HAZARDOUS POLLUTANT  SURVEY

                   FOR THE

           ACRYLONITRILE INDUSTRY


                SECOND DRAFT
                Prepared for:

OFFICE OF AIR QUALITY PLANNING AND STANDARDS
    U.S.  ENVIRONMENTAL PROTECTION AGENCY
   RESEARCH TRIANGLE PARK, NORTH CAROLINA
           Contract No.  68-02-3061
                June  1981
                Prepared by:

   ENERGY AND ENVIRONMENTAL ANALYSIS, INC.
           1111 North 19th Street
         Arlington, Virginia  22209

-------
                             TABLE OF CONTENTS
                                                                      Page
1.    SUMMARY	1-1
     1.1  Monomer Production	1~1
     1.2  Acrylic Fiber Production	1-3
     1.3  ABS/SAN Resin Production	1-4
     1.4  NBR Production	1-6
     1.5  Existing State and Federal  Regulations	1-7
     1.6  Preferred Method of Sampling and Analysis  of AN Emissions  .  1-8
2.    INTRODUCTION	2-1
3.    DESCRIPTION OF INDUSTRY	3-1
     3.1  Source Category	3-1
          3.1.1  Monomer Production	3-1
          3.1.2  Acrylic Fiber Production 	  3-4
          3.1.3  ABS/SAN Resins	3-7
          3.1.4  NBR Elastomers	3-14
          3.1.5  Other Uses of AN	3-17
     3.2  Growth in AN Production	3-18
          3.2.1  Acrylic Fibers	3-18
          3.2.2  ABS/SAN Resins	3-20
          3.2.3  NBR Elastomers	3-20
          3.2.4  Other Areas	3-20
     3.3  Process Description 	  3-21
          3.3.1  AN Monomer	3-21
          3.3.2  Acrylic Fibers	3-25
          3.3.3  ABS/SAN	3-30
                 3.3.3.1  Emulsion Process	3-31
                 3.3.3.2  Suspension  Process	3-33
                 3.3.3.3  Continuous  Mass (Bulk) Process	3-35
                 3.3.3.4  SAN by Emulsion Process 	  3-37
          3.3.4  NBR Elastomers	3-37

-------
                        TABLE OF CONTENTS  (Cont'd)
                                                                      Page
4.    AIR EMISSIONS DEVELOPED IN SOURCE  CATEGORY  	  4-1
     4.1  Plant and Process Emissions  	  4-1
          4.1.1  AN Monomer	4-1
          4.1.2  Acrylic Fibers	4-4
          4.1.3  ABS/SAN Resins	4-9
          4.1.4  NBR Elastomers	4-9
5.    EMISSION CONTROL SYSTEMS 	  5-1
     5.1  Storage and Fugitive Emission Controls	5-1
          5.1.1  Current Controls	5-1
                 5.1.1.1  Process Fugitive  Emission  Controls	5-1
                 5.1.1.2  Storage Tank  Controls	5-2
                 5.1.1.3  Transfer and  Handling  Controls	5-2
          5.1.2  Available Control	5-3
                 5.1.2.1  Fugitives	5-3
                 5.1.2.2  Storage Tanks	5-3
          5.1.3  Loading Facilities	5-8
     5.2  Process Controls	5-9
          5.2.1  AN Monomer	5-9
          5.2.2  Acrylic Fibers	5-13
          5.2.3  ABS/SAN Resins	5-18
          5.2.4  NBR Elastomers	5-20
6.    EMISSION DATA	6-1
     6.1  Availability of Data	6-1
     6.2  Sample Collection and Analysis	6-1
7.    STATE AND LOCAL REGULATIONS	7-1

-------
                              LIST OF TABLES



                                                                      Page

2-1  List of Recipients of Section 114 Letters	  2-2

2-2  List of Plants Visited	  2-3

3-1  1979 AN Production and Consumption for End Use Categories .  .   .  3-2

3-2  Current AN Capacity	  3-3

3-3  AN Production, Capacity, and Capacity Utilization, 1969-1979.   .  3-5

3-4  Acrylic Fiber Producers and Capacities	  3-6

3-5  Acrylic Fiber Production and Capacity Utilization 	  3-8

3-6  End Use of Acrylic and Modacrylic Fibers	  3-9

3-7  ABS/SAN Producers and Capacities	  3-11

3-8  ABS Production and Capacity Utilization 	  3-12

3-9  Trends in Domestic Consumption of ABS Resins by End User,
     1973-1979	  3-13

3-10 Nitrile Elastomer Producers and Capacities	  3-15

3-11 Nitrile Elastomer Production and Capacity Utilization 	  3-16

3-12 Estimated Growth in AN End Markets and AN Demand From 1980
     to 1985	  3-19

3-13 Stream Codes for Figure 3-1	  3-23

3-14 Acrylic and Modacrylic Fiber Production Routes	  3-26

4-1  Summary of Estimated AN Emission Factors From Monomer
     Production	  4-2

4-2  Uncontrolled and Controlled Emissions and Emission Factors
     for AN and VOC From a Model Sohio Process Plant	   4-5

4-3  Summary of Estimated Total AN Emissions From Monomer
     Production	   4-6

-------
                        LIST OF TABLES (Continued)
                                                                      Page

4-4  Estimated Uncontrolled Emissions From Model Plants Producing
     45,500 Mg/yr of Acrylic Fiber 	  4-7

4-5  Summary of Estimated Total AN Emissions From Acrylic Fiber
     Production	4-10

4-6  AN Emissions From ABS/SAN Resin Operations	4-11

4-7  AN Emissions From Nitrile Elastomer Operations	4-13

5-1  Model Storage Tanks and Emissions Reductions	  5-4

5-2  Potential Emissions Reduction For Small Fixed Roof AN
     Storage Tank	  5-5

5-3  Potential Emissions Reduction For External Floating Roof
     Tank	  5-6

5-4  Control Devices Currently Used by the AN Industry	  5-10

5-5  Controls and Estimated Costs for AN Emission Reduction at
     New Monomer Plants	  5-]2

5-6  Existing Controls in the Acrylic Fibers Industry	  5_14

5-7  Estimated Emissions From Model Plants Producing 45,500 Mg/yr
     of Acrylic Fiber Using Various Control Options	  5_16

5-8  Control Options and Estimated Costs for Acrylic Fiber By
     Solution and Suspension Polymerization	  5-17

7-1  Attainment Status 	  7.5

-------
                              LIST OF FIGLRES

                                                                       Page
3-1  Flow Diagram for a Sohio AN Plant	2-22
3-2  Acrylic Fiber Processes	3-28
3-3  Emulsion ABS/SAN Process	3-32
3-4  Suspension ABS Process	3-34
3-5  Bulk ABS Process	3-36
3-6  Nitrile Elastomer Process	3-38

-------
                               1.   SUMMARY

     The acrylonitrile (AN) industry (segments  listed  under Standard  Industrial
Classification (SIC) codes 2821, 2822,  2824,  and 2869)  was  exanined  during
this survey.  The  study was directed at gathering background  information
necessary to assess the need for developing a national  emission  standard
for hazardous air pollutants (NESHAP) for AN  under Section  112 of  the  Clean
Air Act.  The AN industry includes AN monomer synthesis operations,  acrylo-
ni trile-butadiene-styrene (ABS) and styrene-acrylonitrile  (SAN)  plants,
acrylic fiber manufacturing facilities, and nitrile elastomer (NBR)  plants.
1.1  MONOMER PRODUCTION
     There are presently four companies producing AN monomer  at  six  domestic
locations.  Texas  has the largest  number of plants (3  on-line; 1  being
built).  Other States with ronomer plants are Louisiana, Ohio, and Tennessee.
Total domestic production in 1979  was 915,000 Mg (94 percent  of  present
972,700 Mg/yr industry capacity) with sales valued at  $218  million.   The
monomer is used as a feedstock in  the synthesis  of ABS/SAN  resins, NBR,
acrylic fibers, adiponitrile (a nylon intermediate) and acrylamide (a
specialty chemical for use in the  paper industry and wastewater  treatment).
The monomer industry has teen growing at an average annual  rate  of about
6 percent from 1969 through 1979.   Projected  growth in  the  years  between
1980 and 1985 is expected to be slightly lower at 3.8  to 4.4 percent.
     All domestic  AN monomer is produced by the  Sohio  process.   This  is  a
catalytic vapor phase reaction for the  ammoxidation of propylene to  AN.   It
was developed and  patented by the  Standard Oil  Company of Ohio  (Sohio).
This process emits AN to the ataosphere from  reactor startups, absorber  and
column vents, and  fugitive sources (AN  transfer  and handling, storage
tanks, and process equipment).   Total AN emissions to  the atmosphere in
1980 from monomer  operations were  estimated by the Chemical Manufacturers
                                   1-1

-------
Association (CMA) at 747 Mg/yr.  This estimate is based on the installation
of a new incinerator at one plant (American Cyananid in l/estwego,  LA) and
recent plant operating irnprovenents nade by other manufacturers.   The C?1A
estimates were compiled by its AN study team fomed to  evaluate a  report on
the AM industry by the Stanford Research Institute (SRI).   This 1979 SRI
report estimated AN monomer emissions at 2297 Mg/yr.
     The largest AN emission sources from the monomer synthesis process are
the reactor, the absorber, and column vents.  Reactor startup emissions are
very large  (4545 kg/hr) for several short periods each  year (1  hour, 3 to 4
times per year).  Several plants send reactor gases through the absorber
and quench  columns, then to an incinerator or directly  to  the incinerator
(to avoid explosion hazards in the absorber lines).  This  control  procedure
can reduce  reactor startup emissions by as much as 99 percent.   Absorber
and column  vent AN emissions can be controlled through  thermal  incineration
of AN vapors  (catalytic incineration is a second choice) for 98 percent
emission reduction.  The thermal incineration of absorber  and column vent
gases may be a candidate for best available technology  (BAT) for emission
control at  these sources.
     Fugitive AN emissions sources at monomer plants include AN handling
and transfer  (truck, railcar, or barge loading) and process equipment
(pumps, valves, seals, etc.).  These sources can have significant  emis-
sions.  Process equipment fugitive emissions have been  shown to be reduced
as much as  65 percent by adopting a regular inspection  and maintenance
schedule.   Many plants are very conscientious in complying with the Occupa-
tional Safety and Health Administration (OSHA) regulations limiting worker
exposure to AN.  These regulations have encouraged good work practices and
workplace ventilation.  They do not specifically address control  of AN
emission sources.  It has been demonstrated in EPA studies on control of
vinyl chloride fugitive emissions that, even with OSHA  regulations in
place, there is a need for a quarterly maintenance schedule to  reduce
fugitive process emissions.
     Storage tank emissions can be reduced as much as 98 percent through
the use of more effective roof seals (retrofitted according to existing
roof and seal  type) or a vapor recovery system.  The seals could save a
great deal  of previously lost product, justifying their installation and
                                      -2

-------
offsetting their cost.  Emissions from transfer and handling of AN during
loading operations can be large.  They nay be reduced greater than 98 percent
by a vapor recovery system or an incinerator.
1.2  ACRYLIC FIBER PRODUCTION
     Acrylic and nodacrylic fibers are produced by five companies at six
locations in the Lhited States.  There are fiber operations in Alabama,
Florida, South Carolina, Virginia (2), and Tennessee.  The total  U.S. fiber
production in 1979 was 345,000 Mg.  This increased to 353,000 fig  in 1980
(90 percent of the present 390,000 Mg/yr capacity).  Fiber sales  in 1979
had an estimated value of $450 million.
     Acrylic fibers are valued for their similarities in warmth and softness
to natural wool and their ability to be processed into  1 ightweight, high-bulk
yarn.  They are used  in apparel manufacture and in home furnishings.   The
industry has averaged a 3.6 percent growth rate from 1969 through 1979.
Projected growth through 1985 is expected to be much slower (estimated at
only 1 percent for the period 1980 to 1985) with most of that growth  created
by an expanding apparel market.
     Fibers are manufactured by the polymerization of AN with other monomers.
The polymers may be formed by suspension (reaction in aqueous medium) or
solution (reaction in solvent) reactions.  Polymers synthesized by suspen-
sion reaction must be coagulated, then dissolved in solvent for the spinning
process.  Solvent reaction products are already in solution and go directly
to spinning.  The dissolved polymer is extruded into a  hot air stream or
into water to remove  the solvent and form fiber strands.  Fiber is then
stretched, dried, crimped, and cut.  Bales of cut fiber are shipped to end
users for dyeing and yarn production.
     The AN emission  sources from the acrylic fiber manufacturing processes
include the polymer reactor flash tank, slurry steam stripper, crimper
exhaust, and dryers.  There are fugitive emissions from AN storage tanks
and process equipment handling streams containing AN.  Total domestic AN
emissions in 1980 from fiber plants were estimated at 1361 Mg/yr by the  CMA
(SRI 1979 estimate was 4698 Mg/yr).  The AN emissions from the reactor
flash tank and slurry steam stripper can be reduced by 98 percent using  an
AN recovery scrubber system (scrubber with secondary control).  The AN
recovery scrubber vent may be connected in series with a chilled condenser
                                 1-3

-------
or vent gases could be incinerated.  A thorough slurry steam stripping
procedure and polymer washing step removes residual AN monomer such that
dryer and crimper exhausts do not contain significant amounts of ML
Plants using this system account for the lowered emission estimates by the
Ci'lA.  Controlling fugitive emissions at fiber facilities may be accomplished
in the same manner as at monomer synthesis plants.  This involves improved
seals on all storage  tank roofs, vapor control  for AN handling and transfer
procedures, and quarterly inspection and maintenance of process equipment
in AN service.  Federal regulations concerning worker exposure to AN apply
at fiber plants, but  they do not specifically limit AN emissions to the
atmosphere.
1.3  ABS/SAN RESIN PRODUCTION
     The U.S. has seven companies at 13 different locations producing ABS
resins.  There are ABS plants in Massachusetts (2), Ohio (2), Kentucky,
Illinois, West Virginia, Connecticut, Michigan, Missouri, California,
Indiana, and Louisiana.  The total ABS production from these plants in 1979
was  540,000 Mg (64 percent of the total U.S. capacity of 843,000 Mg/yr)
with industry sales valued at $548 million.   The ABS industry has had an
average annual growth rate of 7.1 percent from 1971 to 1979.  Industry
growth from 1980 to 1985 is anticipated at about 5 percent per year.  Major
uses for ABS are in forming plastic pipes and pipe fittings, appliances,
business machines, and automotive parts.
     Monsanto Company and Dow Chemical are the U.S. manufacturers of SAN
resins.  They produced 56,000 Mg of SAN in 1979.  The SAN resin industry
has experienced an average annual growth rate of 6 percent from 1971 to
1979.  These resins are predominately used in compounding with ABS to give
the manufacturer the  desired flexibility and strength in the end product.
Some SAN resin is used in injection molding of housewares, small appliance
parts, automotive lenses, and packaging containers.
     The total AN emissions to the atmosphere in 1980 from ABS/SAN operations
was estimated by the CMA at 2901 Mg/yr (SRI 1979 estimate was 3084 Mg/yr).
Emission points and emission rates of AN differ according to the polymerization
procedures used in the manufacture of ABS/SAN resins.

                                    1-4

-------
     Resin synthesis nay be achieved by emulsion, suspension, or continuous
nass (bulk) procedures.  Emulsion polymerization of ABS/SAN accounts for
the large majority of domestically produced resin.  These processes can
release AN to the atmosphere from reactor vents, monomer steam stripper
(required for SAM production in conjunction with AN recovery scrubber),
coagulation tank vents, vacuum filtration operations,  and polymer dryer
exhausts.  There are fugitive AN emissions from storage tanks, transfer and
handling operations, and process equipment.  The AN process emission sources
have been reduced by 80 percent at one plant using high monomer conversion
technology (HMCT) in the reactor in conjunction with an AN recovery scrubber
at  the vacuum filtration step.  This procedure removes residual  AN in the
polymer, reducing all AN emissions in downstream processing steps.   Ducting
vent gases from AN emission sources to an incinerator/boiler has also been
used for emission control.  This control  method achieves 98 percent reduction
of  AN emissions vented to the system.  The selection of BAT for process
emission sources at ABS/SAN facilities is dependent upon several  factors.
The applicability of HMCT for the production of the variety of ABS/SAN
resins being manufactured is not known at this time.  A plant will  have to
conduct pilot studies to determine if the technology is applicable  to its
process.  There are no data at present to indicate that this technology
cannot be used for many types of resins being produced by the industry.
Thermal incineration is a commonly employed AN emission control  method.  It
reduces emissions and can supply some plant energy needs.  A combination of
HMCT with thermal incineration of vent gases would give the greatest AN
emission reduction (estimated at 99.9 percent), but may not be applicable
at  all emulsion processes.  Either control method could be used separately,
as  appropriate to the facility, to give good emission control process
emission sources.  Fugitive emission controls, like those in the other  AN
industry segments, could successfully reduce these emissions from this
process.
     The suspension, mass-suspension, and bulk processes for ABS/SAN production
can emit AN to the atmosphere from reactor vents, SAN slurry steam strippers,
thin film evaporators, and fugitive sources.  The SAN steam stripper usually
employs a chilled styrene scrubber to recover unreacted monomers.  Emissions
are reduced while improving process operations.  This equipment r.ay achieve
                                    1-5

-------
about 72 percent AN emission reduction.  Scrubber vent gases sent to an
incinerator/boiler can be controlled at 98 percent efficiency.   These
controls could yield a total 98 percent or greater reduction of AN emissions
from the SAN process.  The other processes may have emission points vented
to an incinerator for 98 percent control  efficiency.   These systems are
potentially BAT for these operations.  Fugitive emissions would be controlled
as previously discussed.
1.4  NBR PRODUCTION
     Latexes and NBR are manufactured by five companies at seven locations
in the  U.S.  There are plants in Ohio (3), Texas, Delaware, Louisiana,  and
Kentucky.  Domestic production in 1979 totaled 75,360 Mg (82 percent of the
present 92,000 Mg/yr capacity) with a merchant sale value of $91  million.
Latexes are used in adhesives, paper coatings, and fabric saturants.  Syn-
thetic  rubber materials are highly resistant to oil  and function over a
wide temperature range.  They are valued in the automotive industry for
hoses,  gaskets, and seals.  Combined with other polymers such as polyvinyl
chloride (PVC) elastomers may be used in a variety of products (footwear,
luggage, or recreational items).
     The industry has had an average annual growth rate of 1 percent per
year from 1969 through 1979.  Capacity utilization has fluctuated between
72 and  96 percent during this period.  Production for NBR and latexes is
expected to increase 2 to 3 percent per year through  1985.
     Nitrile rubbers and latexes are produced by emulsion polymerization of
AN with butadiene (BD).  The basic procedures are similar to the emulsion
polymerization processes used for ABS/SAN resins.  The NBR reactions are
carefully controlled to produce the desired polymer chain, then quickly
stopped.  Monomer conversion to polymer can vary between 75 and 90 percent.
Residual ronomers are steam stripped from the latex slurry, condensed,  and
recycled to the process.  The NBR process can emit AN to the atmosphere
from the reactor pre-mix tanks, reactor vents, the monomer condenser,
coagulation tank vents, and polymer dryers.  There are fugitive emissions
from AN storage tanks and process equipment in AN service.
     Control  of AN emissions from NBR processes revolves around the  latex
stripper and AN recovery scrubbers.  A well designed  steam stripper systen
                                   1-6

-------
can remove 99 percent of residual  monomer from the slurry.   This greatly
reduces potential AN emissions from coagulation tanks and polymer dryers.
A chilled water AN scrubber in series with an amnonia chiller can recover
unreacted monomer stripped from the slurry for recycle to the pre-mix
tanks.  The estimated efficiency of the scrubber system is  in excess of
96 percent.  Connecting pre-mix tank vents and reactor vents  to  the  scrubber
system would further reduce AN emissions.  This system potentially represents
BAT for reducing AN emissions from these NBR emission sources.   Incinerating
scrubber and coagulation tank vent gases in the steam boiler  and partially
recycling dryer exhaust could further reduce AN emissions from  this  process.
Fugitive emissions could be significantly reduced by regularly  scheduled
inspection and maintenance of process equipment.
1.5   EXISTING STATE AND FEDERAL REGULATIONS
      There are no State regulations that are specifically directed at
limiting AN emissions to the atmosphere.  The AN industry has plants in 18
different States.  Regulations limiting volatile  organic compound (VOC)
emissions from process operations  and storage tanks have been enacted in
eight of these States.  Three States have VOC emission regulations covering
storage tanks only.  The remaining States do not have any VOC emission
control regulations.
      Worker exposure to AN i s limited by OSHA regulations to  2  ppm over a
time-weighted 8-hour averaging period or 10 ppm averaged over any 15-minute
period of the work day.  These regulations protect worker safety, but do
not specifically control or limit  AN emissions to the atmosphere.  The area
monitoring systems often employed  to assure workplace compliance with these
regulations may not detect all fugitive emissions - some AN fugitive emissions
can occur in places where workers  are not exposed.  The plant may comply
with  the regulations by simply venting vapors to  the air.  The  OSHA regulations
could be augmented by equipment inspection and maintenance.
      The EPA is currently developing VOC emission standards for the  synthetic
organic chemical  manufacturing industry (SOCMI).   This program  is engaged
in formulating new source performance standards (NSPS) for  VOC  emissions
from  new SXMI processes and Control Technology Guidelines  (CTG) documents
for regulating fugitive VOC emissions from SOCMI operations in  ozone
                                   -7

