tend to offset the recovered ethylene value. The reinjected inert
gases, plus any incremental increase in VOC due to a lower selectivity, are
eventually discharged through the main process vent.
With the recycle ethylene stream assumed to have the same conversion as the
main process stream in the purge reactor, the ethylene use efficiency for the
model plant is 46%. The net overall VGC emission removal efficiency is 97,3%
if the raain-process-vent control is catalytic oxidation, and is 99.4% if the
control is thermal oxidation (see Table V-l).
-------
V-4
b. Stripper Purge Vent Combined with Main Process Vent -- In this option the
stream is combined with the larger main-process-vent stream and fed to its
control device. If the main-process-vent control is a catalytic oxidation unit
that is located in the process so as to recover energy, a compressor would be
required to compress the gas to the operating pressure. Under this condition
it probably is best to consider only the compression and recycle option because
in that option some ethylene value can be recovered.
The VOC reduction efficiency will be the same as for the main process vent, 95%
for catalytic oxidation and 99% for thermal oxidation (see Table V-l).
3. Fugitive-Emissions
Controls for fugitive emissions from the synthetic organic chemicals manufac-
turing industry are discussed in a separate EPA document. Emissions from
pumps, process valves, and pressure-relief devices can be controlled by an
appropriate leak-detection system and with repair and maintenance as needed.
Controlled fugitive emissions were calculated with the appropriate factors
given in Appendix C and are included in Table V-l.
There are measures that are taken in some EO plants for safety reasons that
are also effective in reducing fugitive emissions; for example:
a. using trategic location of hazardous vapor detectors in the plant area to
detect EO leaks,
b. equipping EO pumps with double mechanical seals having liquid buffer zones
and alarms to indicate a failure of either seal,
c. using pressurized nitrogen in labyrinth shaft seals of centrifugal EO com-
pressors to prevent leakage to the atmosphere,
d, using leak detectors for critical flanges in EO piping,
e. paying extra attention to the maintenance of EO piping because of the
danger of fires from leaks.
f. collecting EO leakage or drainage from sampling operations and pump vents,
absorbing in water, and then discharging to the sewer.
-------
V-5
4. Storage and Handling Sources
It is important to control the EO vapors in the storage and handling areas,
Source D, Fig. III-l, because of health and safety hazards. The displaced vapors
from the filling of storage tanks and tank cars can be controlled by use of an
1 4
aqueous scrubber. ' A flare has also been indicated as being used for this
service.
An aqueous scrubber is usually a packed tower fed with process water, and the
effluent is processed in the plant desorption unit for EO recovery. EO removal
efficiency is essentially complete. A VOC removal efficiency of 99.5% for total
VOC is used for the controlled emission in Table V-l.
A flare for an EO plant would normally be designed for process emergency vent-
ing conditions. If storage and handling emissions were disposed to such a flare,
caution must be taken because EO requires only an ignition source to cause safety
problems. Use of flares as a control device is discussed in a separate EPA docu-
ment.
5. Secondary Sources
No control system has been identified for the model plant. Control of secondary
7
emissions is discussed in another EPA report.
B. OXYGEN-OXIDATION PROCESS
1. CO Purge Vent
The stream from the CO purge vent (Vent A, Fig. III-2) has a VOC concentration
of only 0.28%, the remainder being CO and water. For the model plant this amounts
to 62 kg of VOC per hour. Control of this stream by thermal or catalytic oxidation
would not be practical; in the model plant it is left uncontrolled (Table V-2).
A theoretical design for controlling this emission that has been proposed con-
sists of two-stage flashing of the CO absorber effluent before it is introduced
into the reactivator. The ethylene will tend to flash out preferentially. This
flash gas is then compressed and returned to the absorber. A preliminary estimate
indicates that a 60% efficiency may be an attainable value for this type of control.
-------
Table V-2. VOC Controlled Emissions for Oxygen-Oxidation-Process Model Plant
Source
CO purge vent
Argon purge vent
Stripper purge vent
Fugitive
Storage and handling
Secondary
Total
Stream
Designation
(Fig.III-2)
A
B
C
D
E
F
Control Device
or Technique
None
Used as fuel
Compression and recycle
Detection and correction
of major leaks
Aqueous scrubber
None
Total VOC ,roo
_ . . VOC Emi!
Reduction (%) Ratioa (g/kg)
4.0
99.9 0.011
99.7 0.0084
62 0.21
99.5 0.013
0.013
4.3
s s ions
Rate (kg/hr)
62
0.17
0.13
3.3
0.21
0.2
66.0
fTl
g of emission per kg of ethylene oxide produced.
-------
V-7
4
One producer uses the CO purge vent in another process. The ultimate fate of
the contained VOC is thermal oxidation. When the other process unit is down,
the vent emissions are oxidized in an existing incinerator.
Argon Purge Vent
The emissions from the argon purge vent (Vent B, Fig. III-2) are high in ethylene
and will support combustion. In the process variation in which methane is used
in the reactor, this vent stream will also contain appreciable methane, making
it an even better fuel gas. The emissions from this vent can be disposed of
145
readily in a fire box or fuel header. At least several producers ' ' do use
it as a fuel.
For the model plant the emission control will be to use the gas as fuel (Table V-2)
An alternative method of control would be to burn the gas in a flare. The VOC
reduction efficiency when the vent gas is burned as fuel can be greater than
99.9%.6
3. Stripper Purge Vent
The emissions from the stripper purge vent (Vent C, Fig. III-2) are high in ethy-
lene and will support combustion but may require supplemental fuel for flame
stability. In the process variation in which methane is used in the reactor,
the emissions will also include some methane. In older processes, this vent
emission has been disposed of in a boiler fire box; it also could be flared.
In newer installations it is compressed and returned to the reaction cycle in
the C0? absorber feed.
For the model plant (Table V-2) the emission control evaluated will consist of
compression and recycling. This type of control has two negative effects on
the process. The reinjected carbon dioxide requires additional energy for its
removal via the CO absorption system, and the reinjected argon must be vented
in the argon purge vent. The net VOC reduction efficiency is 99.7% when the
argon purge vent (Vent B) is used as a fuel gas and the C02 purge vent (Vent A)
is uncontrolled.
-------
V-8
4. Fugitive Emissions
Controls for fugitive emissions from the synthetic organic chemicals manufactur-
ing industry are discussed in a separate EPA document. Emissions from pumps,
process valves, and pressure-relief devices can be controlled by an appropriate
leak-detection system and with repair and maintenance as needed. Controlled
fugitive emissions were calculated with the appropriate factors given in
Appendix C and are included in Table V-2.
Some measures that are taken in one EO plant for safety reasons are also effec-
tive in reducing fugitive emissions; for example:
a. using strategic location of hazardous vapor detectors in the plant area to
detect EO leaks,
b. equipping EO pumps with double mechanical seals having liquid buffer zones
and alarms to indicate a failure of either seal,
c. using pressurized nitrogen in labyrinth shaft seals of centrifugal EO com-
pressors to prevent leakage to the atmosphere,
d. using leak detectors for critical flanges in EO piping,
e. paying extra attention to the maintenance of EO piping because of the danger
of fires from leaks,
f. collecting EO leakage or drainage from sampling operations and pump vents,
absorbing in water, and then discharging to the sewer.
5. Storage and Handling Sources
It is important to control the EO vapors in the storage and handling areas,
Source E, Fig. III-2, because of health and safety hazards. The displaced vapors
from the filling of storage tanks and tank cars can be controlled by use of an
1 4
aqueous scrubber. ' A flare has also been indicated as being used for this
service.
An aqueous scrubber is usually a packed tower fed with process water, and the
effluent is processed in the plant desorption unit for EO recovery. EO removal
efficiency is essentially complete, arid a removal efficiency of 99.5% for total
VOC is used for the controlled emission in Table V-2.
-------
V-9
A flare for an EO plant would normally be designed for process emergency venting
conditions. If storage and handling emissions were disposed to such a flare,
caution must be taken because EO requires only an ignition source to cause safety
problems. Use of flares as a control device is discussed in a separate EPA docu-
6
ment.
6. Secondary Sources
No control system has been identified for the model plant. Control of secondary
7
emissions is discussed in another EPA report.
-------
V-10
C. REFERENCES*
1. J. F. Lawson, IT Enviroscience, Trip Report for Visit to Union Carbide
Corp., South Charleston, WV, Dec. 7, 1977 (on file at EPA, ESED, Research
Triangle Park, NC).
2. J. W. Blackburn, IT Enviroscience, Control Device Evaluation. Thermal Oxidation
(July 1980) (EPA/ESED report, Research Triangle Park, NC).
3. D. G. Erikson and V. Kalcevic, IT Enviroscience, Inc., Fugitive Emissions
(September 1980) (EPA/ESED report, Research Triangle Park, NC).
4. J. F. Lawson, IT Enviroscience, Trip Report for Visit to Celanese Chemical
Co., Clear Lake City, TX, June 21 and 22, 1977 (on file at EPA, ESED, Research
Triangle Park, NC).
5. J. F. Lawson, IT Enviroscience, Trip Report for Visit to BASF Wyandotte
Corp., Geismar, LA, July 11, 1977 (on file at EPA, ESED, Research Triangle
Park, NC).
6. V. Kalcevic, IT Enviroscience, Control Devices Evaluation. Flares and the Use
of Emission as Fuels (in preparation for EPA, ESED, Research Triangle Park, NC)
August 1980).
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
1. Air-Oxidation Process
Table VI-1 shows the environmental impact of reducing VOC emissions by applica-
tion of the described control systems (Sect. V) to the model plant. From an
energy standpoint a typical uncontrolled air-oxidation-process EO plant will
produce a heat surplus of 6 to 11 MJ per kg of EO and will require power of
about 5 MJ per kg of EO.
a. Main Process Vent -- Emissions from the main process vent can be controlled by
a catalytic oxidizer or a thermal oxidizer with the environmental impacts
described below:
Catalytic oxidation — The catalytic oxidizer reduces VOC emissions by
12.9 Gg/yr for the model plant. It uses supplemental fuel to preheat the waste
gas and electrical power for blowers, lighting, and instruments, with a total
energy requirement of 2.5 MJ per kg of EO. If heat recovery equipment is
installed and approximately 62% of the available energy from the combustion
gases is recovered, this would amount to 2.2 MJ per kg of EO, or a net energy
usage of 0.3 MJ/kg for the model plant.
Thermal oxidation — The thermal oxidizer reduces VOC emissions by 13.4 Gg/yr
for the model plant.
The thermal oxidizer uses supplemental fuel to heat the waste gas stream and
electrical power for blowers, lighting, and instruments. The total energy
required is 9.9 MJ per kg of EO produced. If heat recovery equipment is
installed and approximately 62% of the available energy from the combustion
gases is recovered, this would amount to 8.2 MJ per kg of EO, or a net usage of
1.7 MJ/kg for the model plant.
b. Stripper Purge Vent -- The compression and recycle of the emissions from this
vent back to the process, together with a catalytic oxidizer on the main
process vent, reduces VOC emissions by 0.9 Gg/yr for the model plant. Electri-
-------
Table VI-1; Environmental Impact of Controlled Air-Oxidation-Process Model Plant
Emission
Source
Main process vent
Stripper purge vent
Fugitive
Storage and handling
Secondary
Total3
Stream or
Vent
Designation
(Fia. III-1)
A
B
C
D
E
Control Device
or Technique
Catalytic oxidation
Thermal oxidation
Compression and recycle
Catalytic oxidation
Thermal oxidation
Detection and correction
of major leaks
Aqueous scrubber
None
VOC Emission
(%)
95
99
97.3
95
99
50
99.5
Reduction
(Mo/vr)
12,850
13,390
886
865
902
47
593 3
i
to
14,376
aBased on catalytic oxidation for the main process vent and compression and recycle for the stripper purge vent.
-------
VI-3
cal energy, 0.03 MJ per kg of EO produced, is required for compressing this
stream, and ethylene equivalent to 1.8 g/kg of EO is recovered.
c. Fugitive Emissions — The control methods previously described for these emis-
sions are major leak detection and repair of equipment components. Application
of these systems results in a VOC reduction of 0.05 Gg/yr for the model plant.
d. Storage and Handling -- The aqueous scrubber reduces VOC emissions from storage
and handling by 0.59 Gg/yr and recovers EO equivalent to 2.6 g per kg of EO
produced for the model plant. The electrical energy and process water required
for the aqueous scrubber system are negligible.
2. Oxygen-Oxidation Process
Table VI-2 shows the environmental impact of reducing VOC emissions by applica-
tion of the described control systems to the model plant. From an energy
standpoint a typical uncontrolled oxygen-oxidation process EO plant will
produce a heat surplus of about 9 MJ per kg of EO produced and will require
power of 1.5 to 2.5 MJ/kg of EO. This does not include the energy for the
oxygen supply plant.
a. Argon Purge Vent -- The use of emissions from this vent as fuel gas will reduce
VOC emissions by 1.5 Gg/yr for the model plant, and the heating value will be
equal to 0.56 MJ/kg of EO produced.
b. Stripper Purge Vent -- The compression and recycle of the emissions from this
vent back to the process, together with use of the argon purge vent emissions
as fuel, reduce VOC emissions by 0.4 Gg/yr and recover ethylene equivalent to
1.4 g per kg of EO produced for the model plant. The electrical energy consump-
tion for the compressor per kg of EO produced is relatively small.
c. Fugitive Emissions — The control methods described for these emissions are
major leak detection and repair of equipment components. Application of these
systems results in a VOC reduction of 0.05 Gg/yr for the model plant.
-------
Table VI-2. Environmental Impact of Controlled Oxygen-Oxidation-Process Model Plant
Emission
Source
CO purge vent
Argon purge vent
Stripper purge vent
Fugitive
Storage and handling
Secondary
Stream or
Vent
Designation
(Fig. III-2)
A
B
C
D
E
F
Control Device
or Technique
None
Used as fuel
Compression and recycle
Detection and correction
of major leaks
Aqueous scrubber
None
VOC
m
99.9
99.7
62
99.5
Emission Reduction
(Mq/yr)
1488
384
47
357
Total
2276
H
I
-------
VI-5
d. Storage and Handling -- The aqueous scrubber reduces VOC emissions from storage
and handling by 0.36 Gg/yr and recovers EO equivalent to 2.6 g per kg of EO
produced for the model plant. The electrical energy and process water required
for the aqueous scrubber system are negligible.
B. CONTROL COST IMPACT
This section gives estimated costs and cost-effectiveness data for control of
VOC emissions resulting from the production of ethylene oxide. Details of the
model plant (Figs. III-l and III-2) are given in Sects. Ill and IV. Cost
estimate sample calculations are included in Appendix D.
Capital cost estimates represent the total investment required for purchase and
installation of all equipment and material required 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 ethylene oxide production lost dur-
ing installation or startup, research and development, or land acquisition.
Bases for the annual cost estimates for the control alternatives include
utilities, operating labor, maintenance supplies and labor, recovery credits,
capital charges, and miscellaneous recurring costs such as taxes, insurance,
and administrative overhead. The cost factors used are itemized in Table VI-3.
1- Air-Oxidation Process
a. Main-Process Vent (Vent A, Fig. III-l) — This is the major process emission
source. Two emission controls for this vent have been evaluated, a catalytic
oxidizer and a thermal oxidizer. Both controls are evaluated with and without
heat recovery options. Heat recovery is based on the use of a waste heat
boiler on the exit gas. Recuperative recovery could be used to preheat the
feed streams and thereby reduce the supplemental fuel requirements. The emis-
sion control cost estimates for these systems for the model plant are shown in
Table VI-4. The installed capital cost, net annual cost, and cost-effective-
ness variations with capacity are shown in Figs. VI-1 to VI-3. See Appendix D
for the cost estimate sample calculations for a catalytic oxidizer and a
thermal oxidizer, based on complete installations as described in the control
device evaluation reports on catalytic oxidation and thermal oxidation. '3
-------
VI-6
Table VI-3. Annual Cost Parameters
Operating factor
Operating labor
Fixed costs
Maintenance labor plus
materials, 6%
Capital recovery, 18%
Taxes, insurances,
administration charges, 5%
Utilities
Electric power
Natural gas
Heat recovery credits
(equivalent to natural gas)
Ethylene recovery credit
8760 hr/yr'
$15/man-hr
29% of installed capital cost
$8.33/GJ ($0.03/kWh)
$1.90/GJ ($2.00/thousand ft3
or million Btu)
$1.90/GJ ($2.00/million Btu)
$0.287/kg
Process downtime is normally expected to range from 5 to 15%. If the hourly
rate remains constant, the annual production and annual VOC emissions will be
correspondingly reduced. Control devices will usually operate on the same
cycle as the process. From the standpoint of cost-effectiveness calculations,
the error introduced by assuming continuous operation is negligible.
Based on 10-year life and 12% interest.
-------
Table VI-4. Emission Control Cost Estimates for Model Plants Using Air-Oxidation or Oxygen-Oxidation Process
Emission
Source
Main process vent (A)
Stripper purge vent (B)
Stripper purge vent (C)
*(C) - (A) « (B).
bVent designation shown on
Control
Catalytic oxidizer
With waste heat boiler
Without heat recovery
Thermal oxidizer
With waste heat boiler
Without heat recovery
Coif>res8ion and recycle
Compression and recycling
Fig. ixx-1.
cReduction percent is based on 46» VOC reduction in
Vent designation shown on
IB
Fig. III-2.
Installed
Capital
Cost
$2,600,000
1,500,000
3,100,000
1,400,000
500,000
55,000
the purge reactor
(B) a
Annual Operating Costs Total VOC '
Utilities
$1,170,
1,170,
4,310,
4,310,
49,
4,
000
000
000
000
000
,000
and 95% in
Han- Catalyst
power Replacement
Air-Oxidation Process
(A) Emission Cost
ri-trA normrnru riot- Reduction rrr, i 1.
t ixea Recovery wee • - — ~ Effectiveness
Costs Credits Annual Mg/yr Percent (per Mq)
$36,000 $86,000 $754,000 $1,470,000 $ 576,000 12,850 95 $ 45
18,000 86,000
36,000
18,000
5,000
Oxygen-Oxidation Process
4,000
the main-process-vent catalytic
435,000 None 1,709,000 12,850 95 133
899,000 3,531,000 1,714,000 13,390 99 128
406,000 Hone 4,734,000 13,390 99 354
145,000 118,000 81,000 886 97. 3C 91
16,000 55,000 ($31,000)' 382 99.7 (81)'
oxidizer control. <^
H
»J
savings.
-------
VI-8
6000
o
o
o
X
8-
u
•a
-------
VI-9
11,000
10,000 -
80 100
200
300
400 500
800
Plant Capacity (Gg/yr)
(a) Catalytic oxidizer without heat recovery
(b) Catalytic oxidizer with waste heat boiler (100-psig steam)
(c) Thermal oxidizer without heat recovery
(d) Thermal oxidizer with waste heat boiler (250-psig steam)
Fig. VI-2. Net Annual Cost vs Plant Capacity for
Emission Control for Air-Oxidation Process
-------
VI-10
I
w-
-------
VI-11
b. Stripper Purge Vent (Vent B, Fig. III-l) -- Two control methods for this stream
are discussed in Sect. V, compression and recycle to the purge reactor and com-
bining the stream with the main-process-vent stream to its control device.
Only the compression and recycle control has been evaluated; the cost estimate
for the model plant is shown in Table VI-4. The installed capital cost, net
annual cost, and cost-effectiveness variations with capacity are shown in
Figs. VI-1 to VI-3. The recovery credits are based on the estimate that 46% of
the recycled ethylene is reacted in the purge reactor, and the overall VOC
removal efficiency is based on the main process vent being controlled by a
catalytic oxidizer with 95% efficiency.
Combining this vent stream with the main-process-vent (Vent A) stream to its
control device would increase the VOC load to the control by only about 7%.
This incremental increase would not significantly change the cost analysis made
in Sect. B.I.a for the main-process-vent control options.
c. Fugitive Sources -- A control system for fugitive sources is defined in
Appendix C. Another EPA report covers fugitive emissions and their applicable
controls for the synthetic organic chemicals manufacturing industry.
d. Storage and Handling Sources (Vent D, Fig. III-l) -- The system for controlling
storage and handling emissions is an aqueous scrubber. Another EPA report
covers storage and handling emissions for all the synthetic organic chemicals
manufacturing industry.
e. Secondary Sources — No control system has been defined for secondary emissions
from the model plant.
2. Oxygen-Oxidation Process
a. CO Purge Vent (Vent A, Fig. III-2) -- No control system has been defined for
this vent in the model plant.A theoretical design for controlling the emis-
sions is discussed in Sect. V, but additional data would be necessary before
the emission control cost could be estimated.
-------
VI-12
b. Argon Purge Vent (Vent B, Fig. III-2) — The emissions from this vent can be
used readily as a fuel gas. The cost to pipe this stream to an existing fire
box or to incorporate it into a fuel header system would be very site specific,
depending on a particular facility situation. Nevertheless, the cost would be
relatively small compared to the fuel value credit for this vent, and there
would in many cases be a cost-effectiveness savings.
c. Stripper Purge Vent (Vent C, Fig. III-2) -- One emission control for this
source has been evaluated, compression and recycle to the process C0_ absorber.
The emission control cost estimate for the model plant is shown in Table VI-4.
The installed capital cost, net annual cost, and cost-effectiveness variations
with capacity are shown in Figs. VI-4 to VI-6. The recovery credits are based
on an estimate that 50% of the recycled ethylene is reacted. The overall VOC
removal efficiency is based on the unreacted ethylene exiting with the argon
purge vent (Vent B) and the argon purge used as a fuel gas. Disposal of emis-
sions from this vent directly as a fuel gas or to a flare is also discussed in
Sect. V as controls. Disposal as a fuel gas, although comparable in VOC reduc-
tion efficiency to compression and recycle, would not be as cost effective
because it does not provide for recovery of the contained ethylene value.
Flaring this stream would not provide for recovery of ethylene content nor
would there be a heat content value.
d. Fugitive Sources — A control system for fugitive sources is defined in
Appendix C. Another EPA report covers fugitive emissions and their applicable
controls for the synthetic organic chemicals manufacturing industry.
e. Storage and Handling Sources (Vent D, Fig. III-2) — The system for controlling
storage and handling emissions is an aqueous scrubber. Another EPA report
covers storage and handling emissions for all the synthetic organic chemical
manufacturing industry.
f. Secondary Sources — No control system has been defined for secondary emissions
from the model plant.
-------
VI-13
100
80
60
Stripper Purge
Vent Compression
and Recycle
40
20
i
30
40
60 80 100
Plant Capacity (Gg/yr)
200
400
Fig. VI-4. installed Capital Cost vs Plant Capacity for
Emission Controls for Oxygen-Oxidation Process
-------
VI-14
40
o
o
o
X
m
18
cn
m
8
1
to
0
o
i
u)
tn
c
•H
iS
Stripper Purge
Vent Compression
and Recycle
80
120
I I I
i
30
40
60 80 100
Plant Capacity (Gg/yr)
200
400
Fig. VI-5. Net Annual Cost or Savings vs Plant Capacity
for Emission Controls for Oxygen-Oxidation Process
-------
VI-15
40
01
8
ra
ui
3
I
w
4J
01
O
u
£
•rl
80
120
Stripper Purge
Vent Compression
and Recycle
I I I L
i
30
40
60 80 100 200
Plant Capacity (Gg/yr)
400
Fig. VI-6. Cost Effectiveness vs Plant Capacity for
Emission Controls for Oxygen-Oxidation Process
-------
VI-16
C. REFERENCE*
1. Yen-Chen Yen, Propylene Oxide and Ethylene Oxide, Supplement B, Report 2B, A
private report by the Process Economics Program, Stanford Research Institute,
Menlo Park, CA (February 1971).
2. J. A. Keys, IT Enviroscience, Control Device Evaluations. Catalytic Oxidation
(October 1980) (EPA/ESED report, Research Triangle Park, NC).
3. J. W. Blackburn, IT Enviroscience, Control Device Evaluation. Thermal Oxidation
(July 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, Inc., Storage and Handling (September 1980)
(EPA/ESED report, Research Triangle Park, NC).
6. V. Kalcevic, IT Enviroscience, Control Devices Evaluation. Flares and the Use
of Emission as Fuels (in preparation for EPA, ESED, Research Triangle Park, NC)
(August 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.
-------
VI I-1
VII. SUMMARY
Ethylene oxide (EO) is currently manufactured by one of two processes, air
oxidation of ethylene or oxygen oxidation of ethylene. Before 1977 it was also
manufactured by the chlorohydrin process, which was the predominant process
until 1957. The domestic production capacity, including Puerto Rico, of EO for
July 1, 1979, was estimated to be 2783 Gg, with 63% based on the air-oxidation
process and 37% on the oxygen-oxidation process. The 1978 industry utilization
rate was approximately 82%, with about 70% of EO produced being used in the
production of ethylene glycols. The estimated EO consumption annual growth is
4.5 to S.5%.1
Emission sources and uncontrolled and controlled emission rates for the air-
oxidation process model plant are given in Table VII-1 and for the oxygen-
oxidation process in Table VII-2.
The air-oxidation-process major emission source is the main process vent.
Catalytic-oxidation units, located in the process sequence to maximize energy
recovery, are used as controls for this vent. Present catalytic-oxidation
technology can give a 95% VOC reduction efficiency, which may be improved by a
new generation of technology.2 Thermal oxidation could be used to control this
vent with an estimated 99% efficiency but is not as cost effective and may not
fit as well as catalytic oxidation into the overall process energy recovery
design.
The oxygen-oxidation-process major emission that is not controlled is the C02
purge vent, which emits a dilute VOC stream consisting mainly of C02 and water
vapor. A theoretical design consisting of a process change in regenerating the
C02 absorbing fluid has been proposed as a potential method for reduction of
the VOC content,2 but it would require development and its practical removal
efficiency may be only about 60%. The other major emission is from the argon
purge vent, but this stream can readily be used as a fuel gas.
XS A Coaswell "Ethylene Oxide," pp. 654.5031A—654.5033F in Chemical Economics
Handbook Stanford Research Institute, Menlo Park, CA (January 1980).
2J F Lawson IT rn-ir—'-""" THP ReP°rt for Visit to Union Carbide
Corp!. South'cha^^nr WV. Dec": 7, 1977 (on file at EPA, ESED, Research
Triangle Park, NC).
-------
VII-2
Table VII-1. Emission Summary for Air-Oxidation-Process Model Plant
(227,000 Mg/yr)
Emission Source
Main process vent
Stripper purge vent
Fugitive
Storage and handling
Secondary
Total
Stream
Designation
(Fig.III-1)
A
B
C
D
E
VOC Emission
Uncontrolled
1544
104
10.8
68
0.3
1727
Rate (kg/hr)
Controlled
77a
2.8b
5.4
0.34
0.3
85.8
Based on catalytic oxidation as the control.
3Based on compression and recycle to the process as the control.
Table VII-2. Emission Summary for Oxygen-Oxidation-Process Model Plant
(136,100 Mg/yr)
Emission Source
CO purge vent
Argon purge vent
Stripper purge vent
Fugitive
Storage and handling
Secondary
Total
Stream
Designation
(Fig.III-2)
A
B
C
D
E
F
VOC Emission Rate (kg/hr)
Uncontrolled
62
170
44
8.7
41
0.2
326
Controlled
62
0.17
0.13
3.3
0.21
0.2
66.0
-------
VII-3
The emissions from the stripper purge vent in both processes are readily con-
trolled by being compressed and recycled to an appropriate place in the
process. Storage and handling vent streams can be scrubbed for EO recovery and
recycled back to the process. Potential secondary emissions are minor.
-------
A-l
APPENDIX A
Table A-l. Properties of Ethylene*
Synonyms
Molecular formula
Molecular weight
Physical state
Vapor pressure
Vapor density
Boiling point
Melting point
Density
Water solubility
Acetene, ethene
C2H4
28.06
Gas
456 mPa at 0°C
0.98
-103.9°C at 760 mm Hg
-169°C
0.99267 at 20°C/4°C
Insoluble
*From: J. Dorigan et_ al^. , "Ethylene Oxide," p. AII-260 in Scoring of Organic
Air Pollutants. Chemistry, Production and Toxicity of Selected Synthetic
Organic Chemicals (Chemicals D-E), Rev. 1, Appendix A, MTR-7248, MITRE Corp.,
McLean, VA (September 1976).
-------
Table A-2. Properties of Ethylene Oxide*
Synonyms 1,2-Epoxyethane, oxirane
Molecular formula C H.O
2 4
Molecular weight 44.05
Physical state Gas
Vapor pressure 197 kPa at 25 °C
Vapor density 1.52
Boiling point 13.5°C at 746 mm Hg
Melting point -111.3°C
Density 0.8711 at 20°C/20°C
Water solubility Soluble
_
From: J. Dorigan ejb _al., Ethylene Oxide," p. AII-304 in Scoring of Organic
Air Pollutants. Chemistry, Production and Toxicity of Selected Synthetic
Organic Chemicals (Chemicals D-E), Rev. 1, Appendix A, MTR-7248, MITRE Corp.,
McLean, VA (September 1976).
-------
B-l
APPENDIX B
Table B-l. Air-Dispersion Parameters for Air-Oxidation-Process Model
Source
Main process vent
Stripper purge vent
Fugitive emissions*
Storage and handling
(6 spheres, 2
horizontal tanks)
Secondary emissions
Main process vent
Catalytic oxidizer
Thermal oxidizer
Stripper purge vent
(with main process
vent)
Fugitive emissions*
Storage and handling
with
VOC
Emission
Rate
(g/sec)
429
28.9
3.0
18.9
0.083
21.4
4.2
0.8
1.5
0.09
a Capacity of 227,000 Mg/yr
Discharge Flow
Height Diameter Temperature Rate
(m) (m) (K) (m3/sec)
Uncontrolled Emissions
17 0.6 310 37.5
30 0.1 320 0.6
295
9.7 9.7 283
295
Controlled Emissions
15 3 530 76.8
15 3 530 84.9
295
18 0.1 310 0.01
Plant
Discharge
Velocity
(m/sec)
132.5
76.5
10.9
12.0
1.3
*Fugitive emissions are distributed over a rectangular area of 100 m X 400 m.
-------
B-2
Table B-2. Air-Dispersion Parameters for Oxygen-Oxidation-Process Model Plant
with a Capacity of 136,100 Mg/yr
Source
CO purge vent
Argon purge vent
Stripper purge vent
a
Fugitive emissions
Storage and handling
(5 spheres, 3
horizontal tanks)
Secondary emissions
Argon purge vent
Stripper purge vent
a
Fugitive emissions
Storage and handling
VOC
Emission
Rate
(g/sec)
17.2
47.2
12.2
2.4
11.4
0.056
0.05
0.04
0.92
0.06
Discharge
Height Diameter Temperature
(m) (m) (K)
Uncontrolled Emissions
20 0.3 375
20 0.3 310
20 0.1 310
295
9.7 9.7 283
295
Controlled Emissions
18 0.1 310
Flow Discharge
Rate Velocity
(m-Vsec) (m/sec)
6.0 84.2
0.13 1.8
0.025 3.2
0.006 0.8
Fugitive emissions are distributed over a rectangular area of 100 m X 300 m.
|3
Vents used as fuel gas.
-------
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.
Uncontrolled
Emission Factor
Controlled
Emission Factor'
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
(kg/hr)
0.12
0.02
0.021
0.010
0.0003
0.16
0.006
0.009
0.44
0.00026
0.032
(kg/hr)
0.03
0.02
0.002
0.003
o.od'bs
0.061
0.006
0.009
0.11
0.00026
0.019
aBased on monthly inspection of selected equipment; no inspection of
heavy-liquid equipment, flanges, or light-liquid relief valves;
10,000 ppmv VOC concentration at source defines a leak; and 15 days
allowed for correction of leaks.
bLight liquid means any liquid more volatile than kerosene.
*Radian Corp Emission Factors and Frequency of Leak Occurrence for Fittings
in RefineryProcess Units, EPA 600/2-79-044 (February 1979).
-------
D-l
APPENDIX D
COST ESTIMATE DETAILS
This appendix contains sample calculations showing how the costs presented in
this report were estimated.
The accuracy of an estimate is a function of the degree of data available when
the estimate was made. Figure D-l illustrates this relationship. The contin-
gency allowance indicated is included in the estimated costs to cover the
undefined scope of the project.
