UnHedStates     Office of Air Quality     453R94033A
Environmental Protection Planning and Standards     May 1994
Agency        Research Triangle Park NC 27711
Emissions from Epoxy
Resins Production and
Non-Nylon Polyamides
Production -
Background Information
for Proposed Standards
                                   1

-------
                                   EPA-453/R-94-O338
Emissions from Epoxy Resins Production
 and Non-Nylon Polyamides Production

       Background Information for
           Proposed Standards
              Emission Standards DMslon
              . ENVIRONMENTAL PROTECTION AQENCPf
             Offia* of Mr QutUty Pltnnlng »nd Sttndintt
             flMMrah Tri»ngl» Puk. North Cwo/AM 37711

-------
     This report has been reviewed by the Emission Standards
Division of the Office of Air Quality Planning and Standards, EPA
and approved for publication.  Mention of trade names or
commercial products is not intended to constitute endorsement or
recommendation for use.  Copies of this report are available
through the Library Services Office (MD-35), U. S. Environmental
Protection Agency, Research Triangle Park, North Carolina  27711;
from the Office of Air Quality Planning and Standards Technology
Transfer Network, U. S. Environmental Protection Agency, Research
Triangle park, North Carolina  27711;  or, for a fee, from the
National Technical Information Services, 5285 Port Royal Road,
Springfield, Virginia  22161.
                               ii

-------
                  ENVIRONMENTAL PROTECTION AGENCY

                      Background Information
                             and Draft
                  Environmental Impact Statement
    for Epoxy Resins Production and Non-Nylon Polyamide Resins
                            Production

                           Prepared by:
/Bruce  Jordan                                          '  (date/
 Director^  Emission Standards  Division
 U.S. Environmental Protection Agency
 Research Triangle  Park,  North Carolina  27711

 1.  The proposed national  emission standard would limit  emissions
    of hazardous air pollutants  from existing  and new facilities
    that manufacture epoxy resins  and non-nylon polyamide resins.
    The proposed standards implement Section 112 of  the  Clean Air
    Act as amended in 1990 and are based on the Administrator's
    determination  of July  16,  1992 (57 FR 31576)  that epoxy
    resins and  non- nylon polyamide resins manufacturing  generate
    significant emissions  of  certain hazardous air pollutants
    listed in Section 112 (b)  of  the Act,  primarily
    epichlorohydrin .

 2 .  Copies of this document have been sent to  the following
    Federal Departments:   Labor, Health and Human Services,
    Defense, Transportation,  Agriculture,  Commerce,  Interior,  and
    Energy; the National Science Foundation; the Council on
    Environmental  Quality;  members of the State and  Territorial
    Air Pollution  Program  Administrators;  the  Association of
    Local  Air Pollution  Control  Officials;  EPA Regional
    Administrators;  Office of Management and Budget;  and other
    interested  parties.

 3.  The comment period for review  of this document is 60 days.
    Mr. Randy McDonald,  Chemicals  and Petroleum Branch,  telephone
     (919)  541-5402,  may  be contacted regarding the date  of the
    comment period.

 4 .  For additional information contact :

    Mr. Randy McDonald (MD-13)
    Chemicals and  Petroleum Branch
    U. S.  Environmental  Protection Agency
    Research Triangle Park, N. C.   27711
    Telephone:   (919)  541-5402
                                iii

-------
Copies of this document may be obtained from:

U. S. EPA Library (MD-36)
Research Triangle Park, N.C.  27711

National Technical Information Service
5285 Port Royal Road
Springfield, Virginia  22161
                            IV

-------
                        TABLE OF CONTENTS
                                                             Page

LIST OF FIGURES	      ix

LIST OF TABLES	       X

CHAPTER 1.  SUMMARY                                           1-1

            1.1  INFORMATION GATHERING  	     1-1
            1.2  EMISSION DATA ANALYSIS	     1-2
            1.3  CONTROL DEVICE DESIGN AND COST ANALYSIS      1-2
            1.4  ECONOMIC IMPACT ANALYSIS 	     1-3

CHAPTER 2.  INTRODUCTION  	     2-1

            2.1  BACKGROUND AND AUTHORITY FOR STANDARDS  .     2-1
            2.2  SELECTION OF POLLUTANTS AND
                 SOURCE CATEGORIES  	     2-5
            2.3  PROCEDURE FOR DEVELOPMENT OF NESHAP   .  .    .2-6
            2.4  CONSIDERATION OF COSTS	     2-9
            2.5  CONSIDERATION OF ENVIRONMENTAL IMPACTS  .    2-10
            2.6  RESIDUAL RISK STANDARDS	    2-11

CHAPTER 3.  INDUSTRY DESCRIPTION  	     3-1

            3.1  GENERAL	     3-1
            3.2  SOURCE CATEGORY DESCRIPTIONS 	     3-2
                 3.2.1  Basic Liquid Resins  (BLR)
                        Production	     3-2
                 3.2.2  Wet Strength Resins  (WSR)
                        Production	     3-7
            3.3  BASELINE EMISSIONS	    3-13
                 3.3.1  BLR Production	    3-13
                 3.3.2  WSR Production	    3-18
            3.6  REFERENCES FOR CHAPTER 3	    3-19

CHAPTER 4.  EMISSION CONTROL TECHNOLOGIES 	     4-1

            4.1  CONDENSERS	     4-4
                 4.1.1  Design	     4-5
                 4.1.2  Vent Condensers	     4-6
                 4.1.3  Application to the Source
                        Categories	     4-8
            4.2  SCRUBBERS	    4-12
                 4.2.1  General Gas Absorbers	    4-12
                 4.2.2  Design	    4-13
                 4.2.3  Scrubber Applications 	    4-14
            4.3  CARBON ADSORPTION  	    4-15
                 4.3.1  Design	    4-15
                 4.3.2  Applicability	    4-17

-------
                  TABLE OF CONTENTS  (continued)
                                                             Page
            4.4  THERMAL DESTRUCTION	    4-19
                 4.4.1  Flares	    4-19
                 4.4.2  Thermal and Catalytic Oxidizers  .    4-21
            4.5  OTHER CONTROL MEASURES	    4-23
                 4.5.1  Vapor Containment 	    4-23
                 4.5.2  Operational Practices 	    4-24
            4.6  LEAK DETECTION AND REPAIR  (LDAR) AND
                 PRESSURE TESTING PROGRAMS  	    4-25
            4.7  STORAGE TANK CONTROLS	    4-26
            4.8  REFERENCES FOR CHAPTER 4	    4-26

CHAPTER 5.  MACT FLOORS, REGULATORY ALTERNATIVES, AND
            MODEL PLANTS	     5-1

            5.1  MACT FLOORS	     5-1
                 5.1.1  BLR MACT Floors .........   .  5-2
                 5.1.2  WSR MACT Floors	     5-3
            5.2  REGULATORY ALTERNATIVES AND IMPACTS   .  .     5-4
                 5.2.1  BLR Regulatory Alternatives  ...     5-4
                 5.2.2  WSR Regulatory Alternatives  ...     5-7
            5.3  MODEL PLANTS	     5-9
            5.4  REFERENCES	     5-9

CHAPTER 6.  ENERGY AND ENVIRONMENTAL IMPACTS  	     6-1

            6.1  ENERGY IMPACTS	     6-1
            6.2  AIR QUALITY IMPACTS	     6-1
            6.3  SOLID WASTE AND WASTEWATER IMPACTS  ...     6-1

CHAPTER 7.  COSTS	     7-1

            7.1  CONTROL TECHNOLOGIES EXAMINED FOR
                 PROCESS VENTS  	     7-1
                 7.1.1  Condenser Cost Model	     7-2
                 7.1.2  Gas Absorber Cost Model	     7-7
                 7.1.3  Auxiliary Equipment Costs  ....    7-11
                 7.1.4  Process Modifications 	    7-13
            7.2  CONTROL TECHNOLOGIES EXAMINED FOR
                 EQUIPMENT LEAKS  	    7-14
                 7.2.1  Capital Costs	    7-14
                 7.2.2  Annual Costs	    7-14
            7.3  ENHANCED MONITORING COSTS FOR
                 WASTEWATER	    7-16
            7.4  RESULTS OF THE COST ANALYSIS	    7-16
            7.6  REFERENCES FOR CHAPTER 7	    7-20
                                VI

-------
                  TABLE OF CONTENTS (continued)

                                                             Page

CHAPTER 8.  ECONOMIC IMPACT ANALYSIS  	     8-1

            8.1  INTRODUCTION	     8-1
                 8.1.1  EIA Objectives	     8-1
                 8.1.2  Background	     8-2
                 8.1.3  Summary of Estimated Impacts  .   .     8-3
                 8.1.4  Organization of EIA	     8-8
            8.2  OVERVIEW OF ECONOMIC IMPACT ANALYSIS .   .     8-8
                 8.2.1  Overview of Distributional
                        Impacts	     8-8
                 8.2.2  Economic Impact Studies 	    8-10
                 8.2.3  Industry Profile  	    8-10
                 8.2.4  Primary Impacts	    8-11
                 8.2.5  Capital Availability Analysis .   .    8-18
                 8.2.6  Evaluation of Secondary Impacts  .    8-22
                 8.2.7  Affected Plants	    8-22
            8.3  INDUSTRY PROFILE 	    8-22
                 8.3.1  Product Descriptions and
                        End Uses	    8-23
                 8.3.2  Market Structure  	    8-29
                 8.3.3  Market Outlook	    8-33
                 8.3.4  Foreign Trade	    8-35
                 8.3.5  Financial Data	    8-37
            8.4  PRIMARY ECONOMIC IMPACTS AND CAPITAL
                 AVAILABILITY ANALYSIS  	    8-40
                 8.4.1  Introduction	    8-40
                 8.4.2  Estimates of Primary Impacts  .   .    8-40
                 8.4.3  Capital Availability Analysis .   .    8-43
                 8.4.4  Limitations of Estimated Primary
                        Impacts	    8-44
                 8.4.5  Summary of Primary Impacts  ...    8-45
            8.5  SECONDARY ECONOMIC IMPACTS	    8-46
                 8.5.1  Introduction	    8-46
                 8.5.2  Labor Impacts	    8-46
                 8.5.3  Energy Use Impacts	    8-48
                 8.5.4  Foreign Trade Impacts 	    8-48
                 8.5.5  Regional Impacts  	    8-51
                 8.5.6  Limitations of Estimated
                        Secondary Impacts 	    8-51
                 8.5.7  Summary of Secondary Impacts  .   .    8-51
            8.6  POTENTIAL SMALL BUSINESS IMPACTS ....    8-51
            8.7  ECONOMIC COSTS	    8-52
                 8.7.1  Economic Costs of Emission
                        Controls:   Conceptual Issues  .   .    8-52
                 8.7.2 .Other Costs Associated with
                        NESHAP	    8-57
                 8.7.3  Changes in Economic Surplus
                        as a Measure of Costs	    8-58
                 8.7.4  Estimates of Economic Costs ...    8-58

                               vii

-------
                  TABLE OF CONTENTS  (continued)

                                                              Page

APPENDIX A - AFFECTED PLANTS AND EMISSION CONTROL COSTS  .     A-l

APPENDIX B - TECHNICAL DESCRIPTION OF ANALYTICAL METHODS      B-l

APPENDIX C - ESTIMATION OF INDUSTRY SUPPLY AND DEMAND  .  .     C-l

APPENDIX D - SENSITIVITY ANALYSES 	     D-l

REFERENCES FOR CHAPTER 8	     D-6

APPENDIX E - EVOLUTION OF THE BACKGROUND INFORMATION
             DOCUMENT 	  .....     E-l

APPENDIX F   INDEX TO ENVIRONMENTAL CONSIDERATIONS   ...     F-l

APPENDIX G   MODEL PLANT CHARACTERISTICS  	  .     G-l

APPENDIX H   TABULAR COSTS	     H-l
                               viii

-------
                         LIST OF FIGURES

                                                             Page

Figure 3-1.  Epoxy resin process flow diagram 	     3-4

Figure 3-2.  Non-nylon polyamide (wet strength)  resin
             flow diagram	    3-11

Figure 8-1.  Partial Equilibrium Analysis of Basic Liquid
             Resins and Wet Strength Resins	    8-12

Figure B-l.  Domestic Market Supply Shift Due to Emission
             Control Costs  	     B-4
                               IX

-------
                          LIST OP TABLES
                                                              Page
TABLE 3-1.   EMISSIONS INFORMATION FOR FACILITIES THAT
             PRODUCE BASIC LIQUID RESINS  (BLR)   	      3-8

TABLE 3-2.   BLR CONSUMPTION PATTERNS	      3-9

TABLE 3-3.   EMISSIONS INFORMATION FOR FACILITIES THAT
             MANUFACTURE WET STRENGTH RESINS  (WSR)   ...     3-14

TABLE 4-1.   SUMMARY OF EXISTING EMISSION CONTROL DEVICES
             FROM EMISSIONS SOURCES IN BLR MANUFACTURING
             PROCESSES	      4-2

TABLE 4-2.   SUMMARY OF EXISTING EMISSION CONTROL DEVICES
             ON EMISSION SOURCES IN WSR MANUFACTURING  .  .      4-3

TABLE 5-1.   MACT FLOORS	      5-3

TABLE 5-2.   IMPACTS OF MEETING MACT FLOORS AND  REGULATORY
             ALTERNATIVES FOR BLR SOURCE CATEGORY    ...      5-6

TABLE 5-3.    IMPACTS OF MEETING MACT FLOORS AND REGULATORY
             ALTERNATIVES FOR WSR SOURCE CATEGORY ....      5-9

TABLE 6-1.   ENERGY IMPACTS	      6-2

TABLE 6-2.   SOLID WASTE AND WASTEWATER IMPACTS	      6-3

TABLE 7-1.   BLR MANUFACTURING COST ANALYSIS RESULTS   .  .     7-18

TABLE 7-2.   WSR MANUFACTURING COST ANALYSIS RESULTS   .  .     7-19

TABLE 8-1.   SUMMARY OF ESTIMATED ECONOMIC IMPACTS   ...      8-4

TABLE 8-2.   ESTIMATES OF ANNUALIZED ECONOMIC COSTS  ...      8-7

TABLE 8-3.   BIASES RESULTING IF MODEL ASSUMPTIONS ARE
             VIOLATED	     8-17

TABLE 8-4.   EPOXY END-USE CONSUMPTION	     8-24

TABLE 8-5.   COMPANIES PRODUCING UNMODIFIED EPOXY RESINS
             (BASIC LIQUID RESINS), 1980-1991 	     8-31

TABLE 8-6.   COMPANIES PRODUCING EPI-BASED NON-NYLON
             POLYAMIDE RESINS (WET STRENGTH RESINS),
             1988 and 1990	     8-32

TABLE 8-7.   U.S. TRADE IN BASIC LIQUID RESINS  (BLR)   .  .     8-36

-------
                    LIST OF TABLES (continued)

                                                             Page

TABLE 8-8.   U.S. TRADE IN WET STRENGTH RESINS  (WSR)   .  .    8-38

TABLE 8-9.   FINANCIAL DATA FOR RESIN PRODUCERS  .....    8-39

TABLE .8-10.  ESTIMATED PRIMARY IMPACTS ON BLR AND WSR
             MARKETS  .	    8-42

TABLE 8-11.  ESTIMATED EMPLOYMENT LOSSES	    8-47

TABLE 8-12.  ESTIMATED ENERGY USE REDUCTIONS	    8-49

TABLE 8-13.  ESTIMATED IMPACTS ON NET EXPORTS	    8-50

TABLE 8-14.  EMPLOYMENT OF RESIN PRODUCERS	    8-53

TABLE 8-15.  ESTIMATES OF ANNUALIZED ECONOMIC COSTS  ...    8-59

TABLE 8-16.  ESTIMATES OF THE ANNUALIZED EMISSION
             CONTROL COSTS  	    8-62

TABLE A-l.   CONTROL COSTS AT BLR PLANTS  ........     A-2

TABLE A-2.   CONTROL COSTS AT AFFECTED WSR PLANTS  ....     A-3

TABLE A-3.   OPTION I CONTROL COSTS AT AFFECTED WSR  PLANTS    A-4

TABLE B-l.   BASELINE INPUTS	    B-12

TABLE C-l.   VARIABLES AND DEFINITIONS OF PRIMARY DATA  .     C-6

TABLE C-2.   ESTIMATED PRODUCTION FUNCTION COEFFICIENTS  .     C-9

TABLE C-3.   ESTIMATED DERIVED DEMAND COEFFICIENTS   ...    C-10

TABLE D-l.   SENSITIVITY ANALYSIS:  ESTIMATED PRIMARY
             INPUTS ON THE BLR MARKETS WITH  LOW
             ELASTICITY OF DEMAND 	     D-2

TABLE D-2.   SENSITIVITY ANALYSIS:  ESTIMATED PRIMARY
             INPUTS ON THE WSR MARKETS WITH  LOW
             ELASTICITY OF DEMAND 	     D-3

TABLE D-3.   SENSITIVITY ANALYSIS:  ESTIMATED PRIMARY
             INPUTS ON THE BLR MARKETS WITH  HIGH
             ELASTICITY OF DEMAND	  .     D-4

TABLE D-4.   SENSITIVITY ANALYSIS:  ESTIMATED PRIMARY
             INPUTS ON THE WSR MARKETS WITH  HIGH
             ELASTICITY OF DEMAND 	     D-5


                                xi

-------
                    LIST OF TABLES (continued)

                                                              Page

TABLE E-l.   EVOLUTION OF THE BID	      E-2

TABLE F-l.   CROSS-INDEXED REFERENCE SYSTEM TO HIGHLIGHT
             THE ENVIRONMENTAL IMPACT PORTIONS OF THE
             DOCUMENT	      F-2
                               xii

-------
                           1. 0   SUMMARY

     The draft Background Information Document  (BID) chapters
included in this package refer to the control of hazardous air
pollutants (HAP's) from the manufacture of basic liquid epoxy
resins  (BLR)  and wet-strength resins  (WSR).  These preliminary
chapters provide estimates of uncontrolled and baseline HAP
emissions and present information on applicable control
techniques.  Additionally, recommended regulatory alternatives
and their economic impacts are estimated for individual
facilities and entire industries.  The chapters have just
recently been drafted, and none of the information has yet been
peer-reviewed by industry.  Thus, the information contained
herein is preliminary and will presumably be revised based on
comments.  The following paragraphs are intended to provide
general information regarding the regulatory approach and provide
assumptions and methods used to arrive at the information
presented in the chapters.  The discussion below is organized
into the following sections:   (1) Information Gathering,
(2) Emission Data Analysis,  (3)  Control Device Design and Cost
Analysis, and (4) Economic Impact Analysis.
1.1  INFORMATION GATHERING
     Section 114 information request responses were used to
establish the current level of HAP emissions for both industries.
Often, telephone conversations and additional written
correspondence were needed in addition to the completed
Section 114 responses to establish emissipn levels.  All such
pieces of correspondence are contained in the project docket,
which is designated as docket No. A-92-37.  In addition,
references not readily available to the public  (e.g., drafts,
                               1-1

-------
memoranda, and unpublished books) are also included in the
docket.  Published materials are referenced at the end of each
BID chapter.
1.2  EMISSION DATA ANALYSIS
     Emission stream characteristics of the various HAP emission
events that occur in the two industries were established using
data from the Section 114's and from making general assumptions
based on vapor-liquid equilibrium.  In some cases, the
information reported in the Section 114's contained discrepancies
between the total HAP emission rate reported and the
characteristics of the emission streams, such as flow rate, HAP
concentration, and duration.  In such situations, the emission
stream characteristics were used to establish the HAP emission
rates.  In many instances,  uncontrolled HAP emissions were not
reported.  The stream characteristics and total emission rate
were in these cases estimated by assuming that the streams were
saturated with HAP's at the conditions reported.  Hazardous air
pollutant emissions from storage tank working and breathing
losses were estimated using AP-42 emission estimation equations.
Emissions from equipment leaks were estimated using Synthetic
Organic Chemical Manufacturing Industry (SOCMI) emission factors.
     Finally, a source of data that frequently was unavailable at
WSR facilities was the exact number of fugitive components per
affected unit  (i.e., valves, pumps, and flanges) and the duration
of contact with HAP's and the weight percent of HAP's at the time
of contact.  When these counts and the associated information
were not furnished by a facility in the Section 114 responses, a
model equipment count scenario was assumed.
1.3  CONTROL DEVICE DESIGN AND COST ANALYSIS
     The following HAP control techniques were designed and
costed out for either individual BLR facility emission sources or
WSR model emission sources:   (1) condensation systems,
(2) thermal incineration systems with caustic scrubbers for acid
gas emissions,  (3) water scrubbers, (4) steam strippers for
wastewater,  (5) floating roofs for storage tanks, and  (6) leak
detection and repair  (LDAR) programs to control HAP emissions
                               1-2

-------
from equipment leaks.  The first three techniques were designed
and costed out according to the methodology contained in the
OAQPS Cost Manual, 4th edition.  Steam strippers were costed out
using an OAQPS cost algorithm.  The installation of floating
roofs was evaluated using information provided in the Draft
Volatile Organic Liquid Storage CTG (May 1992) for cleaning,
degassing, and installation.  Various levels of LDAR programs
were evaluated for cost and emissions reductions using memoranda
from the Equipment Leaks Negotiated Regulation (Reg Neg)
background file.
1.4  ECONOMIC IMPACT ANALYSIS
     The regulatory alternatives that are presented in BID
Chapter 6 were evaluated for economic feasibility by dividing the
annualized cost of control by the emission reductions achieved
from baseline.  A preliminary estimate of nationwide impacts of
implementing these regulatory alternatives is also presented in
BID Chapter 6 in the "Total" rows in Tables 6-5 and 6-6.   A
detailed cost analysis is contained in BID Chapter 8, Costs.
                               1-3

-------
                        2.0  INTRODUCTION

2.1  BACKGROUND AND AUTHORITY FOR STANDARDS
     According to industry estimates, more than 2.4 billion
pounds of toxic pollutants were emitted to the atmosphere in 1988
("Implementation Strategy for the Clean Air Act Amendments of
1990," EPA Office of Air and Radiation, January 15, 1991).  These
emissions may result in a variety of adverse health effects,
including cancer, reproductive effects, birth defects, and
respiratory illnesses.  Title III of the Clean Air Act Amendments
of 1990 provides the tools for controlling emissions of these
pollutants.  Emissions from both large and small facilities that
contribute to air toxics problems in urban and other areas will
be regulated.  The primary consideration in establishing national
industry standards must be demonstrated technology.  Before
national emission standards for hazardous air pollutants  (NESHAP)
are proposed as Federal regulations, air pollution prevention and
control methods are examined in detail with respect to their
feasibility, environmental impacts, and costs.  Various control
options based on different technologies and degrees of efficiency
are examined, and a determination is made regarding whether the
various control options apply to each emissions source or if
dissimilarities exist between the sources.  In most cases,
regulatory alternatives are subsequently developed that are then
studied by EPA as a prospective basis for a standard.  The
alternatives are investigated in terms of their impacts on the
environment, the economics and well-being of the industry, the.
national economy, and energy and other impacts.  This document
summarizes the information obtained through these studies so that
                               2-1

-------
interested persons will be able to evaluate the information
considered by EPA in developing the proposed standards.
     National emission standards for hazardous air pollutants for
new and existing sources are established under Section 112 of the
Clean Air Act as amended in 1990 [42 U.S.C. 7401 et seq., as
amended by PL 101-549, November 15, 1990], hereafter referred to
as the Act.  Section 112 directs the EPA Administrator to
promulgate standards that "require the maximum degree of
reduction in emissions of the hazardous air pollutants subject to
this section (including a prohibition of such emissions, where
achievable) that the Administrator, taking into consideration the
cost of achieving such emission reductions, and any non-air
quality health and environmental impacts and energy requirements,
determines is achievable ... ."  The Act allows the Administrator
to set standards that "distinguish among classes, types, and
sizes of sources within a category or subcategory."
     The Act differentiates between major sources and area
sources.  A major source is defined as "any stationary source or
group of stationary sources located within a contiguous area and
under common control that emits or has the potential to emit
considering controls, in the aggregate, 10 tons per year or more
of any hazardous air pollutant or 25 tons per year or more of any
combination of hazardous air pollutants."  The Administrator,
however, may establish a lesser quantity cutoff to distinguish
between major and area sources.  The level of the cutoff is based
on the potency, persistence, or other characteristics or factors
of the air pollutant.  An area source is defined as "any
stationary source of hazardous air pollutants that is not a major
source."  For new sources, the amendments state that the "maximum
degree of reduction in emissions that is deemed achievable for
new sources in a category or subcategory shall not be less
stringent than the emission control that is achieved in practice
by the best controlled similar source, as determined by the
Administrator."  Emission standards for existing sources "may'be
less stringent than the standards for new sources in the same
                               2-2

-------
category or subcategory but shall not be less stringent, and may
be more stringent than--
      (A) the average emission limitation achieved by the best
performing 12 percent of the existing sources  (for which the
Administrator has emissions information), excluding those sources
that have, within 18 months before the emission standard is
proposed or within 30 months before such standard is promulgated,
whichever is later, first achieved a level of emission rate or
emission reduction which complies, or would comply if the source
is not subject to such standard, with the lowest achievable
emission rate  (as defined by Section 171) applicable to the
source category and prevailing at the time, in the category or
subcategory for categories and subcategories with 30 or more
sources, or
      (B) the average emission limitation achieved by the best
performing five sources  (for which the Administrator has or could
reasonably obtain emissions information) in the category or
subcategory for categories or subcategories with fewer than
30 sources."
     The Federal standards are also known as "MACT" standards and
are based on the maximum achievable control technology previously
discussed.  The MACT standards may apply to both major and area
sources, although the existing source standards may be less
stringent than the new source standards, within the constraints
presented above.  The MACT is considered to be the basis for the
standard, but the Administrator may promulgate more stringent
standards which have several advantages.  First, they may help
achieve long-term cost savings by avoiding the need for more
expensive retrofitting to meet possible future residual risk
standards, which may be more stringent  (discussed in
Section 2.6).  Second, Congress was clearly interested in
providing incentives for improving technology'.  Finally, in the
Clean Air Act Amendments of 1990, Congress gave EPA a clear..
mandate to reduce the health and environmental risk of air toxics
emissions as quickly as possible.
                               2-3

-------
     For area sources, the Administrator may "elect to promulgate
standards or requirements applicable to sources in such
categories or subcategories which provide for the use of
generally available control technologies or management practices
by such sources to reduce emissions of hazardous air pollutants."
These area source standards are also known as "GACT" (generally
available control technology)  standards, although MACT may be
applied at the Administrator's discretion, as discussed
previously.
     The standards for hazardous air pollutants (HAP's), like the
new source performance standards (NSPS) for criteria pollutants
required by Section 111 of the Act- (42 U.S.C. 7411), differ from
other regulatory programs required by the Act (such as the new
source review program and the prevention of significant
deterioration program) in that NESHAP and NSPS are national in
scope (versus site-specific).   Congress intended for the NESHAP
and NSPS programs to provide a degree of uniformity to State
regulations to avoid situations where some States may attract
industries by relaxing standards relative to other States.
States are free under Section 116 of the Act to establish
standards more stringent than Section 111 or 112 standards.
     Although NESHAP are normally structured in terms of
numerical emissions limits, alternative approaches are sometimes
necessary.  In some cases, physically measuring emissions from a
source may be impossible or at least impracticable due to
technological and economic limitations.  Section 112(h) of the
Act allows the Administrator to promulgate a design, equipment,
work practice, or operational standard, or combination thereof,
in those cases where it is not feasible to prescribe or enforce
an emissions standard.  For example, emissions of volatile
organic compounds (many of which may be HAP's, such as benzene)
from storage vessels for volatile organic liquids are 'greatest
during tank filling.  The nature of the emissions  (i.e, high
concentrations for short periods during filling and low
concentrations for longer periods during storage) and the
configuration of storage tanks make direct emission measurement
                               2-4

-------
impractical.  Therefore, the MACT or GACT standards may be based
on equipment specifications.
     Under Section 112(h)(3), the Act also allows the use of
alternative equivalent technological systems:  "If, after notice
and opportunity for comment, the owner or operator of any source
establishes to the satisfaction of the Administrator that an
alternative means of emission limitation" will reduce emissions
of any air pollutant at least as much as would be achieved under
the design, equipment, work practice, or operational standard,
the Administrator shall permit the use of the alternative means.
     Efforts to achieve early environmental benefits are
encouraged in Title III.  For example, source owners and
operators are encouraged to use the Section 112(i)(5) provisions,
which allow a 6-year compliance extension of the MACT standard in
exchange for the implementation of an early emission reduction
program.  The owner or operator of an existing source must
demonstrate a 90 percent emission reduction of HAP's  (or
95 percent if the HAP's are particulates) and meet an alternative
emission limitation, established by permit, in lieu of the
otherwise applicable MACT standard.  This alternative limitation
must reflect the 90 (95) percent reduction and is in effect for a
period of 6 years from the compliance date for the otherwise
applicable standard.  The 90 (95) percent early emission
reduction must be achieved before the otherwise applicable
standard is first proposed, although the reduction may be
achieved after the standard's proposal (but before January 1,
1994)  if the source owner or operator makes an enforceable
commitment before the proposal of the standard to achieve the
reduction.  The source must meet several criteria to qualify for
the early reduction standard, and Section 112(i)(5)(A) provides
that the State may require additional reductions.
2.2  SELECTION OF POLLUTANTS AND SOURCE CATEGORIES
     As amended in 1990, the Act includes a list of 189 HAP's.
Petitions to add or delete pollutants from this list may be
submitted to EPA.   Using this list of pollutants,  EPA will
publish a list of source categories (major and area sources) for
                               2-5

-------
which emission standards will be developed.  Within 2 years of
enactment  (November 1992),  EPA will publish a schedule
establishing dates for promulgating these standards.  Petitions
may also be submitted to EPA to remove source categories from the
list.  The schedule for standards for source categories will be
determined according to the following criteria:
     "(A) the known or anticipated adverse effects of such
pollutants on public health and the environment;
     (B) the quantity and location of emissions or reasonably
anticipated emissions of hazardous air pollutants that each
category or subcategory will emit; and
     (C) the efficiency of grouping categories or subcategories
according to the pollutants emitted, or the processes or
technologies used."
     After the source category has been chosen, the types of
facilities within the source category to which the standard will
apply must be determined.  A source category may have several
facilities that cause air pollution, and emissions from these
facilities may vary in magnitude and control cost.  Economic
studies of the source category and applicable control technology
may show that air pollution control is better served by applying
standards to the more severe pollution sources.  For this reason,
and because there is no adequately demonstrated system for
controlling emissions from certain facilities, standards often do
not apply to all facilities at a source.  For the same reasons,
the standards may not apply to all air pollutants emitted.  Thus,
although a source category may be selected to be covered by
standards, the standards may not cover all pollutants or
facilities within that source category.
2,3  PROCEDURE FOR DEVELOPMENT OF NESHAP
     Standards for major and area sources must  (1) realistically
reflect MACT or GACT; (2) adequately consider the cost, the non-
air quality health and environmental impacts, and the energy
requirements of such control; (3) apply to new and existing
sources; and (4) meet these conditions for all variations of
industry operating conditions anywhere in the country.
                               2-6

-------
     The objective of the NESHAP program is to develop standards
to protect the public health by requiring facilities to control
emissions to the level achievable according to the MACT or GACT
guidelines.  The standard-setting process involves three
principal phases of activity:   (1) gathering information,
(2) analyzing the information, and (3) developing the standards.
     During the information-gathering phase, industries are
questioned through telephone surveys, letters of inquiry, and
plant visits by EPA representatives.   Information is also
gathered from other sources, such as a literature search.  Based
on the information acquired about the industry, EPA selects
certain plants at which emissions tests are conducted to provide
reliable data that characterize the HAP emissions from well-
controlled existing facilities.
     In the second phase of a project, the information about the
industry, the pollutants emitted, and the control options are
used in analytical studies.  Hypothetical "model plants" are
defined to provide a common basis for analysis.  The model plant
definitions, national pollutant emissions data, and existing
State regulations governing emissions from the source category
are then used to establish "regulatory alternatives."  These
regulatory alternatives may be different levels of emissions
control or different degrees of applicability or both.
     The EPA conducts studies to determine the cost, economic,
environmental, and energy impacts of each regulatory alternative.
From several alternatives, EPA selects the single most plausible
regulatory alternative as the basis for the NESHAP for the source
category under study.
     In the third phase of a project, the selected regulatory
alternative is translated into standards, which, in turn, are
written in the form of a Federal regulation.  The Federal
regulation limits emissions to the levels indicated in the
selected regulatory alternative.
     As early as is practical in each standard-setting project,
EPA representatives discuss the possibilities of a standard and
the form it might take with members of the National Air Pollution
                               2-7

-------
Control Techniques Advisory Committee, which is composed of
representatives from industry, environmental groups, and State
and local air pollution control agencies.  Other interested
parties also participate in these meetings.
     The information acquired in the project is summarized in the
background information document (BID).  The BID, the proposed
standards, and a preamble explaining the standards are widely
circulated to the industry being considered for control,
environmental groups, other government agencies, and offices
within EPA.  Through this extensive review process, the points of
view of expert reviewers are taken into consideration as changes
are made to the documentation.
     A "proposal package" is assembled and sent through the
offices of EPA Assistant Administrators for concurrence before
the proposed standards are officially endorsed by the EPA
Administrator.  After being approved by the EPA Administrator,
the preamble and the proposed regulation are published in the
Federal Register.
     The public is invited to participate in the standard-setting
process as part of the Federal Register announcement of the
proposed regulation.  The EPA invites written comments on the
proposal and also holds a public hearing to discuss the proposed
standards with interested parties.  All public comments are
summarized and incorporated into a second volume of the BID.  All
information reviewed and generated in studies in support of the
standards is available to the public in a "docket" on file in
Washington, D.C.  Comments from the public are evaluated, and the
standards may be altered in response to the comments.
     The significant comments and EPA's position on the issues
raised are included in the preamble of a promulgation package,
which also contains the draft of the final regulation.  The
regulation is then subjected to another round of internal EPA
review and refinement until it is approved by the EPA
Administrator.  After the Administrator signs the regulation, it
is published as a "final rule" in the Federal Register.
                               2-8

-------
2.4  CONSIDERATION OF COSTS
     The requirements and guidelines for the economic analysis of
proposed NESHAP are prescribed by Presidential Executive
Order 12291  (EO 12291) and the Regulatory Flexibility Act  (RFA).
The EO 12291 requires preparation of a Regulatory Impact
Analysis (RIA) for all "major" economic impacts.  An economic
impact is considered to be major if it satisfies any of the
following criteria:
     1.  An annual effect on the economy of $100 million or more;
     2.  A major increase in costs or prices for consumers;
individual industries; Federal, State, or local government
agencies; or geographic regions; or
     3.  Significant adverse effects on competition, employment,
investment, productivity, innovation, or on the ability of United
States-based enterprises to compete with foreign-based
enterprises in domestic or export markets.
     An RIA describes the potential benefits and costs of the
proposed regulation and explores alternative regulatory and
nonregulatory approaches to achieving the desired objectives.  If
the analysis identifies less costly alternatives, the  RIA
includes an explanation of the legal reasons why the less costly
alternatives could not be adopted.  In addition to requiring an
analysis of the potential costs and benefits, EO 12291 specifies
that EPA, to the extent allowed by the CAA and court orders,
demonstrate that the benefits of the proposed standards outweigh
the costs and that the net benefits are maximized.
     The RFA requires Federal agencies to give special
consideration to the impact of regulations on small businesses,
small organizations, and small governmental units.  If the
proposed regulation is expected to have a significant impact on a
substantial number of small entities, a regulatory flexibility
analysis must be prepared.  In preparing this analysis, EPA takes
into consideration such factors as the availability of capital
for small entities, possible closures among small entities, the
increase in production costs due to compliance, and a comparison
                               2-9

-------
of the relative compliance costs as a percent of sales for small
versus large entities.
     The prime objective of the cost analysis is to identify the
incremental economic impacts associated with compliance with the
standards based on each regulatory alternative compared to
baseline.  Other environmental regulatory costs may be factored
into the analysis wherever appropriate.  Air pollutant emissions
may cause water pollution problems, and captured potential air
pollutants may pose a solid waste disposal problem.  The total
environmental impact of an emission source must, therefore, be
analyzed and the costs determined whenever possible.
     A thorough study of the profitability and price-setting
mechanisms of the industry is essential to the analysis so that
an accurate estimate of potential adverse economic impacts can be
made for proposed standards.  It is also essential to know the
capital requirements for pollution control systems already placed
on plants so that the additional capital requirements
necessitated by these Federal standards can be placed in proper
perspective.  Finally, it is necessary to assess the availability
of capital to provide the additional control equipment needed to
meet the standards.
2.5  CONSIDERATION OF ENVIRONMENTAL IMPACTS
     Section 102(2)(C) of the National Environmental Policy Act
(NEPA) of 1969 requires Federal agencies to prepare detailed
environmental impact statements on proposals for legislation and
other major Federal actions significantly affecting the quality
of the human environment.  The objective of NEPA is to build into
the decision-making process of Federal agencies a careful
consideration of all environmental aspects of proposed actions.
     In a number of legal challenges to standards for various
industries, the United States Court of Appeals for the District
of Columbia Circuit has held that environmental impact statements
need not be prepared by EPA for proposed actions under the Clean
Air Act.  Essentially, the Court of Appeals has determined that
the best system of emissions reduction requires the Administrator
to take into account counterproductive environmental effects of
                               2-10

-------
proposed standards as well as economic costs to the industry.  On
this basis, therefore, the Courts established a narrow exemption
from NEPA for EPA determinations.
     In addition to these judicial determinations, the Energy
Supply and Environmental Coordination Act  (ESECA) of 1974
(PL-93-319) specifically exempted proposed actions under the
Clean Air Act from NEPA requirements.  According to
Section 7(c)(l), "No action taken under the Clean Air Act shall
be deemed a major Federal action significantly affecting the
quality of the human environment within the meaning of the
National Environmental Policy Act of 1969"  (15 U.S.C. 793(c)(l)).
     Nevertheless, EPA has concluded that preparing environmental
impact statements could have beneficial effects on certain
regulatory actions.  Consequently, although not legally required
to do so by Section 102(2)(C) of NEPA, EPA has adopted a policy
requiring that environmental impact statements be prepared for
various regulatory actions,  including NESHAP developed under
Section 112 of the Act.  This voluntary preparation of
environmental impact statements, however, in no way legally
subjects the EPA to NEPA requirements.
     To implement this policy, a separate section is included in
this document that is devoted solely to an analysis of the
potential environmental impacts associated with the proposed
standards.  Both adverse and beneficial impacts in such areas as
air and water pollution, increased solid waste disposal, and
increased energy consumption are discussed.
2.6  RESIDUAL RISK STANDARDS
     Section 112 of the Act provides that 8 years after MACT
standards are established (except for those standards established
2 years after enactment, which have 9 years), standards to
protect against the residual health and environmental risks
remaining must be promulgated, if necessary.  The standards would
be triggered if more than one source in a category or subcategory
exceeds a maximum individual risk of cancer of 1 in 1 million.
These residual risk regulations would be based on the concept of
providing an "ample margin of safety to protect public health."
                              2-11

-------
The Administrator may also consider whether a more stringent
standard is necessary to" prevent--considering costs, energy,
safety, and other relevant factors--an adverse environmental
effect.  In the case of area sources controlled under GACT
standards,  the Administrator is not required to conduct a
residual risk review.
                               2-12

-------
                    3.0  INDUSTRY DESCRIPTION

     This chapter presents a description of a segment of the
polymers and resins industry.  Section 3.1 provides a general
description of basic liquid epoxy resins (BLR) and wet strength
resins  (WSR).  Section 3.2 presents source category descriptions.
Section 3.3 presents baseline hazardous air pollutant  (HAP)
emissions for each source category.  Finally, Section 3.4 lists
the references for Chapter 3.
3.1  GENERAL
     The scope of source categories to be covered by this NESHAP
are defined as the manufacture of BLR  (diglycidyl ether of
bisphenol-A [DGEBPA]) and WSR (EPI modified non-nylon
polyamides).  Basic liquid epoxy resins does not include
specialty epoxy resins (epoxy resins that are not BLR) or
modified epoxy resins (BLR resins that have been blended with
solvents, reactive diluents, or other resins).  Wet strength
resins include those resins that are made with dibasic esters,
dicarboxylic acids, amines, and epichlorohydrin  (EPI).
     In general, BLR are plastic materials that become hard,
infusible solids upon the addition of a hardening agent.1  These
resins are characterized by the presence of a three-member cyclic
ether group commonly referred to as an epoxy group.  Usually the
epoxy groups are situated at the ends of the molecule, but they
can be attached at other points, as in the case of specialty
epoxy resins.   To become useful, epoxy resins must be cross-
linked with amines, anhydrides,  or other curing agents to form
thermoset,  three-dimensional structures.  The epoxy groups
provide the points for the curing reaction to take place.  Upon
addition of the curing agent,  the epoxy groups open up to begin a
                               3-1

-------
reaction that permanently cross-links .the resin to the curing
agent and to itself.  It is the cross-linked structure that gives
the epoxy resins desirable properties such as high chemical and
corrosion resistance, good mechanical and strength properties,
outstanding adhesion, low shrinkage upon curing, and
flexibility.2  These properties make the cured resin a desirable
ingredient in adhesives, coatings, and other plastics
applications.
       Epichlorohydrin modified non-nylon polyamide resins  (WSR)
are used primarily by the paper industry as an additive to
increase the tensile strength of paper products.  'Natural
polymers such as cellulose and protein are not highly
crosslinked; their fibers, which are composed of hydrophylic
polymer chains, can change position or become completely
separated upon the addition of water.3'4  The added resins are
used to form a stable polymeric web among the natural fibers.
For this reason, these resins are commonly referred to as WSR.
       Production methods used in the two source categories
include both batch and continuous operations.  The sizes of the
facilities range from those that make several megagrams of resin
per year (Mg/yr) to those that produce over 50 thousand Mg/yr.
The HAP emissions associated with the production processes
include EPI, methanol, and hydrochloric acid (HC1).  This chapter
presents source category, product, and process descriptions,
lists of facilities, descriptions of potential emission sources,
and current levels of emissions in each source category.
3.2  SOURCE CATEGORY DESCRIPTIONS
3.2.1  Source Category 1:  BLR Production
     Diglycidyl ether of bisphenol-A can be made from excess EPI
and bisphenol-A in two different processes, the "two-step
process" and the "conventional process."  Current domestic
production is believed to utilize the two-step process, which is
described below.
     3.2.1.1  Two-Step Process.^  This process involves two major
steps, coupling and dehydrochlorination.  In the coupling step,
                               3-2

-------
EPI is added to bisphenol-A to form dichlorohydrin ether, an
intermediate, as shown in the following reaction:

  Cl     O                        Cl  OH              OH   Cl
   I    / \                       II                II
  2CH2CH — CH2 + HO — X — OH - ^H2C — CHCH2O — X — OCH2CH— CH2
     The reaction is catalyzed with either caustic soda or a
compound such as methyltributyl ammonium chloride,
methyltriphenyl phosphonium bromide, or trimethylsulfonium
iodide.
     The dichlorohydrin intermediate formed in the coupling
reaction is dehydrochlorinated in the next step with caustic soda
to produce BLR in the following reaction:

   Cl  OH              OH   Cl
    II                II.
 H2C — CHCH2° — x — OCH2CH — CH2 + 2NaOH - *•
              O                    O
             / \                  / \
          H2C — CHCH2O — X — OCH2CH — CH2  +  2NaCl  + 2H2O

Sodium chloride (NaCl) ,  water, and BLR are the final reaction
products.
     Excess EPI is removed either before or after the
dehydrochlorination step.  If EPI is removed prior to
dehydrochlorination,  it is distilled off in two stages.  This
two- stage process is shown in Figure 3-1, option A.  After the
first distillation, glycerol dichlorohydrin is added to the
reaction mixture.   After the second distillation, a solvent
mixture of methyl isobutyl ketone and isopropanol is added to the
mixture and the solution is reacted with a 15 percent caustic
soda solution.  The resulting organic solution is purified by
neutralizing with a dilute acid solution and by washing with
water.  The solvent is removed by distillation.  The liquid epoxy
resin is filtered to remove any entrained organic salt.
     If excess EPI is left in the solution during
dehydrochlorination,  as shown in Figure 3-1, option B, the amount
of caustic solution added is carefully controlled to limit

                               3-3

-------
                  ffl
                  CO
                  '
       0)
       Q.
       Q.
03
i














c
.2
i
"5
Q

"<
* m
O
fe
O
9
OC
cf
•Sim
"— i ^M
Q.L
0


z 55
o ~
o S.

o
•5
S
DC



IO
S *





in
1 """"





„.
= ,
5

c
ID
o te
* 8.
9 .9-
£ » <
« E 2
£ 5 c •» 	
^g S •
< o



^^

c <
O CM
S. S 	 	
o »





CB
O
Q.
I
C
1
S

*



O)
i
.



