6 EPA
            United. States
            Environmental Protection
            Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EPA-450/R-94-020
February 1994
            Air
               Control of Volatile Organic
               Compound Emissions from
               Batch Processes - Alternative
               Control Techniques Information
               Document

-------

-------
Control of Volatile  Organic Compound
    Emissions from  Batch Processes
                 Alternative Control Techniques
                   Information Document
                  Emission Standards Division
            U.S. ENVIRONMENTAL PROTECTION AGENCY
                  Office of Air and Radiation
             Office of Air Quality Planning and Standards
            Research Triangle Park, North Carolina 27711

                      February 1994

-------
     This report is issued by the Office of Air Quality Planning
and Standards (OAQPS) to provide information to State and local
air pollution control agencies.  Mention of trade names or
commercial products is not intended to constitute endorsement or
recommendation for use.  Copies of this report will be
available - as supplies permit - from the Library Services Office
(MD-35),  U. S. Environmental Protection Agency, Research Triangle
Park, North Carolina 27711. Springfield, Virginia  22161.
                                ii

-------
                        TABLE OF CONTENTS

                                                            Page

1.0  INTRODUCTION	  .      1-1

2.0  BATCH PROCESS DESCRIPTIONS 	      2-1

     2.1  UNIT OPERATIONS IN BATCH PROCESSING	      2-1

          2.1.1  Reactors	      2-2
          2.1.2  Solid/Liquid Separation  	     2-12
          2.1.3  Drying	     2-21
          2.1.4  Distillation	     2-26
          2.1.5  Extraction	     2-30
          2.1.6  Crystallization	     2-31
          2.1.7  Storage	     2-34
          2.1.8  Transfer Operations	     2-35
          2.1.9  Equipment Leaks	     2-36
          2.1.10 Wastewater 	  ....     2-36


     2.2  EXAMPLE INDUSTRY DESCRIPTIONS 	     2-37

          2.2.1  Synthetic Resin Manufacturing	     2-38
          2.2.2  Pharmaceuticals Industry
                   Description  .............     2-44
          2.2.3  Pesticide Manufacturing  ........     2-54
          2.2.4  Synthetic Organic Chemicals
                   Manufacturing Industry (SOCMI)  ....     2-62

3.0  EMISSION ESTIMATION METHODOLOGIES  	      3-1

     3.1  PROCESS VENT EMISSIONS	      3-1

          3.1.1  Drying	      3-2
          3.1.2  Tank and Reactor Purging	      3-4
          3.1.3  Vapor Displacement Losses  	      3-8
          3.1.4  Vessel Heating	     3-11
          3.1.5  Gas Evolution	     3-14
          3.1.6  Sparging	     3-15
          3.1.7  Batch Pressure Filtration	  .     3-16
          3.1.8  Emissions from Vacuum Generating Equipment  3-20

     3.2  EVAPORATIVE LOSSES FROM WASTEWATER  	     3-23
     3.3  STORAGE TANK EMISSIONS	     3-23
     3.4  EQUIPMENT LEAKS 	 ........     3-24
                               ill

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

                                                            Paoe

4.0  CONTROL TECHNOLOGIES	      4-1

     4.1  CONDENSERS	      4-2

          4.1.1  Design	      4-2
          4.1.2  Specific Systems and Applications  ...      4-5

     4.2  SCRUBBERS	     4-16

          4.2.1  General Gas Absorbers	     4-16
          4.2.2  Design	     4-17
          4.2.3  Specific Systems and Applicability ...     4-18

     4.3  CARBON ADSORPTION 	     4-19

          4.3.1  Design	     4-19
          4.3.2  Applicability	     4-21

     4.4  THERMAL DESTRUCTION	     4-23

          4.4.1  Flares	     4-23
          4.4.2  Thermal and Catalytic Oxidizers  ....     4-28

     4.5  SOURCE REDUCTION MEASURES 	     4-34

          4.5.1  Vapor Containment  	     4-34
          4.5.2  Limiting the Use of Inert Gas	     4-34
          4.5.3  Use of Closed Processing Equipment ...     4-36
          4.5.4  Material Substitution/Improved
                 Separation Techniques  	     4-36
          4.5.5  Improved Process Design  	     4-36

5.0  ENERGY AND ENVIRONMENTAL IMPACTS	      5-1

     5.1  ENERGY IMPACTS	      5-1
     5.2  AIR QUALITY IMPACTS	      5-3
     5.3  WASTEWATER AND SOLID WASTE IMPACTS  	      5-3
                                iv

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

                                                            Page

6.0  DESCRIPTION OP CONTROL OPTIONS	     6-1

     6.1  BACKGROUND	     6-1
     6.2  TECHNICAL BASIS FOR OPTIONS 	     6-3

          6.2.1  Approach	     6-3
          6.2.2  Options Methodology	    6-14

     6.3  PRESENTATION OF FLOWRATE REQUIREMENTS	    6-17

     6.4  IMPACTS OF APPLYING OPTIONS	    6-21

          6.4.1  Industries Covered  	    6-21
          6.4.2  Model Processes	    6-22
          6.4.3  Baseline Assumptions/Extrapolations   .  .    6-23

7.0  IMPLEMENTATION	    7-1

     7.1  DEFINITIONS AND APPLICABILITY 	     7-2

          7.1.1  Definitions	     7-1
          7.1.2  Applicability	     7-4

     7.2  FORMAT OF STANDARDS	     7-7
     7.3  TESTING	     7-7
     7.4  COMPLIANCE MONITORING REQUIREMENTS   	    7-11
     7.5  REPORTING/RECORDKEEPING REQUIREMENTS  	    7-12

APPENDIX A.   PHYSICAL DATA	     A-l

APPENDIX B.   CALCULATIONAL ISSUES   	     B-l

APPENDIX C.   SAMPLE CALCULATIONS 	     C-l

APPENDIX D.   COST CALCULATIONS	     D-l

APPENDIX E.   MODEL EMISSION STREAM CALCULATIONS  ....     E-l

APPENDIX F.   MASS EMISSIONS	     F-l

APPENDIX G.   BATCH PROCESSING EXAMPLE RULE 	     G-l

-------
                         LIST OF FIGURES
Figure 2-1.

