United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EPA-450/3-85-014
April 1985
Air
Sources of Ethylene
Oxide Emissions
;£«fega!r*<^> ^fegdiVK^^4^<^:'^
at^^
§§•&#£?*:
g.A^^^^>-.'v^^g>^(Sa>^xr-' - -\~S^*"-f_^ [^ £•,•*&&!'— fj -• .-"''•
-------
EPA-450/3-85-014
Sources of Ethylene Oxide Emissions
David W. Markwordt
Emission Standards and Engineering Division
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
April 1985
-------
This report has been reviewed by the Emission Standards and Engineering Division of the Office of Air
Quality Planning and Standards, EPA, and approved for publication. Mention of trade names or commercial
products is not intended to constitute endorsement or recommendation for use. Copies of this report are
available through the Library Services Office (MD-35), U.S. Environmental Protection Agency, Research
Triangle Park, N.C. 27711, or for a fee from the National Technical Information Services, 5285 Port Royal
Road, Springfield, Virginia 22161.
-------
EXECUTIVE SUMMARY
The objective of this report is to present the results of.a preliminary
source assessment study conducted for ethylene oxide (EO) in which data
were developed that will aid in evaluating the need for further regulatory
development for EO.
The EO source assessment was focused primarily on air emissions of
EO. Processes which could potentially produce air emissions were identified
and the levels of total EO generated were quantified. The EO emitted to
air was estimated for individual sources from EO production sites (including
captive use), ethoxylation sites, and major sterilizer/fumigator facilities.
Emission categories included leaking equipment, cooling towers, vents,
storage and loading operations, and effluent treatment facilities. Data
were obtained from formal Information requests to operating companies.
National emissions of EO are estimated at approximately 5,000 megagrams
per year. Production/captive use, ethoxylation, medical facilities, and
sterilization/ fumigation sites account for 31, 4, 8, and 57 percent of
total emissions, respectively (see Table 1). Currently, 15 plants produce
approximately 2.5 million megagrams of EO annually. Approximately,
80-90 percent of EO production is captively used on site. The bulk of
emissions from producer and ethoxylator sites are from equipment leaks.
Less than 0.1 percent of EO production is used in sterllizer/fumigator
process, but nearly all of the uncontrolled EO is emitted to the atmosphere.
Control technologies currently used to reduce air emissions of EO
were Identified and evaluated to estimate what level of control exists at
EO sources. The technologies currently applied, along with other potentially
i« •
17
-------
feasible technologies, were assessed to define what constituted effective
control for EO emissions from the various sources. The controls assessment
included an examination of control device applicability and effectiveness
and control device cost.
Application of available control techniques for all £0 emission
sources (excluding medical facilities) could reduce emissions by approxi-
matley 87 percent. The control costs per megagram of EO emissions range
from 460 to 2,500 depending on the emission source and applied technique.
The decision to initiate regulatory development will depend to some
extent on the estimated health effects attributed to exposure to ambient
levels of EO. Emission parameters developed for EO dispersion modeling
are summarized in Appendix B.
Table 1. NATIONAL EMISSIONS - ETHYLENE OXIDE (Mg/yr)
Source
A. Production/Captive Use
1. Equipment Leaks
2. Cooling Towers
3. Vents
4. Storage and Loading
5. Effluents
B. Ethoxylation
C. Medical Facilities
D. Sterilization/Fumigation
Baseline Emissions
1,540
1,200
190
100
40
10
200
400
2,800
4,940
Potential
Controlled
% Total Emissions
31 900-500
700-300
50
100
40
10
4 85-50
8 400
57 30 .
100 1,420-980
IV
-------
Table of Contents
Secti on Page
1 INTRODUCTION 1-1
2 INDUSTRY PROFILE 2-1
2.1 Production 2-1
2.2 Captive Feed Use 2-11
2.3 Ethoxylation 2-18
2.4 Sterilization/Fumigation 2-19
3 NATIONAL EMISSIONS 3-1
3.1 Production Facilities
3.1.1 Equipment Leaks 3-5
3.1.2 Cooling Towers 3-5
3.1.3 Vents 3-9
3.1.4 Storage & Loading 3-12
3.2 Ethoxylation 3-13
3.3 Sterilization/Fumigation 3-13
4 CONTROL TECHNIQUES 4-1
4.1 Equipment Leaks 4-1
4.2 Cooling Towers 4-4
4.3 Vents 4-5
4.4 Steri lization/Fumi gation 4-5
5 CONTROL TECHNIQUES COSTS 5-1
5.1 Equipment Leaks 5-1
5.2 Steri lization/Fumi gation 5-1
References 6-1
Appendices
A - Equipment Leak Control Costs for Ethylene Oxide
Model Plants
B - Dispersion Modeling Parameters Results
-------
LIST OF TABLES
Table Page
1-1 Emission Control Levels 1-4
2-1 Ethylene Oxide Facilities 2-2
2-2 Ethyl ene Oxide Derivatives 2-3
2-3 Ethylene Oxide (EO) Production Sites 2-4
Capacity and Captive Uses
3-1 National Emissions - Ethyl ene Oxide 3-3
3-2 Average Facility Emissions 3-4
3-3 Model Plant Components for (EO) PSroduction Facilities 3-6
3-4 Ethyl ene Oxide Equipment Leak Emissions 3-7
Ethyl ene Oxide Production: Air - Oxidation
3-5 Ethyl ene Oxide Equipment Leak Emissions 3-8
Ethyl ene Oxide Production: Oxygen - Oxidation
3-6 Control Techniques & Emission Rates 3-10
Air Oxidation
3-7 Control Techniques & Emission Rates 3-11
Oxygen Oxidation
4-1 Equipment Leaks - Control Techniques 4-2
4-2 Control Efficiencies: CTG & NSPS 4-3
Equipment Components
5-1 Preliminary Control Efficiency 5-2
and Cost Effectiveness
Ethylene Oxide Emission Sources
5-2 Equipment Leak Control Costs 5-3
Ethyl ene Oxide Production
5-3a Equipment Leaks 5-4
Emissions and Emission Reduction Estimates
Etnylene Oxide Production: Oxygen - Oxidation
5-3b Equipment Leaks 5-5
Emissions and Emission Reduction Estimates
Ethyl ene Oxide Production: Air - Oxidation
5-4 Equipment Leak Control Costs: Ethoxylation 5-6
5-5 Cost Effectiveness - DEOXX™ System 5-7
vi
-------
LIST OF FIGURES
Figure Page
2-1 Basic Operations That May Be Used in Production of 2-7
Ethylene Oxide by Air Oxidation
2-2 Basic Operations That May Be Used in Production of 2-10
Ethyl ene Oxide by Oxygen Oxidation
2-3 Basic Operations That May Be Used in The Production 2-13
of Ethyl ene Glycol, Diethyl ene Glycol and Triethyl ene
Glycol by Conventional Noncatalyzed Hydration of
Ethyl ene Oxide
2-4 Basic Operations That May Be Used in Production of 2-15
Glycol Ethers from Ethyl ene Oxide
2-5 Basic Operations That May Be Used in Production of 2-17
Ethanolamines from Ethylene Oxide
VII
-------
i. INTRODUCTION
1.1 OVERVIEW
This report provides preliminary information on sources of air
emissions of ethylene oxide to assist EPA in regulatory decision-making.
Potentially significant sources are identified and described, including
ethylene oxide production, its captive end-use in ethylene glycol, glycol
ether, and ethanolamine production, linear alcohol ethoxylate production,
and sterilization/fumigation processes. Source descriptions include
general information on each industry and production process, as well
as emission estimates, and control technology.
•
This report also presents emission estimates of EO and control costs
for each emission source. Emission parameters were developed to model
ambient dispersion concentrations of EO; these parameters are summarized
in Appendix B. Emissions were estimated for two control levels: Level 1
represents the baseline level of emission control which would be expected
without further regulatory activity, and Level 2 consists of estimated
emission control for each emission source beyond baseline levels.
The EO emissions, control data, and costs of control in this report
are based on Information obtained from formal Information requests to
operating companies, existing EPA reports and other reports published on
the EO manufacturing industry, and from telephone contacts with control
equipment vendors. Emission estimates are based on- emission data provided
by industry and emission factors developed by EPA.
1-1
-------
1.2 SOURCE ASSESSMENT METHODOLOGY
This section describes the general methodology used for the source
assessment for the EO production, EO captive-feed uses, ethoxylate produc-
tion, and sterilization/fumigation processes. These areas are addressed
below, followed by a brief discussion of the uncertainties involved in
emission and cost estimates.
1.2.1 Ethylene Oxide Production/Capitive-Feed Uses
Detailed emission data and corresponding process parameters from
each EO oxide production site were provided to the EPA from each operator.
EO flow rates (Kg EO/103 Kg EO capacity), both inlet and outlet to control
devices, were developed for each vent source based on plant data. This
information provides control efficiencies, a means of comparing the
relative magnitude of each emission source, and the basis for emission
rates for dispersion analysis.
The cost of equipment leak control assumed the new source performance
standard level of control and cost assumptions; costs are in October 1983
dollars and were estimated on a model plant basis consistent with the
emission estimates. Detailed costs are-provided in Appendix A.
1.2.2 Ethoxylate Production
Data from only one plant forms the basis for emission estimates for
ethoxylation production. It was assumed that the major source of emissions
is equipment leaks. The model plant for ethoxylate producer facilities
was based on comparison of equipment counts from the one ethoxylate plant
and the EO production sites. For emission estimates it was assumed that
1-2
-------
ethoxyl ate producers have about 10 percent of the components of EO
production sites (based on data from one facility); the estimate probably
overstates emissions from these facilities. Data are needed for a more
accurate estimate.
Costs were calculated by averaging the equipment leak control costs
for oxygen and air oxidation plant sites and multiplying by 10 percent; costs
are in October 1983 dollars.
1.2.3 Steri1ization/Fumigation Processes
It was assumed that the typical sterilization and fumigation facility
emits all the EO used in the process to the atmosphere. Emission estimates
were then based on national estimates of the EO consumed in these processes.
The DEOXX™ system manufactured by Chemrox Inc. is the control technique
considered for sterilization/fumigation processes. Information including
control efficiency and costs were provided by the vendor; control costs
are based on one industrial application.
1.3 NATIONAL EMISSIONS
National emissions of EO are estimated at approximately 5,000 megagrams
per year (see Table 1-1). Production/captive use, ethoxylation, medical
facilities, and sterilization/fumigation sites account for 31, 4, 8, and
57 percent of the total emissions, respectively. Emissions associated
with growth in the industry were not estimated.
1.3.1 Production/Captive Use
Equipment leaks contribute approximately 75 percent of EO emissions
at production/captive use sites. Control techniques are summarized in
this document which may reduce these emissions by 42 to 75 percent (see
Table 1-1).
1-3
-------
Cooling towers contribute approximately 13 percent of emissions ana
nay be reduced by approximately 80 percent with appropriate control
techniques.
Vents, storage and loading operations, and effluent treatment facilities
together contribute approximately 10 percent of total production/captive
use site emissions. These sources appear to be wel 1-controlled; additional
control may not be necessary.
1.3.2 Ethoxylatlon
Emissions from equipment leaks may be reduced by 42 to 75 percent,
depending on the degree of control implemented.
1.3.3 Steri1i zati on/Fumi gati on
Add-on control devices may reduce emissions by as much as 99+ percent.
Source
Table 1-1. EMISSION CONTROL LEVELS
Emission Impacts
B
C
Production/Captive Use
1. Equipment Leaks
2. Cooling Towers
3. Vents
4. Storage and Loading
5. Effluents
Ethoxylation
Steri1ization/Fumigation*
Baseline
(Mg/yr)
1,540
1,ZOU
190
100
40
10
200
2,800
Control 1 ed
(Mg//yr)
900 - 500
700 - 3UO
40
100
40
10
85 - 50
32
Reduction
(%)
42 - 68
42 - 75
80
0
0
0
58 - 75
99
4,540
1,020 - 580
78 - 87
*(Excluding medical facilities)
1-4
-------
2. INDUSTRY PROFILE
Twelve companies at 15 locations produce ethylene oxide in the United
States; 2.5 million megagrams were produced in 1983. Most ethylene oxide
is used at the production site to produce ethylene glycols, glycol ethers,
and ethanolamines. Nameplate ethylene oxide captive feed requirements are
80 percent of nameplate ethylene oxide production capacity. Captive feed levels
are approximately 90 percent of actual ethylene oxide production rates.