0
S
DC


i i

o c
IS —
c «
CM Q
4




0> 0
CD >!<*
« 5
** «P
W Q)
o









i
Q.
h.
w
2 E
S|
Z n
o 1
CM % C.
-H-1 «'S
3 2
* 5 OT
o
i«
Offi
^s

z
o
in
*— x«
in

•§
if
i«
o 5



•










                             CD






































4
c
0
I

























_o
i
o
_o
£
u
o
i
•g
•o
£
o
i
^
^

S
&
CO


111
1























o
1
0
u
o
•S
>.
i
f
h»
•o
g
3
0
CO
c
i
IE
a «
c 2
.2 u
o!1^1
m
CQ
0)
o
O
a
0)
m
6
4J
rt
rt
•H
TJ
»
O
rH
m
CO
0)
u
o
a
c
•H
m
0)
*""
6-
o
ft
[VI
H
I
n
0)
3
0")
•H







                                                     3-4

-------
hydrolysis of EPI.  After the dehydrochlorination reaction takes
place, water and EPI are removed from the reaction medium as an
azeotropic mixture.  After condensation, this overhead mixture is
separated into a water layer and an EPI layer by decantation.
After distillation, the organic product is dissolved in an
organic solvent such as toluene, xylene, or methyl isobutyl
ketone (MIBK).  The solution is further purified in the same
method as described above for option A.
     3.2.1.2  Product Specifications.  The effect of mole ratio
of EPI to bisphenol-A on the average molecular weight of the
liquid resin and on the average number of repeating units  (n) is
shown below:6'7
EPI : bisphenol-A
2:1
2:1
1.4:1
1.1:1
Molecular weight
370
450
791
3,750
n
0.1
0.4
2
12
Typically, commercial BLR has an average molecular weight of 370
and consists of about 87 percent of the polymer molecules with
n = 0, 11 percent of the polymer molecules with n = 1, and
1.5 percent of the polymer molecules with n = 2.^  The molecular
weight and chain length have an effect on the physical state of
the resin.  Pure BLR (n = 0) is a solid with a melting point of
43°C  (109°F).9  The higher the molecular weight, the higher the
melting point and the more solid the resin becomes.  Commercial
BLR is a super-cooled liquid with the potential for
crystallization depending on purity and storage conditions.  By
warming, the crystallized resin is restored to its original
form.1^  it does not need to be dissolved in solvent for it to be
liquid, although it is commonly blended with solvents or diluents
to reduce the viscosity before use.
     Basic liquid epoxy resins are widely used in casting,
tooling, adhesives, and coatings applications.  They are just one
                               3-5

-------
ingredient in a formulation that includes curing agents that
cross-link the epoxy resin, diluents or solvents that reduce the
viscosity, inexpensive fillers that reduce the formulation cost,
plasticizers that make the resin more flexible, and other resins
that alter the properties of the finished resin formulation.
Diglycidyl ether of bisphenol-A is used to make higher molecular
weight epoxy resins to be used as ingredients in powder coatings,
resin solutions, or resin dispersions.  It is also used to make
numerous types of modified resins by chemically attaching it to
other resins such as phenolic, urea, melamine, furane, polyester,
vinyl, polyurethane, and silicone; or by reacting it with other
organic compounds such as oils or fatty acids to make epoxy
esters or other useful polymers.11
     3.2.1.3  Emission Sources in Source Category 1.  As
mentioned in Section 3.2.1.1, EPI may be removed either before or
after the dehydrochlorination step.  In either case, the
manufacture of BLR is initiated in a reactor where EPI emissions
occur during EPI charging and reaction.  The dichlorohydrin
intermediate formed in this reaction is dehydrochlorinated in the
next step.  If excess EPI is removed prior to dehydrochlorination
(option A) emissions occur from the two distillation stages.  If
the excess EPI is left in solution  (option B) a
dehydrochlorination reaction takes place.  Epichlorohydrin and
water vapors are removed from the reactor, condensed, and sent to
a decanter.  In this step, potential emissions of EPI are from
noncondensed EPI vapors.  Further reaction, washing and
distillation steps occur with potential emissions of EPI and
solvent.  Currently, the facilities in the source category do not
report the usage of other HAP solvents (such as MIBK) in these
steps.
     Working and breathing losses are noted for storage tanks,
and equipment leaks are potential fugitive HAP emission sources.
Finally, wastewater is generated from the steam stripper bottoms
and from vacuum seals; this water is routed to the wastewater
treatment system, which contributes to emissions of EPI from
volatilization.
                               3-6

-------
     3.2.1.4  Facilities Included in Source Category 1.
Facilities that are included in this source category have
production rates on the order of 45 gigagrams per year  (Gg/yr)
(100 million Ib/yr).   As of 1990, only three U. S. companies
produced BLR on a large scale:  Ciba-Geigy Corporation in
Mclntosh, Alabama; Dow Chemical U.S.A. in Freeport, Texas; and
Shell Chemical Company in Deer Park, Texas.  All three of these
facilities are considered major sources according to the EPA
criterion of having the potential to emit 10 tons per year of any
one HAP or 25 tons per year of combined HAP's.  As of
January 1987, total U. S. production capacity of liquid BLR epoxy
resin was approximately 220 Gg/yr (4.8 x 108 lb/yr).12  Table 3-1
presents the list of facilities that manufacture BLR and provides
information on HAP emissions related to this source category.
The emission estimates are discussed in Section 3.3.
     3.2.1.5  End-Product Uses and Future Growth.  Diglycidyl
ether of bisphenol-A is used to make higher molecular weight
epoxy resins to be used as ingredients in powder coatings, resin
solutions, or resin dispersions.  It is also used to make
numerous types of modified resins by chemically attaching it to
other resins such as phenolics, urea, melamine, furane,
polyester, vinyl, polyurethane, and silicone.  One 1987 reference
indicates that approximately 75 percent of the epoxy resins
currently used worldwide are derived from BLR.13  Table 3-2
shows consumption patterns for BLR.
     In recent years, the market trends have increasingly been
toward the growth of specialty epoxy resins, many of which are
not derived from BLR.  Overall growth for epoxy resins is
estimated at approximately 2 to 3 percent through 1995.  The
coatings segment, of which BLR is a part, has a projected growth
rate of l to 2 percent through this period.14
3.2.2  Source Category 2;  WSR Production
     3.2.2.1  Product Description.  Wet strength resins can be
categorized into two groups:  (1) dimer acid-based and (2) EPI
cross-linked.  The second group of resins is identical to the
first group except that the resins in the second group are
                               3-7

-------




CJ
H
S
PQ
W
ry*
g
CJ
En
p
^t
g
H
B
CO
W
H
H PH
,J J
H CQ
CJ —
En CO
O CO
En En1
Ct!
H H
E-" D
rf CX
§ j
O
Cn
CO
2^
0
H
CO
CO
H
S
E3



H
t
rO
Efl
J
9
^H


I-S-
S ~
•o e
li
I §

03 C
1 U
g 0.
"« 5?
E •§

s
jr^ O
-P 'I u
S ^ -^
|| £
5 I
"8 •« S" s
•e > c §
o 2 .S a
^ JJ j£J d<
0)


4>
y
'>
$
•o
S
O
U
Ui
g -si
111

Il|












Emission sources



tQ U] Uj U] U] U]
° § m § § §
---» a




S!S;S;S;S! S



§g
-
*•* v* ^


w> ON oo oo t-^
ON ON ON ON ON




§ g *3
* * •= -a .2
a> i> P* P" "£ £j
j|IIJI|

1 1 § 1 1 1
O\ N -^ I""-" 3



09
V

B
«
g
 « 5 «
c us .5 g S
•S S || fe «
'1 £ § « S
» '§ S §• o J g.
8 ex OT > 55 £ rt
ft. -N cd
Jl
1 §
Q *
NN M< NH MM &
& w u S Q
» ^ V § °.
-T f-



I I
S E £ S
O '** **3 ^
8 S A
^ S
2 Z

§S ^* 1
*
*™^


° g g i


fi f-a
Q 09 t> 09
.£, "- .£, "
<2- ^ w ^ ^-
S3ea32c'g+.2
1 lsi-1 6 1-1 ^ tl g
||1 §|1 2|£J D

11 v 1 1
ON GO







•o
"S
eg
*5

S


ll s «
> _ « u
1 S s1 I 1
ll
? $
% Q
SI |g S S
= 8 ~" m » =.
- 1 s




O O m oo
ON ON Ox fl"








£ g


fe fe
c c
•O T) O
§ g „.!
•i -s 11
? ^ l| «
^2 se ;l ^
^ js ^ ^ c e Q
Q£ PH p-< tt< !> >--1

§ § § §
xo" **C oC oo"
t** xO











T3 C oo


*H JS ^
c ^s bo .2 «
w W ft C -O g
I 2 «l| t
S 2 B -2 § 5 T
S AH W ? i "9 W
iS
JS 1
c *
                                                 o
                                                ^v


                                                J=
                                                fS

                                                •o

                                                S
                                                o
                                                T3


                                                 i1
                                                1
                                                 Cu
                                                 S
                                                 XI
                                                 JS


                                                 g

                                                 fe
                                                 03
                                                 Q.
                                                 U
3-8

-------
              TABLE 3-2.  BLR CONSUMPTION PATTERNS21
Market
Bonding and adhesives
Flooring, paving, aggregates
Protective coatings
Reinforced uses
Electrical laminates
Other
Tooling, casting, molding
Export
Other
Total
Million Ib
1989
25
25
193
57
26
30
86
41
483
1990
28
26
195
55
31
28
68
33
464
aFrom Modern Plastics, January 1990. p. 114.

EPI-modified.  This NESHAP considers only the manufacture of the
EPI modified or "cross-linked" resin.  The EPI cross-linked
resins have been referred to as WSR, primarily because these
resins are used to form a stable polymeric web among cellulosic
or protein fibers.  This web provides a permanent fiber
organization without modifying the desired characteristics of the
fiber and, in doing so, imparts tensile strength to paper
products.  One of the more important resins in this group is the
EPI-modified diethylene triamine (BETA) hexanedioic acid polymer.
This amino polyamide polymer can be formed from BETA and
hexanedioic acid  (i.e., adipic acid).  In some cases, dibasic
ester is substituted for adipic acid.  While the polymer formed
is the same, methanol, instead of water, is formed as a
byproduct.  Both manufacturing routes for EPI-modified diethylene
triamine  (DETA) hexanediodic acid polymer are shown below:
Dibasic ester route
Adipic acid route
Addition of EPI forms the cross-linked structure in the following
way:
     3.2.2.2  Process Description.   Figure 3-2 shows the WSR
production process.  For the production of the WSR, the initial
reaction is carried out with adipic acid, or dibasic ester and
                               3-9

-------
                                  o      o
                                  II      II
       NH2—R—NH—R—NH2 + CH3 —C—R—C—CH3 -
           DETA (R:  (CH)2)         DIBASIC ESTER
                                00-
                                II       II
            — R — NH — R — NH — C — R — C — NH—
                                               + CH3OH
                                     O       O
                                     II       II
          NH2—R— NH— R— KH2+OH— C— R— C— OH

             EETA (R: (CH)2)       ADIPIC ACID (R: (OH2)4)

                          O
                                 O
                        II         II
— R —NH — R. — NH —  C — R — C — NH —
                                                 H20
    R
       NH
o        o
II        II
C	 R	C
               NH
                   CH-
                   CH  - OH
                    CH-
     R
            NH
N
         R
O
II
C
NH
DETA.  A  second reaction then occurs with the addition of EPI.
The second  reaction between the polyamide and EPI forms EPI
adducts.
     Each of  the feedstocks is transferred from storage tanks  to
dedicated weigh tanks, where the required quantities are
measured.   Depending upon the capacity of the production line,
                                3-10

-------
•

?
o
a
o
o>
o
CD
O
v

w
a
>•
^
Q.


1 ,


_O
a
0
O
S

oc
g
u
<
III
(E
M
t '
                   Q.
                   UJ
       111
           <
                                           O
                                           W
                                           CM
                                      •S i  o

                                      S I*  £
                                      E S-- 2
                                      * 2 D. o
                                      * a iu <
reactor as the
          =  S
          gl
            ^
          s  §
          I*
          ll
          S  a
          jf
            01

          !i
          !!

    o
    CM
..XI

ila

ill
g * s

= 11
o o -o
                                                    ~ x. ui
                                                    •S < o
                                                    g u «
                                                    § a|
                                                    ^ s S
                                                    •o •= £
                                                    a c o
                                                    c o j.
                                                          Cn
                                                          rt
                                                          •H
                                                          -o


                                                          o
                                                          CO
                                                          CQ

                                                          0)
                                                          U

                                                          O
                                                          ^1

                                                          CX


                                                          C
                                                          •H

                                                          CQ
                                                          CU
                                                          ^


                                                          X!
                                                          4J


                                                          &
                                                          CU
                                                          tn
                                                          CN
      m

      0)
       CD
       •H
                                                  O  £
          CD p


          ll


        III
  !5 S

« : s
1. « t>

II s-
i> o o
° "5 •£
£5 S
a ^ TI
                            3-11

-------
the volumes of the weigh tanks can range from a few hundred
gallons to several thousand gallons.  The process of charging the
mixing tank or reactor is normally a gravity process because the
weigh tanks are usually located above the mixing tank or reactor.
In some instances, charging may occur by vacuum filling the
reactor or by pressurizing the weigh tanks with nitrogen to force
the feedstocks through piping into the receiving vessel.
Charging rates are typically about 190 liters per minute (1/min)
(50 gallons per minute [gal/min]).  As the reaction progresses,
the glycidyl groups begin crosslinking and the viscosity of the
reaction mixture increases.  Reaction times for each batch can
range from 2 to 24 hours and reaction temperatures can range from
40° to 70°C (104° to 158°F).  The reactors are generally equipped
with a temperature control jacket, an agitator, sampling ports
(manholes), and a pressure relief mechanism.  The pressure relief
system may be either a vent to the vapor recovery system or
pressure relief valves equipped with rupture disks.  The reactor
sizes used in this process usually range from 1,900 to 7,600 1
(500 to 2,000 gal).  At a predetermined viscosity, sulfuric acid
or hydrochloric acid is added to the reaction mixture to reduce
the pH of the reaction mixture and to stop formation of glycidyl
amine groups.   If no acid is added, the cross-linking will
continue until a gel is eventually formed.  After the reaction
has ceased due to the addition of acid, water and any excess acid
(as well as unreacted feedstocks)  are removed from the bottom of
the reactor.  If not used immediately, the WSR must be maintained
in an acidified solution or an increase in viscosity leading to
gel formation will result.  The commercial WSR are shipped as
acidified solutions that are reactivated by addition of base just
before use in the paper mill.
     3.2.2.3  Emission Sources in Source Category 2.  The
manufacture of WSR typically takes place in a single batch
reactor.  Emissions associated with the batch reactor are
charging, reaction heatup, and acid addition.  Epichlorohydrin
can potentially be emitted during each of these events, and HC1
(a HAP) can be emitted if it is used to halt the crosslinking
                               3-12

-------
reaction.  Emissions of methanol are possible if a dibasic ester
rather than a dibasic acid is reacted with BETA.  The HAP
emissions due to working and breathing losses occur from the
storage tanks and displacement emissions are noted for weigh
tanks dedicated to the manufacture of WSR.  Equipment leaks are
also a source of HAP emissions.
     3.2.2.4  Facilities Included in Source Category 2.
Table 3-3 presents a summary of emissions from facilities that
manufacture WSR.  There are at least 17 facilities in the nation
that manufacture WSR.  Of these facilities, nine are considered
major sources because they have the potential to emit more than
10 tons/yr of any one HAP or more than 25 tons/yr of combined
HAP's from within the fenceline of the entire facility.  In 1990,
an estimated 198 Gg/yr (4.5 x 108 Ib/yr) of resin was
                                                   rj
manufactured using approximately 14 Gg/yr  (3.1 x 10' Ib/yr) of
combined EPI and HC1.
3.3  BASELINE EMISSIONS
3.3.1  Source Category 1;  BLR Production
     Based on information reported from the three facilities that
make up Source Category 1 (BLR production), uncontrolled HAP
emissions (i.e., ignoring existing controls) constitute
approximately 700 Mg/yr  (1.54 x 106 Ib/yr).  The baseline HAP
emissions (representing the current level of control)  from the
three facilities are estimated to be 130 Mg/yr (2.8 x 105 Ib/yr),
or approximately 19 percent of the uncontrolled value.  Potential
wastewater emissions make up the largest fraction of uncontrolled
emissions, while equipment leak emissions make up the largest
fraction of controlled,  or baseline, HAP emissions.
     Uncontrolled emission estimates were calculated based on
data reported in Section 114 information requests.  Process vent
streams from the resin finishing and purification stages were
assumed to be saturated with EPI at the outlet conditions of the
vacuum manifolds prior to the control devices.  Because two of
the facilities currently use water scrubbers to control these
streams,  the overall control efficiency was estimated considering
the fate of EPI in the wastewater.  Both facilities that use
                              '3-13

-------

8
2
f£j
f..l
CO
EH
W
"*"
FFACTUF
K!
i

PH

CO
W
H
?OR FACILIT

n
H
EH
O
pLj
&
H
rfi

O
H
C/J
CO
S
w
m
ro

a
EH


Jjjf
l|i
•= 1 »
fl"g *




S
S
•o
s
c
a
c
§•8 i -
l^i-

B.
K


Emission sources
o, jT 5.
^ Si -C*
3

Si
<
X
e-
c -2 K
11*
£ ~
o.




'5

o
g
«
-1
BB
w w
o o

£§;






Carbon Bed
Carbon Bed

VO O4



1*3 W


PROCESS VENTS
1) EPI Charging
2) H2SO4 Addition

«



•iiL
ID


ce




	 ,
U
allaway Chem
olumbus, GA
U U
-

S
*
B
w










58



w


STORAGE
B
w










oo
8



W

£>
EQUIPMENT LEAKS






















1!
*tR
P r^





II
Water Scrub
Water Scrub

32



BB
U3 W


PROCESS VENTS
1) EPI Charging
2) H2SO4 Addition

•



B
H







	 ,
S
S> 7j
xf
O w
«

S

I

p;





t
Water Scrub

S


•p
§


STORAGE
S
\










R



w

Ji
EQUIPMENT LEAKS






















B B B
WWW
m >O O
VO (S -<
WlSJiR
r* oo r^






Condenser
Condenser
Condenser

*n vo oo



B B B
WWW


PROCESS VENTS
1) EPI Charging
2) Heatup
3) H2SO4 Charging

49



cu
w









o
ft
§ g
DU
m

S
I
127 EPI 1










S



Si


STORAGE
1,240 EPI ||










§



W

JO
EQUIPMENT LEAKS






















B B B
WWW
oo rr» **
*
ts •"• •*
S oo r~






Condenser
Condenser
Condenser

§00 —
o «•>



a. o<
WWW


PROCESS VENTS
1) EPI Charging
2) Heatup
3) H2SO4 Charging

«



(X
U







^
o
eorgia Pacific
sachtree City,
O &
-

8
*
VO










S



w


STORAGE
S










5



S


EQUIPMENT LEAKS






















BBS
W W til
f» f) O
tfttRS*
£££





° 3 2
Cat Incinera
Cat Incinera
Cat Incinera

VI vo 00



B B B
WWW


PROCESS VENTS
1) EPI Charging
2) Heatup
3) H2SO4 Charging

00



B
w









'3 5
i*
f!
o o
>o

«

W







S
Cat Incinera

R



B
w


STORAGE
!










-
*^


w

j=
EQUIPMENT LEAKS






















B B B
WWW
— — o
tRSRSR
PS?






Condenser
Condenser
Condenser

-» — o



EBB
WWW


PROCESS VENTS
1) EPI Charging
2) Heatup
3) H2S04 Charging

fld



a.
W









eorgia Pacific
olumbus, OH
O 0
vo

I














a
/,
§


STORAGE
VO










2



B
w

ja
EQUIPMENT LEAKS























-------
Ti
 Q)
-H
JJ
C!
O
U
ro

ro
-

I
lib
3*1*
I i 1
•1 j «
31 °-



4)

1


•g
a
o 2"
O ^ *~
c 5 2 >,
3 = 'B -C
£2.2:2
< - 6
a *

OH
a



Emission sources
A* & l-i
Sw •£>
Z> — ~

O.
a
c
S'= K.

(5 •§ —
Q.

.§•
O
tt.



£
Z
§
o
•?
SEE
WWW
«c^-
PS?






4) 4) 1)
G C C


u u u
3SS
CS —i



E E E
WWW


ff
PROCESS VENT.
1) EPI Charging
2) Heatup
3) H2S04Chargi



E






u
"Set
<£o
.2 g"

800
3
O W



1
s
2












>0
-<



W



STORAGE
E
a
0
°\
•H












o
0\



s
Jj
w
i?
EQUIPMENT LE




















§U
2 !S












0,^1
-*



S u
w a



PROCESS VENT.
Si
£s












88




£ 0
w a



STORAGE



_,"=>•
w a





09
g
i z

•5 J

Is
00


I
Si













•*• -
t-^^
1—1


S u
u a

CO
<
EQUIPMENT LE




















Jd
Q^ M \J
w S a-
"|a
„„






Sl^ U
4) 4>
c c c

•O *O "O
U O U
«* o
m S >o
*


a
§ii



PROCESS VENT.


X
a S a






s
« w
E ^
8 E
S 3
§•§
ffi-3:
Ov


I


















a
g



STORAGE
sii
H£s












(S t*» ^f
ts « ^p



«
sli
J
sfl
s
EQUIPMENT LE




















K_ « «-
«*•>
P

E
S

>
•t
V
c

•at
al
o -o ts o vo
m cs


a
j; S S S u
S w w w ffi

e

PROCESS VENT.
1) MeOHDistilla
2) EPI Charging
3) Heatup
4) HCI Addition


a
sll





M
OB
g •*_
S £,
e x

11
e


1
li
C4 ^












— m
o^
^^

a
io
SB



STORAGE























WASTEWATER
S|i
3g8












-» m
->^t~- cs
^^

a
sli
jj
CO
<
EQUIPMENT LE





















-------
T3
 a)
 3
 C
-H
.U
 C
 O
 u
m
W

i
Ills-
3-s 1"
1 "" c
IP




4>
.—
S
'c
o
U
c
o «
c B o fc
3 S3 '» -O
eu S-|£
< ~ s


%


Emission sources
X f £

X
c
'5 t> ,£>
o.

£>
i3
3
u.



0
Z
§
'5*
SB
9 E E S o
•g W W W B3
* VO — tS —

Soo oo *o
ui t- oo
E E E g
3333
3333
O O O O
<» CO SO 1C

^ ^ "? "se
0} d) dl O
09 09 W «5
g g g g
•g fg fg fg |
° 3 ° 3 0 3 r°. 3
tJ <** j U ft U o* W fii


^ ^ o\ t""~
cs —


SB
5 W W W E
c
o

1; — c1 c
g|}j
piii

X
W S K



CO
3 j
'i •*
« B-
c E.

ii
	 .
*••<

i
I
S











•*



sc
I


STORAGE





















WASTEWATER
E|B
'22











U^ ^ fQ
•^- — 2


SB

'a

EQUIPMENT LE

















E E S
WWW
VI O\ f^

?
c^


E
w
3

EQUIPMENT LE


















0


WASTEWATER

















W § W
M \o *°


Pt=P






III
III
O O O
CO CO CO


^ O f^)
!? 2^



WWW

en
PROCESS VENT
1) EPI Charging
2) Heatup
3) Acid Addition


W






O
Jl
3 a
i!
m
""

1
S
0
OQ
•—i











o
00



g


STORAGE
W
r-












>o
f~
^^


W
a

EQUIPMENT LE


















«


WASTEWATER

















E E E
WWW
oo vn —
^- 00 — '

t- t~- t~-






s s s
JS XI J2
III
O O O
to co sn



-^ 
-------
•o
 (U

 a
•H
4J
 G
 O
 U
en

W
111*
'$ 1 1 £
w „
"3 c 4>

si*
o
4)
•o


J
c
8 -a S
13.2 S,
1 l-jji

a '
1

Emission sources
** u" >.
a Si
3
1 If
o.

*o



o
2
§

,o
s
gsg
ggtS


r- t- r-



s s s
Xi -0 X>
X) X) .£>
ggg
CO CO CO

-v,n
•* — Ov


E E E
u u u

PROCESS VENT
1) EPI Charging
2) Heatup
3) Acid Addition


E

fO
E?
l!
81
aas
>0
™*

O
i

s
£1











m
0 fS Ov


E E E
U) U U

PROCESS VENT
1) EPI Charging
2) Heatup
3) Acid Addition


a


S
11
s!
au

™

S
s

E
a
«











t—
—


S

STORAGE
s
S!
M










*
S>
fS

S
U
i
EQUIPMENT LE

















4)

WASTEWATER


































PROCESS VENT


S


1
U
a
5

-«

S
*

E
S











S



w

STORAGE



















WASTEWATER
E
r--
5?











r-
«i


g
|
EQUIPMENT LE















s
8











t"
OO

7:









£
^
g
H






•o
'i
o
o
o
c
ce
S
g
•o
§
8
1
1
e
•2
£
'i
^
M
C
•c
9
•o
|

f
S' '?
S
0 J
§€
3
li
i!
8. a
S I
3 C
1|
•o c
_3 0
* O
!S
5 J5
.l|
S 8

£
13
E
O
c
2
£
I

*§
£
**-
o
.2
s

p
i
0
0
G0
C

2
4)
T3
E
i t
00 * «

i ll
e os
1 S-Q
C g ^
1 1-S ,;
I III
tS o* g
•o t §S
=; S S 5
K . ^ J ^
§ 0,^0 c
12 ,2 ex t- >
W 3 ,O gg B
Estimates were made based on a ratio of 0.08 11
Based on a model fugitive component count dev
Estimated based on a compound-specific factor i
Estimates assumed to be similar to information
None, other than what is generated from proces
Bstimated based on scale up from production da
Storage is in 55-gallon drums.
Information is claimed as CBI.
Control efficiencies reported in questionnaire.
?ounds of HCI in 31.5 wt% HCI solution.
« *O O "O «> **- M)«C .« .^

-------
water scrubbers send the scrubber effluent back to the process.
All wastewater is then steam stripped.  Therefore, the fraction
of EPI transferred from the process vents to the wastewater also
includes the stripping efficiency (estimated to be 94 percent) of
EPI and the subsequent disposal of the stripper bottoms.15  Note
that for the facility that uses a flare to control process
emissions, the overall HAP baseline estimate includes the
subsequent formation of HC1.   Therefore, the net HAP reduction is
much lower than the 98 percent VOC control efficiency usually
assigned to a flare.
     As shown in Table 3-1, emissions from wastewater do not
include the fraction of EPI transferred from process vents via
scrubber effluent.  Other sources of wastewater include extractor
(wash) water, steam stripper bottoms, and vacuum system
discharges (i.e., condensed steam and water from ejectors).
Control of EPI in wastewater is achieved through hydrolysis and
biodegradation.  Modeling of volatilization emissions was done
using Chemdat 7.16
     Uncontrolled HAP emissions from storage tanks were estimated
based on AP-42 tank equations.17  Baseline emissions from tanks
were estimated using the same control efficiencies as those
allowed for process vents, since the tanks were in all cases
reported to be controlled by the same types of devices servicing
process vents  (i.e., flares and scrubbers).
     Uncontrolled HAP emissions from equipment leaks were
estimated based on Synthetic Organic Chemical Manufacturing
Industry  (SOCMI)  average emission factors.1   Baseline HAP
emissions were estimated by considering the types of Leak
Detection and Repair (LDAR) programs currently in use and by
using the control efficiencies for various component types based
                                                               1 Q
on monitoring intervals specified in the Section 114 responses. y
3.3.2  Source Category 2:  WSR Production
     There are an estimated 31 Mg/yr  (69,000 Ib/yr) uncontrolled
HAP emissions from the source category.  The baseline HAP
emissions representing the current level of control from these
facilities is estimated to be approximately 28 Mg/yr
                               3-18

-------
(61,000 Ib/yr),  or 88 percent of the uncontrolled value.
Equipment leaks appear to be the largest source of HAP emissions
in both cases.
     Uncontrolled estimates were calculated based on data
reported in Section 114 information requests.  For process vents,
such as reactor, weigh tank, and distillate receiver displacement
emissions, the volumes of gas displaced were assumed to be
saturated with HAP's at the process conditions.  Controlled HAP
emissions were calculated using reported condenser outlet
conditions and assuming that the streams exiting the condensers
were again saturated with HAP's at these conditions.  Emissions
of HAP's from wastewater occur only if water scrubbers are used
to control vents.  Emissions of HAP's from wastewater were
estimated using the volatilization fraction of 29 percent for EPI
in the wastewater.20  Uncontrolled emissions of HAP's from
storage tanks and equipment leaks were calculated using AP-42
fixed roof tank equations and SOCMI average emission factors,
respectively.  Model process fugitive component counts were used
to estimate emissions from facilities that did not report
fugitive component counts.  These model counts, and their bases,
are detailed in Chapter 5 of this document.
3.4  REFERENCES FOR CHAPTER 3
 1.  Epoxy Resins:  Market Survey and User's Reference.
     W. H. Bowie, ed.  Harvard MBA School.  1959.  p. 1.
 2.  Encyclopedia of Polymer Science and Engineering, Volume 6.
     pp. 329-330.
 3.  Papermaking Additives.  Kirk-Othmer Encyclopedia of Chemical
     Technology, 3rd Edition, Vol. 16.  New York, John Wiley and
     Sons.  1981a.  pp. 803-825.
 4.  Britt,  K. W.  Handbook of Pulp and Paper Technology, 2nd
     Edition.  New York.   Van Nostrand Reinhold Company.  1970.
     pp 377-380.
 5.  Stanford Research Institute (SRI) International.  Process
     Economic Program  (PEP) Report 38A, February 1984.
 6.  Reference 2, pp. 325 and 327.
                              3-19

-------
 7.  Kirk Othmer Encyclopedia of Chemical Technology, 3rd
     Edition,  Vol. 7.  New York, John Wiley and Sons.  p. 276.

 8.  Reference 7,  pp. 274,  276.

 9.  Reference 2,  p. 326.

10.  Reference 2,  p. 326.

11.  Reference 1.   p. 1-11.

12.  SRI International Chemical  Economics Handbook.  April 1988.
     580.0631J.

13.  Muskopf and McCallister.  Epoxy Resins.  Ullman's
     Encyclopedia of Industrial  Chemistry, Vol. A.9. VCH.  1987.
     p. 598.

14.  Chemical  Marketing Reporter.  August 17,  1992.  p. 9.

15.  Wastewater and Wastes Enabling Document.   Revised
     Version 1.0.   Not yet released.

16.  Hazardous Waste Treatment,  Storage, and Disposal Facilities
     (TSDF)  Air Emission Models, Second Edition.  Review Draft.
     April 1989.

17.  Estimating Air Toxics Emissions From Organic Liquid Storage
     Tanks.   Publication No. EPA-450/4-88-004.  October 1988.
     p. 25.

18.  Internal  Instruction Manual for BSD Regulation Development.
     Leaking Equipment- -Pumps,  Valves,  Connectors, Compressors,
     Safety Relief Valves.   July 1992.   p. 2-5.

19.  Reference 18.  pp. 4-6, 4-8.

20.  Reference 15.
                               3-20

-------
               4.0  EMISSION CONTROL TECHNOLOGIES

     This chapter describes the technologies used to control
hazardous air pollutant  (HAP) emissions from the types of sources
found in these industries.  The chapter is structured so that
each section provides descriptions of the control devices,
examples of their use by polymer and resin manufacturers, and
representative HAP emission control levels from test data or
calculations.  The following control technologies are discussed:
(1) condensers,  (2) scrubbers,  (3) carbon adsorption beds, and
(4) thermal destruction systems.  In the non-nylon polyamide
industry segment, the resins are manufactured using batch
processing.  The BLR resins are manufactured primarily in the
continuous mode; therefore, the control device technologies
discussed will deal with controlling emissions from both batch
operations and continuous operations.
     Tables 4-1 and 4-2 present information on the types of
control devices used on various emission sources at facilities in
each of the two source categories.  The rest of this chapter
presents qualitative information of control devices that are or
could be used to control HAP emissions.  Sections 4.1 through 4.4
describe condensers, scrubbers, carbon adsorption beds, and
thermal destruction systems, respectively.  Section 4.5 describes
other control measures including vapor containment, operational
practices,  and vacuum loading.  Sections 4.5 and 4.7 briefly
describe leak detection and repair (LDAR)  programs and storage
tank controls, respectively.  Section 4.8 lists the references
for this chapter.
                               4-1

-------
 TABLE 4-1.   SUMMARY OF EXISTING BLR EMISSION CONTROL DEVICES
      ON  EMISSIONS SOURCES IN BLR MANUFACTURING PROCESSES
Emission
sources
l. Process
vents


2 . Storage
tanks


3 . Equipment
leaks
4 . Wastewater
Control
devices
Water scrubbers
Water scrubbers with
recirculation and
hydrolysis
Carbon adsorption
Refrigeration and
flare
Water scrubbers
Carbon adsorption
Flare
Water scrubbers with
recirculation
LDAR
Biological treatment
*Reported
efficiencies
99%
NR
97%
98%
99%
97%
98%
NR
Variable0
98%
Estimated
efficiencies
97%a
>99%

90%b -
97%a

97%

93-98%
aEstimated by assuming that the EPI, if captured by the water scrubber,
 would still volatilize from the wastewater system.  The fraction of
 EPI emitted from wastewater was estimated with WATER7.
bNet HAP efficiency  (considers formation of HC1)
cThe reduction efficiency is different for each equipment component
 (i.e.,  pumps, valves).  Reduction efficiencies range from 100% to 29%.
                                  4-2

-------
   TABLE 4-2.  SUMMARY OF EXISTING EMISSION CONTROL  DEVICES
             ON EMISSION SOURCES IN  WSR MANUFACTURING
Emission source
1 . Reactor vents




2 . Storage tanks
3 . Equipment leaks
4 . Wastewaterc

Control devices
Condensers
Catalytic
incinerator
Water scrubbers
Carbon
adsorption
Water scrubber
LDAR
Biological
treatment
Reported
efficiencies,
percent
-90
99
99-100

99
99
Variable
NRd

Estimated
efficiencies,
percent
70a

71b


71*

71b

aThe estimated efficiency of condensers were calculated based on the
 fraction of saturation  at the condenser exit gas temperature.
"Estimated by assuming that the EPI, if captured by the water scrubber,
 would still volatilize  from the wastewater system.  The  fraction of EPI
 emitted is estimated to be 29 percent for generic wastewater treatment
 systems.
cAny wastewater generated is the result of using an add-on control device
 for process vents  that  creates a wastewater effluent stream, such as  a
 water scrubber,  or a regenerative carbon adsorption unit.
dNR = Not  reported.
                                  4-3

-------
4.1  CONDENSERS
     Condensers can generally be classified as surface noncontact
and direct contact condensers.  Surface condensers are usually
shell-and-tube heat exchangers, in which cooling fluid flows
inside tubes and the gas condenses on the outside of the tubes.
Direct-contact condensers are those which allow for intimate
contact between the cooling fluid and the gas, usually in a spray
or packed tower.  Although direct contact condensers may also be
part of a solvent recovery system, an extra separation step is
usually involved in separating what was the cooling liquid from
the newly-formed condensate.  An exception to this situation is
the direct contact condenser which uses cooling fluid identical
to the desired condensate; in this case, no separation is
necessary.
     In principle, condensers work by lowering the temperature of
the gas stream containing condensables to a temperature at which
the desired condensate's vapor pressure is lower than its
entering partial pressure.  Typical uses for condensers in resin
manufacturing are on the vents of reactors, distillation columns,
solvent recovery systems, and storage tanks.  Note that the
function of condensers servicing reactors and distillation
columns is often to reflux material.  In these situations, the
material in the reactor or distillation column is allowed to boil
out, either at atmospheric pressure or under vacuum and is
condensed by contact with surfaces that are below the boiling
point of the material.  If the condensers function efficiently,
and if there are no noncondensables in the gas stream entering
the condenser, all the material in the gas stream condenses out
and is returned to the reactor.  These reflux condensers or
primary condensers are not considered to be emission control
devices.  Such applications or process steps often use secondary
condensers, which operate at still lower temperatures and
function primarily as product recovery and air emission control
devices for the emission streams leaving the primary condensers.
                               4-4

-------
HAZARDOUS AIR POLLUTANTS FROM EPOXY RESINS AND NON-NYLON
               POLYAMIDE RESINS  PRODUCTION
      BACKGROUND INFORMATION FOR PROPOSED STANDARDS
                    Preliminary Draft

-------
4.1.1  Design
     The control efficiency attained by a 'Condenser is a function
of the outlet gas temperature.  A typical exhaust gas from a
batch reactor contains a large amount of noncondensable material,
such as air or nitrogen, as well as some fraction of volatile
material.  Before this volatile material can condense, the gas
stream must be cooled to the saturation point of the condensable
material.  Heat transferred from the gas stream during this stage
is called sensible heat.  Cooling the gas stream further, after
complete (100 percent) saturation is reached, ensues condensation
of the volatile material.  Heat removed during the condensation
process is called latent heat.  Both kinds of heat, which, in
refrigeration terminology, usually are summed and reported as
tonnes (each ton is 12,000 Btu/hr) , must be considered in the
design of a condenser.  The heat requirement for a surface
condenser is calculated using the following equation:
                                                            ,4-1,
where :
        Q = heat requirement, Btu/hr;
        U = overall heat transfer coefficient, which is based on
            the tube and shell heat transfer, Btu/hr- ft2- °F;
        A = heat transfer surface area, ft2; and
     AT^jyj = log mean difference in temperature between the
            cooling fluid and the exhaust gas at each end of the
            shell and tube exchanger,  °F.