Capital costs given in this report are based on a screening study, as indicated
by Fig. D-l, based on general design criteria, block flowsheets, approximate
material balances, and data on general equipment requirements. These costs
have an accuracy range of +30% to -23%, depending on the reliability of the
data, and provide an acceptable basis to determine the most cost-effective
alternate within the limits of accuracy indicated.
A. MAIN-PROCESS VENT AIR-OXIDATION PROCESS
To determine the cost estimates for controlling the vent emissions from the
air-oxidation process main-process vent (Vent A, Fig. III-l), emission flow
details for the model plant were taken from Tables IV-1 and IV-2 and the waste-
gas flow calculated as shown in Table D-l:
10,859 Ib-moles/hr X 359 scf/lb-mole 4- 60 min/hr = 65,000 scfm of waste gas.
Multiplying the VOC components by their heating value and dividing by the SCF
of waste gas gives the heating value-.
<337 X 22,!70 + 2992 X 21,770 + 75 X 12,630) . ^^ = ig ^^ Qf ^
bU
The waste-gas flows of EO plants larger and smaller than the model plant were
estimated by direct ratio to the model plant:
100-Gg/yr plant = 100/227 X 65,000 = 29,000 scfm of waste gas.
500-Gg/yr plant = 500/227 X 65,000 = 143,000 scfm of waste gas.
-------
IWPoRMATIOSJ USED BY ESTIMATOR
ESTIMATE. TYPE
. PROS.
covr
SCREEKllMG^
(PREUM. EWCq- ^tUOY)
PHA«bE IT
(PRE.LIM. PROC. EMG,.)
PHASE nr
(COMPLETE PROCESS
EWCj. DESI^U")
•
•
•
•
\
^
\
\\
\
\
\\
3 1 E 3 4
APPROX. COST
EUGR.E E^T.
\
EST/MATED COST
' ALLOWANCE
MAX. PROS.
COST
(•/• OP TOTAU
PROBABUE
CAP. COST)
-fco -4o -ZO 0 2o 4C
RAUG(E - PROBA.BLE.
ACTUAL PROJECT
COST C^
feo
o jo zo *>
ft> Au.owA.uce
TO iMCLUDE
4c
Fig. D-l. Precision of Capital Cost Estimates
LAT6W
-------
D-3
Table D-l. Emission Flow Data
Flow Rate
Component
Ethane
Ethylene
EO
N2
0,,
2
CO
2
V
Total
(A)
Molecular
Weight
30
28
44
28
44
18
Weight
Percent
0.09.
0.80
0.02
BO. 22
3.07
15.65
0.15
100
(kg/hr)
153
1,357
34
136,110
5,209
26,553
255
169,671
(Ib/hr)
337
2,992
75
300,071
11,484
58,539
562
374,060
(B)
Adjusted
(Ib/hr) a
337
2,992
75
262,593
58,539
562
325,098
(Ib-moles/hr)
11
107
2
9,378
1,330
31
10,859
^Adjusted to a no-oxygen basis by deducting 3.26 Ib of N2 per Ib of oxygen, as well as
deducting the contained oxygen.
* (A) = (C).
-------
D-4
1. Catalytic Oxidizer
The costs given in Table D-2 were developed from the control device evaluation
report.
2. Thermal Oxidizer
The costs given in Table D-3 were developed from the control device evaluation
2
report. In Sect. V-A.l.b of this report the stated oxidation conditions are
870°C (1600°F) and 0.5-sec residence time for 99% destruction.
B. STRIPPER PURGE VENT AIR-OXIDATION PROCESS
For the model plant it was calculated that a 250-hp compressor handling
1100 scfm (240 acfm) would be required. The estimated December 1979
installation cost for the system is $500,000. The cost and VOC reduction data
for this vent are shown in Table D-4.
C. STRIPPER PURGE VENT OXYGEN-OXIDATION PROCESS
For the model plant it was calculated that a 20-hp compressor handling 53 scfm
(12 acfm) would be required; the size of this compressor can be very site
specific, depending on the plant operating conditions. One producer calculates
that this compressor would be 200 hp for their conditions. The estimated
December 1979 installation cost for the system is $55,000. The control cost
data are shown in Table D-5.
-------
Table D-2. Catalytic Oxidizer Control Cost Data for Air-Oxidation Process
Plant
Size
(Gg)
Waste-
Gas Rate
(scfm)
Installed
Capital
Cost3
b
Fixed
o
Catalyst
Annual
Fuel
Cost (X
Power
1000)
.Man-
power
Credit
Net
No Heat Recovery
100
227
500
29,000
65,000
143,000
$ 780
1,500
2,9009
$ 226
435
841
$ 41
86
176
$ 478
1086
2392
$ 37
85
186
With 100-psig Steam Waste
100
227
500
29,000
65,000
143,000
1,400
2,600
4,800b
406
754
1,392
41
86
176
478
1,086
2,392
37
85
186
$18
18
18
$ 800
1710
3613
voc
Reduction
(Mg/yr)
$ 5,660
12,850
28,300
Cost
E f f ect i veness
(per Mg)
$141
133
128
Heat Generator
36
36
36
$ (647)
(1468)
(3233)
351
579
949
5,660
12,850
2,830
62
45
34 7
en
Interpolated from Fig. IV-2. of ref 1.
From Table VI-5 of this report, 29% of installed capital.
CFrom Table II-l of ref 1. Flue gas to waste gas ratio is 1.58; this value times the waste-gas rate gives the flue-gas
rate. Then interpolating from Fig. A-2, the purchase cost of catalyst is obtained. Adding a 20% installation cost and
dividing by a 3-yr catalyst life (example, p. B-3) gives the annual catalyst cost.
dFrom Table II-l of ref 1 15.9 Btu/scf is required at $2.00/thousand ft (Table VI-5 of this report); this yields the
annual cost.
Calculated as in examples on p. B-3 of ref 1.
fFrom Fig. II-2 of ref 1. Flue-gas heat content at 1150°F is 22 Btu/scf and at 500°F exit boiler temperature the heat
recovery is 62%, or 13.6 Btu/scf of flue gas or 13.6 X 1.58 = 21.5 Btu/scf of waste gas. Then from the waste-gas
flow and the value of steam from Table VI-5-the credit is calculated.
^Assumed to be two half-size units at 85% of individual unit cost.
-------
Table D-3. Thermal Oxidizer Control Cost Data for Air-Oxidation Process
Plant
Size
-------
Table D-4. Stripper Purge Vent Control Cost Data for Air-Oxidation Process
Plant
Size
(Gg)
100
227
500
VOC
Reduction
(Mg/yr)
385
875
1925
Installed
Capital
Cost
$305
500
800
Annual Cost (X 1000)
Fixed
Cost
$ 88
145
232
Utilities
$ 22
49
108
Manpower
$5
5
5
Recovery
Credit
$ (52) a
(118) a
(260) a
Net
$63
81
85
Cost
Effectiveness
(per Mg)
$164
93
44
Savings.
D
-------
Table D-5. Stripper Purge Vent Control Cost Data for Oxygen-Oxidation Process
Plant
Size
(Gg)
40
136
250
VOC
Reduction
(Mg/yr)
112
382
707
Installed
Capital
Cost
$26
55
80
Annual Cost (X 1000)
Fixed
Cost
$ 8
16
23
Utilities
$1
4
7
Manpower
$4
4
4
Recovery
Credit
$(16)a
(55) a
(101) a
Net
$(31)
(31)
(67)
Cost
Effectiveness
(per Mg)
$(30)
(81)
(95)
lumbers in parentheses reflect a savings.
D
00
-------
D-9
D. REFERENCES*
1. J. A. Key, IT Enviroscience, Control Device Evaluation. Catalytic Oxidation
(October 1980) (EPA/ESED report, Research Triangle, Park, NC).
2. J. w. Blackburn, IT Enviroscience, Control Device Evaluation. Thermal Oxidation
(July 1980) (EPA/ESED, Research Triangle Park, NC).
3. J. Starkey, Northern Petroleum Co., letter dated May 2, 1979, to EPA with
information on 60 processes.
*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 oaraaraoh that reference number is indicated on the material involved.
wSen ?heReference appears on a heading, it refers to all the text covered by
that heading.
-------
E-l
APPENDIX E
EXISTING PLANT CONSIDERATIONS
Table E-l lists some emission control devices reported to be used by industry.
To gather information for the preparation of this report, three site visits
were made to manufacturers of EO. Trip reports have been cleared by the
companies concerned and are on file at EPA, ESED, in Research Triangle Park,
NC. — Some of the pertinent information concerning process emissions from
existing EO plants is presented in this appendix. Pertinent information was
also obtained from the Chemical Manufacturers Association and from some
producers who submitted comments in response to the draft of this report issued
in November 1978.
A. PROCESS EMISSIONS FROM EXISTING PLANTS
1. BASF Wyandotte Corporation, Geismar, LA
The ethylene oxide facility consists of two trains of equipment, one con-
structed in 1957 but with later major revisions and the other constructed in
1967. The following are reported emissions:
a. Tail Gas Absorber The light ends from this residual absorber are recycled to
the ethylene oxide absorber. Its composition is shown below:
Component
r\
°2
C H
2 2
CH4
co2
N_
2
Ar
H 0
Amount
Unit 1
1.0
26.5
11.5
54.2
0.4
5.8
0.4
(vol %)
Unit 2
0.9
21.8
9.5
62.5
0.5
4.8
-------
Table E-l. Emission Control Devices Used by Some Domestic Ethylene Oxide Producers
Control Devices Used
Source
By Union Carbide Corp.'
(Air-Oxidation)
By Celanese Chemical Co.
(Oxygen-Oxidation)
By BASF Wyandotte Corp.C
(Oxygen-Oxidation)
Main process vent
CO purge vent
Argon purge vent
Stripper purge vent
Fugitive
Storage and handling
Secondary
aSee ref 1.
See ref 2.
See ref 3.
Catalytic oxidation
NA
NA
Compression and recycle
Engineering design and
operating measures
Scrubbers and flare
NR
Not applicable (NA)
Transferred to another process
Utility boiler fuel with flare
alternate
Compression and recycle
Not reported (NR)
Scrubber
Cooling tower mist eliminators
NA
None
Utility boiler fuel
Compression and recycle
NR
Flare
NR
H
ro
-------
E-3
Other characteristics are as follows:
Temperature: unit 1, 85°F; unit 2, 85°F
Pressure: unit 1, 48 psig; unit 2, 54 psig
Flow rate: unit 1, 1050 lb/hr; unit 2, 800 Ib/hr
This overhead stream is vented to the atmosphere for approximately 5 to 10 min
three to four times a year. This conditioning is caused by a gas overload on
the compressor.
b. Main Process Vent — This vent, which is a slipstream of the recycle stream to
the reactors, is piped to the utility boilers. The composition is as follows:
Component
0
2
C2H4
CH4
co2
C H
26
N2
Ar
H_0
2
Amount
Unit 1
4.1
23.0
47.0
9.0
0.1
9.0
7.0
0.9
(vol %)
Unit 2
5.5
23.0
51.0
11.0
0.1
3.7
4.8
0.9
The flow rate for unit 1 is 1800 Ib/hr and for unit 2 is 650 Ib/hr.
c. Carbon Dioxide Purge Vent -- This vent goes to the atmosphere. The outlets from
both systems have a temperature of 215°F. The composition of this vent from
the two plants is as follows:
Component
Water
co2
Ethylene
Weight
Unit 1
53.3
46.5
0.2
Percent
Unit 2
53.3
46.5
0.2
-------
E-4
The flow rate for unit 1 is 31,360 Ib/hr and for unit 2 is 23,300 Ib/hr, On a
water-free basis the CO from the strippi
mole % ethylene and 0.44 mole % methane.
water-free basis the CO from the stripper overhead contains approximately 0.84
2. Celanese Chemical Company, Pasadena, TX
The ethylene oxide facility consists of a single train of equipment with four
parallel reactors. The emissions from four process sources are given in
Table E-l.
B. TOTAL INDUSTRY EMISSIONS
Emissions from industry were estimated based on the control measures reported
by plants and on the assumption that similar control measures exist for the
other plants. It was estimated that secondary emissions are uncontrolled for
all plants and that maintenance programs required for safety reasons in EO
plants result in controlled fugitive emissions.
Based on the above, total emissions from all plants during 1978 were approxi-
mately 8 Gg. The emissions from these plants would have been approximately
113 Gg if they had been uncontrolled.
C. 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 retro-
fit emission control systems in existing plants than to install a control
system during construction of a new plant.
-------
E-5
Table E-2. Emission Data for Process Vents
Capacity
Component
CO
2
C H
2 6
C2H4
CH,,
4
N
2
0_
2
Ar
Total
co2
C H
2 4
Total
N2
Weight
Percent
Adsorption/Desorption
11.7
0.3
29.4
24.0
8.4
5.7
20.4
Carbon Dioxide Removal
99.44
0.56
Vent Absorber
99.9+
(lb/1000
85%
System Vent
1.36
0.02
2.80
2.80
1.00
0.67
2.37
11.02
System Vent
628.0
3.5
"'• " '
631.5
Vent
443 (Ib/hr)
Ib of EO)
100%
1.39
0.04
3.49
2.84
1.00
0.63
2.42
11.81
678.3
3.8
« i.i m-i_ -i
682.1
Pressure Swing Adsorption Vent
CH
4
N
2
CO.
2
C
2
c"
3
C~4~
Total
91.2
0.6
2.0
4.9
0.2
0.9
8.52
0.05
0.20
0.47
0.02
0.10
9.36
8.84
0.06
0.19
0.48
0.02
0.10
9.69
-------
E-6
E. REFERENCES*
1. J. F. Lawson, IT Enviroscience, Inc., Trip Report for Visit to Union Carbide
Corp., South Charleston, WV, Dec. 7, 1977 (on file at EPA, ESED, Research
Triangle Park, NC).
2. J. F. Lawson, IT Enviroscience, Trip Report for Visit to Celanese Chemical Co.,
Clear Lake City, TX, June 21 and 22, 1977 (on file at EPA, ESED, Research
Triangle Park, NC).
3. J. F. Lawson, IT Enviroscience, Trip Report for Visit to BASF Wyandotte Corp.,
Geismar, LA, July 11, 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.
-------
5-i
REPORT 5
VINYL ACETATE
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.
-------
5-iii
CONTENTS OF REPORT 5
Page
I- ABBREVIATIONS AND CONVERSION FACTORS 1-1
II. INDUSTRY DESCRIPTION II-l
A. Introduction II-l
B. Vinyl Acetate Usage and Growth II-l
C. Domestic Producers II-l
D. References II-4
III- PROCESS DESCRIPTION III-l
A. Introduction III-l
B. Ethylene Vapor-Phase Process: Bayer III-l
C. Acetylene Process III-5
D. References III-6
IV. EMISSIONS IV-1
A. Typical Plant IV-1
B. Sources and Emissions IV-1
C. Reference IV-6
V. APPLICABLE CONTROL SYSTEMS V-l
A. Inert-Gas Purge Vent V-l
B. C02 Vent V-l
C. Emergency Vents V-l
D. Light-Ends and Inhibitor-Tank Vents V-3
E. Storage and Handling Sources V-3
F. Current Emission Controls V-3
G. References v'4
VI. IMPACT ANALYSIS VI"1
A. Environmental Impact VI~1
B. References VI~4
APPENDICES OF REPORT 5
A. PHYSICAL PROPERTIES OF VINYL ACETATE, ACETALDEHYDE, METHYL ACETATE, A-l
AND ETHYL ACETATE
B. EXISTING PLANT CONSIDERATIONS B-l
-------
5-v
TABLES OF REPORT 5
Number
II-l Vinyl Acetate Usage and Growth
II-2 Vinyl Acetate Capacity
IV-1 Emissions from Uncontrolled Typical Plant
IV-2 Typical Plant Storage
V-l Emissions from Controlled Typical Plant
VI-1 Environmental Impact of Control
VI-2 Emission Ratios for Typical Plant and Industry
A-l Physical Properties of Vinyl Acetate
A-2 Physical Properties of Acetaldehyde
A-3 Physical Properties of Methyl Acetate
A-4 Physical Properties of Ethyl Acetate
B-l Emission Control Devices Currently Used
B-2 Direct Emissions (Union Carbide)
B-3 Secondary Emissions (Union Carbide)
B-4 Direct Emissions (Celanese
B-5 Secondary Emissions (Celanese)
B-6 Direct Emissions (Du Pont)
Page
II-2
II-3
IV-2
IV-4
V-2
VI-2
VI-3
A-l
A-2
A-3
A-4
B-2
B-3
B-3
B-5
B-5
B-6
FIGURES OF REPORT 5
Number
III-l Process Flow Diagram
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 UTiits (SI) abbreviations
and conversion factors for this report.
To Convert From
Pascal (Pa)
Joule (J)
Degree Celsius (°C)
Meter (m)
Cubic meter (m3)
Cubic meter (m3)
Cubic meter (m3)
Cubic meter/second
(m3/s)
Watt (W)
Meter (m)
Pascal (Pa)
Kilogram (kg)
Joule (J)
To
Atmosphere (760 mm Hg)
British thermal unit (Btu)
Degree Fahrenheit (°F)
Feet (ft)
Cubic feet (ft3)
Barrel (oil) (bbl)
Gallon (U.S. liquid) (gal)
Gallon (U.S. liquid)/min
(gpm)
Horsepower (electric) (hp)
Inch (in.)
Pound-force/inch2 (psi)
Pound-mass (Ib)
Watt-hour (Wh)
Standard Conditions
68°F = 20°C
1 atmosphere = 101,325 Pascals
PREFIXES
Multiply By
9.870 X 10"6
9.480 X 10"4
(°C X 9/5) + 32
3.28
3.531 X 101
6.290
2.643 X 102
1.585 X 104
1.340 X 10"3
3.937 X 101
1.450 X 10~4
2.205
2.778 X 10"4
Prefix
T
G
M
k
m
M
Symbol
tera
giga
mega
kilo
milli
micro
Multiplication
Factor
1012
109
106
103
10'3
10"6
Example
1 Tg = 1 X 1012 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 pg = 1 X 10"6 gram
-------
II-l
II. INDUSTRY DESCRIPTION
A. INTRODUCTION
Vinyl acetate production was selected for study because preliminary estimates
indicated total emissions of volatile organic compounds (VOC) from the industry were
relatively high and because an increase in consumption was expected to continue.
Vinyl acetate is a flammable liquid boiling at 72.2°C. It tends to polymerize
and must be suitably inhibited during storage and handling. Appendix A gives
the pertinent physical properties of vinyl acetate.
B. VINYL ACETATE USAGE AND GROWTH
The production of vinyl acetate in the United States for 1976 was 656 Gg. Vinyl
acetate consumption is expected to increase at an average annual rate of 5 to
7.5% and to reach 880 to 1012 Gg in 1982.:
The uses of vinyl acetate and their expected growth rates are given in Table II-l.
The major end uses are in the paint and the paper manufacturing industries. Twenty
three percent of production is exported.1
C. DOMESTIC PRODUCERS
Currently vinyl acetate is being produced at seven manufacturing sites. Some
capacity data for these sites were obtained from the literature,1 some from
site visits,2,3 and some from responses to EPA requests for information.4 6
Table II-2 lists the current producers, their plant locations, their capacities,
and the processes employed. Vinyl acetate production requires either captive
or merchant supply of the hydrocarbon raw material (ethylene or acetylene);
therefore a Gulf Coast location is economically favored.
With a total domestic industry capacity of 1060 Gg/yr and a demand projection
of 880 to 1012 Gg/yr for 1982 it is likely that there will be some capacity
added within this five-year period.7'8 The economics of the ethylene and the
acetylene processes indicates that the ethylene process will more likely be
chosen for the added capacity and that it will ultimately replace the acetylene-
based production.
-------
II-2
a
Table II-l. Vinyl Acetate Usage and Growth
End Use
Polyvinyl acetate
emulsions and resins
Polyvinyl alcohol
Polyvinyl butyral
Vinyl chloride
Production for
1976
(%)
47
17
5
4
Average Growth.
for 1976—1982
(%/yr)
5—7
7—9
3—5
3—5
copolymers
Ethylene/vinyl acetate 4 12—15
resins
Other uses 1 3—5
Exports 23 5—7
£
Average 5—7.5
See ref. 1.
Projection from 1981—1982 is estimated to be at same rate as
from 1976—1981.
£»
Weighted arithmetic average.
-------
II-3
Table II-2. Vinyl Acetate Capacity
Producer
Borden Inc.
Celanese Chemical Co.
E. I. du Pont de Nemours and Co.
National Distillers and
Chemical Corporation
U. S. Industrial Chemicals
Co. Division
National Starch and
Chemical Corp.
Union Carbide Corp.
Total
Location
Geismar, LA
Bay City, TX
Clear Lake, TX
LaPorte, TX
Deer Park, TX
Long Mott, TX
Texas City, TX
Capacity
(Gg/yr)
68a
180+b
180a
180°
270d
236
159f
1060
Process
Acetylene
Ethylene, vapor
Ethylene , vapor
Ethylene, vapor
Ethylene , vapor
Acetylene
Ethylene, vapor
phase
phase
phase
phase
phase
See ref. 1.
DSee ref. 2.
"See ref. 4.
See ref. 5,
"See ref. 6.
See ref. 3.
-------
II-4
D. REFERENCES*
1. H. E. Frey et al., "CEH Marketing Research Report on Vinyl Acetate, Polyvinyl
Alcohol," pp. 580.1871A--580.1872Z in Chemical Economics Handbook, Stanford Research
Institute, Menlo Park, GA (September 1977).
2. S. W. Dylewski, Hydroscience, Inc., Trip Report for Visit to Celanese Chemical
Company, Bay City, TX, Sept. 28 and 29, 1977 (on file at EPA, ESED, Research
Triangle Park, NC).
3. S. W. Dylewski, Hydroscience, Inc., Trip Report on Information Obtained on Vinyl
Acetate Process Used at Union Carbide Corporation, Texas City, TX Plant, Dec. 8, 1977
(on file at EPA, ESED, Research Triangle Park, NC).
4. D. W. Smith, letter to EPA from E.I. du Pont de Nemours and Company, Inc., La
Porte, TX, Sept. 18, 1978, in response to EPA's request for information on the
vinyl acetate process.
5. K. G. Carpenter, letter from U.S. Industrial Chemicals Company, Deer Park, TX,
Aug. 17, 1978, in response to EPA's request for information on the vinyl acetate
process.
6. E. W. Bousquet, letter from National Starch and Chemical Corporation, Long Mott,
TX, Aug. 22, 1978, in response to EPA's request for information on the vinyl
acetate process.
7. "Chemical Profile on Vinyl Acetate," p. 9 in Chemical Marketing Reporter,
Oct. 6, 1975.
8. "Chemical Profile on Vinyl Acetate," p. 9 in Chemical Marketing Reporter,
Oct. 9, 1978.
*A reference located at the end of a paragraph usually refers to the entire para-
graph. If another reference relates to certain portions of the paragraph, the
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 two processes are currently in use for the manufacture of
vinyl acetate. The ethylene—acetic acid—oxygen vapor-phase (ethylene vapor-
phase) process accounts for about 92% of the production (see Table I1-2). The
acetylene—acetic acid vapor-phase (acetylene) process accounts for the balance
of the production. This has changed markedly from 1970, when the acetylene
process accounted for most of the world's production.1
Both Bayer AG (Bayer) and U.S. Industrial Chemicals (U.S.I.) developed ethylene
vapor-phase processes separately and have offered to license to others. The
ethylene-based processes are favored over the acetylene process because ethylene
is a lower cost raw material.2
Other processes have been used, such as the liquid process based on ethylene.
Still other processes show promise for the future, such as the one that uses
only ethylene and oxygen as raw materials.2
B- ETHYLENE VAPOR-PHASE PROCESS: BAYER
1. Chemistry
Ethylene reacts with acetic acid and oxygen in the presence of palladium
catalysts1 to form vinyl acetate as shown in reaction 1. About 90% of the
ethylene reacts in this fashion. About 10% of the ethylene is oxidized to C02
according to reaction 2. Less than 1% of the ethylene is oxidized to other
species such as acetaldehyde, ethylacetate, and methyl acetate, of which
acetaldehyde is predominant (see reaction 3).
Reaction 1:
H2C=CH2
(ethylene)
CH3COOH
(acetic acid)
1/202
(oxygen)
CH3COOCH=CH2 + H20
(vinyl acetate) (water)
-------
III-2
Reaction 2:
H2C=CH2 + 302 * 2C02 + 2H20
(ethylene) (oxygen) (carbon dioxide) (water)
Reaction 3:
H2C=CH2 + 1/202 > CH3CHO
(ethylene) (oxygen) (acetaldehyde)
2. Process Description
The process flow diagram shown in Fig. III-l represents a typical continuous
ethylene vapor-phase process under the Bayer license.1 4 Recycled and fresh
acetic acid (Stream 1) and recycled and fresh ethylene (Stream 2) are fed to a
vaporizer.
The heated gaseous mixture is combined with oxygen to form the desired reaction
mixture (Stream 3). The organic content of the mixture is controlled at about
85% and the oxygen content is controlled at about 7% to keep the gas composition
outside the explosive range. Other components of the gas stream are C02, H20,
and inert gases.2
The gaseous reaction mixture (Stream 3) is fed to the reactor containing a fixed-
bed suspended catalyst that includes palladium, gold, and salts of potassium.
Because of the low oxygen and high ethylene content of the reaction mixture,
only about 10% of the ethylene is converted per pass.2
The reactor discharge (Stream 4) is first passed through an energy recovery
step to make use of the heat liberated in the reaction and then to the first
gas-liquid separator.
The gas, mainly ethylene laden with acetic acid and vinyl acetate vapors
(Stream 5), is scrubbed with acetic acid to absorb the condensable materials
and to combine them with the reactor liquid stream.
-------
CCJ,
STR.IPPE.R.
/^J
SEPARATORS V/Y
t
'
1
1
1
ACETI
PECOX
\
C ACID
'ERV COLUMKJ
WATER
STRIPPER
(Q TO WATER
I
1
P
1 r
H
1
1 L,
V y
VA/ HX0
SEPARATOR.
^ . .,_.
>
7\
5
0^
^-
1
^X
r
INHIBITOR
Li<3(WT-eKao;
COUUMKl
1
if®
^- —
v- — _
1
- M/\MDUKJG,
@ -FU&tTIVE
VlklYL ACETATe
COLUMKJ
VIMYL
STORAAE
Fig. III-l. Flow Diagram for Manufacture of Vinyl Acetate by the Ethylene Vapor-Phase Process
-------
III-4
The resultant gas (Stream 6) is recompressed and then recycled back to the
reactor (Stream 7). A purge of recycle gas is fed to the carbon dioxide
removal system where it is scrubbed with water and fed to the carbonate system
to absorb the CO . The inert gases ethane, nitrogen, and argon that accompany
the feeds to the reactor are purged at vent A. The carbonate stream is strip-
ped of the absorbed C02 by the release of pressure and the application of heat
in the C02 stripper. The gas stream resulting from the stripping operation is
composed mainly of C02 and is discharged at vent B.
The reactor liquid (Stream 8), still under pressure, is passed through a pressure
let-down valve and a second gas-liquid separator. The flash gas (Stream 9) is
recompressed and then returned to the reactor; the remaining liquid (Stream 10)
is sent to distillation.
During emergency and planned shutdowns pressure-relief valves will release some
VOC at the vent marked C.
The distillation steps recover, in sequence, unreacted acetic acid for recycle
(Stream 11), water that is made in the reaction and that is used in gas scrub-
bing (Stream 12), light ends such as acetaldyhyde (Stream 13), and vinyl acetate
as a finished product (Stream 14). The low-boiling nature of acetaldehyde and
other light ends may result in some discharge during distillation. Polymerization
inhibitors, introduced during distillation, are dissolved in mix tanks which
generate some VOC emission. These emissions are discharged at vent D.
Polymer wastes, formed in reaction and in distillation, are carried in acetic
acid in stream 11, and after two stages of vaporization the solids are dis-
charged at stream I. Water formed in the reaction is discharged at stream J.
The small amount of acetaldehyde and ethyl acetate is discharged at streams K
and L, respectively.
Process Variations
The U.S.I, process is quite similar to the Bayer process. The difference is
in the reaction, where the U.S.I, process produces a wider variety of by-products,
and in the scrubbing steps, in which a glycol diacetate scrubber is used in
place of the water scrubber. The steps for the two processes can essentially
be represented by the same flow sheet.5
-------
III-5
C. ACETYLENE PROCESS
Acetylene and acetic acid react readily in the presence of zinc acetate to form
vinyl acetate according to the following reaction:
HCHCH + CH3COOH catalyst> CH3COOCH=CH2
(acetylene) (acetic acid) (vinyl acetate)
The reaction is conducted in the vapor phase, with both raw materials being
present in excess. The unconverted raw materials are recovered and recycled.
Ethylene, present as an impurity in the acetylene and being essentially inert,
must be purged from the reaction system. This is done by discharging a portion
of the unreacted acetylene. Acetaldehyde is also formed in the reaction and
must be disposed of. The steps following the reaction are simpler than in the
ethylene vapor-phase process because of the absence of oxidation by-products.6
Since the industrial use of this process is now minimal and is expected to be
phased out, it will not be discussed further in this study.
-------
III-6
D. REFERENCES*
1. D. Rhum, "Poly (vinyl acetate)," pp. 317—353 in Kirk-Othmer Encylcopedia of
Chemical Technology, 2d ed., vol 21, edited by A. Stenden e_t al., Interscience,
New York, 1970.
2. Yen-Chen Yen, Vinyl Acetate Supplement^, pp. 21—99 in Report No. 15, A private
report by the Process Economics Program, Stanford Research Institute, Menlo
Park, CA (June 1972).
3. S. W. Dylewski, Hydroscience, Inc., Trip Report for Visit to Celanese Chemical
Company, Bay City, TX, Sept. 28 and 29, 1977 (on file at EPA, ESED, Research
Triangle Park, NC).
4. S. W. Dylewski, Hydroscience, Inc., Trip Report on Information Obtained on Vinyl
Acetate Process Used at Union Carbide Corporation, Texas City, TX Plant, Dec. 2, 197?
(on file at EPA, ESED, Research Triangle Park, NC).
5. J. W. Pervier, Houdry Division Air Products and Chemicals, Inc., "Vinyl Acetate
via Ethylene," pp. VAC-1--VAC-6 in Survey Reports On Atmospheric Emissions
from the Petrochemical Industry, Vol IV, EPA-450/3-73-005-d (April 1974).
6. J. W. Pervier, Houdry Division Air Products and Chemicals, Inc., "Vinyl Acetate
via Acetylene" pp. VI-1--VI-7 in Survey Reports on Atmospheric Emissions from
the Petrochemical Industry, Vol IV, EPA-450/3-73-005-d (April 1974).
*A reference located at the end of a paragraph usually refers to the entire para-
graph. If another reference relates to certain portions of the paragraph, the
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 atmos-
phere, 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 form 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.
&. TYPICAL PLANT
The typical plant developed for this study has a capacity of 160 Gg/yr, based
on 8760 hr* of operation per year. Although not an actual operating plant,
it is typical of plants recently built. The plant utilizes the ethylene
vapor-phase process licensed by Bayer and best fits today's manufacturing and
engineering technology for vinyl acetate production.
The quality of the raw materials used in this study is typical of production
from recently built facilities. The composition of ethylene was taken as 99.9%
ethylene and the balance assumed to be ethane. The composition of oxygen was
taken as 99.4% 02 and the balance assumed to be nitrogen and argon.
B. SOURCES AND EMISSIONS
Uncontrolled emissions rates from process and storage sources in vinyl acetate
production are summarized in Table IV-1 and are discussed below. The discharge
locations are shown in Fig. III-l. Emissions presented in this section are
based on plant trips, letters, and engineering judgment based on an understand-
ing of the process.
*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
corresondingly 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 in negligible.
-------
IV-2
Table IV-1. VOC Emissions from Uncontrolled Sources in
Typical Vinyl Acetate Plant
Emission Source
Inert-gas purge vent
CO vent
Emergency vents
Light-ends and inhibitor
mix tank vents
Storage
Acetic acid
Vinyl acetate
Total
Stream
Designation
(Fig. III-l)
A
B
C
D
E
F
Emissions
Ratio*
(g/kg)
4.39
0.31
0.013
2.8
0.16
1.95
9.62
Rate
(kg/hr)
80.18
5.66
0.26
51.1
2.92
35.61
175.73
g of emission per kg of vinyl acetate.