Figure 2-2.
Figure 2-3.
Figure 2-4.
Figure 2-5.
Figure 2-6a.

Figure 2-6b.
Figure 2-7.
Figure 2-8a.
Figure 2-8b,
Figure 2-9.
Figure 2-10,
Figure 2-11,
Figure 2 -12,
Figure 2-13,

Figure 4-1.
Figure 4-2.
Figure 4-3.
Figure 6-1.
Figure 6-2.
Figure 6-3.
Figure 6-4.
Figure 6-5.
Basic Design of a kettle-type
  batch reactor 	
Plate-and-frame filter press  .  .
Agitated pressure Nutsche filter
Top-suspended centrifugal filter
Vacuum tray dryer 	
Counter-current air-heated rotary
  dryer 	
Cross-sectional view  	
Tumble (double-cone) dryer  . .  .
Batch fractionator  	
Vacuum generating equipment . .  .
Vacuum crystallizer 	
Wet-strength resins production
Aspirin manufacturing 	
Process schematic for Heptachlor
Process schematic for Chlorendic
  Anhydride 	
Filter cake drying curve  	
Dryer emission stream solvent content  .  .  .
Closed-loop drying system 	
Model Batch Process  	
Flowrate profile  	
Concentration profile 	
Emissions profile 	
Flowrate, concentration and emissions  profiles
Page

 2-3
2-14
2-16
2-19
2-23

2-24
2-24
2-25
2-28
2-28
2-33
2-39
2-49
2-61

2-63
 4-9
 4-9
4-14
 6-4
 6-6
 6-7
 6-8
 6-9
                                vi

-------
                   LIST OF FIGURES  (continued)
                                                             Page
Figure 6-6.   Dependence of control rfov-iao cogt on
                emission intermittency  .........    6-12
Figure 6-7.   Annual Mass Emission Total C/E Curve   ...    6-16
Figure 7-1.   Example Option Analysis ..........    7-14
                               VI1

-------

-------
                        1.0  INTRODUCTION

     The purpose of this document is to provide information on
alternative control techniques for volatile organic compound
(VOC) emissions from batch operations.  Although the control
techniques information applies to batch processing in all
industries, the document focuses primarily on batch processes in
the following six industries:  plastic materials and resins
(described by Standard Industrial Classification (SIC) Code
2821),  Pharmaceuticals (SIC 2833 and 2834),  gum and wood
chemicals  (SIC 2861),  cyclic crudes and intermediates (SIC 2865),
industrial organic chemicals (SIC 2869),  and agricultural
chemicals  (SIC 2879).   This document contains information on
emissions, controls, control options, and costs that States can
use in developing rules based on reasonably available control
technology.  The document presents options only, and does not
contain a recommendation on reasonably available control
technology.
                               1-1

-------

-------
                 2.0  BATCH PROCESS DESCRIPTIONS

     This chapter identifies and describes the most common unit ,
operations found in batch processing and provides descriptions of
industries that typically use batch processing.  The unit
operations section of this chapter provides descriptions of the
equipment (i.e., reactors, filters, dryers, distillation columns,
extractors,  crystallizers, and storage/transfer devices) used to
perform batch processing steps.  In the industry description
section, four industries were selected to illustrate how these
unit operations are combined to produce polymers and resins,
pharmaceutical products, pesticides, and synthetic organic
chemicals.
     Whereas the unit operations section provides general
information on equipment operation and sources of VOC emissions,
the industry description section focuses in detail on equipment
arrangements, process flows, operating conditions, and sources of
emissions.  Whenever possible, information is provided that can
be used, in conjunction with the procedures described in
subsequent chapters, to estimate VOC emissions from the five
example batch processes.  Moreover, the readers may use these
examples as a guide in evaluating emissions from other specific
batch processes that use these same or similar unit operations.
2.1  UNIT OPERATIONS IN BATCH PROCESSING
     The unit operations discussed are commonly used to produce,
separate, and prepare chemical products or intermediates on a
batch basis.  For each unit operation, a discussion is provided
of the equipment used to accomplish that operation, key equipment
design considerations, principles of equipment operation, and
factors affecting emissions.
                               2-1

-------
 2.1.1  Reactors
     The heart of many batch production cycles  is the reaction
 step.  In this step,  feedstocks  are  combined  under  the  proper
 operating conditions  and allowed to  react to  form a product.
 Catalysts may be used to initiate or accelerate the reaction or
 to optimize the generation of the desired product.   Solvents are
 often used to provide a reaction medium for solid reactants or to
 ensure that the product remains  in solution.  Following the
 reaction step, the catalysts, solvents, byproducts,  and excess
 reactants are usually separated  from the desired product.  The
 purpose of this section is to describe (1) the  design and
 operation of the equipment used  to accomplish the reaction step
 and  (2) the emissions and factors affecting emissions from this
 equipment.
     2111  Reactors;  Design  and  Operation.  Kettle-type
 reactors are used often in batch production processes.
 Figure 2-1 shows a schematic of  this type of reactor.   These
 reactors may range from 10 gallons to thousands of  gallons in
 volume.  As can be seen from Figure  2-1, the reactors are
 equipped to provide a range of capabilities that may be required
 during the batch reaction step.  This equipment includes:  a
 jacket for heating and cooling,  hookups for charging raw
 materials and for discharging the contents of the reactor, an
 agitator and recycle  line for mixing, control systems for
 temperature and pressure,  a condenser system for controlling vent
 losses, a return line for refluxing condensables, a steam ejector
 for vacuum operation, a nitrogen supply for padding and purging
 the reactor,  and a manway for taking samples and adding solid
 catalysts,  reactants, and other solid materials to the reactor.
     A typical reaction cycle consists of the following steps:
 (1) charging solvents, catalysts, and reactant(s) to the reactor;
 (2) operating the reactor;  and (3)  discharging the contents of
the reactor and preparing for the next reaction cycle.   There are
many variations to this typical reaction cycle.   This description
is not intended to address  these many variations; instead, it is
                               2-2