The amount of heat transferred, Q, may be calculated by
approximating the sensible and latent  heat change when a gas
stream containing condensable material is cooled:
                                    o
                                  + mhv                      (4-2)
where:
      Q  =   heat requirement, Btu/hr;
      o
      m  =   mass flow rate of cooling fluid, Ib/hr;

                               4-5

-------
     Cp  =   heat capacity of cooling liquid, Btu/lb-°F;
     AT  =   temperature difference between inlet cooling
             temperature and condensate temperature, °F; and
     hy.  =   the specific latent heat associated with a phase
             change, BTU/lb

    Based on the above discussion,  the amount of material that
can be condensed from a gas is limited by the following factors:
(1) the emission stream properties, including the heat capacity
and latent heat of condensation of the volatile material and the
stream temperature and (2) the heat transfer characteristics of
the condenser, most notably surface area.  By controlling these
factors, nearly any amount of cooling can be imparted a gas
stream.
     In practice, however, the design of condensers and the
amount of cooling that occurs is based more on economics.
Cooling fluid can range from water at ambient temperature to
brine, which can be cooled to temperatures approaching or below
the freezing point of pure water, to a low-temperature
refrigerant.  The lower the temperature of the cooling liquid,
the more expensive the system becomes.  In some applications, the
condensing system is staged so that certain condensables that may
be present in the stream, i.e., water, will be condensed out at a
higher temperature, and the remainder of the gas can be cooled
further to condense out lower boiling point materials without the
problem of ice formation.
     A second consideration when appreciable water vapor is
present in the gas stream is whether the water will combine with
the condensable material to form a low-boiling azeotrope.  In
such a situation, the saturation temperature of the azeotrope is
lower than the condensing temperature of either pure compound and
the system must be designed accordingly.
4.1.2  Vent Condensers
     4.1.2.1  Reactor Vent Condensers.  The most common
application of condensers in resin manufacturing is the use of
the simple shell-and-tube heat exchanger to control batch reactor
                               4-6

-------
vents.  Epichlorohydrin cross-linked non-nylon polyamide  (wet
strength) resins  (WSR) are manufactured in batch reactors.
Emissions of HAP's can occur from all reactor processing and
transfer steps, including charging, reaction, discharging, and
cleaning.
     In the production of epichlorohydrin non-nylon polyamides
and many small specialty resins, condensers are typically the
only equipment that control HAP emissions from batch reactors.
In many cases, however, the condensers serve primarily as process
equipment and not control devices.  Only during non-reflux
conditions do the condensers function as control devices.  In
these non-reflux cases, the control efficiency of the condensers
is often greatly increased with the substitution of refrigerant
or brine for cooling water.  Estimation of the expected control
efficiency of a condenser is, especially for single-component
systems, easier than the verification of other control technology
efficiencies, such as carbon adsorption, gas absorption,
incineration, etc., as these technologies require that the outlet
gas pollutant concentrations be measured.  To estimate condenser
efficiency, the outlet gas temperature is the only value that
must be known in addition to the inlet conditions.  By assuming
that the vapor phase of the material is in equilibrium with the
liquid at condenser outlet temperature, the percent by volume HAP
discharged from the condenser may be calculated by dividing its
partial pressure by the total pressure.
     4.1.2.2  Potential HAP Emissions Using Condensers.  Listed
below are some of the HAP's typically found in polymer and resin
batch reactors and distillation columns and the percent by volume
of a saturated gas stream at various temperatures.
volume
percent
at 40°C
35.4%
4.8
% volume
percent
at 20°C
12.5%
1.7
% volume
percent
at 5°C
5.3%
0.6
 Methanol
 Epichlorohydrin
     The percent volumes were calculated by dividing the vapor
pressure by the total system pressure  (atmospheric, in this
                               4-7

-------
example).  If a gas stream saturated in methanol enters the
condenser at 40°C and atmospheric pressure, the volume percent
methanol would be 35.4 percent.  Cooling the stream to 20°C
reduces the concentration of methanol to 12.5 percent, a
reduction of 65 percent.  However, dropping the temperature to
5°C would reduce the concentration to 5 percent  (a reduction of
85 percent).  This shows that using a colder condenser is more
effective than just cooling to ambient temperature.  This
information also points out that different HAP's need to be
cooled to different temperatures to achieve equivalent
efficiencies.  For example, a gas stream saturated with
epichlorohydrin needs only to be cooled to 40°C to achieve the
same HAP concentration as a stream saturated with methanol
at 5°C.
4 .1 .'3  Application to these Source Categories
     Currently, the majority of facilities that manufacture WSR
use vent condensers to control vapor displacement and heat up
emissions in their batch reactors.  Additionally, two
applications for condensers exist in DGFBPA manufacturing:  (1) a
refrigerated condenser used to recover process vent and storage
losses prior to flaring, and (2) a vent condenser to control
displacement emissions for a feedstock mixing stage.  This
section presents information on existing condenser applications
and on potential applications in these source categories.
     4.1.3.1  Existing Applications.
     4.1.3.1.1  WSR.  Currently, seven of the seventeen
facilities that manufacture WSR make use of a reactor vent
condenser to control both displacement emissions and reactor
heat-up emissions.  All these vent condensers are operated with
cooling water, which was assumed for the purposes of estimating
baseline emissions to cool the gas streams to 20°C.  Efficiencies
for the various facility applications vary, depending on what
temperature the displacement occur,  on the temperature range for
the reactor heat up emissions,  and on what HAP's are being
emitted during the emission events.   For displacements, data
indicate that material is fed into the reactors at approximately
                               4-8

-------
45°C.  Therefore, for EPI, cooling the displacement streams to
20°C result in efficiencies of 72 percent.  During the exothermic
reaction, the reactor temperatures range from the incoming 45° to
70°C.  A control efficiency of 84 percent was estimated for a
cooling water condenser, based on the net reduction from the
average EPI concentration in the temperature range.  It should be
noted that in certain instances where methanol is formed as a by-
product of the pre-polymer manufacturing process, the reactor
vent condenser is used as a reflux condenser to condense the
methanol distillate.
     4.1.3.1.2  BLR.  Currently one facility makes use of a
refrigerated condenser to recover HAP compounds from a manifold
of sources that include all process and storage tank vents.  The
plant reports that the practical temperature limit of the gas
stream existing this condenser is at -8°C, due to icing.  A
recovery efficiency of approximately 70 percent was reported for
this situation.   Also, one BLR manufacturer reports the use of a
cooling water condenser to control displacement emissions of EPI
from a pre-mix stage.  However, the temperature of the material
in the pre-mix stage is also at ambient temperature (assumed to
be 20°C).  No control efficiency was therefore assigned to this
situation.
     4.1.3.2  Potential Applications.
     4.1.3.2.1  WSR.  Often, weigh tank displacements or methanol
distillate receivers could have vapor return lines back to the
batch reactors.  The condensers serving the reactors would then
also have some effect on emissions from these sources.
     4.1.3.2.2  BLR.  Emissions from the resin purification and
finishing stages tend to be saturated with EPI at elevated
temperatures of 60°C or more, making them ideal candidates for
control using refrigeration.
     4.1.3.3  Refrigeration Systems for Vapor Displacement from
Process Vessels or Storage Tanks.  In 'some situations, shell-and-
tube condensers are used to control HAP emissions from vapor
displacement from process vessels or storage tanks.  Such
applications are usually for solvent or product recovery
                               4-9

-------
purposes, since it is often desirable to recover material that
would otherwise be emitted as a HAP.  This is especially true for
products that require expensive feedstocks and solvents.
     Vapor recovery systems are often designed so that the
recovered material cost offsets the energy and capital costs of
the systems themselves.  In many cases, however, the recovered
material cost is insignificant compared to the cost of purchasing
and operating the recovery systems.  In such a case, the decision
to install a solvent recovery system as opposed to another type
of system is based on other factors, such as control
effectiveness and concerns about waste handling and ultimate
disposal of the HAP.
     While refrigeration systems are not often used solely to
control single vapor displacement events such as reactor charging
and extractor (mixer-settler) charging, they are often feasible
for controlling collected displaced vapors from a number of
sources.
     It has been shown that facilities that have a large number
of storage tanks, for example, can use staged refrigeration
systems which employ pre-cooler sections.  Often, the precooler
operates at a temperature just above the freezing point of water.
This condenser (usually an indirect shell and tube heat
exchanger) rids the vapor stream of as much water as possible
that would otherwise collect on heat transfer surfaces as ice and
lower the heat transfer potential of colder surfaces.  After the
vapor passes through the initial indirect condenser  (pre-cooler),
it enters the main condenser section, which can cool the gas
stream to very low temperatures, on the order of -100°F.
     Low-temperature refrigeration systems such as the one
described above have been used to control vapor displacement
emissions from multiple sources such as working losses from a
tank farm. Often, the mixtures are separated by distillation
although only one or two pure components may be recovered for
reuse.
     Perhaps the most important issue to consider when evaluating
the need for such a system is the required size of the unit.  For
                              4-10

-------
a number of storage tanks or for a number of process vents from
one manufacturing area, the system is most effective when it can
control the stream having the maximum vapor concentration at a
constant flow rate.  Maintaining a constant vapor loading rate is
crucial to efficient operation of the condenser.  Consider a
condenser which has been sized to control displacement emissions
from the loading of a storage tank.  At any times other than
loading, the flowrate through the condenser will be virtually
zero.  To avoid over- cooling the coolant during these times, the
compressor must go into a bypass mode which results in
inefficient use of the system.
     To optimize the efficiency of the system, the displaced
vapors or process vents of finite duration must be staggered or
controlled so that the system receives a fairly constant vapor
inlet loading.  The use of this system for process vents from
batch processes is more difficult than for processes that have
continuous emissions with constant properties.  No examples in
this industry have been found where multiple process or storage
tank vents are routed to a refrigeration system.
     4.1.3.4  Combination Vapor Compression and Condensation.  In
some situations, condensation is aided by compressing the gas
stream containing VOC's to some elevated pressure in conjunction
with the use of a condenser.  The purpose of this compression
step is to condense out the same amount of material at a higher
condenser operating temperature.  For example, consider the
simple calculation used to estimate the vapor phase mole fraction
of the HAP:
                          Y,
                           HAP   PTQTAL
                                  TOTAL                     (4_3)
     A low value of Y-£&p is desired at the outlet of the
condenser.  This can be achieved by reducing the vapor pressure
of the HAP  (by lowering the gas temperature),  by increasing the
total pressure of the system, or by a combination of both.
     Most applications that use a combination vapor compression-
condensation system use liquid ring compressors with
                               4-11

-------
recirculating ring fluid.  These compressors are available for
numerous ranges of flowrates and discharge pressures.  Currently,
no facilities in this source category use such systems for
control of HAP's.
4.2  SCRUBBERS
4.2.1  General Gas Absorbers
     Scrubbers, or gas absorbers, function by providing an
intimate contacting environment for a gas stream containing
material that is soluble in the contacting liquid.  The rate of
mass transfer from the gas to the liquid depends upon a driving
force related to the actual HAP concentrations in the gas and
liquid versus the equilibrium-defined HAP concentration in the
two media at each point along the contacting path.  The most
common types of scrubbers found in polymer and resin industries
are packed towers and spray chambers. '
     4.2.1.1  Liquid Gas Absorbers or Water Scrubbers.  Gas
absorbers are limited primarily by the solubility of the vapor in
the liquid stream.  Most of the scrubbers found in industry use
water as the scrubbing medium, so the decision to use these
devices depends largely on the solubility of the HAP's in water.
In general, compounds containing nitrogen or oxygen atoms that
are free to form strong hydrogen bonds, and having one to three
carbon atoms are soluble; those compounds with four or five
carbons are slightly soluble; and those with six carbon atoms or
more are insoluble.2  While common solvents such as methanol,
isopropyl alcohol and acetone are very soluble in water, solvents
like toluene and epichlorohydrin are not.  However, many non-
nylon polyamide producers and large producers of BLR report
control of epichlorohydrin process emissions with water
scrubbers.
     4.2.1.2  Chemical Scrubbers.  Some facilities use chemical
scrubbers to control some pollutants.  Instead of using a liquid
medium to absorb material out of the gas phase, chemical
scrubbers use the liquid medium to react with material in the gas
phase.   The hydrolysis of EPI is an example of a mechanism
occurring in a chemical scrubber.  Another good example is an
                              4-12

-------
emergency destruction scrubber for a compound such as phosgene
(COC12).   Phosgene, when reacted with slightly basic water,
hydrolyzes to HC1 and C02.   Although these product gases still
require control, their toxicities are much lower than that of the
initial reactant.  Chemical scrubbers are often used as emergency
back-up devices.
4.2.2  Design
     The design of a scrubber involves the estimation of the
ratio of gas-to-liquid mass flow rates and the appropriate amount
of contacting area necessary to achieve the desired removal.  A
necessary piece of information, which can be difficult to obtain
without experimental work,  is the equilibrium curve depicting
equilibrium mole fractions of the HAP in the solvent in the vapor
and liquid phases at the contacting temperature.  The equilibrium
curve, as the name implies, is not a straight line, but
approximations may be used and the curve may be assumed to be
straight in some situations.  For water scrubbers, the Henry's
law constant at the water temperature is often used to define the
slope of the equilibrium curve.
     The estimation of the physical properties of a scrubber
design, such as the number of transfer units (NOG) and the height
of a transfer unit (HQG)  for a packed tower, may be based on the
reported removal efficiency of a system and the reported
liquid-to-gas mass velocities.  The EPA publication
EPA-450/3-80-027, Organic Chemical Manufacturing Volume 5:
Adsorption. Condensation, and Absorption Devices. December 1980,
contains the methodology that can be used to estimate such
parameters.3  Note that verifying the efficiency of a scrubber
without using test data is more difficult than verifying the
efficiency of a condenser since there are more variables to
consider and the equilibrium data at the required temperature are
not always available.  It is perhaps for this reason that
unrealistically high water scrubber efficiencies may sometimes be
reported.
                               4-13

-------
4.2.3  Scrubber Applications
     4.2.3.1  WSR.  Six of the seventeen WSR facilities use water
scrubbers to control reactor vent emissions.  Reported control
efficiency for these devices are high, at 99 percent or better.
However, an overall control of 71 percent for EPI was assigned to
these devices for purposes of estimating controlled  (baseline)
emissions to account for the fraction of EPI that will volatilize
out of the wastewater during treatment.
     4.2.3.2  BLR.  Scrubbers are commonly used by facilities in
the BLR source category to control process vent emissions
containing multiple HAP's.  Some producers of BLR use water
scrubbers and claim control efficiencies of 98 percent or better
for HAP's from process vents from resin finishing and
purification stages containing epichlorohydrin.  If the scrubber
effluent is sent directly to wastewater treatment, the overall
control efficiency for HAP's can be below 98 percent, however,
since some fraction of HAP's is expected to volatilize out of the
wastewater.  However, specific information on facility
bio-treatment units were reported from the three BLR facilities.
This data, when inputted to the EPA's WATER? model, yielded
estimated destruction efficiencies ranging from 93 to 99 percent.
Therefore, overall removal efficiencies from scrubbers were still
estimated to be high for the BLR facilities.  Additionally, one
of the BLR manufacturer recycles scrubber effluent directly to
the process, eliminating the majority of volatilization
emissions.  In this case,  the scrubber effluent is used as
extractor water makeup.
     Based on information submitted in Section 114 responses,
there are questions relating to the concentration of EPI in the
scrubber effluent, which often does not agree with the amount of
EPI reportedly removed.  It is possible that some EPI undergoes
hydrolysis to HC1 and C02  upon contact with the scrubber water;
however, no test data is available to confirm either the scrubber
efficiencies or to explain the potential fate of the EPI.
                              4-14

-------
4.3  CARBON ADSORPTION
     Cairbon adsorbers function by capturing vapor that is present
in a gas phase on the surface of granular activated carbon.
Adsorbers can be of the fixed-bed design or fluidized-bed design.
Fixed-bed adsorbers must be regenerated periodically to desorb
the collected organics from the carbon.  Fluidized-bed adsorbers
are continuously regenerated.  Most processes that use carbon
adsorbers use the fixed-bed type.  Some use nonregenerative
units, which are contained in 55-gallon drums and are used mostly
for controlling odor from small process vents.  Such units are
returned to their distributors for disposal after they can no
longer adsorb effectively.
4.3.1  Design
     Carbon adsorption is usually a batch operation involving two
main steps, adsorption and regeneration.  This system usually
includes multiple beds so that at least one bed is adsorbing
while at least one other bed is being regenerated, thereby
ensuring that emissions will be continually controlled.  A blower
is commonly used to force the HAP-laden gas stream through the
fixed carbon bed.  The cleaned gas is then exhausted to the
atmosphere.  A gradual increase in the concentration of organics
in the exhausted gas from its baseline effluent concentration
level signals it is time for regeneration.  The bed is shut off
and the waste gas is routed to another bed.  Low-pressure steam
is normally used to heat the carbon bed during regeneration,
driving off the adsorbed organics, which are usually recovered by
condensing the vapors and separating them from the steam
condensate by decantation or distillation.  After regeneration,
the carbon bed is cooled and dried to improve adsorption.  The
adsorption/regeneration cycle can be repeated numerous times, but
eventually the carbon loses its adsorption activity and must be
replaced.  Typically, facilities replace a portion of the carbon
bed on an annual basis.
     The efficiency of an adsorption unit depends on the type of
activated carbon used, the characteristics of the HAP, the HAP
concentration, and the system temperature, pressure, and
                               4-15

-------
humidity.  Overall HAP removal efficiencies depend on the
completeness of regeneration, the depth of the carbon bed, the
time allowed for contact, and the effectiveness of recovery of
desorbed organics.  Carbon adsorption is not suitable for gas
streams with a high concentration of organics, with organics with
boiling points greater than 250aC or molecular weights greater
than 200, with relative humidities greater than 50 percent, with
high levels of entrained solids, or with temperatures over 100°F.
Adsorbing organics from gas streams with high concentrations of
organics may result in excessive temperature rise in the bed due
to the accumulated heat of adsorption; this can be a serious
safety problem due to the risk of a bed fire.  High molecular
weight organics and organics with high boiling points are
difficult to remove from the carbon under normal regeneration
temperatures.  The continuing buildup of these compounds on the
carbon will greatly decrease the operating capacity and will
result in frequent replacement of the carbon.  Plasticizer or
resins should also be prevented from entering the carbon bed,
since they may react chemically on the carbon to form a solid
that cannot be removed during regeneration.
     Entrained solids in the gas stream may cause the carbon bed
to plug over a period of time.  These solids are generally
controlled by a cloth or fiberglass filter.
     Gas streams with high relative humidities will affect the
adsorption capacity of the bed.  Humidity control can be achieved
by cooling and condensing the water vapor in the gas stream.  The
relative humidity can also be decreased by adding dry dilution
air to the system,  but this will usually increase the size and
thus the cost of the adsorber required.
     The adsorption capacity of the carbon and the effluent
concentration of the adsorber are directly related to the
temperature of the inlet stream to the adsorber.  Normally, the
temperature of the inlet stream should be below 100°F or the
adsorption capacity will be affected.  Inlet stream coolers are
usually required when emission stream temperatures are in excess
of 100°F.
                               4-16

-------
     When designing and installing carbon bed adsorber systems
there are several safety factors that need to be considered.
Fixed carbon beds can spontaneously combust whenever the gas
stream contains oxygen and compounds easily oxidized in the
presence of carbon, such as ketones, aldehydes, and organic
acids.  Heat generated by adsorption or by oxidation of organics
in the bed is usually transported from the bed by convection.  If
less convection heat is removed than is generated, the bed
temperature will rise.  Higher temperatures will further increase
the oxidation decomposition, and hot spots exceeding the
autoignition temperature of the carbon may develop in the bed.
If an adsorber is shut down for an extended period and not
regenerated sufficiently upon startup, reintroduction of the
organic-laden stream may also lead to bed combustion.  There are
preventive measures that can be taken to ensure safe operation of
carbon adsorbers.  Using adequate cooling systems, regularly
inspecting valves to prevent steam leaks, and using adsorbers
only on low concentration streams all will ensure safe operation.
In addition, beds used for adsorbing ketones should not be dried
completely after regeneration.  Although not drying them may
reduce adsorption capacity somewhat, it is an effective safety
measure because the water acts as a heat sink to dissipate the
heat of adsorption and oxidation.
     Carbon adsorption systems normally are designed for gas
velocities between 80 and 100 ft/min.4  The maximum rate of
recovery of organics is dependent upon the amount of carbon
provided and the depth of the bed needed to provide an adequate
transfer zone.  The required amount of carbon may be estimated
from an adsorption isotherm, which is generally available for
different compounds at various partial pressures.
4.3.2  Applicability
     Carbon adsorbers have a limited and varied use in the BLR
industry.  Emissions from reactor vents and storage tanks are
known to be controlled by carbon bed adsorbers at one facility.
     In one application, the adsorber is preceded by a condenser
on a process vent.  Since condensers are more efficient on
                               4-17

-------
saturated streams and carbon bed adsorbers are more efficient on
dilute streams, a condenser followed by a carbon bed adsorber may
be an effective control system.  Test data from one source
indicated a 97 percent efficiency for removal of epichlorohydrin
from a waste stream controlled with this system.
     Nonregenerative carbon adsorbers are also used in this
industry to control emission from storage tanks or to control
displacement emissions from small reactors.  These systems are
extremely simple in design.  When the activated carbon becomes
spent, it is replaced with a new charge.  The spent carbon can be
reactivated offsite and eventually re-used.  Carbon canisters,
normally the size of 55-gallon drums, can be used to control
small vent streams  (less that 500 actual cubic feet per minute
[acfm])  with low organic concentrations.  One advantage of these
systems is that they appear to be immune to normal fluctuations
in gas streams that are common to batch processes.
     For all practical purposes, it is difficult to estimate the
efficiency of a carbon adsorption system.  Although EPA has
generally accepted a control efficiency of 95 percent for streams
containing compounds that are considered appropriate for
adsorption, the actual control efficiency attained by a
particular system is largely dependent upon the amount of time
elapsed and the amount of material adsorbed since the last
regeneration or replacement.  Note also that it is more difficult
to predict the amount of material that has been adsorbed for the
intermittent streams with variable characteristics typical of
batch processes than for continuous emission streams with
constant properties.
     As mentioned previously,  most applications of carbon
adsorbers are as secondary control devices following condensers.
Because of the highly flammable nature of many solvents, the
industry trend is away from using these devices as primary
control devices.
     One issue that has been raised with regard to batch
processing emissions in particular is the possibility of
stripping VOC out of the activated carbon and back into the
                               4-18

-------
emission stream during low concentration-high flowrate periods.
For example, consider a batch vent stream 'from a reactor entering
a carbon bed.  During a processing event such as material
charging at elevated temperatures, volatile compounds will be
sorbed onto the surfaces of the activated carbon.  During reactor
purge, however, a high flowrate, low VOC concentration will be
introduced into the bed that may strip volatile materials back
out of the bed and into the exit gas stream.
4.4  THERMAL DESTRUCTION
     It is usually possible to route process vents to an
incinerator or flare for control.  In most cases, these devices
also are not sized for peak concentrations of HAP's, but for
constant flow and concentration of HAP's.  Incineration systems
are usually quite costly and must operate continuously; therefore
the use of such a system is limited to those applications where a
number of vents may be controlled.  Note also that the byproduct
combustion gases in most cases must also be controlled, thereby
increasing costs.
4.4.1  Flares
     Flaring is an open combustion process that destroys HAP
emissions with a high temperature oxidation flame to produce
carbon dioxide and water.  Good combustion in a flare is governed
by flame temperature, residence time of components in the
combustion zone and turbulent mixing of components to complete
the oxidation reaction.
     4.4.1.1  Design.  There are two major types of flares:
elevated flares and ground flares.
     The elevated flare is a single burner tip elevated above
ground level for safety reasons.  The vented gases are burned in
the diffusion flame.  The gas stream containing a HAP enters at
the base of the flame, where it is heated by already-burning fuel
and the pilot burner at the flare tip.  Fuel flows into the
combustion zone, where the exterior of the microscopic gas pocket
is oxidized.  The rate of reaction is limited by the mixing of
the fuel and oxygen from the air.  If the gas pocket has
sufficient oxygen and residence time in the flame zone, it can be
                               4-19

-------
completely burned.  The diffusion flame receives its combustion
oxygen by diffusion of air into the flame from the surrounding
atmosphere.  The high volume of fuel flow in an elevated flare
requires more combustion air at a faster rate than simple gas
diffusion can supply, so steam is injected to increase gas
turbulence in the flame boundary zones, drawing in more
combustion air and improving combustion efficiency.  The steam
injection promotes smokeless flare operation by minimizing the
cracking reactions that form carbon.  Steam flow can be
controlled manually, but automatic control by sensors monitoring
flame characteristics gives a faster response to the need for
steam and a better adjustment of the quantity required.
     Ground flares have multiple-gas-burning heads that are
grouped in an enclosure and are staged to operate based on vent
gas flow to the flare.  The enclosure reduces the luminosity,
noise, and allows the flare to be located at ground level.  The
size, design, number, and arrangement of the burner heads depend
on the flare gas characteristics.  Unlike elevated flares, an
exterior source of steam or air to enhance combustion is rarely
required for ground flares.  Stable combustion can be obtained
with some gases that have heat contents as low as 50 to
60 Btu/ft.  Reliable and efficient operation can be attained at
up to 100 percent of design capacity.  The number of burner heads
and their arrangement into groups for staged operations depend on
the discharge characteristics of the emission source gas.  To
ensure reliable ignition,  pilot burners with ignitors are
provided.  The burners are enclosed in an internally insulated
shell and must be high enough to create enough draft to supply
sufficient air for smokeless combustion of the waste gas for
dispersion of the thermal plume.5
     4.4.1.2  Applicability.   Although flares are not as widely
used for controlling emissions from batch resin processes as
other control devices, they are adjustable and can be useful for
these processes.   In many cases,  however,  they require a        . •
considerable amount of auxiliary fuel to combust gases that
contain dilute concentrations of organics or organics that have
                              4-20

-------
low heats of combustion.  Flares are capable of handling the
highly variable flows that are often associated with batch
process operations.  Steam-assisted elevated flares may be used
to control emission from high-concentration, intermittent vent
streams.  Elevated flares are used to control emissions during
emergency venting or during process upsets, such as startup and
shutdown.  These intermittent emissions are characteristic of
normal batch process operations with the exception that they may
be more concentrated than normal batch emissions.  Ground flares
have less capacity than elevated flares and are usually used to
burn gas continuously.  Ground flares can operate efficiently at
up to 100 percent of design capacity.  The burner heads can also
be specifically sized and designed for the materials in the flare
gas.
     One of the BLR manufacturers surveyed reports the use of a
flare for emission control from a plant manifold.  It is used to
incinerate a mixture of toluene, epichlorohydrin, ethylene oxide,
and propylene oxide vapors which are emitted from several
streams, of which the BLR process and storage emission streams
are a part of.  The reported destruction efficiency is
98 percent, although the flare does create significant amounts of
HC1.  An overall control efficiency of 60 percent was assigned to
this device to account for a net HAP reduction.
4.4.2  Thermal and Catalytic Oxidizers
     Thermal and catalytic oxidizers may be used to control
emission streams of VOC's and HAP's, although they are not
especially suited for intermittent or noncontinuous flows,
especially on an economic basis.  Because they operate
continuously, auxiliary fuel must be used to maintain combustion
during episodes in which the organic HAP load is below design
conditions.  For groups of sources, however, the use of devices
may be reasonable.
     '4.4.2.1  Design.  The design and operation of incinerators
is most influenced by the necessary combustion temperature,
residence time and desired destruction efficiency.  Most
incinerators are capable of achieving better than 90 percent
                               4-21

-------
control efficiency.  Catalytic incinerators, which operate at low
temperatures compared to thermal oxidizers  (500° to 1000°F as
opposed to >1000°F) generally achieve from  90 to 95 percent
destruction, while thermal oxidizers usually achieve at least
98 percent.  The destruction efficiency of  a thermal oxidizer can
be affected by variations in chamber temperature, residence time,
inlet HAP concentration, compound type, and flow regime  (mixing).
Test results show that thermal oxidizers can achieve 98 percent
destruction efficiency for most VOC compounds at combustion
chamber temperatures ranging from 700° to 1300°C  (1400° to
2374°F) and residence times of 0.5 to 1.5 seconds.  These data
indicate that significant variations in destruction efficiency
occurred for C-^ to Cc alkanes and olefins,  aromatics  (benzene,
toluene, and xylene), oxygenated compounds  (methyl ethyl ketone
and isopropanol), chlorinated organics (vinyl chloride), and
nitrogen-containing species  (acrylonitrile  and ethylamines) at
chamber temperatures below 760°C (1400°F).^  This information,
used in conjunction with kinetics calculations, indicates the
combustion chamber parameters for achieving at least a 98 percent
destruction efficiency are a combustion temperature of 870°C
(1600°F) and a residence time of 0.75 seconds  (based upon
residence in the chamber volume at combustion temperature).
Thermal oxidizers generally can burn contaminated streams
(i.e., streams that could poison catalyst in a catalytic
incinerator) but may need to achieve combustion temperatures
approaching 2000°F to effectively oxidize such compounds as
halogenated solvents.  Incinerators usually operate with excess
air to ensure sufficient combustion.  These devices also contain
refractory linings that cannot withstand sudden warmup or
cooldown periods.7
     4.4.2.2  Specific Applications.  Incinerators often are used
to control multiple process vents that can be manifolded
together.   For example, processes that are  contained within one
building or processing area are sometimes vented together and
routed to an incinerator.  For some of these vents,  a primary
control device such as a condenser is located upstream.  Note
       . '                      4-22

-------
 that the stack gases resulting from combustion often contain acid
 such as HC1  and may require an exhaust gas  control  device such as
 a caustic scrubber.
      There are also some incineration units that can handle low
 flow rates (in the range of 10 to 500 scfm) .   These units can be
 applied to single emission streams,  such as reactor vent
 emissions.®
      The use of thermal  and catalytic incinerators  in the
 industry is  limited.   One manufacturer of WSR controls EPI
 emissions from reactor and storage tank vents by piping them to a
 manifold system with vents from other processes and sending it
 all  to a common catalytic incinerator.   This  manufacturer reports
 a destruction efficiency of 99 percent with minimal acid  gas
 generation.
 4.5   OTHER CONTROL MEASURES
 4.5.1  Vapor Containment
      Vapor containment is only a control when in conjunction with
 a control device.   It is discussed below because some of  the
 facilities in the source category report more efficient results
 when these systems are used prior to existing control devices.
      Probably one of the less  expensive and more effective
 methods of controlling displaced vapors from such events  as
 vessel charging and from storage tank working losses is to use
 vapor return lines to vent the vapors back  to the vessel  from
 which the liquid was originally taken.   Assuming that emissions
 from this vessel ultimately are controlled,  essentially
 100  percent  control of the vapors at the point sources is
 achieved,  and there do not appear to be many adverse effects from
 the  standpoint of safety or convenience.
      The use of vapor return lines to control displacement
 emissions from storage tanks is fairly common in this source
 category.  As displacement events are one of  the major sources of
 HAP  emissions,  this control method has the  potential to make a
.significant  reduction in the overall emission rate.
      4.5.1.1  Variable Volume  Containment Devices.   The use of a
 variable volume vessel capable of containing process vents and
                               4-23

-------
continuously releasing an exhaust with constant stream properties
of flowrate and HAP concentration to a control device also is a
method of vapor containment.  One of the BLR facilities uses an
expandable vapor containment device to capture various emission
streams from process and storage vents resulting from other resin
processes.  The exhaust from the breather balloon is routed to a
control device.  Presumably, the advantage to using such a
containment system would be to minimize the fluctuation in
emission stream characteristics that might normally occur in a
manifold of vents.  A disadvantage to this system is that it
could be an explosive hazard.
4.5.2  Operational Practices
     4.5.2.1  Limiting the Use of Inert Gas.  Many applications
in batch resin manufacturing require the use of inert gas for
blanketing and purging of equipment for safety purposes.  By
eliminating purging of reactor vessel headspaces, HAP emission
have been shown to decrease significantly in the WSR industry.
At one facility, process emissions changed from 55,000 Ib/yr EPI
to 2,350 Ib/yr by eliminating the nitrogen purge.  There are also
other practices, however, such as the blowing of lines to move
material and the sparging of large volumes of liquids, that could
be changed so as to reduce the amount of inert gases in the
streams and thereby make the streams more suitable for control by
devices such as condensers.
     The blowing of lines with nitrogen to move material, for
example, could be replaced by simple pumping and/or setting the
lines on an incline.  Blowing cannot be totally eliminated,
however, because the vapor that may be contained in the vapor
space in the lines may need to be purged at various times before
maintenance.  Reductions from these changes have not been
quantified for the WSR industry.
     4.5.2.2  Eliminating Direct Water-contacting Vacuum
Generating Devices.  The replacement of steam ejectors and vacuum
pumps with vacuum generating devices that do not create an
opportunity for contact of the vapor laden gas stream with seal
fluid would also eliminate emissions from all wastewater
                               4-24

-------
generated in WSR manufacturing and from significant portions of
wastewater generated in BLR manufacturing.  Recognizing that the
control of HAP's with water may only "transfer" emissions to
another source, the use of vacuum pumps, steam ejectors, and
water scrubbers may become less viable for processing of HAP's.
4.5.3  Vacuum Loading
     Vacuum loading of volatile material into process vessels
such as WSR batch reactors will reduce emissions from
displacement events in cases where there is no volatile residual
material left in the reactor.  Several facilities in the WSR
source category report the use of vacuum loading as a means of
limiting displacement emissions.
4.5.4  Process Changes
     In WSR manufacturing, the replacement of dibasic ester with
adipic acid in the initial polyamide manufacturing reaction will
eliminate methanol emissions from the process.  Likewise, the
substitution of H2S04 for HC1 to halt crosslinking will eliminate
HC1 displacement and storage emissions.  Another argument for
substituting H2S04 for HC1 is that HC1 converts any excess EPI to
dichloropropenols, which are carried with the resin in aqueous
solution.  When base is added to activate the resin, the
dichloropropenols would be converted to EPI in situ.   This
appears to be a worker exposure concern, however.
4.6  LEAK DETECTION AND REPAIR  (LDAR) AND PRESSURE TESTING
     PROGRAMS
     Emissions of HAP's from process components such as pumps,
valves, flanges, and sampling connections that happen to leak can
be minimized by instituting a periodic leak detection and repair
program  (LDAR).  Several types of LDAR programs exist; the New
Source Performance Standards  (NSPS) Subpart W prescribes a
quarterly monitoring program of pumps, valves, flanges, and
connections.   The leak definition is 10,000 ppmv.
Alternatively, the Equipment Leaks Negotiated Regulation, "Reg
Neg", prescribes a more stringent LDAR program.1^  A leak
definition is 500 ppm, and the monitoring interval for components
varies between quarterly and monthly monitoring.  For batch
                               4-25

-------
processes, the Reg Neg allows the substitution of pressure

testing.
4.7  STORAGE TANK CONTROLS

     Besides add-on controls such as refrigerated condensers and

scrubbers, emissions from storage tanks (working and breathing

losses) can be controlled with external and internal floating

roofs.  Tank sizes range from 30,000 to 200,000 gallons for BLR

manufacturing, and 10,000 to 100,000 gallons in WSR

manufacturing.  Floating roofs are applicable to these sizes of

tanks.
4.8  REFERENCES FOR CHAPTER 4

 1.  Wastewater and Wastes Enabling Document Draft.
     EPA/OAQPS/ESD.  May 1992.

 2.  Solomons, T. W. Graham.  Organic Chemistry, 2nd Edition.
     John Wiley and Sons, New York.  1976, 1978, 1980.  p. 80.

 3.  Organic Chemical Manufacturing, Volume 5:  Adsorption,
     Condensation, and Absorption Devices.  Publication
     No. EPA-450/3-80-027.  December 1980.

 4.  VIC Manufacturing Company.  Carbon Adsorption/Emission
     Control Benefits and Limitations.

 5.  Pacific Environmental Services.  Chemical Processing Plants
     Summary Report for Technical Support in Development of a
     Revised Ozone State Implementation Plan for Memphis,
     Tennessee.  June, 1985.

 6.  Memo and attachments from Farmer, J., EPA:ESD, to
     distribution.  August 27, 1980.  29 p.  Thermal incinerator
     performance for NSPS.

 7.  EPA-450/3-83-008.  Guideline Series.   Control of Volatile
     Organic Compound Emission from the Manufacture of High-
     Density Polyethylene, Polypropylene,  and Polystyrene Resins.

 8.  Bedoya, J.G., In-Process Technology,  Inc.  Letter to
     B. Shine, MRI, summarizing cost and cost-effectiveness for
     small incineration units.  March 28,  1991.