Of the ethylene oxide that 1s not used by producers (< 12 percent
production), approximately 99.9 percent is sold to detergent manufacturers
where ethylene oxide 1s ethoxylated for use in making surfactants. Less
than .1 percent of ethylene oxide is used in the sterilizer and fumigator
industries.
The number of facilities involved in production and consumption of
ethylene oxide is summarized in Table 1. A complete breakdown of ethylene
oxide derivations is provided 1n Table 2.*
2.1 Production. Name plate capacity for the 15 production sites totalled
3.3 million megagrams (7.3 x 109 Ibs.) 1n 1983. Production site locations,
capacities, and captive uses are summarized in Table 3.2»3
2.1.1. Process Description.4 Ethylene oxide is produced by the direct, vapor
phase oxidation of ethylene over a silver catalyst at 10 to 30 atm. pressure
and 200 to 300°C. The main reaction is as follows:
C^ = Cl^ + 1/2 02 —• CH£ - CHg + 106.7 kJoules* (1)
Ag Cat \ '
0
(ethylene oxide)
Calculated at 250°C, 15 atm.
2-1
-------
Table 2-1
Ethyl en e Oxide Facilities
Facility
Producer (Captive Use)
Sterilizers/Fumi gators
Ethoxyl ators
Number of
Facilities
15
100-45U
50
End-Use
(Mg/year)
2.5 x 106
2.8 x 103
0.4 x 10«
National
Emissions
(Mg)
1.6 x 103
2.8 x 1U3
0.2 x 103
Carbon dioxide and water are the only significant by-products formed,
according to the following reactions:
* 5/2 02 —* 2C02 + 2H20 (3)
(ethylene oxide)
The main reaction also produces small amounts of acetaldehyde (less
than 0.1 percent of EO product) and trace amounts of formaldehyde.
Both the ethyl ene reaction rate and purge losses increase with higher
concentrations of ethyl ene and oxygen. Thus, depending on the process and
the oxldant, the ethylene concentration in the reactor Inlet will vary
between 5 and 40 volumn percent. The Inlet oxygen concentrations will
be controlled between 5 and 9 volume percent to stay out of the explosive
range. Normally, air-based reactors operate at the low end of this '
concentration range while oxygen-based reactors operate at the high end.
By controlling the appropriate operating conditions, modern reaction
systems are able to maintain selectivlties (mols ethylene oxide formed per
"Calculated at 250°C, 15 atm.
2-2
-------
ro
i
TABLE 2-2.
OxiJc Dufivut iwun
Illlllllil bllul U...lli.l»l >
Klliyliriu- tlni-l
_(.'untiiuu|il itiii_
Vzf "
60Z
152
I2Z
5Z
L i „ . Elliyli-mi UXI
ass" .~—ISKT cuiiuuimaion
•>H»I«. |^«»M«««U>«« —
oxide
ion
IMMtflt mt IMt******
IMM*«M*fl%Mj
MIMl f MUMM MM*
.•«MM«i« urri ""rrann iirar
MlaUII* MfU»Mui«4IWHt»«A-
IIMI4 Mill* MJI««|M*I (M« «Ut*H — » »****•• l«l» MM
sz
<5Z
.IZ
•I »* I »l !»»*•* !**—»* •'
• •»•« t.'««l*^«l
UU****iw»**l
fcft HI !«**««_« f«W*l««l..-«« -t
•«••••'— ™ ™ '||itlUutl»U**«« m(iMW«i 111?*****
|.|**d MHH «**«»;«•
,[^u
I (.«».*.. »-l«* *•**••
-------
TABLE 2-3. ETHYLENE OXIDE (EO) PRODUCTION SITES*
CAPACITY AND CAPTIVE USES
1982 Production Capacity (MM Ibs/yr)
Producer
BASF
Celanese
Dow
Oow
ICI Americas
Northern
Petrochemical
Olin
PPG
Shell
Sun Olin
Texaco
Texas Eastman
Union Carbide
Union Carbide
Union Carbide
NAMEPLATE
Locati on
Geismar, LA
Clear Lake, TX
Freeport, TX
Plaquemine, LA
Bay port, TX
Morris, IL
Bradenburg, KY
Beaumont, TX
Gei sma r , LA
Claymont, DE
Port Noches, TX
Longvfew, TX
Ponce, P.R.
Seadrift, TX
Taft, LA '
TOTALS
EO CAPTIVE FEED
Ethyl ene
Oxide
480
425
300
450
500
230
110 .
170
700
110
700
195
650
920
1,300
7,200
6,900
Ethyl ene
Glycol
325
400
300
450
450
350
50
200
350
-
500
180
840
850
1,350
6,600
4,700
Di ethyl ene
Glycol
35
50
30
50
50
35
5
20
30
-
50
20
75
75
120
645
540
Triethylene
Glycol
—
10
50
-
-
-
5
2
10
-
20
2
75
-
-
174
150
Glycol
Ethers
^
-
-
120
-
-
70
13
50
-
40
230
490
_
1,013
1,070
tthanol-
ami nes
—
-
-
125
25
-
30
-
-
-
240
-
-
230
-— . i_ -
650
470
CAPACITY REQUIREMENTS
*As reported in the literature.
-------
lOu rnols ethylene reacted) between 65 and 75 percent for air-based processes
and between 70 and 80 percent for oxygen-based processes. The selectivity range
for the air process is lower because of the purge reactors which, of necessity,
operate with low ethylene concentrations and at high conversions per pass.
These high reaction efficiencies are achieved by operating with low ethylene
conversions per pass and adding controlled, trace amounts of chlorinated
inhibitors to moderate the reaction.
Since the reactor exit gas contains substantial amounts of unconverted
ethylene, it must be recycled back to the reactor to achieve a high over-
all ethylene conversion in both the air and oxygen-based processes. The
exit gas to be recycled is first scrubbed using water to absorb the ethylene
oxide, which if not removed from the feed, would inhibit the reaction and
also result in excessive oxidation losses (to C02 and H20).
In the air-based process, a considerable amount of nitrogen is introduced
1n the air feed to the reaction. Accordingly, a large portion of the scrubbed
reactor exit gas must be withdrawn to remove this nitrogen. This purge
stream is usually mixed with additional air to Increase the oxygen concen-
tration and then passed through a secondary (or purge) reactor, similar to but
smaller than the first (or main) reactor to convert most of the contained
ethylene. In large plants, a third stage of reaction is used to achieve an
overall ehtylene conversion well in excess of 95 percent of the total feed.
In the oxygen-based process, since high-purity oxygen is used, much less
inart gas is introduced into the reaction system. As a result, only a limited
purging 1s required and the unconverted ethylene can be more completely
recycled. To avoid a build-up of by-product carbon dioxide, a portion of the
recycled gas 1s first sent to a carbon dioxide removal system where C02 is
2-5
-------
preferentially absorbed and vented with a minimum loss of ethylene. The
reduced purge permits the 02-based reactor to operate at a high ethylene
concentration without excessive vent losses and allows the use of different
diluent gases.
In the air based process, nitrogen must be used as the diluent for the
oxygen and ethylene since any other diluent would be vented together with
the nitrogen. The oxygen process can, however, use other diluents (such
as methane) which allow the safe use of high concentrations of both
ethylene and oxygen, thus leading to increased productivity at high
selectivities.
Both the air- and oxygen-based processes use shell and tube reactors
with the catalyst contained in the tubes and a cooling medium in the shell.
Boiling water is now the preferred shell-side coolant for new reactors
because of improved safety over non-boiling heat transfer oils and boiling
kerosene, which were used in the past.
2.1.2 Air Oxidation.5 Figure 2-1 1s a typical flow diagram for a continuous
air-oxidation process. Ethylene and air (SI and S2) are added to a recycle
ethylene stream (S3), which feeds one or more primary reactors operated
in parallel. The airiethylene feed ratio, usually about 10:1, 1s varied
with the recycle gas to ensure 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 maintained at 220 to 280°C and 1 to
3 MPa (42>° to 536°F; 10-30 atm). The unreacted ethylene is separated
from the reaction products and recycled through the reactor until consumed.
The effluent from the primary reactor (S4) is cooled by the recycle
stream from the main absorber (S3) to about 38°C 1n a shell-and-tube heat
*5i = Steam 1
2-6
-------
~!
STEAM _tocxjujr
" * ' f'fc^v*———i
n
PURG.t
- FUGITIVE eMi*t»iOKj-we*Mx PLWI
fe - S&COMDMW £MI*;»IOHJ POT&JIIAI.
eiMvceue
OKioe..
STORAC^C.
ii .•
Figure 2-1. Production of Ethylene Oxide by Continuous Air-Oxidation Process5
-------
exchanger; it is then compressed before entering the main absorber. The
effluent 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 (S5).
Unabsorbed gas passing overhead from the main absorber is split into
two unequal portions. The larger portion (S3) recycles through the
reactor effluent cooler and joins fresh reactor feed. The smaller portion
(S6) 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 (S7) 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 (VA) is 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 volatile organic chemicals from both absorbers are combined
(S8). The mixture is fed to the top of the desorber, where the crude
ethylene oxide (S9) is distilled off the top and compressed for further
refining. A stripper removes carbon dioxide an'd inert gases overhead
(VB), and the EO, stripped of carbon dioxide (S10), is fed to the midsection
of the refiner, where it is distilled overhead to 99.5 mole % EO. The
product (Sll) is stored under a nitrogen atmosphere. The secondary
reaction of ethylene that produces C0£ 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.
2-8
-------
2.1.3 Oxygen Oxidation.6 Figure 2-2 is a typical flow diagram for a
continuous oxygen-oxi dati on process. A higher concentration of ethyl ene
allows more ethylene to be converted per pass without exceeding the 30%
ethyl ene 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 conversion. The higher
ethylene conversion eliminates the need for the purge-reactor absorber
system required by the air-oxidation process.
As shown by the oxygen-oxi dati on flow diagram, Fig. 2-2, the purge
reactor and purge absorber of the air-oxidation process are replaced by a
C02 absorber and reacti vator. EO, along with some C02, is dissolved in
the water stream leaving the base of the main absorber (S6) 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 (S8, Fig. 2-1).
Part of the unabsorbed gas overhead from the mai n absorber of the
oxygen-oxidation process (S7) passes through a C02 absorber before being
recycled back to the reactor feed (S8). Carbon dioxide must be removed
to maintain favorable catalyst activity and favorable conversion to EO.
The C02 absorbent, usually potassium carbonate '(S9), is heated by the
bottoms from the reactivator and then fed to the top of the reactivator,
where it is stripped of carbon dioxide and recycled to the C02 absorber.
Small amounts of gaseous impurities in the feed, such as argon, must
be removed since they will accumulate in the closed system. Some of the
recycle gas stream is purged through the argon purge vent (VB). This
emission flow rate is automatically regulated by the argon concentration.
2-9
-------
C0t PURC^G
(o)- Puc\iTtve
oveR^^.^. PLM4T
0-SECOMOM\V CIAISMOVI
POTEM1IM.