-------
nonattainnent areas.  The NSPS regulations nay apply to rononer'plants
being built or expanded (plants already have permits and regulations are in
draft stages now), but are not applicable to the existing AN enission
sources.
     There are currently 33 plants using or manufacturing AN in the U.S.
Many plants (24) are located in industrialized areas that are nonattainment
for ozone.  The remaining nine plants are in attainment areas.  The draft
CTG documents for controlling fugitive organic emissions from SOCMI opera-
tions would apply to plants in nonattainment areas where the State has
requested an extension to reach attainment.   Acrylic fiber plants  are not
covered in this CTG.  The draft CTG document outlining control methods for
volatile  organic liquid storage (VOL) in fixed and floating roof tanks
would apply to all AN facilities in nonattainment areas.
1.6  PREFERRED METHOD OF SAMPLING AND ANALYSIS OF AN EMISSIONS
     The EPA has not developed specific reference method (RM) sampling and
analysis  procedures for AN point source emissions.  There are VOC  sampling
and analysis methods (Federal  Register, October 3, 1980, RM 25) that should
provide valid data with some modifications.   Data collected to date has
been gathered by a variety of sampling techniques (carbon absorption/
desorption and flask grab samples) and analyzed by a gas chroma to graph.
The use of RM 25 sampling techniques with a properly prepared gas  chromato-
graph (for qualitative analysis of AN) should yield reliable, valid quantitative
AN emi ssions data.
                                     1-8

-------
                            2.  INTRODUCTION

     A study of the AN industry was perfomed  in  this  survey.   The  AN
industry includes AN monomer synthesis, acrylic and modacrylic  fiber opera-
tions, ABS/SAN resin manufacture,  and NBR production.   These  industries  are
classed by the U.S. Department of  Commence under  SIC codes  2821,  2822,
2824, and 2869.  Monomer operations produced 915,000 Mg of  AN in  1979
(195,000 Mg of this AN was exported), supplying the principal  feedstock  for
the remaining industry segments.  Acrylic fibers  manufacture  in 1979 consumed
approximately 43 percent (345,000  Mg) of domestically  available AN  monomer.
These fibers are used in apparel and home furnishings.   The ABS/SAN plastic
resins industry was the second largest consumer of AN  monomer,  using
157,000 Mg in 1979.  The ABS/SAN thermoplastics are formed  into automotive
parts, appliances, plastic pipe and pipe fittings, and small  consumer  pro-
ducts (luggage, electronics, reacreational  items,  and  furniture).   The  NBR
processes consumed 88,000 Mg of AN in 1979 for the production of  automotive
parts (hoses, seals, gaskets), footwear, luggage,  and  recreational  items.
Plants contacted in this study are listed in Tables 2-1 and 2-2.   The  study
has not closely examined manufacture of adiponitrile or acrylamide  since
these processes are very small segments of the industry with  low  AN emission
estimates.
     The goal of the survey was to determine the  need  for a NESHAP  for  the
AN industry under Section 211 of the Clean Air Act. The requirements under
Section 112 of the Clean Air Act direct the EPA to develop  proposed regula-
tions establishing emission standards for any  air pollutant added  to the
list of hazardous air pollutants by the Administrator.  The proposed NESHAP
Policy and Procedures for Identifying, Assessing,  and  Regulating  Airborne
Substances Posing a Risk of Cancer (40 CFR Part 61) states  that the standard
developed for hazardous air pollutants must, in the judgment  of the Adminis-
                                    2-'

-------
           TABLE 2-1.  LIST OF RECIPIENTS OF SECTION 114 LETTERS
           Plant and Location
     Product
American Cyanamid
  Westwego, Louisiana

DuPont
  Be a u mo n t, Texas

Monsanto Company
  Al vin, Texas

B.F. Goodrich/Abtec
  Louisville, Kentucky

ftonsanto Company
  Indian Orchard, ftessachussetts

American Cyanamid
  Milton, Florida

Tennessee Eastman Co.
  Kingsport, Tennessee

B.  F. Goodrich
  Akron, Ohio
AN Mo none r


An Monomer


AN Monomer


A3S/SAN


ABS/SAN


Acrylic Fibers


Acrylic Fibers


Nitrile Elastomers
                                     2-2

-------
                    TABLE 2-2.  LIST OF PLANTS VISITED
     Plant and Location
      Product
Borg-Warner Chemicals
  Washington, West Virginia

Monsanto Plastics and Resin Co
  Addyston, Ohio

Uniroyal Chemical
  Painesville, Ohio

Monsanto Textiles Co.
  Decatur, Alabama

Badische Co.
  Williams burg,  Virginia
ABS/SAN


ABS/SAN


Nitrile elastomer


Acrylic Fibers


Acrylic Fibers
                                    2-3

-------
trator, "provide an ample margin of safety to protect the public health
from such hazardous air pollutants."  This study has collected background
information toward the goal  of developing NESHAP regulations for the Afl
industry if AN is listed by the Administrator as a hazardous air pollutant.
     The specific objectives of this survey were to confirm and update Afl
emissions data, evaluate process AN emission control technology for each
industry segment, determine possible AN emission reductions using control
techniques observed during the study, and examine the applicability of the
proposed NESHAP generic standard for control of fugitive emissions  from the
AN industry.  This information will be supplied to the EPA for making
exposure/risk evaluations for the AN industry and determining the feasibil-
ity of standard development.  The data base formed can then be used to
assess emission control methods and alternatives in conjunction with their
environmental and economic impacts for each industry segment.
     The study objectives were met through an extensive search of the
pertinent literature (previous EPA studies, EPA library files, and  the
Library of Congress) in conjunction with telephone contacts with trade
associations, plants in each industry segment, government and State agencies,
and industry experts.  A series of plant visits to plants wi th a variety of
process types was conducted to get firsthand knowledge of process operations
and emission control programs.  Information requests under Section  114 of
the Clean Air Act were sent to selected plants to enlarge the data  base.
                                     2-4

-------
                       3.  DESCRIPTION OF INDUSTRY

     Acryloni trile monomer is used as a feedstock chemical  by the  acrylic
fiber, plastics, synthetic rubber, and chemical  industries.   The various
end uses of AN monomer are presented in Table 3-1.  The  acrylic fiber
industry is currently the largest single consumer of AN.   It has large
markets in the apparel and home furnishing industries.   In  the plastics
industry, AN is polymerized with styrene and butadiene  for  the production
of ABS/SAN resins.  These are used in the automotive, construction,  and
electronic industries.  The production of nitrile rubbers  is another major
use of AN monomer.  These synthetic rubbers  are  temperature  and solvent
resistant.  They are used principally by the automotive  industry for under-
the-hood seals and hoses.  Other markets for AN  monomer  are  in the  production
of adiponitrile (intermediate feedstock in Nylon 6/6 manufacture),  acrylamide
(used in wastewater treatment and packaging  materials),  and  nitrile  barrier
resins (formed into lightweight food and specialty containers).  There  is
also a considerable export market for raw AN.
3.1  SOLRCE CATEGORY
3.1.1  Monomer Production
     There are four companies currently producing AN monomer at six  domestic
locations.  The location and capacity of each monomer facility is  listed  in
Table 3-2.  The current domestic production  capacity is  972,700 Mg/yr.
Monsanto Company is the largest producer with 390,000 Mg/yr of capacity  at
its Alvin and Texas City, Texas, plants.  Other producers  are DuPont,
Vistron Corporation, and American Cyanamid.
     Capacity in the industry is projected to increase  276,000 Mg/yr to  a
total of 1,248,700 Mg/yr by 1982 as a result of new plant  construction  and
expansion.  The Monsanto Company will increase capacity  by 95,000 Mg/yr
                                   3-1

-------
         rABLE 3-1.  1979 AN PRODUCTION AND CONSUMPTION FOR END USE CATEGORIES
Major marke ts
Fibers
ABS Resins
SAN Resins
Nitrile elastomers
Exports
Adi ponitrile (from AN)
Ac ryl amide
Barrier resins
Mi seel laneous
Lhaccounted
Actual AN production
1979 Production
(Mg)
345,000
540,000
56,000
75,400
	
N/A
N/A
N/A
N/A

Total AN
consumed
(Mg)
345,000a
140,400b
16,80CC
26,100d
195,000
86,600e
28,500f
ll,400f
36,500f'9
26,000h
915,300
% Total
AN consumption
37.7
15.3
1.8
2.9
21.3
9.5
3.1
1.2
4.0
3.1
100.0
aSased on 1 Ib AN to 1 Ib fiber.
bBased on 0.26 Ib AN to 1  Ib ABS.
cBased on 0.30 Ib AN to 1  Ib SAN.
dBased on 0.35 1 b AN to 1  Ib dry elastomer and 0.32 Ib  AN  to  1  Ib  latex.
eBased on Monsanto facility in Decatur,  Alabama,  at full capacity.
rBased on 1977 growth rates given by SRI.
9Fumigants for tobacco, super absorbants,  fatty amine  production,  cyanoethylation
 of alcohols and amines.   SRI estimates  33,000 Mg produced  in  1977.
L_
 This category contains any inventory accumulated.
SOURCES:  Textile Organon, February 1980;  1980 Facts and Figures of  the  Plastics
          Industry; U.S.  International  Trade  Commission, Synthetic Organic
          Chemicals,  1979; SRI Chemical  Economics Handbook

-------
                    TABLE 3-2.   CURRENT AN  CAPACITY
                                                         Operating capacity
       Company             Plant location                     (Mg/yr)


American Cyanamld          testwego,  LA                       120,200
  Industrial  Chemicals
  Divi sion

E.I. duPont de Nemours     Beaumont,  TX                       158,700
  & Company,  Inc.          Memphix,  TN                        122,400
  Petrochemicals Division

Monsanto Company           Al vi n, TX                          199,500
  Monsanto Chemical         Texas City, TX                     190,500
  Intermediates Co.

Standard Oil  of Ohio       Lima, OH                            181,400
  Vistron Corporation
  Subsidiary
Total                                                        972,700
SOIRCE:  SRI Directory of Chemical  Manufacturers,  1980
                                       3-3

-------
when a third reactor is brought on-line in 1981 at the Texas City plant.
Vistron is currently constructing a 181,000 Mg/yr plant at Green Lake,
Texas, scheduled to cone on-line by 1982.
     3.1.1.1  Production and Capacity Utilization.  Data on AN production
over the past decade are presented in Table 3-3.  Total monomer production
in 1979 was 915,000 Mg.  The monomer producing companies captively consumed
approximately 55 percent of their production (499,000 fig).  The remaining
416,000 Mg was sold on the merchant market at an average price of about
53 cents per kilogram  (current 1981 price is 86 cents/Kg).  The total  1979
sales  for the monomer  industry were $218 million.   Exports amounted to
195,000 Mg, or 21 percent of production.
     Over the period 1969 to 1979, acryloni trile production grew at an
average annual rate of 6 percent (525,000 to 915,000 Mg).  Growth over this
period led to a series of production capacity increases (from 1971 to  1978)
through plant expansions or new plant construction.  Monomer production
during this period was erratic, however, owing to fluctuations in the
acrylic fibers and ABS/SAM resin markets,  feedstock shortages, and the
1974-75 economic recession.  Capacity utilization had averaged 85 percent
through the decade, with 92 percent capacity utilization in 1979.
3.1.2  Acryl ic Fibers
     The acrylic fiber manufacturing industry is the largest consumer of AN
monomer.   It uses approximately 48 percent of domestically available AN.
                                                             2
The  industry consumed  approximately 345,000 Mg of AN in 1979.
     Acrylic fibers are valued for their resemblance in warmth and softness
to natural wool and their ability to be processed into high-bulk, lightweight
yarns.   They are used in apparel manufacture (hosiery, sweaters, pile
fabrics, and craft yarns) and hone furnishings (carpets, rugs, furniture
coverings, and blankets).  There is also a significant export market for
acrylic fibers.
     Acrylic and rodacrylic fibers are produced by five companies at six
locations in the United States.  The largest producers are Monsanto Company
and  DuPont.  Other producers are American Cyanamid, Badische Corporation,
and  Eastnan Kodak.   These plants have a total fiber production capacity of
392,000 Mg/yr in 1979.  Individual plant capacities are  listed in Table 3-4.
                                       3-4

-------
           TABLE 3-3.   AN PRODUCTION, CAPACITY, AND CAPACITY
                          UTILIZATION 1969-1979

Year
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
Capaci ty
(Mg/yr)
523,800
569,200
612,300
612,300
716,700
716,700
762,000
762,000
762,000
972,700
972,700
Production
(Mg)
523,800
471,300
444,000
505,600
614,200
640,300
550,900
688,500
746,600
794,800
915,000
Capaci ty util ization
( °'\
\'°)
100
83
73
83
86
89
72
90
98
82
94
SOURCES:  SRI, Chemical Economics Handbook,  U.S.  International  Trade
          Commission, Synthetic Organic Chemicals,  1979
                                     3-5

-------
              TABLE 3-4.  ACRYLI.C FIBER PRODUCERS AND CAPACITIES
ipany
                      Plant location
                         Annual  acrylic
                         fiber capacity
                            (Mg/yr)
               Annual  modacrylic
                 fiber capacity
                    (Mg/yr)
rican  Cyanamid
'ibers  Division

i sche  Company

 . duPont de  Nemours
i  Company
extile  Fibers  Dept.

 tman  Kodak  Company
'ennessee Eastman
livi sion

isanto  Company
bnsanto Textiles
ompany

  TOTAL
                      Pensacola, FL
Wil liansburg,  VA

Camden, SC
Waynes bo ro, VA
                      King sport, TN
                      DecaUir, AL
                             54,400
31 ,800

72,600
63,500
                            136,000
                                                  358,300
                                                                         4,500
                                                  18,100
                     11,300
                                                  33,900
RCE:  SRI Directory of Chemical Manufacturers, 1980
                                          3-6

-------
     Monsanto, DuPont, and American Cyanamld all produce fibers from captive
supplies of acryloni trile.  The Badische and Tennessee Eastman Companies
purchase acrylonitrile on the open market.  Vistron is the only AN monomer
manufacturer which is not engaged in acrylic fiber production.
     3.1.2.1  Fiber Production and Capacity Utilization.  The production of
acrylic and modacrylic fibers totaled 345,000 fig worth approximately $450 mil-
lion in 1979, an increase of 5 percent from 1978 shipments.   Over the
10-year period 1969 to 1979, domestic acrylic fiber production has averaged
a 3.6 percent growth rate per year.  Fiber production increased to 353,000 Mg
        4
in 1980.   These data are given in Table 3-5.  Information on end uses  of
acrylic fibers is presented in Table 3-6.
     The data in Table 3-6 show that although total production increased
substantially between 1969 and 1979, most of this growth has occurred in
export trade.  Exports increased from 41,000 Mg in 1977 to 101,000 Mg in
1979.  This  increase was  primarily a result of lower fiber production costs
in the United States owing to U.S. government price controls on crude oil
(lowering costs of monomer feedstocks) and the weakened dollar (making  U.S.
products less expensive overseas). '°  Trade conditions have changed,
however, and  the export market is decreasing.  Ebmestic fiber shipments
fell  to 248,000 Mg in 1979 after reaching a high of 281,000 Mg in 1977.
This downturn in home markets is due to the declining use of acrylic fibers
in the carpet and rug markets, where lower cost nylon and polyester fibers
have taken control.  Producers see apparel applications as the only growth
market for acrylic fibers.
     Capacity utilization rates for the acrylic fiber industry were low in
the early 1970's, but increases have occurred owing to gains in export
trade and a  reduction of  production capacity from 392,000 Mg in 1979 to
376,000 Mg by late 1980.   These two factors led to an increase in capacity
utilization  from 86 percent to 90 percent,  No firms are planning expansions
or new facilities at this time.
3.1.3  ABS/SAN Resins
     The manufacturers of ABS and SAN resins are the second largest consumers
of AN monomer.  This industry consumed approximately 157,000 fig of AN  in
1979.
                                     3-7

-------
     TABLE  3-5.   ACRYLIC  FIBERS  PRODUCTION  AND  CAPACITY  UTILIZATION

Year
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
Caoaci ty
'(Mg)
293,000
299,000
317,000
342,000
357,000
374,000
375,000
394,000
394,000
392,000
380,000
Production
(Mg)
241,800
223,100
247,300
306,600
336,600
286,300
238,000
281,700
321,500
329,200
345,200
Capaci ty util ization
(V }
(,o )
83
75
73
83
94
77
63
71
82
84
90
SOURCE:   Textile  Organon,  November 1980 and  earlier years
                                        3-8

-------
          TA3LE 3-6.   END LSES OF  ACRYLIC  AND MODACRYLIC  FIBERS
                                     (Mg)
     End use                                Year

                            1975       1976       1977       1978       1979


Hosiery, sveaters,
  and craft yarn           87,700    95,300   119,000    109,700    108,300

Pile fabrics               32,100    33,500    39,700     37,400    29,500

Other ci rcular and
  flat knit                30,800    38,400    50,700     42,200    44,800

Blankets                   15,400    15,600    22,800     25,000    20,600

Other broad woven           6,400     9,800    12,300     13,000    19,000

Carpet face yarns          37,500    39,600    33,000     28,500    21,800

All other                   2,100     2,500     3,600      3,500      3,700

  Total  domestic          212,000   234,600   281,100    259,300    248,000

  Exports                  39,100    49,900    41,100     68,400    101,400

Total production          251,100   281,700   321,500    329,200    345,200


SOLRCE:   Textile Organon, February 1980
                                       3-9

-------
     The major applications of ABS resins are in pipe and pipe fittings,
automotive components, and appliances.  These products comprised 65 percent
of domestic AN consumption in 1979.  Other uses of ABS are in business
machines and telephones, electronic products, recreational equipment,
furniture, and luggage.  Major uses of SAN resins are in ABS compounding,
automotive instrument  lenses, and  in injection molded parts for housewares
and packaging.
     The average AN content of ABS and SAN resins is 25 and 30 percent by
                     g
weight, respectively.   There are  numerous types of these resins with dif-
ferent physical properties (chemical and temperature resistance) that vary
with AN content.
     There are seven companies at  13 locations producing ABS resins. They
are listed in Table 3-7.  Total domestic ABS capacity is estimated at
843,000 Mg/yr.  The  largest producers are Borg-Warner, Monsanto, and Dow
Chemical, accounting for 78 percent of total U.S. capacity.   The remaining
capacity is divided among small producers (U.S.S. Chemicals, Abtec Chemical,
Mobil Chemical, and Carl Gorden Industries).  Monsanto is the only ABS
producer having a captive supply of AN monomer.  Planned capacity changes
in the ABS industry will have the  net effect of increasing the current
capacity of 843,000 Mg to 886,000  Mg/yr by 1982.  Borg-Warner is construct-
ing a 68,000 Mg/yr ABS facility at Port Bienville, Mississippi, which is
slated to come on-line in 1982.  Mobil Chemical recently announced its
withdrawal from the ABS industry;  Mobil's Joliet, Illinois, plant will be
converted to polystyrene manufacture.
     SAN resins are produced by Dow Chemical and Monsanto.  Total present
capacity is about 84,000 Mg/yr.  Production of SAN resins in 1979 was
56,000 Mg.  Average annual growth  rate was about 6 percent per year between
1971 and 1979.  No changes in SAN  resin capacity are anticipated in the
near future.
     3.1.3.1  ABS/SAN Production and Capacity Utilitization.  Information
on production and sales of ABS and SAN resins is presented in Table 3-8;
ABS consumption by end use is given in Table 3-9.  Production of ABS resins
was 540,000 Mg in 1979 (64 percent of capacity), representing an increase
of 5 percent over 1978 production.   During the period 1971 to 1979, produc-
tion increased at an average rate  of 7.1 percent.  Merchant sales (domestic
                                      3-10

-------
                TABLE 3-7.   ABS AND SAN PRODUCERS AND CAPACITIES

Company
tec Chemical Company
rg-Warner Corporation
^nrn-Wainna^ Cham i rale
Location
Louisville, KY
Ottawa, IL
VJachi nrr+nn WV
ABS capacity
(Mg/yr)
45,400
281,200
SAN capacity
(Mg/yr)
—
J.S.A.

n Chemical  U.S.A.            Allyns Point, CN
                            Ironton, OH
                            Midland, MI              136,000             61,200
                            Pevley, MO
                            Torrance, CA

"1  Gorden Industries, Inc.  Worcester, MA              4,500              —
Hammond Plasticis  Division

jil  Chemical  Corporation    Joliet, ILa               24,900
3etrochemicals Division

isanto Company              Addyston, OH
Monsanto Plastics  and       Muscatine, IA            238,100             22,700
Resins Company              Springfield, MA

> Chemicals                 Scotts Bluff, LA         113,400              —
                                                     843,500             83,900
 n's facility has been converted to polystyrene production in 1981.