 9.  Dynamac Corporation Assessment of Epichlorohydrin Uses,
     Occupation Exposure and Releases.  Final Draft.  July 11,
     1984.  p. 123.  Prepared for the Economics and Technology
     Division, Office of Toxic Substances, U. S. Environmental
     Protection Agency.
                              4-26

-------
10.   Internal Instruction Manual for BSD Regulation Development,
     Leaking Equipment--Pumps,  Valves,  Connectors, Compressors,
     Safety Relief Valves, U. S. Environmental Protection Agency,
     Research Triangle Park,  N.C.  July 1992.
                               4-27

-------
   5.0  MACT FLOORS, REGULATORY ALTERNATIVES, AND MODEL PLANTS

     This chapter presents information on the current level of
control of hazardous air pollutant  (HAP) emissions from basic
liquid resins  (BLR) and wet-strength resin  (WSR) manufacturing
facilities and presents regulatory alternatives and associated
emission reductions.  Additionally, model plants were developed
for both source categories to illustrate the processes typically
employed by facilities located in each source category.  Both
source categories are small  (there are three BLR facilities and
17 WSR facilities), therefore actual plant parameters were used
to evaluate impacts for existing sources.  Section 5-1 is a
discussion of the current HAP control levels and the maximum
achievable control technology (MACT) floors.  Section 5-2
presents the proposed regulatory control alternatives for the
industries.  Section 5-3 comprises the discussion and
presentation of the model plants.
5.1  MACT FLOORS
     The Clean Air Act Amendments of 1990 state that HAP emission
standards for existing sources cannot be less stringent than the
average emission limitation achieved by either  (1) the best
performing 12 percent of existing sources or (2) the best
performing 5 sources in the category or subcategory with fewer
than 30 sources.  For new sources in a category or subcategory,
the amendments state that "the maximum degree of reduction in
emissions that is deemed achievable for new sources in a category
or subcategory shall not be less stringent than the emission
control that is achieved in practice by the best controlled
similar sources...."  These statements define the MACT floors for
existing and new sources,
                               5-1

-------
     According to the statute, MACT floors must be calculated for
each "source".  In this standard, we have chosen to define
equipment leaks as a portion of the source separate from other
emission points, such as process vents, storage, and wastewater
volatilization emissions.  Therefore, we have developed two MACT
floors for each source category, one for combined emissions of
process vents, storage, and wastewater, and one for equipment
leaks.
     The MACT floors for process vents, storage tanks, and
wastewater portion of the source and the equipment leaks portion
were developed for each source category in the following way:
     1.  For existing BLR production facilities, the pounds  (Ib)
of HAP emitted for each portion of the source was divided by the
actual BLR production rate of the facility to yield a value, in
units of Ib HAP emitted per million  (MM) Ib of product, that was
then averaged across the three facilities to obtain the MACT
floor.  The MACT floor for new sources could not be based on
Ib HAP emitted per MMlb product because the MMlb product factor
would correspond directly with an actual facility's value, and
therefore the production rate, which is reported to be
confidential in all cases, could be obtained.  The MACT floor for
new sources, therefore is currently set as a technology-based
standard equivalent to the best performing existing source.
     2.  In WSR manufacturing, the MACT floor was calculated
using the same method as was used in BLR.  The five facilities
with the lowest Ib HAP emitted per MMlb product factors were
averaged together and the resulting value was considered the MACT
floor.  As with BLR. manufacturing, for new sources, the MACT
floors are technology-based and equivalent to the best performing
existing source.
     Table 5-1 presents the MACT floors in tabular form for BLR
and WSR, respectively.  The development of MACT floors for both
source categories is discussed in more detail below.
                               5-2

-------
           TABLE 5-1.    MACT FLOORS,  Ib HAP/MM product
Source category
1. BLR
2. WSR
Process vents,
storage,
wastewater
130
10
Equipment
leaks
570
120
Total
700
130
5.1.1  BLR MACT Floors
     The MACT floors for BLR plants were calculated based on
information supplied by the three facilities that make up the
source category.  Currently, process vent and storage tank
emissions are controlled by scrubbers at Dow and Shell, and a
condenser and flare in series at Ciba-Geigy.  Emission factors
for individual facilities provide the rationale for the MACT
floors but are considered confidential business information  (CBI)
and, as such, cannot be discussed in great detail in this
chapter.  Memoranda describing the development of the MACT floors
are kept in the Emission Standard Division's CBI files.  The
discussion below provides a brief description of the methodology
used for the estimation of the MACT floors.
     Generally, process vent emissions and storage tank emissions
were reported to be controlled using add-on devices (i.e., carbon
adsorbers, condensers,  scrubbers, etc.).  In all cases, reported
emission stream data was used to calculate the quantity of HAP
emissions released at each facility from process and storage
vents.  Wastewater was assumed to have the potential to emit
epichlorohydrin (EPI)  in the wastewater collection and treatment
system.  The USEPA model WATER7 was used to estimate the level of
emissions from wastewater at each of the three facilities.1
Uncontrolled equipment leak emissions were estimated using
Synthetic Organic Chemical Manufacturing Institute (SOCMI)
average emission factors and by factoring typical control
efficiencies of leak detection and repair (LDAR)  programs.2
                               5-3

-------
5.1.2  WSR MACT Floors
     The MACT floors for the WSR plants were calculated based on
data supplied by the nine facilities considered to be major
sources in this source category.  Process vents are generally the
only emission points that are controlled by most plants, usually
with scrubbers or condensers.  Those facilities that are
controlling emissions from process vents with scrubbers reported
high control efficiencies for their devices.  Because no specific
information on wastewater collection and/or treatment was
available, the revised wastewater enabling document was used to
estimate the fraction of EPI in the scrubber liquor that is
expected to volatilize from the effluent.^  This fraction is
estimated to be 29 percent.  In several cases, facilities that
are considered to be area sources have devices achieving the
highest levels of HAP control for process vents and storage tanks
in the industry.  However, these efficiencies could not be
factored into the MACT floors because they are associated with
area sources.  All storage tanks located at major sources are
uncontrolled except one, which is controlled by a water scrubber.
Equipment leak and wastewater emissions in the source category
are not controlled by any facilities.  Again, the equipment leak
HAP emissions were estimated using the SOCMI average emission
factors.4
5.2  REGULATORY ALTERNATIVES AND IMPACTS
     The MACT floors, which approximate the best level .of current
control practices, can be perceived as the least stringent
regulatory alternatives that might be chosen within the
regulatory framework.  Therefore, any control requirement imposed
on these source categories will provide a level of control at
least as stringent as the MACT floors.  Beyond the MACT floors,
regulatory alternatives were chosen to evaluate the costs and
impacts associated with requiring increasing levels of
stringency.  These alternatives are discussed below for each
source category.
                               5-4

-------
5.2.1  BLR Regulatory Alternatives
     The BLR manufacturing regulatory alternatives are presented
below.
     5.2.1.1  Existing Sources.
     5.2.1.1.1  Process vents, storage, and wastewater.  For
process vent, storage, and wastewater emissions, one of the
facilities currently will not meet the MACT floor of 130 Ib
HAP/MMlb production.  The impacts associated with this facility
implementing process changes to meet the MACT floor were
calculated and are presented in Table 5-2.  These costs and
associated emission reductions are based on the replacement of a
flare with a water scrubber and the treatment of the scrubber
effluent at the facility's existing biological wastewater
treatment system.
     In order to establish a more stringent regulatory
alternative than the MACT floor, the current level of control on
all facilities was examined and areas where more control could be
required were identified.  Once the minimum MACT floor
requirement was implemented, all three facilities were found to
be achieving high levels of control; only one area was identified
as a possible additional controlled emission point:  the
generation of HAP-containing scrubber effluent.  The method used
to control this wastewater was recirculation of scrubber effluent
to the process.  This measure reduced emissions down to a maximum
of 125 Ib HAP/MMlb production.  Therefore, the regulatory
alternative beyond the MACT floor corresponds to this mass
emission limit.
     5.2.1.1.2  Equipment leaks.  For the equipment leaks portion
of sources in BLR production, the MACT floor was calculated to be
570 Ib HAP/MMlb production.  Only one facility is below the MACT
floor, and must reduce HAP emissions in order to reach the floor.
This facility must comply by upgrading their current LDAR program
to meet the level required by the equipment leaks negotiated
regulation ("Reg Neg") as adopted from the Hazardous Organic
NESHAP.
                               5-5

-------

Pi
m
pi
o
CO
w
H
|


W
q
2
EH

3 1
C9 CJ °ti
(Do**? V^
£ i§

»-
•K § ^
o > 2
U--SS*
1
CO
8
,_.., (ft
§
s
•*
c*T
W -Ofc
8|
"3 d
"5, v*
U

Emission
reduction,
Mg/yr
o 2"
j3 S *H
J'w Bn
_w 15
"?? _ C3 fcN
S ftj 0 ^.
e S 'jg oo
c §
D
F source
o
. o
'•H
o

g
vi"



§g
^
«« ^

§8
o
t*~ ON
2 2


O ON
2 2




t*" ON

00 00
N fS



a J fc1
^ Co ^ t^ ^S ^3
52 4i ^ ^^ DJO ^^
Jn Oft ^ ^H! 45 C^
O 5^ ^^ .j^ fi^ ^j
£ -2 I i i
B« w !» i i
~

g
«



O O
*•*




^
"-"


«n O
o\ r-




iri oo
•* 00

O O
0 o



•w r. "rt *-(
i U "3 §
6 < So °o
•§" S (2 &
ff , ,
W i i
ri
5-6

-------
     One regulatory option above the MACT floor was developed for
existing sources.  This option requires all three facilities in
the source category to meet the requirements of "Reg Neg" for
equipment leaks.  Table 5-2 also presents impacts for equipment
leaks.
     5.2.1.2  New Sources.  New source MACT floor is the
technology-based equivalent of the level of control achieved by
the best performing source for process vents, tanks, and
wastewater, and is the level of control required by "Reg Neg" for
equipment leaks.  This technology-based requirement will be to
route all emission sources from the process and the storage tanks
to a water scrubber achieving a 99 percent removal of HAP's.  The
scrubber effluent must then be recirculated to the process.
Wastewater is also required to be controlled by 99 percent.
5.2.2  WSR Regulatory Alternatives
     The WSR regulatory alternatives are presented below.
     5.2.2.1  Existing Sources.  For this source category as with
the BLR source category, MACT floors were developed for the
process vent, storage and wastewater portion of the source and
for the equipment leak portion of the source.  The MACT floor for
process vents, storage, and wastewater was estimated to be 10 Ib
HAP/MMlb production, based on the average of the best performing
five sources.  The regulatory alternative beyond the floor was
chosen as 5 Ib HAP/MMlb protection for this source.  As an
alternative means of complying with the process vent,  storage,
and wastewater source MACT floor, facilities may choose to
instead implement the requirements of the LDAR for equipment
leaks.  This alternative means of compliance has been suggested
because the reductions and costs of implementation of the Reg Neg
LDAR are in many cases much more reasonable than those incurred
from requiring compliance with the process vent, storage, and
wastewater regulatory alternatives,  and the emissions reductions
are much greater.
     Since equipment leak emissions were calculated for each
facility from a model component count and scaled on production,
all sources are identical to and therefore meet the MACT floor of
                               5-7

-------
120 Ib HAP/MM product.  The regulatory alternative above the MACT
floor is 7 Ib HAP/MMlb production, a level that corresponds to
the level of control achieved by the Reg Neg LDAR program for a
model component count and model system.  The impacts estimated
for facilities meeting the MACT floor and associated regulatory
alternatives beyond the floors are presented in Table 5-3.
     5.2.2.2  New Sources.  The technology-based new source MACT
floor will be equivalent to running the batch reactor with a
water-cooled condenser (operating at 25°C) and to not generate
methanol or use HC1 or any other HAP to quench the cross-linking
reaction.  As with existing sources, an alternative means of
compliance with the new source MACT floor for process vents,
storage, and wastewater,  facilities will be the implementation of
a Reg Neg LDAR program.
5.3  MODEL PLANTS
     Model plants have been developed for both BLR production and
WSR production.  Because much of the information related to the
BLR manufacturing process has been claimed to be CBI, the model
plant for BLR manufacturing serves to illustrate the process and
the emission sources typical in BLR manufacturing.  The model
plant for WSR serves not only to show the process typically found
in this source category,  but also provides the model component
counts used to calculate impacts of implementing Reg Neg LDAR.
Model plants are described in Appendix G.
5.4  REFERENCES
1.  Hazardous Waste Treatment, Storage, and Disposal Facilities
     (TSDF) Air Emission Models, Second Edition.  Review Draft.
    April 1989.
2.  Internal Instruction Manual for BSD Regulation Development.
    Leaking Equipment--Pumps, Valves, Connectors, Compressors,
    Safety Relief Valves.  July 1992.
3.  Wastewater and Wastes Enabling Document.  Revised
    Version 1.0.
4.  Reference 2.
                               5-8

-------

CO
£
OS
o
fa

lATIVES
1


o;
o
EH -
J
D
0
w
OS

Q o!
a O
< O

CO EH
OS  S
°r
u
!
— j V»
3
3
"^

«"
| g
~-
•2 ^r
'I***

^



1 i &
."> t3 "on
B •§ S
w s
*"

S 8 -
~ .2 ^
VI VI ^
s
•g
« n

*" fb .2 -51
"TH ^ W5 W)
§ * -a S
o S
5 °




u
£
1
t*«
o
a
o
t!
£

§
!«"
tr\



O r*

r* °
tf> f~





2 g








fS tn










00 OO





tT
00
2 8
2 ,2 &-rt
- S u. ts "Z!
W ^^ IT^ w G
o g 
12 t-
g H ^ d
'e ^ ^ '5
§ *^ /^ o*
fi ^ PM w
G. • -
•g -H ts
ff
5-9

-------
              6.0  ENERGY AND ENVIRONMENTAL IMPACTS

     The energy and environmental impacts associated with
applying maximum achievable control technology  (MACT) to HAP
emission sources from basic liquid resins  (BLR) and wet strength
resin  (WSR) manufacturing are presented in this chapter.  Details
of the options selected for each source category are presented in
Chapter 6.
     This analysis includes the national energy burden of
operating the control devices used to meet various regulatory
options for process vents, storage, and wastewater.  Note that no
energy and environmental impacts are associated with equipment
leaks regulatory options.  Solid waste and wastewater impacts
were also evlauated because of the use of a scrubber for one of
the regulatory options.
6.1  ENERGY IMPACTS
     Table 6-1 presents the energy impacts for each of the
process vents, storage, and wastewater regulatory options.  The
energy burden was estimated by calculating the electricity
requirement for refrigerated condensers for WSR manufacturing,
and the electricity requirement for pumps associated with water
scrubbers for BLR manufacturing.
6.2  AIR QUALITY IMPACTS
     No adverse air quality impacts are anticipated due to the
use of the control devices examined for this NESHAP.  Therefore,
no air quality impacts were estimated.
6.3  SOLID WASTE AND WASTEWATER IMPACTS
     Wastewater will be produced from the use of a water scrubber
used for one of the regulatory options.  Table 6-2 presents the
wastewater impacts associated with the options.
                               6-1

-------
                  TABLE 6-1.   ENERGY IMPACTS
Source Category:  BLR Manufacturing

Portion of source
Process vents, storage,
and wastewater
- MACT floor

— Regulatory
Option 1
Uncontrolled
emissions,
Mg/yr


42

42
Baseline
emissions,
Mg/yr


18

18
Emission
reduction,
Mg/yr


7

9

Energy burden,
lO^Btu/yr


1.5

1.5
 Source Category:   WSR Manufacturing
Portion of source
Process vents, storage,
and wastewater
- MACT floor
— Regulatory
Option 1
Uncontrolled
emissions,
Mg/yr

8

8
Baseline
emissions,
Mg/yr

4

4
Emission
reduction,
Mg/yr

2

3
Energy burden,
lO^Btu/yr

4

9
                              6-2

-------
         TABLE 6-2.   SOLID WASTE  AND WASTEWATER IMPACTS
Source Category:   BLR Manufacturing
Portion of source
Process vents, storage,
and wastewater
- MACT floor
— Regulatory
Option 1
Uncontrolled
emissions,
Mg/yr

42

42
Baseline
emissions,
Mg/yr

18

18
Emission
reduction,
Mg/yr

7

9
Wastewater
generated,
GPY

3,200,000

0*
aNo wastewater will be generated this option because the scrubber liquor will be recycled back to
 the process.

Source Category:   WSR Manufacturing

   NONE
                                  6-3

-------
                            7.0  COSTS

     This chapter presents cost estimates for the application of
regulatory alternatives selected for  the source categories
included in the Polymers and Resins II national emission
standards for hazardous air pollutants  (NESHAP).  Control device
models, assumptions, and documentation used in the development of
the costs are explained in detail.  Additionally, the nationwide
impacts of implementing various regulatory alternatives are
contained in this chapter.  The costs for the application of
control technologies to process vents  (i.e., condensation and gas
absorption) and equipment leaks are presented in Sections 7.1 and
7.2.  Section 7.3 contains a summary  of all analyses, and
Appendix H contains the tabular costs for compliance with all the
regulatory alternatives.  Finally, Section 7-4 lists the
references cited in this chapter.
7.1  CONTROL TECHNOLOGIES EXAMINED FOR PROCESS VENTS
     Two technologies, condensation and gas absorption, were
examined for controlling hazardous air pollutant  (HAP) emissions
from process vents in basic liquid resins (BLR) and wet strength
resin  (WSR) manufacturing.  Documentation of the cost models for
each control device, including parameters necessary for sizing
the device, capital and annual costs, and any assumptions made
during the development of the models, is included in this
section.  Sections 7.1.1 and 7.1.2 include documentation of each
of the control device cost models, Section 7.1.3 contains the
costs for auxiliary equipment, and Section 7.1.4 explains the
process modifications available for WSR manufacturers.
                               7-1

-------
7.1.1  Condenser Cost Model
     Condensing HAP's was considered as a method of control for
WSR production.  This section explains the cost model used to
develop the costs for meeting the MACT floor and regulatory
alternatives using condensation.  The parameters necessary for
sizing the condenser, and the capital and annual costs are
included in the following sections.  Also, any assumptions made
during the development of the cost model are included.  The cost
methodology presented below was taken from the OAQPS Control Cost
Manual.1
     7.1.1.1  Parameters Necessary for Sizing the Condenser.
This section details the development of parameters necessary to
size a condenser, including the volume fraction of HAP entering
the condenser, required condenser control efficiency,
noncondensables present in the emission stream, condenser heat
load during events and nonevents, inlet and outlet flowrates, and
required refrigeration.  The development of each parameter is
described below.
     Volume fraction of HAP entering the condenser.  All of the
emission streams that were evaluated for control with
condensation systems were saturated.  Therefore, the volume
fraction of HAP's in the inlet gas stream were estimated assuming
saturation at equilibrium conditions.
     Volume fraction of HAP at condenser outlet.  The volume
fraction of HAP at the condenser outlet is estimated using the
following equation:
          HAPvff  =  [(1 - CE)(HAPvf)]/(l - HAPvf * CE)
where:
      HAPv£f = Volume fraction of HAP at the exit of the
               condenser;
          CE = Control efficiency; and
       HAPV£ = Volume fraction of HAP at the inlet of the
               condenser.

     Uncontrolled and Controlled emissions.  The ideal gas law is
used to estimate emissions on a per unit time basis as follows:
                               7-2

-------
                  Ib/min = (PV/RT)  * HAP *
where :
    P = pressure of emission stream at  the inlet of the
        condenser, mmHg;
    V = volumetric flowrate of emission stream, acfm;
    R = gas constant, 998.97 mmHg ft3/lbmol K;  '
    T = temperature of emission stream  at the  inlet of the
        condenser, K;
MWjj^p = molecular weight  of the HAP; and
HAP   = HAPvf  (volume fraction at condenser inlet or
                (volume fraction at  condenser outlet) .
     Condenser exit flowrate.  The condenser exit flowrate  is
calculated by taking the inlet gas flowrate  (acfm) and
multiplying by the ratio of the volume fraction of
noncondensables at the inlet and outlet of the condenser.   This
value is corrected for standard temperature and pressure at the
condenser outlet.  The equation is given below.
      Exit  FR =  FRiMnoncond.vfi/noncond.vff ) * (Tf/T.j_) * (Pi/Pf )
where :
           FRj_ = inlet flowrate  (acfm) ;
   noncond.vf^ = noncondensable volume fraction at the  inlet of
                 the condenser; and
   noncond.vff = noncondensable volume fraction at the  outlet of
                 the condenser.

     Condenser heat load during events.  The condenser  heat load
consists of sensible and latent heat, which account for the heat
given off by the gas stream as it passes through to the condenser
and is cooled to the condenser outlet temperature.  The heat load
is broken down into three parts:   (1) sensible and latent heat of
condensables that were condensed,  (2) sensible heat of
condensables that were not condensed, and  (3) sensible  heat of
noncondensables.  Each part is broken out below.
     (1)   Sensible and latent heat of condensables that were
condensed are calculated using the following equation:
                               7-3

-------
                    (El -  E2)*[HV + (Cp * AT)]
where:
   El - E2 = amount of HAP's that are condensed, Ib/event;
        El = HAP emissions at the inlet of the condenser,
             Ib/event;
        E2 = HAP emissions at the outlet of the condenser,
             Ib/event;
        Hv = heat of vaporization of the HAP, Btu/lb;
        Cp = heat capacity of HAP, Btu/lb °F, and;
        AT = change in temperature, °F.

     (2)  Sensible heat of the condensables that were not
condensed is calculated as follows:
                          E2  *  (C_ * AT)
     (3)  Sensible heat of noncondensables is calculated as
follows:
    {[cond. exitfr*(l-HAPvff)*Pi]/(RT)} * MWair * Cp air * AT
where:
     cond. exitfr = condenser exit flowrate, scfm.

     Heat load during nonevents.  The heat load during nonevents
(no emission stream going through the condenser) was assumed to
be 10 percent of the heat load during an event  (calculated
above).  Because there is no flow through the condenser unless
there is an emission event occurring,  the system requires only
enough cooling to keep the surfaces cold.
     Tons of refrigeration.  Tons of refrigeration needed for a
year is also calculated.  This value is based on the heat load
required during an event multiplied by the amount of time an
event is occurring, plus the amount of heat load required during
a nonevent (10 percent of heat load during an event) multiplied
by the amount of time there are nonevents.
     7.1.1.2  Capital Costs.  The capital cost specific to the
condenser is the refrigeration unit.  The refrigeration unit cost
is based on the required condenser exit temperature, and the
                               7-4

-------
required refrigeration capacity during an emission event, as
follows2:
       unit cost =  (9.73 -  (0.012 * Tcon ) +  (0.584 * InR)
where:
   Tcon  = temperature of the condenser outlet gas, °F; and
       R = refrigeration capacity, tons  (12,000 Btu/hr).

The above refrigeration unit cost was converted to June 1992
dollars using the Chemical Engineering equipment cost index.
Using the refrigeration unit cost calculated above, the equipment
cost of a packaged refrigeration system can be calculated.  The
equipment cost is estimated to be 25 percent more than the cost
of the refrigeration unit.  Included in the additional 25 percent
is the condenser, recovery tank, and the necessary connections,
piping, and instrumentation.  The purchased equipment cost is
8 percent greater than the packaged equipment cost and includes
sales tax and freight.
     Direct capital costs.  The purchased equipment cost of each
of the control devices is assumed to be 18 percent greater than
the equipment cost.  Direct costs of the control device are
foundation, handling, electrical, piping, insulation, and
painting.  These costs add 30 percent to the purchased equipment
cost.
     Indirect capital costs.  Indirect capital costs add an
additional 15 percent to the purchased equipment cost for
engineering,  startup, construction and field expenses, and
contractor fees.
     7.1.1.3  Annual Costs.   The costs for the electric
compressor motor used in the condensation system is calculated
using the following equation:
                   Ce = R *  E * 0S * pe / 0.85
where:
     Ce = electricity costs  for the compressor, $/yr;
      R = heat load, tons (1 ton = 12,000 Btu/hr);
      E = electricity, kW/ton;
     0S = system operating time, hr/yr; and
                               7-5

-------
     pe = electricity cost, $/kWh.

The electricity requirement used in this equation varies
according to the condenser exit temperature.  Below are the
electricity requirements for various temperature ranges.^
                                       Condenser exit
        Electricity (E.  kW/ton)        temperature. °F
                    1.3                       40
                    2.2                       20
                    4.7                      -20
                    5.0                      -50
                   11.7                      -100

     Direct annual costs.  The direct annual costs are broken
into labor and utilities categories.
     Labor.  The direct costs are operating labor, materials, and
utilities.  Labor is broken up into three types of workers:
operators, supervisors,  and maintenance technicians.   The
operator and maintenance technician costs are based on the number
of shifts per day, and days per year that the control device must
operate.  For the manufacture of BLR, these numbers were assumed
to be 3 and 365, respectively.  The calculation of the operator
and maintenance technician labor costs are as follows:
           Costs ($)  = 0.5  *  A *  shifts/day * days/year
where:
    0.5 = 30 min. per shift; and
      A = rate for operator, $15.64/hr; or rate for maintenance
          technician, $17.21/hr.

The cost for the supervisor is 15 percent of the operator's labor
cost.
     Materials.  Material costs were estimated to be the same as
the maintenance technician's labor costs.
                               7-6

-------
     Utilities.  Electricity costs for the collection fan were
estimated using the OAQPS Control Cost Manual; the electricity
cost is $0.059/kwh.
     Indirect annual costs.  Indirect annual costs include
overhead, administrative, taxes, insurance, reporting and
recordkeeping requirements, and capital recovery.  Overhead is
60 percent of the total labor and materials cost.
Administrative, property taxes, and insurance are 2, 1, and
1 percent of the total capital investment, respectively.
     7.1.1.4  Enhanced Monitoring Costs.  The enhanced monitoring
costs associated with condensation include both capital and
annual elements.
     The capital costs included the purchase of a thermocouple to
measure the outlet gas temperature and a datalogger to record the
measurements.  The cost of these two devices was estimated to be
$3,000.4  The cost of an initial performance test of the
condenser was not factored into the capital costs because the
condenser was only used to estimate impacts in the WSR source
category, and an initial performance test on these batch
processes has not been recommended.
     The annual costs associated with enhanced monitoring
included labor, monitoring supplies, and reporting and
recordkeeping.   The labor cost was estimated to be 0.5 hours per
day multiplied by the number of days per year the process was
operated and the labor rate ($17.50 per hour).5  The monitoring
supplies were assumed to be $500 per year, and the reporting and
recordkeeping requirements were estimated to be $17,500 per
year.6
7.1.2  Gas Absorber Cost Model
     The use of gas absorbers to control process vent emissions
containing EPI was examined.  This section presents the details
of the design and cost calculations for scrubbers.  The
guidelines for these design and costing procedures are contained
in the Handbook of Control Technologies for Hazardous Air
Pollutants.7
                               7-7

-------
     7.1.2.1  Parameters Necessary for Sizing the Gas Absorber.
The major parameter that determines the sizing of the scrubber is
the ratio of moles of scrubber liquid to moles of scrubber gas,
according to the following equation:
                       Laol  = AF  * m *  Gmol
where :
           liquid flowrate,  Ibmoles/hr;
      AF = absorption factor;
       m = slope of equilibrium curve; and
                       / Ibmoles/hr.
     The slope of the equilibrium curve, m, for EPI in water can
be calculated from the Henry's law constant, which is
1.8 atm/mol fraction.
     The Handbook of Control Technologies for Hazardous Air
Pollutants suggests as a rule of thumb that the absorption factor
is generally in the range of 1.25 to 2.0.  Because water
scrubbers are currently in use, an absorption factor ratio was
used based on reported scrubber operating data from an actual
facility.  This ratio is 4.5.  With the variables of AF, Gmol,
and m, known or assumed, I^Q^ may be calculated from the above
equation.
     Dimensions of the scrubber were calculated based on a
fraction of flooding conditions.  The flooding correlation below
is for randomly packed towers.  The abscissa value is calculated
from the equation:
                       Ab -  (LG) (DG/DL)0"5
where :
      L = liquid mass flowrate, Ib/hr;
      G = gas mass flowrate, Ib/hr;
     DG = gas density, lb/ft^; and
     DL = liquid density, Ib/ft  .

Once the abscissa is known, the corresponding ordinate is read
off a chart and input into the following formula to obtain the
gas stream flowrate at flooding:
                               7-8

-------
              Ord =  t(Garea)2  (a/€3)(ML°-2)]/DGDLgc
where :
    G-ar-oa = gas stream flowrate based on column  cross  sectional
     CLi. tiCL   "*
            area  (at flooding conditions), Ib/ft2*sec;
      a, e = packing factors;
       ^L = viscosity of solvent, cp; and
       gc = gravitational constant.

     Garea, is then multiplied by 0.6 to obtain  the optimum Garea
(i.e.,  60 percent of flooding).  The column cross -sectional area
and diameter are then calculated using the following equations:
    Acolumn - G/0600 Garea)
    Dcolumn =
     Column height is calculated using the required removal
efficiency, which was reported in excess of 99 percent  for the
scrubbers used in BLR manufacturing.  The number of transfer
units is calculated using the following equation:
     Nnr  =    InF (HAP0/HAP_) (1 - 1/AF) + X/AFl
      OG              6  1 ? l/AF      -
     where HAPe, HAPQ are the inlet and outlet concentrations,
     respectively.
     HOG, the height of a transfer unit, is calculated  from the
     following mass transfer relationships:
     HQG  =    HG+(1/AF)HL
where :
    HG = [b(3,600 Garea)c/(L»)d] (SCG)0'5
    HL = Y(L»1//iL")s(SCL)0-5
     b,c,d,Y, and s = empirical packing constants  (Packing
constants for 2" Rashig rings were assumed) .
     L" = liquid flowrate,  lb/hr-ft2;
    /iL" = liquid viscosity, lb/ft-hr;
    SCG = Schmidt number for the gas stream; and
    SCL = Schmidt number for the liquid stream.

     The Schmidt number for EPI in air is calculated to be:

                               7-9

-------
     \L    =    (O.lBcpl (2.42 lb/hr-ft/cp)    =    1.47
     pDg       (0.0763 lb/ft3) (0.3871 ft2/hr)

Gas diffusivity,  D , for EPI in air was calculated to be 0.3871
ft2/hr.
     The Schmidt number for the liquid stream is calculated  to
be:
                    ju.    _ (Icp) (2 .42) _
     NSCL "        pDL=  (6.241b/fti) (3.805 X 10'5)    * 1,019

DL the diffusivity of EPI in water, was calculated to be
3.805 X 10"5 ft2/hr.
     Total column height and packing costs are estimated using
the following correlations:

        HTOT = (NOG) (HOG) + 2  + >25 Dcolumn
    vpacking =  (?F/4) (Dcolumn)  x HTcolumn

     Pressure drop through the column is calculated  from  the
following equations:
            P =  (g x 10'8) [10(rL"/DD] (3600 Garea)2/DG
where :
         P = pressure drop, lb/ft2 -ft; and
       g , r = packing constants .

     Total pressure drop through the column is then  expressed  as:
    PTOTAL = PC X HTOT

     7.1.2.2  Capital Costs.  The  costs of absorber  columns,
platforms, and ladders were estimated based on regression
correlations obtained from Figures 5-6 and 5-7 of  the  controlling
Air Toxics Handbook, 1986. 8  For the column,  the cost  equation
is:
                 $ = 1.6 (weight of column)  + 318
                               7-10

-------
The column costs were adjusted to June 1992 dollars using the
Chemical Engineering equipment cost indices.  For the platform
and ladders, the cost is:
                   $  =  1.78  (diameter)  +  1,410.
Packing cost is: $9.7/ft^ (volume packing).

     Purchased equipment costs are made up of the above costs,
plus instrumentation and controls, which are estimated to be
10 percent of the above costs.  Total capital cost is 2.20 times
the purchased equipment costs, according to the 1991 Handbook of
Control Technologies for Hazardous Air Pollutants.
     7.1.2.3  Annual Costs.   Utility costs specific to scrubbers
are water  ($0.30/1,000 gallons), and waste water treatment cost
($2/1,000 gallons).
     7.1.2.4  Enhanced Monitoring Costs.  Enhanced monitoring
capital and annual costs were estimated for gas absorption.
     The capital costs included an initial performance test at an
estimated cost of $24,420.10  Also, instrumentation must be
purchased including a flowmeter and datalogger at an estimated
cost of $3,000.1:L
     The enhanced monitoring annual costs included labor at an
estimated 0.5 hours per day multiplied by the labor rate  ($17.50
per hour)  and the number of days per year the process operated.12
Maintenance materials were estimated to be $500 per year, and
reporting and recordkeeping were estimated to be $17,500 per
year.13
7.1.3  Auxiliary Equipment Costs
     This section describes the cost methodology used to develop
costs for auxiliary equipment.  Auxiliary equipment costs are
part of the capital costs of the control device.  Auxiliary
equipment that is common to both control devices considered in
this analysis include costs of the fans, pumps, and manifolds.
The following paragraphs describe the cost methodology for each
of these auxiliary components.
     Fan costs.   The cost of the fan is expected to be minimal,
based on the expected range of flowrates through the control
                               7-11

-------
devices.  Fan costs are made up of the fan, motor, and starter.
The "following cost equations were used to obtain fan capital
costs.
     Cost of fan: = 57.9 (df^)1*38 (392/363.7)
     Cost of starter and motor =   30.65  [(FR) UP)]0-256
where:
          dfan = diameter of fan, in;
          FR = flowrate of stream, scfm; and
          AP = Pressure drop across fan, in H00.
                                             4b
     The cost of fan was taken from of the 1991 Handbook of
Control Technologies for Hazardous Air Pollutants, and adjusted
in June 1992 dollars using the Chemical Engineering equipment
cost indices.  The starter and motor costs are from the same
reference,  where
                       pmotor -  235 hP°'256
where:
               pmotor = cost of fan motor, belt and starter; and
               hP = horsepower requirement.
     The horsepower requirement for the fan was calculated using
the equation:

     hP   =    1.17 x IP'4 (FR)(AP)
                         •n
where:
    AP = pressure drop through system, in H20;
    FR = flowrate; and
     r/ = efficiency, assumed to be 60 percent.
     Pump costs.  The pump cost was calculated from a regression
correlation developed from Perry's Chemical Engineer's Handbook,
Fifth Edition.14  The pump cost from the regression analysis was
adjusted to June 1992 dollars using the Chemical Engineering
equipment cost indices.
     Manifold costs.  The manifolding costs include one automatic
stainless round damper for the collection main, two elbows per
emission source, and a cost for the estimated length of the
0.25 in. thick stainless round duct (300 ft. was the assumed

                               7-12

-------
length of the collection main).  The cost estimates for each of
these pieces of the manifold are based on the flowrate of the
waste gas stream.  The diameter of the collection main was
calculated using the following equation:
             Velocity =  Flowrate/Cross-sectional area
where:
    Velocity = assumed 2,000 ft/min; and
    Flowrate = flowrate of the waste gas stream, acfm.
Using the relationship between the diameter and the flowrate, the
costs for the manifold equipment were estimated.  The cost for
each automatic damper in June 1992 dollars is:
                $/damper =  (3.45  *  FRA0.5)  +  1,155
The cost per foot for the collection main in June 1992 dollars
is:
                      $/ft  =  0.81  *  FRX0.5
The cost for each elbow in June 1992 dollars is:
                     $/elbow = 1.34 * FR^O.5
The cost of the iranifold equipment is the sum of the automatic
damper cost, the cost per foot of the collection main multiplied
by the length of the collection main (300 ft), and the cost for
two elbows per emission source multiplied by the number of
sources.
7.1.4  Process Modifications
     Process modifications are included as new source regulatory
alternatives for the WSR source category, and are described in
this section.  In WSR manufacturing, methanol (a HAP)  can be
crated as a byproduct during the production of the prepolymer
(eg., dibasic ester and diethylenetriamine create methanol), and
hydrogen chloride (HC1), a HAP, can be used to halt the
crosslinking reaction.  However,  several facilities have
demonstrated that the use of these HAP's in WSR manufacturing is
not essential.   For example, most facilities in the source
category form the prepolymer by reacting a dicarboxylic acid and
diethylenetriamine.   The byproduct of this reaction is water,
instead of methanol.   Also,  several facilities use sulfuric acid
rather than HC1 to' halt  the crosslinking process.  It is for
                               7-13

-------
these reasons that the elimination of methanol and HC1 from the
WSR manufacturing process has been suggested as a regulatory
alternative for new sources.  No costs have been calculated for
these modifications however.
7.2  CONTROL TECHNOLOGIES EXAMINED FOR EQUIPMENT LEAKS
     The control technology used to estimate control of HAP
emissions from equipment leaks is a formal leak detection and
repair  (LDAR) program that meets the requirements of the
Equipment Leaks Negotiated Regulation, or "Reg Neg."  All of the
costs necessary for implementing an LDAR program are described
below and were obtained from the Internal Instruction Manual for
BSD Regulation Development.15
     The cost model required the user to input the number of each
type of component in the plant for BLR manufacturing (a model
component count is used for WSR manufacturing),  the weight
percent of HAP in the stream, and the amount of time each plant
operated (shifts per day and days per year).  The amount of time
each component is expected to be exposed to HAP varied for the
two source categories (8 hr/shift, 3 shifts/day, 365 days/yr for
BLR; 8 hr/shift, 3 shifts/day, 261 days/yr for WSR was assumed).
7.2.1  Capital Costs
     The capital costs of the LDAR program included costs of a
monitoring instrument, compressor vent systems,  pressure relief
devices, caps for open-ended lines, and sample connections.  The
costs for each of these were multiplied by the number of each of
the components and added together.  The cost of each piece of
control equipment is:
     Monitoring instrument =   $6,500/instrument;
     Compressor vent system =  $6,280/system;
     Pressure relief devices = $3,960/device;
     Open-ended lines >=        $102/line; and
     Sample connections =      $413/connection.
7.2.2  Annual Costs
     Two types of annual costs are associated with the LDAR
program:  Capital recovery costs and annual operating costs.
                               7-14

-------
     Capital Recovery Costs.  Capital recovery costs are
calculated by multiplying the capital costs for each piece of
control equipment by the capital recovery factor.  The capital
recovery factor for each piece of control equipment was taken
from the Internal Instruction Manual, adjusted to 7 percent
interest, and is as follows:
     0.55 for pump seals;
     0.21 for monitoring instruments; and
     0.14 for all other equipment.1°
For valves, sample connections, and pumps, the cost of monitoring
and repair also factor into the capital costs.  The monitoring
costs for both liquid and gas valves, sample connections and
pumps includes a subcontractor monitoring fee  ($2.50 per
component), and a 40 percent administrative fee, in addition to
the capital recovery costs described above.  The repair costs
include an initial leak frequency (0.14 for gas valves, 0.065 for
liquid valves, 0.021 for sample connections, and 0.2 for pumps),
the fraction of the total components needing repair (75 percent
for pumps and 25 percent for all other components), four hours
for repair at $22.50 per hour, and a 40 percent administrative
fee.  For pumps there is a cost for replacing seals of $180.00
per seal.
     Annual operating costs.  The annual operating costs are
split into three categories:  maintenance, miscellaneous, and
labor costs.
     Maintenance.   The maintenance cost is 5 percent of the
capital cost for compressors, pressure relief devices, open-ended
lines,  and sample connections.  For monitoring instruments, it is
$4,280, and the annual maintenance cost for pumps is based on the
number repaired.
     Annual miscellaneous costs.  The annual miscellaneous costs
are 4 percent of the capital cost for monitoring instruments,
compressors, pressure relief devices, open-ended lines, and
sample connections.   The cost for pumps is 80 percent of the
annual maintenance cost for pumps.
                               7-15

-------
     Labor.   The annual labor costs are divided into monitoring
and repair costs.  The monitoring costs include component count,
number of monitorings per year (12 for the Reg Neg, except
flanges and pressure relief devices need one monitoring per
year), a $2.00 monitoring fee, and a 40 percent administrative
fee.  For pumps an additional term for weekly visual inspections
is required.  The costs for this term are calculated based on the
pump count,  0.5 min. per pump each week, $22.50 per hour labor
charge, and a 40 percent administrative fee.  The repair cost
includes the component count for gas and liquid valves, pumps,
and sample connections, the leak frequency of each of these
components,  the number of monitorings per year, the fraction
requiring repair, 4 hr for a repair, $22.50 per hour labor, and a
40 percent administrative fee.
     Recovery Credits.  The recovery credit is calculated by
multiplying the quantity of HAP's controlled by the cost per unit
mass of the controlled HAP.  The cost of EPI for BLR
manufacturing, and EPI, methanol, and hydrogen chloride for WSR
                                                    1 7
manufacturing was assumed to be $1,590 per megagram. '
     The total annual costs are the sum of the first year annual
cost and the annual operating cost minus the recovery credit.
7.3  ENHANCED MONITORING COSTS FOR WASTEWATER
     Although no control technologies were evaluated for
wastewater for either source category there are capital and
annual costs associated with the current control technologies
employed in the BLR manufacturing industry.
     The capital cost includes an initial performance test to
demonstrate adequate biodegradation of HAP's in the system at an
estimated cost of $10,000.18  This is the only capital cost
associated with wastewater treatment and hence the only annual
cost associated with wastewater is the capital recovery cost of
the initial performance test.  The annual cost is $1,424 per year
at 7 percent interest and assuming another performance test will
not be needed for 10 years.
                               7-16

-------
7.4  RESULTS OF THE COSTS ANALYSIS
     The nationwide impacts were estimated for both source
categories using the same methodology.  That is, the control
devices were applied, in all cases, to baseline emission streams,
This methodology allowed the emission reductions on the current
level of control to match the control level specified by the
individual regulatory alternatives.  Details of the nationwide
impacts, including capital and annual costs for each regulatory
alternative, and the national costs and incremental cost
effectiveness from baseline is included in Table 7-1 for BLR
manufacturing and Table 7-2 for WSR manufacturing.  The tabular
costs associated with each regulatory alternative for the two
source categories are located in Appendix H to this BID.
                              7-17

-------



CO
EH
ID
rr\
t/J
w
CO
H
CO
1
EH
CO
O
U
S5
H
(V
f-M
B
u
MANUFA
..q
m
•
H
i
W
^





1 ' s
|!l|
•5 SB
l.lf
8 ^
*§

•»
f
O
u
•a
c
a"
QC
O O

1 JH
O



S e1
•2 § fe
S -B ^>
||\
S -S3 ^
DQ E •*
u
1 •
£ §3 -2 -^
|K|S
r> o
2 K





>o oo
^- oo
S §
«M ^— (



o o
oo oo

§ J-
^ ^^ ™ C
.& g ^ <2 U
tu .2 i i

-------









CO
EH

W
CO
H
CO
CO
o
u
CJ
J2J
H
g
EH
U
<
fe
CO
fM
i
M

*c3 co
V

Ills*
St)
« in
o
8"
0)

0 |^
1
V
1
1 -
O Q

3°
'§•**
u




c e"
'% • >>
IP
881.
•Ja .2 -^
S » ">
en 'i ^
ID
1 •
E ^3 *53 "So
® 33 ••§ ^S
Portion of source


°
p">




o o
O ri


O O

0 0
»n vo








r-i m



00 00
jiH !*


^
PI"




g
o r^


o S

00
O f"i








o 2
fl f^
cs ?s



1 1-
L i B
a j§ i i
7-19

-------
7.3  REFERENCES FOR CHAPTER 7

 l.  OAQPS Control Cost Manual,  First Supplement,
     U.  S. Environmental Protection Agency, Research Triangle
     Park, NC.  Chapter 8,  Refrigerated Condensers.  November
     1991.