-------
IV-3
1- Inert-Gas Purge Vent
The major process source of VOC emissions is released at vent A. Nitrogen,
argon, and ethane, which are inert in the reaction, constitute the inert gases.
These gases carry along the ethylene that is present in the recycle stream.
The quantity of emission from this stream is directly related to the inert-gas
content of both feed gases.
2. C02 Vent
The C02 generated in the reactor by oxidation of ethylene is released at vent B
and carries with it some ethylene and ethane.
3. Emergency Vents
Equipment failure, planned shutdowns, and startups contribute some VOC emissions.
These releases occur at vents C.
4. Light-Ends and Inhibitor Mix Tank Discharges
The reaction also generates acetaldehydes and other low-temperature-boiling
materials. The processing of these materials results in some VOC emission.
The makeup of an inhibitor solution and addition to a distillation stream
result in some VOC emission. These losses occur at vent D.
5. Fugitive Emissions
No data on fugitive emissions are presented in this study. Fugitive emissions
for the entire synthetic organic chemical manufacturing industry (SOCMI) are
covered by separate EPA documents.
6- Storage and Handling Emissions
Emissions result from the storage of acetic acid and vinyl acetate (Vents E and
F). Storage tank sizes and conditions for the typical plant are given in Table IV-2
The storage emissions in Table IV-1 are based on fixed-roof tanks, half full,
and a diurnal temperature variation of 11°C and on the emission equations from
AP-42.1
VOC emissions due to handling are not presented in this sutdy but are contained
in a separate EPA storage and handling document covering the entire SOCMI.2
-------
IV-4
Table IV-2. Storage-Tank Data for Typical Vinyl Acetate Plant
Storage
Tank
Designation
Bulk storage
In-process
In-process
Bulk storage
Contents
Acetic acid
Acetic acid
Vinyl acetate
Vinyl acetate
No. of
Tanks
Required
2
2
2
2
Tank
Size
(m3)
5150
675
940
7150
Turnovers
Per Year
12
182.5
182.5
12
Bulk
Temperature
(°C)
37.8 *
37.8
37.8
37.8
*Temperature estimated for purpose of emission calculations.
-------
IV-5
7- Secondary Emissions
Secondary VOC emissions can result from the handling and disposal of the process-
waste liquid streams such as.- high-temperature-boiling polymer wastes (Discharge I),
wastewater (Discharge J), light-ends waste (Discharge K), and ethyl acetate
purge (Discharge L). Evaluation of the potential emissions from disposing of
these and other wastes from the entire SOCMI will be covered by a future EPA
secondary emissions document.
-------
IV-6
c REFERENCE*
j C. C. Masser, "Storage of Petroleum Liquids," pp. 4.3-6 to 4.3-11 in Compilation
of Air Pollutant Emission Factors, AP-42, Part A, 3d ed. (April 1977).
2. D- G. Erikson, Hydroscience, Inc., Storage and Handling Report (on file at EPA,
ESED, Research Triangle Park, NC) (October 1978).
:£ reference located at the end of a paragraph usually refers to the entire para-
graph. If another reference relates to certain portions of the paragraph, the
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- INERT-GAS PURGE VENT
The inert-gas purge emissions (Vent A, Table V-l) are of fuel quality and can
be disposed of readily in a fire box or thermal oxidizer, which also may be
employed for disposing of the emissions discussed below. With a properly designed
and operated thermal oxidizer a VOC reduction of 99% or greater can be achieved.
Data and documentation to support this conclusion will be presented in a future
EPA report on emission control systems.
Flaring of the inert purge gas is practiced in some plants,-1'2 however, in the
typical plant this stream is extremely small. Unless the flare diameter is
designed for the gas rate, the efficiency may be as low as 95%.3 (This is fur-
ther discussed later in this section.)
B. C02 VENT
The C02 vent gas (Vent B) has a very low fuel value and will not self-sustain
combustion; it can, however, be fed as a fume to the thermal oxidizer mentioned
above and as is practiced in some plants.2'4 With proper design and operation
a VOC reduction of 99% or greater can be achieved.
The C02 vent gas can also be flared; however, the fuel value of all the process
vent streams combined would not be sufficient for a stable flame to be maintained
in a flare, and supplemental fuel would be required. With proper design and
operation an emission reduction of at least 99% is expected.3
C. EMERGENCY VENTS
Venting during periodic unscheduled or emergency-equipment outages require the
safe handling of a large quantity of VOC (Vents C). It is estimated that a
flare properly designed for emergency releases and properly maintained can result
in an emission reduction of 98%.3 The gas venting rate during emergency-equipment
outage will require a flare much larger than the one that would be required to
control the emissions during normal operation.
-------
V-2
Table V-l. VOC Emissions from Controlled Sources in
Typical Vinyl Acetate Plant
Stream Emission
Designation Control Device Reduction
Source (Fig. III-l) or Technique (%)
Inert-purge vent
C02 vent
Emergency vents
Light-ends and inhibitor
mix tank vents
Storage
Acetic acid
Vinyl acetate
A Fire box/thermal 99
oxidizer
b
B Thermal oxidizer/ 99
flare
b
C Flare 98
D Fire box/thermal 99
oxidizer
E None
F Floating roof 96
Emissions
a
Ratio Rate
(g/kg) (kg/hr)
0.044 0.80
0.003 0.05
0.0005 ' 0.009
0.028 0.51
0.26 4.75
0.10 1.83
g of emission per kg of vinyl acetate.
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-3
D. LIGHT-ENDS AND INHIBITOR-TANK VENTS
The light-ends and the inhibitor-tank vent gas emissions (Vents D) are of fuel
quality and can be disposed of readily in a fire box or thermal oxidizer. A
VOC emission reduction of 99% or greater can be achieved as discussed above.
Flaring the vent gas with proper design and operation can also result in an
emission reduction of at least 99%.3
E. STORAGE AND HANDLING SOURCES
1- Acetic Acid Storage
Emission of acetic acid due to tank breathing and filling is small and is uncon-
trolled in this typical plant (Vent E).
2. Vinyl Acetate Storage
Internal floating-roof tanks* are used in the industry for emission control in
vinyl acetate storage and are used in the typical plant. The controlled vinyl
acetate emissions given in Table V-l were calculated with the modified AP-42
floating-roof storage-tank emission equations.5 7
3. Handling
Control of emissions due to handling are not presented in this study but are
included in a separate EPA storage and handling document covering the entire
SOCMI.8
F. CURRENT EMISSION CONTROLS
Emission control devices currently used by some domestic vinyl acetate producers
are shown in Appendix B.
^Consist of internal floating covers or covered floating roofs as defined in
API 25-19, 2d ed. (fixed-roof tanks with internal floating device to reduce
vapor loss).
-------
V-4
G. REFERENCES*
1. S. W. Dylewski, Hydroscience, Inc., Trip Report Information Obtained on Vinyl
Acetate Process Used at Union Carbide Corporation, Texas City, TX, Plant, Dec. 8,
1977 (on file at EPA, ESED, Research Triangle Park, NC).
2. D. W. Smith, letter to EPA from E.I. du Pont de Nemours and Company, Inc., La
Porte, TX, Sept. 18, 1978, in response to EPA's request for information on the
vinyl acetate process.
3. V. Kalcevic, IT Enviroscience, Inc., Control Device Evaluation. Flares and
the Use of Emissions as Fuels (in preparation for EPA, ESED, Research Triangle
Park, N.C.
4. S. W. Dylewski, Hydroscience, Inc., Trip Report for Visit to Celanese Chemical
Company, Bay City, TX, Sept. 28 and 29, 1977 (on file at EPA, ESED, Research
Triangle Park, NC)
5. C. C. Masser, "Storage of Petroleum Liquids," pp. 4.3-6 to 4.3-11 in Compilation
of Air Pollutant Emission Factors, AP-42, Part A, 3d ed. (April 1977).
6. Chicago Bridge and Iron, SOHIO/CBI Floating Roof Emission Testing Program,
Supplemental Report (Feb. 15, 1977).
7. Equation for floating-roof withdrawal loss derived by R. Burr, EPA, ESED,
Research Triangle Park, NC.
8. D. G. Erikson, Hydroscience, Inc., Storage and Handling Report (on file at EPA,
ESED, Research Triangle Park, NC) (October 1978).
*A reference located at the end of a paragraph usually refers to the entire para-
graph. If another reference relates to certain portions of the paragraph, the
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
1- Typical Plant
The environmental impact of the application of the described control systems to
the typical plant would be a VOC emission reduction of 1625 Mg/yr compared to
an uncontrolled plant, as shown in Table VI-1.
2. Industry
Emission sources, control levels, and emission ratios for the typical plant are
summarized in Table VI-2. From emission data reported by producing vinyl acetate
plants13 the emission ratios for the industry have been estimated and are
also shown in Table VI-2. These values show that the industry processes as
represented by the Bayer process are about 96% controlled. With a 1978 produc-
tion level of 740 Gg and the assumption that the emissions from the industry
are at the same ratio as for the Bayer process, the emissions from industry are
estimated to be 247 Mg.*
*Fugitive, secondary, storage, and handling emissions are not included.
-------
VI-2
Table VI-1. Environmental Impact of Controlled Sources in
a Typical Vinyl Acetate Plant*
Source
Inert-gas purge vent
Stream
Designation
(Fig. III-2)
A
Control Device
or Technique
Fire box/thermal
Total VOC
Emission Reduction
(%) (Mg/yr)
99 695.4
vent
oxidizer
Thermal oxidizer/
flare
99
*Fugitive, secondary, and handling emissions are not included.
49.1
Emergency vents
Light-ends and inhibitor
mix tank vents
Storage
Acetic acid
Vinyl acetate
Total
C
D
E
F
Flare 98
Fire box/thermal 99
oxidizer
None
Internal floating-roof 96
tank
4.
443.
433.
1625
0
0
5
-------
VI-3
Table VI-2. Emission Ratios for Typical Plant and for
Industry for Bayer Process
Emission Ratio (g/kg)
Emission Source
Inert-gas purge vent
CO vent
Emergency vent
Light ends ^
> Tank vents
Inhibitor MixJ
Total direct emissions
Reduction
Typical
Uncontrolled
4.39
0.31
0.025
2.8
7.525
Plant
Controlled
0.044
0.003
0.0005
0.028
0.0755
99%
Industry
0.079
0.153
0.001
0.030
0.071
0.334
95.6%
g of emission per kg of vinyl acetate.
Average data; see refs. 1—3.
-------
VI-4
B. REFERENCES*
1. S. W. Dylewski, Hydroscience, Inc., Trip Report for Visit to Celanese Chemical
Company, Bay City, TX, Sept. 28 and 29, 1977 (on file at EPA, ESED, Research
Triangle Park, NC).
2. S. W. Dylewski, Hydroscience, Inc., Trip Report for Visit to Union Carbide
Corporation, Texas City, TX Plant, Dec. 8, 1977 (on file at EPA, ESED, Research
Triangle Park, NC).
3. D. W. Smith, letter to EPA from E.I. du Pont de Nemours & Company, Inc., La Porte,
TX, Sept. 18, 1978, in response to EPA's request for information on the vinyl
acetate process.
*A reference located at the end of a paragraph usually refers to the entire
paragraph. If another reference relates to certain portions of the paragraph,
the 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 Vinyl Acetate*
Synonyms
Molecular formula
Molecular weight
Physical state
Vapor pressure
Vapor specific gravity
Boiling point
Melting point
Density
Water solubility
Acetic acid, vinyl ester,
acetic acid, ethenyl ester
C .H O
462
86.09
Liquid (polymerizes)
107.5 mm of Hg at 25°C
3.0
72.2 to 72.3°C at 760 mm of Hg
-93.2°C
0.9317 at 20°C/4°C
Insoluble; soluble in hot HO
*Prom: J. Dorigan et al., "Vinyl Acetate," p. AIV-286 in
Appendix IV, Rev. 1 (Chemicals O—Z) , to Scoring of Organic Air
.Pollutants. Chemistry, Production and Toxicity of Selected
Synthetic Organic Chemicals, MTR-7248, MITRE Corp. , McLean, VA
(September 1976).
-------
A-2
Table A-2. Physical Properties of Acetaldehyde*
Synonyms Acetic aldehyde, Ethyl aldehyde
Molecular formula C9H/io
Molecular weight 44.05
Physical state Liquid
Vapor pressure 923 torrs at 25°C
Vapor specific gravity 1.52
Boiling point 20.8°C at 760 mm
Melting point -121°C
Density 0.7834 at 18°C/4°C
Water solubility Infinite (hot)
_
From: J. Dorigan e_t al, "Acetaldehyde" p. AI-6 in Appendix I, Rev. 1,
(Chemicals A-C), to Scoring of Organic Air Pollutants. Chemistry, Production
and Toxicity of Selected Synthetic Organic Chemicals, MTR-7248, MITRE Corp.,
McLean, VA (September 1976).
-------
A-3
Table A-3. Physical Properties of Methyl Acetate*
Synonyms Acetic acid, methyl ester
Molecular formula C_H O
362
Molecular weight 74.08
Physical state Liquid
Vapor pressure 212.5 mm at 25°C
Vapor specific gravity 2.55
Boiling point 57.8°C
Melting point -98.1°C
Density 0.9330 at 20°C/4°C
Water solubility Very soluble
*
From: J. Dorigan ejt al^ "Methyl Acetate" p. AIII-148 in Appendix III, Rev. I,
.(Chemicals F-N), to Scoring of Organic Air Pollutants. Chemistry, Production
and Toxicity of Selected Synthetic Organic Chemicals, MTR-724B, MITRE Corp.,
McLean, VA (September 1976).
-------
A-4
Table A-4. Physical Properties of Ethyl Acetate*
Synonyms Acetic ester, ethyl etharate
Molecular formula ^.H O
Molecular weight 88.10
Physical state Liquid
Vapor pressure 92.5 mm at 25°C
Vapor specific gravity 3.04
Boiling point 77.06°C at 760 mm
Melting point -83.58°C
Density 0.8946 at 25°C/4°C
Water solubility Soluble (89 gm/liter)
_
From: J. Dorigan et ad, "Ethyl Acetate" p. AII-234 in Appendix II, Rev. 1,
(Chemicals D-E), to Scoring of Organic Air Pollutants. Chemistry, Production
and Toxicity of Selected Synthetic Organic Chemicals, MTR-7248, MITRE Corp.,
McLean, VA (September 1976}.
-------
B-l
APPENDIX B
EXISTING PLANT CONSIDERATIONS
Table B-l lists process control devices reported to be in use by industry.
To gather information for the prepation of this report site visits were made
to manufacturers of vinyl acetate. Trip reports have been cleared by the
companies concerned and are on file at EPA, ESED, in Research Triangle Park, N.C.1,2
A- PROCESS EMISSIONS FROM EXISTING PLANTS
l- Union Carbide Corp., Texas City, Texas1
Union Carbide Corp. at Texas City, Texas manufactures vinyl acelate (VA) by
oxidation of ehtylene and acetic acid. The plant, of single train design,
was started in 1975 and has a capacity of 350 MM Ib/yr. Acetic acid is
received by barge; ethylene and oxygen are supplied by other facilities at
the Texas City site. The product is shipped by marine tankers and barges.
Barger AG and U.S. Industrial Chemicals both have patents on the vapor pahse
oxidation of a mixture of ethylene and acetic acid. The direct VOC emissions
reported before and after the emission control device (ECD) are shown by
Table B-2. There are several liquid streams that are removed from the process
where secondary emissions may be experienced. These streams are tabulated
in Table B-3.
2- Celanese Chemical Co., Bay City, Texas2
The Celanese Chemical Company at Bay City, Texas, manufactures vinyl acetate
(VA) by oxidation of ethylene and acetic acid. Acetaldehyde and acetic acid
are also manufactured in separate plants at this site. The capacity is stated
to be 400 MM Ib/hr. The VA unit is connected to the acetaldehyde and acetic-
acid units by raw material flow piping, liquid purge, and gas purge streams
which aid in reducing emissions but make it difficult to segregate the emissions
specific for the VA unit. Acetic acid is supplied by barge shipments in
addition to the on-site manufacturing facilities. Ethylene is received by
pipeline from Monsanto and Phillips Petroleum facilities located nearby.
VA product is shipped mostly by barge and a minor amount by tank car and tank
truck.
-------
Table B-l. Emission Control Devices Currently Used by Some Domestic Vinyl Acetate Producers by
the Bayer Process
Source
Control Devices Used by
Celanese Chemical Corp.'
Bay City, TX
Union Carbide Corp.
Texas City, TX
Dupont
LaPorte, TX
Inert-gas purge vent
CO purge vent
Emergency vents
Light-ends vent
Mix-tank vent
To another process
Thermal oxidizer
Flare
To another process
To atmosphere
Flare
Catalytic combustion
Flare
Flare
No data
Flare
Thermal oxidizer
No data
To atmosphere
No data
See ref. 2.
DSee ref. 1.
"i
"See ref. 3.
Part of feed to another process; ultimately vented to a flare.
a
i
-------
B-3
Table 8-2. Direct Emissions (Union Carbide)
jSource
C02 purge
Xnerts purge*
Polutant
Ethylene, ethane
Ethylene, ethane
VA, acetaldehyde
Ib VOC/1000
Before ECD
2.63
13.0
Ib Product
After ECD
.0.067
Zero
M
Emergency relief, vent header, reactor start-up, VA column vents,
CO purge, emergency vent.
Table B-3. Secondary Emissions (Union Carbide)
Steam Rate
Description Discharged to Pollutant lb/1000 Ib Product
Waste polymer incinerator Acetic acid, 25.0
ethylidene
diacetate
polymers,
light metal
acetates
Slowdown tower
(from inerts purge),
flare seal pot, Vinyl Acetate 0.104
reactor condensate Wastewater
blowndown, reactor
product water purge,
polymer incinerator
scrubber tails.
By-product residue Boiler fuel Vinyl acetate, 10.64
acetaldehyde
Samples Drummed Acetic acid, 0.009
(off site vinyl acetate
disposal)
acetaldehyde
-------
B-4
Barger AG and U.S. Industrial Chemicals both have patents on the vapor phase
oxidation of a mixture of ethylene and acetic acid. The direct emissions
reported before and after the emission control device (ECD) are shown on
Table B-4. There are several points where liquids are removed from the process
where emissions may be experienced as a secondary emission. These are
indicated by Table B-5.
3. Du Pont, Inc., La Port, Texas3
The nominal capacity of the Du Pont vinyl acetate plant at La Porte, Texas
is 1.1 MM Ib/day. The direct emissions are reported as indicated by Table B-6.
Secondary emissions may occur from the following reported waste streams:
A combination of liquid organic purges from the process. These streams
average about 212 pounds per hour and are 98 wt % organic. These streams
are combined with several others from two processes and incinerated in a
natural gas fired incinerator.
Purge of reaction water from the process. The stream is 31 gals per minute
and is combined with a flow from another process and fed to a waste disposal
well system. The stream is MJ.03 wt % organic.
Purge of high boilers to the gas fixed incineration. The flow is approximately
1270 Ib/hr and essentially all organic.
B. RETROFITTING CONTROLS
The primary difficulty associated with retrofitting may be in finding space
to fit the control device into the existing plant layout. Because of the
costs associated with this difficulty it may be appreciably more expensive
to retrofit emission control systems in existing plants than to install a
control system during construction of a new plant.
-------
B-5
Table B-4. Direct Emissions (Celanese)
Source
Emergency vent from
reactor
Inhibitor mix tank
Shutdown
Pollutant
Ethylene, HOAC, VA
Vinyl acetate
Ethylene, HOAC, VA
Ib VOC/1000 lb Product
Before BCD After BCD
0.0125
0.096
0.037
Zero
No BCD
Zero
Table B-5. Secondary Emissions (Celanese)
Description
Vinyl acetate
heavy ends
Azeotrope column side
stream, water stripper,
reactor inerts purge
Water stripper residue
carbonate system
blowdown
Samples
Sample flush
Discharged to
Incinerator
Deep well
Deep well
Deep well
Pollutant
Confidential
Another process Confidential
Mostly water
VOC
VOC
Steam Rate
Confidential
Confidential
Confidential
4600 Ib/yr
18,300 Ib/yr
-------
B-6
Table B-6. Direct Emissions (Du Pont)
Source Composition (wt %) Flow
Inert gas purge vent 45% ethylene 1100 Ib/hr
10% vinyl acetate
2.5% acetic acid
40% inerts
2.5% misc. VOC
CO purge vent 0.3% ethylene 600 Ib/hr
99.7% CCL
-------
B-7
C. REFERENCES
S. W. Dylewski, IT Enviroscience, Trip Report for Union Carbide Plant,
Union Carbide Corp., Texas City, Texas, Dec. 8, 1977 (on file at EPA,
ESED, Research Triangle Park, NC).
S. W. Dylewski, IT Enviroscience, Trip Report for Celanese Chemical Plant,
Celanese Chemical Co., Bay City, Texas, Sept. 28, 1977 (on file at EPA,
ESED, Research Triangle Park, NC) .
D. W. Smith, Du Pont, Inc., letter to D. R. Goodwin, EPA, dated Sept. 18, 1978.
*A reference located at the end of a paragraph usually refers to the entire
paragraph. If another reference relates to certain portions of the paragraph,
the 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
ACETALDEHYDE
R. J. Lovell
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 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.
D28A
-------
6-iii
CONTENTS OF REPORT 6
Page
I- ABBREVIATIONS AND CONVERSION FACTORS 1-1
II. INDUSTRY DESCRIPTION II-l
A. Reason for Selection II-l
B. Acetaldehyde Usage and Growth II-l
C. Domestic Producers II-3
D. References II-7
HI- PROCESS DESCRIPTION III-l
A. Introduction III-l
B. Acetaldehyde from Ethylene III-l
C. Acetaldehyde from Ethanol III-8
D. Acetaldehyde from Acetylene III-8
E. Acetaldehyde from Saturated Hydrocarbons III-9
F. Acetaldehyde from Synthesis Gas III-9
G. References 111-10
IV. EMISSIONS IV-1
A. Acetaldehyde from Ethylene by the Two-Step Air-Oxidation Process IV-1
B. Acetaldehyde from Ethylene by the Single-Step Oxygen-Oxidation IV-5
Process
C. References IV-7
V. APPLICABLE CONTROL SYSTEMS V-l
A. Acetaldehyde from Ethylene by the Two-Step Air-Oxidation Process V-l
B. Acetaldehyde from Ethylene by the Single-Step Oxygen-Oxidation V-4
Process
C. Control Measures Currently Used V-4
D. References v~5
APPENDICES OF REPORT 6
A. PHYSICAL PROPERTIES OF ACETALDEHYDE, METHYL CHLORIDE, ETHYL CHLORIDE, A-l
AND CHLOROFORM
B. FUGITIVE EMISSION FACTORS B-l
C. EXISTING PLANT CONSIDERATIONS C-l
-------
6-v
TABLES OF REPORT 6
Number page
H-1 End Usage of Acetaldehyde II-2
11-2 Acetaldehyde Capacity 11-4
IV-1 Uncontrolled Emissions for Two-Step Air-Oxidation Process IV-2
IV-2 Composition of Uncontrolled Emissions from Two-Step Air-Oxidation IV-4
Process
V-l Controlled Emissions for Two-Step Air-Oxidation Process V-2
A-l Properties of Acetaldehyde, Methyl Chloride, Ethyl Chloride, and A-I
Chloroform
c~l Control Methods Currently Used by the Domestic Acetaldehyde Industry C-2
c~2 High- and Low-Pressure Scrubber Emissions from Celanese Plant C-3
c"3 High- and Low-Pressure Scrubber Emissions from Texas Eastman Plant C-5
c~4 Estimated 1979 Acetaldehyde Industry Emissions C-6
FIGURES OF REPORT 6
Number Page
n-l Locations of Plants Manufacturing Acetaldehyde II-5
III-l Flow Diagram for Two-Step Air-Oxidation Process III-3
III-2 Flow Diagram for Single-Step Oxygen-Oxidation Process III-6
-------
1-1
ABBREVIATIONS AND CONVERSION FACTORS
EPA policy is to express all measurements used in agency documents in metric
units. Listed below are the International System of Units (SI) abbreviations
and conversion factors for this report.
To Convert From
Pascal (Pa)
Joule (J)
Degree Celsius (°C)
Meter (m)
Cubic meter (in3)
Cubic meter (m3)
Cubic meter (m3)
Cubic meter/second
(
-------
II-l
II. INDUSTRY DESCRIPTION
A. REASON FOR SELECTION
Acetaldehyde production was selected for study because preliminary estimates
indicated that the production process causes significant emissions of volatile
organic compounds (VOC).
Acetaldehyde is a colorless, mobile liquid having a pungent suffocating odor
that is somewhat fruity and pleasant in dilute concentrations. Some physical
properties of acetaldehyde are given in Appendix A.
B. ACETALDEHYDE USAGE AND GROWTH
The current production capacity of acetaldehyde in the United States is 621 Gg/yr,
2
with the 1976 production of the order of 440 Gg or 71% of this capacity. Acetal-
dehyde is a chemical intermediate used in the manufacture of the products shown
in Table II-l. — From 90 to 95% of the acetaldehyde produced is used captively
by the producer.
Peak production of acetaldehyde occurred in 1969, when 749 Gg was produced.
From 1969 to 1975 production declined to a low of 408 Gg/yr. The decline was
largely due to phase-out of n-butanol and 2-ethylhexanol produced from acetal-
4
dehyde. These chemicals are presently made from propylene by the 0X0 process.
Early in 1976 it was estimated that the 1976 acetaldehyde production would reach
612 Gg/yr and that production would continue to grow to 703 Gg/yr by the year
7
1980. This would indicate an average annual growth of 3.5%. However, 1978
statistics indicate that actual 1976 production was between 431 and 445 Gg/yr
2
and that projected production rates were not obtained.
The price of acetaldehyde during the period 1950 to 1973 ranged from $0.20/kg
to $0.22/kg. Since 1973, due largely to the increased cost of hydrocarbon crack-
ing feedstocks for production of ethylene, the basic raw material used in the
production of acetaldehyde, the price of acetaldehyde has increased an average
of 15% per year to the current price of $0.44/kg.8 In 1976 acetic acid produc-
tion consumed 60% of the acetaldehyde produced. With the increasing cost of
-------
11-2
Table II-l. End Usaqe of Acetaldehyde'
End Use
Acetaldehyde Consumption
fnr 1976 (%)
Acetic acid
Synthetic pyridine derivatives
Peracetic acid
c
Acetate esters
Pentaerythritol
Other uses
60
40
See refs
-7.
Acetic acid is used principally for manufacture of vinyl acetate,
cellulose acetate, terephthalic acid, acetic anhydride, acetate
esters, chloroacetic acids, and dyestuffs.
"By the Tischenko process.
Includes crotonaldehyde, chloral, 1,3-butylene glycol, lactic
acid, and glyoxal.
-------
II-3
ethylene feedstock, methanol carbonylation has become the preferred process for
manufacture of acetic acid. Thus no growth in acetaldehyde consumption for
manufacture of acetic acid is expected.
The manufacture of pentaerythritol, peracetic acid, and synthetic pyridine deriva-
tives and the manufacture of acetate esters fay the Tischenko process account
for the remaining 40% of the acetaldehyde consumed. This group may show strong
growth in some products (pentaerythritol, used to Manufacture synthetic lubri-
cants), but even some of them may be produced by alternate processes (pyridine).
The combined growth of the products in this group in not expected to take up
the slack of lost acetic acid growth.
The future of acetaldehyde growth appears to depend on the development of a
lower cost process based on synthesis gas and an increase in demand for prod-
ucts produced by processes based on acetaldehyde.
C. DOMESTIC PRODUCERS*
1 9
There are currently five plants producing acetaldehyde in the United States.
Table 11-2 lists the producers, locations, capacities, and raw materials; Fig. II-l
shows the plant locations.
Commercial processes for the production of acetaldehyde include,- the direct
oxidation of ethylene, the oxidation or dehydrogenation of ethanol, the addition
of water to acetylene, and the partial oxidation of hydrocarbons, Acetaldehyde
was first commercially produced in 1911 by hydration of acetylene. As the
demand increased, ethanol-based processes became the principal method used. In
the 1960s, the Hoechst-Wacker process for direct oxidation of ethylene was com-
mercialized and by 196B became the principal method of acetaldehyde production
in the United States. In 1967 there were 18 domestic plants, with a combined
capacity of 754 Gg/yr. In that year 43% of the acetaldehyde capacity was based
on ethylene, 31% on ethanol, 25% on propane-butane, and 1% on acetylene and
*Since this report was first prepared (January 1979) Celanese has reportedly
increased its capacity from 363 Gg/yr to 431 Gg/yr. Also it has been reported
that domestic consumption of acetaldehyde is expected to decline at a rate of
approximately 3%/yr during the 1978—1983 period.
-------
II-4
Table II-2. Acetaldehyde Capacity'
Company
Celanese Chemical Co.
Celanese Chemical Co.
Texas Eastman Co.
Publicker Industries, Inc.
Shell Chemical Co.
Total
Location
Bay City, TX
Clear Lake City, TX
Long view, TX
Philadelphia, PA
Norco, LA
1978
Production
Capacity
(Gg/vr)
136
227b
227
29C
2
621
Basic
Raw
Material
Ethylene
Ethylene
Ethylene
Ethanol
By-product
See ref 2.
DSee ref 10.
'Publicker Industries, Inc., isolates an estimated 2.3 Gg/yr as acetaldehyde;
the balance runs from ethanol through acetaldehyde (not isolated) to acetic
acid.
-------
II-5
(1) Celanese Corp., Bay City, TX
(2) Celanese Corp,, Clear Lake City, TX
(3) Texas Eastman Co., Longview, TX
(4) Publicker Industries Inc., Philadelphia, PA
(5) Shell Chemical Co., Norco, LA
Fig. II-l. Locations of Plants Manufacturing Acetaldehyde
-------
II-6
4
other processes. In 1966 Tennessee Eastman discontinued production at their 123-Gg/yr
ethanol-based Kingsport plant, and in 1977 Union Carbide discontinued production at
their 91-Gg/yr ethanol-based facilities in West Virginia. Of the five plants in opera-
tion today 94.9% of the acetaldehyde capacity is based on ethylene, 4.7% on ethanol,
and 0.4% recovered as by-product from other processes.
-------
II-7
>. REFERENCES*
1. 1978 Directory of Chemical Producers, United States of America, Chemical Informa-
tion Services, Stanford Research Institute, Menlo Park, CA, p. 415.
2. "CEH Manual of Current Indicators -- Supplemental Data," p. 201 in Chemical
Economics Handbook, Stanford Research Institute, Menlo Park, CA (October 1978).
3. P. L. Morse, Acetaldehyde, Report No. 24, A private report by the Process
Economics Program, pp. 13 — 17, Stanford Research Institute, Menlo Park, CA
(April 1967).
4. E. M. Klapproth, "Acetaldehyde -- Salient Statistics," pp. 601.5020A--601.5020D
in Chemical Economics Handbook, Stanford Research Institute, Menlo Park, CA
(May 1976).
5. H. J. Hagemeyer, "Acetaldehyde," pp. 97—112 in Kirk-Othmer Encyclopedia of
Chemical Technology, 3d ed., vol 1, edited by M. Grayson et al., Wiley-Interscience,
New York, 1978.
6. F. S. Wagner, Jr., "Acetic Acid," p. 141 in Kirk-Othmer Encyclopedia of Chemical
Technology, 3d ed., vol. 1, edited by M. Grayson et al., Wiley-Interscience,
New York, 1978.
7. "Chemical Profile on Acetaldehyde," p. 9 in Chemical Marketing Reporter, May 10,
1976.
8- "Current Prices of Chemicals and Related Materials," Chemical Marketing Reporter
214(16), 62 (Oct. 16, 1978).
9. J. J. Cudahy and J. F. Lawson, IT Enviroscience, Inc., Trip Report to Celanese
Plant, Celanese Chemical Company, Clear Lake City, TX, Sept. 22, 1977 (on file
at EPA, ESED, Research Triangle Park, NC).