 !RCE:   1980 Facts and Figures of the Plastics Industry
                                             3-11

-------
TABLE 3-8.  ABS PRODUCTION AND CAPACITY UTILIZATION

Year
1971
1972
1973
1974
1975
1976
1977
1978
1979
Production (Mg)
ABS
311,100
367,400
397,800
388,700
291,200
448,100
484,900
514,400
540,200 .
SAN
34,900
44,900
54,400
54,000
49,400
51,200
52,200
54,400
56,200
SOURCES:  1978, 1980 Facts and Figures of the Plas-
          tics Industry; Synthetic Organic Chemicals,
          U.S. International Trade Commission, 1971,
          1972
                              3-12

-------
 TABLE 3-9.  TRENDS  IN  DOMESTIC  CONSUMPTION  OF  ABS  RESINS  BY END USES 1973-1979

id use
ipl lance
itomoti ve
islness machines
and telephones
•n sumer
electronics
i rn i tu re
iggage & cases
idifiers
pes & fi ttings
icreational
ital
Consumption
(Mg)
48,100
69,900

18,100

9,100
6,800
10,900
6,800
115,700
69,900
406,000
1973
% of total
11.8
17.2

4.5

2.2
1.7
2.7
1.7
28.5
17.2
100.0
1979
Consumption %
(Mg)
122,005
77,100

29,500

18,100
11,300
11,300
11,300
165,600
83,900
571,500
of total
21.4
13.5

5.2

3.2
2.0
2.0
2.0
29.0
14.6
100.0
1973-79
average
annual
growth
rate (%}
16.9
1.6

8.4

12.1
8.8
0.6
8.8
6.2
3.0
5.8
)URCE:   1978,  1980 Facts  and  Figures  of  the  Plastics  Industry
                                           3-13

-------
and export) of ABS resins in 1979 totaled about 508,000 fig (94 percent of
                                                     q
domestic production) and were valued at S548 million.   Exports totaled
35,000 Mg  (approximately 6.5 percent of production).
     Capacity utilization rates for ABS/SAN producers have been low for
several years, and were under 70 percent for both plastics in 1979.  The
low utilization of capacity is attributable to the optimistic expectations
by leading ABS producers in the 1970's.  A large excess capacity was built
before downturn in the ABS markets.    The ABS market has been losing
ground to  competing thermoplastics such as PVC, polypropylene, and high-impact
polystyrene owing to the higher production costs of ABS.
3.1.4  NBR Elastomers
     Latexes and NBR are made by the copolymerization of acrylonitrile  and
butadiene.  These synthetic rubber materials are highly resistant  to oils
and function over a wide range of temperatures.  The ratio of AN to butadiene
varies depending on the characteristics desired in the final  elastomer
product.   The AN content typically varies between 30 and 40 percent of  the
elastomer by weight.
     Elastomers are used mainly in the automotive industry in hoses, seals,
and gaskets.  They may be combined with other polymers such as PVC to
produce a  variety of products (footwear, luggage, and recreational  items).
Nitrile latexes are used as adhesives, paper coatings, and fabric  saturants.
     Elastomers and latexes are manufactured domestically by five  companies
at seven locations as listed in Table 3-10.  Total  estimated  domestic
capacity is 92,000 Mg/yr.  The two largest producers, BF Goodrich  Company
and Goodyear Tire and Rubber Company, account for 61,000 Mg/yr of  capacity.
Uhiroyal,  Inc., Reichold Chemicals, and Copolyrer Rubber and Chemical
Corporation operate the remaining individual  elastomer facilities.   The
Copolymer plant was built in 1969 and was the last capacity addition in the
industry.   All of these companies purchase AN monomer on the  merchant
ma rke t.
     3.1.4.1  Production and Capacity Utilization.   Production of  NBR grew
at an  average  annual  rate of one  percent from 1969  to 1979.  Total  domestic
production in  1979 was 75,360 Mg, requiring approximately 26,000 Mg of  Mi
monomer.   This was a decrease from a peak production of nearly 88,000 Mg in
1974 as shown  in Table 3-11.  Merchant sales of $91 million accounted for
58,000 Mg, or 77 percent of production, with the remaining output  consumed

                                      3-14

-------
         TABLE 3-10.   NITRILE ELASTCJ1ER PRODUCERS AND CAPACITIES
                                                                Capaci ty
     Producer                       Location                    (Mg/yr)


Co polymer Rubber and              Baton Rouge,  LA                5,000
  Chenical  Corporation

BF Goodrich Company               Akron, OH                     14,000
  Chenical  Division               Louisville,  KY                28,000

Goodyear Tire and Rubber          Akron, OH                      3,000
  Company                         Houston,  TX                    16,000
  Chemical  Division

Reichold Chemical Company         Cheswold, DE                   10,000

Ihiroyal, Inc.                    Painesville,  OH               16,000
  Chemical  Division                                            	

  Total                                                          92,000


SOURCE:   SRI Directory of Chemical  Manufacturers, 1980;  "Enission,  Process,
         and Control Technology Study of the ABS/SAN,  Acrylic  Fiber,  and  NBR
         Industries," Pullman Kellogg, April 1979
                                    3-15

-------
TABLE 3-11.  NITRILE ELASTOMER PRODUCTION AND CAPACITY UTILIZATION

Year
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
Production
(Mg)
70,000
68,100
66,500
74,600
84,300
88,000
55,300
72,500
70,100
73,400
75,400
Capacity util ization
(V }
(/a)
76
74
72
81
92
96
60
79
76
80
82
                                     3-16

-------
captively.  Exports were 20 percent of production and were valued at S21
million.
     Capacity utilization generally varied between 72 and 96 percent from
1969 to 1979, although this figure fell to 60 percent during the 1974-75
recession.  Current capacity utilization is 82 percent.
3.1.5  Other Uses of AN
     There are several additional domestic markets for AN.  The most notable
markets are in the manufacture of adiponitrile, acrylamide, and nitrile
barrier resins.
     Adiponitrile (AD) serves as an intermediate feedstock chemical  in the
production of Nylon 6/6.  This chemical is produced by several  different
processes.  Monsanto Company, at its Decatur, Alabama plant, operates the
only process that uses AN.  It is estimated that this process consumed
72,500 Mg of AN to produce 67,000 Mg of AD in 1977.12  This accounted for
about 10 percent of AN consumption in that year.   Total  AD production was
426,000 Mg for all processes in 1977.13
     Acrylamide manufacture consumed 23,500 Mg of AN to produce 31,000 Mg
of product in 1977.  Markets for acrylamide are in wastewater treatment
flocculants (products used to aid in sewage dewatering) and other products
for the water treatment industry.  These applications accounted for 35 to
40 percent of all acrylamide used in 1977.   Acrylamide is also employed in
the manufacture of paper-treating products.  The paper industry accounted
for an estimated 25 percent of acrylamide consumption in 1977.   Other
                                                                        14
markets for acrylamide are drilling mud additives and textile treatment.
There are currently three companies in the U.S. manufacturing acrylamide.
American Cyanamid operates facilities at Linden,  New Jersey, and New Orleans,
Louisiana, with a total estimated capacity of 36,000 Mg/yr.  Dow Chemical
operates a 23,000 Mg/yr plant at Midland, Michigan,  and Nalco Chemical runs
a 4,500 Mg/yr plant at Garysville, Indiana.
     The production of nitrile barrier resins is another market for AN.
These resins are valued for high gas barrier and chemical resistance
properties.   They are used in food containers and packages and for packaging
abrasive or caustic materials (agricultural chemicals, industrial solvents,
and cleaning products).   The consumption of AN in nitrile barrier resin
production was approximately 9,000 Mg in 1S77.    Nitrile barrier resins
                                     3-17

-------
were first developed in the mid-1970's for use in lightweight, impact
resistant beverage containers to replace glass bottles in the carbonated
beverage market.  The Food and Drug Administration (FDA), however, after
conducting studies on the leaching of residual AN rononer from the container
into the beverage, determined AN use in these containers would create a
health hazard.  Based on the data from these studies, the FDA banned the
use of nitrile barrier resins for beverage containers in February 1977.
The FDA also set allowable threshold AN levels in other types of food
containers.  This action damaged major market potential  for nitrile barrier
resins.  Vistron Corporation has since been successful in developing container
resins for other food product use which meet FDA restrictions on AN.  These
materials are currently used in resealable lunchneat containers  and margarine
tubs.  Other applications are in small blow-molded containers for liquid
paper and nail polish bottles.17'18'19
3.2  GROWTH  IN AN PRODUCTION
     Annual average growth of 4.4 percent in AN monomer production is
estimated through 1985 as listed in Table 3-12.  This figure Is  derived
from AN use  in consuming industries and compares favorably with  most recent
                       20
estimates of AN growth.    These estimates are more modest than  predictions
made several years ago, reflecting the current business slowdown, recent
changes in the U.S. foreign trade position, and competition for  AN-derived
products from other less expensive substitutes such as polyvinyl chloride
                              21 22
(PVC), polystyrene, and nylon.  '
3.2.1  Acrylic Fibers
     Production of acrylic fibers is projected to increase at a  rate of
                                23 24 25
1  percent per year through 1985.  '  '    Domestic markets, with the excep-
tion of apparel, are expected to show little growth because of strong com-
petition from lower cost nylon and polyester fibers.  Apparel markets are
                                                                 26
expected to increase approximately two to three percent per year,   mainly
through those products where the wool-like properties of acrylic fibers are
desired (sweaters and outer garments).  Exports, which reached historic
levels in 1979, fell  to 17 percent of fiber production (84,000 Mg in 1980"7)
owing to the establishment of a permanent anti-dumping duty  by the  European
    "*                                                                   9n
Economic Community which led most U.S. producers to raise export prices.
                                     3-18

-------
 TABLE  3-12.   ESTIMATED GROWTH IN AN END MARKETS AND AN DEMAND FROM 1980 to  1985

End Use
Fibers
ABS resins
SAN resins
Nitrile elastomers
Exports
Adi ponitrile
Ac ryl amide
Barrier resins
Mi seel laneous
Unaccounted
Total
Estimated annual rate
1979 AN
Consumption
(Mg)
345,000
140,400
16,800
26,100
195,000
86,600C
28,500
11,400
36,500
26,500
of growth from
Estimated Annual
Growth Rate
In Production and
AN Consumption
1.0
5.0
5.0
2.5
N/A
N/A
9.0d
12. Od
5.0d

1980e
1985
Estimated
AN Consumption
(Mg)
371,000
179,200
21,400
30,300
175,000b
86,600
47,800
22,500
48,900
30,200
1,012,900
4.4%
 EEA estimate from Table 3-1.
 EEA estimate.
GAssumes no extra capacity has been added.
 Growth rates from SRI Chemical Economics Handbook.
eAcrylonitrile production in 1980 was reported to have fallen to 316,000 Mg  (U.S
 International Trade Commission).
                                           3-19

-------
Over the longer term, increasing feedstock costs for U.S. producers following
oil decontrol and a strengthened U.S. dollar will  suppress growth In this
     29
area.    Production in 1985 is projected at 371,000 Mg.
3.2.2  ABS/SAN Resins
     Growth in ABS and SAN resins is projected to  be 5 percent per year
through 1985.  '    Higher growth rates predicted  by industry sources   nay
be optimistic because of current low capacity utilization, which keeps unit
prices high, and competition from less expensive substitutes such as PVC
and polystyrene which compete with ABS resins in the pipe and automotive
markets.  Healthier growth rates will depend on how well ABS can hold off
its competitors in these areas and whether new markets can be developed.  A
promising area is the use of ABS in refrigerator liners.  One manufacturer
expects this market could stimulate an additional  90,000 Mg/yr of demand
                     33
within several years.    Both ABS and SAN are consumed mainly in durable
goods  (housing, automotive, appliance) which are suffering due to current
economic conditions.  Projected growth for ABS will  probably be slow in the
near term and should improve with business conditions.
3.2.3  NBR  Elastomers
     Latex  and NBR production is projected to increase two to three percent
per year through  1985.   '  '    The greatest potential growth area is
expected to  be as a plastics modifier, mainly in blending with PVC.
Growth will  be low because nitrile rubbers are well  established in most of
their  markets and increases will be determined by general economic condi-
tions.  The  Rubber Manufacturers Association (RMA) has estimated production
                     3fl
of 88,000 Mg in 1985.    This vould require roughly 33,000 Mg of AN in
1985.  Capacity utilization would be 96 percent at this  production level.
No expansions are currently planned, however.
3.2.4  Other Areas
     No data are  available on current production levels  for  adi ponitrile,
acrylamide, nitrile barrier resins, and other miscellaneous  uses of acrylo-
nitrile.  Production and growth rate estimates made in  1977  have been
employed to estimate 1985 production and AN consumption  for  the  latter
three  markets.  The consumption of AN for adiponitrile  production  is esti-
mated  at 86,500 Mg, based on Monsanto's Decatur, Alabama, adiponitrile
                                    39
facility operating at full capacity.    No expansion of  the  'tmsanto  facility
                                    3-20

-------
or other new capacity for adiponitrile production fron AN is known, although
some sources feel an increase in capacity at 'fonsanto is possible.
Acrylamide and nitrile barrier resins could each denonstrate strong growth.
Aery 1 amide use in drilling muds could increase substantially with tertiary
oil recovery.
     Monsanto Company filed  suit against the FDA ban on AN use in beverage
containers in 1979, and in light of new data, the Court remanded the deci-
sion of the FDA  Commissioner for reconsideration.  This data concerns AN
levels in a new  generation of beverage containers developed by Monsanto.
The FDA is still reconsidering the new evidence and will issue a ruling in
                   41
the summer of 1981.    A favorable ruling for producers could open up AN
use for the beverage container market and stimulate uses in other food
packaging.
     AN monomer  exports, after reaching a record 195,000 Mg in 1979, fell
to  176,700 fig in 1980.  Exports are expected to average around this level
through 1985.  Declining markets in Europe, where the EEC is pressuring
U.S. producers to curtail chemical shipments, and in Mexico and Latin
America, where AN capacity has recently been increased, will be offset by a
rapidly growing  Asian market for AN.  This market has developed rapidly for
U.S. producers in the past several years as a result of successful competi-
tion against the ailing Japanese chemical industry.  Shipments to Japan and
Korea alone increased from 27,000 Mg in 1979 to 64,000 Mg in 1980.  Exports
in  1985 are estimated to reach 175,000 Mg.
3.3  PROCESS DESCRIPTION
3.3.1  AN Monomer
     All domestic AN is produced by ammoxidation of propylene.  This process
was developed by Standard Oil of Ohio (Sohio).  The Sohio process involves
the vapor-phase  catalytic ammoxidation of propylene as represented by the
         42
equation:

     2CH2 = CH - CH3 + 2NH3  + 302   catalyst  2CH2 = CH - CN + 6H20    (1)
A representative schematic of the Sohio process is shown in Figure 3-1.
                                         43
Process streams are listed in Table 3-13.
                                   3-21

-------
                                                                                     FIGURE 3-1
                                                           FLOW DIAGRAM FOR A  SOHIO PROCESS AN  PLANT
to
 I
ro
l>lll fell
  fell lei
                                                        0
                        Cooliiii
                         talk
                   AM.O.II  \  O
                   $tou|<
                                               II2S04
                                               O
                                                                 ©I
                                               ot
N —
-»
(E
>
j

^**=S**0«> Well PonJ J

                               -J—
                                                                (o
                                                              b'epWell
                        UaiuUi mini p/oducl tlQM

                        0«iiuUi *ll ulli«i lti««n HUM

0

                                                                                                                                                             f laic
                                                                                                                                                                        SldtV Ui
                                                                                                 Incmeuloi |

                                                                                                        I
                                                                                                                      Pioducl
                                                                                                                                                             f ugihve
                                                                                                                                                          .  IniintuliJi
                                                                                                                                                         >  r«
                                                                   (2)     I'llllllll ll Mill-Ill
                                                             I         >  luiidini: diii'ily
                                                             I	     Vcul
                                                                                                                                           PloduCl
                                                                                                                                              lnt faciilly
                                                                       -*•  lli.i

-------
           TABLE 3-13.   STREAM CODES FOR FIGURE 3-1
Stream Number                      Description
      1                   Propylene feed
      2                  Ammonia feed
      3                  Process air
      4                  Hot reactor product
      5                  Cooled reactor product
      6                  Quencher bottoms
      7                  Quenched reactor product
      8                  Sulfuric acid
      9                  Stripping steam
     10                  Wastewater column volatiles
     11                  Wastewater column bottoms
     12                  Absorber vent gas
     13                  Acrylonitrile plant wastewater
     14                  Absorber bottoms
     15                  Water recycle
     16                  Crude acetonitrile
     17                  Crude acrylonitrile
     18                  Recovery column purge vent
     19                  Acetonitrile column bottoms
     20                  Acetonitrile
     21                  Hydrogen cyanide
     22                  Light ends column purge vent
     23                  Light ends column bottoms
     24                  Product acrylonitrile
     25                  Heavy ends
     26                  Product column purge vent
     27                  Flare
     28                  Fugitive emissions
     29                  Incinerator stack gas
     30                  Deep well pond emissions
     31                  Storage tank emissions
     32                  Product transport loading  facility vent

-------
     Propylene liquid feed (Stream 1)  (a mixture of propylene/propane; 90
to 97 weight percent propylene stored at 21°C and a pressure of 1,050 kPa)
is vaporized and preheated by process  steam prior to its introduction in
the reactor.  Agricultural grade liquid ammonia (99.8 weight percent minimum
and stored at 21C and a pressure of 890 kPa) is also vaporized and preheated
before entering the reactor (Stream 2).  Air (Stream 3) is filtered and
compressed for use in the reaction.  The propylene/ammonia/ai r reactants
are then fed into the reactor in the approximate molar ratio of 1:1.06:8.4.
The three percent excess ammonia forces the reaction towards completion and
the 18 percent excess air helps to continuously regenerate the catalyst.
Several catalysts have been used since the process was developed.   A bismuth-
molybdenum catalyst (Catalyst 41) introduced in early 1972 substantially
increased the yield from propylene and reduced the production of by-products
such as acetonitrile, hydrogen cyanide, and ammonium sulfate.  With the
present technology and good raw materials, the process yields 0.8  Ib AN/lb
                  44
of propylene feed.
     The reactants are fed continuously to a gas fluidized bed catalytic
reactor operating at a pressure of 135 to 310 kPa and a temperature of 400
to 510°C.  Reactor temperature is controlled by internal cooling coils
which generate about 75 percent of the plant's process steam requirements
                                                                 45
while removing the heat of reaction (approximately 21.9 kJ/g AN).     Hot
reactor gases  (Stream 4) are sent through a waste heat boiler which produces
about 25 percent of the process steam  requirements.  The product gases are
thus cooled from 510 to 230°C prior to entering the quencher.
     The quencher operates at pressures from 135 to 205 kPa and 40 to 50°C.
Streams entering the quencher are the cooled reactor effluent (Stream 5),
sulfuric acid  (Stream 8), and wastewater column volatiles (Stream 10).  The
reactor gases are cooled further and catalyst fines are removed by the
scrubbing action of the quencher.  Quencher bottoms (Stream 6) are fed to a
wastewater column where the volatile components of the stream are  steam
stripped (Stream 9).  The wastewater column bottoms (Stream  11) are sent  to
the deep well  pond.  The quenched reactor effluent  (Stream 7) is then ready
to be  fed into the absorber.  A typical plant may have two reactors and
quench neutralizes feeding a single recovery and purification train.