WATCR
f
Figure 2-2. Production of Ethylene Oxide by Continuous Oxygen-Oxidation Process"
-------
2.2 Captive Feed Use5,6,7
2.2.1 Process Description/Ethylene Glycol. Ethylene glycol 1s manufactured
by the addition of water to ethylene oxide (EO) In a noncatalyzed pressure
hydration process. The ethylene glycol (EG) formed will react with additional
ethylene oxide to form diethylene glycol (DEG), trlethylene glycol (TEG),
and other higher homologs. The chemical equations are as follows:
CH2 ^_ _ _ _ CH2OH
CH2 ^
(ethylene oxide)
H20
(water)
CH2OH
(ethylene glycol)
(1)
CH2OH
CH2OH
(ethylene glycol)
CH2
CH2OH
CH2 CH2-0-CH2CH2OH
(ethylene oxide) (diethylene glycol)
(2)
CH2OH
CH2-0-CH2CH2OH
(diethylene glycol)
CH2
CH2
(ethylene oxide)
CH2-0-CH2CH2OH
(3)
CH2-0-CH2CH2OH
(trlethylene glycol)
The normal weight ratios of co-products formed are 87 to 88.5 wt. percent
ethylene glycol, 9.3 to 10.5 wt. percent diethylene glycol, and 2.2 to
2.5 wt. percent trlethylene glycol. These three products constitute an
overall yield of 92.5 to 95.5 percent of theoretical, based on the ethylene
oxide feed.
2-11
-------
Figure 2-6 represents a typical noncatalyiea ethylene oxide Mydracion
process. The continuous process is carried out in the liquid phase, ana
tne reactions are strongly exothermic. Theoretical ly, 0.7i kg of EG 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.
Refined liquid EO (SI), makeup water (S2), and recycle water are
mixed under pressure (1.3bO kPa), preheated, ana fed to the hydrolyzer.
The feed solution (S3) contains approximately 8 kg of water per kg of EO.
The reactor effluent, heated by the exothermic heat of hydration, exists
(S4) the hydrolyzer 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
(S5) to remove light impurities from the system. The remainder of the vapor
and the vapors from the remaining evaporator effects are conuensed and
recycled (S6). The evaporator calandria and the condenser on the final
evaporator effect are vented (A) to remove noncondensable gases.
2.2.2 Process Description/Glycol Ethers. The ethyl ene glycol monoethers
are produced by the following sodium hydroxide catalyzed chemical reactions.
Only the reactions with methoinol are sho.wn; however, the primary alcohols,
ethanol . and butanol , react similarly to produce the ethyl and butyl
ethers.
CH20H + CH2 - CH2 - >
(methanol) (ethyl ene oxide) (ethyl ene glycol monomethyl ether)
2-12
-------
on.«od
VJMCft
•t;
pun*t i
V«Mt /
8N
m
Into
VIA
fcuC. h
ien |/
1 f- -v*
1/-J
^->Sllh
^
, —
h^
^
Hf OROCf t C R
1 M»
i
WATER
REMOVAL.
VAC
It-
IDS
tov*
COOUIJ^
TCKNER
''t <
^
A \
ToUX*J-
a AOOWM
Figure 2-3. Production of Ethylene Glycol by Noncatalyzed Hydration of Ethylene Oxide5
-------
(ethyl ene glycol (ethyl ene oxiae) (diethylene glycol monomethyl etner)
monornethyl ether)
+ Che — CH£
(dietiiylene glycol mono- (ethyl ene oxide) (triethylcne glycol monoetnyl ether)
methyl ether)
Figure 2-4 represents a continuous process for the manufacture of the
glycol ethers. The sodium hydroxide catalyst (SI) and one of the anhydrous
primary alcohols (S2) - are blended in the mix tank. The material from
the alcohol catalyst storage tank is combined with ethylene oxide (S3)
and with the recycled alcohol (S4) and is then ted to the reactor. The
exothermic reaction is carried out at an elevated pressure (2.5 x 10& to
4.6 x 106 Pa) and temperature (200 to 230°C).
The product (S5) exits the reactor and is sent to the alcohol
distillation column, where excess alcohol is distilled overhead and
recycled (S6) for future reaction. The alcohol column bottoms (S7) are
then sent to the monoethylene glycol ether column, where monoethylene
glycol ether is vacuum distilled and sent (S8) to product storage.
Diethylene glycol ether and triethylene glycol ether are vacuum distilled
consecutively in two more distillation columns. The diethylene glycol
ether product (S9) and triethylene glycol ether product (S10) streams are
sent to their respective storage tanks.
2-14
-------
:, 5
Figure 2-4. Production of Glycol Ethers from Ethylene Oxide
-------
2.2.3 Process Description/Ethanolamines. Ethanolamines ara produced by
the following series:
wrg
(ammonid)
NH2CH2CH2CH
CH2 — CH2 - :
(ethylene oxide)
(monoethanolamine (MEA))
CH2 - CH2 - > NH(CH2CH2OH)2
(MEA)
(ethylene oxide) ( di etnanol ami ne
MH(CH2CH2OH)2
CH2 CK2 > N(CH2CH2OH)3
(DEA) (ethylene oxide) (triethanolamine (TEA);
Tha process is noncatal-ytic, strongly exothermic, and is carried out in the
liquid phase in the presence of water. The distribution of products is
dependent on the ratio of ammonia to ethylene oxide used; excess ammonia
favors MEA formation.
The continuous manufacture of ethanol amines is shown schematically in
Figure 2-5. Ethylene oxide (SI) and aqueous ammonia (S2) are fed to a
reactor. The reaction conditons usually are a temperature range of 50 to
100°C, a pressure of 1 to 2 MPa, and an.excess of 28 to 50 percent aqueous
ammonia. The reactor effluent (S3) is stripped'of unreacted ammonia ana
some water (S4) in an ammonia stripper operated under pressure. This
ammnionia, togetiiar with fresh faed (S5), is absorbed in recycled watsr in
the ammonia absorber and fed back to the reactor (S2). The noncondensable
overhead gas (S6) from the ammonia stripper is scrubbed of amonia in an
2-15
-------
IJ
I
OK 10II
,©
:r
AMI 4Okl I *.
,-Lr.~\
o
l
U
U»O
Ol-
E.lll
COl-t'MU
-v..
©
1TZI-
ETI4A.UOI.AMIIJC.
COUUMM
DISPOSAL.
OeVIYDRATIOUr,
COUUMU
• MOJO-
HZh ®
I / VAC-
MOUO-
CTHAJJCLAMlue
COLUMU
-MM-
ETI ( A.W1 OV-AV Ai k.| r.
2--r>. Production of KUinnolnmtncn by tlio Kthylcnc Oxide—Ammonln Pf
-------
ammonia scrubber wi t:i recycle water (S7) ana is vented (A). Inert gases
enter the system with the ethylene oxiae feea, which is storea under a
nitrogen pressure pad.
The ammonia stripper bottoms (S9) are vacuum aistilled in a series
of distillation columns to sequentially remove overhead water (S7), which
is recycled, and MEA, DEA, and TEA (Slu, li, 12), which are products.
Noncondensables 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 (S13) 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.
2.3 Ethoxylation.3 Detergent alcohol ethoxylates are produced by reacting
detergent linear alcohols with ethylene oxide in the presence of a base
catalyst such as potassium hydroxide. The general reaction may be represented
as follows:
KOH
ROH + n CH2 CH2 > RO(CH2CH20)nH
^*'
(ethylene oxide)
The molar ratio of ethyl ene oxide to alcohol in the final product may vary
from 2 to 4U. Ethoxylates produced for subsequent conversion to alcohol
ether sul fates usually contain 3 moles of ethyl ene oxide per mole of
alcohol. Products made for direct use as nonionic surfactants usually
contain 6 to 12 moles of ethylene oxide per mole of alcohol.
2-18
-------
2.4 Sterlizatlon/Fumigation. The type of equipment used for EO fumigation/
sterilization varies with the application. The basic types of equipment
are discussed in the following sections.
Vacuum Chambers. Vacuum chambers are pressure vessels with a vacuum
pump to remove air from the chamber before sterilization begins and to
remove some of the EO/air mixture after sterilization. Though the units
vary widely in size and design features, the operating procedure is
essentially as follows:
1. Contaminated material is loaded into the chamber.
2. The hermetically sealed door is closed.
3. Air 1s vacuumed from the chamber.
4. The sterUant pure (100%) EO, 12 percent EO/88 percent Freon, or
10 percent EO/90 percent carbon dioxide 1s Introduced into the chamber
to a- set pressure or concentration and for a specified time period. Pure
(100%) EO is-used with negative pressure; EO mixtures are used with
positive pressure. Pressure, concentration of sterilant, and time period
are adjusted for the Individual situation.
5. An exhaust vacuum removes the EO or EO/gas mixture from the chamber.
The EO or EO/gas mixture is vented through a vent line to the atmosphere
or to a sewer drain.
6. Fresh air 1s drawn Into the chamber until atmospheric pressure 1s
reached.
7. The door 1s opened and the treated material removed.
8. The treated material may be transferred to an aeration cabinet which
circulates heated air around the material until residual EO has escaped.
(Aeration cabinets are used almost exclusively 1n hospitals).
2-19
-------
Small counter-top models with capacities less than O.lm3 (<4ft3) are
most commonly used in health care and health diagnosis facilities. In
hospitals they are used in areas such as operating rooms. One industrial
use is in the manufacture of contact lenses. EO is supplied either in
single-dose cartridges of pure (100%) EO or 1n pressurized cylinders of
12 percent EO/88 percent Freon. Small chambers generally vent directly
into the atmosphere through a length of tubing. Some models vent into a
sponge kept damp in a bucket of water.
Intermediate-sized chambers of from 0.1 to 2.8m3 (4-100 ft3) are
used primarily in hospital central supply facilities. They are also used
in research and industrial facilities, libraries, museums, and beehive
fumigation facilities. An EO mixture Is supplied in pressurized cylinders.
Intermediate-sized chambers of this type may vent emissions to the
atmosphere or the emissions may be mixed with water which is routed to a
sewer drain.
Large chambers with capacities greater than 2.8 m3 (100 ft3) are used
primarily for industrial sterilization of medical products, spices, and
other products. They may be as large as 85 m3 (3000 ft3) 1n capacity and
are custom-made. In such large capacity custom chambers, an EO mixture
or pure (100%) EO is fed from pressurized cylinders or from large tanks.
Emissions from large sterilizers are typically evacuated using once-through
water-ring vacuum pumps. For these pumps, the amount of ethylene oxide
absorbed In the water 1s generally less than 30 percent due to the
relatively low gas retention time. Ethylene oxide absorbed In the water can
desorb to the workplace above the break between the liquid pipe and drain.
2-20
-------
3. NATIONWIDE EMISSIONS
Chapter 3 provides nationwide ethylene oxide emission estimates from
each of the processes described in Chapter 2. Estimates are provided for
the following emission sources: Equipment leaks (valves, pumps, compressors,
etc.) cooling towers, vents, storage and loading operations, ethoxylation,
steri lization/fumi gati on, and effluents. These estimates represent base-
line (present) emission levels.
Detailed emission data and corresponding process parameters from each
ethyl ene oxide production site were provided to the EPA from each operator.
Information was provided including emission stream characteristics (flpw
rate, temperature, composition, etc.) to and from control devices and
vents to atmosphere where no device was used. Data were provided for all
process vent emission sources, storage and loading operations, and effluents
for the entire plant site including captive uses. Equipment component
counts for pump seals, compressors, flanges, valves, pressure relief
devices, sample connections, and open ended lines were provided for all
equipment handling streams with any ethyl ene oxide present.
The equipment leak component counts from all the plants were used to
develop model plants for both the oxygen and air oxidation processes. Emission
factors for the synthetic organic chemical manufacturers industry (SOCMI) were
applied to these model plants to establish baseline emissions and controlled
emission levels. This approach assumes standard industry practice for control
of equipment leaks for the baseline case.
National emissions for cooling towers, vents, storage and loading facil-
ities, and effluents for production facilities are the emissions reported
by the plants to EPA. Equipment component leak emissions were estimated
using EPA emission factors. Ethylene oxide flow rates (kg E.O./103 kg E.O.
3-1
-------
capacity), to and from control devices, were developed for each emission
source based on plant data. This information provides control efficiencies
and a means of comparing the relative magnitude of each emission source.
