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

-------