                                    3-24

-------
     The AN and by-products (hydrogen cyanide (HCN) and acetoni trile
are recovered by water absorption in a tray-type absorber column.  Recycle
water (Stream 15) from the acetonitrile column enters the top of the absorber,
scrubbing AN, ACN, and HCN from the quenched reactor gases (Stream 7)
entering at the bottom.  The absorber vent gas (Stream 12) is vented to the
atmosphere or sent to a control device.  Absorber bottoms (Stream 14)  are
sent to a recovery column for separation of crude AN (Stream 17) and crude
HCN (Stream 16).  The crude AN stream is sent to crude AN storage and  con-
tains about 80 percent AN, 12 percent HCN, and about eight percent impurities
(water and organics).
     The crude ACN (Stream 10) is sent to the ACN column.  The ACN is
extracted and may be incinerated or purified and sold.   The  water recycle
(Stream 15) is sent to the absorber and the ACN bottoms (Stream 19)  are
sent to the deep well pond.
     Crude AN from the crude tank storage (Stream 17) is fed  to the  light
ends column to remove low-boiling organic components.  By-product HCN
(Stream 21) from the light ends column is incinerated or sold.  Light  ends
column bottoms  (Stream 23) are sent to the product column.  Product  column
bottoms (heavy ends; Stream 25) are incinerated.  Vent streams from  the
recovery, light ends, and product columns (Streams 18, 22, and 26, respec-
tively) are sent to a flare for emission control.
     The product AN stream (Stream 24) is 99+ percent pure.   It is sent to
storage to await shipment by tank truck, barge, or railroad  car.
3.3.2  Acrylic Fibers
     The basic manufacturing process is the same for both acrylic fibers
(containing at least 85 percent by weight of AN) and modacrylic fibers
(containing less than 85 percent, but more than 35 percent AN).  This
section refers to the two fiber designations as "acrylics" except where
specifically differentiated.  Each manufacturer uses a somewhat different
patented process of monomer polymerization and fiber spinning and finishing
to form acrylic fibers.   Table 3-14 outlines the different acrylic fiber
production routes used by each domestic manufacturer.
     The physical  and chemical properties of each of the different polyacrylonitrile
fibers will  vary depending upon:

                                     3-25

-------
                                                 TABLE  3-14.   ACRYLIC  AND  MOUACRYL1C  FIBER  PRODUCTION ROUTES
Acrylic fibers
Amer
ican Cyanamid
Company
Registered
trade name
Creslan
Typical chemical
composition of
fiber
89-90% Acrylonitrile
8% Methyl Methacrylate
2-5% Other
Polymerization
medium
Suspension
Type of
polymerization
operation
Continuous
Type of
spinning
process
Wet
Spinning
solvent
Aqueous
NaSCN
lladische Company
Monsanto Company
                                 Zefran
L.I.  diiPont De Nemours & Co.     Orion
Acrllan
87-90% Acrylonitrile
8% Methyl Acrylate
2-5% Other
88-91% Acrylonitrile
7% Methyl Acrylate
2-5% Other
08-91% Acrylonitrile
8-10% Vinyl Acetate
2-4% Other
Solution


Suspension


Suspension
                                                                          Continuous
                                                                          Continuous
                                                   Batch
                                                                       Wet
                                                                       Dry
                    Wet
                                 Aqueous
                                 (IMF
                                                                                    DMAc
Moda cjry__l |c_ JM bers

t.l. diil'ont Oe Nemours 4 Co.     Orion  FLR


Eastman Chemical Products,  Inc.  Verel
Monsanto Company
Elura
                                  SEF
66-70% Acrylonitrile          Suspension
30% Vinyl Chloride
0-4% Other
37% Acrylonitrile             Solution
40% VinylIdene Chloride
20% Isopropylacrylamide
3% Methyl Acrylate
76-81% Acrylonitrile          Suspension
20% Vinyl Acetate
2-4% Other

79-81% Acrylonitrile          Suspension
8% Vinylidene Chloride
9% Vinyl Bromide
2-4% Other
                                                                          Continuous
                                                                          Batch
Batch
                                                                          Batch
                                                                                                                                Dry


                                                                                                                                Dry



                                                                                                                                Wet



                                                                                                                                Wet
                                 DMF
                                                                                                           Acetone
                                                                                    DMAc
                                 OMAc
SOURCE:  Pullman Kellogg Report, April  T9/9,  p.  8047.
a                    H        R
 In addition to Elura  and  SEF  , Monsanto  also  has  two other modacrylic  fibers,  known as  type  65  and  type  67.

-------
     1.  the inherent chemical structure of the fiber polymer;
     2.  the kind and amount of modifier used to tailor fiber properties
         for a specific application;
     3.  the fiber spinning method; and
     4.  the degree of stretching and other finishing operations.48

Although there are minor variations in acrylic fibers, they generally have
a combination of properties that make them suitable for many applications
            49
in textiles.    A block diagram of the acrylic fiber production process
showing the alternate polymerization reaction routes is presented  in  Figure  3-2.
     Most acrylic fibers are copolymers of Ml and at least one other  monomer.
These copolymers may be formed using suspension or solution polymerization
in a batch or continuous reaction mode.  The most commonly used comonomers
are acrylate and methacrylate esters and vinyl acetate for acrylic  fibers,
and vinyl and vinylidene chloride for modacrylic fibers.   These comonomers
generally have a specific function in the polymer.  A third comonomer is
often added to improve dye ability.  Other agents may be added (during
polymerization or before spinning) in small quantities (totalling  less than
5 percent).  These can be polymerization catalysts or finishing agents
performing static reducing, lubricating, heat or light stabilizing, delustrant,
or optical brightening functions.
     Acrylic fibers are produced by either suspension or solution  polymerization,
Suspension polymerization is the more common process in the Lhited  States,
while solution polymerization is more prevalent in Europe.  In either type
of process, conditions such as monomer feed ratios and temperatures must be
carefully controlled so the reaction will yield a polymer with the  desired
chain construction and length.
     In suspension polymerization, the monomers are dissolved or suspended
in the form of small  globules.  These are dispersed by vigorous agitation
throughout a water solution, which contains a monomer soluble catalyst.
The monomers are polymerized at temperatures with typical monomer  conversion
                                                  52
to polymer at about 65 to 85 percent of feedstock.    The polymer  formed is
insoluble and, thus, can be filtered from the aqueous reaction medium.
After filtering, the polymer is washed to remove impurities, refiltered,
                                    3-27

-------
                                                                 FIGURE 3-2
                                                       ACRYLIC  FIBER PROCESSES
         Solution
         Route
         Suspension
         Route
CO
I
oo
Polymerization/
Spinning Solvent
l r
AN Monomer ^
*• Polymerization t
	 	 *• Reactor
Comonorners
Vaci
Flash
JUftl
Tank
AN RECOVERY
r
t Stripper1 °°P*

Spinning
Solvent
i
AN Monomer ._
* Polymerization k

Comonomers
1
Waler,
Stabiliser
Vacuum
Flash
i
Tank


Filter
\
Fi
Finis


Tank
J
Spinning Dope ' '
SOLVENT RECOVERY
1
• • A
Spinning TOW Fiber Fiber nvein* f>tx< Selling
* Bath * Wash + Slrelching * "* * * Crimping * Dryer
J
' '

ber
hlng
. ^ Windinf 	 *- Cullinj Slaple ^ Balinf

                                                   r   i
                             ED
                                                    Tow
                                                               Yam
          Sluiry Slnppeii art AdJ on Devices thai maybe inltie Process.
Baled
Staple
AN tmlSiiim Homl lui
Solulion Houle

AN Emission Puinl lor
Suspension Route

-------
milled into snail pellets, and dried.  The dried polymer pellets must then
be redissolved in a fiber-spinning solvent.  Monomers are recovered from
the filtration and washing steps and recycled to the polymerization reactors.
The pelletizer is the second greatest source of uncontrolled AN emissions
from suspension polymerization, while by far the greatest AN emissions are
released from the polymer dryer.
     In solution polymerization, AN and other ccmonomers are dissolved in
the spinning solvent and reacted to form the polymer in solution.   The
polymerization/spinning solvent must be highly polar, but can be an organic
(dimethylformamide (DMF), dimethylacetanide (DMAC), or acetone), or an
inorganic solvent, (usually a concentrated aqueous solution of zinc chloride,
sodium thiocyanate, or nitric acid).    An initiator (e.g., azobisi sobutyoni-
trile) is usually added to begin the polymerization reaction.  Since the
monomers and polymer are soluble in the polymerization solvent which is
also used as the spinning solvent, only negligible quantities, if any, of
AN are emitted during solution polymerization.
     The prepared polymers must be made ready for fiber manufacture or
spinning.  Dried polymer pellets made by the suspension route must be
redissolved in a spinning solvent.  Polymer made by the solution reaction
is ready for spinning without the filtering, washing, pelletizing, and
drying steps.  The concentration of polymer (usually 10 to  35 percent  for
polymer of a molecular weight between 80,000 and 170,000) in the spinning
dope is limited by a workable viscosity at an economic temperature for the
particular solvent and extrusion (spinning) equipment.  The polymer must  be
basically linear as branching can adversely affect spinning.  Fblymer
solutions (or spinning dopes) are then filtered and degassed under vacuum.
     The spinning dope is extruded through spinnerets in wet or dry spinning
processes to form fiber tow.  The polymer solution (generally 10 to 25 per-
cent polymer) is extruded through spinnerets having typically 1,000 to
3,000 (some having as many as 60,000) holes, each of 50 to  over 100 micro-
meters (jim) in diameter.  Fiber is formed at a rate of 2 to 200 meters per
minute.   The wet process tow is extruded into a bath of spinning solvent
diluted with water so that the fibers coagulate at the proper rate.  Water
is added countercurrently to the tow to wash out the solvent.  The main
advantaae of this wet spinning is that tow can be further processed on a
    .       .   .   53,54,55
continuous  basis.
                                   3-29

-------
     Dry spinning extrudes rare concentrated (20 to 25 percent poiyner)
preheated solutions through spinnerets having typically 20C to 600 (some
having as many as 900) holes, each wi th a diameter of 50 to 150 ^m.  Fiber
is extruded at a rate of 200 to 400 m/ninute.  The tow is extruded into a
column of hot circulating air which is 50 to 110°C higher than the boiling
point of the spinning solvent.  As the solvent evaporates, the fibers
solidify.  The fibers are collected as tow while the vaporized solvent is
                       r -t c C
recovered and recycled.  '
     Spinning and washing are the only significant sources of emissions
from solution polymerization processes if the unreacted AN monomer has not
been removed previously.  Spinning is a negligible source of emissions from
suspension polymerization processes as the unreacted AN has already been
removed during pelletizing and drying.
     After wet or dry spinning, the fiber tow is passed through a  heated
bath in order to (1) remove the about 10 percent residual  solvent  for
recovery required for an economic operation and (2) heat the fiber so it
can be drawn.
     Drawing involves stretching the fiber tow 4 to 10 times its original
length, in order to orient the polyacrylonitrile chains along the  fiber
axis and thus improve fiber properties.  Stretching is normally accomplished
using air, hot water, or steam at 70 to 110°C and rollers rotating at
increasing speeds.  After stretching, the fibers are crimped to provide a
texture similar to natural fibers and dried, which also stabilizes or sets
the crystalline structure (dimensions) and other physical  properties  and to
improve dyeability.  Fibers may undergo finishing (i.e., are coated with
lubricating and antistatic agents) to facilitate textile processing operations
Finally, fibers may be cut into staple (6 to 15 on in length) and  baled,
left uncut and sold as tow, or wound on bobbins and sold as continuous
filament yarn.  '    Emissions of AN from finishing operations are negligible.
3.3.3  ASS/SAN
     Three types of polymerization processes are used for the commercial
production of ABS plastics.  The majority of ABS resins are produced  pri-
marily by emulsion polymerization although suspension and bulk polynerization
are also used.
                                      3-30

-------
     3.3.3.1  Emulsion Process.  The emulsion process consists of three
distinct polymerizations.  A polybutadiene substrate latex is prepared,
styrene and acrylonitrile are grafted onto the polybutadiene substrate, and
styreneacrylonitrile copolymer is formed.  The latter two reactions may
take place simultaneously in the same vessel   (used for graft ABS polymers
with low polybutadiene content) or in separate vessels and then latex
blended.  A flow sheet for the emulsion process is shown in Figure 3-3.
     Butadiene is first charged to a water jacketed batch pressure reactor
(1) containing emulsifiers.  Additional emul sifiers, catalysts, and demine-
ralized water are then added.  Reaction temperature may be varied from 5 to
70°C depending on the desired structure of the polymer.  It is controlled
by circulating cooling water in the jacket.  The rate of reaction is limited
by the ability of the cooling jacket to remove the heat of polymerization.
The time required for this synthesis step ranges from 12 to 24 hours.
     Reaction conditions and chemical recipes are important to end product
properties.  Latex particle size is determined by the emulsifier and quan-
tities of the various monomers used in the reaction.  Excess crosslinking
and ring-formation may be prevented by the addition of a shortstop such as
sodium di thiocarbonate after 70 to 90 percent monomer conversion (monomer
conversion in emulsion reactions can be greater than 96 percent).  The
unreacted butadiene and monomers are stripped from the latex in a flash
stripper (2).  The reactor, stripper, and recovery system vents are directed
to a butadiene flare system.  The polybutadiene latex is stored for use in
graft reactions.
     The next synthesis step (3) is the grafting of styrene and acrylonitrile
into the polybutadiene substrate which serves as the "backbone" latex.  The
amount of substrate used in the graft reaction is determined by the physical
properties desired in the final polymer and usually ranges from 10 to
60 percent by weight of the polymer.  The graft reaction may be either
batch or continuous.  In a batch reaction, emulsifiers, such as sodium
oleate, initiators (cumene hydroperoxide), and activators (sodium pyrophso-
phate, dextrose, and ferrous sulfate) are prepared separately and charged
to reactors which have been purged of oxygen.  A free-radical initiator,
such as potassium persulfate, may be used to  start the reaction.  A chain-
transfer agent, such as terpinolene, is also  needed.  Appropriate quantities
                                   3-31

-------
                                               AN to Atmospnere
                                                     t
      Sutafliene,
      £riiulsifi«rs,
      Initiator],
      Water
                                                     0
                                                 To Incinerator
                                                                            AN Storage
                                                                            Emissions
                    Butadiene
                  to Comoressor
                                                                Acrylomtnle, —
                                                                Styrene,
                                                                Emulsifiers.
                                                                Cataiysts, Water
   Batch Pressure
      Reactor
                                                                             AN to AtmosDhere

                                                                          J
                  Flash Stripper

                   Styrene
AN  4—
Storage
Emissions
• Acrylonitnle
   Emulsifiers
 and Initiators
                                                                                   AN to Atmospnere
ALTERNATIVE
     SAN PATH
                                         FIGURE 3-3
                               EMULSION  ABS/SAN PROCESS
                                                                                                   Packaging
                                                                  3-32

-------
of styrene and AN monomers are added and mixed to form an emulsion while
the temperature is raised and maintained at the reaction temperature (usually
between 50 and 75°C).  In continuous processes, monomers, water, emulsifiers,
and soap are metered and pumped in continuously and overflowed continuously
from the reactors.  Reaction conversion of monomers is around 95 percent
for the batch polymerization process and slightly lower for continuous
processes.  Vapors from the reactors are pumped to an AN absorber (4).
Absorber bottoms are recycled and the tops are sent to an incinerator (5).
     The copolymerization of AN and styrene is achieved by feeding the
monomers to a reactor vessel (6) along with deionized water, emulsifiers,
and catalysts. Monomer conversion efficiencies to SAN range from 90 to
98 percent.  Polymer solids are recovered from latex by coagulation and
flocculation.  The emulsifiers are chemically destroyed by adding dilute
salt or acid solution and the polymer solids are allowed to agglomerate
(7).   Calcium chloride, sodium chloride, sulfuric acid, and hydrochloric
acid are commonly used coagulation agents. The agglomerated polymer is
dewatered by screening (8), centrifuging (9), and vaccum filtration (10).
Solids are mechanically blended (11) with dyes, antioxidants, and other
additives.  The polymer sheets from these operations are then pelletized
and packaged (12).
     ABS plastic  is a blend of graft rubber and SAN resin.  The graft latex
may be mixed with SAN resin either in the latex stage before coagulation or
mechanical ly mixed at the finishing stage.
     The potential AN emission sources from this process are storage tanks,
graft  reactors, absorbers, latex treatment vessels, and polymer dryers.
These  emission pints are illustrated in Figure 3-3.
     3.3.3.2  Suspension Process.  Although most ABS is produced by the
emulsion process, suspension polymerization is also used to produce specia-
lized  plastics.   The suspension process, shown in Figure 3-4, begins wi th a
polybutadiene rubber which is so lightly crosslinked, it is soluble in the
monomers.  The polybutadiene is typically synthesized by the batch emulsion
process previously described.
     Polybutadiene is dissolved in styrene and AN monomers (1) to produce a
solution free of crosslinked rubber gels.  A free-radical initiator such as
an organic peroxide or an azo compound is added to the solution along with
                                     3-33

-------
                               FIGURE 3-4
                       SUSPENSION ABS PROCESS
                                                     AN Emissions
                                         Storage Tanks •
Poiybutadiene,
Rubber
          Initiators
          and Agitation
                Suspending Agent,
                Water
                                          Hot Air
                                                                 Storage
                                           3-34

-------
chain-transfer agents (2).  It is then heated to 30 to. 120°C for a period
of  two to eight hours with sufficient shearing agitation to prevent cross-
linking and maintain the desired polymer particle size.   After 25 to 35
percent monomer conversion, the polymer syrup is transferred by a high
viscosity gear pump to a suspension reactor (3) where it is dispersed in
water with agitation.  Suspending agents are added to the reactor and it is
heated to 100 to 170°C for about six to eight hours.
     After achieving the desired monomer conversion, the  products are
cooled and pumped to a dewatering system (4), usually a  continuous centrifuge.
The polymer beads are washed and dewatered to less than  10 percent moisture.
Wet product beads are then conveyed to a conventional hot air dryer (5)
where they are dried to less than 1 percent moisture.  Suspension ABS
polymer beads range in size fron 0.4 to 1.2 mm in diameter.  The dry beads
are stored in silos prior to compounding.
     Fbssible AN emission sources from the suspension process are the
prepolymerizer, the suspension reactor, the dewatering system,  and the
dryer.  Emission points are illustrated in the flow diagram.
     3.3.3.3  Continuous Mass  (Bulk) Process.  The continuous mass (bulk)
process has recently begun to  achieve more commercial importance.  It was
previously used only by Dow Chemical Company, but several  other manufacturers
(Borg-Warner and Monsanto) have entered into preparation  of ABS plastics by
this procedure.  The reaction  does not proceed in water,  so treatment of
wastewater is minima! and less energy is used per unit weight of product
since no dewatering or polymer drying is needed.  It is  a more  mechanically
complex process with less product flexibility and lower  monomer conversion
(requiring devolatilization of the polymer), but these problems are
outweighed by the process advantages.  A block flow diagram for the process
is given in Figure 3-5.
     A monomer-sol uble plybutadiene substrate is dissolved in styrene and
AN with initiators and modifiers (1).  The mixture is polymerized through
phase inversion to approximately 30 percent monomer conversion (2).  This
prepolymerized syrup is then pumped into a specially designed bulk polymerizer
(3) where monomer conversion is taken to between 50 and  80 percent.  (No
water is used in this procedure.)  The bulk polymerizer  is operated continuously
at 120 to 180°C with polymer residence times of one to five hours.  Monomer
                                  3-35

-------
                                             FIGURE  3-5

                                         BULK ABS PROCESS
Poiybutadiene
   Rubber
                                          AN
                                           A
                            AN Storage
                       AN	

                       Styrene

                       Initiators and Modifiers
                            Bulk Polymerizer
                                                       Devolatilizer
Water Bath
                                                                                           Pelletizer
                                                   3-36

-------
vapors evaporated by the heat of reaction are condensed and recycled with
fresh monomer feed (4).  The polymer melt is pumped to a devolatil izer (5)
where remaining unreacted monomers are removed under vacuum at temperatures
in excess of 150°C.   Normally, about five to 30 percent of the feed  stream
is removed as unreacted monomer and recycled.  The ABS polymer is removed
from the devolatil izer via a melt pump or extruder and chopped into  pellets
(6).  The AN emission points include the bulk polymerizer, the monomer
vapor condenser, and the devolatil izer as shown in Figure 3-5.
     3.3.3.4  SAN by Emulsion Process.  The emulsion copolymerization of
styrene and AN to SAN resin is achieved commercially by both  batch and
continuous processes.  A typical reaction system is shown in  Figure  3-3.
Portions of the monomers (68.5 parts styrene; 31.5 parts AN),  an  initiator-
emulsifier solution  (potassium persulfate, rosin soap, and desalted  water),
and a chain-transfer agent (tert-dodecyl mercaptan) are charged  to the
reaction and then purged with nitrogen.  Polymerization temperature  is
controlled at about 80°C by reflux cooling of the  reaction mixture.   The
remaining portions of reactants are added to the reaction system  continuously.
The reaction is allowed to proceed for about one hour after final  addition
of reactants (total  cycle time is one to three hours) with monomer conversions
reaching 97 percent or higher.  The SAN latex slurry may be directly blended
with a SAN grafted-rubber latex for ABS production or coagulated  to  recover
SAN copolymer.   Coagulation of the latex is accomplished by adding electrolyte
(dilute salt or acid solution) or freezing.  Polymer is then  washed  and
dried.
3.3.4  NBR Elastomers
     Nitrile elastomers are copolyners of AN and butadiene.  They are
produced by emulsion polymerization in batch or continuous processes.  The
process block flow diagram is shown in Figure 3-6.
     The emulsion polymerization process uses water as a carrier  medium.
This offers the advantages of safety, inexpensive  product and monomer
recovery, and ease of product purification compared with organic  solvents.
Nitrile rubbers are generally manufactured as either hot or cold  types
depending on temperatures at which the polymerizations are controlled.
These temperatures are about 40°C for hot types and 5°C for cold  types.
Cold varieties  have  lower content of branched and  cross! inked polymer
chains and can be better processed.
                                     3-37