-------
to
                                                                Pressure
                                                               Relief Valve
                           Condensed
                            Steam and
                             Organics
                                                          Steam ^   I
                                           Steam-Jet
                                            Ejector

                                          Nitrogen
                  Solvent, Raw Material
                  and Reactant Addition
       Steam

Cooling Water
   or Coolant
             Cooling Water
                or Coolant
                                               X. Reactor
                                                 Jacket
                               Condensed
                                 Steam
                                                Drain Valve

                      Figure 2-1.   Basic  design of a kettle-type batch reactor.
                                                                              Recycle
                                                                              Line
                                                                                Reactor
                                                                                Discharge
                                                                             NCOIOC It IMNi to* 1-1 I IOM>

-------
intended to provide a more general description of the operations
involved in this stage of the process.
     2.1.1.1.1  Reactor charging.  The reaction cycle is normally
initiated by charging solvents, catalysts, and raw materials into
the reactor.  For the purpose of this description, "raw
materials" refers to compounds that are combined with other
reactants to produce the desired product or intermediate.  The
initial charging step may be accomplished in many ways.  If the
reactor is tied into a vacuum system, the materials can be
"pulled" into the reactor by reducing the pressure in the reactor
below atmospheric.  Steam jet ejectors, as shown in Figure 2-1,
or vacuum pumps may be used for this purpose.  The hookup
connection for material addition is then used to introduce
materials into the reactor.  Drums containing solvents,
catalysts, and raw materials can be hooked up to the reactor
using flexible (i.e., flex) hoses.  A dip-leg is inserted into
the drum and connected to one end of the flex hose.  The other
end of the flex hose is connected to the reactor hookup
connection.  When the valve located at the reactor (in the hookup
line shown in Figure 2-1) is opened, material in the drum flows
through the flex hose and into the reactor due to the pressure
differential.
     Solvents, catalysts, and raw materials can also be pumped
into the reactor through the hookup connection.  Portable pumps
are often used for this purpose.  During the charging process,
the valves in the vent line (the manual block valve and the
control valve) are normally opened to prevent reactor pressure
from increasing.  The condenser is usually operated to reduce
material losses through the vent line when volatile compounds are
present in the reactor.
     The manway may also be used to introduce materials into the
reactor. Solid materials are usually added to the reactor in this
manner.  The manway normally contains bolts that can be removed
to open it, but some manways have latches to allow for quick
opening and closing.  Solids are usually poured through the
manway opening into the reactor.  Once the transfer is complete,
                               2-4

-------
the manway is closed and bolted so that the reactor can be
operated under pressure.  In some cases, the middle part of the
manway is constructed of heavy glass or plexiglass so the
operator can view the inside of the reactor during operation.
     The use of flexible hoses and quick-disconnect fittings are
typical of manual-type operations where many different products
may be manufactured in the same vessel.  At other facilities
where the same batch product is manufactured routinely in the
same vessel, the reactor may be equipped with dedicated lines for
transferring materials into the reactor.  The dedicated lines
connect the reactor with storage or weigh tanks containing
solvents and raw materials.  These storage tanks are often
located at a higher level than the reactor so that material will
flow by gravity into the reactor once valves in the transfer
lines are opened.  In some cases, a reactor may be equipped with
dedicated lines for charging certain materials that are used
often (e.g., common solvents) and also equipped with a general
hookup line for less common materials.
     2.1.1.1.2  %eac_tant addition and reactor operation.-  The
most complex step in a typical batch reaction cycle is often the
reactant addition step.  This step involves the introduction of a
reactant or reactants with the materials already charged into the
reactor (e.g., solvents, catalysts, and initiators).  The
manufacture of some products involves only a single reactant
addition step.  The manufacture of other products is more
complex, requiring several steps.  In these cases, several
intermediates may be generated during the reaction cycle, and
different reactants may be reacted with each subsequent
intermediate.
     The operation of the reactor during the reactant addition
step is affected primarily by two factors:  (1) the kinetics of
the specific reaction and  (2) the capabilities of the reactor
design.  The reaction kinetics define the desired operating
conditions.  However, limits on the conditions that can be
obtained during operation are often defined by the reactor
design.  Four important operating variables that are monitored
                               2-5

-------
and controlled during the reactant addition step are:  the
addition rate of the reactant, the reactor temperature, the
reactor pressure, and the degree of mixing.  The addition rate is
closely tied to the reactant concentration, which optimizes the
generation of the desired product.  If undesirable by-products
are generated when a reactant is available in excess, it would be
necessary to monitor the addition rate closely to avoid operating
with a reactant concentration that is too high.  In other cases,
the reactant concentration is not critical and, therefore, tight
control of the reactant feed rate is not required.
     The reactant addition step may be accomplished in the same
way that materials are added to the reactor during the charging
step.  When the reactant feed rate must be tightly controlled, a
metering pump is sometimes used.  If the reaction rate is known,
the reactant feed rate can be adjusted using the pump to maintain
a proper concentration of the reactant.  In other cases,
operating parameters such as temperature and pressure determine
how rapidly the reactant is added to the reactor.  For example,
if the reaction is exothermic, the cooling capacity of the
reactor may determine how rapidly a reactant can be fed.   The
monitoring and control of operating variables other than the
reactant addition rate are discussed in more detail below.
     The reactor shown in Figure 2-1 is equipped with a typical
temperature control system.  The reactor is "jacketed" so that
either cooling water or steam can be circulated around the shell
of the reactor.  For example, steam may be required initially to
heat reactor contents to elevated temperatures due to kinetic
considerations, while cooling water is required at a later time
to quench or stop the reaction at a desired conversion level.  A
thermocouple is typically inserted into the side of the reactor
and used to monitor the reactor temperature.  The temperature
read by the thermocouple is normally transmitted to controllers
that manipulate the action of the automatic cooling water and
steam valves.
     Figure 2-1 also shows the scheme used to control the reactor
pressure.   A sensor located on top of the reactor measures
                               2-6