10. A. K. Rafie and S. L. Soder, "Acetaldehyde—Product Review," pp 601.5020A--
601.5020M in Chemical Economics Handbook. Stanford Research Institute, Menlo
Park, CA (March 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
INTRODUCTION
In the United States 94.9% of the acetaldehyde capacity is based on the Hoechst-
Wacker process, which consists of two-step air oxidation of ethylene. Approxi-
mately 4.7% is based on oxidation of ethanol, and the remaining 0.4% is recovered
as by-product from other processes. Hydration of acetylene and oxidation of
saturated hydrocarbons (butane and propane), once processes of major importance
in the United States, have given way to the more economic ethylene-based process.
A rhodium catalyzed process capable of converting synthesis gas directly into
2
acetaldehyde in a single step was reported in 1974. This process may become
important in the future as coal gasification methods are perfected.
ACETALDEHYDE FROM ETHYLENE
The direct liquid-phase oxidation of ethylene to acetaldehyde by means of a
palladium chloride-cupric chloride catalyst was discovered in 1956. The com-
mercial process was developed and licensed by the consortium of Hoechst and
Wacker in Germany in the late 1950s and early 1960s. Since 1960 all reported
4
new plants have used the Hoechst-Wacker technology.
Direct oxidation of ethylene to form acetaldehyde is accomplished through a
series of oxidation-reduction reactions. The catalyst is an aqueous solution
of palladium chloride and cupric chloride. The reaction of ethylene with an
aqueous palladium chloride solution to form acetaldehyde is represented by Eq. (1);
C H + PdCl + HO =»- CH CHO + Pd + 2HC1 (1)
^* 4 tt £t J
(ethylene) (palladium (water) (acetaldehyde) (metallic (hydrochloric
chloride) palladium) acid)
The palladium is reoxidized to palladium chloride by cupric chloride as shown
by Eq. (2).
Pd + 2CuCl2 > Pdcl2 + 2CuCl (2)
(palladium) (cupric chloride) (palladium (cuprous chloride)
chloride)
-------
III-2
The cuprous chloride formed is then reoxidized with oxygen or air [Eq. (3)]:
2CuCl + 1/20 + 2HC1 - > 2CuCl + HO (3)
£* £ £t
(cuprous (oxygen) (hydrochloric (cupric chloride) (water)
chloride) acid)
The net result can be represented by the overall Eq. (4):
C2H4 + 1/202 2, 2 ^ CH3CHO (4)
(ethylene) (oxygen) (catalyst) (acetaldehyde)
The process is carried out with a large excess of cupric chloride and only small
"catalytic" amounts of palladium chloride. Catalyst life is practically infinite;
however, HC1 is consumed by side reactions and thus must be continually added.
Two variations of the process have been developed. In the two-step process,
air instead of oxygen is used, and ethylene and air react separately with the
catalyst solution in separate reactors. In the first-stage reactor, ethylene
reduces the cupric chloride catalyst by a combination of Eqs. (1) and (2). The
cuprous chloride in the liquid phase is separated from the product gases and
recycled with the catalyst solution to the second-stage reactor. Cuprous chloride
is reoxidized [Eq. (3)] by air in the second-stage reactor and then returned to
the first-stage reactor.
In the single-step process a mixture of ethylene and pure oxygen reacts with
the catalyst solution in a common reactor according to the reaction shown by
Eq. (4). The reaction product is separated, the unreacted gas is recycled, and
the consumed ethylene, oxygen, and hydrochloric acid are replaced. The single-
step process has not been employed in the United States.
3
1. Two-Step Air-Oxidation Process
a. Model Process -- The model two-step process is represented by Fig. III-l. After
ethylene (Stream 1) is fed to a tubular reactor, it reacts under pressure (approxi-
mately 820 kPa at 130°C) with the catalyst solution to form acetaldehyde and
cuprous chloride. The pressure of the solution (Stream 2) is reduced in a flash
-------
PEW1TOR
ETUVLEWE
7>
OFF-AIR
SEPARATOR
OXIDIZER
AIR
COMPRESSOR
A- OFF-AIR vtMT
B- OFF-(=(*A VEKlT
F - HAKlDLlKJiq
H - FUGITIVE
J- SECONDARY
K- SECOWD6«V
FROM PUAKTT
FROM WA'bTe PRODUCTS =>.C. STREAM
FROM WASTEWATER
ACETALDEI-IYDE
PRODUCT
FIKJAU
Dl'sTlLLATlOKl
COLUMN!
UGWT-ENDS
DISTILLATION
COLUMkl
nr
TOWER
CRUDE
DISTIULAT.
COLUMN
fOFF-
' ABSORBER
STEAM
MAKEUP
WATER
OFF-AIR
ABSORBER
I
U)
Fig. III-l. Flow Diagram for Uncontrolled Typical Plant Producing Acetaldehyde from
Ethylene by Two-Step Air-Oxidation Process
-------
III-4
tower, from which the evaporated acetaldehyde and water (Stream 3) are sent to
the crude distillation column. The catalyst solution (Stream 4), which contains
cuprous chloride equivalent to the amount of acetaldehyde formed, is fed to a
second-stage tubular oxidizer, where the cuprous chloride is reoxidized with
air (Stream 5) to cupric chloride. Unreacted ethylene and a portion of the
organic by-products contained in the catalyst solution are oxidized to carbon
dioxide and water. The catalyst mixture (Stream 6) passes to the off-air sepa-
rator, where gases and uncondensed vapors (Stream 7) are separated from the
catalyst solution. Hydrochloric acid (Stream 8) is added to replenish that
lost through by-product reactions. The regenerated catalyst solution (Stream 9)
is then returned to the first-stage reactor for further reaction with ethylene.
The gases from the separator (Stream 7) pass to an absorber for recovery of
residual quantities of acetaldehyde, along with other water-soluble components.
The unabsorbed gases and vapors (Stream 10) are vented (Vent A), and the absorber
liquid (Stream 11) is fed to the crude distillation tower.
In the crude distillation column acetaldehyde and organic impurities are removed
overhead. This stream is condensed and the condensate (Stream 12) is passed to
the light-ends distillation column. Uncondensed vapors and gases (Stream 13)
are passed to the off-gas absorber. The stripped water bottoms (Stream 14) are
recycled to the reactor system and to the absorber units. The light-ends distil-
lation column separates the low-boiling-point light-organic impurities (Stream 15)
from the acetaldehyde and high-boiling-point materials (Stream 16). The off-gases
(Stream 15) pass through an absorber for recovery of residual quantities of
product and are vented (Stream 17, Vent B). The absorber liquid (Stream 18)
returns to the crude distillation column.
The acetaldehyde (Stream 16) enters the final distillation column, where the
purified product is removed overhead (Stream 19) to product storage. A side-cut
stream (Stream 20) consisting of heavier chlorinated organic by-products is
sometimes separated. The bottoms (Stream 21) are the remaining water and higher-
boiling-point organic impurities.
Process yields of 94 to 95.2% have been reported.
-------
III-5
Process Variations — Applications of the Hoechst-Wacker technology will result
in similar plants, although process modifications are possible and are often
employed. The manner of product purification may vary somewhat from that described
for the model process.
A small stream of oxidized catalyst (Stream 6) is often withdrawn and passed
through a separate catalyst regenerator reactor, where it is heated to about
160°C to decompose by-products.
The side-cut of chlorinated by-products (Stream 20) from the final distillation
column may not be separated but instead may be combined with the wastewater for
treatment and disposal.
A purge stream may be withdrawn from the absorber water circuit (Stream 14) and
discharged or treated separately.
Single-Step Oxygen-Oxidation Process
The model single-step process is represented by Fig. III-2. High-purity ethylene
(Stream 1) and pure oxygen (Stream 2) are fed to the reactor filled with the
catalyst solution. The reaction takes place at about 130°C and 405 kPa. Vaporized
reaction products, evaporated water, and unreacted ethylene and oxygen (Stream 3}
are separated from the catalyst solution (Stream 4) by the demister. A side
stream of catalyst (Stream 5) is treated with oxygen and heated to about 170°C
to decompose the by-products. Hydrochloric acid is continually added (Stream 6)
to replenish that lost through by-product reactions.
The reaction products (Stream 3) are quenched and then passed to the absorber,
where acetaldehyde vapors are cooled and absorbed with water to separate it
from the unreacted gases. The major portion of the gas is recycled (Stream 7)
to the reactor, and a small amount (Stream 8) is vented (Vent A) to prevent
accumulation of gaseous contaminants. The crude acetaldehyde solution (Stream 9)
from the bottom of the absorber is passed to the extractive distillation column,
where the acetaldehyde solution (Stream 10) is separated from the low-boiling-
point light-organic components (Stream 11) by extractive distillation with water.
The off-gases from the column overhead are vented (Vent B).
-------
O/YG.EK1
HCI
WATER.
PUR.G,E.
OUEMCH
TOWeR
REACTOR
0
I
COMPRESSOR
P1WAU
DISTlULATlOkl
COLUMKJ
OFF (3,AS
|(e)
<8>
STM.
EX.TRAC-TIVE
V_^ DIST1LLATIOM
COUJMM
PRODUCT
I3> PRODUCTS'? > LOAD.t-S
^/ «o\nF-riiT^'
®
VEWT
. VEMT
(D, - STORAGE EMISSiOMS
vg) - MAKJDUMGi EMlSSIOkJS
(H) - PUNITIVE EM>SSIOKI6 PROM
Q) - SECONDARY EMISSIOMS PROM
PRODUCTS SIDE-CUT ~~
® - SECOMDARY EMl=,'b\OUS FROM
WATER
Fig. III-2. Flow Diagram for Uncontrolled Typical Plant Producing Acetaldehyde from
Ethylene by Single-Step Oxygen-Oxidation Process
-------
III-7
Acetaldehyde (Stream 12) is separated from water and purified in the final dis-
tillation column. A side-cut stream is taken to partially withdraw the high-
boiling chlorinated by-products (Stream 13), and those remaining are discharged
3--5
with the wastewater (Stream 14). Process yeild is reported to be 94 to 95.2%.
3- Process Comparison
The first-stage reactor of the two-step process operates at a significantly
higher pressure than that in the single-step process, which results in a higher
4
reaction rate and makes it possible for a single-pass operation to be used.
Since the gases are not recycled, ethylene can be used that is considerably
lower in purity than is necessary for the single-step process. The process
yields (kg of acetaldehyde produced per kg of pure ethylene fed to the reactor)
3--5
are reported to be equal for both processes.
In the two-step process emissions from vent A consist primarily of inert gas
(99%) since the process uses air for the oxygen supply and nearly all of the
oxygen is consumed. The volume of the single-step process from vent A is very
small and consists mostly of VOC since pure ethylene and oxygen are fed to the
reactor. The total amount of ethylene lost is essentially the same with each
process.
The two-step process produces relatively small amounts of heavily contaminated
wastewater, since most of the water is removed in the crude distillation step
and recycled. The single-step process produces considerably larger amounts of
dilute wastewater since the water used in absorption of the product is discharged
on final product purification.
Published data indicate that the single-step process consumes 4 g of HC1 per kg
of acetaldehyde produced, whereas the two-step process consumes 15 g/kg. If
it could be concluded from these data that less chlorinated by-product waste
material is produced by the single-step process than by the two-step process, the
single-step process would be more attractive from an environmental standpoint.
The economics of the two processes are similar, with the oxygen-based single-
step process having better economy for smaller plants (less than 100 Gg/yr) and
-------
III-8
the air-based two-step process having the advantage for large plants (greater
than 100 Gg/yr).3
C. ACETALDEHYDE FROM ETHANOL
Before the Hoechst-Wacker ethylene-based process was developed, acetaldehyde
4
was produced principally by the catalytic oxidation of ethanol. Preheated air
and alcohol vapors are passed over a silver catalyst at 480°C to produce acetalde-
hyde according to the following equation:
CH3CH2OH + 1/202 * CH3CHO + H2 CH3CHO
(acetylene) (water) (catalyst) (acetaldehyde)
-------
III-9
E. ACETALDEHYDE FROM SATURATED HYDROCARBONS2
Acetaldehyde is formed as a coproduct in the vapor-phase oxidation of saturated
hydrocarbons, such as butane. This process was of significant commercial import-
ance in the United States until it was rendered uncompetitive by rising costs
of petroleum-based feedstocks. Oxidation of butane yields acetaldehyde, formal-
dehyde, methanol, acetone, and mixed solvents as major products; other aldehydes,
alcohols, ketones, glycols, acetals, epoxides, and organic acids are formed in
smaller concentrations. The cost of feedstocks and problems in product separa-
tion and recovery make it unlikely that new plants will be built based on this
process.
F. ACETALDEHYDE FROM SYNTHESIS GAS
A process for converting synthesis gas directly to acetaldehyde in a single
step using a rhodium catalyst was reported in 1974. Synthesis gas is passed
over a 5% rhodium on a silicon oxide catalyst bed at a temperature of 300°C and
a pressure of 2000 kPa to form acetaldehyde by the following equation:
CO + H2 >• CH3CHO + other products
(carbon (hydrogen) (catalyst) (acetaldehyde)
monoxide)
The principal coproducts formed are acetaldehyde, 24%; acetic acid, 20%; and
ethanol, 16%.
-------
111-10
G. REFERENCES*
1. 1978 Directory of Chemical producers, United States of America, Chemical Information
Services, Stanford Research Institute, Menlo Park, CA, p. 415.
2. H. J. Hagemeyer, "Acetaldehyde," pp. 97—112 in Kirk-Othmer Encyclopedia of
Chemical Technology, 3d ed., vol. 1, edited by M. Grayson et al., Interscience,
New York, 1978.
3. R. Jira, W. Blau, and D. Grimm, "Acetaldehyde Via Air or Oxygen," Hydrocarbon
Processing 55(3), 97--100 (March 1976).
4. P. L. Morse, Acetaldehyde, Report No. 24, A private report by the Process Economics
Program, pp. 3 and 10, Stanford Research Institute, Menlo Park, CA (April 1967).
5. Hoechst-Uhde Corp., "Acetaldehyde from Ethylene (Aldehyd GmbH)," Chemical Week
46(11) 135 (1967).
6. Veba-Chemie AG, "Acetaldehyde," Hydrocarbon Processing 56(11), 118 (1977).
*A reference located at the end of a paragraph usually refers to the entire para-
graph. If another reference relates to certain portions of the paragraph, the
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 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. ACETALDEHYDE FROM ETHYLENE BY THE TWO-STEP AIR-OXIDATION PROCESS
1. Typical Plant
The typical plant for this study has a capacity of 113.5 Gg/yr, based on 8760 hr
of operation per year. Although not an actual operating plant, it is typical
of existing plants. The plant utilizes the model two-step process (Fig. III-l)
and fits today's acetaldehyde manufacturing and engineering technology for that
process.
The Celanese Clear Lake, TX, plant and the Texas Eastman Longview, TX, plant
have two production units each with reported capacities of about 113.5 Gg/yr
per unit. The Celanese Bay City plant has one unit with a reported capacity of
136 Gg/yr (see Table II-2). Foreign plants based on ethylene have capacities
ranging from 21.8 to 134.3 Gg/yr.
2. Sources and Emissions
Uncontrolled emission rates and sources for the model plant are summarized in
Table IV-1 and are described below. The process emission rates are in the range
2--4
of actual emission data reported by existing plants. The off-air and off-gas
absorbers (scrubbers) for the purpose of this report are considered to be integral
components of process equipment important to the efficiency of the process and
not emission control devices as such. Potential storage, handling, and fugitive
emissions were calculated from characteristics of the model process that were
2--4
based on data on existing plants.
-------
IV-2
Table IV-1. Total VOC from Uncontrolled Emissions Produced by
Model Plant Using Two-Step Air-Oxidation Process
Emission Source
Off-air absorber vent
Off -gas absorber vent
b
Intermittent
Storage
Handling
Fugitive
Secondary
Stream
Designation
(Fig.III-1)
A
B
B
D
F
H
J,K
Emissions
Ratio
(g/kg)
2.27
2.79
0.005
0.17
0.47
0.58
c
Rate
(kg/br)
29.4
36.1
0.06
2.3
6.1
7.5
c
g of total VOC per kg of acetaldehyde produced.
Average rate for entire year, based on one startup per year.
CSecondary emissions were not calculated; potential for significant
secondary emissions exists.
-------
IV-3
a. Off-Air Vent Emissions -- The off-air absorber vent (Vent A, Fig. III-l) is a
principal source of emissions from the acetaldehyde production plant. Nitrogen
and other inert or unreacted components of the air fed to the second-stage reactor
and gases or unabsorbed vapors generated in the catalyst oxidation or regenera-
tion process are discharged from this vent. The VOC components, principally
acetaldehyde, methyl chloride, and ethyl chloride, amount to only 0.15 wt % of
the total flow from this vent. The compositions of the model-plant uncontrolled
emissions are given by Table IV-2. The average emission during normal operation
is given in Table IV-1. The estimated flow from this vent is 271 m /min.
b. Off-Gas Vent Emissions — The off-gas vent (Vent B, Fig. III-l) discharges gases
and low-boiling-point VOC vapors separated in product purification operations.
The VOC components (see Table IV-2) make up 4.8 wt % of the total discharge
from this vent. The average emission during normal operation is given in Table IV-1,
The estimated flow from this vent is 8.9 m /min.
c. Intermittent Air Emissions -- The acetaldehyde plant is normally operated con-
tinuously and is shut down for annual maintenance. During startup the amount
of ethylene in the off-gas (Vent B) may run as high as 25 to 50% for 2 to 6 hr.
After the catalyst is activated, the ethylene content decreases to normal levels.
The intermittent emissions reported in Table IV-1 were calculated based on one
startup per year.
d. Fugitive Emissions -- Process pumps and valves are potential sources of fugitive
emissions. The model plant is estimated to have 19 pumps handling VOC, with 12
used for light-liquid service and 7 for heavy-liquid service. The model plant
is estimated to have 648 process valves handling VOC, with approximately 30%
used for heavy-liquid service, 50% for light-liquid service, and 20% for gas/
vapor service. The model plant is estimated to have 16 relief valves, with
approximately 80% used for gas/vapor service and 20% for light-liquid service.
The fugitive-emission factors from Appendix B were applied to determine the
fugitive emissions shown in Table IV-1.
e. Storage and Handling Emissions — Emissions result from the storage and handling
of acetaldehyde. Sources for the model plant are shown in Fig. III-l. Not
shown on the process flow diagram are surge tanks and catalyst, crude aldehyde,
-------
IV-4
Table IV-2. Uncontrolled Emission and Waste Composition for Model Two-Step Process
Component Formula
VOC
Ethylene C H
Acetaldehyde C2H4°
Methyl chloride CH3C1
Ethyl chloride C2H5C1
Methylene chloride CH Cl
Chloroform CHC13
Acetic acid C2H4°2
Chloroacetaldehyde C H OC1
Acetyl chloride C?^2(~>C''~2
Chloral C2HOC13
Paraldehyde fC2H4O) 3
Other organic
Total VOC
Other gases
Carbon dioxide CO
Nitrogen N
Oxygen O
Argon A
Total other gases
Water H2O
Total stream
Emission Ratio
(g/kg)a
Vent A Vent B
Off-Air Off -Gas
1.00
1.47 0.19
0.76 0.37
0.04 0.78
0.10
0.35
2.27 2.79
37.75 22.65
1430.55 30.15
11.67 0.50
24.39 0.67
1504.36 53.79
1-79 0.82
1508.44 57.58
Waste Discharge Ratio
(g/kg)b
Discharge K Discharge J
Wastewater Side Cut
7.8
13.9 0.6
5.5
4.2 5.0
2.1 3.4
1.6
4.0 2.0
25.8 24.3
795.6 25.5
821.4 49.8
*g of off-air or off-gas per kg of acetaldehyde.
Dg of discharge K or J per kg of acetaldehyde.
-------
IV-5
and process-water storage tanks. These tanks are operated at positive pressure
and are vented back to the process and therefore do not contribute to the storage
emissions. Emissions from the storage of the side-cut organic by-products (Dis-
charge J) and/or wastewater (Discharge K) are considered under secondary emissions.
The low boiling point and relatively high vapor pressure of acetaldehyde require
that it be stored and handled in pressurized tanks, which are padded with nitrogen.
The model plant has three 1079-m spherical storage tanks 12.8 m in diameter.
The tanks are maintained at ambient temperature and between 207 and 377 kPa
pressure. The calculated average storage emissions based on 45 turnovers per
year are given in Table IV-1.
All of the acetaldehyde produced in the model plant is shipped in pressurized
tank cars. In the uncontrolled plant, tank cars are vented before they are
filled. During filling, the tank car is vented back to the acetaldehyde storage
tank. The calculated handling emission from the tanks being vented before they
are filled is given in Table IV-1.
f- Secondary Emissions -- The model plant discharges approximately 10.6 m of process
wastewater per hour that contains 334-kg/hr VOC (Discharge K, Fig. III-l) and
approximately 0.6 m of chlorinated by-product waste per hour that contains
315-kg/hr VOC (Discharge J). These streams can be significant sources of secondary
emissions resulting from desorption or evaporation before they are ultimately
treated and disposed of. Probable secondary emissions were not calculated.
B. ACETALDEHYDE FROM ETHYLENE BY THE SINGLE-STEP OXYGEN-OXIDATION PROCESS
There are no acetaldehyde plants in the United States employing the single-
step oxidation process; thus no actual emission data could be obtained. The
following inferences are based on a comparison of the process chemistry, model
process flow sheets, and published data pertaining to the two processes.
1. Sources and Emissions
5-~7
The process yields reported for both processes are equal, which indicates
nearly equal amounts of total carbon content in the air emissions, wastewater,
and by-products generated by each process. The amount of HCl consumed by the
-------
IV-6
two-step process is reported to be 2.6 to 3.7 times that used by the single-
5 7
step process. ' This indi
by the single-step process.
5 7
step process. ' This indicates that fewer chlorinated by-products are generated
2. Purge-Gas Vent Emissions
The purge-gas vent (Vent A, Fig. III-2) is required to purge contaminants from
the recycle gas stream. The volume of the vent stream should be relatively
small, depending on the level that contaminates can be tolerated in the reaction
system. Since high-purity oxygen and ethylene are fed to the reactor, the purge
stream contains principally unreacted ethylene, acetaldehyde, by-product carbon
dioxide, and water vapor, together with small quantities of oxygen, argon and
nitrogen from the oxygen feed, ethane from the ethylene feed, and chlorinated
by-product compounds. The concentration and quantity of total VOC (ethylene and
acetaldehyde plus the chlorinated compounds) discharged from this vent are expected
to be relatively high.
3. Off-Gas Vent Emissions
The off-gas (Vent B, Fig. III-2) discharges the low-boiling-point light organic
by-products separated from the process stream during product purification. The
vent stream volume would be expected to be small, containing largely chlori-
nated by-products and small quantities of ethylene, acetaldehyde, and inert
gases.
4. Secondary Emissions
Since water absorption is used for separating the crude product from the recycle
gas stream, the quantity of wastewater discharged would be much higher than
that reported for the two-step process. The VOC contained in the water, however,
would be less than that of the two-step process since less by-product waste is
produced. The potential for secondary emissions therefore should be less.
5. Other Emissions
The intermittent, fugitive, and storage and handling emissions are expected to
be similar to those of the two-step process.
-------
IV-7
C. REFERENCES*
1. P. L. Morse, Acetaldehyde, pp. 15 and 16 in Report No. 24, A private report by
the Process Economics Program, Stanford Research Institute, Menlo Park, CA
(April 1967).
2. J. J. Cudahy and J. F. Lawson, IT Enviroscience, Inc., Trip Report to Celanese
Plant, Celanese Chemical Co., Clear Lake City, TX, Sept. 22, 1977 (on file at
EPA, ESED, Research Triangle Park, NC).
3. J. J. Cudahy and J. F. Lawson, IT Enviroscience, Inc., Trip Report to Texas Eastman
Plant, Texas Eastman Chemical Co., Longview, TX, Nov. 16, 1977 (on file at EPA,
ESED, Research Triangle Park, NC).
4. P. E. Hime, Celanese Chemical Co., response to Texas Air Control Board 1975
Emissions Inventory Questionnaire for Celanese Chemical Co., Bay City, TX, Plant,
Mar. 19, 1976.
5. R. Jira, W. Blau, and D. Grimm, "Acetaldehyde Via Air or Oxygen," Hydrocarbon
Processing 55(3), 97--100 (March 1976).
6. P. L. Morse, Acetaldehyde, pp. 3 and 10 in Report No. 24, A private report by
the Process Economics Program, Stanford Research Institute, Menlo Park, CA
(April 1967).
7. Hoechst-Uhde Corp., "Acetaldehyde from Ethylene (Aldehyd GmbH)," Chemical Week
46(11), 135 (1967).
*A reference located at the end of a paragraph usually refers to the entire para-
graph. If another reference relates to certain portions of the paragraph, the
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. ACETALDEHYDE FROM ETHYLENE BY THE TWO-STEP AIR-OXIDATION PROCESS
1. Off-Air Vent Emissions
The emissions from the off-air vent (Vent A) remain uncontrolled for the typical
plant. The total flow from this vent is 16,275 m3/hr (19,527 kg/hr), containing
29.4-kg/hr, or 0.15 wt %, VOC. The stream contains less than 0.8 wt % oxygen,
with the bulk of the emission being nitrogen gas (see Table IV-2).
Typical of existing plants in the United States, "" a water scrubber (absorber)
is employed to recover residual amounts of acetaldehyde from the off-air stream
before it is vented. These scrubbers are operated under high pressure, and
chilled water is used to achieve maximum absorption efficiency.
The gases vented are estimated to have a heating value of approximately 37 kJ/m .
A minimum heating value of "115 Btu/ft3" (ref. 4) (4300 kJ/m ) is required to
successfully flare the gas.4 Thus the use of a flare or other thermal oxidation
control devices for destruction of the residual VOC in this stream would require
very large amounts of auxiliary fuel. Combining this stream with the emission
stream from vent B would raise the heating value to only about 90 kJ/m . Emissions
from the scrubber off-air are not controlled in existing plants or in the typical
plant described (see Table V-l).
2. Off-Gas Vent Emissions
The emissions from the typical plant off-gas vent (Vent B) are controlled by a
flare. An emission reduction of 99%, typical of emission reductions achieved
by a properly designed flare, was used to calculate the controlled emission
rate given in Table V-l. Usually the ethylene feedstock is produced on-site in
an adjacent plant. The flare system associated with the ethylene plant might
be used jointly for both the ethylene and the acetaldehyde plants.
The total flow from vent B of the typical plant is 531 m /hr (743 kg/hr) and
contains about 36-kg/hr, or 4.8 wt %, VOC. The heating value of the mixture is
about 1735 kJ/m3. Flaring of this emission requires that about 45-m /hr natural
gas be added for minimum combustion conditions to be achieved. Flaring produces
undesirable HC1 emissions.
-------
Table V-l. Total Controlled VOC Emissions for Typical Plant Using Two-Step Air-Oxidation Process
Emission
Source
Off-air absorber vent
Off-gas absorber vent
b
Intermittent
Storage
Handling
Fugitive
Secondary
Stream
Designation
(Fig.III-1)
A
B
B
D
F
H
J,K
Control Device
or Technique
None
Flare
Flare
Recycle
Recycle
Repair and
maintenance
c
Emission
Reduction
99
99
100
99
91
c
Emissions
Ratio
(gAg)
2.27
0.03
Negligible
Negligible
0.005
0.2
c
Rate
(kg/hr)
29.4
0.36
Negligible
Negligible
0.06
2.6
c
g of total VOC per kg of acetaldehyde produced.
Intermittent startup emissions are discharged from vent B, which is controlled by a flare.
'Secondary emissions and emission control measures were not defined.
I
tvj
-------
V-3
A thermal oxidation system with heat recovery is a possible control alternative.
Although auxiliary fuel would be required to maintain stable combustion, thermal
oxidation would produce emission reduction efficiencies of greater than 99%.
Combustion of the mixture, however, produces acid gas emissions, which must be
considered in design of the equipment. Thermal oxidizer systems and efficiencies
are discussed in a separate EPA document.
3. Intermittent Air Emissions
Intermittent high emissions from vent B produced during plant startup are reduced
by 99% by the flare system used for control of emissions from vent B (see Table V-l).
A flare system is ideally suited for burning vent streams of widely varying
quantity and composition. If a thermal oxidation system with heat recovery
were used to control emissions from vent B, the increased heat rate during
startup could be a problem in sizing and operation of the oxidation system.
4. Fugitive Sources
Controls for fugitive sources are discussed in a separate EPA document covering
fugitive emission from the synthetic organic chemicals manufacturing industry.
Control of emission from the pumps and valves can be attained by an appropriate
leak detection system followed by repair maintenance. Controlled fugitive
emissions have been calculated with the factors given in Appendix B and are
included in Table V-l. The factors are based on the assumption that major
leaks are detected and repaired.
5. Storage and Handling Sources
In the typical plant acetaldehyde storage emissions are controlled by the dis-
charge from the storage-tank pressure-relief valves being returned to the off-
gas absorber. Emissions from the absorber (Vent B) are flared, which results
in an overall reduction of essentially 100% (see Table V-l).
Handling emissions from the venting of incoming empty tank cars are controlled
by venting to the off-gas absorber. While the tank cars are being filled they
are vented back to the acetaldehyde storage tank. A control efficiency of 99%
obtained with the absorber system was used to calculate the controlled emissions
given in Table V-l.
-------
V-4
6. Secondary Emissions
Secondary emissions can result from evaporation of VOC contained in aqueous
wastes going to wastewater treatment (Discharge K) and from processing or dis-
posal of the side-cut by-products (Discharge J). Considerable potential exists
for large secondary emissions because of the quantity and concentration of VOC
contained in these streams. Some form of pretreatment would be required to
lower the organic concentration of the wastewater stream before it is to be
7 7
treated by conventional biodegradation. Side-cut organics can be incinerated.
One plant disposes of both streams in a deep-well system. Control of secondary
emissions will be discussed in a future EPA report.
B. ACETALDEHYDE FROM ETHYLENE BY THE SINGLE-STEP OXYGEN-OXIDATION PROCESS
No acetaldehyde plants using the single-step oxidation process have been built
in the United States; thus no data are available on the emissions and/or emis-
sion controls used.
Emissions from the purge-gas vent (Vent A) should have sufficient heating value
to permit control by flaring or other forms of thermal oxidation. Emissions
from vent B and intermittent, storage and handling, and fugitive emissions should
be controllable by the same measures used in the two-step process. Secondary
emissions from the single-step process are less of a problem since there are
fewer by-products formed and the wastewater is rather dilute and treatable by
biological degradation.
C. CONTROL MEASURES CURRENTLY USED
The emission control measures now in effect in domestic acetaldehyde plants
are discussed in Appendix C.
-------
V-5
D. REFERENCES*
1. J. J. Cudahy and J. F. Lawson, IT Enviroscience, Inc., Trip Report for Celanese
Plant, Celanese Chemical Company, Clear Lake City, TX, Sept. 22, 1977 {on file
at EPA, ESED, Research Triangle Park, NC).
2. J. J. Cudahy and J. F. Lawson, IT Enviroscience, Inc., Trip Report for Texas
Eastman Plant, Texas Eastman Chemical Co., Longview, TX, Nov. 16, 1977 (on
file at EPA, ESED, Research Triangle Park, NC).
3. P. E. Hime, Celanese Chemical Co., response to Texas Air Control Board 1975
Emissions Inventory Questionnaire for Celanese Chemical Co., Bay City, TX Plant,
March 19, 1976.
4. J. F. Straitz, III, "Make the Flare Protect the Environment," Hydrocarbon
Processing 56(10), 134 (1977).
5. 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).
6. J. W. Blackburn, IT Enviroscience, Inc., Control Device Evaluation. Thermal
Oxidation (July 1980) (EPA/ESED report. Research Triangle Park, NC).
7. R. Jira, W. Blau, and D. Grimm, "Acetaldehyde Via Air or Oxygen," Hydrocarbon
Processing 55(3), 97--100 (March 1976).
*A reference located at the end of a paragraph usually refers to the entire para-
graph. If another reference relates to certain portions of the paragraph, the
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.