Data from only one plant forms the basis for emission estimates for
ethoxylation produciton. It was assumed that the major source of emissions
is from leaking equipment. The model plant for ethoxylate producer
facilities was based on comparison of equipment counts from the one
ethoxylate plant and the ethylene oxide production sites. For rough
emission estimates it was assumed that ethoxylate producers have about 10
percent of the components of ethylene oxide production sites; the estimate
probably overstates emissions from these facilities. Data 1s needed for
a more accurate estimate.
It was assumed th'at the typical sterilization and fumigation facility
vents all the ethylene oxide used in process to the atmosphere. Emission
estimates were then based on national estimates of the ethylene oxide
consumed in these processes.
Total nationwide ethylene oxide emissions from production, ethoxylation,
and fumigant/sterilant uses are summarized 1n Table 3-1. Approximately 5,000
metric tons of ethylene oxide are being emitted per year. The average
•
producer, sterllizer/fumlgator, medical facility, and ethoxylator facilities
emit approximately 100, 6 to 28, 0.06, and 4 metric tons per year of ethylene
oxide respectively (see Table 3-2).
3.1 Production Facilities
Emission sources at production facilities Include equipment leaks, cooling
towers, vents, storage and loading, and effluent treatment facilities.
3-2
-------
TABLE 3-1. NATIONAL EMISSIONS .
ETHYLENE OXIUE
(Mg/Yr)
Emissions
Source
A. Ethylene Oxide Production Facilities1
1. Equipment Leaks
2. Cooling Towers
3. Vents
4. Storage & Loading
5. Effluents
Subtotal
Subtotal
b. Surfactant Production
(Equipment Leaks)
C. Funrfgant/Sterilant Use*
Subtotal
TOTAL
Mg/Yr
1,18U
190
100
40
10
i,520
< 200
2,60G-3,9003
% Subtotal
7tt
13
6
3
< 1
100
100
100
4,320-5,620
1 Includes captive use of ethylene oxide in production of (a) ethyl ene,
aiethylene, and triethylene glycol, (b) glycol ethers, (c) ethanolamines
(d) polyethylene glycol and nonionic surface active agents
2 Assumes all ethyl ene oxide is emitted to atmosphere
3 Reinhart, Joe. Officer of Pesticides Programs, U.S. Environmental Protection
Agency. Memo to Carolyn Smith, Radian Corporation, May 19b3.
3-3
-------
Table 3-2. AVERAGE FACILITY EMISSIONS
(Mg of EO/yr)
Facility
Production
Steril izers/Fumi gators
Medical Facilities
Number of
Facilities
15
10U-450
6,000-10,000
National
Emissions
(Mg)
1,520
2,BuO
400
tiTii ssions Per
Facility
(Mg)
100
6-2b
<0.0fa
Surfactant Production
5U
200
Total 4,920
3-4
-------
3.1.1 Equipment Leaks
Equipment leaks are those emissions that result when process fluid (either
liquid or gaseous) leaks from plant equipment. There are many potential sources
of equipment leaks in a typical synthetic organic chemical plant. The following
sources are considered in this report: pumps, compressors, in-line process
valves, pressure relief devices, open-ended valves, sampling connections, and
flanges. Emissions which result from leaks in these types of equipment are
generally random occurrences which cannot be predicted.
A complete description of each type of equipment leak is addressed in the
background information document: "VOC Fugitive Emissions in Synthetic Organic
Chemicals Manufacturing Industry - Background Information for Proposed Standards,"
EPA-450/3-80-033a, November 1980. The number of equipment components subject
to leakage for both the oxygen and air oxidation model plants are summarized
in Table 3-3. Emission factors, component emissions, ana total model plant
emissions are summarized for both oxygen and air oxidation model plants
in Tables 3-4 and 3-5.
3.1.2 Cooling Towers
Cooling towers are found in the oxygen-oxidation plants where more heat
is released than in the air oxidation process. The purpose of these towers
is to cool the plant's recirculated effluent from the ethylene oxide stripper
column. The cooling process is achieved by evaporation when the process
couling water and air are contacted. The emission rate is inversely propcrtiorial
to the efficiency of the stripper column 1n removing ethylene oxide. Approximately
0.07 Kg of ethylene oxide are emitted for each 1000 Kg of ethyl ene oxide capacity.
3-5
-------
Table 3-3. Model Plant Components for Ethylene Oxide Production Facilities
Number of Components 1
0? - Oxidation2
Air - Oxidation3
Pump Seals
Compressors
Flanges
Valves
Press. Rel. Dev.
Sample Conn
Open Ended Lines
Packed
Mechanical
D - Mechanical
Gas
Liquid
Gas
Liquid
Gas
Liquid
Gas
Liquid
16
1U
3
1738
177
693
26
15
18
22
38
122
11
Ib
6
1214
200
581
60
40
55
90
70
122
1 Includes captive feed processing components
2 Average of six facilities component counts: Average EO capacity = 1 x 105 mtons/yr
3 Average of five facilities component counts: Average EO capacity = 3.6 x 105 mtons.
3-6
-------
Table 3-4. Ethyl ene Oxide Equipment Leak Emissions
Ethylene Oxide Production: Air-Oxidation
Model Plant
Emissions
Source
Pump Seals
Light Liquid
Heavy Liquid
Valves
Number of
Sources
27
Emission Factor a Adjustment Annual
(kg/day/source) Factor" Emissions, (Mg/yr)c
1.19
.5
5.8
Gas
Light Liquid
Heavy Liquid
Safety/relief valves
Gas
Open-ended lines
Compressors
Sampling connections
Flanges
TOTAL
2UO
581
60
192
6
145
1,214
0.13
0.17
2.5
0.04
5.47
0.36
0.02
.5
.5
.5
.5
.5
.5
.5
4.8
18.0
27.4
0.0d
5.9
2.4e
4.4
68.7
a"Control of Volatile Organic Compound Leaks from Synthetic Organic Chemical and
Polymer Manufacturing Equipment," EPA-450/3-83-006, March 1984, p. 2-21.
bThe emission factor assumes 100 percent organic in stream and must be adjusted
for ethyl ene oxide emission estimates to account for the different percent of
ethylene oxide in various streams. Assume the average stream contains 50 percent
ethyl ene oxide.
cFor estimating purposes, one operating year was assumed to be 365 days.
dAll open-ended lines (in model unit) are assumed to be controlled at Daseline;
therefore, there are no associated emissions.
eSeventy-f1ve percent of sampling connections are assumed to be controlled at
baseline; therefore, the emissions are based on 36 of 145 sampling connections.
3-7
-------
Table 3-5. Ethyl ene Oxide Equipment Leak Emissions
Ethylene Oxide Production: Oxygen-Oxidation
Model Plant
Emissions
Source
Pump Seals
Light Liquid
Heavy Liquid
Valves
Number of
Sources
30
Emission Factor a Adjustment Annual
(kg/day/source) Factorb Emissions, (Mg/yr)c
1.19
.5
6.5
Gas
Light Liquid
Heavy Liquid
Safety/relief valves
Gas
Open-ended lines
Compressors
Sampling connections
Flanges
TOTAL
177
693
26
160
3
40
1,738
0.13
0.17
2.5
0.04
5.47
0.36
0.02
.5
.5
.5
.5
.5
.5
.5
4.2
21.5
11.8
0.0d
2.9
0.66
6.4
53.9
a"Control of Volatile Organic Compound Leaks from Synthetic Organic Chemical and
Polymer Manufacturing Equipment," EPA-450/3-83-006, March 1984, p. 2-21.
bThe emission factor assumes 100 percent organic in stream and must be adjusted
for ethyl ene oxide emission estimates to account for the different percent of
ethyl ene oxide in various streams. Assume the average stream contains 50 percent
ethyl ene oxide.
cFor estimating purposes, one operating year was assumed to be 365 days.
dAl 1 open-ended lines (in model unit) are assumed to be controlled at baseline;
therefore, there are no associated emissions.
eSeventy-five percent of sampling connections are assumed to be controlled at
baseline; therefore, the emissions are basea on 5 of 40 sampling connections.
3-8
-------
3.1.3 Vent Emission Sources
3.1.3.1 Air Oxidation
The main process vent is the largest vent source of ethylene oxide
emissions in the air oxidation plant. The vent gases also contain nitrogen,
unreacted oxygen from the air feed, ethane and ethylene from the ethylene feed,
and by-product C0£. The main process gases are vented to a thermal or catalytic
oxidizer; inlet and outlet (to atmosphere) mass flow rates of ethylene oxide are
summarized in Table 3-6.
The ethylene oxide stripper vent releases inert gases and ethylene
which were absorbed into the main and purge absorber waters. The stripper vent
stream is combusted in a boiler resulting in approximately 100 percent control
of ethylene oxide.
In some plants, vent streams are routed to an absorber to recover ethylene
oxide. The vent stream from the absorber is then fed to a boiler or plant
flare; emission levels are approximately zero.
3.1.3.2 Oxygen Oxidation
The volume of the main process vent or argon vent of the oxygen oxidation
process is much less than that of the corresponding vent in the air process.
This vent contains mainly argon and nitrogen from the oxygen feed, ethane from
the ethylene feed, and very small amounts of .ethylene oxide (< 0.01 mole %).
The ethylene content of the main process vent is sufficient to support
combustion and is routinely routed to a boiler or incinerator; ethylene
oxide emissions are approximately zero from this source. Inlet and
outlet mass flow rates of ethylene oxide are summarized in Table 3-7.
3-9
-------
t->
O
Principal Source
VENT SOURCES:
A. Main Process
B. Vent Absorber
Table 3-b. CONTROL TECHNIQUES & EMISSION RATESa
AIR OXIDATION
Control Units
a. Thermal Oxidation
b. Catalytic Oxidation
a. Boiler
b. Plant Flare
Control
Efficiency
80
100
100
99
E.O.: Flow Rate
(kg/ltPkg Capacity)
Control Unit
Inlet
.2
.005
Avg. .1U
4 x 10-5
0
Avg. 4 x
Outlet
.04
0
0
C. E.O. Stripper
Boiler'
100
unknown
unknown
aBased on data provided by ethylene oxide producers
-------
Table 3-7 . CONTROL TECHNIQUES & EMISSION kATESa
UXYtiEM OXIDATION
E.O.: Flow Rate
Principal Source Control Units
I. VENT SOURCES:
A. Argon a.
b.
c.
B. C02 Stripper a.
b.
c.
C02 Regenerator a.
C. E.O. Stripper a.
b.
c.
E.O. Reabsorber a.
b.
E.O. Purification a.
b.
c.
Boiler
Boiler
Incinerator
Carbonate Flasher and
Vent Condenser
C02 Recovery
Methanol Unit
N.A. (to sale)
Uncontrolled
Recylce to reabsorber
Cooling tower
Recycle
Incineration
f^O Scrubber
Vent absorber
Recycle to reabsorber
Control
Efficiency
100
100
99.9
92
100
100
100
N.A.
0
100
99.95
86
yf*.99+
(kg/KPkg Capacity)
Control Unit
Inlet
.0011
.0016
.002
Avg. .0016
.75
.00*
.00063
.002
Avg. .19
.04
unknown
.002
.065
unk
11.4
—
Avg. 2.9
Outlet
0
0
0
0
.06
0
0
.06
unknown
0 or .002d
0
trace
.0004
—
.0004 - .0024
aBased on data provided by eth>lene oxide producers
-------
More than 99 percent of the C02 from the C02 absorber is processed to
recover C02- Also, a carbonate flasher followed by a vent condenser is
being used to reduce atmospheric emissions. The average ethylene oxide
emission rate from the C02 vent is approximately 0.06 Kg per 1000 Kg of
ethylene oxide capacity (Table 3-7).
The ethylene oxide stripper purge vent is composed of inert gas,
ethylene, and ethylene oxide that were dissolved in the main absorber water
during the recovery of ethylene oxide from the reaction gases. After
ethylene oxide is desorbed from the water, the purge vent gases are stripped
from the ethylene oxide. Typically, ethylene oxide in this stream is scrubbed
with water via an absorber and recyled to the process.