-------
*.H
     AN Storage
                                                                                         Butadiene
                                                                                       . to Storage
                                                                                         or Recycle
AN to
Conaensaticn
ana Storage
                                 FIGURE 3-5
                             NITRILE ELASTOMER
                                   PROCESS
                                                                          Sawning, Washing,
                                                                             Dewatering
                                                      3-38

-------
     Butadiene and AN monomers are piped to agitated polymerization reactors
(1) along with additives and soap.  The water not only serves as a reaction
inediun, but also effectively transfers the heat of reaction to the cooled
reactor surfaces.  The additives include catalysts (cunene hydroperoxide as
an oxidizing component), and sodium formaldehyde sulfoxylate with EDTA
(ferrous sulfate complexed with ethylenedianine-tetraacetic acid) as the
reducing component) and modifiers (alkyl mercaptans).
     The rosin soap solution forms small bubbles called micelles.  These
micelles act as  tiny reaction vessels for the polymerization processes.
Inside the micelles free-radicals formed in the water phase by the initiators
(organic peroxide or potassium persulfate) create polymer nuclei from the
monomers trapped in the bubble.  Droplets of monomer outside the micelles
supply the reaction sites with raw materials.  The small polymer particles
formed in the micelles continue polymerization with the supplied monomers
and grow into the desired polymer chain.  The soap bubbles cover the rubber
particles formed and keep the mixture in liquid form (prevent agglomeration
and plugging of  the reactor).
     The reaction is allowed to proceed for 5 to 12 hours.  A shortstop
solution (sodium bi sulfate or potassium dimethyl dithiocarbonate) is added
to terminate the reaction at a predetermined point, usually after 75 to
90 percent conversion (depending upon the desired moleclar weight of the
product).  The reaction latex is then sent to a blowdown tank (2) where
antioxidants are normally added.
     The latex is subjected to several vacuum flash steps (3) where most of
the unreacted butadiene is released.  It is then steam stripped under
vacuum (4) to remove the remaining butadiene and most unreacted AN.  The
unreacted monomers are sent to recovery and recycle.  Stripped latex at
about 110 to 130°F is pumped to blend tanks (5).
     Gases released in the flash steps and stripper overhead contain butadiene.
These are sent to a partial  condenser and separator (6) where butadiene
vapor is condensed and sent to liquid storage.  Lhcondensed butadiene vapor
from the separator flows to an absorber (9) where it is absorbed by counter-
current contact with chilled oil.  The absorber bottoms are pumped to a
flash tank and dissolved butadiene is released and returned to the compressor.
The hot lean oil  is then cooled, chilled, and returned to the top of the
absorber.
                                      3-39

-------
     Ihreacted AN in flash vapors and latex stripper overhead Is recovered
by sending these gases to a water absorber (10).  Absorber bottoms and the
liquid phase of the latex stripper overhead are pumped to a steam stripper
(11).  The overhead vapor stream from this stripper is condensed in a
decanter.  Phase separation is allowed to take place and the AN phase is
decanted to storage while the water-rich phase with residual AN is returned
to the stripper.
     Latex is pumped from the blend tanks (5) to a coagulation tank (6)
where the emulsion is broken by the addition of dilute inorganic salt
solution (sodium chloride or aluminum sulfate) or weak organic acid.  The
slurry of fine polymer crumb is then filtered to remove coagulating chemicals
(liquor  is recycled) and may be reslurried for further purification.  Crumb
is dewatered in an extruder (12), then hot air dried (13).   Dried rubber is
weighed, pressed into bales, and prepared for shipment.
     If  latex is the desired product, the final processing  steps (coagulation,
screening, washing, and drying) are omitted.  The initial steps are essentially
identical with those of solid rubber production.
     Emissions of AN may occur from the storage tanks, reactors, and absorbing
and stripping columns, as shown in Figure 3-6.
                                      3-40

-------
REFERENCES TO CHAPTER 3

1.    U.S. International Trade Commission.   Synthetic  Organic  Chemicals,
     U.S. Productions and Sales.  1979.

2.    Stanford Research Institute.  1980  Chemical  Economics  Handbook.   Menlo
     Park, California.  1980.  p. 607.5032J.

3.    1980 Kline Guide to the Chemical  Industry,   p. 172.

4.    Textile Economics Bureau, Inc.  Textile Organon.   New  York.   February 1981.

5.    The Coming Squeeze on Petrochemical  Exports.   Chemical Week.
     October 1, 1980.  pp. 26-31.

6.    A Shaky Recovery for Chemical Producers.   Chemical  and Engineering  News.
     December 22, 1980.  pp. 30-35.

7.    Chemical and Engineering News.  December  1,  1980.   p.  12.

8.    Reference 2.  p. 607.5032J.

9.    Reference 1.

10.  ABS Market:  A Mix of Optimism  and  Problems.   Chemical and  Engineering News
     July 30, 1979.  pp. 8-9.

11.  Pullman-Kellogg, Inc.   Emission,  Process,  and  Control  Technology  Study
     of the ABS/SAN, Acrylic Fiber,  and  NBR Industries.   EPA  Contract  No.
     68-02-2619, Task 6.  April  1979.

12.  Reference 2.  p. 607.5032W.

13.  Reference 12.

14.  Reference 2.  p. 607.5032X.

15.  Reference 2.  p. 607.5032Y.

16.  Reference 2.  p. 607.5032Z.

17.  AN Maneuvers for a Comeback.  Modern Packaging.   April 1979.

18.  Acrylonitrile's Still a Worry.  Modern Packaging.   June  1978.

19.  High-Clarity Acrylonitrile  Scores for Small  Containers.   Modern  Plastics.
     December 1978.

20.  Chemical Marketing Reporter.  February 4,  1980.

21.  Reference 2.  p. 607.5031D.
                                        3-41

-------
22.  Reference 11.

23.  Reference 2.  p.  543.3521F.

24.  Reference 7.

25.  EEA contacts with  industry.

26.  Reference 23.

27.  Reference 4.

28.  EEC Moves Against  U.S. Acrylic  Fibers.   Chemical  and  Engineering  News.
     May 12, 1980.  p.  8.

29.  The High Cost of Oil Decontrol  is  Hitting  Home.   Chemical Week.
     February 11, 1981.

30.  EEA estimate.

31.  Reference 10.

32.  Letter from Harvey,  L.A, President, Borg-Warner Chemicals, to  Editor,
     Chemical and Engineering News.  October  22,  1979.   pp. 4-5.

33.  1979 Annual Report.  Borg-Warner Corporation.

34.  Reference 11.

35.  Reference 21.

36.  Lean Years Ahead  for the Rubber Industry?   Rubber World.  January  1980.
     pp. 27-28.

37-  C.H. Kline, Inc.   Elastomers, A Forecast to  1983.   1979.

38.  Reference 36.

39.  Stanford Research  Institute.  1980 Director  of Chemical  Producers  of  the
     United States, Supplement.  Menlo  Park,  CA  1980.

40.  Reference 12.

41.  Monsanto Company v. David  Kennedy  and Joseph  A. Califano, Jr.   No.  77-2023.
     U.S. Appellate Court D.C.  Circuit.  November  6, 1979.

42.  Hobb,  F.D.  and Key, J.A.   Hydroscience,  Inc.  Emission Control  Options  for
     the Synthetic Organic Chemical  Manufacturing  Industry:   Acrylonitrile
     Product Report.   Report from EPA Contract  No. 63-02-2577.
     August 1978.  p. III-l.
                                      3-42

-------
43.   Hughes, T.W. and Hour, D.A.   Source  Assessment:   Acrylonitrile
     Manufacture (Air Emissions).   EPA-600/2-77-107.   September 1977.
     pp. 10-20.

44.   Reference 43.   p. 13.

45.   Reference 43.   p. 17.

46.   Reference 43.   p. 19.

47.   Click, C.N. and D.O. Moore.   Emission,  Process,  and  Control  Technology
     Study of the ABS/SAN, Acrylic  Fiber,  and  NBR  Industries.   Pullman-
     Kellogg.   Final Report for  EPA  Contract No. 68-02-2619,  Task 6.
     Houston,  Texas.  April 1979.   p. 80.

48.   Considine, D.M.  Chemical and  Process Technology Encyclopedia.
     McGraw-Hill Book Company.   New  York.  1974.   pp.  27-28.

49.   Encyclopedia of Polymer  Science and  Technology.   Volume  1:   Acrylic
     Fibers,  p. 352.

50.   Reference 3.  pp. 343-344.

51.   Work, R.W.  Man-Made Textile  Fibers.  In:   Riegel's  Handbook of
     Industrial Qhemistry.  Kent,  J.  (ed).   Van  Nostrand  Reinhold.  New York.
     1974.  p. 332.

52.   Development Document for Proposed  Effluent  Limitations Guideline  and
     New Source Performance Standards for the  Synthetic Resins  Segment of
     the Plastics and Synthetic  Materials Manufacturing Point Source
     Category.  U.S. Environmental  Protection  Agency.  EPA #940/1-73-010.
     September 1973.  p. 71.

53.   Reference 11.  pp. 80-85.

54.   Reference 3.  pp. 348-351.

55.   Reference 1.  p. 83.

56.   Reference 3.  p. 3.
                                     3-43

-------
               4.  AIR EMISSIONS DEVELOPED  IN SOURCE CATEGORY

4.1  PLANT AND PROCESS EMISSIONS
4.1.1  AN Monomer
     Major sources of AN emissions from AN  monomer manufacturing operations
are the absorber vent, column vents, product storage and handling vents,
and fugitive emissions.  The AN emissions estimates in this report for the
various sources at a model 180,000 Mg/yr monomer manufacturing plant are
based on emission factors given by Hobbs and Key.   Table 4-1 lists the
                                                                2 3
emissions factors as provided by the CMA and SRI for comparison. '   These
factors were not used for emissions estimates in this report because the
data were not identified as controlled or uncontrolled emissions and the
control methods that may have been used were not discussed.
     The absorber vent gas consists of the  gross reactor effluent after
stream neutralization and recovery of nitriles.   The vent gas is composed
of approximately 97.25 weight percent inerts, 1.5 percent carbon monoxide,
                                             4
1.25 percent total VOC, and 0.001 percent AN.   The AN emission factor for
this vent is estimated to be 0.01 kg AN/kg  AN produced.   Model plant emis-
sions of AN from this source would be around 18 Mg/yr.   Vent gas composition
is dependent upon the catalyst used, reactor operating conditions, raw
material purity, production rate, and absorber operating conditions.
Emissions of AN increase markedly when the  absorber column overhead tempera-
ture increases.   Vent emissions of AN and other hydrocarbons are also
influenced by the air feed rate to the reactor under normal operating
conditions.   Absorber vent emissions increase during' reactor startups, when
the stream composition varies continuously  until normal  operating conditions
are reached.   The emission rate of AN during reactor startup is approximately
3.84 kg AN/hr as opposed to 2.05 kg AN/hr during normal  operations.  During
startuo conditions,  the absorber vent stream also has a higher concentration
                                       4-1

-------
Source of estimate
Plant
Absorber (oxidizer)
Controlled
Column vents (flare)
Controlled
llandl ing/Loading
Fugitive
Storage
Wastewater pond
TOTALS

Emission factor (kg/Mg
Chemical Manufacturers Association3
DuPont Sohio Monsanto
Beaumont Memphis Lima Alvin Texas City
0.02
0.07
0.07
0.02
0.01
0.09
0.26
1.20 0.52 0.02 0.12
0.07 0.01 0.01 0.02
<0.04 0.02 0.01 0.01
0.02 0.02 0.04 0.04
<0.10 0.40 0.63 0.30
<0.04 -
1.49 0.97 0.71 0.49
of monomer)

American Cyanami
West Wego
0.03
0.01
0.11
0.01
0.72
0.008
0.89
SRlb
d Model
plant
0.046
0.50
0.14
0.26
0.81
0. 10
1.85
Hydroscience
Model plants
0.002
0.05
0.11-0.17
0. 16
Total all storage
-
M.32
 Source:  (Chemical Manufacturers Association, 1980).  Emission controls for each plant are not specified, some sources
 are uncontrolled.  Fugitive emissions low owing to OSHA regulations.
 Source:  (Suta, 1979).  Emission factors were supplied to Stanford Research Institute (SRI) by EPA for use in the
 population exposure analysis.
cSource:  llobbs and Key, 1978.  Hydroscience, Inc., developed emission factors for the various processes using feasible
 control options by a model plant analysis.  Only absorber vent and column vent factors assume controls -- all other-
 factors given are estimates of uncontrolled emissions.
 Memphis and Lima plants are not controlled, American Cyanamid now installing controls.
Emission factor for DuPont, Memphis = 1.20 kg/Mg;  American Cyanamid,  Westwego = 3.0 kg/Mg.

-------
of VOC and some catalyst fines.  This is a result of bypassing the quencher
and absorber (or absorber only) during the startup process.  These conditions
last for about 1 hour, a reactor normal ly being shut down and restarted
about four times per year.   Uncontrol led startup AN emissions from a two-
reactor production line vculd be approximately 31  kg AN per year.
     Vent gases from the recovery, acetonitrile, acetonitrile purification,
light ends, hydrogen cyanide, and product columns  are composed of non-
condensables dissolved in the column feeds, uncondensed VOC, and, for
vacuum columns, air that has leaked  into the columns and is removed by the
vacuum jet systems.  There  is little reliable data on the composition of
emission streams or the quantities of the various  stream constituents;
therefore, a combined emission factor for all column vents was estimated at
5 kg AN/Mg AN produced.   The estimated uncontrolled model  plant AN emissions
from the combined column vents would then be approximately 900 Mg AN/yr.
     Product storage tanks  and handling facilities are another source of AN
emissions at AN monomer plants.  The combined uncontrolled AN emissions
factor for the vents from crude AN,  AN run, and AN product tanks is about
1 Kg/Mg AN produced.  This  translates into approximately 180 Mg/yr of AN
emitted by the uncontrolled model plant.  The uncontrolled AN emission
factor for AN handling operations is approximately 0.272 Kg/Mg AN produced.
An uncontrolled AN handling facility at a model plant can thus be expected
to emit about 49 Mg AN/yr.
     Fugitive emissions occur from leaks in the pumps, compressors, and
valves that handle process  streams containing AN.   The model plant is
assumed to have 25 pumps and 600 valves handling AN out of an estimated 50
and 1200 total, respectively.  The estimated fugitive AN emissions factor
is 0.159 Kg AN/Mg AN produced.   Lhcontrol led fugitive AN emissions would
thus be 28.6 Mg/yr for the  model plant.
     Fbssible secondary AN  emissions sources are the flare and incinerator
stacks and the deep well pond.  Emissions of AN from the flare and incinerator
stacks are minimal, these streams being composed mostly of nitrogen, carbon
dioxide,  and water.  Plant  wastewater sent to the  deep well pond has less
than 0.02 percent by weight of AN.   This fact, in conjunction with a VOC
emissions control  efficiency of 92 percent afforded by a heavy lube oil
layer covering the deep well pond, is the basis for assuming that AN emis-
sions are very small .

                                   4-3

-------
     Table 4-2 summarizes the emission factors, characteristics, and emissions
from the possible sources of AN emissions at the model plant.  Emission
factors for total VOC are included for comparison.
     Estimates of total emissions  of AN from monomer plants were prepared
                                           7 8
by SRI and CMA and are listed in Table 4-3.  '   Emission estimates from SRI
                                                                       q I Q
were developed using process-specific emission factors supplied by EPA.
These factors were developed by EPA in previous studies conducted by Hydro-
science and Pullman-Kellogg.  The CMA study team for AN reviewed the SRI
report and updated AN emission estimates.  The CMA figures reflect the use
of plant-specific data as opposed to the SRI model plant approach.   Recent
improvements in process and control methods are incorporated in the figures.
The CMA figures have not yet been thoroughly evaluated.
4.1.2  Acrylic Fibers
     The only nonnegligible uncontrolled AN emissions from acrylic fiber
production for suspension polymerization followed by wet spinning occur at
the pelletizer, polymer dryer, and crimper;  for solution polymerization
followed by wet spinning, emissions occur at the spinning and washing
steps.  The emission factors presented in Table 4-4 were developed by Click
and Moore of Pullman Kellogg   for model plants producing 45,400 Mg
(100 million pounds) of acrylic fiber per year.  Both uncontrolled model
plants assume an acrylonitrile (AN) polymerization reactor conversion of 87
percent followed by vacuum flashing to release 80 percent of the unreacted
AN for monomer recovery by absorption/stripping.  The suspension polymerization
model plant is assumed to have two filter/wash stages each removing 37.5
percent of the AN in the incoming polymer slurry.   The solution polymeriza-
tion model plant is assumed to remove 70 percent of the remaining unreacted
AN for recovey during the washing step following spinning.   State implemen-
tation plans (SIPs) do not address acrylonitrile emissions specifically,
but do address emissions of all volatile organic compounds (total VOC).
Therefore, the uncontrolled model plant, which includes typical AN recovery
for economic reasons, is also a good representation of a plant satisfying
SIPs.  As the price of AN monomer feedstock continues to increase, additional
recovery/emission control measures become more attractive.   Also, improved
enclosure and capture of emissions is helpful  in limiting OSHA worker
exposure to AN.

                                   4-4

-------
                     lAfiLE  1-2.   UNCONTROLLED AND CONTROLLED EMISSIONS AND EMISSION FACTORS FOR AN AND VOC FROM A MODEL  PLAN!
                                                                USING THE SOIIIO PROCESS
 i
en
Einissi
Kg emission;
Unconlrol led
Source
Absorber vent
(Normal) .
(Startup)
Column vents (total)
Storage vents
Crude acryloni tri 1e
Aery loni tri le run tanks
Acryloni tri le storage
Hand) ing
Acryloni tri le .
Acryloni tri lef
fugi tlve
Secondary
Incinerator
Deep wel 1 pond
TOTALS
AN
0.10
0.187
5.00
0.101
0.149
0.751
0.167
0.105
0.159
_
6.719
Total VOC
100.0
0.244
10.00
0.101
0.149
0.751
0. 167
0.105
0.326
0.360
13.00
125.203
Ion factor AN emissions
i/Mg AN produced Mg/yr
Controlled Uncontrolled
AN
0.0010
0.0019
0.0500
0.0010
0.0015
0.0075
0.0017
0.0011
0.0154
_
0.081
Total VOC
1.00 18.0
0.0025 0.031
0.100 900.0
0.0010
0.0015
0.0075 Total storage 180.2
0.0017
0.0011 Total handling 48.96
0.0310 28.6
0.360
1.100
2.606 1179.791
Control led

0.18
0.00031
9.0
1.802
0.504
2.86
-
14.346
              Model  plant operating 8760 hrs/yr at 180,000 Mg capacity.
             'Vented to atmosphere to prevent possible explosion in absorber lines.  Yearly averaged value; 4 startups each of 2  reactors.   Each  startup
              lasts  1  hour.
              By  lank  car.
              lly  barge.
             efsl initiled total emissions assuming model plant has 25 out of 50 pumps and 600 out of 1200 valves handling acryloni tri le.
             Source:   llobhs  and Key, 1978.   llydrosc ience, Inc.