-------
APPENDIX A
Table A-l. Properties of Acetaldehyde, Methyl Chloride, Ethyl Chloride, and Chloroform
Acetaldehyde'
Methyl Chlorideb
Ethyl Chloride
Chloroform
Synonyms
Molecular formula
Molecular weight
Physical state
Dens ity
Vapor pressure
Vapor specific gravity
Boiling point at 1 atm
Water solubility
Acetic aldehyde,
ethyl aldehyde
C2H4°
44.05
Clear liquid
0.7834 g/ml at 18°C
1.23 atm at 25°C
1.52
20.8°C
Infinite
Chloromethane
CH3C1
50.49
Gas
0.9159 at 20°C/4°C
2.83 atm at 25°C
1.78
-24.2°C
4.9 g/liter
Chloroe thane ,
muriatic acid
64.52
Liquid or gas
0.8978 at 20°C/4°C
20 mm at 21 °C
2.22
12.3°C
5.7 g/liter
Trichloromethane
CHC13
119.39
Liquid
1.4984 at 15°C
200 mm at 25.9°C
4.12
61.26°C
8.0 g/liter 3
— • . h
aFrom: J. Dorigan et _al., "Acetaldehyde", p. AI-6 in Scoring of Organic Air Pollutants. Chemistry, Production and
Toxicity of Selected Synthetic Organic Chemicals (Chemicals A-C), Rev. 1, Appendix I, MTR-7248, MITRE Corp., McLean, VA
(September 1976) .
b"Methyl Chloride" ibid. (Chemicals F-N), p AIII-174.
c"Ethyl Chloride," ibid. (Chemicals D-E), p AII-254.
d"Chloroform," ibid. (Chemicals A-C), p A AI-265.
-------
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 Factor5
(kg/hr)
Pump seals ,
Light-liquid service
Heavy-liquid service
Pipeline valves
Gas/vapor service
Light-liquid service
Heavy-liquid service
Safety/relief valves
0.12
0.02
0.021
0.010
0.0003
0.03
0.02
0.002
0.003
0.0003
Gas/vapor service
Light-liquid service
Heavy- liquid service
Compressor seals
Flanges
Drains
0.16
0.006
0.009
0.44
0.00026
0.032
0.061
0.006
0.009
0.11
0.00026
0.019
Based on monthly inspection of selected equipment; no inspection of
heavy-liquid equipment, flanges, or light-liquid relief valves,-
10,000 ppmv VOC concentration at source defines a leak; and 15 days
allowed for correction of leaks.
Light liquid means any liquid more volatile than kerosene.
*Radian Corp., Emission Factors and Frequency of Leak Occurrence for Fittings
in Refinery Process Units, EPA 600/2-79-044 (February 1979).
-------
C-l
APPENDIX C
EXISTING PLANT CONSIDERATIONS
Table C-l lists process control devices reported to be in use by industry. To
gather information for the preparation of this report site visits were made to
manufacturers of acetaldehyde (AcH). Trip reports have been cleared by the
companies concerned and are on file at EPA, ESED, in Research Triangle Park,
NC.1'2
A. PROCESS EMISSIONS FROM EXISTING PLANTS1
1. Celanese Chemical Co., Clear Lake, TX
The acetaldehyde production facility consists of two identical continuous trains
of equipment to produce acetaldehyde by the two-stage process for air oxidation
of ethylene. The total published capacity of the Celanese Clear Lake, TX, acetal-
dehyde plant is 500 million pounds per year. The process was licensed from
Aldehyde GmbH, a jointly owned company formed by Hoechst, A. G. and Wacker-Chemie
GmbH. The trains were built in 1967 and 1971.
Emissions from the process can be from two sources: the high-pressure vent absorber
and the low-pressure vent absorber. Emission compositions and flow data were
reported in Table C-2.
The methane and ethane concentrations in these streams are mainly from the
methane used as an inert-gas blanket on the acetaldehyde storage tanks. Methane
padding is no longer used and consequently the methane and ethane concentrations
in the low-pressure vent absorber vent gas will be much lower than that shown.
Both the high-pressure and low-pressure vent absorbers are used for product
recovery and are very important for process efficiency. The low-pressure vent
absorber is also fed an organic stream from the light-ends column and a scrubbing
water stream from the finishing column bottoms, as well as vent streams from
the process, the acetaldehyde storage tanks, and the acetaldehyde tank car loading
systems. The unabsorbed vent gases from the low-pressure vent absorber are fed
to a process flare. The vent gas from both acetaldehyde process trains accounts
for about 70 to 80% of the load to this flare.
-------
Table C-l- Control Methods Currently Used by the Domestic Acetaldehyde Industry'
Control Methods For
Company and Location
Celanese
Bay City, TX
Celanese
Clear Lake City, TX
Texas Eastmen
Longview, TX
b
Off-Air
Vent
None
None
None
b
Off- Gas
Vent
Flare
Flare
None
Product
Storage
Recycle
Recycle
Recycle
Product Side-Cut
Handling Organics
Recycle c
d e
Recycle Deep well
Recycle c
Wastewater
c
Deep well
f
*For those plants producing acetaldehyde from ethylene by the two-step air-oxidation process; see Table II-2,
"'The off-air and off-gas absorbers (scrubbers) for the purpose of this report are considered as integral
components of process equipment and not as emission controls as such.
"Not reported.
Empty tank cars are vented to a flare before they are refilled.
"Combined with wastewater.
Specific data not reported; wastewater from the Longview facility goes to anerobic lagoons.
n
i
NJ
-------
C-3
Table C-2. High- and Low-Pressure Scrubber
Emissions from Celanese Plant
Waste Gas
C2H4°
H20
N2
°2
CO,
C2H4
CH3C1
C2H5C1
CH2C12
CHC13
N2
°2
co2
CK4
C2H6
Emissions
(wt %)
High Pressure Scrubber Emissions
(Off-Air Vent)
0.3
0.2
95.3
1.9
2.3
100.0
Low Pressure Scrubber Emissions
(Off-Gas Vent)
0.9
1.9
2.1
0.3
0.7
37.5
0.9
41.9
11.9
2.9
101.0
(lb/1000 Ib of AcH)
4.3
2.7
1,540.2
30.9
37.8
1,615.9
0.5
1.1
1.1
0.1
0.4
20.6
0.5
23.0
6.5
1.6
55.4
-------
C-4
2
2. Texas Eastman, Longview, TX
The Texas Eastman acetaldehyde production facility at Longview was also licensed
from Aldehyde GmbH and with regard to process emissions appears essentially
identical to the Celanese acetaldehyde production facility at Clear Lake, TX.
Typical emissions data reported for the Texas Eastman facility are given in
Table C-3.
Both scrubbers recover product and are very important for process efficiency.
Vent streams from process and product storage plus tank car loading emissions
are also directed to the low-pressure scrubber. No additional control device
is used for control of emissions from the high- and low-pressure scrubbers.
B. TOTAL INDUSTRY EMISSIONS
Table C-4 lists the estimated emissions for the acetaldehyde industry for 1979.
This estimate is based on emission data received from the major acetaldehyde
1 2
producers ' and/or emission rates calculated for the typical plant. It is
estimated that the total emissions from all plants during 1979 were approximately
1.1 Gg.
The emissions from these plants would have been 2.4 Gg during 1979 if they
had been uncontrolled.
C. RETROFITTING CONTROLS
The primary difficulty associated with retrofitting may be in finding space to
fit the control device into the existing plant layout. Because of the costs
associated with this difficulty it may be appreciably more expensive to retrofit
emission control systems in existing plants than to install a control system
during construction of a new plant.
-------
C-5
Table C-3. High- and Low-Pressure Scrubber
Emissions from Texas Eastman Plant
Waste Gas
C2H4°
CH Cl
co2
N2
C2H5C1
H2°
Argon
C2H4
co2
N2
C2H4°
CH3C1
C2H5C1
H2°
Argon
Emissions
db/hr)
High-Pressure Scrubber Emissions
(Off-Air Vent)
1
20
457
37,668
<100
47
649
Low-Pressure Scrubber Emissions
(Off -Gas Vent)
5
741
1,158
<2
6
24
21
20
{vol %)
0.0017
<0.05
0.8
97.8
<0.2
0.2
1.2
0.3
27.8
68.3
<0.1
0.2
0.7
1.9
0.8
-------
C-6
Table C-4. Estimated 1979 Acetaldehyde Industry Emissions
Total VOC Emissions5 (Gg/yr)
Source Uncontrolled Current Controlled
Off-air vent A 0.745 0.745
Off-gas vent B 1.119 0.240
Storage 0.077 Neg
Handling 0.208 0.001
Fugitive 0.253 0.128
Secondary d d
Total 2.402 1.114
afiased on estimated total 1978 production of 448 Gg.
Based on uncontrolled emissions reported by industry and/or emission rates
calculated for uncontrolled typical plant (Table IV-1).
Current control represents the degree of emission control obtained by industry
in 1976, based on control measures reported by industry (Table C-l) and the
control efficiencies described for typical plant controls (Table V-l).
Available data insufficient to estimate secondary emissions; however, it is
believed that secondary emissions are currently small because of control
measures taken by industry.
-------
C-7
D. REFERENCES*
1. J. J. Cudahy and J. F. Lawson, IT Enviroscience, Inc., Trip Report for Celanese
Plant, Celanese Chemical Company, Clear Lake City, TX, Sept. 22, 1977 (on file
at EPA, ESED, Research Triangle Park, NC).
2. J. J. Cudahy and J. F. Lawson, IT Enviroscience, Inc., Trip Report for Texas
Eastman Plant, Texas Eastman Chemical Co., Longview, TX, Nov. 16, 1977 (on
file at EPA, ESED, Research Triangle Park, NC).
3. A. K. Rafie and L. S. Soder, "Acetaldehyde—Product Review," p 601.50201
in Chemical Economics Handbook, Stanford Research Institute, Menlo Park, CA
(March 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.
-------
7-i
REPORT 7
ETHANOLAMINES
T. L. Schomer
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 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.
D52E
-------
7-iii
CONTENTS OF REPORT 7
Page
I. ABBREVIATIONS AND CONVERSION FACTORS 1-1
II. INDUSTRY DESCRIPTION II-l
A. Introduction II-l
B. Ethanolamines Usage and Growth II-l
C. Domestic Producers II-l
D. References II-5
III. PROCESS DESCRIPTIONS III-l
A. Introduction III-l
B. Ethanolamines from Ethylene Oxide and Ammonia III-l
C. Process Variation III-3
D. References III-4
IV. EMISSIONS IV-1
A. Emissions IV-1
B. References IV-3
V. IMPACT ANALYSIS V-l
A. Industry V-l
B. References V-2
APPENDICES OF REPORT 7
Page
APPENDIX A. PHYSICAL PROPERTIES OF ETHANOLAMINES A-l
-------
7-v
TABLES OF REPORT 7
Number
II-l
II-2
A-l
Usage of Ethanolamines
Ethanolamines Capacity
Physical Properties of Ethanolamines
Page
II-2
II-3
A-l
FIGURES OF REPORT 7
Number
II-l
III-l
Locations of Plants Manufacturing Ethanolamines
Production of Ethanolamines by the Ethylene Oxide—Ammonia
Process
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)
Multiply By
9.870 X 10"6
9.480 X 10"4
(°C X 9/5) + 32
3.28
3.531 X 101
6.290
2.643 X 102
1.585 X 104
1.340 X 10"3
3.937 X 101
1.450 X 10~4
2.205
2.778 X 10"4
Standard Conditions
68°F = 20°C
1 atmosphere = 101,325 Pascals
PREFIXES
Prefix
T
G
M
k
m
Symbol
tera
giga
mega
kilo
milli
micro
Multiplication
Factor
1012
109
106
103
10"3
10"6
Example
1 Tg = 1 X 1012 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 ug = 1 X 10~6 gram
-------
II-l
II. INDUSTRY DESCRIPTION
A. INTRODUCTION
Ethanolamines production was chosen for study as part of the family of products
produced from ethylene oxide and because preliminary estimates indicated that
emissions of volatile organic compounds (VOC) were relatively high. This study
is presented as an abbreviated product report because information from producers
indicates that process VOC emissions are negligible.
Three ethanolamine products were studied: monoethanolamine (MEA), diethanolamine
(DBA), and triethanolamine (TEA). They are low-vapor-pressure colorless liquids
at room temperature or slightly above. Physical property data are given in
Appendix A.
B. ETHANOLAMINES USAGE AND GROWTH
Table II-l gives a breakdown of the use of ethanolamines. In 1978 the demand
for ethanolamines was 165 Gg. The production ratio in recent years has been
32% MEA, 31% DBA, and 37% TEA.1 MEA is used mainly for scrubbing acid gases
from gas streams, DBA is used chiefly in fatty alkanolamides for liquid detergents
and textile chemicals, and TEA is used in the production of fatty acid soaps
for dry cleaning and in cosmetics.2
Through 1983 ethanolamine consumption is expected to increase at the rate of 3%
per year,- no significant new markets are expected.1 The availability and cost
of ethanolamines depend on the availability and cost of the raw materials ethylene
oxide and ammonia. All the domestic producers have captive ethylene oxide produc-
tion.2
C. DOMESTIC PRODUCERS
As of early 1979 there were four domestic producers of ethanolamines in five
different locations.1 Table II-2 lists the producers and their rated capacities.
Figure II-l shows the plant locations. Dow Chemical started up the Plaquemine, LA,
facility with a rated capacity of 56.7 Gg/yr,- they mothballed a facility in
Freeport, TX, that had a capacity of 22.7 Gg/yr. The Freeport plant is not likely
to reopen.1 In 1979 Jefferson Chemical started up a new plant in Port Neches, TX,
with a rated capacity of 68 Gg/yr.1
-------
II-2
Table II-l. Usage of Ethanolamines*
End Use Percent of Total Usage
Detergents (textile, toilet goods, metal and 40
other specialty surfactants)
Gas conditioning and petroleum use 25
Other (including agricultural intermediates 15
and cement grinding aids)
Export 20
*See ref. 1.
-------
II-3
Table II-2. Ethanolamines Capacity0
Company
Dow Chemical Co.
Dow Chemical Co.
Jefferson Chemical Co.
Olin Chemical Co.
Union Carbide
Total
Location
Midland, MI
Plaquemine, LA
Port Neches, TX
Brandenburg, KY
Seadrift, TX
Annual Production
Capacity
(Gg) (1979)
22. 7b
56.7
68.1
13.6
104.3
265.4
See ref. 1.
This unit mostly produces isopropanolamines but can also
produce ethanolamines.
-------
II-4
1. Dow Chemical Co., Midland, MI
2. Dow Chemical Co., Plaquemine, LA
3. Jefferson Chemical Co., Port Heches, TX
4. Olin Chemical Co., Brandenburg, KY
5. Union Carbide, Seadrift, TX
Pig. II-l. Locations of Plants Manufacturing Ethanolamines
-------
II-5
D. REFERENCES*
1. "Chemical Profile on Ethanolamines," p. 9 in Chemical Marketing Reporter
(Apr. 9, 1979).
2. J. L. Blackford, "Ethylene Oxide," pp. 654.5032R,S in Chemical Economics^Handbook,
Stanford Research Institute, Menlo Park, CA (September 1976).
*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
III. PROCESS DESCRIPTIONS
INTRODUCTION
Ethanolamines are produced commercially in the United States by the liquid-phase
reaction of ethylene oxide and aqueous ammonia. Other methods of synthesizing
ethanolamines have been developed, such as the reaction of ethylene chlorohydrin
with ammonia or the hydrogenation of formaldehyde cyanohydrin; however, these
methods are not commercially practiced.1
ETHANOLAMINES FROM ETHYLENE OXIDE AND AMMONIA
The ethanolamines are produced by the following series of chemical reactions:1
NH3 + CH2-CH2
0
(ammonia) (ethylene oxide)
NH2CH2CH2OH
(MEA)
NH(CH2CH2OH)2
(DEA)
CH2-CH2
0
(ethylene oxide)
NH2CH2CH2OH
[monoethanolamine (MEA)]
-> NH(CH2CH2OH)2
[diethanolamine (DEA)]
CH2-CH2
0
(ethylene oxide)
N(CH2CH2OH)3
[triethanolamine (TEA)]
The process is noncatalytic and is carried out in the liquid phase in the presence
of water. The reactions are strongly exothermic, about 100 kJ per g-mole of
ethylene oxide reacted.2 The distribution of products that can be obtained is
dependent on the ratio of ammonia to ethylene oxide used. Excess ammonia favors
MEA formation. Also, desired product distribution can be achieved by recycling
MEA and/or DEA to be further reacted with ethylene oxide.1
A typical flow diagram for the continuous manufacture of ethanolamines is shown
by Fig. III-l. Ethylene oxide (stream 1) and aqueous ammonia (stream 2) are
fed to a reactor. The reaction conditions usually are a temperature range of
50 to 100°C, a pressure of 1 to 2 MPa, and an excess of 28 to 50% aqueous ammonia.1
The reactor effluent (stream 3) is stripped of unreacted ammonia and some water
-------
a
STBIPPCE.-
Dl-
E.TVt
COl-UUU
ji
A.MWOMIA
*6scjReeB
H2h,@
-VA.C.
^ — , — •*
(
^
•nz.*-
e.THAU<
COV-UM
/-^vAc-
^
r0n
DEHYDRATIOU
couumu
>"\
MCUO-
TO
t3ISPO=>Al_
/*
VAC.
WOUO-
E-THAUCLM/IUE
COLUV/U
T12.1-
ET H A u. OCJ\V^ IU E.
H
h-t
M
I
to
Fig. III-l. Production of Ethanolamines Toy the Ethylene Oxide—Ammonia Process
-------
III-3
(stream 4) in an ammonia stripper operated under pressure. This ammonia, together
with fresh feed (stream 5), is absorbed in recycled water in the ammonia absorber
and fed back to the reactor (stream 2). The noncondensable overhead gas (stream 6)
from the ammonia stripper is scrubbed of ammonia in an ammonia scrubber with
recycle water (stream 7) and is vented (A). Inert gases enter the system with
the ethylene oxide feed, which is stored under a nitrogen pressure pad.3
The ammonia stripper bottoms (stream 9) are vacuum distilled in a series of dis-
tillation columns to sequentially remove overhead water (stream 7), which is
recycled, and MEA, DBA, and TEA (streams 10, 11, 12), which are products. Non-
condensables from the vacuum distillation columns are vented (B) from the
vacuum-jet discharges, and the vacuum-jet waste waters are discarded to waste
treatment. The bottoms residue (stream 13) from the triethanolamine column is
sent to waste treatment or is sold. The product storage tanks are ordinarily
equipped with steam-heating coils to keep the products liquid and are padded
with a dry inert gas such as nitrogen to prevent product discoloration.
C. PROCESS VARIATION
No detailed descriptive information appears to have been published on modern
units for ethanolamines production. Various producers may be using widely dif-
ferent operating conditions and different distillation sequences.2 It is not
considered likely that these variations would have any appreciable effect on
emission sources.
-------
III-4
D. REFERENCES*
1. A. W. Hart, "Alkanolamines," pp. 810—814 in Kirk-Othmer Encyclopedia of Chemical
Technology, 2d ed., vol. 1, edited by A. Standen et al., Wiley, New York, 1963.
2. H. W. Scheeline, "Ethylene Glycols, Glycol Ethers and Ethanolamines," Report
No. 70, A private report by the Process Economics Program, Stanford Research
Institute, Menlo Park, CA (August 1971).
3. J. F. Lawson et al., IT Enviroscience, Inc., Emissions Control Options for the
Synthetic Organic Chemicals Manufacturing Industry Ethylene Oxide Product
Report (on file at EPA, ESED, Research Triangle Park, NC) (November 1978).
*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
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 at-
mosphere, participate in photochemical reactions producing ozone. A relatively
small number of organic chemicals have low or negligible photochemical re-
activity. 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.
As indicated in Fig. III-l, there are two potential process emission sources.
The ammonia scrubber vent (A) purges the small amount of nitrogen gas entering
the system with the ethylene oxide feed, about 1 g per kg of ethanolamines.1/2
There are no VOC emissions reported for this vent,1—3 and the ammonia content
is reported as approximately 100 ppm.2 The vacuum distillation system vacuum-
jet vents (B) purge the gases that may enter the system by leakage or that may
be used for control. The overhead distillate products are either water or low-
volatility, infinitely water-soluble organics; therefore the VOC emissions have
been calculated to be at most only a trace.1
There are two waste streams that are potential secondary emission sources. One
is the condensed steam from the vacuum jets. One producer reports this flow to
be about 8 g per kg of ethanolamines and to contain 1% organics.1 Such streams
are sent to biological treatment.1'2 The organics are low-volatility, infinitely
water-soluble materials and as such should not have a significant potential for
secondary emissions.4 The other waste stream is the bottoms (stream 13, Fig. III-l)
from the TEA column. One producer reports that this stream quantity is 22 g
per kg of ethanolamines produced and that it is sold as a waste product.1 Even
if this material is disposed of by being burned or sent to landfill, its potential
contribution to secondary emissions would be minor.
-------
IV-2
Fugitive and storage emissions are considered typical for the synthetic organic
chemicals manufacturing industry and are not discussed in this abbreviated report.
Fugitive and storage VOC emissions for the entire SOCMI are covered by separate
EPA reports.5'6
-------
IV-3
B. REFERENCES*
1. Olin Chemicals, letter dated May 17, 1978, in response to EPA's request for
information on emissions data on ethanolamines production facilities.
2. Dow Chemical U.S.A., letter dated Sept. 15, 1978, in response to EPA's request
for information on emissions data on ethanolamines production facilities.
3. Jefferson Chemical Company, Inc., letter dated May 9, 1978, in response to EPA's
request for information on emissions data on ethanolamines production facilities.
4. J. J. Cudahy and R. L. Standifer, IT Enviroscience, Inc., Secondary Emissions Report
(on file at EPA, ESED, Research Triangle Park, NC) (October 1979).
5. D. G. Erikson, IT Enviroscience, Inc., Storage and Handling Report (on file at EPA,
ESED, Research Triangle Park, NC) (October 1978).
6. D. G. Erikson, IT Enviroscience, Inc., Fugitive Emissions Report (on file at EPA,
ESED, Research Triangle Park, NC) (March 1979).
*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. IMPACT ANALYSIS
A. INDUSTRY1—3
The ethanolamines industry does not contribute any significant process or
secondary VOC emissions.
-------
V-2
B. REFERENCES*
1. Olin Chemicals, letter dated May 17, 1978, in response to EPA's request for
information on emissions data on ethanolamines production facilities.
2. Dow Chemical U.S.A., letter dated Sept. 15, 1978, in response to EPA's request
for information on emissions data on ethanolamines production facilities.
3. Jefferson Chemical Company, Inc., letter dated May 9, 1978, in response to EPA's
request for information on emissions data on ethanolamines 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.
-------
APPENDIX A
Table A-l. Physical Properties of Ethanolamines
Product
Synonyms
Molecular formula
Molecular weight
Physical state
Vapor pressure
Boiling point
Melting point
Liquid specific
gravity
Water solubility
Monoeth ano1amine
Ethanolamine, 2-amino-
ethanol, B-ethanol-
amine, colamine
C2H7N°
61.08
Liquid
800 Pa at 60°C (6 mm Hg)
170°C at 101.3 kPa
(760 mm Hg)
10.3°C
1.0180 at 20°C/4°C
Infinite
Diethanolamine
DEA, di(2-hydroxyethy1)-
amine, bis-hydroxy-
ethylamine, diethylol-
amine, diolamine
C4H11N°2
105.14
Liquid-solid
667 Pa at 138°C (5 mm Hg}
269.1°C at 101.3 kPa
(760 mm Hg)
28.0°C
1.0919 at 30°C/20°C
Infinite
Triethanolamine
2,2 ' , 2 "-Trihydroxy-
ethylamine
C6H15N°3
149.19
Liquid-solid
<1.3 Pa at 20°C
(0.01 ram Hg)
277°C at 20.0 kPa
(150 mm Hg)
21.2°C
1.1241 at 20°C/4°C
Infinite
J. Dorigan, B. Fuller, and R. Duffy, "Ethanolamine," p. AII-230 in Scoring of Organic Air Pollutants.
Chemistry, Production and Toxicity of Selected Synthetic Organic Chemicals (Chemicals D-E), MTR-7248,
Rev. 1, Appendix, Mitre Corp., McLean, VA (September 1976).
3Ibid., "Diethanolamine," p. AII-74.
'Ibid., "Triethanolamine" (Chemicals O-Z), p. AIV-260.
-------
8-i
REPORT 8
ETHYLENE GLYCOL
Ralph Love11
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
November 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.
D55E
-------
8-iii
CONTENTS OF REPORT 8
Page
I. ABBREVIATIONS AND CONVERSION FACTORS j^
II. INDUSTRY DESCRIPTION H-l
A. Reason for Selection II-l
B. Ethylene Glycols Usage and Growth II-l
C. Domestic Producers II-3
D. References II-6
III. PROCESS DESCRIPTION III-l
A. Introduction III-l
B. Model Process Ethylene Glycol from Ethylene Oxide III-2
C. Other Processes III-6
D. References III-8
IV. EMISSIONS IV-1
A. Model Plant IV-1
B. Sources and Emissions iv-1
C. References IV-6
V. APPLICABLE CONTROL SYSTEMS V-l
A. Types of Controlled Emissions V-l
B. References V-4
VI. IMPACT ANALYSIS VI-1
A. Environmental and Energy Impacts VI-1
B. Control Cost Impact VI-3
C. References VI-10
VII. SUMMARY VII-1
APPENDICES OF REPORT 8
A. PHYSICAL PROPERTIES OF GLYCOLS
B. AIR-DISPERSION PARAMETERS
C. LIST OF EPA INFORMATION SOURCES
D. FUGITIVE-EMISSION FACTORS
E. COST ESTIMATE SAMPLE CALCULATIONS
F. EXISTING PLANT CONSIDERATIONS
-------
8-v
TABLES OF REPORT 8
Number
II-l Ethylene Glycol Usage and Growth
II-2 Ethylene Glycols Capacity as of 1980
IV-1 Uncontrolled Emissions from Ethylene Glycol Model Plant
V-l Controlled Emissions from Ethylene Glycol Model Plant
VI-1 Environmental Impact of Controlled Model Plant
VI-2 Annual Cost Parameters
Vii-l Emission Summary for Model Plant
A-l Properties of (Mono)-Ethylene Glycol
A-2 Properties of Diethylene Glycol
A-3 Properties of Triethylene Glycol
B-l Air-Dispersion Parameters
F-l Emission Controls Currently Used
II-2
II-4
IV-3
V-2
VI-2
VI-4
VII-2
A-l
A-2
A-3
B-l
F-2
FIGURES OF REPORT 8
Number
II-l Location of Plants Manufacturing Ethylene Glycol
III-l Flow Diagram for Ethylene Glycol
III-2 Flow Diagram for Ethylene Glycol Process Variations
VI-1 Capital Cost vs Plant Capacity for Surface-Type Condensers
VI-2 Annual Cost vs Plant Capacity for Surface-Type Condensers
VI-3 Cost Effectiveness vs Plant Capacity for Surface-Type Condensers
II-5
III-3
III-5
VI-6
VI-7
VI-8
-------
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 10~8
9.480 X 10~4
(°C X 9/5) + 32
3.28
3.531 X 101
6.290
2.643 X 102
1.585 X 104
1.340 X 10"3
3.937 X 101
1.450 X 10"4
2.205
2.778 X 10"4
Prefix
T
G
M
k
m
M
Symbol
tera
giga
mega
kilo
milli
micro
Multiplication
Factor
1012
109
106
103
10"3
- 10"6
Example
1
1
1
1
1
1
Tg
Gg
Hg
km
mV
= 1
= 1
= ]_
= 1
= 1
= 1
X
X
X
X
X
X
10 12 grams
109 grams
10 6 grams
103 meters
10"3 volt
10 6 gram
-------
II-l
II. INDUSTRY DESCRIPTION
A. REASON FOR SELECTION
Ethylene glycol was selected for study because preliminary estimates indicated
that total emissions of volatile organic compounds (VOC) from its manufacture
were high and because an increase in consumption was expected to continue.
The manufacture of (mono)-ethylene glycol (EG) from ethylene oxide results in
production of diethylene glycol (DEC) and triethylene glycol (TEG) as principal
co-products. These compounds are practically odorless, stable, colorless liquids,
having greater densities and viscosities and higher boiling points than water.1
Some physical properties of the ethylene glycols are given in Appendix A.
B. ETHYLENE GLYCOLS USAGE AND GROWTH
The 1980 production capacity of ethylene glycol in the United States was
2442 Gg/yr.2'3 The 1979 production was 2066 Gg, or 85% of this capacity.4 The
estimated production capacity of diethylene glycol is 238.6 Gg/yr,5 with the
1978 production being 168.8 Gg, or 71% of this capacity.4 The triethylene glycol
production capacity is estimated to be 82.6 Gg/yr,6 with the 1978 production
being 54.4 Gg, or 66% of this capacity.4 Consumption during 1979—1984 is
expected to increase at an average annual rate of 4% for ethylene glycol,2 4%
for diethylene glycol,5 and 2% for triethylene glycol.6 If announced new plant
constructions remain on schedule and no shortage of ethylene develops, the supply
will be ample to meet projected demands through 1984.2
The uses of ethylene glycols and their expected growth rates are given in
Table II-l.2'5'6 Ethylene glycol, the most important of the glycols, was first
commercially manufactured in 1925. Today its consumption rate makes it one of
the more important of the synthetic organic chemicals. The largest use (approxi-
mately 43%) of ethylene glycol is for permanent-type antifreeze for liquid-cooled
motor vehicles.
High-purity ethylene glycol is used to manufacture polyethylene terephthalate
ester fibers and films. Approximately 40% of the ethylene glycol produced is
used in polyester fiber production, and about 6% is used in polyester films and
resins, the fastest growing use.2
-------
II-2
Table ll-l. Usage and Growth of Ethylene Glycols
1979 Use
End Use
Average Growth
for 1979—1934
(%/yr)
Ethylone Glycol
Antifreeze
Polyester fibers
Polyester film
Polyester bottles
Otherb
Total
Diethylene Glycol
Unsaturated polyester resins,
polyester polyols for polyurethanes
Triethylene glycol
Morpholine
Natural-gas dehydration
Textile agents
Udex extraction solvent
Dioxane
Plasticizers and surfactants
Exports
Otherd
Total
Triethylene Glycol
Natural gas dehydration
Vinyl plasticizer
Solvent
Humectant
Unsaturated polyester resins,
polyester polyols for polyurethanes
Exports
Other
Total
43
40
3.5
2.5
11
100
35
13
8
7
6
&
6
19
100
34
16
15
14
6
15
100
2
4
9.4
16
4.5
3.8
Sea ref 2.
Other uses include asphalt-emulsion paints, heat-transfer agents, low-ores-
sura laminates, brake fluids, glycol diacetate, low-freezing dynanite,
solvents, extractants for various purposes, solvent mixutre for cellulose
esters and ethers, cosmetics, lacquers, alkyd resins, printing inks, wood
stains, adheaives, leather dyeing, textile processing, tobacco, and deicing
fluid (see ref 2) .
CSee ref 5.
Other uses include blending into antifreeze, plasticizer for cork adhesives,
coupling agent for cosmetics and soaps, and as a humectant for tobacco
(see ref 5) .
See ref 6.
-------
II-3
Diethylene glycol is usually manufactured as a co-product of ethylene glycol
production. Most diethylene glycol (35%) is used to manufacture polyurethane
and unsaturated polyester resins. This market is expected to grow 8 to 9%
annually.5
Triethylene glycol, also a co-product of ethylene glycol production, is used
principally as a natural-gas dehydrant, which consumes 34% of the triethylene
glycol manufactured. If natural-gas price ceilings are lifted and gas produc-
tion increases, triethylene glycol consumption would be expected to increase.
No new uses for triethylene glycol are foreseen.6
C. DOMESTIC PRODUCERS
Eleven producers were operating 14 ethylene glycol plants at the end of 1979.2
Chemical Exchange Co. and Dixie Chemical Co. reportedly recover diethylene glycol
and triethylene glycol from purchased glycol bottoms.5'6 Table II-2 lists the
producers, locations, and capacities for each of the principal glycol compounds
produced. Figure II-l shows the plant locations.