-------
pressure in the reactor headspace.  This pressure reading is
transmitted to controllers that operate flow valves on the vent
line (during atmospheric pressure operation) or the ejector inlet
line (during vacuum operation).  Both lines may be fitted with
condensers to minimize losses of volatile materials when they are
purged from the reactor.  The condensed materials are refluxed
back to the reactor through the return line.  As discussed
earlier, the reactant addition rate is sometimes governed by ,
operating variables such as the reactor pressure.  The cooling  .
capacity of the condenser and the sizing of the vent line both
affect the operating pressure.   In some cases, the reactant feed
rate must be slowed to prevent overpressuring of the reactor or
to reduce material losses through the vent line.
     The degree of mixing is another operating variable that must
be controlled during many reaction processes.  The reactor shown
in Figure 2-1 is equipped with an agitator for mixing.  In some
designs, a variable speed motor is installed so that the mixing
rate can be adjusted.  In addition to the agitator, mixing can be
accomplished using the recycle line and the reactor transfer
pump.  The contents of the reactor are mixed by pumping material
at the bottom of the reactor through the recycle line and back
into the top of the reactor.  The valve located in the recycle
line can be manually throttled to control the recycle flow rate.
This recycling process may be conducted with or without the
agitator running, depending on the mixing needs of the specific
reaction.
     2.1.1.1.3  Discharging reactor contents.  Once the reaction
step is complete, product purification steps are usually
required.  These steps may involve a number of unit operations
such as crystallization, distillation, filtration, and others.
Some of these steps, such as solvent recovery, may be conducted
in the reactor vessel.  Other steps require more specialized
equipment.  These unit operations will be discussed in subsequent
sections.  For the purpose of this section, it  is assumed that
the contents of the reactor are discharged  following  the reaction
step and the reactor is prepared  for the next batch.
                               2-7

-------
     The contents of  the  reactor  can be discharged by gravity or
by using a transfer pump.   If a pump is not available, nitrogen
pressure or air pressure  may be used to transfer material.  Flex
hoses can be used to  connect the  transfer line to the next
equipment piece, such as  a  batch  distillation column.  At one
end, the flex hose would  be attached to the connection shown on
the discharge side of the pump.   On the other end, the flex hose
would be attached to  a similar connection on the distillation
column.  The contents of  the reactor would be transferred to the
still by opening valves in  the transfer line at the reactor and
distillation column and starting  the transfer pump, if necessary.
     After the contents have been transferred, it is often
necessary to thoroughly clean the reactor.  This cleanup is
essential if a different  product  is to be produced in the next
batch.  If another batch  of the same product is planned for the
reactor, cleanup may  not  be required.  In cases where the reactor
needs thorough cleaning,  it is often washed with water to remove
residual product or catalyst.  If water-insoluble compounds must
be removed from the reactor, a solvent rinse is often required
prior to starting the next batch.  Both of these steps generate
waste streams.  Wastewater generated during water washing is
often discharged to onsite wastewater treatment facilities; in
some cases,  it may be sent to the public sewer.  Solvent from the
reactor rinse step is usually collected and stored in waste
solvent containers for disposal.  The waste solvent may be
disposed of by methods such as incineration or it may be purified
for reuse by distillation.  These operations may be conducted
either onsite by the  facility or  the solvent may be sent to a
commercial reclaimer.
     Following the water wash or  the solvent rinse, the reactor
is heated until dry.   Steam is used to heat the reactor jacket
and evaporate residual water or solvent remaining on the inside
reactor walls.  If the reactor is washed in water, the
evaporating water vapor may be allowed to flow out the vent line
into the atmosphere.   If a solvent rinse is used,  normally the
condenser is operated and the condensed solvent is reclaimed for
                               2-8

-------
disposal.  In this case, some solvent may escape through the vent
with noncondensables flowing out the vent line.  Once the reactor
is clean and dry, it can be closed rind prepared for the next
reactor charge.
     2.1.1.2  Emissions and Factors Affftefcino Emissions From
Reactors.  The potential for VOC emissions exists during all
steps of the reaction cycle that were discussed above.  Emissions
are discussed below in order according to the chronology of the
steps presented above.
     2.1.1.2.1  Charging.  During the charging process, volatile
compounds may be lost through the vacuum system, the vent line,
or the manway.
     Vacuum operation duringcharging.  If materials are being
charged into the reactor using a vacuum system, volatile
compounds may be pulled into tne vacuum system, which typically
will be a steam jet ejector or a water seal vacuum pump.  These
compounds either leave the jet ejector system with the steam
condensate, leave with vacuum pump seal water, or are vented from
the vacuum system with noncondensables.  The amount of materials
lost through the vacuum system depends on the volatility of the
compounds in the reactor and the duration of vacuum system
operation.  The steam condensate or pump seal liquid may be
combined with other waste streams and treated onsite, or it may
be discharged into a public sewer.  Emissions may occur during
the collection and treatment of these wastewater streams and are
referred to as secondary emissions.  A description of these
emissions is discussed  in this document.
     Atmospheric venting during charging.  If the reactor vent
line is left open to the atmosphere during charging, volatile
compounds may be vented along with the inert gases being
displaced from the reactor through the vent line.  As  the
material is pumped into the reactor, the rising liquid surface
causes the displacement of the vapor occupying the shrinking
headspace.
     Manway emissions during charging.  Emissions can also  occur
when the manway  is open for charging solids into the  reactor  if
                               2-9