An alternative to stripping inerts from the ethylene oxide stream is to
vent these gases from the reabsorber tower when ethylene oxide is reabsorbed
in water. The process sequence is reactor, scrubber, stripper (ethylene oxide
stripped from water), and ethylene oxide reabsorber. The vent stream from the
reabsorber can be incinerated.
Inert gases are also purged from the ethylene oxide purification tower.
This stream can be scrubbed, vented to an absorber, and recycled to a reabsorber;
atmospheric emissions from this vent are negligible (Table 3-7).
3.1.4 Storage & Loading
Ethylene oxide is a gas at ambient termeratures. Therefore, ethylene
oxide is stored at approximately 10°C in insulated or underground tanks to
maintain it as a liquid. Tanks are generally blanketed with nitrogen for
safety purposes. Vapors displaced during storage and rail car loading
operations are recycled to the process or scrubbed prior to flaring or incin-
eration at most facilities. Emissions from storage and loading are reduced
to negligible amounts with the exception of one producer facility where a
3-12
-------
caustic scrubber is utilized emitting approximately 39 Mg per year.
3.2 Ethoxylation
Most ethoxylation facilities produce surface active agents (surfacants).
Ethylene oxide 1s received by pressurized railcar, stored, then reacted
with various chemicals in a closed process. The primary source of ethylene
oxide emissions is assumed to be equipment leaks, although minimal data has
been collected. Other emissions sources at these facilities are assumed to be
negligible. Control techniques specified in both the synthetic organic chemical
manufacture industry (SOCMI) fugitive emissions NSPS and CT6 would effectively
reduce ethylene oxide emissions by approximately 75 and 40 percent respectively.
It was assumed that component counts at these facilities (minus flanges)
were approximately 10 percent of those at ethylene oxide production facilities.
3.3. Fumi gati on/Steri11zatlon
Ethylene oxide emissions result from fumigation and sterilation operations
when chambers are purged to the atmosphere. Emissions from these facilities 1s
based on estimates of ethylene oxide used by these facilities. Approximately
2,800 metric tons of ethylene oxide is used 1n these Industries and ultimately
vented to the atmosphere.
3-13
-------
4. CONTROL TECHNIQUES
4.1 Equipment Leaks
Primary control methods for equipment leaks Include leak detection and
repair (LDAR) methods and control equipment. Leak detection methods are used
to Identify equipment components that are emitting significant amounts of
volatile organic chemicals. Emission from leaking sources may be reduced by
three general methods: repair, modification, or replacement of source. In
the case of open-ended lines, however, equipment leaks are treated more
effectively by Installation of control equipment.
The U.S. Environmental Protection Agency has proposed new source standards
(NSPS) and published control guidelines (CTG) for new and existing facilities in
the synthetic organic chemical Industry. Both the NSPS and CTG are applicable
to ethylene oxide production and ethoxylator facilities. A summary of the
CTG and NSPS control levels are Included 1n Table 1 for comparison. The CTG and
NSPS background information document, can be consulted for a detailed description
of the Individual component control techniques (see Table 4-1 footnotes).
Control efficiencies for Individual equipment components under the CTG and
NSPS level of control are provided In Table 4-2.
4-1
-------
Table 4-1 EQUIPMENT LEAKS
CONTROL TECHNIQUES
Emission
Source
Control Technique
CTG1NSPS2
Pump Seals (Light Liquid)
Valves (Gas)
(Light Liquid)
LDAR3
(quarterly monitoring)
LOAR
(quarterly monitoring)
LDAR
(monthly monitoring)
LDAR
Uonthly monitoring)
Safety/relief valves
(Gas)
LDAR
(quarterly monitoring)
Performance
standard*
Open-ended Lines
Compressors
Sampling connections
Caps
LDAR
(quarterly monitoring)
none
Caps
Seal System
Closed Purge
System
1 "Control of Volatile Organic Compound Leaks From Synthetic Organic Chemical and
Polymer Manufacturing Equipment," EPA-450/3-83-006, March 1984.
2 "voc Fugitive Emissions 1n Synthetic Organic Chemicals Manufacturing
Industry - Background Information for Proposed Standards, "EPA-450/3-80-033a,
November, 1980.
3 LDAR - Leak Detection and Repair Model
4 No detectable emissions
4-2
-------
Table 4-2. Control Efficiencies: CTG & NSPS
Equipment Components
Emissions
Source
Pump Seals
Light Liquid
Valves
Gas
Light Liquid
Safety/relief valves
Gas
Open-ended lines
Compressors
Sampling connection
Emission Factor3
(kg/day/spurce)
(uncontrolled)
1.19
0.13
0.17
2.5
0.04
5.47
0.36
Present Control
Efficiency
CTG NSPS
0.33
0.64
0.44
0.44
1.0
U.33
0.0
U.61
U.7J
0.59
1.0
1.0
1.0
1.0
4-3
-------
4.2. Cooling Towers
Two basic techniques are available for controlling cooling tower emissions.
Emissions can be reduced by either using a sealed heat exchanger or using a
more efficient stripper column(s).
4.2.1. Heat Exchanger
A heat exchanger used in place of a cooling tower would virtually eliminate
ethylene oxide emissions. It was assumed that the incremental cost of utilizing
a heat exchanger in place of a conventional cooling tower is prohibitive.
Therefore, heat exchangers are not considered a viable approach to reducing
cooling tower emissions.
4.2.2. Stripper Column Design
4.2.2.1. New Sources
Recirculated effluent from the ethylene oxide stripper column to the
cooling tower results in ethylene oxide emissions when water contacts the
atmosphere. Emission can be reduced by utilizing higher efficiency trays and
packing internals to Increase stripper column effectiveness. More efficient
stripper columns require less energy for processing; this energy saving may
off-set the cost of additional trays or more costly packing.
4.2.2.2. Existing Sources
Although existing stripper columns can be redesigned to incorporate
additional trays, it 1s most likely too costly. Because packing trays are welded
in place, a retrofit of an existing column requires cutting out olo trays to
replace with new ones.
A stripping efficiency of existing towers can be increased by replacing
packing materials with more efficient packing.
4-4
-------
4.2.2.3. Emission Reduction
An emission reduction of roughly 80 percent appears possiole for new
facilities utilizing both optimum tray and packing design. Existing facilities
with only new packing intervals would experience less than an 8U percent
emission reduction.
4.3 Vents
As discussed in section 2.1.3., nearly all vent streams rrom etnylene
oxide production facilities are recycled to the process or combusted. In the
oxygen oxidation process the C02 stripper or regenerator vent is processed
to recover COg. Present (baseline) control levels are summarized in Table 6
and 7 of Section 3.
4.4 Sterilization and Fumigation
Ethylene oxide emissions vented from sterilizer and fumigator chambers
may be combusted or processed for recovery.
4.4.1 DEOXX™.
The DEOXX™ system was designed by Chemrox, Inc. to chemically process
ethyl ene oxide vapors to saleable ethyl ene glycol. The system has been
certified by New York State as best available control technology (BACT)
ana appears to have established the 99 percent efficiency level as BACT
in New York. The DEOXX™ system is under patent review. DEQXX™ ethylene
oxide emission control systems have been or are scheduled to be installed
in the states of New York, Maryland, Michigan, Utah, Hawaii, California,
Florida, Connecticut, Texas, Rhode Island, Illinois, Pennsylvania,
Puerto Rico, South Carolina, and Minnesota.11^
4-5
-------
5. CONTROL TECHNIQUE COSTS
Annualized control costs, emission reductions, and cost effectiveness for
each emission source are summarized in Table 5-1. The cost analysis was calculated
on a model plant basis for equipment leak sources.
5.1 Equipment Leaks
5.1.1. Ethylene Oxide Production
Total annual equipment leak control costs are provided in Table 5-2 for
both the oxygen and air oxidation model plants. Model plant emission sources,
emissions, and emission reductions are summarized in Table: 5-3a and 5-3b. The
annualized cost per metric ton of ethylene oxide reduced is $1,800 for oxygen -
oxidation plants and $1,200 for air-oxidation facilities. These costs and
emission reductions reflect the new source performance standard level of
control; detailed costs are provided in the Appendix A.
5.1.2. Ethoxylation
Ethoxylation equipment leak control costs are provided in Table 5-4.
Equipment counts were assumed to be 10 percent of ethylene oxide production
facilities. Annualized cost per metric ton of ethylene oxide reduced is
approximately $2,500.
5.2 Steri 1 i zati on/Fumi gati on
Capital costs, annualized costs, and cost effectiveness for a DEOXXTM
system treating approximately 90 metric tons of ethylene oxide is summarized
in Table 5-5. Annualized costs per metric ton of ethylene oxide reduced is
approximately $450.
5-1
-------
Table 5-1.
Preliminary Control Efficiency and Cost Effectiveness
Ethylene Oxide Emission Sources
Model Plant
Facility/Source
I. Production
A. Equipment Leaks
1. Oxygen Oxidation
2. Air Oxidation
B. Cooling Towers
C. Vents
1 . Argon
2. C0£ Stripper
3. E.O. Stripper,
Reabsorber, or
Purification
II. Et ho xyl a ti on/Equipment
Leaks
III. Sterilizers & Fumi gators
Emission Annual Cost
Control Technique Reduction (1983 $)
NSPS SXMI Std
for Equipment Leaks
Repacking stripper
column
Boiler/Incinerator
Carbonate Flasher and
Vent Condenser
Recycle to Process
NSPS SXMI Std
for Equipment Leaks
DEOXXTM
65 64,30Ua
80 63,600a
80 unknown
99.9
92
99.9
70
12.400C
99 39,800d
Emiss. Reduction Cost Effect
(Mg/yr) ($/Mg)
35b 1,800
54b 1,200
20 unknown
- Baseline -
- Baseline -
- Baseline -
5C 2,500
87 450
aFrora Table 5-2
bFrora Table 5-3
cFrom Table 5-4
dFrom Table 5-5
-------
TABLE 5-2. EQUIPMENT LtAK CONTROL COSTS3 EThYLENt UXIUL PRODUCTION
Model Plants
Total Annual Cost
(Oct
Source/Process
Pumps
Valves
Safety/Relief Valves
Open-Ended Lines
Compressors
Sampling Connections
Monitoring Instruments
TOTAL COST
Product Recovery Credit
NET ANNUAL COST
Oxygen Oxidation
10,090
57,040
7,840
0
7,200
1.700
7,270
91,140
+(26,800)
64,340
Air Oxidation
8,890
51,200
16,800
0
14,4uu
6,120
7,270
104,680
+{ 41,040)
63,640
aDetailed costs are provided in Appendix.