 2.  Reference 1,  p. 8-21.

 3.  Reference 1,  p. 8-30.

 4.  Memo to R. McDonald,  CPB, from B. Shine, MRI, Polymers and
     Resins II NESHAP,  Enhanced Monitoring Costs, July 14, 1993.

 5.  Reference 4.

 6.  Memo to D. Stackhouse, SDB, from S. Angyal, MRI Magnetic
     Tape NESHAP--Format of Standard, Enhanced Monitoring
     Requirements  and Costs, July 26, 1993.

 7.  Handbook of Control Technologies for Hazardous Air
     Pollutants, U. S.  Environmental Protection Agency,
     Cincinnati, OH.  Publication No. EPA-625/6-91-014.
     June 1991.

 8.  Handbook of Control Technologies for Hazardous Air
     Pollutants, U. S.  Environmental Protection-Agency, 1986.

 9.  Reference 7,  p. 4-52.

 10. Reference 4.

 11. Reference 4.

 12. Reference 4.

 13. Reference 6.

 14. Perry, Robert H.  Perry's Chemical Engineer's Handbook,
     Fifth Edition.  New York, McGraw Hill.  1973.  Figure 6-5.

 15. Internal Instruction Manual for ESD Regulation Development,
     Leaking Equipment--Pumps, Valves, Connectors, Compressors,
     Safety Relief Valves,  U. S. Environmental  Protection Agency,
     Research Triangle Park, NC.  July 1992.

 16. Reference 15, p. 4-16.

 17. Memo, to D. Markwordt, CPB, from D. Whitt, Radian, HON
     Equipment Leaks, May 20, 1991.

 18. Reference 4.
                               7-20

-------
                  8.0  ECONOMIC IMPACT ANALYSIS

8.1  INTRODUCTION

     The U.S. Environmental Protection Agency  (EPA) is reviewing
National Emission Standards for Hazardous Air  Pollutants  (NESHAP)
for the basic liquid epoxy resin  (BLR) and wet strength resin
(WSR)  industries.  These industries emit several of the hazardous
air pollutants  (HAPs) identified by the Clean Air Act Amendments
of 1990.!
     Section 317 of the Clean Air Act requires EPA to evaluate
regulatory alternatives through an Economic Impact Analysis
(EIA).   Accordingly, this EIA has been conducted to satisfy the
requirements of the Clean Air Act.
8.1.1  EIA Objectives
     There are two primary objectives of this  EIA.  The first
objective is to describe the distribution of adverse impacts
associated with the NESHAP among various members of society.  The
second objective is to adjust estimated emission control  costs so
that these reflect the economic costs associated with the
standard.
     Neither the benefits nor the costs associated with the
NESHAP will be distributed equally among different members of
society.  Since this study is focused on costs, emphasis  is
placed on estimating and describing the adverse impacts
associated with the NESHAP.  Those members of  society who could
potentially suffer adverse impacts include:
     •     Producers whose facilities require emission controls.
     •     Buyers of goods produced by industries requiring
          controls.
    1 These HAPs are MeOH, HC1 and EPI.

                               8-1

-------
     •    Employees at plants requiring controls.
     •    Individuals who could be affected indirectly such as
          residents of communities proximate to controlled
          facilities, and employees of industries that sell
          inputs to or purchase inputs from directly affected
          firms.
Because of potential distributional impacts, and because of other
policy issues, impacts on both energy consumption and foreign
trade are also considered in this study.
     Economic costs generally do not correspond to emission
control costs because the latter do not reflect market
adjustments that occur because of higher production costs caused
by the installation, operation and maintenance of emission
controls.  A second purpose of this EIA is to make appropriate
adjustments to estimated emission control costs so that they
reflect the economic costs of the NESHAP.
8.1.2  Background
     8.1.2.1  Affected Markets.  EPA expects the NESHAP to affect
two of the industries included in Standard Industrial
Classification code 2821.  They are:
     •    Basic Liquid Epoxy Resin  (Diglycidyl Ether of Bisphenol
          A or DGEBPA).
     •    WSR (Epichlorohydrin Cross-Linked Non-Nylon polyamide
          resins).
     8.1.2.2  Regulatory Alternatives.  The Clean Air Act
Amendments of 1990 stipulate that HAP emission standards for
existing sources must at least match the percent reduction of
HAPs achieved by either a)  the best 12 percent of existing
sources,  or b) the best five sources in a category or subcategory
consisting of fewer than 30 sources.  This minimum standard is
called a MACT floor.
     The NESHAP considered in this EIA is the MACT floor for WSR
plants.  The MACT floor for these plants requires controls on
storage tanks and process vents.  This EIA considers an
alternative to the MACT Floor for WSR plants.  This alternative,
which we refer to as Option I,  requires controls only on
equipment leaks.  The NESHAP for BLR plants is the MACT floor for

                               8-2

-------
storage tanks and process vents, but requires more  stringent
controls than the MACT floor for equipment leaks.
     Both the BLR and WSR industries consist of  fewer  than
30 sources.  Thus, definition b) was used in both cases  to
construct the MACT floor for existing sources.   For new  sources
the Amendments stipulate that the MACT floor be  set at the
highest level of control achieved by any similar source.
     There are currently three facilities producing substantial
amounts of BLR.  The MACT floor for existing sources was
constructed by averaging the percentage reduction of HAPs
achieved for each source type by each facility.  A  source  type is
a piece of equipment or component of production  which  produces
HAPs.  The MACT floor requires controls on the following BLR
source types:  process vents, storage tanks and  equipment  leaks.
As noted above, the NESHAP considered in this EIA requires
controls at BLR plants more stringent than the MACT floor  for
equipment leaks.
     There are 17 existing WSR plants.  The MACT floor was
constructed by averaging the percentage reduction of HAPs
achieved by the five best controlled sources for each  source
type.  The MACT floor for WSR plants requires controls on  storage
tanks and process vents, but no additional controls on equipment
leaks.
     As noted earlier, Option I, which is an alternative to the
MACT Floor for the WSR industry, requires controls  on  equipment
leaks, but no controls on either storage tanks or process  vents.
We consider Option I because it results in larger emission
reductions at considerably lower costs than the  MACT floor.
8.1.3  Summary of Estimated Impacts
     8.1.3.1  Primary and Secondary Impacts.  Table 8-1
summarizes the estimates of the primary and secondary  economic
impacts associated with the NESHAP.^  Primary impacts  include
price increases, reductions in market output levels, changes  in
    2 Table 9-1 summarizes the results of the MACT floor for the wet strength
resin industry.  We describe the impacts of Option I in the text that follows.

                               8-3"            •  •

-------
     TABLE 8-1.  SUMMARY OF ESTIMATED ECONOMIC IMPACTS
     Analysis
             Estimated Impacts
Primary Impacts

Price Increases


Market Output
Value of Domestic
Shipments


Plant Closures
Secondary Impacts

Employment
Estimated price increases are 0.05 percent
for BLR and 4.19 percent for WSR.

Estimated reductions in market output are
0.08 percent for BLR and 3.73 percent for
WSR.

Estimated changes range from a decline of
0.03 percent for BLR to an increase of
0.31 percent for WSR.

No plant closures are expected in the BLR
industry and one plant closure is predict-
ed for the wet strength resin industry.
Energy Use
Net Exports
Regional Impacts
No significant employment losses are ex-
pected.

Estimated industry-wide use to decline by
0.08 percent  ($8,500) in the BLR industry
and by 3.73 percent  ($45,000)  in the wet
strength resin industry.

Estimated trade impacts are small.   Net
exports of BLR predicted to decline by
about $25,000.  Lower volume of wet
strength resin exports expected to be off-
set by higher post-control prices.

No significant regional impacts are ex-
pected.	
                            8-4

-------
the value of shipments by domestic producers, and plant closures.
Secondary impacts include employment losses, reduced energy use,
changes in net exports, and potential regional impacts.
     The estimated primary impacts on the BLR market are small.
For example, we estimate that the market price will increase by
just 0.05 percent, and that market output will fall by about
0.08 percent.  The estimated impacts of the MACT Floor on price
and output in the WSR market are somewhat larger than those for
the BLR market.  We estimate an increase in price of 4.19 percent
and a decrease in U.S. production of 3.73 percent.  Note,
however, that we expect a slight increase in the value of ship-
ments by domestic WSR producers.  This occurs because the
estimated price increase more than offsets the lower production
volume.  Our analysis predicts no plant closures in the BLR
industry, but one WSR plant closure is possible.  This predicted
closure, however, may be due to some "worst case" assumptions
adopted in our analysis.
     The analysis of the primary impacts on the WSR market under
the implementation of Option I yields substantially less adverse
impacts than the MACT floor.  The increase in WSR price is
estimated at 0.22 percent (compared to a 4.19 percent increase
under the MACT floor).  We estimate that market output will
decrease by just 0.20 percent under Option I, with an associated
increase (due to the slight increase in price) in the value of
domestic shipments of $7,000 (0.02 percent).  While one plant
closure is possible under the MACT floor, none is expected under
Option I.
     The estimates of secondary impacts reported in Table 8-1
follow the estimates of primary impacts described above.  We
expect only small employment losses and reductions in energy use.
These findings, of course, are consistent with our estimates of
small impacts on market output.  We estimate that the reduction
in net exports of BLR will be small, about $25,000, and that
higher post-control prices will offset a slightly lower volume of
WSR exports.  Finally, we expect no significant regional impacts.
                               8-5

-------
     The secondary impacts of Option I on the WSR industry are
also smaller than those of the MACT Floor.  Employment  (pro-
duction job) losses are almost negligible  (0.10 production jobs)
and energy use is expected to decline by 0.20 percent.  The
estimated trade impacts are negligible.  WSR exports are
estimated to fall by  .21 percent.  Also, no significant regional
impacts are expected.
     8.1.3.2  Financial Analysis.  Our financial analysis
indicates that capital and annual emission control costs are
small relative to the financial resources of the firms producing
BLR and WSR.  As a result, we do not expect that it will be
difficult for these firms to raise the capital required to
purchase and install emission controls.
     8.1.3.3  Sensitivity Analyses.  In Appendix D of the report,
we examine the sensitivity of the estimated primary impacts to
our estimates of market demand elasticities.  The results
reported in Appendix D indicate that the primary impacts
summarized in Table 8-1 are relatively insensitive to reasonable
ranges of demand elasticity estimates.  However, analysis
conducted assuming a "low" elasticity of demand yields slightly
less adverse impacts, including no plant closures in the WSR
industry.
     8.1.3.4  Potential Small Business Impacts.  All of the
affected BLR and WSR producers are large companies and none
satisfies the criteria for a small business.  Consequently, we do
not expect any significant small business impacts to result from
implementing the NESHAP.
     8.1.3.5  Economic Costs.  Table 8-2 reports estimates of the
economic costs associated with the NESHAP.  The estimated
annualized economic costs are $120 thousand for the BLR industry
and $465 thousand for the WSR industry under the MACT floor.  The
economic costs associated with Option I, $51 thousand, are
considerably lower than those of the MACT floor..  These estimates
measure changes in economic surplus and include the costs
associated with higher prices of imports to the U.S. economy.
                               8-6

-------
 TABLE 8-2.   ESTIMATES OF ANNUALIZED ECONOMIC COSTS
             (thousands of 1992 dollars)
Industry
BLR
Wet Strength Resin
MACT Floor
Option I
Loss in
Consumer
Surplus
141
1,607
87
Loss in
Producer
Surplus
-3
-841
-22
Loss in
Residual
Surplus
-19
-300
-13
Loss in
Surplus
Total
120
465
51
Economic costs are computed as the change in economic
surplus associated with the NESHAP.  The estimates
include the costs of higher prices of imported products
                         8-7

-------
8-1-4  Organization of EIA
     We describe the analytical methods employed to estimate the
economic impacts associated with the NESHAP in Section 8-2.
Section 3 contains profiles of the two affected industries.  We
report in Section 8-4 estimates of primary economic impacts,
including those on market prices, market output levels, value of
shipments by domestic producers, and plant closures.  Section 8-5
presents estimates of secondary impacts, including the effects on
employment, foreign trade, energy use and regional economies.  We
describe potential adverse impacts of small businesses in
Section 8-6.  In Section 8-7, we report estimates of the economic
costs associated with the NESHAP.
     There are four appendices to this section.  We describe the
model plants used in the analyses and report estimates of
emission control costs and other baseline data in Appendix A.
Appendix B provides a detailed technical description of the
analytical methods employed to estimate economic impacts and
costs.  We describe an econometric model of the resin industry in
Appendix C.  We report in Appendix D the results of sensitivity
analyses in which we consider ranges of demand elasticity
estimates.
8.2  OVERVIEW OF ECONOMIC IMPACT ANALYSIS
     We assess the economic impacts associated with the NESHAP by
conducting studies of the affected industries.  These industries
are the BLR and the WSR.  We describe the analytical methods
employed in these studies below.
8.2.1  Overview of Distributional Impacts
     As noted earlier in the introduction to this section,
several groups might potentially suffer from adverse impacts
associated with the NESHAP.  These groups include:
     •    Resin producers.
     •    Resin buyers.
                               8-8

-------
     •    Employees at affected plants.
     •    Individuals affected indirectly by the NESHAP.
We describe the potential adverse impacts affecting each of these
groups below.
     8.2.1.1  Impacts on Producers.  The emission control costs
associated with the standard are likely to reduce the
profitability of at least some of the affected plants.  Indeed,
some affected plants may be forced to shutdown operations in the
face of emission control costs.  Ultimately, the magnitude of the
adverse impacts incurred by affected plants will depend on the
extent to which emission control costs can be passed on to
buyers.  In addition, 'operators of some affected plants might
have difficulty acquiring the capital necessary to purchase and
to install emission control equipment.
     Some plants in affected industries may not suffer adverse
impacts as a result of the implementation of an emission control
standard.  The post-control profitability of an affected plant
will improve if post-control price increases more than offset the
plant's emission control costs.  This could occur if control
costs for some plants are substantially higher, per unit of
output, than those for other plants in the industry.
     8.2.1.2  Impacts on Consumers or Buyers.  Both BLR and WSR
are purchased primarily by firms which use these products as
inputs to produce other goods.  These firms and the consumers of
the goods which they produce are likely to suffer from two
related adverse impacts.  First, post- control prices for resins
produced at the affected plants are likely to be higher as
sellers attempt to pass through some of the costs of emission
controls.  This will cause profits to be smaller, at least in the
short run, for firms which purchase BLR and WSR as inputs.  It
will also cause prices of final goods to be higher as firms
attempt to pass through some of the increase in production costs.
Second, the shift in supply caused by emission control costs is
likely to reduce the amount of resin sold in affected markets, as
well as the level of output sold in markets which use the resin

                               8-9

-------
as an input.  These two effects are related in that post-control
equilibrium prices and output levels in affected markets will be
determined simultaneously.
     8.2.1.3  Indirect or Secondary Impacts.  Two countervailing
impacts on employees of affected plants are likely to result from
the implementation of the NESHAP.  Employment will fall if
affected plants either reduce output or close operations
altogether.  On the other hand, increases in employment
associated with the installation, operation, and maintenance of
emission controls are likely.
     A number of other indirect or secondary adverse impacts may
be associated with the implementation of a standard.  The
indirect impacts we consider in this study include:  foreign
trade effects; impacts on regional economies; and, effects on
energy consumption.
8.2.2  Economic Impact Studies
     The industry segment studies that follow in this report
        •
include six major components of analysis.  These components or
phases of analysis, which are designed to measure and describe
economic impacts, are:
     •    Industry profile.
     •    Direct impacts  (market price and output, domestic
          production and plant closures).
     •    Capital availability analysis.
     •    Evaluation of secondary impacts (employment, foreign
          trade,  energy consumption, and regional and local
          impacts).
     •    Analysis of potential small business impacts.
Each of these phases of analysis is described below.
8.2.3  Industry Profile
     The industry profile provided in Section 8-3 describes
conditions in affected industries that are likely to determine
the nature of economic impacts associated with the implementation
of the NESHAP.  We discuss the following seven topics in the
industry profile:

                               8-10

-------
     •    Product descriptions.
     •    Prices and output.
     •    Market outlook.
     •    Market structure.
     •    Foreign trade.
     •    Financial conditions.
     •    Employment and energy  use.
8.2.4  Primary Impacts
     We employ a partial equilibrium model of the BLR and WSR
industries to estimate the primary impacts of emission control
costs, including market equilibrium price, market output, the
value of domestic shipments,  and the number of potential plant
closures.3  This analysis is  so  named because the predicted
impacts are driven by estimates  of how the affected industries
achieve market equilibrium after the air quality standard is
implemented.
     In a competitive market,  equilibrium price and output are
determined by the intersection of demand and supply.  The supply
function is determined by the marginal (avoidable) operating
costs of existing plants and  potential entrants.  A plant will be
willing to supply output so long as market price exceeds its
average  (avoidable) operating costs.   The installation,
operation, and maintenance of emission controls will result in an
increase in operating costs.   An associated upward shift in the
supply function will occur.
     The procedures employed  in  the market analysis are
illustrated in Figure 8-1.  Constructing the model and predicting
impacts requires completing the  following four tasks.
     •    Estimate pre-control market demand and supply
          functions.
     •    Estimate per unit emission control costs.
     •    Construct the post-control supply function.
    3 The results of the partial equilibrium analyses are also used to esti-
mate employment,  energy and foreign trade impacts and the economic costs
associated with the regulatory alternatives.

                               8-11

-------
    Pre-Control
    Market Data
Specify Demand and
 Supply Functions
    Estimate Pre-Control
     Demand and Supply
    Emissions
  Control Costs
 Discounted Cash
 Flow Parameters
    Estimate per Unit
        Emissions
      Control Costs
                             Construct
                           Post-Control
                          Supply Function
                                                  Solve for Post-Control
                                                   Price and Output, and
                                                      Predict Closures
     Figure 8-1.
Partial Equilibrium Analysis of  BLR and
      WSR Industries.
                                8-12

-------
     •    Solve  for post-control price,  output and employment
          levels,  and predict plant closures.

We briefly des-cribe each of these tasks  below,4
     8.2.4.1   Pre-Control Market Demand  and Supply Functions.
Pre-control equilibrium price and output levels in competitive
markets are determined by market demand  and supply.  Because
estimates of  demand and supply for the relevant industries are
unavailable from the literature, we estimated  these functions as
part of this  study.  Both the market demand and domestic supply
functions were estimated econometrically using time-series
data.5
     Market demand in the household segment was specified as a
function of product price and a time trend to  capture structural
change in demand over time.   However, because of uncertainty
regarding the demand elasticity estimates,  we  report the results
of sensitivity analyses in Appendix D.
     Market supply includes domestic supply and foreign supply of
imports.  We  derived our estimate of domestic  supply elasticity
from a production function in which output of  is expressed as a
function of capital stock held by the industry, material and
labor inputs,  and time.  We assume, in the absence of other
information,  that the supply elasticity  of imports (foreign
supply) is the same as that for domestic supply.7
     8.2.4.2   Per Unit Emission Control  Costs.  Emission control
costs will cause an upward vertical shift of the supply curves in
affected markets.   The height of the vertical  shift for each
    4 See Appendices A, B, and C for more detailed descriptions of the  data
and methods employed in the partial equilibrium analysis.

    5 See Appendix C for detailed descriptions of the data and methods
employed to estimate these functions.  Appendix C also reports the estimated
parameters of the functions.
    6 Our estimates of demand elasticities are -1.50 for BLR and -0.92  for
WSR.
    7 Given this assumption, a one percent change in price causes the same
percentage increase in both domestic and foreign supply.

                                8-13

-------
affected  plant is given by the  after-tax cash flow required to
offset  the  per unit increase  in production costs resulting from
the installation, maintenance,  and operation of emission control
equipment.
     Estimates of the capital,  operating and maintenance costs
associated  with emission control equipment for affected  plants
were obtained from the draft  BID document.  Per unit,  after-tax
costs are estimated by dividing after-tax annualized  costs by
annual  output.  This cost reflects the offsetting cash flow
requirement which, in turn, yields an estimate of the post-
control vertical shift in the supply function.
     Computing per unit after-tax control costs requires,  as
inputs, estimates of the following parameters:
     •    The useful life emission control equipment.
     •    The discount rate  (marginal  cost of capital).
     •    The marginal corporate income tax rate.
Estimates of the expected life  of emission control equipment were
                                             •
obtained  from the draft BID document.   The results presented in
this report are based on a 10 percent  real private discount
rate8 and a 25 percent marginal tax rate.
     8.2.4.3  The Post-Control  Supply Function.  Estimated
aftertax  per unit control costs are added to pre-control supply
prices  to determine the post-control supply prices for domestic
producers.   We construct the  post-control domestic supply
function  by sorting affected  plants,  from highest to  lowest, by
per unit  post-control costs.  We assume that plants with the
highest per unit emission control costs are marginal  (highest
cost) in  the post-control market.9  Because per unit  control
      The discount rate referred to here measures the private marginal cost
of capital to affected firms.  This rate, which is used to predict the market
responses of affected firms to emission control costs, should be distinguished
from the social cost of capital.   The social cost of capital is used to mea-
sure the economic costs of emission controls.  See Section 9.7 for a more
detailed discussion of this issue.
    ^ Note that any other construction of the post-control supply curves
would result in the same or smaller vertical shifts in supply, and according-
ly, the same or smaller economic  impacts.

                                8-14

-------
costs differ across affected plants  within an industry segment,
the post-control domestic supply  function is segmented.  Total
market supply is given by the  sum of domestic and foreign supply.
We assume, of course, that  foreign supply is unaffected by
emission controls.11^
     8.2.4.4  Post-Control  Prices. Output,  and Closures.   The
baseline, pre-control equilibrium output  in an affected market is
taken as the level of observed national  consumption (shipments by
domestic producers minus net exports).  We compute post-control
equilibrium price and output levels  in affected markets by
solving for the intersection of the  market demand curve and the
market post-control, segmented supply curve.  The estimated
reduction in market output  is  given  by the difference between the
observed pre-control output level and the predicted post-control
output level.  Similarly, the  estimated  increase in price is
taken as the difference between the  observed pre-control price
and the predicted post-control equilibrium price.
     Because higher market  prices lead to higher imports, the
reduction in domestic production  is  larger than the reduction in
market output.  Specifically,  the reduction in output for
domestic producers is given by the reduction in market output
plus the increase in imports.  We estimate the number of plant
closures by dividing the predicted reduction in domestic output
by the production levels at plants with  post-control supply
prices higher than the post-control  equilibrium market price.
     8.2.4.5  Reporting Results of Market Analyses.  The results
of the partial equilibrium  market analyses for each of the
affected industries are presented in Section 8-4 of this report.
In particular, estimates of the following are reported:
     •    Price increase.
     •    Reduction in market  output.
     •    Annual change in  the value of  domestic shipments.
    10 This assumption means that no shift in the foreign supply function
occurs as a result of emission controls on domestic producers.  The quantity
supplied by foreign producers, however, increases as market price increases.

                             .  8-15

-------
      TABLE 8-3.   BIASES RESULTING IF MODEL ASSUMPTIONS
      	ARE VIOLATED	

                                               Direction of Bias
                                       Change in   Change in
                                        Industry    Industry      Net
 Assumption  	.	Quantity	Price	Closures

   i)    national market                    +           *          +

  ii)    controlled plants at margin        +           +          +
        in baseline

 iii)    no regulation-induced expan-        +           +          +
	sion of domestic producers	

     Price changes  will  vary depending on the locations  of affected
     plants and the levels of regional trade barriers  and degree of
     product differentiation.
                                 8-17

-------
     •    Number of plant closures.
           V
     8.2.4.6  Limitations of the Market Analysis.  The partial
equilibrium model has a number of limitations. First, a single
national market for homogeneous output is assumed in the
analysis.  However, markets may be regional.  Then each region or
product type will be affected primarily by cost changes of plants
in the region, rather than all plants in the national market.
Output reductions and price effects will vary across regions
depending on locations of affected plants.  In addition, the
assumption of a national market is likely to cause predicted
closures to be overstated to the extent that affected firms are
protected somewhat by regional trade barriers.
     Second,  the analysis assumes that plants with the highest
per unit emission control costs are marginal post-control.  This
assumption produces an upward bias in estimated effects on
industry output and price changes because the control costs of
non-marginal firms will not affect market price.  Predicted
closures will also be overstated.
     Third, the analysis assumes that the implementation of
controls does not induce any domestic producers to expand
production.  An incentive for expansion would exist if some
plants have post-control incremental unit costs between the
baseline price and the post-control price predicted by the
partial equilibrium analysis.  Expansion by domestic producers
will result in reduced impacts on industry output and price
levels.  While plant closures will increase as expanding
producers squeeze out plants with higher post-control costs, net
closures (closures minus expansions) will be reduced.
     Table 8-3 summarizes the biases discussed above.  In most
cases, the assumptions embedded in the market analyses produce an
upward bias in estimated impacts on market quantity, market
price, and net closures.
     Also,  statistical errors in the estimated- demand and supply
functions exist.  We report the statistical properties of the
                               8-16

-------
estimates of these  functions  in Appendix C.11  In addition, it
is likely that uncertainty  in the estimates of emission control
costs exist, causing  control  costs for some plants to be either
overstated or understated.  The control costs used in this EIA are
study estimates and are  accurate within plus or minus 30 percent.
8.2.5  Capital Availability Analysis
     We assume in the market  analysis that affected firms will be
able to raise the capital associated with controlling emissions at
a specified marginal  cost of  capital.  The capital availability
analysis, on the other hand,  examines the variation in firms'
ability to raise the  capital  necessary for the purchase,
installation, and testing of  emission control equipment.
     The capital availability analysis also serves three other
purposes.  First, it  provides information for evaluating the
appropriateness of  the selected discount rate as a proxy for the
marginal cost of capital of the industry; implications for bias in
the partial equilibrium  analysis follow.  Second, it provides
information on potential variation in capital costs across firms.
Third, it provides  measures of the potential impacts of controls
on the profitability  of  affected firms.
     8.2.5.1  Evaluation of Impacts on Capital Availability.  For
each model plant1^  included in the capital availability
analysis, the impact  of  the alternative standards on the following
two measures is evaluated:
     •    Net income/assets.
     •    Long-term debt/long-term debt and equity.
Net income is measured before-tax and is defined to include all
operations, continued and discontinued.
     The ratio of net income  to assets is a measure of return on
investment.  The implementation of emission controls is likely to
reduce this ratio to  the extent that net income falls
    11 See Appendix D for estimates of impacts associated with alternative
estimates of demand elasticities.
    12 The model plants included in the analysis are described later in this
section and in more detail in Appendix A.

                               8-18

-------
(e.g., because of higher operating costs)  and assets increase
(because of investments in emission control equipment).
     The ratio of long-term debt to long-term debt plus equity is
a measure of risk perceived by potential investors.  Other things
being the same, a firm with a high debt-equity ratio is likely to
be perceived as being more risky, and as a result, may encounter
difficulty in raising capital.  This ratio will increase if
affected firms purchase emission control equipment by issuing
long-term debt.
     8.2.5.1.1  Baseline values for capital availability analysis.
Baseline values for net income and net income/assets are derived
by averaging data for as many years as are available between 1988
and 1991.  Data from these four years are employed to reduce
distortions caused by year-to-year fluctuations.  Since changes in
the long-term debt ratio represent actual structural changes,
1990 or 1991 data are used, whichever is the most recent year the
data are available.
     8.2.5.1.2  Post-control values for capital availability
analysis.  Post-control values for the two measures identified
above are computed to evaluate the ability of affected firms to
raise required capital.  The post control values are computed as
follows:
     •    Post-control net income — pre-control net income minus
          the after-tax annualized costs associated with the
          purchase, installation, maintenance and operation of
          emission control equipment.
     •    Post-control return on assets — post-control net income
          divided by the sum of pre-control assets plus
          investments in emission control equipment.
     •    Post-control long-term debt ratio — the sum of
          precontrol long-term debt plus investments in emission
          control equipment divided by the sum of pre-control
          long-term debt, equity, and investments in emission
          control equipment. .
     The calculations are done for a worst-case scenario of the
impact of controls on the measures.  First, the total investment
in emission control equipment is assumed to be debt-financed.
                               8-19

-------
Second, it is assumed that there  is no  increase  in the  price a
company receives for its output.
     8.2.5.1.3  Limitations of the capital-availability analysis.
The capital availability analysis has limitations.   First,  future
baseline performance may deviate  from past  levels.   The financial
position of a firm during the period 1988-1991 may not  be  a good
approximation of the company's position later during the
implementation period, even in the absence  of the  impacts  of
emission control costs.
     Second, a limited set of measures  is used to  evaluate the
impact of controls.  These measures reflect accounting
conventions and provide only a rough approximation of the  factors
that will influence capital availability.
8.2.6  Evaluation of Secondary Impacts
     The secondary impacts that we consider in this study
include:
     •    Employment impacts.
     •    Energy impacts.
     •    Foreign trade impacts.
     •    Regional impacts.
     8.2.6.1  Employment Impacts.  As equilibrium  output in
affected industry segments falls because of control costs,
employment in the industry will decrease.   On the  other hand,
operating and maintaining emission control  equipment requires
additional labor for some control options.   Direct net  employment
impacts are equal to the decrease in employment  due to  output
reductions, less the increase in employment associated  with the
operation and maintenance of emission control equipment.
     Our estimates of the employment impacts associated with the
NESHAP are based on employment-output ratios and estimated
changes in domestic production.  Specifically, we  compute  changes
in employment proportional to estimated changes  in domestic
production.^
    13 See Appendix B for descriptions of the data and methods used to
estimate employment impacts.
                               8-20

-------
     Estimates of  the  labor hours required to operate and
maintain emission  control  equipment are unavailable.
Accordingly, the employment impacts presented in this report are
overstated to the  extent that potential employment gains
attributable to operating  and maintaining control equipment are
not considered.
     The estimates  of  direct employment impacts are driven by
estimates of output reductions obtained in the market analyses.
Biases in these estimates  will likely cause the estimates of
employment impacts  to  be biased in the same direction.
     8.2.6.2  Energy Effects.   The energy effects associated with
the NESHAP include  reduced energy consumption due to reduced
output in affected  industry segments plus the net change in
energy consumption  associated with the operation of emission
controls.
     The method we  use to  estimate reduced energy consumption due
to output reductions is similar to the approach employed for
estimating employment  impacts.14  Specifically, we assume that
changes in energy use  are  proportional to estimated changes in
domestic production.   Estimates of the net change in energy
consumption due to  operating emission controls are
unavailable.15
     8.2.6.3  Foreign  Trade Impacts.  Other factors being the
same, the implementation of the NESHAP will raise the production
costs of domestic  resin manufacturers relative to foreign
producers, causing  U.S. net exports of resin to decrease.
     The extent to  which imports to U.S.  increase will depend on
the supply elasticity  of foreign-produced resin to the U.S. Unfor-
tunately, we have not  identified any estimates of resin import
supply elasticities in the literature and the available data does
not permit us to derive our own estimate.  Accordingly, we assume
       See Appendix B for a more detailed description of this procedure.
    15 We view these as short-run estimates of reduced energy consumption.
In the long run,  resources diverted from the production of BLR and WSR will
likely be directed to producing other goods and services.
                               8-21

-------
that the import supply elasticity is the same as that for
domestically produced resin.
     We report estimates of the dollar value of the increase in
imports associated with the implementation of the standard.
There are two sources of this increase:  (1) the increase in the
quantity of goods imported; and  (2) increases in prices of
imported goods.  The estimates we report reflect the
contributions from both sources.
     8.2.6.4  Regional Impacts.  Substantial regional or
community impacts may occur if a plant that employs a significant
percent of the local population or contributes importantly to the
local tax base is forced to close or to reduce output because of
emission control costs.
     Secondary employment impacts may be generated if a
substantial number of plants close as a result of emission
control costs.  Secondary employment impacts include those
suffered by employees of firms that provide inputs to the
directly affected industry, employees of firms that purchase
inputs from directly affected firms for end-use products, and
employees of other local businesses.
8.2.7  Affected Plants
     The NESHAP is expected to affect three BLR and 17 WSR
plants.  Because only three BLR plants are affected, the analysis
considers plant specific data for this industry.  However,
because of the large number of affected WSR facilities, the
analysis is based on three different model plants that have been
developed to represent the 17 plants in the industry.  Appendix A
describes the characteristics of the affected BLR and WSR plants.
8.3  INDUSTRY PROFILE
     This section describes market conditions for products that
will be affected by the NESHAP.  The affected products are BLR
made from diglycidyl ethers of bisphenol A  (DGEBPA),  which is a
type of unmodified epoxy resin, and WSR, which are
epichlorohydrin based non-nylon polyamide resins.
     We cover the following topics in this industry profile:
     •    Product descriptions and end uses.
                               8-22

-------
     •    Market structure.
     •    Market trends and outlook.
     •    Foreign trade.
 _   •    Financial conditions.
8.3.1  Product Descriptions and End Uses
     8.3.1.1  Basic liquid resins (BLR).  Diglycidyl ether of
bisphenol A (DGEBPA),  from which BLR are made, is a type of
unmodified epoxy resin.  There are several types of unmodified
epoxy resin, but the standard and most common commercial epoxy
resin is DGEBPA.  In fact, DGEBPA is often referred to as
"conventional" epoxy resin.  Epoxy resins are plastic materials
which contain a specific molecular group that reacts with
different curing agents or hardeners resulting in hard, infusible
solids.  These solids have useful properties including good
adhesion to many substrates, low shrinkage, high electrical
resistivity, and good corrosion and heat resistance.  Commonly
used curing agents for DGEBPA include phenolic, urea, melamine,
furane, polyester, vinyl, polyurethane and silicone.
      The primary application of BLR is in protective coatings.
Other applications include electrical laminates, adhesives,
tooling, and flooring.  Industrialized nations are by far the
largest producers and consumers of BLR. Table 8-4 reports
patterns of consumption across end-use categories for the years
1989 and 1990.
     8.3.1.1.1  Protective coatings.  About 40 percent of
domestic BLR sales go to protective coatings markets.  The
primary uses of these coatings are automobile primers and
finishes, maintenance and marine coatings, can coatings, and
other product finishes.  The popularity of BLR in the coatings
industry is due to the high chemical resistance, toughness, and
adhesion properties.
     Epoxy coatings are called high performance coatings.  A high
performance coating or lining is one that is superior to paint in
adhesion, toughness, and resistance to continuing exposure to
industrial chemicals, food products, water, sea water, weather
and high humidity.  These coatings are designed to protect from
                               8-23

-------
             TABLE  8-4.  BLR  END-USE  CONSUMPTION
                     (millions of pounds)
End Use
Protective coatings
Reinforced uses
Electrical laminates
Other
Export
Tooling, casting,
molding
Bonding and adhesive
Fl coring , paving ,
aggregates
Other
Total
1989
193
57
26
86
30
25
25
41
483
Percent
of
total
40.0
11.8
5.4
17.8
6.2
5.2
5.2
8.5

1990
195
55
31
68
28
28
28
33
464
Percent
of
total
42.0
11.9
6.7
14.7
6.0
6.0
6.0
7.1

Source:  Plastics World. "Resin Report," January 1991.
                             8-24

-------
corrosive or otherwise detrimental exposure,  and to slow the
breakdown of industrial structures.   The  coatings need to be safe
for use with materials in which they  come into  contact as well as
be dense and have a minimum of absorption with  contacting
materials.  Also, they should have a  high resistance to the
transfer of chemicals through the coating.   Finally,  they should
maintain a generally good appearance  even though subject to
severe weather and chemical conditions.
     Epoxy surface coatings are the third most  common type of
industrial finish behind alkyds and acrylics.1^  Epoxies tend
to be more expensive, but have more attractive  properties than
other coatings including superior adhesion,  flexibility and
corrosion resistance when used on metallic substrates.  However,
due to their tendency to chalk or discolor upon exposure to
sunlight they are not often.used for  architectural purposes.
Solid DGEBPA low-molecular weight resins  are the most common type
of epoxy resin used in coatings.
     There are several types of epoxy coatings.17  Each has
different properties, but all are resistant and cure by internal
linkage only.  This means they need not be exposed to the air to
cure so that thick coatings can be achieved in  a single
application.  The primary types of epoxy  coatings are amine-cured
epoxies, polyamide cured epoxies, phenolic epoxies,  and coal tar
epoxies.  DGEBPA can also be reacted  with oils  or fatty acids to
make epoxy esters and other polymers.  The esters are used
primarily in floor finishes, primers  for  appliances,  and
maintenance coatings.  Epoxy esters accounted for approximately
5 percent of epoxy coating demand in  1991.
     Amine-cured epoxies are the most chemically resistant of the
ambient temperature cure variety  (they do not require heat or
other processes to cure).  However, they  can be brittle and chalk
quickly when exposed to the weather.   They are  mostly used in
    16 Chemical Economics Handbook, Epoxy Surface Coatings.
    17 The material in this section was taken from Pulp and Paper (1979),
Modern Plastics mid-October Encyclopedia issue.