-------
volatile compounds have been previously  charged  into  the  reactor.
These compounds will  saturate  the vapor  space above the liquid in
the  reactor.  If  these vapors  are less dense than air, they will
flow from the reactor once the manway is opened  due to the
buoyancy effect.  This buoyancy effect will be increased  if the
liquid in the reactor is wanner than room temperature.  The
longer the manway is  left open, the greater the  emissions will be
during this step.
     ffitrogen purging duringcharging.   When toxic or ignitable  "
material is contained in the reactor or  is being charged  into the
reactor, the reactor  headspace is often  purged with an inert gas
such as nitrogen.  The purge may be carried out  when  the  reactor
vent is open to the atmosphere or prior  to opening the manmay for
solids addition.  The purge reduces high concentrations of
volatile material in  the headspace that  could harm workers in the
immediate area or create an explosive mixture, but it increases
the emissions of VOC's.
     2.1.1.2.2  Emjysgipns during reactant addition and reaction.
     Reactant addition.  The reactant addition step essentially
is a charging step, except that the temperature  of the material
in the reactor may begin to increase as  reactant is added.
Emissions occur as a  result of vapor displacement and increase
with the rise in temperature because of  increased volatilization
of material in the reactor headspace.  Emissions from reactant
addition steps are normally emitted through the  reactor vent
line.
     Reactor heatup.   During the reaction,  the contents of the
reactor may begin to  heat up,   if the reaction is exothermic.
External heating may  also be applied to  the reaction.  Emissions
of VOC's and air toxics occur during this step because of the
expansion of headspace gas volume and because of the  (increased
volatilization of VOC's)  due to temperature rise.
     Additional load  is placed on the condenser  system if the
reactor is purged with nitrogen during reaction.   The nitrogen is
routed through the vent line so that condensables in the purge
gas can be refluxed back to the reactor.   Since  the nitrogen
                               2-10

-------
purge reduces the concentration of volatile organics flowing
through the condenser, it lowers the dew point of the stream.  In
addition, the mass flow rate increases and the residence time
decreases.  This combination of effects can result in reduced
condenser efficiency and, therefore, greater emissions of
volatiles.
     Pressure relief.  Volatile organic compounds may also be
emitted through the pressure relief valve during the reaction.
This safety device is used to relieve overpressure in the reactor
to prevent vessel rupture.  The valve is set above any pressure
that should normally be encountered during a normal reaction
process.
     2.1.1.2.3  Emissions from product purification and transfer.
     Vacuum distillation.  After the reaction is complete, excess
solvent may be separated from the product by vacuum distillation.
Emissions from this step will be limited by the exit conditions
of the reactor condenser, or condensers, if a secondary condenser
is used.  Condensers work effectively in these situations since
the uncontrolled streams contain high concentrations of volatile
components that are easily condensed at moderate temperatures and
atmospheric pressure.
     Product transfer.  Following reaction, the contents of the
reactor are discharged for further processing and packaging.  The
transfer of the reactor material contents may be accomplished by
gravity, by pumping,  by pressurizing the reactor, or by
depressurizing the receiver.  The transfer step can create
displacement emissions in the receiving vessel if the material
transferred has a significant VOC concentration or if the
receiving vessel contains VOC.  If material is transferred using
a vacuum pump, emissions may occur from the pump seal water, if
the system is "once-through."  The transfer of material using
nitrogen or air pressure may cause VOC emissions, since the inert
gas used as a carrier will in most cases be vented from the
process lines after the transfer is complete.  Depending upon the
situation, this inert gas may contain significant amounts of
entrained VOC's.
                              2-11

-------
     2.1.1.2.4  Reactor washing.  As discussed earlier,
wastewater and waste solvent streams may be generated during
reactor washing.  If the reactor is washed with water,  the
resulting wastewater stream will be directed either to a
treatment facility (where secondary emissions may occur) or to a
sewer.  If the reactor is rinsed with a solvent, emissions may
occur during the charging or disposal of the waste solvent.
Emissions may also occur during subsequent drying steps.  During
drying, heat is applied to the reactor jacket to evaporate any
residual solvent remaining on the inside reactor walls.   The
evaporating solvent may be routed through the condenser system
for recovery.
2.1.2  Solid/Liquid Separation
     Two general methods ars available for the separation of a
solid/liquid mixture--settling and filtration.  Whereas settling
relies on gravity to effect a separation, filtration uses
external forces to separate the two phases.  Specifically,
filtration uses a permeable medium that retains the solid while
allowing the liquid to pass through.
     In order to force a liquid through a filter medium, a
pressure drop must be applied.  This pressure drop may be
affected by gravity,  centrifugal force, vacuum or positive
pressure.  Centrifugal separation is discussed in
Section 2.1.2.2.  The following section discusses batch
filtration.
     2.1.2.1  Batch Filtration.  The two types of batch
filtration systems most widely used are pressure and vacuum
filters.  Batph pressure filters are used more often than vacuum
filters when filtering fine particles, because pressure
filtration provides the driving force needed to achieve
economical filtering rates.  Batch pressure filters have the
following advantages:
     1.  They allow for rapid filtration of fine slurries, which
would otherwise be filtered at an uneconomically low rate;
     2.  They are compact and offer high filtering area per unit
of plant space occupied; and
                               2-12

-------
     3.  They are flexible in operation and provide this
flexibility at a lower initial cost than other types of filters.
However, a batch vacuum filter may be better suited for
filtration applications that involve solvent vapors that produce
highly combustible atmospheres.1
     There are several types of batch pressure filters.  Two
common types are plate-and-frame and nutsche filters.
     Plate-aqd-frame press.  Figure 2-2 is a simple diagram of a
plate-and-frame filter press, which consists of alternating solid
plates and hollow frames.  Plates and frames are separated by
filter cloth.  The feed slurry enters at the top of the frames,
and the filter cake accumulates within the frames as the slurry
flows downward.  An open filtrate discharge allows the drain
ports to empty into a trough.  In closed discharge filters, drain
ports are located in the corners of each plate.  This drainage
system allows the filtrate to flow in a channel along the length
of the press.
     Slurry is pumped into the filter press until the frames are
full.  This determination is made based on time, a decrease in
feedrate or an increase in backpressure.  Once the frames within
the filter area are full, the discharge ports are opened.  Filter
cake forms on the cloth as the slurry liquid flows through the
cloth.
     The thickness of the filtercake depends on the purpose of
the filtration.  In purification of a dilute slurry to yield a
clean filtrate, the filtercake is thin.  In solids recovery, when
slurries may be 40 percent solids, the frames are usually full
after the cycle.  Frames of varying thickness are available for
different applications.
     After the slurry has passed through the filter, a wash
liquid may be applied.  There are two different methods used for
cake washing.  In simple washing, wash liquid follows the same
path as the slurry.  In through washing, wash liquid enters
alternate plates and is forced through the entire cake by
alternately closed discharge ports.  Cake characteristics
                               2-13