5-3
-------
TABLE 5-3a
.p-
EQUIPMENT LEAKS
EMISSIONS AND EMISSION REDUCTION ESTIMATES
ETHYLENE OXIDE PRODUCTION: OXYGEN-OXIDATION
Model Plant
Emissions
Source
Pump Seals
Light Liquid
Heavy Liquid
Valves
Gas
Light Liquid
Heavy Liquid
Safety/relief valves
Gas
Open-ended lines
Compressors
Sampling connections
Flanges
TOTAL
Number of
Sources
30
177
693 '
2b
0
3
10
1,738
Annual
Emissions, Mg/yra
6.5
4.2
21.5
11.8
0.0
2.9
0.6
6.4
53.9
Control
Technique NSPS
LDAR
LDAR
Performance
Standard
CAPS
Seal System
Closed Purge
System
Present
Control
Efficiency
0.61
0.73
0.59
1.0
1.0
1.0
1.0
TOTAll
Emission
Reductions,
Mg/yr
3.9
3.0
12.6
11.8
0.0
2.9
0.6
34.8
aFrom Table 3.5
-------
TABLE 5-3b
Ul
Ul
EQUIPMENT LEAKS
EMISSIONS AND EMISSION REDUCTION ESTIMATES
ETHYLENE OXIDE PRODUCTION: AIR-OXIDATION
Model Plant
Emissions
Source
Pump Seals
Light Liquid
Heavy Liquid
Valves
Gas
Light Liquid
Heavy Liquid
Safety/ relief valves
Gas
Open-ended lines
Compressors
Sampling connections
Flanges
TOTAL
Number of
Sources
27
200
5bl
60
0
6
36
1,214
Annual
Emissions, Mg/yra
5.8
4.8
18.0
27.4
0.0
5.9
2.4
4.4
68.7
Control
Technique NSPS
LDAR
LDAR
Performance
Standard
CAPS
Seal System
Closed Purge
System
Present
Control
Efficiency
0.61
0.73
0.59
1.0
1.0
1.0
1.0
Emission
Reductions,
Mg/yr
3.5
3.5
10.6
27.4
0.0
5.9
2.4
53.3
aFrom Table 3.4
-------
TABLt 5-4. EQUIPMENT LEAK CONTROL COSTS: ETHOXYLATIUN
Model Plant3
Total Annual
Source/Process Cost, $
Pumps 950
Valves b,400
Safety/Relief Valves 1,200
Open-ended Lines 0
Compressors l.ObU
Sampling Connections 390
Monitoring Instruments 7,270
TOTAL COST 16,290
Product Recovery Credit (3,850)b
NET ANNUAL COST 12,440
Component counts assumed to be 10% of production facility components. Costs are
then lOfc of average costs for oxygen and air oxidation model plants (except for
monitoring instrument costs).
^Emission reduction = 5 mton/yr.
5-6
-------
Table 5-5 COST EFFECTlVENtSS
DEOXX™ SYSTEM
Cost 1983 $ Annual Cost
Capital Cost (Installed) 217,350* 25,500b
Operating Cost — 9,300C
Local Taxes, Insurance -- 3,500d
Administration & Contingency — 800e
Total Annual Costs 39,100
Emission Reduction (Mg/yr) 87f
Cost Effectiveness ($/mton) $ 450
aLetter from Anthony Buonlcore of Chemrox, Inc. to David Markwordt U. S. EPA
dated January 20, 1984. Plant: 87 metric tons of ethylene oxide from 8
sterilizers
^Capital Charges assume 1 = 10% and 20 years equipment life
°Letter from Anthony Buonlcore of Chemrox, Inc. to David Markworat, U.S.
Environmental Protection Agency, dated August 27, 1984.
d!.6% fixed-capital costs
e6% of total operating cost (9,300 + 3,500)
fAssumed no value for recovered chemical by-product
C-7
-------
References
1. Chemical Origins and Markets. Chemical Information Services.
Menlo Park, California. Fourth Edition. 1967, p. 12.
2. Chemical Profile: Ethylene Oxide. Chemical Marketing Reporter.
June 8, 1981.
3. Chemical Products Synopsis - Ethylene Oxide, Mannsvllle Chemical
Products, December 1982.
4. Ozero, Brian R. and ProcelH, Joseph Y., Can Developments Keep
Ethylene Oxide Viable. Hydrocarbon Processing, March 1984,
pp. 55-58. i
5. Kalcevic, V. andJ.F. Lawson. Ethylene Oxide. In: Organic Chemical
Manufacturing, Vol. 9: Selected Processes. (Prepared for U.S. Environ-
mental Protection Agency, EPA-450/3-80-028d). IT Envirosdence.
Knoxville, TN. December 1980. :
6. Field, D.D., et al. Engineering and Cost Study of A1r Pollution
Control for tfie" "Petrochemical Industry, Vol. 6: Ethylene Oxide Manu-
facture by Direct Oxidation of Ethylene. (Prepared for U.S. Environ-
mental Protection Agency, Contract No. 68-02-0255). Houdry Division,
A1r Products and Chemicals, Inc. Marcus Hook, PA. June 1975.
7. Kirk-Othmer Encyclopedia of Chemical Technology. Third Edition, Vol. "9.
Ethylene Oxide. John Wiley and Sons. New York f NY. 1980. pp. 432-471.
8. Detergent Alcohols, Chemical Economics HandboTJk. SRI International,
' February 1981,'p. 601. 5022 T.
9. Letter from Anthony J;- Buonicore of Chemrox Inc. to David W. Markwordt
of fc. S. Environmental Protection Agency, RTP, N. C. January 20, 1984.
10. Letter from Anthony J. Buonicore of Chemrox Inc. to David W. Markwordt
of U. S. Environmental Protection Agency, RTP, N. C. August 27, 1984.
5-8
-------
APPENDIX A
Equipment Leak Control Costs
for
Ethylene Oxide Model Plants
-------
TABLE 1. EQUIPMENT LEAK CONTROL COSTS3 cThYLLUE OXIDE PROuuCTlOM
Model Plants
Total Annual Cost
(Oct 19tt3 i)
Source/Process
Pumps
Valves
Safety/Relief Valves
Open-Ended Lines
Compressors
Sampling Connections
Monitoring Instruments
TOTAL COST
Product Recovery Credit
NET ANNUAL COST
Oxygen Oxidation
10,090a
57,040b
7,840C
0
7,20UC
1,700C
7,270
91,14u
+ (26, 800) d
64,340
Air Oxidation
8,890d
51,200b
lb,800c
0
14,40UC
6,120C
7,270
104,680
( 41,040)"
63,640
aFrom Table 3a and 3b.
bFroni Taole 5a and 5b.
cFrom Table 7a and 7b.
dValue of product recovered = $.35/lb or $77U/Mg. Letter from U. C. rtacauley,
Union Caroide Corporation, to Susan R. Wyatt, U.S. EPA, October 5, 1964.
A-2
-------
Table 2a. Emissions ana Emission Reduction Estimates
Ethylene Oxide Production: Oxygen Oxidation
Model Plant
Emissions Source
Pump Seals
Light Liquid
Heavy Liquid
Valves
Gas
Light Liquid
Heavy Liquid
a» Safety/Relief Valves
u> Gas
Open-ended Lines
Compressors
Sampling Connections
Flanges
TOTAL
Number of
Sources
30
177
693
26
0
3
10
l,7Jb
Emission Factor,
kg/day/source
1.19
0.13
0.17
2.5
0.04
5.47
0.36
0.02
Adjustment
Factor**
.5
.5
.5
.5
.5
.5
.5
.5
Annual
Emissions, Mg/yrb
6.5
4.2
21.5
11.8
0
2.9
0.6
6.4
53.9
Control
Efficiency
0.61
0.73
0.59
1.0
1.0
1.0
1.0
--
Emission
Reductions,
Mg/yr
J.9
3.0
12.6
il.8
0
2.9
0.6
-
34.8
aThe emission factor assumes 100 percent organic in stream and must be adjusted for ethyl one oxide emission estimates
to account for the different percent of ethylene oxide in various streams. Assumed the average stream contains bO
percent ethylene oxide.
estimating purposes, one* operating year was assumed to oe 365 days.
-------
Table 2b. Emissions and Emission Reduction Estimates
Ethylene Oxide Production: Air Oxidation
Model Plant
Emissions Sources
Pump Seals
Light Liquid
Heavy Liquid
Valves
Gas
Light Liquid
Heavy Liquid
Safety/Relief Valves
Gas
Open-ended Lines
Compressors
Sampling Connections
Flanges
TOTAL
Number of
Sources
27
200
581
60
2
b
36
1,214
Emission Factor,
kg/day/source
1.19
0.13
0.17
2.5
0.04
5.47
0.36
.0.02
Adjustment
Factor8
.5
.5
.5
.5
.5
.5
.5
.5
Annual
Emissions, Mg/yrb
6.8
4.8
18.0
27.4
0
5.9
2.4
4.4
68.7
Control
Efficiency
0.61
U.7J
0.59
l.u
1.0
1.0
i.U
Emission
Reductions,
Mg/yr
3.5
3.5
10.6
27.4
U
5.9
2.4
53. J
aThe emission factor assumes 100 percent organic in stream and must be adjusted for ethylene oxide emission estimates
to account for tne different percent of ethylene oxide in various streams. Assumed the average stream contains 5u
percent ethylene oxide.
bFor estimating purposes, one operating year was assumed to be 3b5 days.
-------
Table 3a. ANNUAL COSTS OF A LEAK DETtCTIOM AND REPAIR PRUbRAi'i KOK
PUMP SEALS--MONTHLY MONITORING: OXYGEN-OXIDATION
Monitoring
Number of Pump Seals
Monitoring Time, Man-Mi n/Source
Monitoring Time Per Interval, Win
Number of Times Monitored Annual ly3^
Monitoring Time Total, Man-hours
Instrument
x
x
30
10
300
12/60
6U
Visual
3u
x .5
Ib
x 52/60
1J
Repairs
Number of Pump Seals
Fraction of Leaks Found Annually3
Estimated Number of Repairs Annually
Laoor hours Per Repair
Repair Times Total, Man-Hours
Estimated Number of Repairs Annually
Seal Costs, S/Single Seal
Annual Materials Cost
Aggregation of Costs
Instrument Monitoring Hours
Visual Monitoring Hours
Repair Hours
Total Labor Hours
Labor Charge Rate, $/Man-Hour
Labor Charges
Admin. & Support Costs (0.4 x Labor Charges)
Materials Costs
Annualized Charyes for Initial Repairs
TOTAL ANNUAL COSTS
X
X
X
30
0.408
12.il
16
195
11.8
164
1,935
60
13
+ 195
Z5F
x 21
5,b2»
2,251
1,935
+ 273C
510,087
fraction of sources repaired annually is an output of the LDAR model.
DThe number of times monitored annually is divided by "60" to convert
from man-minute to man-hours.
cFrom Table 4a.
A-5
-------
Table 3b. ANNUAL COSTb OF A LtAK OEThUiON AND RtPAIR PRUbftaM FOR
PUMP SEALS--MONTHLY MONITORING: AIR OXIDATION
Monitoring
Number of Pump Seals
Monitoring Time, Man-Mi n/Source
Monitoring Time Per Interval, Win
Number of Times Monitored Annual lya»k
Monitoring Time Total, Man-hours
Instrument
27
x 10
270
X12/60
b4.0
Visual
27
x 0.5
X52/60
2.7
Repairs
Number of Pump Seals
Fraction of Leaks Found Annually3
Estimated Number of Repairs Annually
Labor Hours Per Repair
Repair Times Total, Man-Hours
Estimated Number of Repair Annually
Seal Costs, $/Single Seal
Annual Materials Cost
Aggregation of Costs
Instrument Monitoring Hours
Visual Monitoring Hours
Repai r Hours
Total Labor Hours
Labor Charge Rate, $/Man-Hour
Labor Charges
Admin. & Support Costs (0.4 x Labor Charges)
Materials Costs
Annualized Charges for Initial Repairs
TOTAL ANNUAL COSTS
27
x 0.408
11.0
x 16
DT>
11.0
x 164
54.0
2.7
176
232.7
21
$4,887
1,955
1,804
248C
$8,b94
aFraction of sources repaired annually is an output of the LDAR model.
bThe number of times monitored annually is divided by "60" to convert from
man-minute to man-hours.
cFrom Table 4b.
A-6
-------
Taolc 4a. COSTS OF INITIAL LtAK UETtLTIGN ANu REPAIR OF PUMP ScALS
ETYLENE OXIDE MODEL UNIT : OXYGEN-OXIDATION
Number of Pump Seals 30
Initial Leak Frequency x 0.088
Estimated Number of Leaks 2.64
Repair time, man-hour/leak x 16
Labor hours required 42.2
Labor charge rate, $/man-hour x 21
Initial leak repair 555
Admin. & Support Costs (0.4 x Initial leak repair) + 354
SUBTOTAL 1,240
Seal costs, $/single seal3 164
Estimated Number of Leaks . 2.64
Materials Cost 433
Subtotal (from above) ' 1,240
TOTAL CAPITAL COSTS 1,673
Capital" recovery factor* x .163
ANNUALIZED COST OF CAPITAL |~~27T
^Fraction of sources repaired annually is an output of the LDAR model.
bThe number of times monitored annually is divided by "60" to convert
from man-minutes to man-hours.