-------
                                      TABLE 4-3.   SUMMARY OF ESTIMATED TOTAL AN EMISSIONS FROM MONOMER PRODUCTION
IMaul and Cily
Am. Cyanamid
Woslwego
dn Pout de Nemours
Beaumont
MonsanLo
Texas Cily
du Pool de Nemours
Memph i s
Sojiio (Vislron)
1 ima
Sohio (Vislron)
Victoria
Monsanto
Alvin
Non-
Attain- attain-
ment ment Extension
County and State area area requested
Jefferson X
LA
llardin X -
TX
Galveston X
TX
Shelby - X
TN
Allen X
Oil"
Victoria - X
TX
Brazoria X
TX
Total AN
emissions
AN emissions (Mg/yr) (Mg/yr)
Process Storage Fugitive EEA SRI CMA
74.2 109.8 3.6 187. 6 578 107
23.4 4.1 56.2 83.7 294 43
49.3 59.4 167.2 275.9 353 94
90.2 328.9 N/A 419.1 367 187
335 17C
-
370 140
Comments
Data obtained from Section 1)4
letters and follow-up phone
conversations.
Fugitive AN emission figure in-
cludes estimated process equip-
ment emissions and settling pond,
ship and barge operations, flare
vent, and storage tank vent.
Process and storage emission fig-
ures calculated from Monsanto
emission factors in 114 response.
Fugitive figure includes data
for equipment leaks.
Fugitive emissions not available.
Did not receive questionnaire.
New plant - not In operation.
Did not receive questionnaire.
TOTAL 2279 717
 I iienjy and Environmental Analysis, Inc. (EEA)




SOURCE:   Plant visit and Section 114 questionnaires.

-------
      TABLE 4-4.   ESTIMATED UNCONTROLLED EMISSIONS FROM MODEL PLANTS
                  PRODUCING 45,500 Mg/Yr OF ACRYLIC FIBER

*
Source
Suspension Polymerization:
Pel letizer
Polymer Dryer
Solvent Scubber Vent
Crimping
Setting Dryer
Total
Solution Polymerization:
Monomer Recovery
Scrubber Vent
Spinning, Washing
Vent
Total
Emission
(kg emi
Mg fiber
AN

1.15
10.5
0.0
0.05
Q.Q
11.7

0.0
9.0
9.0
Factors
ssions/
produced)
Total
VOC

1.15
10.5
1.0
4.05
5.0
21.7

0.1
9.9
10.0
Emi
AN

52
480
0
2.
0
530

0
410
410
ssion Rate
(Mg/yr)
Total
VOC

52
480
46
3 180
230
990

4.6
450
455
Source:   Pullman Kellogg Report, April 1979, p. 103-104.
                                       4-7

-------
     The AN and VOC emissions from an uncontrolled suspension polymeriza-
tion model plant's filtration reslurry tank and AN recovery absorber vent
were estimated as negligible.  AN emissions were also estimated as negligible
from the solvent scrubber vent and setting dryer.  AN emission factors and
emissions were estimated by  Pullman Kellogg   to be 1.15 kg AN emitted per
Mg fiber produced or 52 Mg/yr from the polymer pelletizer, 10.5 kg/Mg or
480 Mg/yr from the polymer dryer, and 0.05 kg/Mg or 2.3 Mg/yr from crimping.
The annual emission rates of AN are for the model plant with a capacity of
45,500 Mg/yr.  AN process emissions estimates total 11.7 kg/Mg or 530 Mg/yr
for this model plant.
     The AN and VOC emissions from the uncontrolled solution polymerization
model plant were estimated as negligible from the solvent purification vent
and from all of the finishing operations.  Emissions of VOC other than AN
were expected from the AN monomer recovery scrubber.   The only process
source with major estimated  AN emissions is the spinning and washing vent
with 9 kg/Mg or 410 Mg/yr.   Although the amount of AN in this stream is
quite high, the concentration is low because of the large amounts of air
(as much as hundreds of cubic feet per minute) used to exhaust the spinning
and washing area.
     Monomer storage tanks are also a source of AN emissions.   Emissions
from uncontrolled (fixed roof) tanks range from about 2 to 30 Mg/yr per tank,
depending on the tank size,  usage patterns, and weather conditions.   Some
plants use control techniques such as floating roofs or venting to nearby
scrubbers to recover monomer and/or reduce emissions from selected tanks to
2 Mg/yr or less.   Equations  are available for calculating emissions from
                                                  12
storage tanks containing volatile organic liquids.
     No industry-specific data on fugitive emissions or the number of their
common sources (pumps, valves, etc.) are available for acrylic fiber plants.
The total fugitive AN emissions from a fiber operation have, therefore, not
been estimated at this time.
     Several estimates have  been made of total national emissions from
acrylic fiber plants.  Pullman Kellogg estimated total AN emissions of
4748 Mg in 1977 by prorating, on the basis of capacity, the emissions
reported by industry for some plants up to estimates for the entire industry
segment.     SRI generated a  national estimate of 4698 Mg/yr using process-
                                     4-8

-------
specific emission factors supplied by  EPA.14'15'15  The  CMA  reviewed the
SRI report and formed an AN study team  to  update  industry AN  emission
estimates.   The national total AN emissions estimated by CMA  are only
1361 Mg/yr, reflecting recent process  and  emission control changes.    The
CMA data have not yet been thoroughly  evaluated by EPA.  The  CMA and SRI
plant-specific estimates are presented  in  Table 4-5.
4.1.3  ABS/SAN
     Plants manufacturing ABS/SAN resins emit AN  to the atmosphere from
process equipment vents, storage tank  vents, and  fugitive emissions from
process streams.  Process equipment AN  emission sources include reactors,
absorbers, and dryers as discussed in  the  ABS process description.   Fugitive
emissions result from pumps, valves, and flanges.  Control methods are
discussed in Section 5.2.3 of this report.
     AN emissions estimates (controlled and uncontrolled emissions) were
made using data obtained from the individual manufacturers through Section 114
requests and plant visits.  Visits were made to the Borg-Warner Chemicals,
Washington, West Virginia, and Monsanto Plastics  and Resin Company,-Addyston,
Ohio, plants.
     The estimated uncontrolled plant  emissions for a model 45,500 Mg/yr
operation were calculated using previously developed data.     Storage tank
losses were estimated using AP-42 emission factors.  Total uncontrolled
emissions for the model ABS/SAN plant  are  491 Mg/yr.  Emssions estimates
for all ABS/SAN operations releasing AN to the atmosphere are given in Table 4-6
4.1.4  NBR Elastomers
     Nitrile rubber manufacturing plants emit AN  to the atmosphere through
process vents, storage tanks, and fugitive emissions sources.  The process
sources include reactors, steam strippers, and product finishing area as
discussed in the NBR process description.   Fugitive emissions are discussed
in Section 5.2.3 of this report.
     AN emissions estimates (controlled and uncontrolled) were made from
data requested from the individual manufacturers  through Section 114 requests
and plant visits.   The estimated uncontrolled plant AN emissions for a
model  23,000 Mg/yr nitrile elastomer manufacturing plant were calculated
                                           18
using previously developed EPA information.    Storage tank  losses were
                                       4-9

-------
TABLE 4-5,   SUMMARY OF ESTIMATED TOTAL AN EMISSIONS FROM ACRYLIC FIBER PRODUCTION

Plant and City
Am. Cymiamid
Milton
Badische
Wi 1 iTamsburg
On Pont
Canute n
Ou_Ponl
Wayneslioro
1 onn. Eastman
Kiiujsport
Monsanto
Oecatur

Non-
Attain- attain-
ment merit Extension
County and State area area requested
Santa Rosa X
FL
James City X
VA
Kershaw X -
SC
Augusta X -
VA
Sullivan - X
TN
Morgan X
AL

Total AN
emissions
AN emissions (Mg/yr) (My/yr)
Process Storage Fugitive EEA SRI CMA
129.5 9.3 75.1 213.9 91 91

336.6 19. 1 N/A 355.7 726 726

451 0.8 N/A 451.8 480 353

357.5 N/A N/A 357.5 338 339
24 1 N/A 25 69 69

0.2 57.7 N/A 57.9 2994 144

TOTAL 4698 1361

Comments
Fugitive is from equalization
basin. Process fugitives not in-
cluded. Storage tank emissions
Include LFQ Basin and three
100,020 gallon tanks.
Fugitive emissions not available.

Fugitive emissions not available.

Storage and fugitive emissions
are not avai (able.
Process assumes AN emissions
supplied are for only one reac-
tor train. Plant has stated
these emissions are not correct.
They wl 1 1 revise.
Fugitive emissions not available.



-------
TABLE 4-6.   AN EMISSIONS FROM ABS/SAN RESIN OPERATIONS
Plant and City
Ab tec (MobayJ
louisvi 1 le
Bory-Warner
Washington
Borcj-Warner
Ottawa
Dow
lorrance
Dow
Midland
Uow
I'ev ley
ATTyns Point
Uow
I ronton
Monsanto
Aildyston
Monsanto
Muscat ine
Monsanto
Spring field
USS Chemical
Scotls Bluff

Attain-
ment
County and State area
Jefferson
KY
Wood X
WV
La Salle
IL
Los Angeles
CA
Midland X
Ml
Jefferson X
MO
New London
CT
Lawrence
OH
llami 1 ton
Oil
Muscatine X
IA
Hamj>den
MA
East Baton Rouge
1 A

Non- Total AN
attain- emissions
ment Extension AN emissions (Mg/yrJ (Mg/yr)
area requested Process Storage Fugitive EEA SRI
X X 15 N/A 3.7 18.7 125
634 0.5 5.3 639 1769
X - - - 90 387
X X - 9
9.1 8.2 N/A 17.3 15
6
X X - - - 8
X - - 8
X X 19.9 64.7 5.0 89.6 163
(332.3) (30.9) (1.8)(365) 363
X X 3.9 5.0 1 10 27
X - - - 154
TOTAL 3084
CMA
15
1769
387
9
15
6
8
8
84
365
10
175
2901
Comments
Storage emissions data not
avai lable.
Emission reductions achieved by
AER system.
Did not receive questionnaire.
EEA estimate based on plant
visit discussion.
Did not receive questionnaire.
Did not receive questionnaire.
Numbers based on old EPA data.
Did not receive questionnaire.
Did not receive questionnaire.
Did not receive questionnaire.
Data from plant visit, 2/4/81.
Did not receive questionnaire.
Numbers based on old EPA data.
Data from Section 114 response,
March 1981.
Did not receive questionnaire.


-------
estimated using AP-42 emission factors.  Total uncontrolled emissions for
the model plant are 228 Mg/yr.
     Estimated total AN emissions from nitrile elastomer operations is
given in Table 4-7.  The emission figures were obtained from the manufac-
turers during this and previous studies.
                                       4-12

-------
TABLE 4-7.   AN EMISSIONS FROM NITRILE ELASTOMER OPERATIONS

Plant and City
Copolymer Rubber
Uaton Rouye
Goodrich
Akron
Goodrich
1 oulsvi 1 le
Goodyear
Akron
Goodyear
HonsLon
Rckhold
Cheswold
.£, Uni royal
i PafnesvTl le
OJ

Attain-
ment
County and State area
East Baton Rouge
LA
S mum it
Oil
Jefferson
KY
Summit
Oil
Harris
TX
Kent X
llE
Lake
OH

Non-
attain-
ment Extension
area requested
X

X

X X

X

X X

-
X X

AN emissions (Mg/yr)
Process Storage Fugitive
0.5 0.9 N/A

112 N/A N/A 1

318 N/A N/A

.

_

5.4 0.2 2.8
32.8 6.1 Included in
process emis-
sion estimate
TOTAL
Total AN
emissions
(Mg/yr)
EEA SRI CMA
1.4 1 1

12 123 123

303 254

91 91

79 79

8.4 not 6
inc.
38.9 53 40
650 594

Comments
Emissions calculated from re-
sponses to EPA inquiries by Pull-
man-Kel logg In 1978. Fugitive
emissions not shown by Copolymer.
Storage and fugitive emissions
are not aval lable.
Storage and fugitive emissions
are not available.
Did not receive questionnaire.

Did not receive questionnaire.

From correspondence with EPA
7/21/78 and 10/24/78.
CMA emissions reported after new
steam stripping equipment was
added.


-------
REFERENCES FOR CHAPTER 4

1.   Hobbs, F.D.  and J.A. Key.  Hydroscience, Inc.  Emissions Control
     Options for  the Synthetic Organic Chemicals Manufacturing Industry.
     Acrylonitrile Product Report for EPA Contract No. 68-02-2577.   August
     1978.  pp.  IV-2, V-2.

2.   Chemical  Manufacturers  Association (CMA).  Acrylonitrile Emission
     Estimates.   Washington,  D.C.  December 19,  1980.

3.   Suta, B.E.   SRI International.  Assessment  of Human Exposure to  Atmospheric
     Acrylonitrile.   Report  for EPA Contract No. 68-02-2835,  Task 20.
     August 1979.

4.   Reference 1.  p. IV-3.

5.   Reference 1.  p. IV-4.

6.   Hughes, T.W. and D.A. Hour.  Monsanto Research Corporation.   Source
     Assessment:   Acrylonitrile Manufacture (Air Emissions).   EPA Contract
     No. 68-02-1874.  September 1977.  pp. 21-22.

7.   Reference 2.

8.   Reference 3.

9.   Letter from  D.C. Mascone, EPA:ESED, to B.E. Suta, SRI  International.
     May 25, 1979a.

10.  Conversation between D.C. Mascone, EPA:ESED, and  B.E.  Suta,  SRI  International
     May 31, 19795.

11.  Click, C.N.  and D.O.__Moore.  Emission, Process and Control  Technology
     Study of the ABS/SAN, Acrylic Fiber, and NBR Industries.   Pullman
     Kellogg. Houston, Texas.  Final  Report for  EPA Contract  No.  68-02-2619,
     Task No. 6.   April  1979.  pp. 87-105.

12.  Control  of  Volatile Organic Emissions from  Volatile Organic  Liquid
     Storage in  Floating and  Fixed Roof Tanks.  Preliminary Draft Final
     Report for  EPA  Contract  No. 68-02-3168.  January  1981.  91  pp.

13.  Wallace, M.J.  Controlling Fugitive Emissions.  Chemical  Engineering.
     August 27,  1979.  pp. 78-92.
 14.  Reference 3.

15.  Reference 9.

16.  Reference 10.

17.  Reference 2.

18.  Reference 11.
                                      4-14

-------
                      5.  EMISSION CONTROL SYSTEMS

     This chapter discusses the air pollution control techniques applicable
to reducing AN emissions from each industry  segment.  Current control
methods and alternative control technologies are covered for each AN emis-
sion point at the various processes.  Section 5.1 presents an outline of
the control techniques which can be applied  to each of the four AN industry
segments.  These techniques are non-process  specific and concern general
plant management practices and AN storage and handling procedures.   They
include control methodologies for storage tank,  loading operation,  and
process fugitive emissions.  Section 5.2 deals with control techniques for
process vent streams and is divided into four sections.   Each section
covers process controls for one of the four  industries either producing or
consuming AN.
5.1  STORAGE AND FUGITIVE EMISSION CONTROLS
     This section discusses emission control methodologies for process
fugitive emissions, storage tank emissions,  and transfer and loading opera-
tion emissions.  These controls are applicable across all four industries
which produce and consume AN.  At present, varying control techniques are
used to reduce emissions from these sources  in the different industries.
Many of these control practices are focused  at minimizing product losses.
Others are required by State and local government regulations.
5.1.1  Current Controls
     5.1.1.1   Current Controls for Process Fugitive Emissions.   There are
currently no regulations or standard work practices for the reduction of
fugitive emissions in the SOCMI, ABS/SAN, and acrylic fiber and NBR indus-
tries.   The OSHA standard (effective in November 1978) is directed at
limiting worker exposure to AN.   It requires that exposure not exceed
                                      5-1

-------
2 pptnv over an 8-hour time-weighted average or 10 ppmv over any 15-minute
period.   This standard does not specifically require AN emissions reduction.
     5-1.1.2  Current Controls for Storage Tanks.  Storage tanks in the AN
industry are either floating or fixed roof bulk  storage tanks, other large
fixed roof storage tanks, or smaller "day" tanks.  Generally, bulk tanks
are external floating roof tanks or fixed roof tanks with internal floating
roofs.  These are employed at monomer facilities for finished product
storage and are used by fiber, ABS/SAN, and nitrile rubber producers as
primary AN monomer storage facilities.  Turnovers normally run 10 to 20 per
year.  Large fixed roof tanks are used for other purposes such as crude AN
storage at monomer facilities.  Smaller "day" tanks are used by the AN-
consuming industries and feed directly into process operations.   These
normally have capacities of approximately 113,000 liters and have more
frequent turnovers with one cycle taking several days or a week.   Day tanks
are often fixed roof tanks controlled with internal floating roofs,  although
sometimes these tanks have fixed roofs with no controls.
     Information available specifically characterizing storage tank controls
in the SOCMI, ABS/SAN, fiber, and nitrile rubber industries is limited.
However, a wide variety of techniques are known to be in use.   For example,
AN monomer producers use floating roofs, chilled water scrubbers,  vent
condensers, absorbers, and variable vapor spaces to control  AN emissions
from storage tanks.  The AN-consuming industries most commonly use internal
or external floating roofs.  Only scattered information on exact control
efficiencies has been obtained; however, chilled water scrubbers and absorbers
will normally achieve a minimum of 90 percent control  efficiency.   Control
efficiencies for internal and external floating roofs are discussed in
Section 5.1.2.2.
     5.1.1.3  Current Transfer and Handling Controls.   Limited information
has been obtained on control techniques used in transfer and loading opera-
tions.  Emissions from these operations could potentially occur at monomer
plants,  where AN is loaded into rail cars or barges, and at ABS/SAN,  fiber,
and elastomer facilities, where monomer is received and unloaded into bulk
storage tanks.
     Emissions from loading operations are controlled at two monomer facil-
ities with the use of flares.   Emissions from unloading operations at

                                       5-2

-------
ABS/SAN, fiber, and nitrile elastomer plants may currently be controlled
where bulk external floating roof tanks are present.   In this case, there
is no vapor space  that must be displaced from the storage tank, and ambient
air may displace the rnonomer within the railcar or barge.  Where fixed roof
tanks are loaded, control may be achieved by a flare,  vapor balance, or
vapor recovery system.  Many States require some form of vapor control
system in their draft SIPs (see Chapter 7).
5.1.2  Available Control
     5.1.2.1  Fuqi tives.  The EPA is currently developing CTG documents for
emissions of VOC in the synthetic organic chemical industry.   These stan-
dards focus on reducing fugitive emissions from process pumps, valves, and
flanges, and give guidelines on how these emissions reductions can be
achieved through the use of available control  technologies for improved
operation, maintenance, and housekeeping practices.  The major focus of
this draft standard is leak detection and repair.  It  is estimated that
through the use of these recommended guidelines that fugitive emissions
from the affected  plants may be reduced by 65 percent.
     Under Section 112 of the Clean Air Act, EPA has proposed generic
procedures for the control of air pollutants determined to be carcinogenic
(Federal Regi ster/Vol. 44, No. 197/October 10, 1979).  Like the fugitive
standards for SOCMI, polymers, and resins, the major thrust of the standard
would be a leak detection air repair program aimed at maintaining specific
performance levels for process pumps, valves,  and flanges.  The generic
standard would also apply to fugitive emissions from storage vessels,
transfer and handling equipment, and process equipment.
     Because AN has a lower vapor pressure than the SOCMI chemicals, the
use of a regularly scheduled leak detection and repair program to all  AN
process, storage, and transfer and loading equipment achieves significant
reduction in fugitive AN emissions.  If comparable abatement to SCCMI is
accomplished,  this would result in.approximately a 65  percent emissions
reduction.  There would be substantial  raw material savings from the imple-
mentation of these programs and, owing to the  price of AN, there could be
potential  net benefits.
     5.1.2.2  Storage Tanks.  Substantial emissions reductions can be
achieved through the retrofit of existing storage tanks with evaporative
                                      5-3

-------
                                              TABLE  5-1.  MODEL STORAGE  TANKS AND EMISSIONS REDUCTIONS

                                                       (assumes July  1900 AN price of 37tf/lb or $814/Mg
en
i

Type of
tank
Siual 1 fixed roof
5x 7 M . 151 M
60 turnovers/yr








F 1 xed roo f -,
0x10 M, 481 M
50 turnovers/yr

Floating roof -.
19x12 H. 3103 M
10 turnovers/yr


Sugijested Annuallzed Credit
Control Emissions improve- Emissions Reduction cost with AN at
ment $8l4/Mg


None 2.29 Mg/yr Internal 0.35 Mg/yr -85X $2928 $1577
contact -1 .94 Hg/yr
floating
roof w/
pr ima ry
liquid
nounted
secondary
seal


Hone 7.22 Mg/yr Same as 0.58 Mg/yr -92% $3094 $5404
above -6.64 Mg/yr

Primary 23.08 Mg/yr Continuous 1.57 Mg/yr -931 $9945 $17.506
seal secondary -21.51 Mg/yr
seal and
fixed roof
To ta 1 remo va 1
cost/year


($1351)










$1510 credit


$7561 credit



               SOURCES:   Draft CTG for VOC Control  in SOCMI  from Floating and  Fixed Roof Tanks

-------
      TABLE 5-2.   POTENTIAL EMISSIONS REDUCTION FOR SMALL FIXED ROOF
                                AN STORAGE TANK
Tank size                 Level of control            Emissions reduction