Ethylene glycol is manufactured principally by the noncatalytic hydration of
ethylene oxide. Diethylene glycol and triethylene glycol are co-products of
this process. All ethylene glycol manufacturers using the process also produce
their own ethylene oxide feedstock.1 Diethylene glycol may be produced by the
reaction of ethylene glycol and ethylene oxide, but this is not done to any
large extent. Approximately 15% of the triethylene consumed is produced by
reacting diethylene glycol with ethylene oxide.7
In 1978 Oxirane began production of ethylene glycol with an acetoxylation process
developed by Halcon International, Inc. With this process ethylene glycol is
produced directly from ethylene with acetic acid in the presence of a catalyst
to form mono- and diacetates. These compounds are further hydrolyzed to ethylene
glycol. The plant operated intermittently during 1978 and 1979 and was shut
down in November 1979 because of severe corrosion problems. Oxirane reportedly
is exploring alternative uses for the plant but has given no time table for
this evaluation.2'8
-------
11-4
Table II-2. Ethylene Glycols Capacity as of 1980'
Production Capacity (Gg/yr)
Company
Location
Ethylene
Glycol.
Diethylene
Glycoic
Triethlene
Glycol
a
BASF Wyandotte
Calcasieu Chemical
Celanese Chemical
Q
Chemical Exchange
Q
Dixie Chemical
Dow Chemical U.S.A.
Eastman Kodak
Northern Natural Gas
Olin
PPG Industries
Shell Chemical
Texaco
g
Union Carbide
Total
Ge ismar, LA
Lake Charles, LA
Clear Lake City, TX
Houston, TX
Bayport, TX
Freeport, TX
Plaquemine, LA
Longview, TX
Morris, IL
Br andenbur g, KY
Beaumont, TX
Geismar, LA
Port Neches, TX
Seadrift, TX
Taft, LA
Penneulas, PR
113
82
227
116
159
82
91
.18
82
154
150
329
567
272
24421
i
15.9
20.4
NA
NA
34.0
8.2
6.8
NA
8.2
11.3
36.3
97.5
238.6
4.5
NA
NA
22.7
0.5
2.3
0.5
11.3
6.8
34.0
82.6
See refs 2, 5, and 6.
Reported plant capacities vary from one reference to another. The total EG capacity
reported by Chemical Marketing Reporter is 2982.4 Gg/yr,- see ref 4.
-i
"Some diethylene glycol capacities are based on 10% of ethylene glycol capacity and
represent only the capability to produce.
Capacities to produce triethylene glycol are flexible.
a
"DEC and TEG are obtained by distilling glycol still bottoms purchased from other producers;
see refs 5 and 6.
PPG closed their Guayanilla, PR, plant in 1978 due to lack of ethylene; see ref 2.
Union Carbide has announced plans to construct a 408-Gg/yr ethylene glycol facility at a
site to be named later; see ref 2.
ICI Americas, Inc. is building an ethylene oxide facility at Bayport, TX. Dow Chemical
has announced another ethylene oxide facility with 1983 startup (no location specified).
Both will produce some ethylene glycol; see ref 2.
-------
II-5
1. BASF Wyandotte, Geismar, LA 9.
2. Calcasieu, Lake Charles, LA 10.
3. Celanese, Clear Lake City, TX 11.
4. Chemical Exchange, Houston, TX 12.
5. Dixie Chemical, Bayport, TX 13.
6. Dow Chemical, Freeport, TX 14.
7. Dow Chemical, Plaquemine, LA 15.
8. Eastman Kodak, Longview, TX 16.
Northern Natural Gas, Morris, IL
01in, Brandenburg, KY
PPG, Beaumont, TX
Shell, Geismar, LA
Texaco, Port Neches, TX
Union Carbide, Seadrift, TX
Union Carbide, Taft, LA
Union Carbide, Penuelas, PR
Fig. II—1. Locations of Plants Manufacturing Ethylene Glycols
-------
II-6
D. REFERENCES*
1. H. W. Scheeline, Ethylene Glycols, Glycol Ethers and Ethanolamines, pp. 35—40
in Report No. 70, A private report by the Process Economics Program, Stanford
Research Institute, Menlo Park, CA (August 1971).
2. R. T. Gerry, "Ethylene Glycol," pp 652.5030A—652.5030R in Chemical Economics
Handbook, Stanford Research Institute, Menlo Park, CA (April 1980).
3. "Chemical Profile on Ethylene Glycol," p. 9 in Chemical Marketing Reporter
(July 24, 1978).
4. "CEH Manual of Current Indicators—Supplemental Data," p. 256 in Chemical
Economics Handbook, Stanford Research Institute, Menlo Park, CA (August 1980).
5. "Chemical Profile on Diethylene Glycol," p. 9 in Chemical Marketing Reporter
(Nov. 12, 1979).
6. "Chemical Profile on Triethylene Glycol," p. 9 in Chemical Marketing Reporter
(Aug. 13, 1979).
7. T. F. Killilea, "Di- and Triethylene Glycols Salient Statistics," pp. 652.5130A—
652.5130F in Chemical Economics Handbook, Stanford Research Institute, Menlo Park,
CA (June 1978).
8. "Process Failure Squeezes Glycol Market," Chemical and Engineering News
57(49), 6 (Dec. 3, 1979).
*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
III. PROCESS DESCRIPTION
A. INTRODUCTION
Ethylene glycol, CH2OHCH2OH, is manufactured on a very large scale throughout
the world by the addition of water to ethylene oxide (EO). The ethylene glycol
(EG) formed will react with additional ethylene oxide to form diethylene glycol
(DEC), triethylene glycol (TEG), and other higher homologs. The chemical equations
are as follows:
CH
CH2
(ethylene oxide)
HoO
(water)
CH2OH
CH2OH
(ethylene glycol)
CH2OH
CH2OH
(ethylene
glycol)
CHo
CH2
(ethylene
oxide)
CH2OH
CH2-0-CH2CH2OH
(diethylene glycol)
CH2OH
CH2-0-CH2CH2OH
(diethylene glycol)
CH2
CH2
(ethylene oxide)
CH2-0-CH2CH2OH
CH2-0-CH2CH2OH
(triethylene glycol)
The normal weight ratios of co-products formed are 87 to 88.5 wt % ethylene
glycol, 9.3 to 10.5 wt % diethylene glycol, and 2.2 to 2.5 wt % triethylene
glycol. These three products constitute an overall yield of 92.5 to 95.5% of
theoretical, based on the ethylene oxide feed.1
In the United States the principal method of manufacture of ethylene glycol is
by noncatalyzed pressure hydration of ethylene oxide.2 In this process a resi-
dence time of 1 hr at 200°C and a pressure of 1380 kPa is common. Present-day
practice is to use the noncatalyzed pressure hydration process because the alter-
native acid catalyst process results in problems with acid residue in the product.1
In the acid hydration process ethylene oxide is converted to ethylene glycol by
contact with a 0.5 to to 1.0% sulfuric acid catalyst solution at 50 to 70°C for
30 min.3
-------
III-2
In both processes diethylene and triethylene glycols are formed as co-products.
The greater the ratio of water to ethylene oxide in the feed, the greater will
be the proportion of (mono)-ethylene glycol in the reactor product. The water:oxide
weight ratios of the order of 8:1 are used. Ratios of the products formed to a
limited extent can be varied to meet market demand by varying the feed ratio
and/or other process variables.l
The Oxirane Corp. plant at Channelview, XX, which began production in 1978
but was shut down in 1979, produced ethylene glycol by the acetoxylation process.
All major production facilities use the conventional ethylene oxide hydration
process.
B. MODEL PROCESS ETHYLENE GLYCOL FROM ETHYLENE OXIDE
1. Process Description
The process flow diagram shown in Fig. III-l represents a typical noncatalyzed
ethylene oxide hydration process. The continuous process is carried out in the
liquid phase, and the reactions are strongly exothermic. Theoretically, 0.71 kg
of EO is required to produce 1 kg of EG; 0.83 kg of EO is required to produce
1 kg of DEG; and 0.88 kg of EO is required to produce 1 kg of TEG.1 The model*
process produces an overall product and co-product yield of 94.8% of theoretical.
Refined liquid EO (stream 1), makeup water (stream 2), and recycle water are
mixed under pressure (1380 kPa), preheated, and fed to the hydrolyzer. The
feed solution (stream 3) contains approximately 8 kg of water per kg of EO.
The reactor effluent, heated by the exothermic heat of hydration, exits (stream 4)
the hydrolizer at 200°C and enters a multiple-effect evaporator system for removal
of water.
A portion of the vapor from the first evaporator effect is purged (stream 5) to
remove light impurities from the system. The remainder of the vapor and the
vapors from the remaining evaporator effects are condensed and recycled (stream 6).
The evaporator calandria and the condenser on the final evaporator effect are
vented (A) to remove noncondensable gases.
*See p 1-2 for a discussion of model plants.
-------
MULTIPLE. - E.FFE.CT E.VA. PO RA To «.
V^
/-
Tl!
1 &.0*i-
* n.stV . I W t
I
u:
DOWM
Fig. III-l. Flow Diagram for Uncontrolled Model Plant Producing Ethylene
Glycol by Noncatalyzed Hydration of Ethylene Oxide
-------
III-4
The concentrated glycol solution then enters the water removal column for final
drying. The remaining water is vacuum distilled overhead, condensed, and recycled
(stream 7). The glycol mixture (stream 8) from the bottom of the column is
passed to the refining section. Vapors from the vacuum producer are vented
(B).
Ethylene glycol (stream 9), diethylene glycol (stream 10), and triethylene glycol
(stream 11) are distilled overhead in separate vacuum distillation columns.
Steam-jet ejectors used to maintain the vacuum on each distillation column are
vented (C). The residual heavy ends discharged from the bottom of the final
distillation column (stream 12) are stored for disposal or for sale as by-products.
A one month's storage capacity for each product is provided in conventional
cone-roof tanks. The tanks are padded with nitrogen to prevent absorption of
atmospheric moisture and are heated in the winter to prevent excessive viscosity.
2. Process Variations
There are existing ethylene glycol plants in which crude EO vapor from the EO
plant desorber4 is fed directly to the EG plant5 as shown in Fig. III-2. The
crude EO vapor (stream 1) is reabsorbed into water (stream 2) by an absorber
that is part of the glycol unit. The unabsorbed vent gases (stream 3) accompanying
the crude EO are vented. The EO solution (stream 4) then enters the hydrolizer.
The effluent (stream 5) from the hydrolizer is passed to a stripper, where the
remaining gases and light hydrocarbons (stream 6) are separated and vented.
The degassed glycol solution (stream 7) then enters the evaporator system.
Thus the emissions normally associated with EO refining operations are carried
over to the glycol plant, where they are discharged.5 The combined emissions
in both cases are essentially the same. The heavy ends (largely glycols) normally
separated in the EO refining operations are likewise carried over to the glycol
plant and ultimately end up in the product or are discharged with the heavy
ends from the glycol refining operation. Thus secondary emissions associated
with glycol production may be increased, while secondary emissions from EO pro-
duction may decrease.
Large amounts of high-temperature steam are required for removing the excess
water required by the hydrolysis reaction in production of glycols. Common
-------
i
VE.WT
T
EVAPC3RATOR.
GLYCOU
COtOVEUTlOUAL PROCESS : RE.FIUED QX.IDC FED T° 5LYCOL UM>T
I
CRUDE EO
(VAPOR)
OXlDt UKIIT -4-
^
-*» €tuvcou UKJVT
CRUQ6 oxAoe.
TO
EVAPORATOR.
6LXCOI-
Fig. III-2. Process Variations for Producing Ethylene Glycol by
Noncatalyzed Hydration of Ethylene Oxide
-------
III-6
practice is to integrate the oxide unit with the glycol unit on an energy basis.
The excess high-pressure steam produced in the oxide unit is often consumed by
the glycol unit, and low-pressure steam from the glycol unit is returned for
consumption in the oxide unit. Similarly, the heat from high-temperature sources
in the glycol unit is often recovered by the feed water to the oxide reactor
steam generators being preheated.5 Schemes for energy utilization and conservation
may vary extensively from one plant to the next. New plants in the future might
be sized and designed for maximum energy utilization, with the glycol unit sized
for the amount of energy available from the oxide unit and with the operating
temperatures and pressures for the evaporators and distillation columns selected
for optimum energy availability.1'5
Many variations in the design and operation of the water removal section of the
plant exist between existing plants. Also, the number of product distillation
steps or columns used by different plants may vary. Some plants do not recover
diethylene glycol or heavier glycols but instead sell the still bottoms to inde-
pendent producers for recovery of the heavier glycol by-products.
C. OTHER PROCESSES
1. Ethylene Glycol Directly from Ethylene
Halcon International developed a new acetoxylation process for making ethylene
glycol directly from ethylene. Ethylene is reacted with acetic acid in the
presence of a catalyst to form mono- and diacetates. These products are further
oxidized to ethylene glycol. The Oxirane Corp. plant at Channelview, TX, which
went on-stream June 16, 1978, but was shut down in November 1979, is the first
plant built based on this technology. Specific process and emissions data are
not available. The future use of the plant is uncertain.6'7
2. Carbonation of Ethylene Oxide
Ethylene oxide, carbon dioxide, and water (with a sodium bromide—sodium bicar-
bonate catalyst) are fed to a carbonation reactor. The ethylene carbonate that
is formed is then hydrolyzed to glycols in the same reactor system. The amount
of water required by this process is much lower than that required by the conven-
tional process. The process yields 98% ethylene glycol. The overall production
cost is estimated to be slightly lower than that for conventional hydration since
-------
III-7
the cost of removing water is considerably less. However, the corrosiveness of
the catalyst solution requires that special materials of construction be used
for the plant. The process has not been commercialized.1
3. Ethylene Glycol from Formaldehyde, Carbon Monoxide, and Water
From 1940 to 1969 Du Pont produced ethylene glycol by the reaction of formal-
dehyde, carbon monoxide, and water, followed by hydrogen reduction of the inter-
mediate glycolic acid to obtain the ethylene glycol. Du Pont shut down the plant
because of pollution problems. PPG is reportedly studying a modification of the
process and Chevron Research recently applied for patents using a hydrogen fluoride
catalyst rather than the sulfuric acid catalyst used by Du Pont.1'8'9
4. Ethylene Glycol from Hydrogenation and Hydrogenolysis of Carbohydrate
In the early 1970s3 ICI United States, Inc., produced ethylene glycol by fermenta-
tion of molasses. The plant is now on standby.8
5. Ethylene Glycol from Synthesis Gas
Union Carbide Corporation is reportedly developing a process for producing ethylene
glycol from synthesis gas.8
-------
III-8
D. REFERENCES*
1. H. W. Scheeline, Ethylene Glycols, Glycol Ethers, and Ethanolamines, pp. 10—49
in Report 70, A private report by the Process Economics Program, Stanford Research
Institute, Menlo Park, CA (August 1971).
2. P. H. Miller, "Glycols," p. 642 in Kirk-Othmer Encyclopedia of Chemical Technology,
2d ed., vol. 10, edited by A. Standen et al., Interscience, New York, 1966.
3. F. A. Lowenheim and M. K. Moran, "Glycols," pp. 397—402 in Faith, Keys, and
Clark's Industrial Chemicals, 4th ed., Wiley-Interscience, New York, 1975.
4. See Fig. III-l in J. F. Lawson and V. Kalcevic, IT Enviroscience, Inc., Ethylene
Oxide (in preparation for EPA, ESED, Research Triangle Park, NC).
5. J. F. Lawson, IT Enviroscience, Inc., Trip Report for Visit to Union Carbide
Corporation, South Charleston, WV, Dec. 7, 1977 (on file at EPA, ESED,
Research Triangle Park, NC).
6. "Oxirane Begins EG Shipments from New Channelview Facility," Chemical Marketing
Reporter 214(1), 14 (July 3, 1978).
7. "Process Failure Squeezes Glycol Market," Chemical and Engineering News,
5J7(49), 6 (Dec. 3, 1979).
8. G. E. Weismantel, "New Technology Sparks Ethylene Glycol Debate," Chemical
Engineering 86(2), 67—70 (Jan. 15, 1979).
9. T. F. Killilea, "Ethylene Glycol Salient Statistics," p. 652.5030C in Chemical
Economics Handbook, Stanford Research Institute, Menlo Park, CA (Oct. 1976).
*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,
participate in photochemical reactions producing ozone. A relatively small
number of organic chemical have low or negligible photochemical reactivity.
However, many of these organic chemicals are of concern and may be subject
to regulation by EPA under Section 111 or 112 of the Clean Air Act since there
are associated health or welfare impacts other than those related to ozone
formation.
A. MODEL PLANT*
The model plant for this study has a total glycol (EG, DEG, and TEG) capacity
of 170 Gg/yr based on 8760 hr of operation annually.** The capacity of existing
production units based on the EO hydration process varies from 15 to about 325 Gg/yr.
The recent trend in the industry has been to construct large-capacity units or
to expand the capacity of existing units.
The model plant utilizes the model noncatalyzed ethylene oxide hydration process
described in Sect. III-B. Although not an actual operating plant, it is typical
of existing plants utilizing the noncatalyzed ethylene oxide hydration process.
Storage tanks for the model plant were sized to provide 1 month of storage capacity
for each product. Characteristics of the model plant that are important in
air-dispersion modeling are given in Appendix B.
B. SOURCES AND EMISSIONS
The process emissions estimated for the ethylene glycol model plant are
based on information given in a trip report of a visit to Union Carbide1 and
*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 calculatons, the error
introduced by assuming continuous operation is negligible.
-------
IV-2
in responses to EPA's requests for information from selected companies (see
Appendix C), together with data from a report published by Stanford Research
Institute2 and an understanding of the process chemistry and yields. The storage
and handling emissions were calculated based on physical properties. The fugitive
emissions due to leaks are based on the data referenced in Appendix D. Uncontrolled
air emission data reported by individual plants producing ethylene glycol vary
widely from plant to plant. These differences appear to be related largely to
differences in plant design and operation and in how the reported air emissions
were determined.
The glycols are water soluble, have low vapor pressures, and boil at higher
temperatures than water. Thus glycol emissions tend to be small. The principal
volatile impurities in the reactor product are ethylene oxide and acetaldehyde.
Both these compounds are infinitely soluble in water and tend to reabsorb in
the condensate from the evaporators or distillation column vents. The quantity
of air emissions is related to the manner in which these streams are handled.
Those plants that vent the evaporator purge stream and the distillation column
steam-jet ejectors directly to atmosphere as vapor have large emissions. If
barometric-type condensers are used to condense and absorb these discharges,
the process emissions are reduced, but the fugitive emissions from the cooling
tower and the secondary emissions from treatment of cooling tower blow-down
water may become significant. If surface-type condensers are used and the conden-
sate is isolated from the cooling water, the cooling tower emissions are eliminated.
However, the secondary emissions from treatment and disposal of the condensate
are increased.
The model-plant uncontrolled emission rates given in Table IV-1 were calculated
based on the model-plant characteristics and operating and emission data from
existing plants. The model-plant overall product and by-product yield is 94.8%,
with an EO conversion factor of 99%. Heavies discharged from the bottom of the
triethylene glycol column account for 2.5% of the losses. Air emission and the
purge stream and cooling tower blow-down losses make up the remaining 2.7%.
1. Process Emissions
Uncontrolled process emissions from the model plant originate from the evaporator
first-effect purge stream (stream 5, Fig. III-l), the evaporator calandria vents
-------
IV-3
Table IV-1. Total VOC from Uncontrolled Emissions from
Production of Ethylene Glycol in a Model Plant
Emission Source
c
Process emissions
d
Storage emissions
Handling emissions
e
Fugitive emissions
Secondary emissions
Total
Designation
(Fig.III-1)
5,A,B,C
D
F
H
K
Emissions
Ratio3
(g/kg)
0.0595
0.0028
0.0010
4.7407
0.3542
5.1582
Rateb
(kg/hr)
1.16
0.05
0.02
92.05
6.88
100.16
ag of emission per kg of total products produced.
For the 170-Gg/yr model plant based on an average glycol production rate of
19,417 kg/hr.
°Due to direct emissions from evaporator calandria vent (A). Emissions from other
process vents (vent B, water removal column vacuum ejector; vent C, glycol
purification column vacuum ejectors; and stream 5, evaporator purge) are absorbed
by the cooling water through use of barometric-type condensers and thus contribute
to fugitive and secondary emissions.
Principally from storage of ethylene glycol; emissions from storage of heavier
glycols are negligible.
6Estiraated sources of fugitive emissions are cooling water (4.6705 g/kg) and
leakage from pumps and valves (0.0702 g/kg); contributors to cooling water
fugitive emissions are stream 5 (4.6653 g/kg), vent C (0.0052 g/kg), and
vent B (negligible).
^Secondary emissions result from treatment of cooling water blowdown; estimated
contributors to cooling water secondary emissions are stream 5 (0.3534 g/kg),
vent C (0.0007 g/kg), and vent B (negligible).
-------
IV-4
(A), the discharge from the water removal column steam-jet ejector (vent B),
and the discharges from the distillation column ejectors (vents C).
The uncontrolled model plant incorporates barometric-type contact condensers to
condense and absorb the vapor from the evaporator purge (stream 5) and the dis-
charge from the various steam-jet ejectors (vents B and C). The emissions from
these sources then circulate with the cooling water. Partial desorption occurs
as the cooling water passes through the cooling water circuit and cooling tower.
The remainder of the contaminants end up in the cooling tower blow-down stream.
The evaporator calandria emissions for the uncontrolled model plant are vented
directly to the atmosphere.
2. Storage Emissions
Due to their hygroscopic properties glycols are normally stored in tanks blanketed
with nitrogen. The tanks are heated to prevent excessive viscosity in cold
weather. Breathing losses are negligible because the tank temperature is con-
trolled. Because of the low vapor pressure of glycols the emissions due to
working losses are small for ethylene glycol and are negligible for the heavier
glycols when calculated by the emission equations from AP-423 (see Table IV-1).
3. Handling Emissions
Emissions from loading of shipping vessels were calculated with the equations
from AP-42.4 Because of the low vapor pressure of the glycols the handling
emissions are small (see Table IV-1).
4. Fugitive Emissions
Process pumps and valves that handle organic compounds under pressure are poten-
tial sources of fugitive emissions. The model plant is estimated to have 7 pumps,
38 process valves, and 3 relief valves handling light organics in the feed and
water removal sections and 6 pumps and 15 valves in sections handling heavy
organics under positive pressure. The fugitive-emission factors from Appendix D
were applied to determine the fugitive emissions shown in Table IV-1.
The largest source of fugitive emissions is from the cooling tower. Partial
desorption of organics contained in the cooling water occurs as the water passes
through the cooling tower. The cooling water system emissions given in Table IV-1
are based on average desorption data reported by plants using barometric condensers.5'
-------
IV-5
5. Secondary Emissions
Secondary VOC emissions can result from the handling and disposal of process
waste liquid streams. The potential sources (source K) that exist for the model
plant are the blow-down water from the cooling tower and from disposal of the
heavy ends (stream 12). Due to the low volatility of the heavy ends stream the
secondary emissions calculated for this stream were negligible. Secondary emissions
from treatment of the cooling tower blow-down water are shown in Table IV-1.
The calculations were based on wastewater treatment by a primary clarifier followed
by an activated-sludge system and were done by the methods described in another EPA
report on secondary emissions.7
-------
IV-6
C. REFERENCES*
1. J. F. Lawson, IT Enviroscience, Inc., Trip Report for Visit to Union Carbide
Corp._, South Charleston, WV, Dec. 7,1977 (on file at EPA, ESED, Research
Triangle Park, NC).
2. H. W. scheeling, Ethylene Glycols, Glycol Ethers and Ethanolamines, Report
No. 70, A private report by the Process Economics Program, Stanford Research
Institute, Menlo Park, CA (August 1971).
3. C. C. Masser, "Storage of Petroleum Liquids," Sect. 4.3 in Compilation of Air
Pollutant Emission Factors, AP-42, Part A, 3d ed., EPA, Research Triangle Park,
NC (August 1977).
4. C. C. Masser, "Transportation and Marketing of Petroleum Liquids," pp 4.5-5 to
4.4-6 in Compilation of Air Pollutant Emission Factors, AP-42, Part A, 3d ed.,
EPA, Research Triangle Park, NC (August 1977).
5. Shell Oil Co., letter dated Jan. 11, 1979, in response to EPA's request for
information on emissions data on ethylene glycol production facilities.
6. BASF Wyandotte Corp., letter dated Nov. 27, 1978, in response to EPA's request
for information on emissions data on ethylene glycol production facilities.
7. J. J. Cudahy and R. L. Standifer, IT Enviroscience, Inc., Secondary Emissions
(September 1980) (EPA/ESED report, 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.
-------
V-l
V. APPLICABLE CONTROL SYSTEMS
A. TYPES OF CONTROLLED EMISSIONS
1. Process and Process-Related Emissions
Process emissions from the uncontrolled model plant are the result of direct
atmospheric discharge from the vents (A) on the multiple-effect evaporator and
from desorption of organics contained in the cooling water. Contamination of
the cooling water results from its direct contact with the discharges from the
steam-jet ejectors associated with the distillation column vents (B and C) and
with the discharge from the evaporator purge stream (5) through use of barometric-
type contact condensers on these streams.
To prevent the cooling water from becoming contaminated, the controlled model
plant uses surface-type condensers to replace the barometric-type condensers.
The condensate from the condenser is discharged to wastewater treatment. Uncon-
densed gases are vented to the atmosphere. A surface-type condenser is also
used to control emissions from the evaporator vents (A) in the controlled model
plant.
The installation of surface-type condensers to isolate the condensate from the
cooling water eliminates fugitive emissions from the cooling tower. However,
the condensate added to the wastewater increases secondary emissions. Also,
the uncondensed gases vented from the surface condensers contain some VOC. The
net reduction in emissions originating from all process sources (vents A, B,
and C and stream 5) by application of surface condensers is 85%,* as indicated
in Table V-l. This includes the overall change in direct process emissions
plus the change in fugitive and secondary emissions related to process sources.
2. Storage and Handling Emissions
Emissions due to storage and handling of glycols remain uncontrolled in the
model plant. Emissions from these sources are slight due to the low vapor pressure
*This number was calculated by subtracting the controlled emissions originating
from all process sources (0.0062 + 0.7720 = 0.7782 g/kg) from the uncontrolled
emissions originating with streams 5, A, B, and C (0.0595 + 4.6705 + 0.3534 +
0.0007 + neg = 5.0842 g/kg) (see Table IV-1) to determine the net reduction in
VOC emissions (5.0842 - 0.7782 = 4.306 g/kg) due to use of surface condensers
instead of barometric condensers. The value 4.306 g/kg is 85% of 5.0842 g/kg.
-------
V-2
Table V-l. Total Controlled VOC Emissions for Model-Plant Ethylene Glycol Production
Source
c
Process emissions
Storage emissions
Handling emissions
. . d
Fugitive emissions
e
Secondary emissions
Total
Stream
Designation
(Fig.III-1)
5,A,B,C
D
F
H
K
Emission
Control Device Reduction
or Technique (%)
Surface Condensers 85
None
None
Inspection and Maintain 66
None
84 (Av)
Emissions
a
Ratio
(g/kg)
0.0062
0.0028
0.0010
0.0240
0.7720
0.8060
Rateb
(kg/hr)
0.12
0.05
0.02
0.47
14.99
15.65
g of emission per kg of total products produced.
For the 170-Gg/yr model plant based on an average glycol production rate of 19,417 kg/hr.
°The net reduction in emissions originating from all process sources (vents A, B, and C and
stream 5) is 85%; this includes overall change in direct process emissions plus change
in fugitive and secondary emission related to process sources.
The reduction in fugitive emissions applies to leakage from pumps and valves, which are
reduced 66% by inspection and maintenance. Installation of surface condensers
eliminated cooling tower fugitive emissions.
6Secondary emissions have increased due to condensate from vent condensers being added
to wastewater. '
-------
V-3
of the product stored and the nitrogen-blanketed controlled-temperature
storage techniques employed. Control options for storage and handling emissions
are discussed in another EPA document.1
Fugitive Emissions
The principal fugitive emissions from the uncontrolled model plant are from
desorption of volatile organics contained in the cooling water as it passes
through the cooling tower. The source of cooling water contamination was elimi-
nated in the controlled model plant by replacement of the contact condensers
used in the uncontrolled model plant with surface-type condensers.
The remaining fugitive emissions result from leaks from pumps and valves even
though much of the equipment is operated under vacuum. Emissions from pumps
and valves can be controlled by an appropriate leak-detection system and repair
and maintenance program. Controlled fugitive emissions calculated with the
factors given in Appendix D are included in Table V-l; these factors are based
on the assumption that major leaks are detected and corrected. Control measures
for control of fugitive emissions are discussed in another EPA report.2
Secondary Emissions
The principal secondary emissions from the controlled model plant result from
desorption of volatile organic compounds contained in the condensate from the
vent condensers. The secondary emission data given in Table V-l were calculated
based on the characteristics and the estimated concentration of the volatile
components in the condensate. Treatment by a conventional clarifier and activated-
sludge system was assumed. No control system has been identified for the secondary
emissions from wastewater treatment. Secondary emissions and their applicable
controls for all the synthetic organic chemicals manufacturing industry are
discussed in another EPA report.3
-------
V-4
B. REFERENCES*
1. D. G. Erikson, IT Enviroscience, Inc., Storage and Handling (September 1980)
(EPA/ESED report, Research Triangle Park, NC).
2. D. G. Erikson and V. Kalcevic, IT Enviroscience, Inc., Fugitive Emissions
(September 1980) (EPA/ESED report, Research Triangle Park, NC).
3. J. J. Cudahy and R. L. Standifer, IT Enviroscience, Inc., Secondary Emissions
(September 1980) (EPA/ESED report. 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.
-------
VI-1
VI. IMPACT ANALYSIS
A. ENVIRONMENTAL AND ENERGY IMPACTS
1. Emission Reduction
An overall emission reduction of 84%, or 740.31 Mg/yr, is achieved (see Table VI-1)
by application of the control systems described in Sect. V to the model plant
described in Sects. Ill and IV.
The principal source of emissions from the uncontrolled model plant are fugitive
emissions from the cooling tower. Contamination of the cooling water results
from the use of contact condensers (barometric-type) on the process vents. The
source of VOC contamination in the cooling water is eliminated by replacing the
contact condensers with surface condensers. The condensate collected is dis-
charged to the wastewater treatment system.
The cooling water required for both types of condensers is essentially the same.
However, the exhaust from the evaporator calandria vents (A) is not condensed
in the uncontrolled model plant. It is estimated that the amount of additional
cooling water required for this condenser would be 152 liters/min.
It is possible that additional heat could be recovered in condensing the evaporator
purge stream or the vapor from evaporator calandria vents through use of addi-
tional heat exchangers. The low-pressure steam generated might be used for the
vacuum distillation operations or for operations in the EO plant. Because the
potential exists for some producers to have excess steam on-site, it may not be
practical for all producers to utilize heat recovery. No credit for heat recovery
is taken for the controlled model plant.
2. 1979 Industry Emissions
The total VOC emissions from the domestic ethylene glycol industry in 1979 are
estimated to be 6300 Mg, including the estimated emissions from process, fugitive,
secondary, storage, and handling sources! The estimate is based on the 1979
level of total glycol production of 2066 Gg by the hydration of ethylene
oxide process. To the extent available, actual emission data reported by the
-------
VI-2
Table VI-1. Environmental Impact of Controlled Ethylene Glycol Model Plant
Vent
Designation
Emission Source (Fig.III-1)
b
Process emissions 5,A,B,C
Storage emissions D
Handling emissions F
Fugitive emissions H
d
Secondary emissions K
Total
VCC Emission
Reduction3
Control Device or Technique (%)
Surface condensers
None
None
Installation of surface con-
densers and detection and
correction of major leaks
None
84 (Av)
. (Mg/yr)
9.11
0
0
802.24
-71.04
740.31
For the 170-Gg/yr model plant based on full-capacity operation.
Direct process emissions only are given here; secondary and fugitive emissions related
to process discharges are given under those categories.
'Fugitive emissions are reduced 794.44 Mg/yr by replacing barometric condensers on process
vents with surface condensers; correcting leaks in pumps and valves provides an additional
7.80-Mg/yr reduction.