-------
                                                    Fntfiis
                                                   Discharge
Figure  2-2.   Plate-and-frame filter press,
                     2-14

-------
determine the appropriate wash method.  Compressed gas, such as
air, may also be used to clean and dry the caJce.
     Solids are discharged by openly auu separating the plates.
Vibration and air blowing may be used to detach the filtercake
from the cloth. These operations may be done manually or
mechanically.
     Nutsche filter.  Another type of pressure filter, the
nutsche filter, can either compress a slurry or apply vacuum to
it in order to create a filter cake.  A typical agitated nutsche
filter is presented in Figure 2-3.  The equipment not only acts
as a filter but can also function as a product dryer after the
slurry has been compressed and filtered into cake form.
     The filter works by pressurizing the slurry with nitrogen to
force the liquid through the filtering medium.  The pressure
needed to help maintain this process until enough liquid has been
extracted is almost entirely a function of the specific particle
characteristics of the product.  Conversely, a vacuum may also be
applied to the nutsche to draw the liquid down through the cake.
Vacuum applications are usually limited to slurries with highly
combustible atmospheres.2  Because particles are spherical to
irregular in shape and generally amorphous,  the type of cake
formation expected will determine the optimal pressure or vacuum
needed to complete filtration.  Experience has shown that
filtration pressures generally range from 20 to 35 psig.  Filters
range in size from 1,000 to 2,000 gallons.  Vacuum filtration
occurs at pressures ranging from 3 to 20 psi.  These filters also
range in size from 1,000 to 2,000 gallons.
     Upon completion of filtration, the filter may or may not go
through a reslurry process, where it is washed and filtered
again.  This option is usually carried out when a highly
specialized product requiring purity is desired or when solvents
were not removed as part of the original slurry filtration
process.
     The nutsche filter is also capable of drying the filter cake
and may be converted into a filter/dryer by only limited
modifications.  The actual drying process carried out in the
                               2-15

-------
Figure 2-3.  Agitated pressure Nutsche filter.
                     2-16

-------
modified filter is usually convective.  Heat is introduced to the
filter/dryer through a hot gaseous medium (usually N2) which is
blown up through the cake until the desired level of dryness is
achieved.  The cake can be agitated or remain static, depending
on the drying characteristics associated with the product.
     2.1.2.1.1  Factors Affecting Emissions from Batcfr
Filtration.
     Plate and frame filters.  Emissions from the plate-and-frame
filter press can occur during filtration, washing, and discharge
steps.  The potential exists for VOC emissions during filtration
and solvent washing from the trough used with open discharge, and
the corner holes associated with closed discharge.  Likewise,
emissions can occur when the filter press is opened to remove
solids.  Vibration and air blowing to detach the solids can also
increase the rate of emissions.  The range of emissions will
depend primarily on the vapor pressures and mole fractions of
each VOC, the operating temperature of the filter, and air
circulation rate.  Because plate-and-frame filter presses offer
no containment,  it is also unlikely that material containing a
high percentage of volatile or toxic solvent will be filtered
using this type of device.
     Nutsche filters.  Pressure filters such as the nutsche
filter shown in Figure 2-3 normally do not emit VOC's during
actual filtration since they are fully enclosed.  However, during
slurry charging or vessel depressurizing, emissions of VOC can
occur.
     Emissions also can occur from all batch process filters if a
compressed gas is used to purge the filter or dry the cake.  The
gas will entrain evaporated solvent and carry it to a vent.
Emission rates will depend on the factors cited above and the
compressed gas purge rate.  Note also that if filtrate from
either operation is discharged to wastewater treatment, there is
also potential for emissions resulting from cross-media transfer
effects.
     2.1.2.2  Centrifugal Reparation.  As mentioned in
Section 2.1.2, filtration is used to separate a solid from a
                               2-17

-------
liquid.  Centrifugal filters (basket centrifuges) make use of the
outward (centrifugal) force that is exerted on an object during
rotation.   This centrifugal force pushes the liquid through the
filter medium and presses the solids against the walls to form a
cake.  In a solid-bowl centrifuge the liquid is separated from
the solid by centrifugal force and is continuously decanted off.
The recovered solid accumulates on the sides of the bowl.  Solid
bowl centrifuges are used to recover small amounts of solids that
are dispersed in large amounts of liquid.  Catalysts, for
example, often are recovered from liquid product in this manner.
     2.1.2.3  Centrifugal Filters;  Design and Operation.
Centrifugal filters are cylinders which contain a rotating basket
at the base of a vertical shaft.  Figure 2-4 depicts a typical
configuration for a basket centrifuge.  The basket may be 0.8 to
1.2 m in diameter and 0.5 to 0.8 m deep.  Its sides are
perforated and covered with a filter medium such as fabric or
woven metal.  An inert gas such as nitrogen is often introduced
into the chamber prior to the addition of slurry to avoid the
buildup of an explosive atmosphere.  Centrifuges must be
carefully operated to avoid air infiltration by vortex
entrainment.  Therefore, they usually are operated under nitrogen
blanket and kept sealed during operation.3
     Feed slurry enters the chamber through an inlet pipe as the
basket rotates at speeds of 600 to 1,800 revolutions per minute
(rpm).  Centrifugal force pushes the mixture towards the wall of
the basket.  The liquid passes through the filter medium and is
discharged through a pipe.  The solid particles form a filtarcake
on the sides of the basket.
     After all of the slurry has been fed to the chamber, a wash
liquid may be introduced to force the remaining slurry liquid
through the cake and filter medium.  The basket continues to spin
in order to remove any residual liquid.
                               2-18