Small  fixed roof tank
  (such as day tank)
  @ 151 M3 capacity

                          Uncontrolled                      —

                          Internal floating roof
                          with primary seal only            76%

                          Internal floating roof
                          with primary and
                          secondary rim-mounted
                          continuous seal                   88% (suggested
                                                                 control)
                                           5-5

-------
        TABLE 5-3.   POTENTIAL EMISSIONS REDUCTION FOR EXTERNAL FLOATING
                                      ROOF TANK
Tank size
Level of control
Emissions reduction
External  floating
  roof tank
  @ 3483  M3 capacity
  (bulk storage)
Uncontrolled
(primary seal only)

Primary and
secondary seal

Fixed roof; contact
internal floater with
primary seal only

Fixed roof; contact
internal floating roof
with primary and
secondary seal
                                                            62%
                                                            86%
                                                            93% (suggested
                                                                 control)
                                           5-6

-------
and working emissions control systems.  The EPA has  recently examined
emissions controls for fixed and floating roof storage tanks.  These con-
trols and potential emissions reductions are  presented in EPA documents
concerned with control of VOC emissions from  storage tanks.2
     Potential emissions reductions achievable through the  implementation
of controls suggested in the EPA studies are  presented in Table 5-1.  These
estimates are based on emissions from benzene storage.  (Benzene has a
vapor pressure of  1.5 psia at 70°F; AN has a  slightly higher vapor pressure
of 1.8 psia at 70  F; therefore, these emission reduction estimates should
be comparable.)
     The most effective controls for fixed roof storage tanks, found in the
EPA study, are a contact internal floating roof with liquid-mounted primary
seals and rim-mounted continuous secondary seals.  The estimates are that
an 85 percent emissions reduction is possible when these controls are
installed on an uncontrolled fixed roof storage tank of 151 M  capacity.
This size is roughly comparable to day tanks  used in the AN-consuming
industries.  Control efficiencies of greater  than 90 percent can be obtained
through the use of recommended EPA controls on large fixed roof storage
tanks.  The suggested controls for external floating roof tanks (assumed to
be equipped with a primary seal) are the addition of a rim-mounted continuous
secondary seal and the construction of a fixed roof.  Estimated AN emissions
reductions are 93  percent with these controls.
     Potential emissions reductions from various levels of control  of fixed
and floating roof AN storage tanks are presented in Tables 5-2 and 5-3.
These estimates were developed from algorithms presented in draft EPA
documents and adjusted for AN vapor pressure, density, and molecular weight.
They are presented to show emissions reductions possible from uncontrolled
tanks and the incremental reductions possible by bringing partially con-
trolled tanks into line with recommended control  options.
     The estimated emission reduction for an  uncontrolled fixed roof tank
(5x7 M, 151 M3) equipped with a contact internal  floating roof and liquid-
mounted primary seals is 76 percent.  An additional  12 percent emission
reduction to 88 percent control  can be achieved with a rim-mounted secondary
seal.
                                       5-7

-------
     For an existing external floating roof tank  (19x12 M, 3483 M  capacity)
with primary seals, a 67 percent emission reduction is estimated with the
addition of a rim-mounted continuous secondary seal.  A fixed  roof built
over a tank with primary seals only has an estimated 86 percent reduction.
If a roof is built in addition to secondary seals, a 93 percent emission
reduction is estimated.
     Substantial emissions reductions could potentially be achieved by
applying EPA suggested controls to existing uncontrolled AN storage tanks
and by adding incremental controls to those tanks already partially con-
trolled.  The increased price of AN affords plants using these con-trols
potential net benefits through conservation of raw material  as shown in
Table 5-1.   The emission reduction and annualized cost estimates from EPA
studies of organic liquid storage in conjunction with a July 1980 AN price
of 81
-------
5.2  PROCESS CONTROLS
5.2.1  AN Monomer
     Standard methods of emissions controls such as thermal oxidizers,
flares, and scrubbers are commonly used by the AN manufacturing  industry.
Table 5-4 is a listing of emission points and control devices currently
used by domestic AN manufacturers.
     Emissions from the absorber vent are commonly controlled by combustion
devices.  The VOC emissions are pyrolyzed, oxidized, and converted to
carbon dioxide and water.  The problem of NO  emissions from combustion of
nitriles can be eased by sophisticated techniques such as staged combus-
tion.  Reduction of VOC emissions by 98 percent can be expected with proper-
ly operated thermal oxidizers.  Catalytic oxidizers can achieve a high
degree of HC emissions reduction without emitting high levels of NO  as
                                                                   /%
they operate at lower temperatures.  However, depending upon the catalytic
oxidizer system and the waste gas components, catalytic oxidizers may also
have a high rate of unburned hydrocarbon passthrough.  Carbon adsorption is
unsuitable for absorber vent emissions control because the dilute gaseous
pollutants are of low molecular weights and will not be efficiently absorbed.
     During reactor startup, the absorber vent gas continuously changes in
composition and has a higher concentration of AN than during normal operation.
Flaring is a control method that is traditionally used for variable flow,
intermittent streams such as occur during startup conditions.   Combustion
of these nitrile vapors leads to high emissions of NO .   Since the absorber
                                                     A
vent stream is more concentrated during reactor startup, an emissions
control system consisting of a water scrubber followed by carbon adsorption
has been recommended for further study.
     Emissions from the column vents and some storage tank vents are commonly
controlled by flaring.  As mentioned before, combustion of nitriles can
lead to very high emissions of NO .  Flaring also has the disadvantage of
                                 /\
allowing large amounts of VOC to pass through if existing weather conditions
are windy.   Thermal or catalytic oxidizers are optional control methods
which may be used for column vents emissions control.
     Scrubbers are frequently used as control devices for emissions from AN
storage and loading.  Gaseous emissions from storage and loading are passed

                                    5-9

-------
                              TABLE 5-4.   CONTROL DEVICES CURRENTLY USED BY THE AN INDUSTRY
         Company
                                                          Control  devices at various emission points
Absorber vent
Column vents
    Storage tank vents
 Deep well  pond
American Cyanamid
DuPont
  Beaumont, TX

  Memphis,  TN
Monsanto
  Alvin, TX

  Texas City, TX
Vistron
Thermal oxidizer

Catalytic oxidizer

None

Thermal oxidizer

Thermal oxidizer
None
Flare

Flare

Flare

Flare and vent
  condenser
Not reported
Flare
Scrubbersc
Scrubbers, refrigerated
  condenser, and flare
Scrubber
Not reported

Not reported
Flare3
Lube-oil  cover

Lube-oil  cover

None

Lube-oil  cover

Lube-oil  cover
None
 Some vents go directly to the atmosphere from conservation vents.

 No deep well pond, part of plant wastewater is concentrated and incinerated; remainder is treated and sent to
 municipal sewer.
Source:  Plant data.

-------
through  absorption  columns  and  the  column  bottoms  recycled  to  the  process.
Economics and competition  in  the  industry  has  led  to  increased  product
recovery.   Venting  emissions  from product  storage  and  handling  directly  to
the atmosphere  from conservation  vents  and  flaring  have  become  undesirable
with  rising  prices.   The use  of  vent  condensers  and  floating  roof  tanks  is
rapidly  becoming more attractive.   Carbon  adsorption and  refrigerated
compression  systems are also  possible methods of product  recovery /ami ssions
control  from product  storage  and  handling.   The  estimated costs and control
efficiencies of these control systems at a  new monomer  plant  are listed  in
Table  5-5.
      Although deep  well ponds emit  large quantities  of  hydrocarbon pollu-
tants, wastewater from AN  manufacture contains relatively small amounts of
    Q
AN.    The heavy lube  oil cover  used at  most plants  for  emissions control
has a  VOC reduction efficiency of 92  percent.  Plant wastewater may be
                              g
treated  by  solvent  extraction.    Some plant wastewater  may  also be diverted
to  the plant incinerators  in  case of  well  injection  pump  failure.
      Other  approaches to emissions  control  from  AN monomer  manufacture
involve  process changes.   Since  VOC emissions occur  from  incomplete con-
version  in  the  reactor, the  industry  has concentrated upon  the development
of  more  selective catalysts.  Escalating costs for  raw materials and control
equipment operation have encouraged the search for more selective catalysts
to  increase  AN  production  while  lowering the amount of  by-products.  This
could  make  waste streams easier  to  handle  and AN emissions  could be more
easily controlled.  A new  process developed  by the  Badger Company has the
potential for becoming the  "cleanest" way  to manufacture AN.  The process
involves the selective absorption of  AN from quenched reactor effluent in a
countercurrent absorption  column  by using  a hydrocarbon solvent.  The AN
and solvent are removed as  the  bottom stream while the  hot  absorber over-
head  gases  are  incinerated.   The  column bottoms  are  sent  to a "lights"
column to remove volatile  by-products.  The  bottoms  from  the  lights column,
containing AN in the  solvent, are sent  to  the product column where nearly
pure AN is  distilled  from  the solvent.  Although the economic feasibility
of  the process has  not yet  been  evaluated,  the process  facilitates the
disposal  of waste products  by incineration.  Thus,  wastewater treatment and
stripper corrosion  problems owing to  the handling  of sulfuric acid are
                                     5-11

-------
                               TABLE 5-5.   CONTROLS AND ESTIMATED COSTS FOR AN EMISSIONS  REDUCTION  AT  NEW MONOMER  PLANTS"
en
 l

Recommended
Control
Source System
Control
Efficiency
(*)
Estimated Costs (1980)
Capital Annual Ized
($1000) ($1000/yr)
Comment
                Reactor startup
Flare (dedicated)
901-
                                       Water scrubber/  .         99
                                       Carbon adsorption
                Absorber vents.4
                  column vents
                Column vents & IICN
                  storage S handling
Thennal oxidizer          98
  lleat recovery
  No heat recovery

Flare                     90
 300           90      Very high  NO  emission expected  -
                       29,000 ppm Nfi  at 100% fuel  nitrogen
                       conversion Including  auxiliary  fuel.

 370           60      Adsorption alone  for  high  concentra-
                       tions  has  an explosive hazard and
                       would  be  very expensive.   Scrubber
                       alone  would attain 90% control at
                       high cost.   Emissions could  be
                       vented to  scrubber for crude storage
                       tanks.

                       Operating  cost for case with heat
4300           90      recovery  fuel  credits for  recovered
2200         6400      steam  In use at sone  installations.

 200           60      Emissions  ducted  to existing emer-
                       gency  flare  system; capital costs
                       Include only ducting  and utilities
                       are  an allocated  portion of  total
                       needed for flare  system.
Crude acrylonl trile
storage
Product storage &
hand! ing
Water scrubber
Water scrubber
99
99
95
240
20
(80)
Being employed at DuPont and
Anerlcan Cyanamld plants.
These scrubbers present a net saving
of product - costs are offset by
product recovery.
                aAl I costs have been updated to real (constant) November 1980 dollars.   Cost  estimates have not been  thoroughly
                 evaluated in  light of current state-of-the-art and economics.

                 Oased on Schwartz, et al ., 1975.  Air Products developed cost estimates for  a  model  plant with 91,OOOMg/yr capacity.

                Clor estimate only.  This  Is not an established or confirmed efficiency for a flare.

                ''liased on Anguin and Anderson, 1979.  Accurex developed cost estimates  for a  model  plant with 91,000 Mg/yr capacity.

                efrom llobbs and Key, 1978.  llydroscience estimated costs (+40% to -30%  accuracy)  for  a model plant with capacity of
                 Jfl6,0()0 My/yr and operating 8760 hrs/yr.

-------
                                                                           1 p
eliminated.  Also eliminated is the operation of ammonium sulfate recovery.
5.2.2  Acrylic Fibers
     The types of control options in use by this industry segment were
outlined in the Pullman  Kellogg study.    Since the tine of that study
(1979), the only reported major changes are that better designed units have
been installed.  The increased price of acrylonitrile, however, has become
a greater  impetus to conservation and recovery measures.
     Table 5-6 lists the AN controls in use by the various manufacturers.
Activated  carbon can be  used to adsorb AN from dilute process streams.
Water scrubbers are commonly used to control streams containing AN — parti-
cularly at process vents or storage tanks.  High conversion reactors can
reduce potential  emissions downstream by leaving less AN monomer unreacted.
However, development of  high conversion technology for a specific process
by a particular manufacturer can be difficult and costly and its use may
affect product characteristics.  Mechanical seals on agitators prevent
fugitive emissions.  Slurry or solution stripping of AN from the polymer
using steam appears to be an effective control measure as it both reduces
potential  emissions downstream by reducing the concentration of unreacted
AN to 150  to 500 ppm and allowing condensation/recovery of AN-rich vapors.
Polymer stripping must be developed for each particular polymer blend and
process to maintain product quality and recover unreacted monomer.  Strippers
and scrubbers are also used to recover AN from water streams for recycle.
AN recovery and subsequent recycle is a normal part of acrylic fiber manu-
facture as it is crucial  for favorable process economics.  Washing and
filtration steps are an  integral part of most suspension polymerization
plants as the polymer is dewatered before drying; the AN content of the
polymer is also reduced  at these steps.  A vacuum flash step following
polymerization can reduce polymer AN content by about 80 percent.  As much
as 70 percent of the unreacted AN in the spinning dope can be recovered by
the solvent concentration and AN recovery sections of a solution polymeri-
zation plant.  A corresponding reduction in the AN emissions from spinning
and washing (the  largest sources in solution polymerization) is achieved.
Chemical  adducts have also been used to lower the residual AN content in
the spinning dope.

                                    5-13

-------
                         TABLE 5-6.  EXISTING AN CONTROLS IN THE ACRYLIC FIBERS INDUSTRY
                                                                                        1
                         American
                         Cyanamld
                                  Dow
                               Badische
Dupont
Monsanto
Tennessee
Eastman
en
i
       AN Removal By
       Act.  Carbon

       ILO Scrubbers
       For Tank &
       Process Vents

       High Conversion
       Reactors
Mech.  Seals for

Agitators

Slurry on
Solution Stripping

AN Stripper

AN Removal By
Washing  &
FiItaration

AN Removal By
Vaccum Flash

AN Recovery From
Dilute Solvent

Chemical Adcluct
                                        x

                                        x
               x

               x
        Moclacrylic  fibers.

-------
     5.2.2.1  Alternative Control Techniques for Acrylic Fibers.  Alternative
control techniques suggested and developed into model plants by PulIman-Kellogg
are summarized in Table 5-7-  The model plant control options for suspension
polymerization followed by wet spinning are:  (1) slurry suspension stripping
to 500 ppm (dry wt.) reducing pelletizer emissions from 1.15 to 0.02 kg/Mg,
polymer dryer emissions from 10.5 to 0.13 kg/Mg, and total model plant
emissions from 11.7 kg/Mg to 0.5 kg/Mg or 27 Mg/yr; (2) a high conversion
(95 percent rather than 87 percent) reactor and an additional wash/filter
stage (increasing AN removal efficiency from about 60 percent with two
stages to about 75 percent with three stages) reducing pelletizer and
polymer dryer emissions to 0.25 and 2.10 kg/Mg, respectively, and total
model plant emissions to 2.4 kg/Mg or 110 Mg/yr.  The model  plant control
options for solution polymerization followed by wet spinning are:   (1) solu-
tion stripping with steam to 500 ppm (dry wt.) reducing the spinning emission
factor from 9 to 0.45 kg/Mg or 20 Mg/yr for the model plant; (2) enclosure
of the spinning and washing section and venting to an incinerator (98 percent
assumed efficiency) reducing AN emissions from 9 kg/Mg to 0.45 kg/Mg or
20 Mg/yr.  The associated estimated costs (roughly updated for these control
options) are presented in Table 5-8.  For each process, control  option 1 is
expected to be less costly on total annualized cost basis.   It should be
noted that the control options are not mutually exclusive -  they could be
employed jointly or separately.  The enclosure of the spinning/ washing
section with movable plastic shields is becoming common even in combination
with slurry stripping due to OSHA regulations on AN concentrations in the
workplace.   The exhaust from the spinning/ washing section,  however, is
usually vented directly to the atmosphere and is a low AN concentration,
high air flow stream.
     5.2.2.2  Emission Reduction Systems for Acrylic Fibers.  Based on the
estimated control efficiencies and costs from the Pullman Kellogg study and
on information from plant visits, polymer stripping appears  to be an excellent
emission reduction system.     Although each manufacturer justifiably considers
their stripping operation unique because of the variations in polymer and
process, polymer stripping in general  is widespread throughout the industry.
Extensive pilot and shakedown work for each product and process line is
required, however, to ensure that polymer stripping does not degrade
                                        5-15

-------
                               TABLE 5-7.   ESTIMATED EMISSIONS FROM MODEL  PLANTS  PRODUCING
                               45,500 Mg/yr OF ACRYLIC  FIBER  USING VARIOUS CONTROL OPTIONS
                                                Emission Factor  (kg An emitted/
                                                       fiber  produced)
                                                         Control    Control
                                        Uncontrolled
                                                   Emission Rate Mg An emitted/yr

                                                               Control   Control
       Source
         Total
   AN     VOC
            Option 1   Option 2
                Total      Total
            AN   VOC   AN  VOC
                            Uncontrol led
                                  Total
                             AN    VOC
              Option 1  Option 2
                 Total      Total
              AN  VOC   AN  VOC
en

cr>
       Suspension Polymerization;
          Stripper Vent
          Pelletizer
          Polymer Dryer
          Solvent Scrubber Vent
          Crimping
          Setting Dryer
            TOTAL
 1.15   1.15
10.5   10.5
  0
 0.05
  0
1
4.05
5
0.3   0.3   	
0.02 0.02  0.25 0.25
0.13 0.13  2.10 2.10
 01       01
0.05 4.05  0.05 4.05
 05       05
                                                                    14
11.7   21.7
           0.5 10.5   2.4 12.4
 52    52
480   480
  0    46
  2.3 180
  0   230

 530  990
0.9   0.9  11     11
5.9   5.9  96    96
 0   46     0    46
2.3 180     2.3 180
 0  230     0   230
                                            23   480   110   560
Solution Polymerization:
Stripper Vent
Monomer Recovery
Scrubber Vent
Spinning Washer Vent
Incinerator Inlet
Incinerator Outlet
TOTAL

0 0.

9 9.
—
—

9.0 10.

1

9
—
—

0
0.3 0.
0 0.

0.15 1.
—
—

0.45 1.
3
1

05
--
--

45

0 0.

9 9.
9 10
0.45 0.

0.45 0.

1

9

5

5

0

410
—
—

410

4.6

450
—
—

460
14 14
0 4.6

6.8 48
— —
— —

20 66


0

410
410
20
20


4.6

450
460
23
23
       Source:  Pullman Kellogg, April  1979,  p.  103-104.

-------
TABIE 5-0.  CONTROL OPTIONS AND ESTIMATED COSTS FOR ACRYLIC FIBER BY SOLUTION AND SUSPENSION POLYMERIZATION'

Polymerization
Process
Solution
Option
1
Control
Methods
Steam Stripping
AN Emission
Reduction
(%)
95
Estimated
Capital
($1000)
3300
Costs (1980)a
Annuali zed
($1000/yr)
640
Comment
Does not affect VOC
   Polymerization
 Suspension
   Polymerization
                                     Enclosure of  spinning/  98
                                     washing  section  - vent
                                     to  incinerator
Slurry stripping
95
                                      2700
3300
                                     a)  High conversion
                                        technology
                                        (2  reactors)

                                     b)  Additional  wash
                                        step
                        79
              3200
                                     emissions;  could
                                     affect product  properti

                            870      Reduces VOC emissions
                                     85%
620      Lower costs with good
         emission control;
         requires evaluation of
         effects of fiber
         properties

720      95% AN conversion not
         always possible
  Rased  on 45,500 Mg/yr capacity,  8,000 operating hrs/yr  model plants developed by Pullman Kellogg (Click and Moore,  1979).
  Pullman Kellogg order-of-magnitude cost estimates  updated  to real  (constant) November 1980 dollars using  the Chemical
  Engineering Plant Cost Index.   Assumed capital  recovery factor of  0.147 (12 yrs, 10% interest, real dollars) and
  oTher  capital  (taxes, insurance,  etc.) factor  of 0.04 used by Hydroscience (Hobbs and Key, 1978) in their study on
  acrylonitrile  monomer production.   Cost estimates  have  not been  thoroughly evaluated in light of current  state-of-
  Ihe-art and economics.