                               8-25

-------
industry where an air-dry coating is required and the need for
chemical resistance is high.  Urea and Melamine are two of the
curing agents used to make amine-cured epoxies.
     Polyamide cured epoxies are somewhat less chemically
resistant than amine-cured epoxy.  However, they have much better
weather resistant qualities.  The coating is non brittle and
fairly flexible.  It has excellent resistance to alkali and to
water. The uses of this type of resin are broad based and include
maintenance coatings for most industries including the chemical,
paper, marine, atomic power, and food industries.
     Epoxy-phenolic systems offer chemical resistance along with
excellent mechanical properties.  When they are heat cured, they
are the strongest and most resistant of the epoxy coatings.  For
this reason, they are used for chemically resistant coatings on
process equipment, tank and drum linings, pipe linings, and for
protection from direct exposure to various chemicals.  They are
also commonly used for exposure to solvents, vegetable and animal
oils, fatty acids, foods, and alkalis.
     Coal tar epoxy coatings have good chemical resistance,
reasonable weather resistance, and outstanding resistance to
fresh and salt water,  brine, and hydrogen sulfide.  More
generally, they are resistant to both acidic and alkaline
conditions.   They are one of the most durable coatings for the
protection of concrete and metal, either under water or above
water, against corrosive elements.  These resins are black and so
have limited decorative use, and, of course, can be used only
where black is acceptable.  They are used throughout the chemical
industry and in the marine industry, both on ships and on
offshore structures.
     The following is a list of some specific uses of epoxy
coatings.
     •    Heavy duty industrial and marine maintenance coatings.
     •    Tank linings.
     •    Industrial floorings.
     •    Coatings for farm and construction equipment.
     •    Aircraft primers.
                               8-26

-------
     •    Floor and gymnasium finishes.
     •    Maintenance coatings.
     •    Metal decorating finishes.
     •    Pipe coatings.
     •    Container coatings.
     •    Electrodeposition primers- for automobiles.
     •    Solder masks.
     •    Beverage and food can coatings.
     •    Appliance primers.
     •    Hospital and laboratory furniture.
     •    Coating for jewelry and hardware.
     •    Impregnating varnishes.
     8.3.1.1.2  Bonding and adhesives.   Because of their
excellent adhesion to many substrates,  epoxy resins are widely
used as high performance adhesives.  For example, because of
their extraordinary adhesion to metal they are used in the
automobile, aircraft and construction industries.  According to
the September 1990 issue of Chemical Marketing Reporter, about
80 percent of epoxy adhesive sales go into the automobile and
construction industries.  According the Chemical Economics
Handbook, production of epoxy adhesives and sealant grew at an
average annual rate of 8 to 8.5 percent from 1983 through 1989.
     8.3.1.1.3  Molding, casting, and tooling.  Uses in this
category include encapsulation of electrical components by epoxy
molding compounds.  Also, epoxy casting resins are used as
prototypes and master models in the manufacture of tools.
Epoxies based on ultraviolet light stable structures are used in
the casting of outdoor insulators switch gear components and
instrument transformers.
     8.3.1.1.4  Laminating and composites.  Epoxy-based laminates
are used in printed wiring boards, such as those used in
computers and complex telecommunication equipment.  Epoxy
compounds are used in filament-wound glass reinforced pipe in oil
field applications, in the manufacture of pressure vessels and
tank and rocket motor casings, chemical plants, water
distribution, and as electrical conduits.  In the aerospace
                               8-27

-------
industry, graphite fiber-reinforced multifunctional epoxy resin
composites are becoming standard.
     8.3.1.1.5  Building and construction.  Epoxies are used  in
flooring, to repair bridges, roads, and  cracks in concrete,  to
coat reinforcing bars, and to perform as binders for patios,
swimming pool decks and the soil around oil well drills.
     8.3.1.2  WSR.  Polyamide-epichlorohydrin or WSRs are a type
of non-nylon polyamide resin sold almost exclusively in the paper
additives market.  Approximately 90 percent of these resins are
used to improve the wet tensile strength of paper products.
Other uses of these resins in the paper industry include floccu-
lent, drainage and drying aids, dry creping aids, cationizing
agent for unmodified  (pearl) potato and tapioca starch  (used  for
dry strength), and as a component of paper surface finishes.
     Paper which has been treated with a WSR shows greater
resistance to rupture or disintegration when exposed to water.
Note that wet strength is defined as tensile strength when the
paper is completely absorbed with water, not water repellency.
There are three primary types of WSR:   (1) Ureaformaldehyde
resins,  (2)  melamine formaldehyde resins,  (3) polyamide-polyamine
epichlorohydrin and modifications.  It is the third category  that
includes EPI-based non-nylon polyamide resins.  Other wet
strength additives include those made from polyacrylamide,
dialdehyde starch, polyacrolein resin, and cellulosic resin.  The
wet strength of paper increases almost linearly with the addition
of WSR up to a point.  Beyond this threshold the addition of  WSR
has little affect on the wet strength of the paper.
     There are a large number of uses for paper which retains
tensile strength when wet.  Examples include tea bags, paper
towels, and paper groceries bags.
     Demand for this type of paper is most likely inelastic
because of its "necessary" nature and the small percentage of
income which is typically spent on these types of products.
Specifically WSR are used for protection against:18
    18 This list was drawn from Pulp and Paper (1979).

                               8-28

-------
     •    Exposure to water of paper products used as drying or
          wiping media.  Examples are paper towels, napkins,
          windshield wiping tissue, industrial wiping towels,
          lens paper, and facial tissue.
     •    Exposure to weather.  Examples are packing cases,
          outdoor posters, building papers, paper bags, maps, and
          mulch paper.
     •    Wrapping for wet materials.  Examples are butcher
          wraps, fruit and vegetable wraps and boxes, frozen and
          prepared food packages, and foil wrapped wet wipes.
     •    Exposure to water by immersion in a processing
          operation.  Examples are photographic paper, copy print
          paper, filter paper, saturating paper, and tea bag
          paper.
     •    Disposables used in place of textiles.  Examples are
          hospital bed sheets, hospital gowns and other sanitary
          single-use garments.
8.3.2  Market Structure
     BLR and WSR are produced by a few large corporations, and
many of these are conglomerates.  Accordingly, market
concentration is relatively high and vertical integration is
common.
     8.3.2.1  BLR.  The major producers of basic liquid epoxy
resins  (BLR) are Dow Chemical Company, Ciba-Geigy Corporation,
and Shell Chemical Company.  They are also the largest producers
of any type of unmodified epoxy resin - each having production
rates on the order of 45.4 million kilograms per year.  These
three large companies have been producing epoxy resin for at
least 12 years.  Shell and Dow Chemical are also major producers
of epichlorohydrin and bisphenol-A which are the primary
feedstocks for epoxy.  While Ciba-Geigy is not similarly backward
integrated, it is a significant player in markets for further
processed epoxy products, including formulated systems,
electronic materials and composite materials.19  These facts
suggest that the market is fairly concentrated.  Together, these
    19 CEH  (1991) Epoxy Resins.

                               8-29

-------
three producers were responsible for approximately 60 percent of
                                          7 O
unmodified epoxy resin production in 1990. u
     BLR and other types of unmodified epoxy resin are reasonably
good substitutes.  For this reason, entry and exit from the
entire unmodified epoxy resin industry is worth examining.
Table 8-5 lists companies who produced unmodified epoxy resin- at
some time over the period 1980 to 1991.  An "X" in the column
under a given year indicates that the company produced epoxy
resin in that year.  These lists of manufacturers comprise only
those companies which responded to inquiries from the Society of
the Plastics Industry.  However, SPI's Committee on Resin
Statistics estimates that these manufacturers account for about
                                          o -i
95 percent of unmodified epoxy production.
     Over the last 12 years the list of manufacturers has had no
more tfian 7 companies on it, and, beginning in 1984, has had only
six. Since 1980, only four companies have entered or exited. The
average duration of manufacture over the 12 years from 1980 to
1991 was 8.3 years.
     8.3.2.2  WSR.  Firms which produced epichlorohydrin based
non-nylon polyamide resins in either 1988 or 1990 are listed in
Table 8-6.  An "X" in the column under a given year means that
the company produced the resin in that year, an "0" means that
they did not.  Over the two years of data, there were two entries
into and one exit from this industry.
     In 1988, Hercules accounted for approximately 80 percent of
the production of EPI-based polyamide. Henkel, Georgia-Pacific,
Borden and Callaway each accounted for about 5 percent of the
market.  Trinova produced very little.  It is unclear exactly
what percentage of the market is controlled by Hercules in 1990.
However, two Borden plants, five Hercules plants, and one Akzo
    20 SPI I and MRI  (1992).
    21 Source:  SPI II.
                               8-30

-------
           TABLE 8-5. COMPANIES PRODUCING UNMODIFIED EPOXY RESIN,
                                      1980-1991

Company Name       80 , 81  82  83   84   85   86  87  88  89  90  91
Celanese Plastics
Ciba-Geigy
Dow Chemical
Reichold Chemical
Shell X
Union Carbide
Morton Industries
Rhone Poulenc Inc.
Interez
Hi-Tek Polymers*
X
X
X
X
X
X
X



X
X
X
X
X
X
X



X
X
X
X
X
X
X



X
X
X
X
X
X




X
X
X
X
X
X




X
X
X
X
X
X


X


X
X
X
X
X


X


X
X
X
X
X


X


X
X
X
X
X





X
X
X
X
X



X

X
X
X
X
X

X



X
X
X

X

X


*      Hi-Tek Polymers was owned by Rhone Poulenc in 1989.

Source: Society of the Plastics Industry 1.
                                        8-31

-------
     TABLE 8-6.  COMPANIES PRODUCING EPI BASED NON-NYLON
               POLYAMIDE  RESINS,  1988 AND  1990
                                                   1988,  1990
 Borden
 Callaway Chemicala
 Georgia Pacific Corp.
 Henkel of America, Inc.
 Hercules, Inc.
 Trinova Corp.
 Pioneer Plastics
 Akzo
XX
XX
XX
XX
XX
xo
ox
ox
aA subsidiary of Exxon Corporation.

Source:  Chemical Economics Handbook and MRI  (1992).
                             8-32

-------
plant together accounted for 80 percent of production in
1990.22
     Raw materials for epichlorohydrin based polyamide resin
include adipic acid, diethylenetriamine, and epichlorohydrin.  No
producer of epi-based polyamide resins also produced these
feedstocks as of 1988.
     As of 1991 Georgia Pacific owned or controlled over 6
million acres of timber and timberlands.  They were also
producing pulp and paper (8 percent of U.S. annual capacity),
containerboard and packaging, uncoated free sheet paper, tissue,
envelopes and other paper products.  Hercules was producing
various paper products in addition to WSR in 1991.
8.3.3  Market Outlook
     While domestic production of both BLR and WSR has
fluctuated, the long-term trend has seen increased output of both
products.  BLR prices have been relatively stable recently, but
WSR prices have fallen.  The demand for both products is expected
to increase moderately over the next few years.
     8.3.3.1  BLR.  Domestic production of unmodified resins has
fluctuated somewhat over the twenty year period 1971 to 1990.
However, production has increased on average by about 7 percent
annually over this time.2^  BLR comprised about 60 percent of
total unmodified epoxy resin production in 1990.  Nominal prices
peaked at $2.89 (per kilogram) in 1984 and since then have
fluctuated around $2.40.24
     Two recently published industry reports predict that
U.S. epoxy resin production will experience healthy growth to the
end of the 1990's.  Network Consulting Inc.  (1992) expects a
4 percent annual growth in domestic production to last through
1997.  The Freedonia Group (1992) predicts North American
    22 MRI (1992).
    23 Computed from data in USITC I.
    24 USITC I.

                               8-33

-------
production, which is dominated by U.S. firms, to grow at an
annual rate of 4.4 percent to 1995.
     The Freedonia Group attributes expected growth to an
expanding export market which they predict will reach 300 million
pounds by 1995.  However, they expect the heavy growth in exports
to be countered by a slower growth in North American consumption,
which they expect to be just 3 percent per year.
     Network Consulting, Inc. (1992) base their projections of
growth on the recovery of the U.S. economy and the advent of
environmental regulations which favor the use of epoxies in high
solids and powder coatings.  In recent years, environmental
pressures have resulted in the rapid development of these epoxy
products because they use substantially less organic solvent but
retain the useful chemical and physical properties of epoxies.
      The Freedonia Group expects epoxy adhesives to grow at
4.4 percent per year to 1995, reaching 37 million pounds.  They
expect epoxy coatings to grow four tenths of a percent faster
than the overall coatings market at 2.8 percent per year to 1995.
This means that by 1995, 235 million pounds of epoxy resin will
be used in coatings.
     Chemical Economics Handbook estimates that between 1983 and
1988 the use of epoxy resins in adhesives and sealants grew at an
average rate of about 8.5 percent per year.  An increase of about
7 percent occurred in 1989.  Between 1989 and 1994, growth in the
consumption of epoxy adhesives is expected to slow slightly to a
6 percent annual rate.  They expect growth to be driven by the
increasing use of resin based composites in aerospace, automotive
and recreational markets.  Also, there is a trend toward using
epoxies instead of more expensive welds, especially in the
automotive market.  Technological advances may also contribute to
the growth of epoxy adhesives.  Improvements have been made in
various properties including adhesion to plastics and toughness,
and new systems have been' introduced that allow faster bonding at
lower initiation temperatures.
     Chemical Economics Handbook estimates that epoxy surface
coatings grew at an average annual rate of 3.5 to 4 percent
                              8-34

-------
between 1986 and 1990.  They expect this growth to slow to 3 to
3.5 percent presently.  Furthermore, they report that the
consumption of epoxy esters will decline due to their adverse
effects on the environment.  On the other hand powder coatings
are projected to grow at 5 to 10 percent annually because of the
high quality of the coating and the lack of adverse environmental
effects.  Consumption of epoxy resins in surface coatings is
projected by CEH to reach 215-220 million pounds by 1995.
     8.3.3.2  WSR.  The growth rate of production for non-nylon
polyamide was more volatile than that of epoxy over the period
1971-1990.  For example, production increased by 73 percent in
1974 and fell by 49 percent in 1975.  However, the average growth
rate of production was somewhat greater than that of epoxy, about
10 percent annually.  The nominal price reached $2.54 per
kilogram  (dry weight basis) in 1985, but has since" declined.
Production and prices both fell sharply in 1990.  Production fell
by about 26 percent in 1990 and price declined substantially to
$1.50 — a level it had not been below since 1972.25
     In 1987, Chemical Economics Handbook reported that the
growing trend to use neutral-cure wet-strength resins as
replacements for formaldehyde based (melamine and urea) resins
would continue through 1992, and that it would account for strong
increases in the demand for WSR.  Urea formaldehyde and melamine
formaldehyde resins are inferior as wet strength additives in
unbleached paper production because of their acid curing
characteristics.   Specifically, they can cause embrittlement and
deterioration of paper as well as reduce absorbency.  Demand for
epichlorohydrin based polyamide was predicted to continue to
increase at a rate of 5 to 6 percent annually from 1987 through
1992.  Production of non-nylon polyamide was somewhat erratic
over the period 1987 through 1990, experiencing 10 percent
decline in volume in 1990.
8.3.4  Foreign Trade
    25 USITC I  (various issues).
                               8-35

-------
   8.3.4.1  BLR.  Table 8-7 shows exports and  imports  of  epoxy
            TABLE 8-7.  U.S. TRADE IN EPOXY RESIN
                    (millions of kilograms)
Year
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
Exports
42.3
45.7
41.5
44.5
40.5
44.0
55.7
69.0
85.1
94.4
97.9
Imports/Exports
.05
.05
.07
.09
.14
.18
.12
.09
.07
.09
.11
Source:  U.S. Department of Commerce, Bureau of the Census.
                             8-36

-------
resin in millions of kilograms per year from 1981 through 1991 as
reported by the U.S. Department of Commerce.  The table also
reports imports divided by exports.
     According to the Bureau of the Census, exports in 1991
amounted to 97.9 million kilograms.  In 1989 the principle
destinations were: Far East countries other than Japan, which
accounted for 27 percent; Canada which accounted for 23 percent;
Western Europe, 21 percent; Japan, 11 percent; and Mexico,
7 percent.
     Imports have been fairly stable and relatively small in
volume through the eighties.  In 1989, an estimated 50 percent of
the imported epoxy resins were used in coatings, and the
remainder went primarily into adhesives and electronic
encapsulation.
     8.3.4.2  WSR.  Table 8-8 shows exports and imports of
nonnylon polyamide resins in millions of kilograms per year from
1981 through 1991. The ratio of exports to domestic production,
import to domestic production and imports to exports are also
reported.
     Exports fluctuated considerably during the 1980's, reaching
a high of 10.2 million kilograms in 1988 and a low of 3.6 million
kilograms in 1982.  However, in most years exports ranged between
4 and 6 million kilograms.  Imports were more stable over the
period hovering around 1 million kilograms until 1988.  However,
in 1989 and 1990 imports of non-nylon polyamide increased.
8.3.5  Financial Data
     Baseline financial data for firms producing resins are
displayed in Table 8-9.  The ratio of net income to assets and
the ratio of long term debt to long term debt plus equity are
reported in the table.  In order to compensate for cyclical
fluctuations, net income over assets figures were averaged over
'the years 1988 through 1991.  The long term debt ratio is
reported for 1991.
     Most of the resin producers are earning between a 4 and
6 percent return on their assets.  Dow is earning by far the
largest return at 9 percent.  Hercules and Shell show the poorest
                               8-37

-------
        TABLE  8-8.  U.S.  TRADE  IN NON-NYLON POLYAMIDE
                    (millions  of kilograms)
Year
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
Exports
5.0
3.6
4.1
4.4
5.3
4.5
6.1
10.2
5.9
6.2
n.a.
Imports
n.a.
n.a.
0.8
1.1
1.1
1.2
1.2
1.1
3.8
4.7
6.1
Imports/
Exports
n.a.
n.a.
.19
.24
.21
.26
.19
.11
.64
.76
n.a.
Sources:  U.S. Department of Commerce  (1992), USITC.
                             8-38

-------
        TABLE 8-9
FINANCIAL DATA FOR RESIN PRODUCERS

Hercules
Shell
Dow
Ciba Geigy
Borden
Exxon
Georgia Pacific
Henkel
NI/A*
(average 1988-1991)
.02
.03
.09
.05
.05
.06
.04
.05
LTD/(LTD+E)
(1991)
.22
.11
.39
.20
.41
.20
.58
.14
Source:  Moody's Industrial Manual 1991, Annual Reports 1991,
     NI = net income
     LTD = long term debt
                              A = assets
                              E = equity
                             8-39

-------
returns at  2  and 3 percent, respectively.
     Shell  and Henkel have the smallest  long  term debt ratios at
less than 0.15 each.   Hercules, Ciba Geigy and Exxon are a little
larger at around 0.2  each.  Dow and Georgia Pacific have the
highest ratios.
8.4  PRIMARY  ECONOMIC IMPACTS AND CAPITAL  AVAILABILITY ANALYSIS
8.4.1  Introduction
     This section presents estimates of  the primary economic
impacts which would result from the implementation of the NESHAP
and the results of the capital availability analysis.  We also
present results for Option I for the WSR industry.  Primary
impacts include changes in market prices and  output levels,
changes in  the value  of shipments by domestic producers, and
plant closures.   The  capital availability  analysis assesses the
ability of  affected firms to raise capital and the impacts of
control costs on plant profitability.
8.4.2  Estimates of Primary Impacts
     As explained earlier in Section 8-2,  we  use partial
equilibrium models of the affected industries to estimate primary
impacts.  The increase in production costs resulting from the
purchase and  operation of emission control equipment causes an
upward, vertical shift in the domestic supply curves.  The height
of this shift is determined by the after tax  cash flow required
to offset the per unit increase in production costs.  Because
control costs vary across plants within  each  industry segment,
the post-control supply curves are segmented.   We assume a worst
case scenario in which plants with the highest control costs  (per
unit of output)  are marginal (highest cost) in the post-control
market.
     Foreign  supply (net imports)  is assumed  to have the same
elasticity  as domestic supply in both markets.2^  Foreign and
    26 The United States is a new exporter of both BLR and WSR.  Trade in
DGEBA is substantial.  Net exports accounted for approximately 15 percent of
production in 1990.  However, trade in wet strength resin is insignificant.
The  small volume of trade is due to the custom of shipping only the polyamide
and  letting the receiving company complete the epichlorohydrin reaction.  In
1988 exports accounted for about 1 percent of domestic wet strength resin
                                8-40

-------
post-control domestic supply  are  added together to form total
market post-control supply.   The  intersection of post-control
market supply curve with market demand determine the new market
equilibrium price and quantity.   The  post-control domestic output
is given by post-control market output less post-control imports.
     Table 8-10 presents the  primary  impacts predicted by the
partial equilibrium analysis  for  the  BLR and WSR industries.  For
example, we estimate that  the NESHAP  will result in a 0.12 cent
per kilogram  (0.05 percent) increase  in the price of BLR and an
annual reduction in domestic  production of about 106 metric tons
(0.08 percent of baseline  production).  We also estimate that the
NESHAP will cause the annual  value  of domestic shipments to fall
by about $108,000  (0.03 percent).   No plant closures are
predicted.
     Table 8-10 also shows the estimated impacts on the WSR
industry,  both for the MACT  Floor  and Option I. Estimated price
and output changes range from very  small impacts associated with
Option I to larger impacts under  the  MACT Floor.  Under the MACT
Floor, estimated increases in price and decreases in domestic
production are approximately  4 percent,  and about 1 plant closure
is possible27.  However, under Option I, price and output
impacts are only 0.22 percent and 0.20 percent, respectively, and
no plant closures are predicted.
     We emphasize that the assumptions we adopt in our analysis
are likely to cause us to  overstate predicted plant closures.
First, we assume that the  plant with  the highest per unit
emission control costs also is the  least efficient in that it has
the highest baseline per unit production costs.  Second, we
assume a national market,  but regional trade barriers might
afford some protection for some plants.   Finally, the production
of WSR is intermittent.  When our analysis predicts a plant
production, and imports were only a small fraction of exports.
     07
        Table 4-1 reports fractions of plant closures.  The 0.63 plant
closure predicted  for the wet strength resin industry means that we estimate
that the marginal  plant would lose 63 percent of its annual production.

                               8-41

-------
        TABLE  8-10.   ESTIMATED PRIMARY IMPACTS ON BLR
                AND WET STRENGTH RESIN MARKETS

IMPACT
Price Change
C/kilograma
Percent
Annual Change in
Domestic Output
Metric Tonsb
Percent
Annual Change in
Value of Domestic
Shipments
$l,000a
Percent
Plant Closures

BLR
.12
.05
-106
-.08


-108
-.03

.00
WET STRENGTH RESIN
MACT Floor Option I
.84
4.19
-7347
-3.73


-123
-.31

.63
.04
.22
-404
- .20


7
.02

.03
a 1992 dollars.
b Wet weight basis.
                             8-42

-------
closure, it means that the plant will  cease production of WSR,
but not close operations altogether.
     The estimated primary impacts  reported above depend on a set
of parameters used in the partial equilibrium model  of the wet
strength and BLR resin industries.  One  of the  parameters, the
elasticity of demand, measures how  sensitive  buyers  are to price
changes.  The estimated impacts reported above  are based on a
demand elasticity of -0.92 for the  wet strength market and
-1.5 for the BLR market.  In Appendix D,  we report the results of
analyses that show the sensitivity  of  the estimated  impacts to
changes in the demand elasticity.   The "low"  elasticity case
adopts a demand elasticity of -0.5  for the WSR  industry and -0.62
for the BLR industry.  The results  show  slightly larger price
increases, smaller reductions in market  output  and less adverse
impacts on domestic producers than  results reported  above.28
The "high" elasticity case uses a demand elasticity  of -1.34 for
the WSR industry and -3.10 for the  BLR industry.   In general,
this case shows slightly smaller price increases but more adverse
impacts on domestic producers.  However,  the  sensitivity analysis
generally shows that the estimated  primary impacts are relatively
insensitive to reasonable ranges of demand elasticity estimates.
     Also, the estimated impacts reported in  Table 8-10 is based
on the assumption that plants with  the highest  emission control
costs (per unit of output) are marginal  (highest cost)  producers
in the post-control market.  This assumption  causes  the adverse
impacts associated with the regulatory alternatives  to be
overstated.
8.4.3  Capital Availability Analysis
     The capital availability analysis involves examining pre-
and post-control values of selected financial ratios.  These
ratios include net income divided by assets and long term debt
divided by the sum of long term debt and equity.   In order to
       Also, the plant closure previously predicted for the wet strength
resin industry under the MACT Floor is no longer predicted when a * low*
elasticity of demand is assumed.

                               8-43

-------
reduce the effects of year-to-year fluctuations in net income, a
four-year average  (1988 through 1991) of net income over assets
was used as the baseline.  Changes in the long term debt ratio
represent structural changes and so are not subject to the same
cyclical fluctuations.  Long term debt ratios from 1991 were used
as the baseline.
     As explained in Section 8-2, these financial statistics lend
insight into the ability of affected firms to raise the capital
needed to acquire emission controls.  They provide estimates of
the changes in profitability which would arise from the
implementation of the NESHAP.
     To calculate the post-control ratio of net income to assets,
annualized control costs were subtracted from pre-control net
income, and capital control costs were added to pre-control
assets.  To calculate the post-control long term debt ratio,
capital control costs were added to pre-control long term debt,
both the numerator and denominator of this ratio.  Note that both
post-control ratios reflect a worst-case assumption that affected
firms are required to absorb emission control costs without the
benefit of higher market prices.
     Financial data are available for all three DEGBPA producers
and 5 of the 7 WSR producers.  The 5 wet strength producers own
15 of the 17 facilities.
     All of the companies that produce BLR and WSR are large
corporations.  As a result,  emission controls costs, which are
relatively small, have no perceptible impacts on the firm's
financial ratios after rounding.  Accordingly, we conclude that
affected companies will not find it difficult to raise the
capital necessary to purchase and install the required emission
controls.
8.4.4  Limitations of Estimated Primary Impacts
     Several qualifications of the estimated primary impacts
presented in this section need to be made.  A single market for
homogeneous output is assumed in the partial equilibrium
analysis.  However, there may be some regional trade barriers
which would protect producers.   Furthermore, the analysis assumes
                               8-44

-------
that plants with the highest per unit emission control costs are
marginal post-control.  This assumption will cause the impacts
presented above to be overstated since market impacts are
determined by the costs of marginal plants.  Finally, some plants
may find that the price increase resulting from regulations make
it profitable to expand production.  This would occur if a firm
found its post-control incremental unit costs to be smaller than
the post-control market price.  Expansion by these firms would
result in a smaller decrease in output and increase in price than
otherwise would occur.
     We have also noted that the estimated primary impacts depend
on the parameters of the partial equilibrium model.  The results
of the sensitivity analyses presented in Appendix D,  which are
based on a larger (more elastic) estimate of demand elasticity,
show slightly more adverse impacts on domestic producers.
     The capital availability analysis also has limitations.
First, future baseline performance may not resemble past levels.
Second, the tools used to measure the impact of controls are
limited in their scope.  Finally, the financial analysis is based
on a worst-case assumption that affected establishments will
fully absorb emission control costs without the benefits of
higher prices.
8.4.5  Summary of Primary Impacts
     The estimated impacts of the NESHAP on the BLR industry is
relatively small.  Predicted price increases, reductions in
domestic output and the value of domestic shipments for the BLR
industry are 0.08 percent or less.  The impacts estimated under
the MACT Floor for the WSR industry are somewhat more adverse.
Predicted price increases, reductions in domestic output and the
value of domestic shipments for the MACT Floor are about 4
percent, and one plant closure is possible.  Under the Option I
Scenario, however, predicted price increases, reductions in
domestic output and the value of domestic shipments are
0.22 percent or less, and no plant closures are expected.  As
noted earlier, these results are likely to overstate the true
adverse impacts.  Finally, because emission control costs are
                               8-45

-------
very small  relative to the financial resources  of  affected
producers,  they should not find it difficult  to raise the capital
necessary to  finance the purchase and installation of emission
controls.
8.5  SECONDARY  ECONOMIC IMPACTS
8.5.1  Introduction
     This section presents estimates of the secondary economic
impacts that  would result from the implementation  of the NESHAP.
Secondary impacts include changes in employment, energy use, and
foreign trade and regional impacts.
8.5.2  Labor  Impacts
     The estimated labor impacts associated with the NESHAP are
based on the  results of the partial equilibrium analyses of the
two resin industries.   These estimated impacts  depend primarily
on the estimates of reduction in domestic production reported
                        O Q
earlier in  Section 8-4. ^  Note that changes  in employment due
to the operation and maintenance of control equipment have been
omitted from  this analysis due to lack of data.  Also,  the
estimated employment impacts reported below do  not include
potential employment gains in industries which  produce substitute
commodities that might benefit from reduced BLR and WSR
production.   Thus,  the changes in employment  estimated in this
section reflect only the direct employment losses  due to reduc-
tions in domestic production of BLR and WSR.
     Table  8-11 presents estimates of employment losses for each
of the two  industries.  As Table 8-11 indicates, the estimated
job losses  are  small (up to two production jobs).   As expected,
the estimated employment losses in the WSR industry are smaller
for Option  I  than for the MACT Floor.  The generally small
impacts occur primarily because only small reductions in output
are expected  to occur as a result of the implementation of the
    2' More specifically, we estimate employment impacts by assuming that
labor use per unit of output will remain constant when the quantity of output
changes.   Production worker hours per dollar of output was calculated from
1989 Annual Survey of Manufactures and a producer price index for chemicals
and allied products obtained from the Economic Report of the President 1991.
See Appendix B for a  more detailed discussion.

                                8-46

-------
            TABLE 8-11.   ESTIMATED EMPLOYMENT LOSSES
IMPACT
Lost Jobs
Percent Loss
BLR
0.34
0.08
Wet Strength
Resin, MACT
Floor
1.84
3.73
Wet Strength
Resin,
Option I
0.10
0.20
NOTE:  Estimates do not include potential employment gains due to  operating
      and maintaining emission controls.
                                 8-47

-------
NESHAP.  Also, the industry is characterized by a relatively high
output/employment ratio or low labor intensity.
8.5.3  Energy Use Impacts
     The approach we employ to estimate reductions in energy use
is similar to the approach employed to estimate labor impacts.
Again, these impacts depend primarily on the estimated reductions
in domestic output reported earlier in Section 8-4.  Note that
the changes reported below do not account for the potential
increases in energy use due to operating and maintaining emission
control equipment.  This omission is due to lack of data.
     Table 8-12 presents changes in the use of energy by each
industry.  As expected, the estimated changes in energy use are
minor because only small reductions in output are expected as a
result of the implementation of the NESHAP.  The change in the
use of energy by the WSR industry differs substantially between
the MACT Floor and Option I.  Much smaller energy use reductions
are expected under Option I.
8.5.4  Foreign Trade Impacts
     Other factors being the same, the implementation of the
NESHAP will raise the production costs of domestic resin
manufacturers relative to foreign producers, causing U.S. imports
of resin to increase and U.S. exports to decrease.  The effects
of the regulation on both the quantity and the value of net
exports  (exports-imports) are reported in Table 8-13.
     The estimated trade impacts are small, both because of small
predicted domestic price increases and because of the relatively
small amount of trade that exists currently for the two products.
For example, we estimate that the implementation of the standard
will result in reduced BLR net exports of about 20 metric tons
annually (about .12 percent of baseline net exports) or about
$38,000 per year.   Note that we predict only a slight change in
the dollar value of WSR exports under both the MACT Floor and
Option I, even though we estimate that the volume of exports will
fall by about 73 metric tons annually under the MACT Floor, and 4
metric tons annually under Option I.  In either case, the higher
                               8-48

-------
          TABLE 8-12.   ESTIMATED ENERGY USE REDUCTIONS
INDUSTRY/ IMPACT
BLR
$1,000 1992
Percent Reduction
WSR (MACT Floor)
$1,000 1992
Percent Reduction
WSR (Option I)
$1,000 1992
Percent Reduction

8.51
0.08
45.55
3.73
2.50
.20
NOTE:  Estimates do not include potential increases  in energy use due to
      operating and maintaining emissions controls.
                                 8-49

-------
          TABLE  8-13.  ESTIMATED IMPACTS ON NET EXPORTS
INDUSTRY/ IMPACT
BLR
Volume (metric tons)
Percent Change (volume)
Value ($1,000 1992)
WSR (MACT Floor)
Volume (metric tons)
Percent Change (volume)
Value ($1,000 1992)
WSR (MACT Floor)
Volume (metric tons)
Percent Change (volume)
Value ($1,000 1992)

-20
-.12
-38
-73
-3.73
1.23
*
-4
-.21
.06
NOTES:      Dollar estimates of  trade impacts are adjusted for higher
           post-control prices.  Changes in trade volumes are reported on
           a wet weight basis.
                                 8-50

-------
post-control prices offset (approximately) the reduced physical
volume of exports.
8.5.5  Regional Impacts
     No significant regional impacts are expected from the
implementation of the NESHAP because estimated employment impacts
are small.
8.5.6  Limitations of Estimated Secondary Impacts
     Our estimates of the secondary impacts associated with the
NESHAP are based on changes in market equilibria predicted by the
partial equilibrium models of the two affected markets.
Accordingly, the caveats we discussed earlier in Section 8-4 for
the primary impacts apply as well to our estimates of secondary
impacts.
     As noted earlier, the estimates of employment impacts do not
include potential employment gains due to operating and
maintaining emission control equipment or employment gains in the
manufacturing of substitute products.  Similarly, the estimates
we report exclude potential indirect employment losses in
industries that supply inputs to the resin industries.  In short,
the reported estimates of employment impacts include only direct
production job losses in the BLR and WSR industries.
8.5.7  Summary of Secondary Impacts
     The estimated secondary economic impacts of the alternative
NESHAP are generally small.  Estimated employment and energy
impacts are small because only small reductions in industry
output are expected.  The estimated trade impacts are minor
because only small domestic price increases are expected and
because baseline trade volumes for the affected products are
small.  No significant impacts on regional economies are
expected.
8.6  POTENTIAL SMALL BUSINESS IMPACTS
                               8-51

-------
     Firms in the  BLR  and WSR industries are classified as  "small
businesses" if they  employ fewer than 750 employees.30  No  BLR
producer satisfies the criteria for a small business.31  The
three BLR producers, Shell,  Ciba-Geigy and Dow Chemical, employed
over 30 thousand people in 1991.  Ciba-Geigy employed over  90
thousand people in 1991.   Employment data are available for 5 of
the 7 wet strength facilities.   Of the five, the company
employing the fewest people was Hercules at approximately 15
thousand employees.  No WSR producer for which we have employment
data comes close to  qualifying as a small business.  Table  8-14
shows the total employment of resin producing companies in  1991.
     The Small Business Administration defines a small business
as one which is not  dominant in its field.  There are three
producers in the DGEBA industry.  Each producer has a substantial
market share.
     The EPA Guidelines for Implementing the Regulatory
Flexibility Act state  that the definition of a small business is
"any business which  is independently owned and operated and not
dominant in its field."  The three corporations producing DGEBA
each have substantial  market share.  Similarly, the producers of
WSR are typically  large conglomerates which employ well over 750
people.
8.7  ECONOMIC COSTS
     Estimates of  the  economic costs associated with the
implementation of  the  NESHAP for the BLR and WSR industries are
presented below in this section of the report.
8.7.1  Economic Costs  of  Emission Controls:  Conceptual Issues
     Air quality regulations affect society's economic well-being
by causing a reallocation of productive resources within the
economy.  Specifically,  resources are allocated to the production
of cleaner air and away from other goods and services that  could
    30 EPA  (1992).  EPA Guidelines for Implementing the Regulatory Flexibili-
ty Act, Revised April 1992, Appendix C.  Small Business Size Regulations, 13
CFR Part 121.
    31 EPA may adopt an alternative definition of a small business if an
alternative size  cutoff can be justified.
                               8-52

-------
         TABLE 8-14.  EMPLOYMENT OF RESIN PRODUCERS
Company Name
Georgia Pacific
Henkel
Hercules
Dow
Borden
Exxon
Ciba-Geigy
Shell
Employment








in 1991
over 52,000
41,000
15,000
62,000
44,000
101,000
91,000
30,000
Source:  1991 Annual Reports
                            8-53

-------
otherwise  be  produced.   Accordingly, the economic  costs of
emission controls can be measured as the value  that  society
places on  those goods and services not produced as a result of
resources  being diverted to the production of improved air
quality.   The conceptually correct valuation of these costs
requires the  identification of society's willingness to be
compensated for these foregone consumption opportunities that
would otherwise be available.^2
     In the discussion that follows, we distinguish  between
emission control costs and the economic costs associated with the
regulatory alternatives.  The former are measured  simply as the
annualized capital and annual operating and maintenance costs of
controls under the assumption that all affected plants install
controls.  As noted above, economic costs reflect  society's
willingness to be compensated for foregone consumption
opportunities.
     Estimates of emission control costs will correspond to the
conceptually  correct measure of economic costs  only  if the
following  conditions hold:
     •     Marginal plants affected by an alternative standard
           must be able to pass forward all emission  control costs
           to  buyers through price mark-ups without reducing the
           quantity of goods and services demanded  in the market.
     •     The prices of emission control resources
           (e.g.,  pollution control equipment and labor)  used to
           estimate costs must correspond to the prices that would
           prevail if these factors were sold in competitive
           markets.
     •     The discount  rate employed to compute the  present value
           of  future costs must correspond to the appropriate
           social discount rate.
     •     Emission controls do not affect the prices of goods
           imported to the domestic economy.
    •3 ^
       Willingness to be compensated is the appropriate measure of economic
costs, given the convention of measuring benefits as willingness to pay.
Under this convention, the potential to compensate those members of society
bearing the costs associated with a policy change is compared with the
potential willingness of gainers to pay for benefits.  See Mishan (1971) .

                                8-54

-------
     8.7.1.1  Market Adjustments.  A plant is marginal if it is
among the least efficient producers in the market and, as a
result, the level of its costs determine the post-control
equilibrium price.  A marginal plant can pass on to buyers the
full burden of emission control costs only if demand is perfectly
inelastic.  Otherwise, consumers will reduce quantity demanded
when faced with higher prices.  If this occurs, estimated control
costs will overstate the economic costs associated with a given
air quality standard.
     The emission control costs estimates do not reflect any
market adjustments that are likely to occur as affected plants
and their customers respond to higher post-control production
costs.  The estimates of economic costs presented later in this
section do reflect estimates of such market adjustments.
     8.7.1.2  Markets for Emission Control Resources.  Other
things being the same, estimated emission control costs will
overstate the economic costs associated with an alternative air
quality standard if the estimates are based on factor prices
(e.g., emission control equipment prices and wage rates) which
reflect monopoly profits earned in resource markets.  Monopoly
profits represent a transfer from buyers to sellers in emission
control markets, but do not reflect true resource costs.
     The extent to which sellers in emission control markets
possess monopoly power has not been investigated.  Consequently,
we assume in this study that emission control resources are
traded in competitive markets.  The estimated economic costs
reported in this section are overstated if this assumption does
not hold.
     8.7.1.3  The Social Discount Rate.  The estimates of
annualized emission control costs presented earlier in this
report were computed by adding the annualized estimates of
capital expenditures associated with the purchase and
installation of emission control equipment to estimates of annual
operating and maintenance costs.  Capital expenditures were
annualized using a 7 percent discount rate.  The private cost of
capital is appropriate for estimating how producers adjust supply
                               8-55

-------
prices  in response to control  costs.33  In order to estimate
the economic costs associated  with the NESHAP,  an appropriate
measure of the social discount rate should be used in the
amortization schedule.
     There is considerable  debate regarding  the use of
alternative discounting procedures and discount rates to assess
the economic benefits and costs associated with public
programs.34  The approach adopted here is a  two-stage procedure
recommended by Kolb and Scheraga (1990).
     First,  annualized costs are computed by adding annualized
capital expenditures  (over  the expected life of emission
controls)  and annual operating costs.  Capital  expenditures are
annualized using a discount rate that reflects  a risk-free
marginal return on investment.35  This discount rate, which is
referred to below as the social cost of capital, is intended to
reflect the opportunity cost of resources displaced by
investments in emissions controls.   Kolb and Scheraga (1990)
recommend a range of 5 to 10 percent for this rate.  We adopt a
midpoint value of 7.0 percent  in this analysis.36
     Second,  the present value of the annualized stream of costs
is computed using a consumption rate of interest which is taken
as a proxy for the social rate of time preference.   This discount
rate, which is referred to  below as the social  rate of time
preference,  measures society's willingness to be compensated for
postponing current consumption to some future date.  Kolb and
Scheraga (1990)  argue that  the consumption rate of  interest
     33 In other words, a discount rate reflecting the private cost of capital
to affected firms should be used in analyses designed to predict market
adjustments associated with emission control costs. The private cost of
capital, assumed to be 10 percent in this analysis, is higher than the 7 per-
cent  social discount rate because it reflects the greater risk faced by
individual procedures related to the risk faced by society at large.

     34 See Lind, et al.  (1982) for a more detailed discussion of this debate.

     35 The risk-free rate is appropriate if the NESHAP, as a program, does
not add to the variance of the return on society's investment portfolio.
     36 The 7 percent discount rate is also consistent with recent OMB recom-
mendations .