-------
                                      Motor
  Casing
 Filtrate
Oitchargt
                                                         Slurry
                                                          (mat
                                   flmovaWt
                                   Valvt Plan
                               Solids
                             Oiscftarg*
  Figure 2-4.  Top-auspended  centrifugal filter.
                           2-19

-------
     At this point the motor speed is reduced, slowing the
rotation to between 30 and 50 rpm.  An adjustable knife is
engaged to scrape the side of the basket and dislodge the
filtercake.  The cake material falls to the bottom of the chamber
where it is discharged through an opening in the basket.
     Manual dumping of filter cake from basket centrifuges also
can occur, especially when top-unload centrifuges are used.
Operators must "scoop out" product into a transfer vessel.  The
vessel is usually purged with high flow rates of inert gas just
prior to this step.
     Another type of basket centrifuge is the Heinkel centrifuge.
The main feature that distinguishes the Heinkel centrifuge from
other basket centrifuges is the inverting filter cloth the
Heinkel employs.  The inverting filter cloth allows top unloading'
in a simplified manner.  Rather than scooping the contents out
manually, the operator(s) can displace the entire filter cloth
and empty its contents.4
     2.1.2.4  Factors Affecting Emissions from Centrifugal
Separation.  Emissions from centrifuges may occur during initial
vessel purging prior to the addition of slurry, and during
discharge.  A potential source of emissions from centrifuges is
created by the inert gas blanket which is used to prevent  the
possibility of an explosive atmosphere.  The inert blanket is
especially necessary in bottom-discharge centrifuges because they
contain metal knife scrapers that move the filtercake away from
the walls.  The mechanical friction associated with metal-to-
metal contact and static electricity discharge are likely
ignition sources.
     The potential for an explosion depends on the type of
centrifuge, the characteristics of the solvent vapor, and  how the
centrifuge is operated.5  A centrifuge is difficult to blanket
with an inert gas during discharge because it cannot stay  sealed.
During discharge, therefore, an inert gas purge  is more effective
in evading explosive conditions because of the higher flow rate
associated with a purge.  Note that during the actual
                               2-20

-------
centrifugation process, an inert gas blanket contributes
significantly lower VOC emissions than a purge.
     As mentioned, the much higher flnw i-mtes associated with
inert gas purges will obviously induce greater emissions.  Purges
are used during bottom-discharge and prior to opening a top-
unload centrifuge for sampling or unloading.
     The solids removed from the centrifuge may still be "wet"
with solvent and therefore be a source of emissions during
unloading and transport to the next process step.  Bottom-
discharge centrifuges can minimize this problem if the solids are
transferred to a receiving care chrough a closed chute and the
receiving cart is covered during transport.  As with other
filters, the emission rate from centrifugal filters will be
influenced by operating temperature, VOC vapor pressures and mole
fractions, inert purge gas Cio* iue," and the use of mitigating
factors such as closed chutes and carts.
2.1.3  Drying                                      '  <
     The term "drying" generally refers to the removal of liquid
from primarily solid material.  However, due to the large amount
of solids (and sometimes a large portion of liquid) dryers can be
large VOC emission sources.  Dryers are used to remove liquids,
usually residual solvent, from centrifuged or filtered product.
This removal is accomplished by evaporating solvent into a gas
stream.  Solvent evaporation is accelerated by application of
heat and/or vacuum to the wet solids.  Circulation of warm air
also speeds the drying process.
     It is important to note the differences between dryers and
evaporators.  Whereas evaporators remove liquids as vapors at
their boiling points, dryers remove the vapor into a gas stream
at temperatures below its boiling point.  Also evaporators are
usually used to remove large amounts of liquid.
     There are several different types of dryers being used by
industry today.  For example, tray, tunnel, rotary, drum and
spray dryers are available.  Selection of dryer type depends
primarily on characteristics of the solid.  Three dryer types
that are commonly used in batch processes are tray, rotary, and
                               2-21

-------
 double-cone  dryers.   The  previous  discussion on the  converted
 nutsche filter/dryer is also considered relevant to  this
 discussion.
      2.1.3.1 Tray Dryera  Figure  2-5  is a  simplified diagram of
 a tray  dryer.  Tray  dryers  are  among the simplest type of dryer,
 although they are labor intensive  because of necessary manual
 loading and  unloading.  The product intended for drying is placed
 on trays that are stacked on shelves.   After all the trays have  
 been  filled, the dryer door is  closed  and the shelves are heated.
 A vacuum is  also pulled within  the dryer to allow for drying at
 low temperatures.  Typically, tray dryers contain 15 to
 20 trays.6
      2.1.3.2  Rotary Dryers;  Design ajid Operation.  Another
 important type of dryer is  the  rotary  dryer.  As shown in
 Figure 2-6,  this dryer consists of a revolving  cylinder that is
 slightly inclined to the  horizontal.   The diameter of the
 cylinder may range from 0.3  to  3 m, and the length may vary from
 1  to  30  m,
      Feed enters at  the elevated end and is  carried  through the
 dryer by the rotation and slope of the  cylinder.   In direct-heat
 rotary dryers, the solids are dried by  direct contact with a
heated gas stream.   This  stream may consist  of air or flue gas
 flowing  at approximately  2.8 m3 per minute.  The  flights shown in
Figure 2-6 lift the  solids and shower them  through the gas
stream.   The solids  and gas may flow cocurrently or  counter-
currently, with countercurrent flow having  a greater heat-
transfer efficiency.
     Due to the nature of the equipment, the outlet  stream for
rotary dryers must be free-flowing and granular.   Sticky feed
materials may be dried if some of  the granular product is
recycled and mixed with the feed.
     2-1.3.3  Tumble  (Double-Cane)  Dryers;   Deaign and Operation.
A batch double-cone dryer is shown in Figure 2-7.  Material to be
dried in this type of dryer must be manually loaded  into the
dryer and manually unloaded after  the drying cycle is complete.
Double-cone dryers may be operated under a vacuum  in which a
                              2-22