A-7
-------
TAbLE 4b. COSTS UF IwIlIAL LEAK utTEUTIOi, ANU RtPAIk OF PuMP StALS
ETHYLENE OXIDE MODEL UNIT : AIR OXIDATION
Number of Pump Seals 27
Initial Leak Frequency x 0.088
Estimated Number of Leaks O
Repair Time/Man-Hour/Leak x 16
Labor Hours Required JO
Labor Charge Rate, $/Man-Hour x 21
Initial Leak Repair 8013
Admin. & Support Costs (0.4 x Initial Leak Repair) + 323
Sub-Total (Labor Charges) 1
Seal Costs, $/Single Seal6
Estimated Number of Leaks x 2.4
Materials Cost 394
Subtotal (from above) 1,129
TOTAL CAPITAL COSTS $1^23
Capital Recovery Factorb x 0.163
ANNUALIZED COST OF CAPITAL ~
aCost of single seal (replacement) includes a 50 percent credit
for the old seal.
^Capital recovery factor is based on a 10 percent interest rate
over a 120-year period.
A-8
-------
Table 5a. ANNUAL COSTS OF A LtAK DETtCTlOfo AND RtPAIR PROGRAM
VALVES--MONTHLY MONITORING: OXYGEN-OXIDATION
Monitoring Instrument
Number of Valves 870
Monitoring Time, Man-Min/Source x 10
Monitoring Time Per Interval, Min. 8,700
Number of Times Monitored Annually x 12/60
Monitoring Time Total, Man-hours 1,740
Repairs
Number of Valves 870
Fraction of Leaks Found Annually3 x .191
Estimated Number of Repairs Annually 166
Labor Hours Per Repair3 x 1.13
Repair Times Total, Man-Hours 188
Aggregation of Costs
Instrument Monitoring Hours 1,740
Repair Hours + 188
Total Labor Hours 1,9*8
Labor Charge Rate, $/Man-Hour 21
Labor Charges 40,488
Admin. & Support Costs (0.4 x Labor Charges) 16,195
Annualized Charges for Initial Repairs + 352b
TOTAL ANNUAL COSTS 57,035
a"Fugitive Emission Sources of Organic Compounds—Additional Information
on Emissions, Emission Reductions, ana Costs," EPA-45U/3-82-010, P.5-8.
% Leaks=(i!rLeaks/yr)/(#Valves)
bFrom Table 6a.
A-9
-------
Taole 5b. AUuUAL COSTS OF A LcAK ufltCTlON AND RtPAIR' PROGRAM
VALVES--MONTHLY MONITORING: AIR-OX10ATIUM
Monitoring Instrument
Number of Valves 781
Monitoring Time, Man-Mi n/Source x 10
Monitoring Time Per Interval, Min. 7.81U
Number of Times Monitored Annually x 12/60
Monitoring Time Total, Man-hours 1,562
Repairs
Number of Valves 781
Fraction of Leaks Found Annually3 x
Estimated Number of Repairs Annually 149
Labor Hours Per Repaira x 1.13
Repair Times Total, Man-Hours Iba
Aggregation of Costs
Instrument Monitoring hours 1,562
Repair Hours + 168 '
Total Labor Hours 1.73U
Labor Charge Rate, $/Man-hour x 21
Labor Charges 36,330
Admin. & Support Costs (0.4 x Labor Charges) 14,53*.
Annual i zed Charges for Initial Repairs + 331°
TOTAL ANNUAL COSTS 51,193
a"Fugitive Emission Sources of Organic Compounds--Additional Information
on Emissions, Emission Reductions, and Costs," EPA-450/3-82-01G, P.5-«.
% Leaks=(#Leaks/yr)/(#Valves)
bFrom Table 6b.
A-10
-------
Taole 6a. COSTS OF INITIAL LtAK DETECTION AND REPAIR OF VALVES
ETHYLZUE OXIDE PRODUCTION: OXYGEN OXIDATION
Service
Gas Light Liquid
Number of Valves 177 693
Initial Leak Frequency3 x 0.114 x 0.065
Estimated Number of Leaks 213 ?3
Repair Time, Man-Hour/Leak3 x 1.13 x 1*13
Labor Hours Requires ZZ76 51
Labor Charge Rate, $/Man-Hour x 21 x 21
Initial Leak Repair ~~4~73 TTU7I
Admin. & Support Costs (0.4 x Initial Leak Repair) + 190 + 428
~~ "
Total Capital Costs 665 1,499
Capital Recovery Factor^ x 0.163 x 0.163
Annual ized Cost of Capital 108 244
a"Fugitive Emission Sources of Organic Compounds—Additional Information on Emissions,
Emission Reductions, and Costs," EPA-450/3-82-01U, p. 5-7.
bCapital Recovery Factor is based on a 10 percent interest rate over a 10-year period.
A-ll
-------
Taole 6b. COSTS OF HUTIAL LEAK uETcCTIOfc AND RtPAIR OF VALVES
hTHYLENE OXIDE PRODUCTION: AIR OXIDATION
Service
Gas Light Liquid
Number of Valves 200 58.1
Initial Leak Frequency3 x 0*114 x 0.065
Estimated Number of Leaks TT 33
Repair Time, Man-Hour/Leak3 x 1.13 x 1.13
Labor Hours Requires 2U 4o
Labor Charge Rate, $/Man-Hour x 21 x 21
Initial Leak Repair ~~54~5 9U3
Admin. & Support Costs (0.4 x Initial Leak Repair) +218 + 361
" I,2b4
Total Capital Costs 764 1,264
Capital Recovery Factor^ x 0.163 x 0.163
Annualized Cost of Capital 125 206
a"Fugitive Emission Sources of Organic Compounds—Additional Information on Emissions,
Emission Reductions, and Costs," EPA-4bO/3-82-010, p. 5-7.
DCapital Recovery Factor is .based on a 10 percent interest rate over a 10-year period.
A-12
-------
Table 7a.
COSTS OF LQUIPMENT CONTROL IN MOUtL UNIT (OCT 19b3$)
Ethyl ene Oxide Production Plants
(Oxygen' Oxidation Process)
CO
1
Source
Safety/Relief Valves (G)a
Open-Ended Lines
Compressors
Sampling Connections
TOTAL
Component Component
Captidl Annual ized
Number Of Cost, Capital Cost
Components $ Source Cost, $ $ /Source
7b 3,765e 2b,355 l,120e
Oc 65e 0 lbe
3 y,470e 26,410 2,40Ue
10d 670fc 6,700 170e
61,465
Annual ized
Cost, $
7,840
0
7,200
1,700
16,700
Component control costs are comprised of 5U percent RD/block valve and 50 percent RD/3-way valve.
D75 percent of gas service safety/relief valves are assumed to be controlled at baseline; therefore, the
emissions reduction and costs of control are based on 7 valves Instead of 2b.
CA11 open-ended lines (in model unit) are assumed to be controlled at baseline; therefore, there ure no
associated costs of control or emissions reduction.
°75 percent of sampling connections are assumed to be controlled at baseline; therefore, the emissions
reduction arid costs of control are based on 10 out of 40 sampling connections.
eFrom Table 8.
-------
TABU 7b.
COSTS OF LQUIPMLNT CONTROL IN MOULL UNIT (OCI
Lthylene Oxide Production Plants
(Air Oxidation Process)
Source
Safety/Relief Valves (G)*
Open-Ended Lines
£ Compressors
Sampling Connections
TOTAL
Component
Capual
Number Of Cost,
Components $ Source
15b J>765e
Oc 65e
6 9,47oe
36d 670*
Capital
Cost, $
56,475
0
56,820
24,120
137,415
Component
Annual ized
Cost
WSource
1.1*0*
lb«
•MUU«
170^
Annual 1
Lost
16.UUO
0
14,400
6,120
37,320
zed
, *
Component control costs are comprised of 50 percent RD/block valve and 50 percent RU/3-way valve.
percent of gas service safety/relief valves are assumed to be controlled at baseline; therefore, the
emissions reduction and costs of control are based on 15 valves Instead of 60.
CA11 open-ended lines (192 1n model unit) are assumed to be controlled at baseline; therefore, there are no
associated costs of control or emissions reduction.
d75 percent of sampling connections are assumed to be controlled at baseline; therefore, the emissions
reduction and costs of control are based on 36 out of 145 sampling connections.
eFrom Table 8.
-------
TABLE 8
COSTS OF CONTROLLING FUGITIVE VOC EMISSIONS USING
EQUIPMENT CONTROLS (Oct 03$)
Factor
for
en
t Source
Pump (LL)
Safety/Relief
Valves (G)
Open-ended Lines
Compressors
Sampling Connections
Control Technique
Seal system/flare
Seal system/enclosed
device
RD/block valves
RD/3-way valve
Soft-seals (0-rlngs)
Connection to flare
2nd valve
Seal system
Closed purge system
Capital
Cost,
I/Source
4,690e
8,4006
2.510J
5.020f
2909
2,8409
659
94709
6709
Amortization
Period,
Years
2/1 OC
2/1 QC
2/1 Qd
2/1 Od
10
10
10
10
10
Capital
Recovery
Factor*
0.58/0.163C
0.58/0 163C
0.58/0 163d
0.58/0 163d
0.163
0.163
0.163
0.163
0.163
Other
Capital
Charges^
0.09
0.09
0.09
0.09
0.09
0.09
0.09
0.09
0.09
Annual 1 z\
Cost,
I/Source
1,650
2,5906
750f
1,490?
73
720
16
2,400
170
computecTUsTng" Tfie TO~percenTlnl;ereslt ra¥e~TT) "arfdThe ambrtfz'atldn perTod" Tn):
CRF - 1 (1 + 1)n
Bother capital charges Include maintenance, taxes, Insurance, etc.; 9 percent of the capital cost Is used for those
charges.
CA 2-year period Is used for the seal; a 10-year period Is used for piping and Installation charges.
dA 2-year period Is used for the disk; a 10-year period Is used for piping and Installation charges.
eFrom Table 9
f Front Table 10.
9Apply CE plant cost Index ration (Oct 1983/1978) 318.2/218.8 • 1 45 to figures In Table 2 of Memo: Subject:
Procedure to Compile Control Cost, From: John Stalling. To: SOCMI Fugitive NSPS File, March 11, 1983
-------
T^',e 9. EQUIPMENT COST FOR CONTROL OF LfclSSIOUS FrtGM A PJMP SEAL
(October 1983- collars)
Capital Costs
Dual seal3
Barrier fluid0
Vent system
Total
-nr.ual Costs
Degassing Reservoir
Vented to a Flare
Dual seal2
Barrier fluid6
Vent system6
Maintenance &
"iscellaneous charges^
Total 9
1015C
165
421
Degrassing Reservior
Vented to an Enclosed
Combustion Device
21cC
4730
707
355
770
756
£3ased on S125u/seal (I960 dollars) average cost quoted by industry.
197& cost « 51030 (seal cost = 51015, installation cost « 3240, single
seal credit = S225). See Table 10 for 1983-dollars.
~rrc/n Chapter 3 of the BID.
cEased on installed capital cost of 2Um of 5.1cm piping (Sb5u), one 5.1cm
check valve (S125), and one 5.1cm block valve (S415) per vent system ana
cne vent per pair of pump seals (1978 Dollars).
i
dBased on 2 year seal life and 10 percent interest (CRF * 0.58) for aual
seal and 10 year amortization period and lu percent interest (CRF »
j.163) for installatin charges.
e8ssea en 10 \zcr equipment life and 10 percent interest (CRF = 0.16':}.
fiascd en 9 percent of total capital ccsts (Chapter if cf the BID).
?;osts ao not include credits for recovered proauct.