                                8-56

-------
probably lies between 1 and 5 percent.  We do not, however,
present estimates of the present value of the costs associated
with the NESHAP in this report.
     The resulting estimates of the present value of the economic
costs associated with the NESHAP can be compared with estimates
of the present value of corresponding benefits in the BCA.  The
social rate of time preference should be employed to discount the
future stream of estimated benefits.
     8.7.1.4  Costs of Imported Goods.  The NESHAP is expected to
cause an increase in prices paid for imports.  From the
perspective of the world economy, higher prices paid for imported
goods represent a transfer from domestic consumers to foreign
producers.  However, from the perspective of the domestic economy
alone, higher prices on imported goods represent an economic
cost.
     Since we do not consider the welfare of foreign producers in
this analysis, we treat expenditures on BLR and WSR due to higher
prices as a cost.  Note that there are two sources of this cost:
(1) higher prices paid for baseline imports; and  (2) higher
prices paid for the additional imports induced by emission
control costs faced by domestic producers.
8.7.2  Other Costs Associated with NESHAP
     It should be recognized that the estimates of costs reported
later in this section do not reflect all costs that might be
associated with the NESHAP.  Examples of these include
administrative, monitoring, and enforcement costs (AME),  and
transition costs.
     AME costs may be borne by directly affected firms and by
different government agencies.  These latter AME costs, which are
likely to be incurred by state agencies and EPA regional offices,
for example, are reflected neither in the estimates of emission
control costs, nor in the estimates of economic costs.
     Transition costs are also likely to be associated with the
alternative standards.  Analyses described in previous sections
of this report, for example, predict that some plants will close
because of emission control costs.  This will cause some
                               8-57

-------
individuals to suffer  transition costs associated with temporary
unemployment and affected  firms  to incur shutdown costs.  These
transition costs are not reflected in the cost estimates reported
later in this section.
8.7.3  Changes in Economic Surplus as a Measure of Costs
     As was noted earlier,  willingness to be compensated for
foregone consumption opportunities is taken here as the
appropriate measure of  the costs associated with the NESHAP.  In
this case, compensating variation is an exact measure of
willingness to be compensated.   In practice, however,
compensating variation  is  difficult to measure; consequently, the
change in economic surplus associated with the air quality
standard is used as an  approximation to compensating variation.
     The degree to which a change in economic surplus coincides
with compensating variation as a measure of willingness to be
compensated depends on  whether the surplus change is measured in
an input market or a final goods market.  The surplus change is
an exact measure of compensating variation when it is measured in
an input market, but it is an approximation when measured in a
final goods market. 7
     The direction of the  bias in the approximation of
compensating variation  when the  surplus change is measured in a
final goods market depends on whether affected parties realize a
welfare gain or suffer  a welfare loss,  but in either case,  the
bias is likely to be small.38  Affected firms (and their
customers) will suffer  a welfare loss as the result of the
implementation of emission controls.   In this case,  the change in
economic surplus will exceed compensating variation,  the exact
measure of willingness  to  be compensated.39
8.7.4  Estimates of Economic Costs
    37 See Just, Heuth, and Schmitz  (1982) for a more detailed discussion.
    38 See Willig (1974).

    39 See Appendix B for a detailed, technical description of the methods
employed to compute changes in economic surplus.

                               8-58

-------
     Estimates of the annualized total economic  costs associated
with the NESHAP are reported in Table 8-15  (for  a  social  cost  of
capital equal to 7.0 percent).  The estimates of total  annual
costs of the NESHAP are $120 thousand for the BLR  industry,
$465 thousand for the WSR industry under the MACT  Floor,  and $51
thousand for the WSR industry under Option  I.
     We measure economic costs as net losses in  economic  surplus.
Table 8-15 shows how losses in surplus are  distributed  among
consumers, domestic producers and society at large.  The  latter
is referred to as "residual" surplus in the table.
     The loss in consumer surplus includes  higher  outlays for
foreign and domestically produced BLR and WSR plus a dead weight
loss due to foregone consumption.  As Table 8-15 indicates,
consumers in each market suffer a loss in surplus.  These losses
are due mostly to higher expenditures on BLR and WSR.
     We compute the loss in producer surplus as  annualized
emission control costs incurred by plants remaining in  operation
plus the dead weight loss in surplus due to reduced output less
increased revenue due to higher post-control prices.  The
estimated losses in producer surplus reported in Table  8-15  are
negative, meaning that domestic producers would  realize a net
gain in economic surplus.  This occurs because higher post-
control market prices more than offset emission  control costs.
     Surplus losses to society at large are computed as
"residual" adjustments to account for differences  in private and
social discount rates and transfer effects  of taxes.  The
estimates of changes in producer surplus reflect a 10 percent
real private rate on emission control capital costs.  Recall that
social costs are discounted at a 7.0 percent real  rate.40
     We note that the distribution of economic costs between
consumers and domestic producers depends, in part, on the way we
have constructed the post-control supply curve.  As explained
earlier, we have assumed that plants with the highest emission
    4  Since the loss  in producer surplus measures the burden of the alter-
native borne by producers, we calculate it using the private cost of capital.

                               8-59

-------
     TABLE 8-15,
ESTIMATES OF ANNUALIZED  ECONOMIC  COSTS
(thousands  of 1992 dollars)

Industry
BLR
Wet Strength Resin
MACT Floor
Option I
Loss in
Consumer
Surplus
141
1,607
87
Loss in
Producer
Surplus
-3
-841
-22
Loss in
Residual
Surplus
-19
-300
-13
Loss in
Surplus
Total
120
465
51
NOTE:  Estimates are computed as the annualized reduction in economic
      surplus to the domestic economy.
                                8-60

-------
control costs (per unit of output) are marginal in the post-
control market.   This assumption is worst case in that it results
in large increases in prices (relative to an alternative
assumption that plants with high control costs are not marginal),
thus shifting the cost burden to consumers and away from plants
that continue to operate in the post-control market.  Any
alternative construction of the post-control supply curve would
result in smaller price increases and shift a larger share of
economic costs away from consumers to domestic producers.  In
other words, smaller price increases would reduce the economic
rent realized by domestic producers in the post-control market.
     Earlier, we explained that economic costs differ from
emission control costs.  Recall that the latter are computed
simply as annualized capital costs plus annual operating and
maintenance costs, assuming that all plants install controls.
Table 8-16 reports estimates of annualized emission control
costs.  These estimates are $145 thousand for the BLR industry
and range from $52 thousand under the MACT Floor to $519 under
Option I for the WSR industry.   The emission control costs
reported in Table 8-16 exceed the economic costs reported in Ta-
ble 8-15  under the MACT Floor scenario.  This occurs because the
estimated economic costs reflect market adjustments away from
marginally expensive production.
                               8-61

-------
       TABLE 8-16.
 ESTIMATES  OF THE ANNUALIZED EMISSION
       CONTROL COSTS
(thousands of 1992 dollars)
BLR
145
145
Wet Strength
Resin
MACT Floor
519
N.A.
Wet Strength
Resin
Option I
N.A.
52
Total
664
197
NOTE:  Estimates are computed as annualized capital costs  plus annual
      operating and maintenance costs,  assuming all plants  continue to
      operate after controls are installed.  Capital costs  are annualized
      at a 7 percent discount rate.
                                8-62

-------
                APPENDIX A




AFFECTED PLANTS AND EMISSION CONTROL  COSTS

-------
                            APPENDIX A
            AFFECTED PLANTS AND EMISSION CONTROL COSTS
                                                                *
     This appendix describes the affected BLR and  WSR plants and
the estimates of emissions and emission  control  costs used in
this study.
AFFECTED PLANTS
     There are three major BLR producers.   Consequently,  we are
able to use plant specific data for baseline emissions,  emissions
reductions, and control costs.  Data  on  production rates at BLR
plants, however, is considered confidential and  is not available
to the public.  We use an average annual production rate of
45,000 metric tons  (wet weight) as baseline output for each of
the three BLR facilities.1
     There are 17 WSR facilities nationwide.  However,  only five
of these plants are expected to incur emission control costs
under the MACT Floor and nine under Option  I.  We  assume that
each of these plants produce 11,600 metric  tons  annually.2
EMISSION CONTROL COSTS
     Table A-l reports emission control  capital  costs and
annualized costs for the three BLR facilities.   Table A-2 shows
the same information for the five affected  WSR plants.
Annualized costs include amortized capital  costs plus the annual
operating and maintenance costs associated  with  emission con-
trols.  Table A-3 shows capital and annualized costs for the nine
WSR plants expected to be affected by Option I  (estimated costs
are the same for all nine plants).
    1 Draft BID, Section 6, Appendix A (wet weight).
    2 Draft BID, Section 6, Appendix A (wet weight).

                               A-l

-------
     TABLE A-l.  CONTROL COSTS AT BLR PLANTS
PLANT
DOW
Ciba-Geigy
Shell
CAPITAL
COSTS
(1992$)
254,873
104,778
67,618
ANNUALIZED
COSTS
(1992$)a
31,207
75,461
38,266
 a  Capital  costs  annualized at a 7 percent
    discount rate.
TABLE A-2.
MACT FLOOR CONTROL COSTS AT AFFECTED
        WSR PLANTS


PLANT ID #
1
2
3
12
14
CAPITAL
COSTS
(1992$)
24,500
37,400
360,000
48,000
39,500
ANNUALIZED
COSTS
(1992$)a
86,000
112,000
91,000
111,000
119,000
 a  Capital costs annualized at a 7 percent
    discount rate.
       TABLE A-3.  OPTION I CONTROL COSTS
      AT AFFECTED WET STRENGTH RESIN PLANTS
NUMBER OF
PLANTS
9
CAPITAL
COSTS
(1992$)
15,348
ANNUALIZED
COSTS
(I992$)a
5,782
 a  Capital costs annualized at a 7 percent
    discount rate.
                       A-2

-------
                 APPENDIX B
TECHNICAL DESCRIPTION OF ANALYTICAL METHODS

-------
                            APPENDIX B
           TECHNICAL DESCRIPTION OF ANALYTICAL METHODS

     This technical appendix provides detailed descriptions of the
analytical methods employed to conduct the following analyses:
     •    Partial equilibrium analysis (i.e., computing post-
          control price, output and trade impacts) .
     •    Estimating changes in economic surplus.
     •    Labor and energy impacts .
     •    Capital availability.

We also present the baseline values used in the partial
equilibrium analysis.
PARTIAL EQUILIBRIUM ANALYSIS
     The partial equilibrium analysis requires the completion of
four tasks.  These tasks are:
     •    Specify market demand and supply.
     •    Estimate the post -control shift in market supply.
     •    Compute the impact on market quantity.
     •    Compute the impact on market price.
     •    Predict plant closures.

Market Demand and Supply
     Baseline or pre- control equilibrium in a market is given by:
          Qd = aPe                                          (B.I)
          Qd = 0P7                                          (B.2)
          Q| = PP7                                          (B.3)
where,    Q  = output;
          P  = price;
          e  = demand elasticity;

                               B-l

-------
          7  = supply elasticity;
          a, 18 and p are constants;
          Subscripts d and s reference demand and supply,
          respectively; and,
          Superscripts d and f reference domestic and foreign
          supply, respectively.

The constants a, 0 and p are computed such that the baseline
equilibrium price is normalized to one.  Note that the market
specification above assumes that domestic and foreign supply
elasticities are the same.
Market Supply Shifts
     Supply price for a model plant will increase by an amount
just sufficient to equate the net present value of the investment
and operation of the control equipment to zero.  Specifically,

         [(C-Q) - (V+D) ] (1-t) +D =k                          (B 5)
                   S
where   C  is the change in the supply price;
        Q  is output;
        V  is a measure of annual operating and maintenance
           control costs.
        t  is the marginal corporate income tax rate;
        S  is the capital recovery factor;
        D  is annual depreciation (we assume straight -line
           depreciation) ;
        k  is the investment cost of emissions controls.

     Solving for C yields the following expression:
         C =     -    +                                        (B.6)
            Q(l-t)     Q
                               B-2

-------
     Estimates of  k  and V were obtained from EPA  (1991).   The
variables, D, I, and S  are computed as follows:
        D  =   k/T                                           (B.7)
and
        S  =   r(l+r)T/((l+r)T-l)                             (B.8)

where   r  is the  discount rate or cost of capital faced by
           producers;
        T  is the  life  of emission control equipment.

     Solving for P in Equation (B.2)  yields the following
expression for the baseline inverse market supply function for
domestic producers.
        P  =    (Qd//3)1/7                                     (B.9)
     Emission control costs will raise the supply price of the
ith model plant by Cj_  (as computed in Equation  (B.6)) .  The
aggregate domestic market supply curve, however, does not
identify the supply  price for individual plants.  Accordingly,  we
adopt the worst-case assumption that model plants with the
highest after-tax  per unit control costs are marginal in the
post-control market.  Specifically,  we write the post-control
supply function as
        P  -    (0^/0)1/7 + C(Ci,qi)                          (B.10)
where q^ is the total output of all model plants of type i.
     The function  C(C.j_,q^)  shifts segments of the pre-control-
domestic supply curve vertically by C^.  The width or horizontal
distance of each segment is q^.  The resulting segmented post-
control domestic supply curve is illustrated in Figure B-l as S2/
compared with pre-control supply S.
      The supply curves in Figure B-l are drawn as linear functions for ease
of exposition.  Because the supply curves are specified as Cobb-Douglas, they
are log-linear.

                                B-3

-------
                            ci
                           Figure B-l.
   Domestic Market Supply Shift Due to Emission Control  Costs

Impact on Market Price and Quantity
                                                •
     The impacts of the alternative standards on market output
are estimated by solving for post-control market equilibrium and
then comparing that output level,  Q2,  to the pre-control output
level, Q.]_.  Because post-control domestic supply is segmented, a
special iterative algorithm was developed to solve for post-
control market equilibrium.   The algorithm first searches for the
segment in the post-control  supply function at which equilibrium
occurs and then solves for the post-control market price that
clears the market.
     Since the market clearing price occurs where demand equals
post-control domestic supply plus foreign supply,  the algorithm
simultaneously solves for the following post-control variables.
     •  Equilibrium market price.
     •  Equilibrium market quantity.
     •  The quantity supplied by domestic producers.
     •  The net quantity supplied by foreign producers.
                               B-4

-------
     We assess the market impacts of control costs by comparing
baseline values to post-control values for each of the variables
listed above.
Trade Impacts
     We report trade impacts as the change in both the volume and
dollar value of net exports.  We assume that exports comprise an
equivalent percentage of domestic production in the pre- and
post-control markets.  We also assume that foreign and domestic
supply elasticities are the same.  As the volume of imports rises
and the volume of exports falls, the volume of net exports will
decline.  However, if demand is inelastic, it is uncertain
whether the dollar value of net exports will rise or fall.  The
dollar value of imports will increase due to increases in both
volume and price.  Exports will decrease in volume, but price
will increase.  If demand is inelastic then the dollar value of
exports will increase.  If the increase in the dollar value of
exports is greater than that of imports then the alternative will
result in an increase in the dollar value of net exports.
     We use the following algorithms to compute trade impacts:
     Change in volume of imports
           Qsf2 - Qs*                                       (B.ll.a)

     Change in dollar value of imports

                                of                       (B.ll.b)
     Change in volume of exports =

      Q* (Qgt - Qs?)                                          (B.n.c)
     Change in dollar value of exports

                                                          (B.ll.d)
      Q."
                               B-5

-------
where the subscript e references exports by domestic producers.

     We report the change in the volume of net exports as
'(B.ll.c) minus (B.ll.a).   We report the change in the dollar
value of net exports as the difference between (B.ll.d) and
(B.ll.b).
     We also report the change in the dollar value of shipments
by domestic producers.  This value, AVS, is given by

     AVS = P2-Qsd2

Plant Closures
     We predict that any plant will close if its post-control
supply price is higher than the post-control equilibrium price.
Post-control supply prices are computed by Equation  (B.10).  We
round fractions of plant closures to the nearest integer.
CHANGES IN ECONOMIC SURPLUS
     The shift in market equilibrium will have impacts on  the
economic welfare of three groups:
     •    Consumers.
     •    Producers.
     •    Society at large.
The procedure for estimating the welfare change for each group is
presented below.  The total change in economic surplus, which is
taken as an approximation to economic costs, is computed as the
sum of the surplus changes for the three groups.
Change in Consumer Surplus
     Consumers will bear a dead weight loss associated with the
reduction in output.  This loss represents the amount over the
pre-control price that consumers would have been willing to pay
for the eliminated output.  This surplus change is given by:

         Ql
         1   (Q/a)1/e dQ - P! •  (Qi-Q2)                       (B.13)
         Q2
     In addition, consumers will have to pay a higher price for
post-control output.  This surplus change is given by:
                               B-6

-------
           (P2 - Px) • Q2                                    (B.14)

     The total impact on consumer surplus, ACS, is given by
 (B.13) plus  (B.14).  Specifically,

              Ql
        ACS  = i    (Q/a)1/6 dQ - P^  + P2Q2                  (B.15)
              Q2
     This change, ACS, includes losses of surplus incurred by
 foreign consumers.  In this report we are only concerned with
 domestic surplus changes.  We have no method for identifying the
 marginal consumer as foreign or domestic.
     To estimate the change in domestic consumer surplus we
 assume that  total consumer surplus is split between foreign and
 domestic consumers in the same proportion that sales are split
 between foreign and domestic consumers in the pre-control market.
 That is, the change in domestic consumer surplus, ACS, is:
                       ACS                                 (B.16)
     While ACS is a measure of the consumer surplus change from
the perspective of the world economy, ACS^ represents the
consumer surplus change from the perspective of the domestic
economy.
Chancre in Producer Surplus
     To examine the effect on producers, output can be divided
into two components:
     •    Output eliminated as a result of controls.
     •    Remaining output of controlled plants.
The total change in producer surplus is given by the sum of the
two components.
     Note that post-tax measures of surplus changes are required
to estimate the impacts of controls on producers' welfare.  The
post-tax surplus change is computed by multiplying the pre-tax
surplus change by a factor of (1-t) where t is the marginal tax
                               B-7

-------
rate.  As a result, every one dollar of post-tax loss in producer
surplus will be associated with a complimentary loss of t/(l-t)
dollars in tax revenues.
     Output eliminated as a result of control costs causes
producers to suffer a dead-weight loss in surplus analogous to
the dead-weight loss in consumer surplus.  The post-tax
deadweight loss is given by:
                         si
                                        (1-t)
                                                           (B.17)
     Plants remaining in operation after controls realize a
welfare gain of P2 — P]_ on each unit of output, but incur a per
unit welfare loss of C^.  Thus, the post-tax loss in producer
surplus for m model plant types remaining in the market is
                       m
                                 (l-t)
              (B.18)
     The total post-tax change in producer surplus, APS, is given
by the sum of  (B.17)  and  (B.18).  Specifically,
     APS =
PiQ.
'*!
£
ciqi
(l-t)
                                                           (B.19)
     Recall that we are interested only in domestic surplus
changes.  For this reason we do not include the welfare gain
experienced by foreign producers due to higher prices.  This
procedure treats higher prices paid for imports as a dead-weight
loss in consumer surplus.  Higher prices paid to foreign
producers represent a transfer from the perspective of the world
economy, but a welfare loss from the perspective of the domestic
economy.
                               B-8

-------
Residual Effect on Society
     The changes in economic surplus, as measured above, must  be
adjusted to account for two effects which cannot be attributed
specifically to consumers and producers.  These two effects  are
caused by tax impacts and differences between private and  social
discounts rates.
     Two adjustments for tax impacts are required.  First, per
unit control costs C^, which are required to predict post-control
market equilibrium, reflect after-tax control costs.  The  true
resource costs of emissions controls, however, must be measured
on a pre-tax basis.  For example, if after-tax control costs
exceed pre-tax control costs, Cj_ overstates the true resource
costs of controlling emissions.
     A second tax-related adjustment is required because changes
in producer surplus have been reduced by a factor of  (1-t) to
reflect the after-tax welfare impacts of emissions control costs
on affected plants.  As was noted earlier, a one dollar loss in
pre-tax producer surplus imposes an after-tax burden on the
affected plant of  (1-t) dollars.  In turn, a one dollar loss in
after-tax producer surplus causes a complimentary loss of  t/(l-t)
dollars in tax revenues.
     A second adjustment is required because of the difference
between private and social discount rates.  The rate used  to
shift the supply curve reflects the private discount rate  (or  the
marginal cost of capital to affected firms).  This rate must be
used to predict the market impacts associated with emission
controls.  The economic costs of the NESHAP, however, must be
computed at a rate reflecting the social cost of capital.  This
rate is intended to reflect the social opportunity cost of
resources displaced by investments in emission controls.4
     The adjustment for the two tax effects and the social cost
of capital, which we refer to as the residual change in surplus,
ARS, is given by:
    4 See Section 7 for a more detailed discussion of this issue.

                               B-9

-------
            m
     ARS = -    (Ci-pc-jjqi  + APS- [t/(l-tn                  (B.20)
where     pc.j_ =  per unit cost of controls for model plant type
                 i, computed as in (B.5) with t=0 and r=social
                 cost of capital .

     The first term on the right -hand -side of (B.20) adjusts for
the difference between pre- and post -tax differences in emission
control costs and for the difference between private and social
discount rates.  Note that these adjustments are required only on
post- control output.  The second term on the right -hand- side of
(B.19)  is the complimentary transfer of the sum of all post -tax
producer surplus .
Total Economic Costs
     The total economic costs, EC, is given by the sum of changes
in consumer and producer surplus plus the change in residual
surplus.  Specifically,
          EC  =  ACSd + APS + ARS       '                   (B.21)
LABOR AND ENERGY IMPACTS
     Our estimates of the labor and energy impacts associated
with the alternative standards are based on input -output ratios
and estimated changes in domestic production.
Labor Impacts
     Labor impacts, measured as the number of jobs lost due to
domestic output reductions, are computed as
              Pi «£ - Q8d2 > Ll                               (B.22)
                       .
                   2000

where AL is the change in employment,  L^_ is the production
worker hours per dollar of output, and all else is as previously
defined.  The number 2000 is used to translate production worker
hours into jobs  (i.e., we assume a 2000 hour work year) .
                               B-10

-------
       Imacts
     We measure  the  energy impacts associated with the
alternative standards as the reduction in expenditures on energy
inputs due to output reductions.  The method we  employ is similar
to the procedure described above for computing labor impacts.
Specifically,

        AE=E1P1(QgdL-Qd2)                                  (B.23)

where AE  is the change  in expenditures on energy  inputs,  E-j_  is
the baseline expenditure on energy input per dollar output and
all else is as previously defined.
BASELINE INPUTS
     The partial equilibrium model described above requires,  as
inputs, data on  the  characteristics of affected  plants and
baseline values  for  variables and parameters that  characterize
each market.  The characteristics of affected plants have been
described earlier in Appendix A.  These include  the number of
plants by model  type and a measure of output for each model
plant .  Appendix A also reports estimates of capital and annual
emission control costs.
     Table B-l reports the baseline values of variables and
parameters for each  market.  The baseline price  of BLR is taken
from the Chemical  Economic Handbook (p. 580.601G);  the baseline
price of WSR is  from Synthetic Organic Chemicals.5 ^  Baseline
domestic output  in each market is computed as the  sum of
production at all  domestic plants (see Appendix  A  from production
rates at BLR and WSR plants) .
     The import  and  export ratios reported in Table B-l were
computed from production and trade data for BLR  (unmodified epoxy
resins) and WSR  (epi-based non-nylon polyamide)  reported in
    ^ Prices were converted to 1992 dollars using the producer price index
for chemicals and allied products  (Economic Report of the President).
    ° Wet  strength resin prices were converted from a dry weight to wet
weight basis using a wet-to-dry weight conversion factor of 7.7 (computed from
data in the Chemical Economics Handbook, p. 580.1000X).
                                B-ll

-------
                 TABLE B-l.  BASELINE INPUTS
Variable/Parameter
Price (P^3-
Domestic Output ^
Import Ratioc
Export Ratiod
Supply Elasticity (e)
Demand Elasticity (7)
Tax Rate (t)
Private Discount Rate (r)
Social Discount Rate
Equipment Life (T) e
Labor (L-,)f
Energy (E-^S
MARKET
BLR
$2.59
135.0
0.028
0.178
3.76
-1.50
0.25
0.1
0.07
10
0.0025
0.031
Wet Strength Resin
$.20
197.2
0
0.010
3.76
-0.924
0.25
0.1
0.07
10
0.0025
0.031
Notes:
Dollars (1992) per kilogram  (wet weight).
"  Thousands of metric tons  (wet weight) .
*~  Imports divided by domestic production.
°-  Exports divided by domestic production.
®  Years.
f  Production worker hours per dollar of output
9  Energy expenditure per dollar of output.
                             B-12

-------
Chemical Economics Handbook  (pp. 580.601M, 580.601K and
580.1000Y).  Imports of WSR were reported to have been
"insignificant."
     We describe the data and procedures employed to estimate
supply and demand elasticities  (7 and e, respectively) in
Appendix C.  Note that we use the estimates of the demand
elasticities reported in Table B-l for the "base case" results
presented in Sections 4, 5, and 7 of this report.  We assess the
sensitivity of the estimated impacts to demand elasticity by
reporting in Appendix D results based on "low" and "high"
estimates.
     We use a marginal tax rate of 25 percent to assess the
impacts of emission controls.  We adopt a 10 percent private
discount rate  (real marginal cost of capital) and a 7.0 percent
social discount rate.  The expected life of emission control
equipment is 10 years.
     Finally, the values for labor hours per unit of output  (L1)
and energy use per unit of output (E-^) were obtained from the
Annual Survey of Manufactures.  Data from the ASM used to derive
these estimates include 1989 annual values for total production
worker hours used, total expenditures on energy; and the value of
shipments.  Recall that these data are available at the 4-digit
SIC code level.  BLR, WSR and other-resin products are included in
SIC code 2821.  For this reason, 1^ and E^_ are the same in both
resin markets.
CAPITAL AVAILABILITY ANALYSIS
     Pre- and post-control values of the following financial
measures are compared in the capital availability analyses:
     •    Net income/assets.
     •    Long-term debt/long-term debt plus equity.
Pre-Control Financial Measures
     Pre-control measures of net income and net income/assets are
computed by averaging data for the period 1988 through 1991 where
these data are available.  The long-term debt ratio is computed
from 1991 data, or the most recent year available.

                               B-13

-------
     All figures are adjusted to 1991 dollars by the producer
price index for chemicals and allied products.  Then, pre-control
values are estimated by:
   i)
n
 1991
  £    n±/4
i=1988
(B.24)
  ii)
 1991
  I
i=1988
                (n/a)/4
(B.25)
 iii)  1

where  n

      ni
       r

      ai
       1

   11987
   e!987
                        -1991
                                                      (B.26)
        average net income
        net income in year i
        average return on assets
        assets in year i
        long-term debt ratio
        long-term debt in 1991
        equity in 1991
Post-Control Values
     To determine the impact of controls, an estimate of the cost
of controls is made.  In order to get an idea of the steady-state
cost, an annualized cost is used.  The annualized cost, AC, for a
plant is:
     AC    =   V + kS                                        (B.27)
where the variables are as defined previously.
     Annualized costs and capital costs are estimated for each
model plant type.  For each establishment, post-control measures
are given by:
 pn =
where pn
              1991  ni -AC
               £    -±—
             i=1988   4
        post-control average net income
                                                     (B.28)
                               B-14

-------
     =  1991  (ni-AC)/(ai+k)                           (B>2g)
       i=1988       4

  pi =    Il991+k                                     (B.30)
AC   =   annualized cost for the company
pr   =   post -control return on assets
 k   =   capital cost for the company
pi   =   post -control long-term debt ratio
                         B-15

-------
               APPENDIX C






ESTIMATION OF INDUSTRY SUPPLY AND DEMAND

-------
                            APPENDIX C
             ESTIMATION OP INDUSTRY SUPPLY AND DEMAND

INTRODUCTION
     This appendix describes the analytical approach and the data
we employed to estimate the supply and demand elasticities used in
this EIA.  We also report and evaluate the statistical properties
of the estimates.
APPROACH
     The approaches we adopt to estimate supply and demand
elasticities are consistent with economic theory and, at the same
time, exploit the available data.  Briefly, we derive an
industrywide estimate of supply elasticity from an estimated
production function.  Because the data required to estimate the
production function are available only at a four-digit SIC level
(SIC 2821 which includes both the epoxy and WSR industries), we
obtain a single estimate of supply elasticity.  We adopt this
single estimate for both the BLR and WSR industries, implicitly
assuming that the two industries face similar production
functions.
     Because both BLR and WSR are used as intermediate inputs to
produce other goods, the demand for these inputs is derived from
the goods they are used to produce.  The data required to estimate
the derived demand functions are available separately for both BLR
and WSR.  As a result, we obtain estimates of demand elasticities
for each of the two industry segments.
Supply Elasticity
     As noted above, we derive an estimate of the market supply
elasticity from an industry-wide estimate of the production
function.  Given the production function, we solve for the dual
cost function.  Then,  exploiting the result that market price is
established at marginal production cost,  we derive the inverse
supply curve as the derivative of the "cost function with respect
to output.  The important result is that the parameters of the
supply function can be stated in terms of the parameters of the
estimated production function.
                                C-l

-------
     We assume that the industry is economically efficient  in  that
production costs are minimized subject to a production constraint.
In equation form, this can be written as:
         minimize   E r^x.^                                    (C.I)
            xi

          subject to:     Q =  f(x.j_)

where     x^   =   factor  inputs  (used to produce resins)
          rj_   =   factor  prices
          Q   =   output  (of resins)

     The solution to this problem is a set of input demand
functions:
          xj   =   g(r±,Q)                                     (C.2)
     If the input demand  functions are substituted back into the
objective function, one obtains a cost function in terms of input
prices and output.
          C   =   h(ri?Q)                                     (C.3)
     Equilibrium in the market is established at the point  where
price equals marginal cost.  That is:
          P   =   dC/dQ   =  h' (rif-Q)                          (C.4)
where P is output price.  Equation  (C.4) is a relationship  between
output and output price and thus represents the industry supply
curve.
     An explicit functional expression for the right-hand side of
(C.4) can be determined if one makes a specific assumption  on  the
form of the production function.  For this analysis, we assume a
multiplicative form for the production function with two variable
inputs and a capital factor.  Because we use time series data  to
estimate the production function, we also include a time factor to
account for changes in technology.  Specifically,

           aK  X  aL  otM                                    ir  en
     Qt = AKt  tA Lt  Mt                                     l<-.a)
                               C-2

-------
where     Qt   is  industry output in year t
          Kt   is  real  capital stock in year t
          Lt   is  production man-hours in year t
          Mfc   is  an index of materials input in year t
          t    is   time in years
          A, QfL,aM,X are  parameters to be estimated.
     Equation  (C.5)  can be written in linear form by taking the
natural logarithms of  both sides.  Thus, linear regression
techniques  can be applied.
     Given  a particular form for the production function,  the
steps described by Equations (C.2) to  (C.4) can be used to derive
the implied supply function.  For this analysis,  we assume that
capital stock  is  fixed.7  The derived supply function can be
written as:
       In Q   =   B0 + 7 In P + B2 In K                      (C.6)
                + B3 lnPL + B4 In PM + B5  In t
where     PL   =   factor price of labor  input
          PM   =   factor price of material input
          K    =   fixed real capital stock
The Bj_ and  7 coefficients are functions  of the oc^,  the
coefficients of the production function.  For example,  7,  the
supply price elasticity,  can be shown to be equal to
     y-                                                     (C.7)
          --
     It is  clear from (C.7)  that it may be necessary to place
restrictions  on  the  estimated coefficients of the  production
function in order to have well-defined supply function
coefficients.  For example,  the sum of the coefficients for labor
and materials  should be less than one.  Otherwise,  7 is undefined
or, if both coefficients are positive, 7 would be  negative.   For
this reason,  the production function is estimated  with the
      This specification, which treats in place capital as  sunk,  is consis-
tent with our objective of modeling how supply adjusts to price changes in the
post-control market.  This response will depend on the behavior of avoidable
production costs and emission control costs.

                                C-3

-------
restriction that the  sum of  coefficients for the inputs should
equal one.  This is equivalent  to assuming long-run constant
costs in the industry, an assumption that seems reasonable on a
priori grounds and appears to be consistent with the data.*3
       Functions
     WSR is used primarily  in the production of pulp and paper
products (SIC 2851) .  BLR is  used to produce a variety of
products; about 50  percent  of industry output is used to sealants
and adhesives  (SIC  2621) and  coatings and paints (SIC 2891) .  As
intermediate inputs,  the demand for both WSR and BLR are derived
from the demand for the products they are used to produce.
     We assume that firms using WSR and BLR as inputs attempt to
maximize profits subject to a production constraint.  The profit
function can be written

     Max TT = P_-g(Q7W) - P-Q  - r -w                          (C.8)
      Q,W     e                  W
where     TT   =   profit;
          Pe  =   the price of the final good  (e.g., pulp and
                  paper products) ;
          Q   =   input use of WSR or BLR;
          W   =   a vector  of other inputs
          P   =   the price of WSR or BLR; and,
          rw  =   a vector  of prices of other inputs.

Note that the function g(Q,W)  defines the production function for
the end product, say Qe.

     The solution to C.8 yields a system of input demand
equations of the form
     Q = h(P,Pe,rw)                                           (C.9)
      The unrestricted estimates of the production function coefficients
summed nearly to unity.  Thus, the restriction on the coefficients is only
marginally binding.

                                C-4

-------
In words, C.9 states that the derived demand for WSR or BLR
depends on its own price, the price of the final good, and the
prices, of other inputs.
     We adopt a multiplicative function form for equation C.9.
Specifically, we write the derived demand function as
where     Q   =   the quantity demanded of WSR or BLR;
          P   =   the price of WSR or BLR;
          Pe  =   the price of the end product; and,
          B,  /?,  0e are  parameters to  be  estimated.

The parameter /3, of course, is the demand elasticity for the
input — either WSR or BLR.
     Note that equation C.10 excludes variables for the prices of
other inputs  (rw of equation C.9).  Unfortunately, data on these
prices are unavailable.  This requires us to adopt the implicit
assumption that the use of WSR and BLR in end products is fixed
by technology.
     Because the markets for WSR and BLR are simultaneous in P
and Q, it is necessary to apply a systems estimator in order to
obtain consistent estimates of the coefficients for the demand
equations.  We employ a two- stage least squares estimator (2SLS)
to estimate the demand equations.  In order to estimate consis-
tent demand equation coefficients, one uses as instruments the
exogenous variables included in the system of demand and supply
equations.  The supply- side instruments used to estimate the
demand functions include capital stock (K) ,  a cost index (Pv)
measuring the weighted -average cost of variable inputs (labor and
materials) , and time.
DATA
     Table C-l identifies the variable names, units of measure,
and variable descriptions for the data available for the
analysis.  Those variables directly relat'ed to a specific SIC

                               C-5

-------
TABLE C-l.   VARIABLES AND DEFINITIONS OF  PRIMARY  DATA

1.
2.
3.

5.
6.
7.
8.
9.
10.
11.
12.
Variable*
YEAR
SIC
PISHIP
VSHIP
CAP
COSTMAT
PIMAT
PRODW
PRODH
PRICE
SALES
IPD
Unit
—
4 -digit
index
millions $
millions
1972 $
millions $
index
millions $
millions
hours
dollars per
kilogram
millions of
kilograms
index
Description
Observation identifier, 1958-1989
Industry identifier
Producer price index for Value of
Shipments (SICs 2821, 2891 and 2851)
Value of industry shipments
Real capital stock (SIC 2821)
Cost of materials inputs (SIC 2821)
Price index for materials inputs (SIC
2821)
Production worker wages (SIC 2821)
Production worker hours (SIC 2821)
Price per kilogram (Type A Liquid
Resin, BLR Resin and Non-Nylon
Polyamide Resin)
Quantity sold by domestic producers
annually (BLR use in protective
coatings, BLR use in bonding and
adhesives and Non-Nylon Polyamide
Resin)
Implicit Price Deflator (1.0 in 1972)
         Items 1-9 obtained from the ASM.  Items 10 and 11 obtained
         from the ITC and the SPI.  Item 12 obtained from 1991
         Economic Report of the President.
                             C-6

-------
were obtained from the Annual Survey of Manufactures  (ASM).9
These data are defined for 4-digit SICs and represent  annual
values which cover the years 1958-1989.  Recall  that both BLR and
WSR belong to SIC 2821 code.  Industry segment price and output
data, obtained from the ITC and SPI, were used to  estimate demand
elasticities.  These data are available for the  years  1971-1990.
     Items 1 through 9 of Table C-l were used to estimate the
production function (see Equation C.5) for SIC 2821.   We formed
the industry output variable, Q, as VSHIP/PISHIP;  this ratio
yields the real value of shipments in SIC 2821.  The capital
stock variable, K, is measured as CAP, the real  value  of capital
stock in millions of 1972 dollars.  Labor input, L, is measured
as PRODH, millions of production worker hours.   The time trend,
t, is measured by the variable YEAR.  Finally, we  measure
materials use, M, as the ratio of COSTMAT/PIMAT;
this ratio yields the real cost of material inputs for SIC 2821
in millions of 1972 dollars.
     Items 3, 10, 11 and 12 were used to estimate  the  derived
demand equations.  The dependent variable for the  WSR  equation,
quantity demanded, is measured as sales of nonnylon polyamides in
millions of kilograms per year.  The "real" price  variable for
this equation is measured as the nominal price of  non-nylon
polyamides (item 10) divided by IPD  (item 12).   The "real" price
of the end-product good is measured as PISHIP for  SIC  2821 (pulp
and paper) divided by IPD.
     We estimate two derived demand functions for  BLR — one for
BLR use in coatings and paints and another for BLR use in
sealants and adhesives.  The dependent variables for each
equation are measured as sales for the respective  use  in millions
of kilograms annually.  The real price variable  for BLR use in
coatings and paints is measured as the nominal price of Type  A
liquid resin divided by IPD; the real price for  BLR use in
coatings and adhesives is measured as the nominal  price of BLR
      We thank Eric Bartlesman of the Federal Reserve Board for providing the
data set to us.

                               C-7

-------
resin divided by IPD.  Both price variables are expressed as
dollars per kilogram.  Finally, the real prices of end-products
are formed by the ratios of PISHIP/IPD for SICs 2891 (coatings
and paints) and 2621 (sealants and adhesives).
     The 2SLS estimates of the derived demand equations require
data for three instrumental variables — time, capital stock and a
cost index for variable inputs.  Time and capital stock are
measured as the variables YEAR and CAP (for SIC 2821).   We form
the cost index for variable inputs as a weighted index of PIMAT
and PRODH  (for SIC 2821), expressed in constant 1972 dollars.
STATISTICAL RESULTS
Production Function/Supply Equation
     A restricted least squares estimator was used to estimate
the coefficients of the production function shown in Equation
(C.5).   A log-linear specification was estimated with the sum of
the a.j_ restricted to unity.  The results are shown in Table C-2.
The equation explains about 96 percent of the variation in the
output variable.  While the coefficients on labor and time are
significant at the 99' percent confidence level, the coefficients
on capital and materials are not statistically significant at
conventional confidence levels.
     Using the estimated coefficients reported in Table C-2 and
the result shown in Equation C.7, we derive a supply elasticity
estimate of 3.76.  Note that the calculation of statistical
significance for the elasticity is not straightforward since it
is a non-linear function of the production function coefficients.
No attempt has been made to assess the statistical significance
of the estimated elasticity.
Demand Equations
     Table C-3 reports estimates of the derived demand equations
for WSR and BLR.  The reported coefficients are 2SLS estimates of
the parameters of Equation C.10.  We have also corrected the
estimates of all three equations for first-order serial
                               C-8

-------
TABLE C-2.
ESTIMATED PRODUCTION FUNCTION COEFFICIENTS
   (t-ratios  in parentheses)
Industry
SIC 2821
Time
.323
(7.118)
Capital
.211
(.632)
Labor
.485
(3.036)
Materials
.304
(1.552)
Adjusted R2
.96
                         C-9

-------
TABLE C-3.
2SLS ESTIMATED DERIVED DEMAND COEFFICIENTS
   (t-ratios  in parentheses)
Industry
Wet strength resin
BLR (in coatings and
paints)
BLR (in sealants and
adhesives')
Own Price
.. (0)
-.924
(4.363)
-1.474
(1.780)
-1.481
(2.620)
End- Product Price
(j8^)
1.136
(1.023)
.097
(.115)
2.040
(2.445)
                         C-10

-------
correlation using the Prais-Winston algorithm10  and the two BLR
equations  for heteroschedasticity.
     We  have estimated the  derived demand function for WSR
consistently with the approach described earlier in this-
appendix.   The estimated own-price coefficient is correctly
signed and highly significant.   The estimated coefficient on the
end-product price (SIC 2621)  is correctly signed but not
statistically significant.11
     The estimated own-price  coefficients for BLR are sensitive
to the instruments used in  the two-stage procedure and to
corrections for autocorrelated errors.  As a result,  it was
necessary  to modify the general approach described earlier in
this appendix.  Specifically,  the estimated equation for BLR used
in coatings and paints includes only the cost index for variable
inputs and includes a time  trend variable as an  explanatory
variable.12  The estimated  equation for BLR used in sealants
and includes both time and  the variable cost index as instru-
ments, but not the capital  stock variable.
     The estimated own-price  coefficients are —1.474 and —1.481,
respectively,  for BLR used  in coatings and paints,  and in
sealants and adhesives.  Accordingly,  we adopt a mid-point demand
elasticity of —1.5 for BLR.   We caution, however,  that this
estimate is not robust.  As noted above, the estimates for BLR
are sensitive to the specification of  instrumental variables and
to corrections for autocorrelated errors.
     We  acknowledge the uncertainty in our estimate of demand
elasticities for WSR and especially for BLR.  Accordingly, we
       The  Prais-Winston algorithm is similar to the more familiar Cochrane-
Orcutt estimator.  However, unlike the Cochrane-Orcutt method, the Prais-
Winston algorithm does not skip the first observation and uses the full
generalized  least squares (GLS) transformation.