-------
product quality and to achieve an economical residual AN  removal.  The
efficiencies of removing  residual AN from the polymer range  from 80 percent
to over 95 percent in practice.  It should be noted, however, that recovery
credits for AN will increase with the price of AN monomer.   Spinning emissions
could also be incinerated (though it would be difficult)  in  addition to
being enclosed and vented.
     Floating roof controls appear to be effective  in reducing emissions
from storage tanks.  A regularly scheduled inspection and maintenance
program could be effective in reducing fugitive emissions in the plant.
5.2.3  ABS/SAN
     Three different processes can be used to produce ABS plastics:  emulsion
polymerization, suspension polymerization, and continuous mass polymeriza-
tion.  SAN resin is generally produced by the emulsion process.   In the
emulsion process, AN emission points are the graft  reactor latex treatment
vessels, absorbers, and polymer dryers.  Suspension process  emission sources
are the prepolymerizer, the dewatering system, and  the dryer.  Emission
points' in the continuous mass process are the bulk  polymerizer,  the monomer
vapor condenser, and the devolatilizer.
     Fugitive emission sources include pumps, valves, and drains.   AN
monomer storage tanks are the largest AN emission sources, and the process
equipment handling lines containing greater than 1  percent of AN are likely
emitters of fugitives.
     Manufacturers using the emulsion process have  taken two basic approaches
to control acrylonitrile emissions:   high monomer conversion technology
(HMCT) and vent incineration.  The high conversion  approach  is an effective
emission control method and economically rewarding.   Monomer is  highly
polymerized, leaving less residual AN monomer in the polymer slurry and
lowering emissions from later processing steps.   High monomer conversion is
achieved using a second reactor the same size as the first reactor.  The
products of the first reactor are transferred to the second-stage reactor
where they are allowed to react further, increasing the monomer conversion
to around 98 percent.
     The HMCT is estimated to have 80 percent emission reduction efficiency.
It requires for effective AN emission reduction that the second reactor and
                                        5-18

-------
an absorber be  installed at  the polymer filter step.  Emulsion polymeriza-
tion processes  normally have another absorber following the graft reactors
as a part of the process.   In order to achieve high conversion without
affecting product quality, it may also be necessary to switch catalysts.
Catalyst switching has been  used by at least one manufacturer.
     The second method, vent incineration, is applicable to all  types of
processes.  Basically, in this approach all emission vents are tied into
one or more incinerators.  These incinerators may be parts of steam genera-
tors or other existing equipment in the plant.  Although the venting of all
AN emission lines to an incinerator does not involve changes which affect
the product, it can be a very complex undertaking.  This is because of the
many ducts which must be installed to carry all  vent gases containing AN to
the incinerators.
     Among the  alternative control techniques, high conversion can be used
in combination  with the incineration of dryer emissions in the emulsion and
the suspension  processes.  Due to the large volume of air exiting the
dryer, it is not possible to incinerate all the  emissions from the dryer.
A possible solution is to recycle the dryer vent and incinerate  only a
small  slipstream.  This procedure, combined with high conversion, can lead
to an AN emissions reduction of greater than 85 percent.
     Emission sources in the SAN emulsion polymerization processes are the
graft reactors, the latex treatment vessels, and polymer dryers.   The
vapors from the reactor and  latex storage vents  are sent to an AN absorber
where AN is recovered by countercurrently flowing demineralized  water.
Vapors from the latex treatment vessels which also contain a small amount
of AN which may be vented to an incinerator or the atmosphere.  The latex
from the SAN reactor is steam stripped and the overhead vapors (containing
some AN) are condensed, decanted, and directed to the AN absorber for
further AN recovery.
     In addition to having AN absorbers, many manufacturers achieve some
relatively high monomer conversion in the SAN reactors,  thus reducing re-
sidual  AN monomer.  High conversion reduces AN emissions by 80 percent.
However, if further emission reduction is desired, the only commonly
applicable method is incineration of all vent streams containing  AN.
                                       5-lg

-------
     The  incineration of  vent  streams, as  in  the ASS processes, would  be
done by connecting all the streams to one or  more incinerators.  They  may
be part of existing  steam generators or other incinerating devices.  Although
complex and extensive ductvvork is involved, careful planning can optimize
costs and overcome safety problems.
     A comprehensive vent incineration system, when applied together with
high conversion, has potential as best possible control technology for the
reduction of AN emissions.   Vent incineration is applicable to all processes
whereas high conversion,  although theoretically applicable, could require
changes in catalysts, etc.   Thus, where possible, vent incineration applied
in combination with  high  conversion appears to give the greatest AN emission
reduction.
5.2.4  NBR Elastomers
     Typical emission sources  from a nitrile  elastomer manufacturing plant
are reactors, the absorbing  and stripping columns, and the finishing section
equipment.  The fugitive  emission sources are the storage tanks and the
process equipment such as pumps, valves, flanges, and drains.
     In the manufacture of nitrile elastomers, high conversion technology
is not applicable since the  desired end product properties can be obtained
only with a specific percentage of monomer polymerization.  Any variation
in the monomer conversion will alter the properties of the product.
     An effective and economically beneficial  method of Mi reduction is to
steam strip the latex.  The  latex formed in the polymerization reactor  is
first flashed to remove residual butadiene and some AN and then steam
stripped.  The stripper tops containing AN are sent to condensers for AN
recovery.  The downstream AN emissions are reduced and unreacted AN is
recovered.  Steam stripping  efficiency is estimated to be greater than
90 percent.
     Among the alternative control  techniques are incineration of reactor
gases released at the flashing step.  The latex is flashed in the post-
reactor section and  then  pumped to the blend  tank.  Semi-tight enclosures
could be used for all the tanks to collect vent gases and these gases could
be combined with dryer exhaust and sent to the incinerator.  Alternatively,
flashing of the latex may be omitted and vent gases from all  operations may
be combined and sent to the  incinerator.
                                       5-20

-------
     The highest AN emission reductions for NBR plants appears to be achieved
by latex stripping with steam, followed by AN recovery and incineration of
vented gases in the finishing sections.  The latex from the reactors is
steam stripped.  The stripper overhead, containing the unreacted AN, is
sent to a water condenser followed by a chilled ammonia condenser for AN
recovery.  Some AN may remain in the latex and emissions from the finishing
process may need to be controlled.  This may be done by routing all  the AN
vent streams containing AN to an incinerator.
                                      5-21

-------
REFERENCES FOR CHAPTER 5

 1.   Guideline Series:  Control of Volatile Organic Fugitive Emissions  from
     Synthetic Organic Chemical, Polymer, and Resin Manufacturing  Equipment.
     Preliminary draft.  U.S. EPA:OAQPS.  January 1981.

 2.   Guideline Series:  Control of Volatile Organic Emissions from Volatile
     Organic Liquid Storage  in Floating and Fixed Roof Tanks.  Preliminary
     draft.   U.S. EPA:OAQPS.  January 1981.

 3.   Summary of Group  I Control Technique Guideline Documents for  Control
     of Volatile Organic Emissions from Existing Stationary Sources.
     EPA-450/3-78-120.  December 1978.  pp. 2.21-2.22.

 4.   Reference 3.  pp. 2.23-2.24.

 5.   Hobbs,  F.D. and J.A. Key.  Hydroscience, Inc.  Emissions Control Options
     for the Synthetic Organic Chemicals Manufacturing Industry.   Arcylonitrile
     Product Report.   EPA Contract No. 68-02-2577.  August 1978.    P. V-6.

 6.   Anguin, M.T. and  S. Anderson.  Accurex Corp.  Acrylonitrile Plant Air
     Pollution Control.  EPA Contract No. 68-03-2567.   February 1979.  p. 1-2..

 7.   Reference 5.  pp. 19, 21.

 8.   Hughes, T.W. and  D.A. Horn.  Monsanto Research Corporation.   Source
     Assessment:  Acrylonitrile Manufacture (Air Emissions).   EPA Contract
     No. 68-02-1874.   September 1977.  p. 38.

 9.   Reference 8.  pp. 3-21.

10.   Mitre Corporation.  Acrylonitrile Manufacture:   Pollutant Prediction and
     Abatement.  EPA Contract No. MTR-7752.  pp. 62-65.

11.   Click,  C.N. and D.O. Moore.  Emission, Process and Control Technology
     Study of the ABS/SAN, Scrylic Fiber, and NBR Industries.  Pullman
     Kellogg.   Houston, Texas.  Final Report for EPA  Contract No.  68-02-
     2619, Task No.  6.  April 1979.  pp. 151-193.

12.   Reference 11.   pp. 87-112.
                                     5-22

-------
                            5.  EMISSION DATA

6.1  AVAILABILITY OF DATA
     The AN emissions data used in this report has been gathered from the
CMA, individual plants (Section 114 requests plant visit questionnaires),
permit files (Regional EPA offices and State offices), and previous EPA
studies (SRI, Hydroscience, and Pullman Kellogg).  These data were developed
by using limited emissions testing and engineering calculations.  The
available point source emission test data were gathered by in-house testing
teams using a variety of sampling methods.  Test methods differed even
within the same company; one plant used a grab sample with gas chromato-
graphic (GC) analysis while another sampled vapors by carbon absorption and
elution to the GC.  The collected data on storage tank and fugitive emissions
were obtained using emission factors in AP-42 and emission factors published
by the American Petroleum Institute.  The emissions data were not verified
during this survey owing to the limited time and scope of this portion of
the NESHAP regulatory development.  There were no specific data on uncon-
trolled AN emissions other than model plant estimates made in other studies
of the AN industry (Pullman Kellogg and Hydroscience).
6.2  SAMPLE COLLECTION AND ANALYSIS
     The EPA does not have a reference sampling and analysis method for
measuring AN emissions from point sources.  The EPA RM 25 is used to measure
total gaeous non-methane organic compounds from petrochemical operations.
It extracts a gaseous grab sample into a clean, evacuated stainless steel
dry ice trap.   The trap condenses organic vapors.  This trap is backed up
by an evacuated stainless steel flask to hold non-condensible gases.   These
samples are returned to the laboratory where they are heated for combustion
of vapors (the flasks contain sufficient oxygen at known pressure for
                                   6-1

-------
pyrolysis) to carbon dioxide.  A special GC measures total organic vapor
content.  The RM 25 method can also use a gas chromatograph that samples
gases continuously from the source and  performs oxidations In the  instrument
for total organics analysis.
     The analysis of AN emissions from  point sources could be accomplished
using a modified RM 25.  The grab sample technique would  probably  be the
most applicable since the GC needed to  do qualitative and quantitative
analysis of mixed component systems may not be readily transported (this is
not always the case).  The stainless steel  traps or glass flasks may be
used for the grab sample.  Eluting the  sample into a GC prepared to separate
and analyze individual  gas components can be accomplished by heating the
flasks below the combustion temperature of the gases.  Total mass  emission
rate can be calculated using measured effluent flow parameters  (volumetric
flow rate, temperature, and pressure) and the concentration of  AN  in the
grab sample.
                                    6-2

-------
                     7-  STATE AND  LOCAL  REGULATIONS

     This chapter discusses State and  local  regulations pertaining to VOC
emissions and their applicability to AN emission  sources.  Federal regula-
tions that could affect AN emission sources  also  are examined.
     There are no States which currently  have  specific regulations limiting
acrylonitrile emissions to the atmosphere; therefore, AN emissions from all
four industry segments (ABS/SAN, fibers,  NBR,  and monomer) are regulated
under State VOC standards.  Facilities manufacturing or using AN are located
in 18 States, 9 of which regulate AN emission  sources by applying the
general provisions of their VOC regulations  for process, storage, and
loading VOC emissions.  The States  regulating  all three types of emissions
are Texas, Ohio, Tennessee, Louisiana, Illinois,  California, Connecticut,
Michigan, and Kentucky.  Nine States have no VOC process, storage, or
loading regulations.  These are South  Carolina, Florida, Mississippi
(Borg-Warner is building a bulk process ABS/SAN plant in Bienville, Mississippi),
Delaware, Virginia, Iowa, West Virginia,  Missouri and Massachusetts.   The
only State that has just VOC storage and  loading  regulations is Alabama.
     The majority of the States which  have process regulations require
85 percent reduction of process VOC emissions.  The States requiring a per-
centage reduction also may require  specific  control techniques.  Recommended
techniques include vent gas incineration, adsorption, vapor recovery systems,
or any other control system which has  been approved by the State for use
and meets emission reduction requirements.   Incineration is the most common
VOC emission control technique required by the State regulations.
     Ohio, which has the highest number of AN  emission sources, requires
this 85 percent VOC emission reduction to be achieved if process VOC source
discharges to the atmosphere are greater  than  18  Kg/day (40 Ibs/day) or
3.6 Kg/hr (8 Ibs/hr).   Under Ohio regulations, emission reductions can  be
                                    7-1

-------
achieved through any of the above-mentioned controls except vapor recovery
systems.  Ohio's regulations apply only to those areas classified as Priority I,
while other areas of the State have no VOC regulations.  Connecticut and
Illinois set mass reduction levels the same as those in Ohio.   VOC emission
sources in these States which have higher VOC emissions than this level
must apply control techniques to achieve 85 percent emission reduction.
     Some States do not require a percentage reduction; they only require
the sources to apply emission control devices.   In the cases of Texas,
which has the second highest number of facilities per State, and Louisiana,
VOC emission reduction is required by burning the VOC vent gas streams in a
smokeless flare or a direct flame incinerator or by applying any other
approved emission control device.   Texas requires VOC emission controls in
both ozone attainment and nonattainment areas.
     VOC emission sources located in Michigan and Tennessee are required to
meet the lowest achievable emission rate (LAER).   Requiring the emission
source to meet LAER instead of applying a specific emission control  device
does not preclude the use of any effective control devices an industry may
want to employ.   The industry must demonstrate to the State during the
permitting process that the control  technique will meet LAER.
     Eleven States have regulations  for storage tank VOC emissions covering
both large and small  storage tanks.   Large tanks range in size from 95 M
             3                                        3
to over 246 M  with the average size being about 150 M  or more.  Generally,
States require large storage tanks to have controls such as double seal
floating roofs or vapor recovery systems unless these tanks are pressure
tanks capable of maintaining a working pressure sufficient to prevent vapor
loss.   Small  tanks range in size from 1 M  to just under 150 M .  There are
two types of small  tanks.   Those averaging from 1 to over 3.8 M  are usually
required to use submerged fill  pipes to prevent excessive vapor emissions;
however, they may be equipped with other control  devices.   Tanks up to
150 M  in size are required to have floating roofs or vapor recovery systems
if they are not pressure tanks.   The control  method used for both large and
small  tanks is dependent upon the vapor pressure of the VOC being stored.
     Ten States have regulations for loading or unloading VOC from any
loading facility.   The control  equipment required depends on the throughput
level  of a loading system.   This level ranges from 39,000 to over 150,000
                                     7-2

-------
liters.  The average capacity of the vehicle or tanks being loaded is
852 liters.  The most commonly used control technique is a vapor collection
and disposal system consisting of an adsorber or condensation system which
recovers 90 percent of the emissions.   This system is equipped with a
loading arm with pneumatic, hydraulic, or other mechanical means which
provides a vapor tight seal between the adapter and the tank or vehicle
hatch.  Other controls which are used at loading facilities are a submerged
fill pipe or an equally efficient approved control device.  Drainage into
the vehicle or tank also must be provided before the removal of the arm so
as to prevent spills during attachment and removal of the loading arm.
     Connecticut is the only State that has established an odor threshold
limit for AN.   The regulation states a facility cannot emit a substance
with an objectionable odor beyond its property line.   The odor threshold
limit for AN is 21.4 ppm.
     Along with the State regulations there are local regulations that
could be applicable to AN emissions.   Kentucky and Tennessee have local
regulations for VOC emissions.   However, Memphis, Tennesse, VOC regulations
do not apply to AN emission sources since the regulations only regulate
specific sources, such as  coating industries.   Jefferson County, Kentucky,
regulates process, storage, and loading emissions.  The chemical reaction
process regulations apply to both of the AN sources in the State and require
84 percent reduction or a vapor recovery system which recovers 90 percent.
Storage tanks in Jefferson County under 1000 liters are not regulated.
Tanks ranging from 1000 to 150,000 liters are required to be equipped with
a submerged filler pipe,  floating roof, or equivalent control equipment.
Those storage vessels which are pressurized are exempt from the standard.
Loading facilities are required to install  vapor recovery systems.
     In addition to State and local regulations, OSHA has established
Federal regulations which may be applicable to AN emission sources.   OSHA
limits worker exposure to AN by establishing a workplace standard of 2 ppm
averaged over 8 hours or 10 ppm averaged over any 15-minute period during
the working day.   This standard went into effect in November 1978 and
allowed industry two years to install  the required engineering controls to
limit exposure.  OSHA is only concerned with worker exposure to AN emissions
which could affect employee safety and health; the regulations do not
                                     7-3

-------
specifically limit AN emissions to the atnosphere.  One engineering wo rk
practice which can be used to comply with the OSHA standard to ensure
employee safety is to vent the airborne AN to a location where employees
will  not be exposed.   These fugitive AN emissions then may become an atmos-
pheric emission problem.  In those cases where the emissions are not vented
to the atmosphere, the standard has led to a substantial reduction in
fugitive emissions.
     The EPA is presently developing air pollution regulations and guidelines
for SOCMI which could have an impact on atmospheric AN emission levels.
Process VOC emissions from new or modified AN monomer plants will be regulated
by the air oxidation  unit process NSPS currently being developed.  The
proposed fugitive  VOC emission standard for SOCMI would apply to fugitive
emissions from new AN monomer facilities.   The SOCMI regulation being
developed for VOL  storage would regulate tanks with a capacity greater than
150,000 liters and a  vapor pressure greater than 1.5 psia and would apply
to new storage tanks  in all  four segments of the AN industry.  The AN
monomer capacity increases planned by Vistron at Green Lake, Texas, and  by
Monsanto at Texas  City, Texas, would not be regulated by either the process,
fugitive, or storage  NSPS.  The new ABS/SAN resin facility being constructed
by Borg-Warner at  Bienville,  Mississippi,  would not be regulated by the
storage NSPS.   These  plants  are not regulated because construction permits
have  been issued to these plants prior to NSPS promulgation.
     There are also CTGs being developed by EPA for process, fugitive, and
storage emissions  from SOCMI and are applicable to existing sources in
nonattainment areas requesting a SIP extension.  The CTGs are designed to
assist the States  in  developing regulations for these emissions.  Since  a
majority of the AN facilities are in nonattainment areas, these facilities
could be affected  by the guidelines if they are adopted by the States.
Table 7-1 lists the attainment status for each facility.
                                      7-4

-------
TABLE 7-1.   ATTAINMENT STATUS
Producer Location
(city, county, state)
AN Monomer Facilities:
American Cyanamid
OuPont
Dupont
Monsanto
Mosanto
Sohio (Vistron)
Sohio (Vistron)
ABS/SAN Facilities:
Abtec (Mobay)
Borg-Warner
Borg-Warner
Dow
Dow
Dow
Dow
Dow
Monsanto
Monsanto
Monsanto
USS Chemical

Acrylic Fiber Facilities
American Cyanamid
Badische
DuPont
OuPont
Kodak
Monsanto

West Wego, Jefferson, LA
Beaumont, Hardin, TX
Memphis, Shelby, TN
Alvin, Brazoria, TX
Texas City, Galveston, TX
Lima, Allen, OH
Victoria, Victoria, TX

Louisville, Jefferson, KY
Washington, Wood, WV
Ottawa, LaSalle, IL
Torrance, Los Angeles, CA
Midland, Midland, MI
Pevley, Jefferson, MO
Allyns Point, New London, CT
Ironton, Lawrence, OH
Addyston, Hamilton, OH
Muscatine, Muscatine, IA
Springfield, Hampden, MA
Scotts Bluff, East Baton
Rouge, LA
:
Milton, Santa Rosa, FL
Williamsburg, Surry, VA
Camden, Kershaw, SC
Waynesboro, Augusta, VA
Kingsport, Sullivan, TN
Decatur, Morgan, AL
Attainment status
Ozone

Nonattainment
Attainment
Nonattainment
Nonattai nment
Nonattainment
Nonattainment
Nonattainment

Nonattainment
Attainment
Nonattainment
Nonattainment
Nonattainment
Attainment
Nonattainment
Nonattainment
Nonattainment
Attainment
Nonattainment
Nonattainment


Nonattainment
Attainment
Attainment
Attainment
Nonattainment
Nonattainment
Extension
Requested









Yes


Yes


Yes

Yes

Yes









Nitrile Elastomer Facilities:
Copolymer Rubber

Goodrich
Goodrich
Goodyear
Goodyear
Reichold
Uni royal
Baton Rouge, East Baton
Rouge, LA
Akron, Summit, OH
Louisville, Jefferson, KY
Akron, Summit, OH
Houston, Harris, TX
Cheswold, Kent, CE
Painesville, Lake, OH
Nonattai nment

Nonattainment
Nonattainment
Nonattainment
Nonattainment
Attainment
Nonattai nment



Yes

Yes

Yes
           7-5

-------