A-16
-------
TABLE 10. RELIEF VALVE CONTROL COSTS
(October 1983 S)
d) Rupture disk systems Total Capital Cost = S2,5C9 re'.ief valve
vntn block valves
Annual II zed Costs Capital Costs
Rupture disk Sibo i 133
Holders, etc. 363 2,226
Maintenance & Misc. 226 _
TOTAL (S/year) 1737 S2t5o»
•
(2) Rupture disk systems with
2-s2/ valve
Assembly
One 7.6 cm stainless steel rupture disk S 282
One 7.6 cm carbon steel disk holder 471
One 0.6 cm dial -face pressure gauge 22
One 7.6 en safety/relief valve 1,800
Two 7.5 cm elbows _ 36
SUBTOTAL S2,bi2
Three-way Valve
une / .b cm, j-way, 2 port valve SI, 624
Installation 783
SUBTOTAL 3
TOTAL COST 55,019
Annual i zed Costs
Rupture disk S
Holcer, valve, etc. 772
•'svvienance & Misc. 552
TOTAL (S/year) si,4be
A-17
-------
APPENDIX B
Dispersion Modeling Parameters
-------
HUMAN EXPOSURE MOUEL
EPA's Human Exposure Model (HEM) is a general model capable of producing
quantitative expressions of public exposure to ambient air concentrations
of pollutants emitted from stationary sources. The input data needed to
operate this model include source data that are typically required for tne
operation of dispersion models (e.g. plant location, height of emission
release point, and velocity of stackgases). Based on the source data, tine
model estimates the magnitude and distribution of ambient air concentrations
ot the pollutant in the vicinity of each source.
The dispersion model used is a Gaussian diffusion model that uses the
same basic dispersion algorithm as EPA's Cl imatological Dispersion Model
(COM), but has been simplified to improve overall efficiency. In general,
COM determines long-term (seasonal or annual) quasi-stable pollutant
concentrations at any ground level receptor using average emission rates
from point and area sources and joint frequency distribution of wind
direction, wind speed, and stability for the same period.
Stability array (STAR) summaries are the principal meteorological
input to the HEM dispersion model. STAR data are standard climatological
frequency-of-occurrence summaries formulated for use in EPA models and
available for major U.S. meteorological monitoring sites from the National
Climatic Center. A STAR summary provides a frequency-of-occurrence of wind
speed, stmospheric stability, and wina direction, classified according to
Pasquill's categories.
Estimates of human exposure are developed using dispersion modeling of
emissions that pairs population density estimates with predicted pollutant
concentrations from the concentration grid described above. Estimates of
B-2
-------
the number and distribution of people residing within 50 km of..each plant
are oased on a slightly Modified version of the Census Bureau's "Master
Enumeration District List--Extended" (MED-X) data base.
Dispersion results for ethylene oxide production facilities are
presented below:
DISPERSION RESULTS
Production
Plants
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Maximum Annual
Average Concentration
(ppm)
0.002
0.002
0.0006
0.004
0.004
0.001
0.0001
0.001
0.003
O.Oil
0.004
0.002
0
0.0002
0.005
(yg/m3)
3.59
3.05
0.98
6.37
6.30
2.29
0.22
2.15
4.92
18.8
7.54
2.96
0
0.29
9.12
B-3
-------
TABLE 1.
Dispersion Modeling Parameters
Ethylene Oxide/Production Facilities
CO
North West Emission Stack/Vent source
1 ant/Emission Source Latitude Longitude Urban 0 Rate Height Area Stack(O)
(deg/mln/s) (deg/min/s) Rural 1 (kg/yr) (meters) (m2) Vent(l)
1
2
3
4
Process 30 11 22 90 59 43
Cooling Tower
Equipment Leaks
Effluent
Process 29 37 24 95 03 34
Cooling Tower
Equipment Leaks
Effluent
Process 28 59 34 95 24 22
Cooling Tower
Equipment Leaks
Effluent
Process 30 19 00 91 16 20
7,400
15,300
61,500
0
6,600
13,500
61,500
0
4,600
—
61,500
0
34,800
25.4
16
76,000
—
25.4
16
76,000
—
25.4'
—
76,000
—
25.4
1
1
-
-
1
1
-
-
1
-
-
-
1
Stack/Vent Emis. Emis.
Diameter Velo- Tenip.
(meters) city (K)
(m/s)
0.23
5.98
—
—
0.23
5.98
—
—
0.23
—
—
—
0.23
17
11.7
0.01
—
17
11.7
0.01
--
17
—
0.01
—
17
346
314
2/3
—
346
314
273
--
346
—
2/3
—
346
Cooling Tower
5
Equipment Leaks
Effluent
Process 29 37 44 95 03 04
Cooling Process
Equipment Leaks
Effluent
113,000
0
7,700
15,900
61,500
0
20,000
—
25.4
16
76 ,000
--
-
-
1
1
-
-
—
—
0.23
5.98
—
--
0.01
—
17
11.7
0.01
—
2/3
—
346
314
2/3
--
-------
TABLE 1. (cont'd)
Dispersion Modeling Parameters
Ethylene Oxide/Production Facilities
00
North West Emission
Plant/Emission Source Latitude Longitude Urban 0 Rate
(deg/m1n/s) (deg/m1n/s) Rural 1 (kg/yr)
6
7
8
9
10
Process 41 24 30
Cooling Tower
Equipment Leaks
Effluent
Process 38 00 10 86 07 16
Cooling Tower
Equipment Leaks
Effluent
Process 30 3 38 94 2 53
Cooling Tower
Equipment Leaks
Effluent
Process 30 11 12 90 59 28
Cooling Tower
Equipment Leaks
Effluent
Process 39 48 20 75 25 40
Cooling Tower
Equipment Leaks
Effluent
3,600
7,300
61,500
0
1,700
3,500
61,500
0
2,600
5,400
61,500
0
10,800
22 ,300
61,500
0
1,700
110,000
61,500
0
Stack/ vent Source
height Area Stack(O) Stack/Vent Emis.
(meters) (m2) Vent(l) diameter Velo-
(meters) city
(m/s)
25.4
16
—
—
25.4
16
—
—
25.4
16
—
—
25.4
16
—
—
25.4
16.8
--
—
1
1
76 ,000
—
1
1
76 ,000
—
1
1
76,000
--
1
1
76 ,000
—
1
1
76,000
—
0.23
5.98
—
—
0.23
5.98
—
—
0.23
5.98
—
—
0.23
5.98
—
—
0.23
7.3
—
--
17
11.7
0.01
—
17
11.7
0.01
—
17
11.7
0.01
—
17
11.7
;
0.01
—
17
40
0.01
—
Emis.
Temp.
(K)
346
314
273
—
346
314
273
—
346
314
273'
—
346
314
273
—
346
283
273
—
-------
TABLE 1. (cont'd)
Dispersion Modeling Parameters
Ithylene Oxide/Production Facilities
North West Emission Stack/ vent Source
Plant/Emission Source Latitude Longitude Urban 0 Rate height Area Stdck(O) Stack/Vent trnis. tmis.
. (deg/min/s) (deg/min/s) Rural 1 (kg/yr) (meters) (m2) Vent(l) diameter Velo.- Temp.
(meters) city (K)
(ro/s) .
11 Process
Cooling Tower
Equipment Leaks
Effluent
12 Process 32 26 44 94 41 17
Cooling Tower
5,400 25.4 - 1
—
113,000 -- 20,000
1,100
3,000 25.4 — 1
6,200 16.8 — 1
0.23 17
—
0.01
0.01
0.23 17
7.3 40
34(i
—
273
273
346
283
Equipment Leaks
no Effluent
at
13 Process 17 59 36 66 44 46
61,500 — 76,000
0 —
5,000 25.4 — 1
0.01
__
0.23 17
273
—
346
Cooling Tower
Equipment Leaks
Effluent
14 Process 28 30 30 96 46 15
113,000 ~ 20, DUD
1 ,000
7,100 25.4 ~ 1
0.01
0.01
0.23 17
273
273
346
Cooling Tower
Equipment Leaks
Effluent
15 Process 29 59 27 90 26 50
113,000 — 20,000
1 ,500
10,000 25.4 — 1
0.01
0.01
0.23 17
273
273
346
Cooling Tower
Equipment Leaks
Effluent
113,000 — 20,000
2,000
0.01
0.01
273
273
-------
oo
I
TABLE 2.
DISPERSION MODELING PARAMETERSa
Ethylene Oxide/Etnoxylator, Sterilizer, ana Fumigator Facilities
Emission Stack/vent
Rate Height
Model Plant/Emission Source (Kg/yr) (meters)
Source
Area
(m2)
Stack (0) Stack/vent Emission Emission
Vent(l) Diameter Velocicy Temp.
(meters) (m/s) (k)
Ethoxyl a tors/Equipment Leaks
6,100
7,000
Sterilizers/Vent Emissions lc
and (Low)
Fumigators 2d 30,000
(High)
25.4
25.4
20,000
0.23
0.23
O.Oi
17
17
273
273
273
aA11 parameters are assumed.
bl(ft of 61.5 Mg/yr
cAssumes 450 facilities.
Assumes 100 facilities.
-------
Table 3.
tthylene Oxide Plant Production Capacity
Plant
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Plant
BSAF
Celanese
Dow
Dow
ICI Americas
Northern
Petrochemical
01 in
PPG
Shell
Sun 01 in
Texaco
Texas Eastman
Union Carbide
Union Carbide
Union Carbide
Location
Geismar, LA
Clear Lake, TX
Freeport, TX
Plaquemine, LA
Bayport, TX
Morris, IL
Brandenburg, KY
Beaumont, TX
Geismar, LA
daemont, DA
Port Noches, TX
Longview, TX
Ponce, P. R.
Seadrift, TX
Taft, LA
Capacity
(103kg EO)
218,000
193,000
136,000
205,000
227,000
105,000
50,000
77,QUO
318,000
50,000
318,000
87,000
295,000
418,000
•
590,000
B-8
-------
Table 4.
Source
Process Vent(s)
Cooling Tower
Effluents
Model Plant Dispersion
Emission Factor
(kg/103 kg capacity)
Air Q£
0.017 .034
N/A .07
.0035 .0
Parameters
Ht./D
~7iT
25. 4/ .23
16 /5.9b
N/A
-
Vel.
(m/sec)
17
11.7
0.01
Temp.
Tit)
346
314
273
Equipment Leaks
(Mg/yr)
113 61.5
N/A
0.01
273
B-9
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
REPORT NO.
EPA-450/3-85-Q14
2.
3. RECIPIENT'S ACCESSION NO.
TITLE AND SUBTITLE
Sources of Ethylene Oxide Emissions
5. REPORT DATE
April 1985
6. PERFORMING ORGANIZATION COOi
AUTHOR(S)
David W. Markwordt
8. PERFORMING ORGANIZATION REPORT NO.
PERFORMING ORGANIZATION NAME AND ADDRESS
Office of Air Quality Planning Standards
U.S. Environmental Protection Agency
Research Triangle Park, N.C. 27711
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
13. TYPE OF REPORT AND PERIOD COVERED
Director for Air Quality Planning and Standards
Office of Air and Radiation
U.S. Environmental Protection Agency
Research Triangle Park, N.C. 27711
14. SPONSORING AGENCY CODE
EPA/200/04
IS. SUPPLEMENTARY NOTES
16. ABSTRACT
The objective of this report is to present the results of a preliminary source
assessment study conducted for ethylene oxide in which emission, cost, and dispersion
data were developed that will aid in evaluating the need for further regulatory
development for ethylene oxide.
KEY WORDS AND DOCUMENT ANALYSIS
a.
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
Air Pollution
Pollution Control
Ethylene Oxide
Air Pollution Control
13/b
18. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (Tliis Report!
Unclassified
21. NO. OF PAGES
80
20. SECURITY CLASS (Tliis page)
Unclassified
22 PRICE
EPA Farm 2220-1 (R«v. 4-77) PREVIOUS EDITION is OBSOLETE
------- |