       Note that we do not report adjusted R2 for the derived demand equa-
tions.  First, Basemann (1962) warns that low multiple correlation coeffi-
cients for simultaneous equation estimators are not evidence of poor fit or
lack of joint significance of the set of explanatory variables.  Second, the
correction for autoregressive errors renders R2 meaningless.
    12 The  estimated coefficient for the time trend variable is .468 and the
associated t-ratio is 5.104.

                                C-ll

-------
assess  the sensitivity  of our estimated economic  impacts by
reporting in Appendix D results corresponding to  "low" and "high"
demand  elasticity cases.   The low demand elasticities are —.50
and —.62,  respectively,  for WSR and BLR; the corresponding high
demand  elasticities are —1.34 and —3.10.  The low and high demand
elasticities are, respectively, minus  and plus two standard
deviations of the mid-point estimates.1^
     ^ we use  the standard error of the estimate for DGEBPA used in coatings
and paints for  the high demand elasticity case and the standard error for
DGEBPA use in sealants and adhesives for the low demand elasticity case.  This
procedure causes us to use relatively higher demand elasticities for both
cases, thus representing "worst" case scenarios.

                                 C-12

-------
     APPENDIX D
SENSITIVITY ANALYSES

-------
                            APPENDIX D
                       SENSITIVITY ANALYSES

INTRODUCTION
     This appendix presents the results of a sensitivity analysis
that explores the degree to which the results presented earlier
in this report are sensitive to estimates of demand elasticity.
SENSITIVITY ANALYSIS:  DEMAND ELASTICITY
     The "base case" results presented earlier in this report are
based on demand elasticities of -1.50 for BLR and -.92 for WSR.
Below, we report results for "low" and "high" demand elasticity
cases.  These alternative cases use the following values for
demand elasticities:
     •    Low demand elasticity:   —.62 for BLR and —0.50  for WSR.
     •    High demand elasticity:   -3.10  for BLR and -1.34  for
          WSR.

     The greater the elasticity of demand  (in absolute value),
the more consumers will reduce the quantity they purchase in
response to a given change in price.  Therefore, we expect that
when we use a higher demand elasticity in the partial equilibrium
analysis, the reduction in market output will be greater and the
price change will be smaller than in the base case.  Similarly,
when we use a lower elasticity, we expect the change in price to
be greater, and the change in market quantity to be smaller,
relative to the base case.
     Tables D-l  through D-4 present estimates of the primary
economic impacts associated with the NESHAP for each of the two
industry segments in the case of low and high demand
elasticities.  Tables D-l and D-2 report results based on low
demand elasticities and Tables D-3 and D-4 report results based
on high demand elasticities.
     In general, the results of the sensitivity analysis are
consistent with the base case results presented earlier in this
report.   For the BLR industry,  no plant closures are predicted,

                               D-l

-------
TABLE D-l.   SENSITIVITY ANALYSIS:   ESTIMATED PRIMARY IMPACTS
       ON THE BLR MARKET WITH  LOW ELASTICITY OF DEMAND



Price
Change (%)
.06
Market
Output
Change
-.04
Change in Value of
Domestic Shipments

($1,000
1992)
51


(%)
.01



Plant
Closures
.00
      Note:  Results are based on a demand elasticity of -.62.
                              D-2

-------
TABLE D-2.   SENSITIVITY ANALYSIS:   ESTIMATED  PRIMARY IMPACTS
       ON THE WSR MARKET WITH LOW ELASTICITY OF DEMAND

Regulatory-
Option
MACT Floor
Option I

Price
Change
(%)
4.60
0.24
Market
Output
Change
(%)
-2.23
- .12
_ Change in Value of
Domestic Shipments
($1,000
1992)
897
48
(%)
2.28
.12

Plant
Closures
.02
.38
   Note:  Results are based on a demand elasticity of -0.50
                              D-3

-------
TABLE D-3.   SENSITIVITY ANALYSIS:   ESTIMATED PRIMARY  IMPACTS
      ON THE BLR MARKET WITH HIGH  ELASTICITY OF DEMAND

Price
Change (%)
.04
Market
Output
Change (%)
-.12
Change in Value of
Domestic Shipments
($1,000
1992)
-293
(%)
-.08

Plant
Closures
.00
       Note:  Results are based on a demand elasticity of -3.10
                              D-4

-------
TABLE D-4.   SENSITIVITY  ANALYSIS:  ESTIMATED PRIMARY IMPACTS
      ON  THE WSR MARKET  WITH HIGH ELASTICITY OF DEMAND

Regulatory
Option
MACT Floor
Option I

Price
Change
(%)
3.86
.20
Market
Output
Change
(%)
-4.95
-.27
Change in Value of
Domestic Shipments
($1,000
1992)
-505
-27
(%)
-1.28
-.07

Plant
Closures
.84
.05
         Note:  Results are based on a demand elasticity of -1.34.
                              D-5

-------
and even in the high demand elasticity case, the estimated

reduction in market output is just 0.12 percent.  However, for

the sensitivity analysis of the WSR market, when a low elasticity

of demand is employed, the plant closure predicted in the

previous analysis is less probable.  Also, when a "low"

elasticity of demand is assumed, the impacts on domestic
production, the value of domestic production, net exports,

employment and energy are reduced.  The estimated impacts of

Option I on the WSR industry are very small, even when a high

elasticity of demand is assumed.


REFERENCES

American Metal Market (1992).  June 3, p. 4.
Basemann, R.L. (1962).  Letter to the Editor, Econometrica, Vol.
     30, No. 4, October.

Chemical Economics Handbook  (1991).  Marketing Research Report
     Epoxy Resins by Michael J. Haley with R. Mulach and Y.
     Sakuma. January

Chemical Economics Handbook  (1989).  Marketing Research Report
     Polyamide Resins (Non-Nylon Types) by Renita A. Gurule and
     David A. Tong with R. Mulach and M. Tashiro.  March.

Chemical Economics Handbook  (1991).  Marketing Research Report
     Epoxy Surface Coatings by Eric Linak with F. Stahel and Y.
     Ishikawa. October

Chemical Marketing Reporter  (1992).  June 8, pg. 15.

Chemical Marketing Reporter  (1990).  September 17, pg. 29.

Hydrocarbon Processing  (1990).  April, p. 25.

Just, R.E., D.L.  Hueth and A. Schmitz  (1982).  Applied Welfare
     Economics and Public Policy. Prentice Hall, Inc., Englewood
     Cliffs, NJ.

Kirk-Othmer Encyclopedia of Chemical Technology  (1985).  Third
     edition, volumes 9, 16, and 6.

Kolb, J.A. and J.D. Scheraga  (1990).  "Discounting the Benefits
     and Costs of Environmental Regulations."  Journal of Policy
     Analysis and Management. Vol. 9, No. 3, Summer.
                               D-6

-------
Lind, R.C., et al. .  (1982).  Discounting for Time and Risk  in
     Energy Policy. Resources for the Future, Inc., Washington,
     DC.

Mathtech  (1985a).  Corporate Income Taxes, Social Costs, and Dis-
     tributional Impacts:  Implications for EIA's.  Prepared for
     the Office of Air Quality Planning and Standards, U.S. Envi-
     ronmental Protection Agency.

Mathtech  (1985b).  Estimated Control Costs and Social Costs:
     Implications  of Investment Tax Credits for EIA's.  Prepared
     for the Office of Air Quality Planning and Standards,  U.S.
     Environmental Protection Agency.

Mishan, E.J. (1971).  Cost-Benefit Analysis. Praeger Publishers,
     Inc,, New York.

Modern Plastics  (1991).  January, pp. 113-122.

Modern Plastics Mid-October Encyclopedia Issue (1991).

Moody's Industrial Manual (1991).

MRI  (1992).  Draft of Industry Profile on Epoxy and Epichloro-
     hydrin based  Polyamide Resins.  Received from L. Sorrels,
     September.

MRI II  (1992).   Draft of the Background Information Document.
     Received from L. Sorrels October 16.

Network Consulting, Inc. (1992).  The U.S. Market for Thermoset
     Resins. Molding Compounds and Processes.  "Highlights  of
     Major Conclusions."

Plastic News (1992.  April 27,  p. 22.

Plastics World (1991).  "Resin Report," January.

Pulp and Paper (1979). Pulp and Paper; Chemistry and Chemical
     Technology.  Third edition Volume III.

SPI I.  Society of the Plastics Industry (1971-1991).  Monthly
     Statistical Report, SPI Committee on Resin Statistics.

SPI II.  Society of the Plastics Industry  (1992).  Telecomm.
     between J. Broyhill and C.  Lawson.

U.S. Department of Commerce (1992).  Bureau of the Census,  Trade
     Data Inquiries and Control Section.  Information obtained by
     request and received in September.
                               D-7

-------
USITC I, Annual Issues (1970-1991).   Synthetic Organic Chemicals
     United States Production and Sales.  Investigation number
     332-135.

Willig, R.D. (1976).  "Consumer Surplus Without Apology," Ameri-
     can Economic Review. 66, No. 4, September, 589-97.
                               D-8

-------
                  APPENDIX E.




EVOLUTION OF THE BACKGROUND INFORMATION DOCUMENT

-------
                           APPENDIX E.

         EVOLUTION OF THE BACKGROUND  INFORMATION DOCUMENT


     The purpose of this study was to provide data to support the
development of the proposed national emission standard for
hazardous air pollutants (NESHAP) for the production of basic
liquid epoxy resins (BLR)  and non-nylon polyamide resins, also
known as wet strength resins (WSR).   To accomplish the objectives
of this program, technical data were gathered on the following
aspects of the industry:  (1) the operation of process equipment,
storage tanks, wastewater treatment,  and transfer equipment
(piping), (2)   the release and controllability of hazardous air
pollutants (HAP's)  emitted into the atmosphere from the above
emission points, and (3)  the types and costs of demonstrated
emission control technologies.   The bulk of the information was
gathered from the following sources:

     1.   Technical literature;

     2.   Plant visits;

     3.   Questionnaires sent to industry;

     4.   Industry representatives;

     5.   State and regional air pollution control agencies; and

     6.   Equipment vendors.

     Significant events relating to the evolution of the
background information document are itemized in Table E-l.
                               E-l

-------
TABLE E-l.  EVOLUTION OF THE BACKGROUND INFORMATION DOCUMENT
Date

6/13/91












7/12/91
7/17/91


7/29/91

9/13/91

4/23/92

6/1/92



10/1/92



5/13/93

Company, consultant, or
agency / 1 ocat i on
Georgia-Pacific
(Ukiah, CA, Eugene, OR,
Crossett, AR, Peachtree City,
GA)
CPS Chemicals
Old Bridge, NJ
Sonoco
Holyoke , MA
Henkel Corporation
Charlotte, NC
Call away Chemical Company
Columbus , GA
Call away Chemical Company
Shreveport , LA
Georgia-Pacific
Columbus, OH, Albany, OR,
Grayling, MI
Pioneer Plastics Corporation
Auburn, ME
The Dow Chemical Company
Freeport , TX
Shell Chemical Company
Deer Park, TX
The Dow Chemical Company
Freeport , TX


Mailed to all known industry-
members and selected vendors


Shell Chemical Company
Deer Park, TX
Nature of action
Section 114 information
request sent by the
U. S. EPA '











































Information sent from plant
regarding the concentration
of epichlorohydrin in
wastewater
Request from U. S. EPA
comment on draft BID
chapters 3, 4,
and 6
Emission test data sent
plant


for



from

                             E-2

-------
                APPENDIX  F.




INDEX TO ENVIRONMENTAL IMPACT CONSIDERATIONS

-------
                           APPENDIX F.

          INDEX TO ENVIRONMENTAL  IMPACT CONSIDERATIONS

     This appendix consists of a reference system which is cross-
linked with the October 21, 1974,  Federal Register (39 FR 37419)
containing the Agency guidelines concerning the preparation of
environmental impact statements.  This index can be used to
identify sections of the document which contain data and
information germane to any portion of the Federal Register
guidelines.
                               F-l

-------
  TABLE P-l.   CROSS-INDEXED REFERENCE SYSTEM TO HIGHLIGHT THE
         ENVIRONMENTAL IMPACT PORTIONS OF THE DOCUMENT
Agency guidelines for
preparing regulatory action
environmental impact
statements (39 FR 37419)
                        Location within Background
                        Information Document
1.   Background and summary
     of regulatory
     alternatives
     Statutory basis for
     proposing standards
                        The statutory basis for proposing
                        standards is summarized in
                        Chapter 2,  Section 2.1.
     Industries affected by
     the regulatory
     alternatives
                        A discussion of the industries
                        affected by the proposed
                        standards is presented in Chapter
                        3.   Further details covering the
                        business and economic nature of
                        the industry are presented in
                        Chapter 8, Section 8.1	   	
     Regulatory
     alternatives
     Control techniques
                        The alternative control
                        techniques are discussed in
                        Chapter 4.
     Regulatory
     alternatives
                        The various regulatory
                        alternatives are defined in
                        Chapter 5, Section 5.2.	
     Environmental impact
     of the regulatory
     alternatives
     Primary impacts
     directly attributable
     to the regulatory
     alternatives
                        The primary impacts on mass
                        emissions and ambient air quality
                        of the regulatory alternatives
                        are discussed in Chapter 5,
                        Section 5.2.  Tables summarizing
                        the environmental impacts are
                        included in Chapter 6.	
     Secondary or induced
     impacts
                        Secondary impacts for the various
                        regulatory alternatives are
                        presented in Chapter 6.	
4
Other considerations
Potential socioeconomic and
inflationary impacts are
discussed in Chapter 8, Section
8.2.
                              F-2

-------
APPENDIX G.  MODEL PLANT CHARACTERISTICS

-------
    APPENDIX G.  MODEL PLANT CHARACTERISTICS - BLR PRODUCTION
     Production Characteristics
     (a) BLR Production
          68,000 Mg/yr (150 x 10^ Ib/yr) design capacity
          45,000 Mg/yr (100 x 106 Ib/yr) actual capacity
     (b) HAP Feedstock Usage
          25,000 Mg/yr (56 x 106 Ib/yr) EPI actual usage
     Operating Time
          Continuous occurring over 8,400 hr/yr
     Emission Stream Characteristics - Process Vents
Emissions
1) Dilutea
2) Richb
Flowrate
40 scfm
30 scfm
Temperature
25°C
60°C
Composition
(Vol %)
0.6 EPI, 1.4 IPA, 0.8 H,0
12 EPI, 40 IPA, 8 H20
aEmissions from reaction and premix stages
"Emissions from resin finishing and purification
4.   Storage Information - EPI
      6   100,000-gallon fixed roof tanks
     10   turnovers/yr each
     Equipment Leaks
     Duration: 7,350 hr/yr
% EPI
Component counts :
Pump seals
Valves
-Liquid
-Gas
Flanges
Open-ended
lines
Sampling
connections
PRV's
1-25
13
444
0
463
5
4
26
26-50
13
783
10
898
0
3
1
51-75
9
520
16
1,158
23
22
18
>75
8
296
37
406
0
8
2
                               G-l

-------
6.    Wastewater

     From steam stripper bottoms and extractor:  40,000  gal/day,
     1,000 mg/L concentration of EPI.

MODEL PLANT CHARACTERISTICS - WET STRENGTH RESIN  (WSR) PRODUCTION

1.    Production Characteristics

     (a) WSR Production

          11,600 Mg/yr  (25.6 x 106 Ib/yr) actual capacity

     (b) HAP Feedstock Usage

          900 Mg/yr  (2 x 106 Ib/yr) EPI actual usage
          663 Mg/yr  (1.4 x 10° Ib/yr) MeOH actual production
          233 Mg/yr  (0.5 x 106 Ib/yr) HC1 actual usage

2 c    Operating Time

           12 hours/batch
            1 batch/day
          175 days/year

3.    Emission Stream Characteristics - Process Vents

     (a) MeOH Distillation Receiver Displacement

          Potential HAP's emitted       =    MeOH

          Flowrate  (scfm)          =    1.5
          T (°C)                   =    20
          Composition  (vol. %)     =    16 MeOH
          Duration  (minutes)       =    120

     (b) Reactor Displacement

          Potential HAP's emitted  =    EPI

          Flowrate  (scfm)          =    7.35
          T (°C)                   =    20
          Composition  (vol. %)     =    1.8 EPI
          Duration  (minutes)       =    15

     (c) Heatup  (Crosslinking Reaction)

          Potential HAP's emitted  =    EPI

          Tx  (°C)                  =    20
          T2  (°C)                  =    50
          Flowrate  (scfm)          =    1.4 x 10"2
                               G-2

-------
          Composition  (vol % @ Tx)  =     1.8 EPI
          Composition  (vol % @ T2)  =     7.9 EPI
          Duration  (minutes)        =     180


     (d) Acid Addition

          Potential HAP's emitted   =     HC1, EPI

          Flowrate  (scfm)           =     8.6
          T  (°C)                    =     50
          P  (mm Hg)                 =     760
          Composition  (vol %)       =     18.5 HC1
                                         7.9 EPI
          Duration  (minutes)        =     5

4.   Storage Information


      EPI        1  20,000-gallon  fixed  roof tank
                 11 turnovers/yr

      MeOH       1  20,000-gallon  fixed  roof tank
                 12 turnovers/yr

      HC1        1  10,000-gallon  fixed  roof tank
                 7  turnovers/yr

5.   Equipment Leaks

     Duration = 2,100 hr/yr
     Component Counts:

      Component	    No./Plant        % HAP

      Pump seals                     2              100
      Flanges                        64              100
      Liquid valves                  14              100
      Gas valves                     1              100
      Pressure relief devices        2              100
      Sample connections             2              100
      Open-ended lines               1              100

6.   Wastewater

       None

Model Plant Designations

     Model Plant 1:  Describes plants which, only  emit  EPI as a
HAP; there are no HC1 or MeOH emissions.  Therefore, the
applicable process vent emission stream  characteristics are 3(b),
(c), and (d) with EPI only.  Also, no storage of MeOH  or HC1 is
considered.
                               G-3

-------
     Model Plant 2;  Describes facilities that use HC1 in acid
addition stages.  Storage of HC1 and the process vent emission
stream characteristics 3(b), (c),  and  (d) with EPI and HC1 are
considered.

     Model Plant 3:  Describes plants which use HC1 and also
produce methanol.   All emission streams are considered.
                               G-4

-------
 APPENDIX H




TABULAR COSTS

-------
 TABULAR COSTS FOR




  EXISTING  SOURCES




BLR SOURCE CATEGORY

-------
Equipment Leaks Options

-------
         2S |
                                          JS JS J! J!

                                          II II

                                          z Z z 2

                                          as !L £ £
                                          22 22

                                          = r r s
                                             f!
SS  ""***3    S«T?SSOSO°     oS£S\e3     S?3S  SSSS^



        S             S g =|         B  § R


                                      glfltl !!,!ll!!I
                                      •sfjlmmmifi

-------
                                                                  1
2S3SS
         SPSSSS
          ?-
          a
II

IS
y ff
a. as
S 2

53
                  **



                  I S



                  II


                  3 S
-X ±

Jl

ff
3 2
                                         as z
                                          s z*
                         ?S
                         gs?
                         as
                             S8S1  S
                                                      12
                                         I*

                                         11,
                                                 ff
                                                til

                                                lit
                                                        ri
                                                   8 A 5
                                                   3
-------
          1  5 ">





          *? 2 2 3 s
          ~ ~ *» «- (
                                              11  1!
                                                            11  11s
                                                            a   a a
'> SK
iss

  3
I &

§1
iill.,
iliili

-------
                                                          883° °°

                                                           2 55 o ^   S3 fiS 99 Q ^o o 21 r- 5s   99999(3 o      o\ H*I^\C->QOO   c~- ~«rs^«ooo
             Sr^^oao   * oft 55  5  • 3 ^ • ^ ^   SS8SS  S       • "i |!#1!^4 fl!?1I?s|??I«|4 * 3   UpSSal   I {1..££?13i ?|1..S§*1'3:  ?a
!                                   ! 1              1                       ?"

-------
                   1*
                   SI
                  *¥S.
                                           s s
f §

£ I
Hi!

-------
88 S S3! 8 Si
gssssis 3

<  a  8s=-
SSS82S5
             a  ss  ss
             *  1:1  si
                » s  n £
t 3 B
• - o!
 as;
: sa 8 25
       '
J£ 5 8 8 S

j aj MJ O O O
                                 I
                                 •s


                                 f,
                                                                                   H

                                                                                   "£ «


                                                                                   8 s

                                                                                   = (3

-------
.Process Vents, Storage, and



     Wastewater Options

-------
Costs for Ciba-Geigy and Dow recycling
scrubber liquor back to process
Ciba-Geigy and Dow
Capital Costs:

Pump with the following specifications (a):
50 GPM, 50 ft head, 1.5 hp, 3550 RPM,
S.G. = 1.0, discharge pipe size = 1"

Piping with the following specifications (b):
400 ft of stainless steel piping, pipe
racks, 1" diameter, schedule 10

Installation Costs (materials & labor) (c):
Assumed the pump installation cost would be
the same as the piping installation costs.
                                                     Cost/Plant
 $1,112
 $2,239
 $4,444
TOTAL CAPITAL COSTS:
 $7,795
Annual Costs:

Electricity:

Capital Recovery Costs:
Assumed 10 year life of pump and piping, and
7 percent interest,

TOTAL ANNUAL COSTS:
 $1,039

 $1,110



 $2,149
Facility

Ciba-Geigy

Dow
Emission
Reduction
  (Ib/yr)

  3,195

  1,930
(a)  Richardson's Cost Manual, Volume IV, p. 100-280,1984.
(b), (c) Richardson's Cost Manual, Volume III, p. 15-47, 1984.

-------
SCRUBBER CALCULATIONS
Ciba Geigy
INPUT:
GAS FLOW RATE (ACFM):                              133
GAS TEMPERATURE (C):                                20
SCRUBBING LIQUOR:                         H2O
MOLWT                                              18
MOLECULAR WEIGHT OF VOC:                          92
REQUIRED EFFICIENCY:                              0.99
GAS VOC CONCENTRATION (PPM):                    11050
SLOPE OF EQUILIBRIUM LINE:                           1.8
ABSORPTION FACTOR:                                 4.5
EMISSION DURATION (MIN):                             60
EMISSIONS/SHIFT:                                       8
SHIFT/DAY:                                            3
DAY/YEAR:                                          365
TMEVAR                                   0.24987872352
OUTPUT:
GAS MOLAR FLOW RATE (LBMOLE/HR):         20.7203180368
GAS MOLECULAR WEIGHT:                        29.69615
GAS DENSITY (LB/FT3):                         0.0771069765
LIQUID MOLAR FLOW RATE (LBMOLES/HR):      167.834576098
LIQUID DENSITY (LB/FT3):                            62.4
CALCULATED ABSCICCA:                      0.17258852531
                          COST EFFECTIVENESS:
                                  3496.84

                          ENERGY(BTU):
                               1469419.31

                          Mass Flux (Ibs/yr)
                                   46107

                          Time Var:
                          0.24987872351939

                          Mg Controlled:    20.723252
READORD:
GAREA AT FLOODING (LB/SECFT2):
GAREA:
ACOLUMN (FT2):
DCOLUMN (FT):

NOG:
0.09690833382
0.47377735865  USING 2 INCH CERAMIC RASCHIG RINGS
0.28426641519
0.60126858271
0.98593738324

5.60148277717
L DOUBLE PRIME(LB/HR.FT2):
SCHMIDT NO. GAS:
SCHMIDT NO. LIQ:
HG:
HL:
HOG:
H COLUMN:
H TOTAL (FEET):

WEIGHT COLUMN (LB):
V PACKING (FT3):

PRESS DROP (LB/FT2.FT):
DELT P TOTAL (IN H20):
5024.41414142
        1.47
        1019
1.71531975393
2.21914209153
2.20846244094
 12.370664327
14.6171486728

729.667307441
9.45750571483

1.51183666295
3.36643629741
"MUST ENTER THESE VALUES*"
"MUST ENTER THESE VALUES***
**
  PACKING CONSTANTS FOR 2 IN RASCHIG RINGS*
"PACKING CONSTANTS FOR 2IN RASHIG RINGS***

"PACKING CONSTANTS FOR 2IN RASHIG RINGS"*
"PACKING CONSTANTS FOR 2IN RASHIG RINGS*"

-------
COST CALCULATIONS

Capital Costs

Direct Costs
ABSORBER PRICE:
PLATFORM AND LADDERS:
PACKING COST:
Enhanced Monitoring Capital Equipment Cost
 $1,803.36
 $1,713.87
  $111.37
 $3,000.00
DUCTWORK:

DUCT PRICE ($ for 100 FT):

FAN PURCHASED COST:
FAN MOTOR

PUMP     (GAL/MIN)
PUMP HP
PUMP HE (FT)
COST
  $934.14

 $1,485.11
  $146.25

    $6.04
    $0.02
   $14.62
  $348.38
STACK PRICE:                                     $1,401.21

PURCHASED EQUIPMENT COSTS:                   $ 10,943.67
10 % INSTRUMENTATION AND CONTROLS:           $ 1,094.37

TOTAL DIRECT CAPITAL COST (7/92) DOLLARS:      $26,483.68
Indirect Costs

Enhanced Monitoring
Initial Performance Test:

Total Indirect Capital Cost
$24,420.00

$24,420.00
TOTAL CAPITAL COST
$50,903.68

-------
ANNUALIZED COSTS:

WATER COST:                                       $951.69 JUNE'85
ELECTRICITY:
FANKWH/yr                       f                 $177.39
PUMP KWH                                       $8,744.71

TOTAL ELECTRICITY COST                          $526.40
OPERATING LABOR:                               $8,562.90
MAINTENANCE LABOR:                            $9,422.48
SUPERVISORY LABOR:                             $1,284.44
MATERIALS:                                      $9,422.48
Enhanced Monitoring Supplies:                            $500.00
Enhanced Monitoring Labor:                           $3,150.00

DIRECT OPERATING COSTS:                       $33,820.38

OVERHEAD:                                     $11,861.89
PROP TAX:                                         $509.04
INSURANCE:                                        $509.04
ADMINISTRATION:                                $1,018.07
Reporting and Recordkeeping Requirements                 $ 17,500
CRF:                                             $7,247.56

TOTAL ANNUALIZED COST:                       $72,465.98

AMOUNT CONTROLLED (MG/YR):                 20.72325222

COST EFFECTIVENESS ($/MG):                  3496.84390787

-------
DOW
Enhanced Monitoring Costs for Process Vents and Storage
Capital Cost Components:

 Indirect Capital Costs

Initial performance tests

Process Vents
1.  scrubber ($)
2.  carbon adsorber ($)

Storage
All costs included in process vents

 Total Indirect Cost:
  Cost
$24,420
$14,470
$38,890
 Direct Capital Costs

Instrumentation
1.  flowmeter and datalogger ($)

 Total Direct Cost:
 $3,000

 $3,000
TOTAL CAPITAL INVESTMENT (TCI):
$41,890
Annual Cost Components:

 Direct Costs

Labor
Labor rate ($/hr)
hr/d
d/yr
Total Labor

Maintenance materials
Flowmeter and CEM

 Total Direct Cost:
 $17.50
   0.5
   360
 $3,150
  $500

 $3,650
  Indirect Costs

Capital recovery costs

Capital recovery factor
years (n)
interest rate (i)
Capital recovery factor value

Capital recovery cost

Reporting and Recordkeeping Requirements
(This cost is applied only once to a facility
because the cost of the requirements will not
increase substanially with the addition of
generic source types or control devices.)

 Total Indirect Cost

TOTALANNUAL COST
   10
   7%
0.142378

 $5,964

$17,500
$23,464

$27,114

-------
Shell
Enhanced Monitoring Costs for Process Vents and Storage
Capital Cost Components:

 Indirect Capital Costs

Initial performance tests

Process Vents
1.  scrubber ($)

Storage
All costs included in process vents

 Total Indirect Cost:
  Cost
 $24,420
 $24,420
 Direct Capital Costs

Instrumentation
1.  flowmeter and datalogger ($)

 Total Direct Cost:
 $3,000

 $3,000
TOTAL CAPITAL INVESTMENT (TCI):
 $27,420
Annual Cost Components:

 Direct Costs

Labor
Labor rate ($/hr)
hr/d
d/yr
Total Labor

Maintenance materials
Flowmeter and CEM

 Total Direct Cost:
 $17.50
   0.5
   360
 $3,150
  $500

  $3,650
 Indirect Costs

Capital recovery costs

Capital recovery factor
years (n)
interest rate (i)
Capital recovery factor value

Capital recovery cost

Reporting and Recordkeeping Requirements
(This cost is applied only once to a facility
because the cost of the requirements will not
increae substantially with the addition of
generic source types or control devices.)

 Total Indirect Cost

TOTAL ANNUAL COST
    10
   7%
0.1423775

  $3,904

 $17,500
 $21,404

 $25,054

-------
DOW
Enhanced Monitoring Costs for Wastewater
Capital Cost Components:

 Indirect Capital Costs
                                                     Cost
Initial performance tests

Wastewater                                         $10,000


 Total Indirect Cost:                                  $10,000


 Direct Capital Costs

 Total Direct Cost:                                     $0


TOTAL CAPITAL INVESTMENT (TCI):                   $10,000


Annual Cost Components:

 Direct Costs

Labor
Labor rate ($/hr)                                      $0.00
hr/d                                                   0
d/yr                                                   0
Total Labor                                           $0

Maintenance materials
Flowmeter and CEM                                   $0

 Total Direct Cost:                                     $0


 Indirect Costs

Capital recovery costs

Capital recovery factor
years (n)                                              10
interest rate (i)                                         7%
Capital recovery factor value                          0.14238

Capital recovery cost                                  $1,424

 Total Indirect Cost                                   $1,424

TOTAL ANNUAL COST                                $1,424

-------
Ciba-Geigy
Enhanced Monitoring Costs for Wastewater
Capital Cost Components:

 Indirect Capital Costs
                                                     Cost
Initial performance tests

Wastewater                                         $10,000


 Total Indirect Cost:                                  $10,000


 Direct Capital Costs

 Total Direct Cost:                                     $0


TOTAL CAPITAL INVESTMENT (TCI):                   $10,000


Annual Cost Components:

 Direct Costs

Labor
Labor rate ($/hr)                                      $0.00
hr/d                                                   0
d/yr                                                   0
Total Labor                                           $0

Maintenance materials
Flowmeter and CEM                                   $0

 Total Direct Cost:                                     $0


 Indirect Costs

Capital recovery costs

Capital recovery factor
years (n)                                              10
interest rate (i)                                         7%
Capital recovery factor value                          0.14238

Capital recovery cost                                  $1,424

 Total Indirect Cost                                   $1,424

TOTAL ANNUAL COST                                $1,424

-------
 TABULAR COSTS FOR




  EXISTING  SOURCES




WSR SOURCE CATEGORY

-------
Equipment Leaks Options

-------
          '3 S •
              « S ~~

            2- ?' | 1*

          l*fI Itsfi

          msiigi?
                            »a-
                            S
*i 
-------
                                                    i—OOO
                                     -?
                                  I  H
                                               it
                                               « o
                                               *. -s

                                          iJl  tit tp!Jl
                      s^s
            3SS
                                       3 5 s S a i!
                                       R 5; S S S« 3
                                       mp«
                                                S 3 S ? S 8 S



*
S


S
3
1
1


g
7
s
i
|


S
1
!
a
|
•s
8
Biul maiBtea
TOTAL .B

|
•8
S
o
1
i
s
1

*


3
If
!
i
It


5
o




nectiOD* ($):
(S):
!
J
i_
a"
|
!
|
i
TOTAL «o
I
•g
Is
  *  *
   llllllllllil
 i . tii!
li-llfriil
= £5f *E^J!£
                                               f
                                               i
ll|   .VI|


|3i|la|j-||
!|
i?
!l
i|
ll

-------
Process Vents and Storage Options

-------
Tabular Costs for MACT Floor Option  (Process Vents and Storage)

Plant #3

Currently have condenser at 20°C.  The costs for this facility
include cooling the exit gas temperature of the condenser
currently on the process vents, and  installation of a new
condenser on the storage vessel.

So, the cost of condenser on process vents  (turning the
temperature down [from 20°C to  10°C]) is:

     Capital:  $500

     Annual:   $22,000a

aEnhanced monitoring costs only.

and, the costs of a condenser at 10°C for storage are:

     Capital:  $24,000

     Annual:   $64,000

Total costs for this facility are:

     Capital:  $24,500

     Annual:   $86,000

The control device was applied  to uncontrolled streams.  However,
only the incremental control above the baseline was considered.

So,

       [(225  *  (1-.72))+(156  *  (l-.84))+(38  *  (1-.72))+127]

-[(225 * (1-.84))+(156 * (l-.91))+(38 *  (1-.84))+(127 *  (1-.44))]

            = 225 - 127 = 98 Ib/yr emission reduction

-------
Plant #8
Currently no control.
Costs of  condenser at  -10°C for process  vents are:
     Capital:   $24,400
     Annual:    $47,000
Costs of  condenser at  -10°C for storage  are:
     Capital:   $31,000
     Annual:    $65,000
The total costs for this  facility are:
     Capital:   $37,400
     Annual:    $112,000
So,
  353- [ (56*(1-.085)) + (37*(1-.089)) + (56*(1-.085)) + (56*(l-.085)) + (12*(1-0 . 91))
          + (5+(1-0.94))+ (18*(1-0.96))+ (109* (1-0.85))+ (60* (1-0.89)]
                 =353 - 46 = 307 Ib/yr emission reduction

-------
Plant #9

Currently have condenser at 20°C.  Cool the condenser down from
20°C to -50°C.

So,  the cost of condenser on process vents (turning the -
temperature down [from 20°C to -50°C])  is:

     Capital:  $360,000

     Annual:   $91,000

Emission reduction = 3,008 Ib/yr

-------
Plant #12

Currently have scrubber.  The costs below assume the scrubber is
removed from the batch reactor and a condenser is added to
control process vent emissions.  A condenser is also added to the
storage tank.

So, condenser at -10°C on process vents costs:

     Capital:  $24,400

     Annual:   $47,000

Condenser at 10°C on storage tank costs:

     Capital:  $24,000

     Annual:   $64,000

The total costs for this facility are:

     Capital:  $48,400

     Annual:  $111,000

     1,817- [(743* (1-0.94)) + (2,548*(1-0.97)) + (1,231*(1-0.98))

                         +(506*(1-0.44))]

          =  1,817  -  429  = 1,388  Ib/yr emission reduction

-------
Plant #14



Currently have scrubber.  same situation as Plant #12.



So, costs for condenser at 10°C on process vents:



    .Capital:  $15,500



     Annual:   $45,000



Costs for condenser at 10°C on storage are:



     Capital:  $24,000



     Annual    $64,000



The total costs for this facility are:



     Capital:  $39,500



     Annual:  $119,000



     975 -  [(511*(1-0.77))+(1,328*(1-0.89))+(1,073*(1-0.95)



                        + (131*(1-0.44)}]



             975  - 391 =  584 Ib/yr emission reduction

-------
           Tabular costs for Regulatory Option #1 for
                    Process Vents and Storage
Plant #2
Currently have scrubber on process vents and storage vessels.
This facility needs 2 * 8.8 = 18 Ib emission reduction to meet
the 5 Ib HAP/MM Ib product cutoff for this regulatory-
alternative.  The costs below assume the scrubber currently in
use at the facility will be disconnected and a condenser
installed.
  63 - [(42 * ((1 - .85) + (10 * (1 - .91)) +  (166 *  (1 -  .85)]

              63 - 32 = 31 Ib/yr emission reduction

So, costs for process vent condenser are:

     Capital:  $24,400

     Annual:   $47,000

Costs for storage tank condenser are:

     Capital:  $31,000

     Annual:   $65,000

Total costs for this facility are:

     Capital:  $55,400

     Annual:   $112,000

-------
Plant #3

With MACT floor option a condenser at 10°C is in place on process
vents and a separate condenser is on storage tanks.  To achieve
the 5 lt> HAP/mm Ib product cutoff the exit gas temperature of  the
condenser already in place was lowered to -10°C. ,
So,

225 -  [(225M1-.96) ) + (156* (1-. 98) + (38* (1-. 96) ) +  (127* (1-. 85) ) ]

             225 - 33 = 192 Ib/yr emission reduction

Costs for condenser on process vents are:

     Capital:  $3,800

     Annual:   $22,000a

Costs for condenser on storage are:

     Capital:  $31,000

     Annual:   $65,000

Total costs for this facility are:

     Capital:  $34,800

     Annual:   $87,000
Enhanced monitoring costs only.

-------
Plant #6


Currently have condenser at 20°C


              2- [(4*(l-.96) + (l*(l-.98)) + (0*(l-.99))]
              •*

             2 - .18 = 1.82 Ib/yr emission reduction


So,  costs for condenser for process vents are:


     Capital:  $3,800


     Annual:   $22,000a
 Enhanced monitoring costs only.

-------
Plant #8



Currently no control.



So,



           353-[(56*(l-.85)+(37*(l-.89))+(56+(l-.85))



              (12*(1- .91) ) + (5*(l-.94) ) + (18*(l-.96) ) +



                    (109M1-.85) )+(60*(l-.89) )]



            = 353 - 46 = 307 Ib/yr emission reduction



Costs of condenser for process vents are:



     Capital:  $24,400



     Annual:   $47,000



Costs of condenser for storage are:



     Capital:  $31,000



     Annual:   $65,000



Total costs for this facility are:



     Capital:  $55,400



     Annual:   $112,000

-------
Plant #9



Currently have condenser at 20°C.





                  3,008  Ib/yr  emission  reduction



The costs of the condenser on process vents is:



     Capital:  $360,000



     Annual:   $91,000

-------
Plant #12

Currently have scrubber.  The costs below assume the scrubber is
disconnected and 2 condensers are installed to control process
vent and storage emissions, respectively.


     1,817 - [ (743* (1-.94)) + (2548*(1-.97) + (1231*(1-.98))

                       +  (506 *  (1 -.85))]

           1817 - 222 = 1595 Ib/yr emission reduction.

So, costs for condenser at -10°C on process vents are:

     Capital:  $24,400

     Annual:   $47,000

Costs for condenser at -10°C on storage are:

     Capital:  $31,000

     Annual:   $65,000

Total costs for this facility are:

     Capital:  $55,500

     Annual:   $112,000

-------
Plant #14



Currently have scrubber.  Same situation as Plant #12.





 975- [(511*(l-.94(+(l,328*(l-.97)) + (1,073*(1-.99) + (131*(1-.85))]



             975  -  101  = 874  lb/yr  emission reduction



So, costs for condenser at -10°C on process vents:



     Capital:  $24,400



     Annual:   $47,000



Costs for condenser at  -10°C-on storage are:



     Capital:  $31,000



     Annual:   $65,000



Total costs for this facility are:



     Capital:  $55,500



     Annual:   $112,000

-------
Plant #15



Currently have scrubber.





       684- [(471*(l-.94)) + (165*(1-.97)) + (953* (1-.98))+223]



             684  -  275  = 409  Ib/yr emission reduction



So, costs for the condenser at  -10°C on process vents are:



     Capital:  $24,400



     Annual:    $47,000



Total costs for the facility are:



     Capital:  $24,400



     Annual:    $47,000

-------
Plant #16



Currently have scrubber.



     619 - [(611*(l-.85))+(217*(l-.97))+(914*(1-.98))+167]



             619  -  283  = 336  Ib/yr emission reduction



So,  costs for a condenser at -10°C on process vents are:



     Capital:  $24,400



     Annual:    $47,000



Total costs for the facility are:



     Capital:  $24,000



     Annual:    $47,000

-------
TECHNICAL REPORT DATA
(Pleat read Instructions on the reverse before completing)
t. REPORT NO. 2.
EPA-453/R-94-033a
4. TITLE AND SUBTITLE
Emissions from Epoxy Resins Production and Non-Nylc
Polyamides Production - Background Information for
Proposed Standards
7. AUTHOR(S)
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Emission Standards Division
Office of Air Quality Planning and Standards
U.S. EPA - Maildrop 13
Research Triangle Park, NC 27711
12. SPONSORING AGENCY NAME AND ADDRESS
Emission Standards Division
Office of Air Quality Planning and Standards
U.S. EPA - Maildrop 13
Research Triangle Park, NC 27711
15. SUPPLEMENTARY NOTES

6. REPORT DATE
in May 1994
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO.
10. PROGRAM ELEMENT NO.
li. CbNf RA
-------