Secondary emissions increase 217.9%, or by 71.04 Mg/yr, due to the VOC in the condensate
that is discharged to the wastewater treatment system.
-------
VI-3
individual plants were used to calculate the emissions (see Appendix C). For
those plants where emission data were not available or where gaps appeared in
reported data, the emissions were assumed to be the average of the reported
emission for each emission source category. Fugitive emissions due to leaks
were assumed to be controlled in 50% of the equipment.
B. CONTROL COST IMPACT
Estimated costs and cost-effectiveness data for control of VOC emissions result-
ing from the production of ethylene glycol are given in this section. Details
of the model plant (Fig. III-l) are given in Sects. Ill and IV. Cost estimate
sample calculations are included in Appendix E.
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 esti-
mates do not include the costs of production lost during installation or startup,
of research and development, or of land acquisition. If the control systems
were retrofitted in an existing plant, difficulty may be experienced in finding
space to accommodate the retrofitted control equipment in the existing plant
layout. Because of these associated costs the cost of retrofitting emission
control systems in existing plants may be appreciably greater than the cost for
a new installation.
Bases for the annual cost estimates for the control alternatives include utilities,
operating labor, maintenance supplies and labor, capital charges, and miscellaneous
recurring costs such as taxes, insurance, and administrative overhead. The
cost factors used are itemized in Table VI-2.
1. Process and Process-Related Emissions
Model-plant process emissions are controlled by surface-type condensers installed
on each process vent. Separate condensers are required on the following: the
evaporator purge (stream 5), the evaporator calandria vents (A), the water removal
column (vent B), the ethylene glycol column (vent C), the diethylene glycol
column (vent C), and the triethylene glycol column (vent C). The uncontrolled
model plant has barometric (contact) condensers on all vents except the evaporator
calandria vents (A). For a new plant the difference in installed costs for
-------
VI-4
Table VI-2. Annual Cost Parameters
Operating factor 8760 hr/yr
Operating labor $15/man-hr
Fixed costs
Maintenance labor plus
materials, 6%
b
Capital recovery, 18%
Taxes, insurances,
administration charges, 5%
29% of installed capital cost
Utilities
Cooling water $0.026/m3 ($0.10/1000 gal)
Wastewater treatment $0.07/m plus $0.22/kg BOD
(greater than 2 million gal/day) ($0.25/1000 gal plus $0.10/lb BOD)
SProcess downtime is normally expected to range from 5 to 15%. If the hourly
rate remains constant, the annual production and annual VOC emissions will be
correspondingly reduced. Control devices will usually operate on the same
cycle as the process. From the standpoint of cost-effectiveness calculations,
the error introduced by assuming continuous operation is negligible.
Based on 10-year life and 12% interest.
-------
VI-5
surface condensers versus barometric condensers is primarily the difference in
purchase costs of each type of condenser. Although the purchase cost of a surface
condenser will be more than that of a barometric condenser, the difference is
small in comparison to the total installed capital cost and is considered to be
negligible for the purposes of this report.
The total estimated cost for the surface condenser on the evaporator calandria
vents (A) is $50,000, which includes the cost of the equipment and of installing
the equipment, piping, and insulation (see Appendix E for sample calculations).
Figure VI-1 was plotted to show the variation of installed cost with plant capacity.
The condensers for the model plant are estimated to have a net annual cost of
$350,000, which includes capital recovery, miscellaneous capital, maintenance,
and utilities. The estimated variation of net annual cost with plant capacity
is shown by Fig. VI-2.
The cost effectiveness of installing and operating emission controls for the
model plant is $480 per Mg of VOC emissions removed. A plot of the estimated
cost effectiveness versus plant capacity is shown by Fig. VI-3.
2. Storage and Handling
Storage and handling controls are not included in the controlled model plant
since the rate of emissions from these sources is low. Control options for
storage and handling emissions are covered in a recent EPA document.1
3. Fugitive Emissions
The principal fugitive emission source was eliminated by preventing the process
cooling water from becoming contaminated through installation of surface condensers
on the process vents. These controls are described under "Process and Process-
Related Emissions."
Fugitive emissions due to leaks in pump and valve seals are controlled in the
model plant by a program of inspection and maintenance. A recent EPA document
describes fugitive emissions and their control measures.2
-------
VI-6
1OO
200
300
350
Plant Capacity (Gg/Yr)
Fig. VI-1. Installed Capital Cost vs Plant Capacity for
Emission Control by Installation of
Surface-Type Condensers
-------
VI-7
800
o 600
o
o» ^
o v*
o ^rC
•= N 400
3 o>
Z « 200
o
o
a
a
•a
o
2
100
Plant Capacity
200
(Gg/Yr)
300
350
Fig. VI-2. Net Annual Cost vs Plant Capacity for
Emission Control by Installation of
Surface-Type Condensers
-------
VI-8
600
o>
S
^»
tn
n
to
o
c
0)
u
o
%-
UJ
(O
o
o
500
400
100 200
Plant Capacity (Gg/Yr)
300
350
Fig. VI-3. Cost Effectiveness vs Plant Capacity for
Emission Control by Installation of
Surface-Type Condensers
-------
VI-9
4. Secondary Emissions
Control options for control of secondary emissions are covered in a recent EPA
document.3 No control system has been identified for the secondary emissions
from the model plant.
-------
VI-10
C. REFERENCES*
1. D. G. Erikson, IT Enviroscience, Inc., Storage and Handling (September 1980)
(EPA/ESED report, Research Triangle Park, NC).
2. D. G. Erikson and V. Kalcevic, IT Enviroscience, Inc., Fugitive Emissions
(September 1980) (EPA/ESED report, Research Triangle Park, NC).
3. J. J. Cudahy, IT Enviroscience, Inc., Secondary Emissions (September 1980)
EPA/ESED report, 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.
-------
VII-1
VII. SUMMARY
Ethylene glycol is manufactured principally by the noncatalytic hydration of
ethylene oxide,1 with diethylene glycol and triethylene glycol produced as
co-products. The domestic production capacity of ethylene glycol for 1980 was
2442 Gg,2 with an industry utilization of approximately 85% of this capacity.
The manufacture of antifreeze consumes about 43% of the ethylene glycol produced,
and 46% is used to manufacture polyester fibers and films. The estimated con-
sumption annual growth rate is 4%.3
Emission sources and uncontrolled and controlled emission rates for the ethylene
glycol model plant are given in Table VII-1. The major emission source from the
uncontrolled model plant is the fugitive emissions from the cooling tower. The
contamination in the cooling water results from use of contact condensers on the
process vents. The emissions are controlled by installing surface condensers,
with the condensate collected and discharged to a wastewater treatment plant.
The emissions from storage and handling of glycols are slight and controls are
normally not applied. Fugitive emissions due to leaks in pumps and valves are
controlled by a program of inspection and maintenance. Secondary emissions
become the major potential source of emissions after installation of surface
condensers to control emissions from process-related sources. Control of secondary
emissions is described in a recent EPA document.4
The total ethylene glycol industry VOC emissions are estimated to be 6300 Mg
in 1979, with most of the uncontrolled VOC emissions coming from fugitive and
secondary sources.
ip. H. Miller, "Glycols," p. 642 in Kirk-Othmer Encyclopedia of Chemical Technology,
2d ed., vol. 10, edited by A. Standen e_t al., Interscience, New York, 1966.
2R. T. Gerry, "Ethylene Glycol," pp 652.5030A—652.5030R in Chemical Economics
Handbook, Stanford Research Institute, Menlo Park, CA (April 1980).
3"CEH Manual of Current Indicators Supplemental Data," p. 256 in Chemical
Economics Handbook, Stanford Research Institute, Menlo Park, CA (August 1980).
4J. J. Cudahy and R. L. Standifer, IT Enviroscience, Inc., Secondary Emissions
(September 1980) (EPA/ESED report, Research Triangle Park, NC).
-------
VI I-2
Table VII-1. Emission Summary for Ethylene Glycol Model Plant'
Emission Source
Process emissions
Storage emissions
Handling emissions
d
Fugitive emissions
. . c ,e
Secondary emissions
Total
Stream
Designation
(Fig.III-1)
5,A,B,C
D
F
H
K
VOC Emission Rate (kg/hr)
Uncontrolled
1.-16
0.05
0.02
92.05
6.88
100.16
Controlled
0.12
0.05
0.02
0.47
14.99
15.65
For the 170-Gg/yr model plant based on an average glycol production
rate of 19,417 kg/hr.
b
Data apply to direct process emissions only.
r»
"Storage, handling, and secondary emissions remain uncontrolled in the
controlled model plant.
Principal source of fugitive emissions in the uncontrolled model
plant is the process cooling water; cooling water contamination is
eliminated by the contact condensers on the process vents being
replaced with surface-type condensers.
3
'Secondary emission increase in the controlled model plant due to
condensate from the vent condensers being added to the wastewater.
-------
A-l
APPENDIX A
Table A-l. Properties of (Mono)-Ethylene Glycol*
Synonyms
Molecular formula
Molecular weight
Physical state
Density
Vapor pressure
Boiling point
Water solubility
Glycol, ethandiol-1,2
C2H6°2
"62.1
Liquid
1.110 g/mlat 25°C
16 Pa at 25°C
197.3°C
Infinite
*From: Properties and Uses of Glycols, Dow Chemical USA, Midland,
MI, 1961.
-------
A-2
Table A-2. Properties of Diethylene Glycol*
Synonyms
Molecular formula
Molecular weight
Physical state
Density
Vapor pressure
Boiling point
Water solubility
Glycol ether
.C4H10°3
10G.1
Liquid
1.113 g/ml at 25°C
1.3 Pa at 25°C
244.8°C
Infinite
*From: Properties and Uses of Glycols, Dow Chemical USA, Midland,
MI, 1961.
-------
A-3
Table A-3. Properties of Triethylene Glycol*
Synonyms
Molecular formula
Molecular weight
Physical state
Density
Vapor pressure
Boiling point
Water solubility
C6H14°4
150.2
Liquid
1.119 g/ml at 25°C
<1.3 Pa at 2B°C
288°C
Infinite
*From: Properties and Uses of Glycols, Dow Chemical USA, Midland,
MI, 1961.
-------
B-l
APPENDIX B
Table B-l. Air-Dispersion Parameters for
Model Plant with a Capacity of 170
Source
Total VOC
Emissions
Ratea
(g/sec)
Vent Vent
Heightb Diameter13
(m) (m)
Gg/yr
Discharge
Tempera-
ture13
(K)
Flow Discharge
Rateb Velocity13
(m /sec) (m/sec)
Uncontrolled Emissions
Process emissions for
evaporator calandria
(3 vents)
Storage emissions
EG (4 tanks)
DEG (1 tank)
TEG (1 tank)
Heavy ends (1 tank)
Handling emissions
Fugitive emissions
Cooling tower
Leaks from pumps
and valves0
Secondary emissions from
cooling tower blowdown
0.321
0.015
Neg
Neg
Neg
0.005
25.192
0.378
1.91
12.2 0.05
12.2 0.2
12.2 0.2
7.3 0.2
7.3 0.2
18.6 5.5
373
313
313
313
313
336
305
0.033 17
220 9.1
Controlled Emissions
Process emissions
Evaporator purge
Evaporator calandria
Water removal column
EG column
DEG column
TEG column
Fugitive emissions from
leaks from pumps and valves'
Secondary emissions from
condensate discharge
e
0.033
Neg
Neg
Neg
Neg
0.130
4.164
12.2 0.08
18.3 0.08
15.2 0.08
13.7 0.08
13.7 0.08
327
327
327
327
327
30
0.002 0.4
STotal of all vents from source.
bAverage for each separate vent from source.
fugitive emissions from leaks are distributed over a 50-m by 150-m area.
dSurface of ground-level wastewater treatment system.
eVent from evaporator purge condenser tied to evaporator calandria vent condenser.
-------
C-l
APPENDIX C
LIST OF EPA INFORMATION SOURCES
Ableson, P. M., Calcasieu Chemical Corp., letter dated Dec. 20, 1978, in response
to EPA's request for information on emissions data on ethylene glycol production
facilities.
Dutcher, V.D., Union Carbide Corp., Texas Air Control Board 1975 Emission Inventory
Questionnaire for Union Carbide Corp, Seadrift, TX, Plant, Sept. 3, 1976.
Fritsch, J. J., Jr., Celanese Chemical Co., Texas Air Control Board 1975 Emissions
Inventory Questionnaire for Celanese Chemical Co., Clear Lake City, TX, plant,
May 19, 1976.
Kovacevich, T. R., BASF Wyandotte Corp, letter dated Nov. 27, 1978, in response
to EPA's request for information on emissions data on ethylene glycol production
facilities.
Lawson, J. F., IT Enviroscience, Inc., Trip Report for Visit to Celanese Chemical
Co., Clear Lake City. TX, June 21 and 22, 1977 (data on file at EPA, ESED,
Research Triangle Park, NC).
Lawson, J. F., IT Enviroscience, Inc., Trip Report for Visit to Union Carbide
Corp., South Charleston, WV, Dec. 7, 1977 (data on file at EPA, ESED, Research
Triangle Park, NC).
Louisiana Air Control Commission, Emission Inventory Questionnaire for Union
Carbide Corp., Taft, LA, plant, Mar. 6, 1975.
Louisiana Air Control Commission Permit No. 373 issued Nov. 8, 1974, to Union
Carbide Corp., Taft, LA, plant for ethylene oxide/glycol facility, unit 2.
Louisiana Air Control Commission Permit No. 476 issued July 9, 1975, to Union
Carbide Corp., Taft, LA, plant for modifications of ethylene oxide/glycol
facilities, unit 1.
-------
C-2
Mullins, J. A., Shell Oil Co., letter dated Jan. 11, 1979, in response
to EPA's request for information on emissions data on ethylene glycol production
facilities.
Rogers, P. F., Houston Chemical Co., Texas Air Control Board 1975 Emissions
Inventory Questionnaire for Houston Chemical Co., Beaumont, TX, plant. May 24,
1976.
Texas Air Control Board Permit No. 1329 issued 1973 to Texas Eastman Co.,
Longview, TX, for ethylene oxide—ethylene glycol plant.
Texas Air Control Board Permit No. C-3361 issued 1975 to Houston Chemical Co.,
Beaumont, TX, for ethylene oxide—glycol expansion.
Texas Air Control Board Permit No. 4273 issued 1976 to Dow Chemical USA,
Freeport, TX, for ethylene glycol facility.
Texas Air Control Board Permit No. 5032 issued 1977 to Union Carbide Corp.,
Texas City, TX, for ethylene oxide/ethylene glycol, unit 1.
-------
D-l
APPENDIX D
FUGITIVE-EMISSION FACTORS*
The Environmental Protection Agency recently completed an extensive testing
program that resulted in updated fugitive-emission factors for petroleum re-
fineries. Other preliminary test results suggest that fugitive emissions from
sources in chemical plants are comparable to fugitive emissions from correspond-
ing sources in petroleum refineries. Therefore the emission factors established
for refineries are used in this report to estimate fugitive emissions from
organic chemical manufacture. These factors are presented below.
Uncontrolled
Emission Factor
Controlled
Emission Factor'
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
(kg/hr)
0.12
0.02
0.021
0.010
0.0003
0.16
0.006
0.009
0.44
0.00026
0.032
(kg/hr)
0.03
0.02
0.002
0.003
0.0003
0.061
0.006
0.009
0.11
0.00026
0.019
aBased on monthly inspection of selected equipment; no inspection of
heavy-liquid equipment, flanges, or light-liquid relief valves;
10,000 ppmv VOC concentration at source defines a leak; and 15 days
allowed for correction of leaks.
Light liquid means any liquid more volatile than kerosene.
*Radian Corp., Emission Factors and Frequency of Leak Occurrence^for Fittings
in Refinery Process Units, EPA 600/2-79-044 (February 1979).
-------
E-l
APPENDIX E
COST ESTIMATE SAMPLE CALCULATIONS
This appendix contains the sample calculations for the estimated costs presented
in this report.
The accuracy of an estimate is a function of the degree of data available when
the estimate was made. Figure E-l illustrates this relationship. A contin-
gency allowance, as indicated on this chart, is included in the estimated costs
to cover the undefined scope of the project.
Capital costs given in this report are based on a screening study, as indicated
by Fig. E-l, based on general design criteria, block flowsheets, approximate
material balances, and data on general equipment requirements. These costs
have an accuracy range of +40% to -30%, depending on the reliability of the
data, and provide an acceptable basis to determine the most cost-effective
alternate within the limits of accuracy indicated.
In all capital calculations, allowances of 35% were added for magnitude, hazard,
and definition contingencies.
This example is based on the use of surface condensers on each process vent in
the model plant (see Sect. VI-B-1). The surface condenser for the evaporator
calandria vents (A) has 181 sq ft of heat-exchange surface area and requires
40 gpm of cooling water. The surface condensers for the other process vents
use the same amount of cooling water as the contact (barometric) condensers in
the uncontrolled model plant. The use of surface condensers increases the flow
to the waste-water treatment plant by 96 gpm and the BOD by 363 Ib/hr. The
increase in energy required by treating this additional waste is considered to
be negligible.
A. INSTALLED CAPITAL COST
Figure A-l of the control device evaluation report on condensation1 shows that
the installed capital cost of a carbon steel condenser having 181 sq ft of area
is $50,000. The installed capital cost of the other surface condensers is
almost the same as for contact condensers for the same service (see Sect. VI-B-1);
the difference is considered to be negligible. Therefore the total installed
-------
E-2
capital cost for the controlled model plant with surface condensers is $50,000
more than for the uncontrolled model plant with contact condensers on all but
one vent.
B. NET ANNUAL COST
From Table VI-2 of this report the total fixed cost, including capital recovery,
is 29% of the installed capital cost:
$50,000 X 0.29 = $14,500/yr.
From Table VI-2 the cost of cooling water is $0.10/1000 gal. The annual cost
of 40 gpm of cooling water is
40 X 60 X 8760 X 0.10 = $210Q/yr
From Table VI-2 the cost of wastewater treatment is $0.25/1000 gal plus
$0.10/lb of BOD. To treat the 96 gpm of condensate from all the surface condensers
and that contains 363 Ib/hr of BOD the annual cost is
96 X 60 X 8760 X 0.25
= $12,600/yr for flow
plus 363 X 8760 X 0.10 = $318,000/yr for BOD.
The total cost of wastewater treatment for the additional condensate is
$12,600 + $318,000 = $330,600/yr.
The annual cost summary is as follows:
Fixed $ 14,500
Cooling water 2,100
Wastewater treatment 330,600
Total $350,000 (rounded)
1D. G. Erikson, IT Enviroscience, Inc., Control Device Evaluation. Condensation
(July 1980) (EPA/ESED report, Research Triangle Park, NC).
-------
E-3
C. COST EFFECTIVENESS
Cost effectiveness is the net annual cost, $350,000, divided by the annual VOC
reduction. From Table VI-1 the net annual VOC reduction achieved by using
surface condensers is
9.11 + 794.44 - 71.04 = 733 Mg/yr (rounded).
The cost effectiveness then is
$350^000 = $480/Mg of VOC (rounded).
-------
F-l
APPENDIX F
EXISTING PLANT CONSIDERATIONS
A. PROCESS CONTROLS
Table F-l shows the control devices or techniques used by some domestic ethylene
glycol producers. For the most part the data available are not current and do
not clearly define the emission controls used or the specific vents controlled.
Since the data sources (see Appendix C) do not specifically address aqueous
wastes nor fugitive and secondary emissions, no data were reported for these
categories by most respondents.
The design and operation of the water-removal section of the various existing
plants vary extensively and therefore the emissions and emission sources reported
also vary. The vapors from the evaporator purge vent are usually condensed.
If contact (barometric-type) condensers are used, the condensate ends up in the
cooling water. If surface condensers are used, the condensate is usually dis-
charged as wastewater. In some plants heat is recovered by the surface condensers.1
The amount of vapor purged from the evaporator may vary, depending on the product
end use or product quality requirements. Some plants purge most of the vapor
from the first-stage evaporator, as indicated for the model plant. Others may
purge a portion of the vapor from the second-stage evaporator. Still other
plants may not identify a purge stream as such but instead increase the amount
of vapor vented from the evaporator calandria vents.
In newer facilities the ethylene oxide and ethylene glycol plants may be inte-
grated, and common cooling towers, emission controls, and energy-saving tech-
niques be employed. Some plants collect the uncondensed gases from the glycol
plant vent condensers and route them back to a flare or thermal oxidizer
associated with the ethylene oxide plant.1'2
Union Carbide uses an air-cooled condenser on the evaporator calandria vents.1
Celanese, as well as several other manufacturers, sells the heavy-ends waste
product stream to an independent company for recovery of by-products.3 Some
manufacturers may be collecting the condensate from some vents and concentrating
the organic for inclusion with the waste product sold. Shell disposes of its
wastewater in a disposal well.4
-------
Table F-l. Emission Control Devices or Techniques Currently Used by Some Ethylene Glycol Producers*
Control Devices or Techniques Used
Producer
BASF, Geismar, LA
Calcasieu, Lake Charles, LA
Celanese, Clear Lake City, TX
Dow, Freeport, TX
Eastman, Longview, TX
PPG, Beaumont, TX
Shell, Geismar, LA
Union- Carbide,
Seadrift, TX
Taft, LA (Unit 1)
Taft, LA (Unit 2)
Texas City, TX (proposed)
Evaporator
Purge
Not reported
Not reported
Barometric
condenser
Not reported
Not reported
Not reported
Barometric
condenser
Not reported
Not reported
Not reported
Not reported
Evaporator
Calandria
Vent condenser
None
None
None
None
None
Barometric
condenser
None
None
Vent condenser
Vent condenser
Water Removal
Column
Barometric
condenser
EO plant
flare system
Barometric
condenser
None
None
None
Barometric
condenser
None
None
Not reported
Not reported
Product
Distillation
Columns
Barometric
condenser
EO plant
flare system
Barometric
condenser
None
None
None
Barometric
condenser
None
None
None
None
Aqueous
Wastes
Not reported
Secondary
treatment
Not reported
Not reported
Biological
oxidation
IT
Not reported i'
Disposal
well
Not reported
Not reported
Not reported
Wastewater
treatment
*See Appendix C.
-------
F-3
B. RETROFITTING CONTROLS
The primary difficulty associated with retrofitting may be in finding space to
fit the control device into the existing plant layout. Because of the costs
associated with this difficulty it may be appreciably more expensive to retrofit
emission control systems in existing plants than to install a control system
during construction of a new plant. The replacement of existing barometric
condensers in such a plant with new surface condensers will be significantly
more costly than the incremental cost in a new plent where this is merely an
alternative.
-------
F-4
C. REFERENCES*
1. J. F. Lawson, IT Enviroscience, Inc., Trip Report for Visit to Union Carbide
Corp., South Charleston, WV, Dec. 7, 1977 (data on file at EPA, ESED, Research
Triangle Park, NC).
2. BASF Wyandotte Corp., letter dated Nov. 27, 1978, in response to EPA's request
for information on emissions data on ethylene glycol production facilities.
3. Personal communication between R. H. Maurer, Celanese Chemical Co., Inc.,
Dallas, TX, and R. J. Lovell, IT Enviroscience, Inc., July 23, 1979.
4. Shell Oil Co., letter dated Jan. 11, 1979, in response to EPA's request for
information on emissions data on ethylene glycol production facilities.
*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.
-------
9-i
REPORT 9
GLYCOL ETHERS
T. L. Schemer
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.
D10L
-------
9-iii
CONTENTS OF REPORT 9
Page
I- ABBREVIATIONS AND CONVERSION FACTORS 1-1
II. INDUSTRY DESCRIPTION !!_!
A. Introduction II-I
B. Glycol Ethers Usage and Growth II-l
C. Domestic Producers U_3
D. References II-6
III. PROCESS DESCRIPTION III-l
A. Introduction III-l
B. Model Process for Manufacture of Ethylene Oxide—Derived III-l
Glycol Ethers
C. Propylene Oxide—Derived Glycol Ethers III-4
D. Process Variations III-4
E. References III-5
IV. EMISSIONS iv-1
A. Model Plant IV-i
B. Sources and Emissions IV-l
C. References IV-10
V. APPLICABLE CONTROL DEVICES V-l
A. Process Sources V-l
B. Fugitive Sources V-l
C. Storage and Handling Sources V-l
D. Secondary Emissions V-l
E. References V-2
VI. IMPACT ANALYSIS VI-1
A. Industry Emissions VI-1
B. References VI-3
APPENDICES OF REPORT 9
Page
A. PHYSICAL PROPERTIES OF GLYCOL ETHER PRODUCTS A-l
B. FUGITIVE-EMISSION FACTORS B-l
-------
9-v
TABLES OF REPORT 9
Number
II-l Usage and Growth of Glycol Ethers
II-2 Glycol Ethers Capacity
IV-1 Uncontrolled VOC Emissions Methanol Process
IV-2 Uncontrolled VOC Emissions Ethanol Process
IV-3 Uncontrolled VOC Emissions Butanol Process
IV-4 Model-Plant Storage Tank Data
IV-5 Composition of Vacuum System Condensate Methanol Process
VI-1 1978 Industry VOC Emissions
A-l Physical Properties of Ethylene Glycol Monomethyl Ether
A-2 Physical Properties of Ethylene Glycol Monoethyl Ether
A-3 Physical Properties of Ethylene Glycol Monobutyl Ether
A-4 Physical Properties of Diethylene Glycol Monomethyl Ether
A-5 Physical Properties of Diethylene Glycol Monoethyl Ether
A-6 Physical Properties of Diethylene Glycol Monobutyl Ether
A-7 Physical Properties of Triethylene Glycol Monomethyl Ether
A-8 Physical Properties of Triethylene Glycol Monoethyl Ether
A-9 Physical Properties of Triethylene Glycol Monobutyl Ether
II-2
II-4
IV-2
IV-3
IV-4
IV-6
IV-8
VI-2
A-2
A-2
A-3
A-3
A-4
A-4
A-5
A-5
A-5
Number
II-l
III-l
FIGURES OF REPORT 9
Manufacturing Locations of Glycol Ethers
Flow Diagram for Uncontrolled Typical Plant
II-5
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
M
Symbol
tera
giga
mega
kilo
milli
micro
Multiplication
Factor
1012
109
106
103
10"3
10"6
Example
1
1
1
1
1
1
Tg =
Gg =
Mg =
km =
mV =
Mg =
1
1
1
1
1
1
X
X
X
X
X
X
10 12 grams
109
106
103
10"
10"
grams
grams
meters
3 volt
6 gram
-------
II-l
II. INDUSTRY DESCRIPTION
A. INTRODUCTION
The production of glycol ethers was chosen for study because of its association
with the production of ethylene oxide, which, because of projected high volatile
organic compound (VOC) emissions, was studied early in the program. Over 90%
of the glycol ethers produced are derived from ethylene oxide; the remaining
amount is derived from propylene oxide.1
This study deals mainly with the production of nine of the ethylene oxide—
derived glycol ethers: the methyl, ethyl, and n-butyl monoethers of ethylene
glycol, diethylene glycol, and triethylene glycol. In most cases these glycol
ethers are produced in the same production facilities.
B. GLYCOL ETHERS USAGE AND GROWTH
The end uses of the ethylene oxide—derived glycol monoethers, the percentage
of total production of each, and their expected growth rates are shown in Table II-l.
From 1975 to the present almost one-third of the consumption of glycol ethers
has been for solvent applications in the protective-coating industry.1 Due to
their chemical structure, the glycol ethers have solvent properties similar to
those of alcohols and ethers.2 Their low evaporation rate makes them well suited
as coalescing agents in water-based surface-coating systems, which are beginning
to replace solvent-based surface coatings and therefore should be very influential
on the growth of glycol ethers. A variety of other uses include solvent applica-
tions in hydraulic fluids, printing inks, metal cleaners, and textile dyeing
processes and their use as chemical intermediates and jet fuel additives.1
The domestic production capacity of the nine major glycol ethers for 1980 is
reported to be 555,000 Mg. Applying the projected average annual growth rate
of 5%/yr during the years 1977—1982 to the 1977 production figure of 297,000 Mg
indicates that ~62% of this production capacity will be utilized in 1980. No
shortage of ethylene oxide is expected during this period.3
The glycol ethers derived from propylene oxide are used as coupling agents in
hydraulic fluids and as solvents by the coating industry for water-based paints.
The 1977 production was reported to be 20,000 Mg, with a projected annual growth
-------
II-2
Table II-l. Usage and Growth of Glycol Ethers*
1977
Production
Split
Projected
Average Annual
Growth (1977-1982)
Compound
End Use
Ethylene glycol
Monomethyl ether
Monoethyl ether
Monobutyl ether
Diethylene glycol
Monomethyl ether
Monoethyl ether
Monobutyl ether
Triethylene glycol
Monomethyl ether
Monoethyl ether
Monobutyl ether
Jet fuel additive; solvent in
protective coatings; additive
to textile and leather dyeing
processes
Production of ethylene glycol
monoethyl ether acetate;
solvent for protective
coatings and printing inks
Solvent for protective coatings;
diluent in hydraulic brake
fluids, rust removers, insecti-
cides, and herbicides
Solvent in wood stains, lacquers,
stamp pad inks, diluent for
hydraulic brake fluids, coalescing
agent for latex paints
Diluent for hydraulic brake fluids,
solvent for protective coatings,
textile printing and dyeing
Production of diethylene glycol
monobutyl ether acetate, coales-
cing agent in latex paints, sol-
vent for stamp pad inks, dyes,
diluent for hydraulic brake fluids
Diluent in brake fluids, solvent in
protective coatings
17
36
26
4
3
1
5.2
4.0
6.7
5.5
>5.7
*From ref 1.
-------
II-3
rate of consumption of 6—6.5% through 1982. On this basis the projected 1982
consumption of glycol ethers from propylene oxide will be approximately
27,000 Mg.1
C. DOMESTIC PRODUCERS
According to 1980 figures there are seven domestic companies producing ethylene
oxide—derived glycol ethers at ten plants. Table II-2 lists the producers and
their rated capacities.1 Figure II-l shows the plant locations.
The three companies that produce propylene oxide—derived glycol ethers are the
Dow Chemical Company, Olin Corporation, and Union Carbide Corporation.
-------
II-4
Table II-2. Glycol Ethers Capacity'
1980 Capacity
(10 Mg)b
Dow, Midland, MI
Dow, Plaquemine, LA
Jefferson, Port Heches, TX
Olin, Brandenberg, KY
PPG Industries, Beaumont, TX
Shell, Geismar, LA
Texas Eastman, Longview, TX
Union Carbide
Ponce, PR
Seadrift, TX
Taft, LA
Total
95
54
18
32
9
25
100
222
555
rrom ref 1.
Capacity amounts are flexible since they depend on the product
mix. Some capacities also include the capability for producing
propylene oxide—derived glycol ethers.
-------
11-5
1. Dow, Midland, MI 6.
2. Dow, Plaquernine, LA 7.
3. Jefferson, Port Heches, TX 8.
4. Olin, Brandenberg, KY 9.
5. PPG Industries, Beaumont, TX 10,
Shell, Geismar, LA
Texas Eastman, Longview, TX
Union Carbide, Ponce, PR
Union Carbide, Seadrift, TX
Union Carbide, Taft, LA
Fig. II-l. Manufacturing Locations of Glycol Ethers
-------
II-6
D. REFERENCES*
1. R. T. Gerry, "Glycol Ethers," pp. 663.5021B—663.5022Y in Chemical Economics
Handbook, Stanford Research Institute, Menlo Park, CA (April 1979).
2. H. W. Scheeline, Ethylene Glycols, Glycol Ethers and Ethanolamines, Report
No. 70, A private report by the Process Economics Program, Stanford Research
Institute, Menlo Park, CA (August 1971).
3. S. C. Johnson, "U.S. EO/EG—Past, Present, and Future," Hydrocarbons Processing
55(6), 109—113 (June 1976).
4. Chemical Products Synopsis, Glycol Ethers, Mannsville Chemical Products,
Mannsville, NY (April 1978).
*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
III. PROCESS DESCRIPTIONS
A. INTRODUCTION
The reaction of ethylene oxide with anhydrous methyl, ethyl, or n-butyl alcohol
is the only process that is commercially practiced in the United States for the
production of ethylene oxide—based glycol ethers.1 Similarly, glycol ethers
derived from propylene oxide are formed from the reaction of alcohols with propy-
lene oxide.
B. MODEL PROCESS FOR MANUFACTURE OF ETHYLENE OXIDE—DERIVED GLYCOL ETHERS
The ethylene glycol monoethers are produced by the following sodium hydroxide
catalyzed chemical reactions.2 Only the reactions with methanol are shown;
however, the primary alcohols, ethanol and butanol, react similarly to produce
the ethyl and butyl ethers.
CH3OH
(methanol)
CH3OCH2CH2OH
CH2-CH2
\/
(ethylene oxide)
CH2-CH2
(ethylene glycol (ethylene oxide)
monomethyl ether)
CH3OCH2CH2OH
(ethylene glycol monomethyl ether)
—>• CH3OCH2CH2OCH2CH2OH
(diethylene glycol monomethyl ether)
CH3OCH2CH2OCH2CH2OH
(diethylene glycol mono-
methyl ether)
CH2-CH2
\ /
V
(ethylene oxide)
CH3OCH2CH2OCH2CH20CH2CH20H
(triethylene glycol monomethyl
ether)
The mono-, di-, and triethylene glycol products are produced simultaneously. The
reaction and recovery operations are continuous.
The model continuous process for the manufacture of the glycol ethers is shown
in Fig. III-l. The sodium hydroxide catalyst (stream 1) (acid catalyst can be
used) and one of the anhydrous primary alcohols methanol, ethanol, or n-butanol
(stream 2) are blended in the mix tank.3 The material from the alcohol catalyst
storage tank is combined with ethylene oxide (stream 3) and with the recycled
-------
o
AL-COHOU^^
fto-1
|TT - It— cw P
L5_F
X. TAUK. ^^-^ I
C.l.VCOL. C.THCR
I'UUI'OUT
TAMK^
:nAru:-(i>jc,
Fig. III-l. Flow Diagram for Uncontrolled Model Plant Producing Glycol Ethers from Ethylene Oxide
-------
III-3
alcohol (stream 4) and is then fed to the reactor. The reaction is carried out
at an elevated pressure (2.5 X 106 to 4.6 X 106 Pa) and temperature (200 to
230°C). The reaction between ethylene oxide and the alcohols is exothermic (20
to 25 kg-cal per g-mole of ethylene oxide reacted).4
Ethylene oxide reacts with some of the ethylene glycol ether to form diethylene
glycol ether and with some of the diethylene glycol ether to form triethylene
glycol ether. The reaction product consists of a mixture of mono-, di-, and
triethylene glycol ethers, as well as some higher molecular weight glycol
ethers.2 The reaction mixture product distribution is influenced by the
alcohol:ethylene oxide ratio in the reaction feed. A higher alcohol:ethylene
oxide ratio reduces the formation of higher glycol ethers.
The product stream (5) exits the reactor and is sent to the alcohol distilla-
tion column, where excess alcohol is distilled overhead and recycled (stream 6) for
future reaction. The column is normally operated at atmospheric pressure
although, for the higher alcohols, it could be operated under a slight vacuum.
The alcohol column bottoms (stream 7) are then sent to the monoethylene glycol
ether column, where monoethylene glycol ether is vacuum distilled and sent
(stream 8) to product storage via one of the two monoglycol ether day tanks.
Similarly, diethylene glycol ether and triethylene glycol ether are vacuum
distilled consecutively in two more distillation columns. The vacuum system
normally consists of a four-stage steam-jet series with surface intercondensers.
The diethylene glycol ether product (stream 9) and triethylene glycol ether
product (stream 10) streams are sent to their respective storage tanks.2 The
heavy ends (stream 11) from the triethylene glycol ether column is disposed
of.3 No data were available on the disposal of this stream,- however, it is
probably incinerated or landfilled.
When product lines are switched, the process equipment is drained to one of the
three pump-out tanks. The contents of the pump-out tank containing the next
product line to be produced is then charged to the process. The columns are
put on total reflux, feed is slowly started, and the product streams are returned
to the pump-out tank until product specification is attained. When specifications
are met, the feed is increased to design levels and the product streams are
sent to the day tanks.
-------
III-4
C. PROPYLENE OXIDE — DERIVED GLYCOL ETHERS
The propylene glycol monoethers are produced by the following reaction. The
most important family of these ethers utilize methanol as the primary alcohol.5
H2C-CH-CH3 + CH3OH - > CH3OCH2CHOHCH3
(propylene oxide) (methanol) (propylene glycol
monomethyl ether)
As in the ethylene oxide — derived process, the propylene glycol monomethyl ethers
formed in the reactor react further with the propylene oxide to form di- and
tripropylene glycol monomethyl ethers. An excess of anhydrous methanol limits
the formation of these higher ethers. In some cases the same equipment is be-
lieved to be used for the production of ethylene oxide.
D. PROCESS VARIATIONS
Jefferson Chemical Company produces only methyl and ethyl glycol ethers from
ethylene oxide. Their production unit has only two distillation towers; there-
fore the unit must be operated in two separate passes to recover both mono- and
di-ethers.6
-------
III-5
E. REFERENCES*
1. J. L. Blackford, "Glycol Ethers," pp. 663.5021B—663.5022Y in Chemical Economics
Handbook, Stanford Research Institute, Menlo Park, CA (November 1976).
2. H. W. Scheeline, Ethylene Glycols, Glycol Ethers and Ethanolamines, Report
No. 70, A private report by the Process Economics Program, Stanford Research
Institute, Menlo Park, CA (August 1971).
3. T. L. Schomer, IT Enviroscience, Inc., Trip Report to Dow Chemical Co.,
Midland, MI Feb. 10, 1978 (on file at EPA, ESED, Research Triangle Park,
NC).
4. R. N. Shreve and L. F. Albright, "Alkylation," Chap. 14, p. 804, in McGraw-
Hill Series in Chemical Engineering,."Unit Processes in Organic Synthesis,"
P. H. Groggins, editor-in-chief, McGraw-Hill Book Co., New York, 1958.
5. J. L. Blackford, "Propylene Oxide," pp. 690.8021B—690.8022C in Chemical Economics
Handbook, Stanford Research Institute, Menlo Park, CA (November 1976).
6. J. F. Cooper, Jefferson Chemical Company, letter dated July 11, 1979, to
David R. Patrick, EPA, 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.
-------
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 atmos-
phere, participate in photochemical reactions producing ozone. A relatively
small number of organic chemicals have low or negligible photochemical re-
activity. However, many of these organic chemicals are of concern and may be
subject to regulation by EPA under Section 111 or 112 of the Clean Air Act since
there are associated health or welfare impacts other than those related to ozone
formation.
A. MODEL PLANT
The glycol ether capacity selected for the model plant is 45,400 Mg/yr. For
this study the nine major glycol ethers that utilize methanol, ethanol, and
n-butanol as raw materials are produced in the same process equipment on a part-
time basis. The relative amounts of each glycol ether produced in the United
States in 1975 formed the basis for the following production capacity split for
the model plant: 10,800 Mg of methyl glycol ethers, 19,000 Mg of ethyl glycol
ethers, and 15,600 Mg of n-butyl glycol ethers.*
The model plant is assumed to operate 8256 hr a year, with 504 hr a year
allotted for process down time required for switching from one alcohol-based
glycol ether product line to another. This is common practice in most of the
industry. However, some production facilities utilize only one of the alcohols
mentioned,1 in which case the plant would not require time for switching product
lines and would not require as many storage tanks. In either case the process
equipment requirements would be the same.
B. SOURCES AND EMISSIONS
Emission sources and uncontrolled emission rates of VOC for the model plant
producing methanol-, ethanol-, and n-butanol-based glycol ethers are summarized
in Tables IV-1, IV-2, and IV-3 and are discussed below.
*In order to minimize the revision time, the 1975 production split that was used
for the original draft of this report is used rather than the 1977 split. For
our purposes the differences are not believed to be significant.
-------
IV-2
Table IV-1. Uncontrolled Emissions of Total VOC from the
Production of Methanol-Based Glycol Ethers3
Source
Catalyst-methanol mix tank
Methanol recovery column vent
c
Vacuum system vent
Storage vents
Fugitive emissions
Secondary emissions
Stream
Designation
(Fig.III-1)
A
B
C
D
E
F
Total VOC
b
Ratio
(g/kg)
0.0097
0.15
0.013
0.24
0.19
0.03
Emissions
Rate
(kg/hr)
0.05
0.82
0.07
0.30G
1.04
0.14
a
Emission ratios and emission rates apply only to methyl glycol ether production,
which is 1960 hr/yr; the annual production is 1.08 X 107 kg for model plant.
g of emission per kg of methyl glycol ethers produced.
See ref 2.
See ref 2 for inert-gas flow from alcohol column.
g
Weighted average of storage emissions for 8760 hr/yr.
-------
IV-3
Table IV-2. Uncontrolled Emissions of Total VOC from the
Production of Ethanol-Based Glycol Ethers3
Source
c
Catalyst-ethanol mix tank
d
Ethanol recovery column vent
e
Vacuum system vent
Storage vents
Fugitive emissions
Secondary emissions
Stream
Designation
(Fig.III-1)
A
B
C
D
E
F
Total VOC
b
Ratio
(g/kg)
0.006
0.093
0.013
0.17
0.19
0.03
Emissions
Rate
(kg/hr)
0.03
0.51
0.07
0.37f
1.04
0.14
Emission ratios and emission rates apply only to ethyl glycol ether production,
which is 3450 hr/yr; the annual production is 1.90 X 107 kg for model plant.
g of emission per kg of ethyl glycol ethers produced.
Calculated from catalyst-methanol mix tank emission (Table IV-1) number and
ratios of vapor pressure and molecular weight of ethanol to methanol.
See ref 2 for inert-gas flow from alcohol column.
6See ref 2.
Weighted average of storage emissions for 8760 hr/yr.
-------
IV-4
Table IV-3. Uncontrolled Emissions of Total VOC from the
Production of Butanol-Based Glycol Ethers3
Source
Q
Catalyst-butanol mix tank
d
Butanol recovery column vent
e
Vacuum system vent
Storage vents
Fugitive emissions
Secondary emissions
C +- T*o 3TT1
•3 L..L CCUll
Designation
(Fig.III-1)
A
B
C
D
E
F
Total VOC
b
Ratio
(g/kg)
0.001
0.016
0.013
0.10
0.19
0.03
Emissions
Rate
(kg/hr)
0.005
0.09
0.07
f
0.17
1.04
0.14
Emission ratios and emission rates apply only to butyl glycol ether production.
Butyl glycol ether production is 2850 hr/yr; the annual production is
1.56 X 107 kg for model plant.
g of emission per kg of butyl glycol ethers produced.
Q-
calculated from catalyst-methanol mix tank emission (Table IV-1) and ratios of
vapor pressure and molecular weight of butanol to methanol.
See ref 2 for inert-gas flow from alcohol column.
GSee ref 2.
Weighted average of storage emissions for 8760 hr/yr.
-------
IV-5
1. Alcohol-Catalyst Mix Tank and Storage Tank Vents
The mix tank and storage tank are blanketed with nitrogen to maintain anhydrous
conditions (vent A, Fig. III-l); otherwise the water present will contaminate
the glycol ethers. This vent stream contains nitrogen and, depending on the
product line being made, one of the three alcohols methanol, ethanol, or
n-butanol. The emission data shown in Table IV-1 for the mix-tank vent for
methyl ethers were provided by industry.2 The emissions from this vent for
ethyl and n-butyl ethers, shown in Tables IV-2 and IV-3 respectively, were
calculated from the methyl ethers emission and the ratios of molecular weight
and vapor pressure of ethanol and butanol to methanol. Much of the heat gene-
rated from the dissolution of sodium hydroxide in alcohol is removed in the mix
tank with cooling water,- therefore no appreciable fluctuation occurs in the
temperature of the vessel contents or in the alcohol emissions.
2. Alcohol Recovery Column Vent
This vent emits alcohol and inert gases from the alcohol column reflux tank
(source B). The purpose of the inert-gas stream is to prevent moisture in the
air from contacting the alcohol. The total uncontrolled VOC emission rates
from this vent during capacity production of the methanol-, ethanol-, and
n-butanol-based glycol ethers are 0.82 kg/hr, 0.51 kg/hr, and 0.09 kg/hr
respectively.
3. Vacuum System Vent2
The vent (C, Fig. III-l) from the vacuum system contains water vapor, inert
gases, and a small percentage of VOC. The total uncontrolled VOC emission rate
is estimated to be 0.07 kg/hr for the model plant operating at capacity. This
rate was estimated by an industrial producer of methyl glycol ethers. Since
this emission is small and emission data were not available for this stream
i
during production of ethyl and butyl glycol ethers, the emission is assumed to
be the same for all three product lines.
4. Storage and Handling Emissions
Emissions result from the storage and handling of the glycol ether products and
the alcohol raw materials. Sources (D) of the losses are shown in Fig. III-l.
All storage tanks are blanketed with nitrogen to maintain anhydrous conditions.
Storage tank parameters for the model plant are given in Table IV-4. Since
-------
IV-6
Table IV-4. Model-Plant Storage Tank Data
Storage Facility
Feed tanks
Recycle tanks
Product day tanks
Product storage tanks
Product pump-out
tanks
Contents
Methanol
Ethanol
Butanol
Ethylene oxide
Methanol
Ethanol
Butanol
Ethylene glycol ethers
Diethylene glycol ethers
Triethylene glycol ethers
Ethylene glycol methyl ether
Ethylene glycol ethyl ether
Ethylene glycol butyl ether
Diethylene glycol methyl ether
Diethylene glycol ethyl ether
Diethylene glycol butyl ether
Triethylene glycol methyl ether
Triethylene glycol ethyl ether
Triethylene glycol butyl ether
Methanol, mono-, di-, and tri-
ethylene glycol methyl ethers
Ethanol, mono-, di-, and tri-
ethylene glycol ethyl ethers
Butanol, mono-, di-, and tri-
ethylene glycol butyl ethers
No. of
Tanks
Required
1
1
1
1
1
1
1
2
2
2
.1
1
1
1
1
1
1
1
1
1
1
1
Tank
Size
757
1136
1893
1325
15
23
11
114
17
26
1135
1702
2270
170
265
378
416
189
151
8
8
8
Turnovers
per Year
6
10
6
21
5
5
5
164a
169b
97C
6
9
6
6
9
6
6
9
6
1
1
1
Bulk
Temperatur*
20
20
20
10
20
20
20
60
60
60
20
20
20
20
20
20
20
20
20
20
20
20
„
32 turnovers for production of methyl ethers, 70 for ethyl ethers, and 62
30 turnovers for production of methyl ethers, 72 for ethyl ethers, and 67
49 turnovers for production of methyl ethers, 32 for ethyl ethers, and 16
for butyl ethers,
for butyl ethers.
for butyl ethers.
-------
IV-7
ethylene oxide is stored in a pressurized tank, none of it is emitted. The
uncontrolled storage and handling emission rates in Tables IV-1, IV-2, and IV-3
are based on fixed-roof tanks, half full, and a diurnal temperature variation
of 22°C and the use of the emission equations from AP-42.3 However, breathing
losses were divided by 4 to account for recent evidence indicating that the
AP-42 breathing loss equation overpredicts emissions.
The losses from the multiuse day tanks are assigned to a product line only during
the period that product is being produced, while the losses from the feed, recycle,
product storage, and product pump-out tanks are assigned to the product line
for which they are used. It is assumed that all feed and product storage tanks
are left half full when the product line they are assigned to is not being pro-
duced; therefore breathing losses from these tanks occur all year. The storage
losses from each product line expressed as emission rates in Tables IV-1, IV-2,
and IV-3 are prorated for the entire year.
5. Fugitive Emissions
Process pumps and valves are potential sources of fugitive emissions. The typical
plant is assumed to have 34 pumps, 300 process valves, and 30 pressure relief
valves handling VOC. The fugitive emission factors from Appendix B were applied
to the valve and pump count to determine the uncontrolled fugitive emission
rate of 4.4 kg/hr from the model plant. It is assumed that the fugitive emissions
rate remains constant throughout the year, regardless of the product line being
produced.
6. Secondary Emissions
Secondary VOC emissions can result from the handling and disposal of process-waste
streams. Two potential sources (F) are indicated in Fig. III-l for the model
plant. The composition of the aqueous condensate stream from the vacuum system
during methanol-based glycol ether production is shown in Table IV-5. Data are
not available on the composition of this stream during ethanol- and butanol-based
glycol ether production; however, the hydrocarbon composition is expected to be
approximately the same during production of all three alcohol-based glycol ethers.
The emission rate from this source is estimated to be 0.01 kg/hr, assuming that
the emissions are the same for all product lines.
-------
IV-8
Table IV-5. Composition of Aqueous Condensate Stream from
Vacuum System During Methanol-Based Glycol Ether Productiona
Production Ratio
Stream Composition (g/kg)
Water 350
Methanol ^0.063
Ethylene glycol methyl ether 'v-0.094
Diethylene glycol methyl ether ^0.063
Triethylene glycol ethyl ether ^0.015
Total hydrocarbons ^0.235
a
From ref 2.
g of component per kg of glycol ether produced.
-------
IV-9
Disposal of the bottoms from the triethylene glycol column represents another
potential secondary emission. The uncontrolled secondary emission of total VOC
from this source for the model plant operating at capacity has been estimated
from data supplied by industry to be 0.13 kg/hr.4
7. Process Variation
One manufacturer indicated that the overheads from the alcohol column are sub-
cooled to 25 to 30°C by the column condenser.2 This subcooling reduces the
vapor pressure of the condensed overheads and therefore reduces the amount of
VOC emissions through the alcohol vent. This procedure is known to be prac-
ticed in the glycol ether industry but not necessarily by all manufacturers.
-------
IV-10
C. REFERENCES*
1. H. W. Scheeline, Ethylene Glycol, Glycol Ethers and Ethanolamines, Report
No. 70, A private report by the Process Economics Program, Stanford Research
Institute, Menlo Park, CA (August 1971).
2. T. L. Schomer, IT Enviroscience, Inc., Trip Report to Union Carbide Corp- , South
Charleston, WV, Feb. 13, 1978 (on file at EPA, ESED, Research Triangle Park,
3. C. C. Masser, "Storage of Petroleum Liquids," p. 4.3-6 in Compilation of Air
Pollutant Emission Factors, AP-42, Part A, 3d ed., U.S. EPA, Research Triangle
Park, NC (August 1977)
4. T. L. Schomer, IT Enviroscience, Inc., Trip Report to Dow Chemical Co., Midland,
HI, Feb. 10, 1978 (on file at 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.
-------
V-l
V. APPLICABLE CONTROL DEVICES
A. PROCESS SOURCES
Process emissions from the model plant occur from vents A, B, and C. (See
Fig. III-l for vent locations and Tables IV-1, IV-2, and IV-3 for uncontrolled
emissions.)
The sum of the process emission rates for the glycol ether processes based on
methanol, ethanol, and butanol are 0.94 kg of VOC/hr, 0.61 kg of VOC/hr, and
0.16 kg of VOC/hr respectively. Since there is a relatively small quantity of
process VOC emissions, due primarily to the low volatility of the products, no
emission control devices have been identified.1'2
B. FUGITIVE SOURCES
Controls for fugitive sources will be discussed in a future document covering
fugitive emissions from the synthetic organic chemicals manufacturing industry
(SOCMI).
C. STORAGE AND HANDLING SOURCES
Controls for SOCMI storage emissions are discussed in a separate EPA document.3
D. SECONDARY EMISSIONS
Secondary emissions for SOCMI are covered by a separate EPA document.4
-------
V-2
E. REFERENCES*
1. T. L. Schemer, IT Enviroscience, Inc., Trip Report to Union Carbide Corp., South
Charleston, WV, Feb. 13, 1978 (data on file at EPA, ESED, Research Triangle Park,
NC).
2. T. L. Schomer, IT Enviroscience, Inc., Trip Report to Dow Chemical Co., Midland, MI,
Feb. 10, 1978 (on file at EPA, ESED, Research Triangle Park, NC).
3. D. G. Erikson, IT Enviroscience, Inc., Emission Control Options for the Synthetic
Organic Chemicals Manufacturing Industry Storage and Handling Report (EPA,
ESED, Research Triangle Park, NC) (October 1978).
4. J. J. Cudahy and R. L. Standifer, IT Enviroscience, Inc., Secondary Emissions
Report (EPA, ESED, Research Triangle Park, NC) (October 1979).
*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.
-------
VI-1
VI. IMPACT ANALYSIS
A. INDUSTRY EMISSIONS
The information available from industry indicates that no emission control devices
are utilized in glycol ether production facilities to control process emissions.1'2
Tables IV-1, IV-2, and IV-3 show that the process emissions from the methyl,
ethyl, and n-butyl glycol ether processes are small.
The production of glycol ethers from ethylene oxide in 1978 is estimated at
281,000 Mg.3 The emissions associated with this production are shown in Table IV-1.
Emission factors from Sect. IV were used to calculate the emissions in Table VI-1.
The production split of methyl, ethyl, and n-butyl glycol ethers is assumed to
be the same as the split reported in Sect. IV-A for 1975.
Fugitive emissions were calculated by using the fugitive emissions factors of
Appendix B and are estimated to comprise over one-third of the total emissions
from the glycol ethers manufacturing industry.
-------
VI-2
Table VI-1. 1978 Industry VOC Emissions from Glycol Ether Production
Using Ethylene Oxide
VOC Emissions (Ma/vr)
Source
. . b
Process emissions
b
Storage vents
c
Fugitive emissions
d
Secondary emissions
Methyl
Glycol
Ethers
11.6
35.5
18.6
2.0
Ethyl
Glycol
Ethers
13.1
41.0
32.9
3.5
Butyl
Glycol
Ethers
2.9
20.4
27.2
2.9
Total
Glycol
Ethers
27.6
96.9
78.7
.? ..-A
211.6
1978 glycol ether production, 281,000 Mg total (67,000 Mg of
methyl glycol ethers; 117,000 Mg of ethyl glycol ethers? and
97,000 Mg of butyl glycol ethers).
Calculated from Tables IV-1 through Table IV-3.
'Controlled fugitive emission factor is 0.28 g/kg.
Secondary emission ratio from Tables IV-1 through IV-3.
-------
VI-3
B. REFERENCES
1. T. L. Schomer, IT Enviroscience, Inc., Trip Report to Union Carbide Corp.,
South Charleston WV, Feb. 13, 1978 (on file at EPA, ESED, Research
Triangle Park, NC).
2. T. L. Schomer, IT Enviroscience, Inc., Trip Report to Dow Chemical Co., Midland, MI,
Feb. 10, 1978 (on file at EPA, ESED, Research Triangle Park, NC).
3. 1978 Directory of Chemical Producers, United States of America, Chemical
Information Services, p. 645, Stanford Research Institute, Menlo Park, CA.
*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
PHYSICAL PROPERTIES OF GLYCOL ETHER PRODUCTS
The data sources for the physical properties given in the following tables are
as follows:
Table A-l: J. Dorigan et al., "Scoring of Organic Air Pollutants. Chemistry,
Production and Toxicity of Selected Synthetic Organic Chemicals
(Chemicals D-E)," MTR-7248, Rev. No. 1, Appendix II, p. AII-292,
Mitre Corp. (September 1976).
Table A-2: MTR-7248, App. II, p. AII-286; also: Welcome to the World of Dow
Products and Services, Catalog of Dow Products, pp. 6 and 7, 1971.
Table A-3: MTR-7248, App. II, p. AII-284; Dow Catalog, p. 6.
Table A-4: MTR-7248, App. II, p. AII-98; Dow Catalog, p. 6.
Table A-5: MTR-7248, App. II, p. AII-92; Dow Catalog, p. 6.
Table A-6: MTR-7248, App. II, p. A-88; Dow Catalog, p. 6.
Table A-7: J. Dorigan e_t al., "Scoring of Organic Air Pollutants. Chemistry,
Production and Toxicity of Selected Synthetic Organic Chemicals
(Chemicals 0-Z)," MTR-7248, Rev. No. 1, Appendix IV, p. AIV-268,
Mitre Corp. (September 1976).
Table A-8: G. G. Hawley, The Condensed Chemical Dictionary, 8th Ed., p. 356,
Van Nostrand Reinhold, New York, 1971.
Table A-9: Ibid., p. 136.
-------
A-2
Table A-l. Physical Properties of Ethylene Glycol
Monomethyl Ether
Synonyms
Molecular formula
Molecular weight
Vapor pressure
Melting point
Boiling point
Density
Physical state
Water solubility
2-Methoxyethanol, ethoxyacetate
C3H8°2
76.1
6.2 mm Hg at 20°C
-81.5°C
124.5°C at 1 atm
0.966 g/ml at 20°C/4°C
Liquid
Infinite
Table A-2. Physical Properties of Ethylene Glycol
Monoethyl Ether
Synonyms
Molecular formula
Molecular weight
Vapor pressure
Pour point
Boiling point
Density
1'hy.sical stato
Water solubility
2-Ethoxyethanol
C4H10°2
90.1
3.8 mm Hg at 20°C
-100°C
. 135.1°C
0.9360 g/ml at 50°C/15°C
Liquid
Infinite
-------
A-3
T.ihlc.' A- I. Phys i c.i I Proper t: i o:; of Ktliy I r>n<; OlyooL
Mnnobii I y I !•:( her
Synonyms 2-Butoxyethanol
Molecular formula r H. .00
b 14 2.
Molecular weight 118.2
Vapor pressure 0.76 mm Hg at 20°C
Pour point -75°C
Boiling point 171.2°C
Density 0.9027 g/ml at 20°C/4°C
Physical state . Liquid
Water solubility Infinite
Table A-4. Physical Properties of Diethylene Glycol
Monomethyl Ether
Synonyms 2~(2-Methoxyethoxy)ethanol
Molecular formula C5H12°3
Molecular weight 120.2
Vapor pressure 0.2 mm Hg at 20°C
Pour point -85°C
Boiling point 194.2°C
Donsity 1.0354 g/ml at 20°C/4°C
Physical state Liquid
Water solubility Infinite
-------
A-4
Tablo A-5. Physical Properties of Diotr.hylono Glycol
M' in'if -t hy I lit he r
2- (2-Ethoxyethoxy)ethanol
Molecular formula C^H-, /,Oo
6 14 3
Molecular weight 134.2
Vapor pressure <1.0 mm Ilg at 20°C
Pour point -90°C
Boiling point 201,9°C
Density 0.9902 g/ml at 20°C/4°C
Physical state Liquid
Water solubility Infinite
Table A-6. Physical Properties of Diethylene Glycol
Monobutyl Ether
Synonyms 2-(2-Isobutoxyethoxy)ethanol
Molecular formula CH 0
o -Lo J
Molecular weight 162.2
Vapor pressure 0.01 mm Hg at 20°C
Pour point -68.1°C
Boiling point 230.6°C
Density . 0.9536 g/ml at 20°C/20°C
Water solubility Infinite
-------
A-5
A-'/. i-:-r/:.ical Properties of Triethylene Glycol
Monomethyl Ether
Synonyms 2-[2-(2-Methoxyethoxy)ethoxy]
ethanol, methoxytriglycol,
methoxytriethylene glycol
Molecular formula C_H O
/ ID 4
Molecular weight 164.2
Vapor pressure <0.01 mm Hg at 20°C
Boiling point 249°C
Density 1.0494 g/ml
Physical state Liquid
Water solubility Infinite
Table A-8. Physical Properties of Triethylene Glycol
Monoethyl Ether
Synonyms Ethoxytriglycol
Molecular formula C H 0(C H O) H
£ J ^ ^t O
Molecular weight 178
Vapor pressure <0.01 Hg at 20°C
Melting point -18.7°C
Boiling point 255.4°C
Density 1.021 g/ml at 20°C/20°C
Physical state Liquid
Water solubility Infinite
Table A-9. Physical Properties of Triethylene Glycol
Monobutyl Ether
Synonyms Butoxytriglycol
Molecular formula C4H9°(C2H4O)3H
Molecular weight 206.3
Vapor pressure <0.01 mm Hg at 20°C/20°C
Physical state Liquid
Water solubility Infinite
-------
B-l
APPENDIX B
FUGITIVE EMISSION FACTORS
Fugitive emission factors established for petroleum refinery operation and
published in AP-42 are based on emission quantities per unit throughput and
therefore are unsatisfactory for use here. The emission factors for each
equipment component used in this report are based on the orginal emission
2--4
studies used to establish the AP-42 factors with assumptions as follows:
1. Pump Seals (including standby pumps)
a. "Uncontrolled" is the average loss measured for mechanical seals.
b. "Controlled" is the average loss for mechanical seals, with major
leaks assumed to be fixed.
Uncontrolled Controlled
Pump seals (kg/day/seal) 1.5 0.16
2. Compressor Seals
a. "Uncontrolled" is the average loss measured for all seals venting to
atmosphere.
b. "Controlled" is the average loss based on the large leaks being
fixed.
Uncontrolled Controlled
Compressor seals (kg/day/seal) 3.9 1.0
3. Valves
a. "Uncontrolled" is the average loss measured for all valves.
b. "Controlled" is the average loss, with the large leaks assumed to be
fixed.
Uncontrolled Controlled
Pipeline valves (kg/day/valve) 0.068 0.006
-------
B-2
4. Pressure Relief Devices
a. "Uncontrolled" is the average loss measured for all valves.
b. "Controlled" is the average loss based on the assumption that the
large leaks are fixed.
Uncontrolled Controlled
Pressure relief devices
(kg/day/valve) 1.1 0.1
REFERENCES*
W. M. Vatavak, "Petroleum Industry," pp. 9.1-1 to 9.1-8 in Compilation of
Air Pollutant Emission Factors, 2d ed., AP-42, EPA, Research Triangle Park,
NC (March 1975).
R. K. Palmer, Hydrocarbon Losses from Valves and Flanges. Report No. 2,
PB-216-682, Joint District, Federal and State Project for the Evaluation of
Refinery Emissions. Air Pollution Control District, County of Los Angeles, CA
(March 1957).
B. J. Steigerwald, Emissions of Hydrocarbons to the Atmosphere from Seals on
Pumps and Compressors. Report No. 6, PR-216-582, Joint District, Federal and
State Project for the Evaluation of Refinery Emissions. Air Pollution Control
District, County of Los Angeles, CA (April 1958).
4
B. J. Steigerwald, Hydrocarbon Leakage from Pressure Relief Valves. Report
No. 3, PB-216-715, Joint District, Federal and State Project for the Evaluation
of Refinery Emissions. Air Pollution Control District, County of Los Angeles,
CA (May 1957).
*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.
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
. FiefrOFT NO.
EPA-450/3-80-028d
2.
3. RECIPIENT'S ACCESSION NO.
T.T.E AND SUBTITLE
Organic Chemical Manufacturing
Volume 9: Selected Processes
5. REPORT DATE
December 1980
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
R. i). liovoll, J. A. Key, R. Ti. Standifor,
V. KalcovJc, .). !•'. l.awson, K. W. Dylowski, T. I.. Schemer
8. PERFORMING ORGANIZATION REPORT MO.
S. PERFORMING ORGANIZATION NAME AND ADDRESS
IT Enviroscience, Inc.
9041 Executive Park Drive
Suite 226
Knoxville, Tennessee 37923
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-02-2577
12
. SPONSORING AGENCY NAME AND AO.DRESS
DAA for Air Quality Planning and Standards
Office of Air, Noise, and Radiation
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
13. T-YPE CXF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
EPA/200/04
IE. SUPPLEMENTARY NOTES
16. ABSTRACT
EPA is developing new source performance standards under Section 111 of
the Clean Air Act and national emission standards for hazardous air pollutants
under Section 112 for volatile organic compound emissions (VOC) from organic
chemical manufacturing facilities. In support of this effort, data were gathered
on chemical processing routes, VOC emissions, control techniques, control costs,
and environmental impacts resulting from control. These data have been analyzed
and assimilated into the ten volumes comprising this report.
This volume presents in-depth studies of several major organic chemical
products.
17.-
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b. IDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
13B
. D'STRiBUTiQN STATEMENT
Unlimited Distribution
19 SECURITY CLASS (This Report)
Unclassified
j21. NO. OF PAGES
! 545
20 SECURITY CLASS (This page/
Unclassified
22. PRICE
Forrr 2:20-1 (K*v. 4-77) PREVIOUS EDi TiON • S O BSOLET E
------- |