-------
                                          14
                                          
-------
                                                              Conosmtw
                                        Dry Solids Discharge
                            (a)
                                        Rights
                             (b)
Figure 2-6.
(a)  Countercurrent air-heated rotary dryer,
(b) Cross-sectional  view.
                            2-24

-------
                                           0)
                                           er
                                           -o
                                           0)
                                           c
                                           o
                                           u

                                           0)
                                           3
                                           O
                                           O
                                           (1)
                                           
                                           O1
2-25

-------
 small  flow rate of air  is allowed  to  leak  in  or  occasionally may
 be  used  to dry material  convectively  with  heated gas.  Tumble
 dryers range  in size  from 20 to  100 gallons.  Flow rates of
 drying gas and drying temperatures vary with  product.
     2.1.3.4  Factgrs Affecting  Emissions  frorp Dryers.  Volatiles
 may be emitted at the feed inlet and  product  discharge areas of
 the dryers, as well as  from the  dryer exhausts.  Tunnel, nutsche
 filter/dryers, and rotary dryers typically use a moving stream of
 heated air to dry the feed material.  This mode  of drying is
 termer! "convective".  Emission streams from convective dryers
 will have large volumes  of noncondensable  gases  throughout the
 drying cycle.
     Tray and double-cone dryers typically are operated under
vacuum,  in which the  heat transferred to the  material being dried
will be  through conduction from  heated surfaces.  Under vacuum, a
smaller volume of air passes through  the equipment due to inward
leakage.  This vacuum exhaust contains VOC's.  The volume of
noncondensables in vacuum dryer  exhaust is small, compared to
convective exhaust, and  increases throughout  the drying cycle.
     Dryers are potentially large emission sources.  Emissions
vary according to dryer  type, dryer size,  number of drying cycles
per year, and amount  and type of solvent evaporated.  Emission
rates vary during a batch drying cycle:  they are greatest at the
beginning of the cycle and least at the end.  The rate of VOC
emissions from a given batch drying operation will be a function
of the duration of the drying cycle and the amount of solvent in
the material.
2.1.4  Distillation
     Distillation is used to separate a mixture of liquids.  The
basis for this separation is the relative volatility (i.e.,
vapor-pressure and boiling point) of  the components.  Within
refining and chemical manufacturing,  distillation is the most
commonly used method for separation and purification of liquids.
     Separation is achieved by the redistribution of the
components between the liquid and vapor phases.   The more
volatile component(s)  concentrates in the vapor phase while the
                              2-26

-------
less volatile component(s)  concentrates in the liquid phase.  The
two phases are generated by vaporization and condensation of the
feed mixture.
     There are several different types of distillation
operations.  In simple operations, the feed is vaporized and
condensed one time.  This usually does not yield a clean
separation.  Fractional distillation involves repeated
vaporization and condensation and results in a sharper
separation.
     2.1.4.1  Batch FractioQatprs:  Design and Qperatj.pQ.  The
batch fractidnator in Figure 2-5 consists of a reboiler and a
sieve-plate column.  The feed mixtv.T? is charged into the
reboiler and heated until it begins to boil.  The initial vapor
that forms is richer in the more volatile component (A) than the
liquid is.  However, the vapui suill uuuuains a significant
amount of both components.   In order to increase the
concentration of A in the vapor-, the vapor stream enters the
column where it is brought into contact with boiling liquid.
     The vapor that exits the top of the column goes to a
condenser and then to the accumulator (reflux drum).  Some of the
condensate in the reflux drum is returned to the column as reflux
at the top of the column.  As it flows down the column, the
liquid contacts the vapors that are moving upward.  Contact
between the two phases occurs in a stagewise manner in a column
which holds horizontal-stacked sieve trays.  Vapor flows up
through the perforations. Liquid flows down through pipes called
downcomers.
     The downcomers are located on alternating sides of each
tray.  Thus the liquid must flow across the tray.  The top of the
downcomer acts a weir, maintaining a minimum depth of liquid on
the tray.  The vapor bubbles up through the layer of liquid.
This contact causes some of the more volatile component  (A) to
diffuse from the liquid, thus enriching the vapor.
     The vapor leaving the top of the column is condensed; part
of the condensate is returned to the column as reflux and the
remainder  is drawn off as product liquid, or distillate.  This
                               2-27

-------
                                                                          Steam
to
i
to
00




S~









Feed -*/
Hmf In  1






i

Q.
cd

i
i

^










1 CW
 ^V-j tjecior
Vapor Phase /OC"" M



-^








i

^T
\
 . f J Accumulator






T3
'3
_O"
1
- T
Reflux

i

Overhead Product
Distillation (Distillate)
Column
i
1 cw
X"j^
f^j Condenser
SY

Vent
 ** *

J Accumulator



Waste Stream

Reboiler A
_t-i_rt_ -~-  _^*n 1
m
n
18
J } 	 ~ Bottom Product

                               (a)
(b)
NCOCC i? SW* fn 2 
                       Figure  2-8.   (a) Batch  fractionator with  (b)  vacuum
                                        generating equipment.

-------
 circulation  continues  until  the  desired  separation is  achieved,
 which  is usually  determined  by the purity  of  the  distillate.
 Specifications  for  this  stream usually state  a maximum
 concentration of  the less volatile component.
     The liquid leaving  the  reboiler is  called the bottom product
 or bottoms.  It is  rich  in the less-volatile  component but is not
 as pure as the  distillate.   Whereas the  vapor is  enriched as it
 moves  upward through the column, the equipment in Figure 2-8 does
 not provide  for enrichment of the liquid stream.   Thus, the
 condensed vapor product  (distillate) will be  more pure than the
 bottom product.   Rectification of the liquid  stream, to yield a
 nearly pure  bottom  product,  requires a more complex column.  Such
 columns are  usually run on a continuous, rather than batch,
 basis.
     The equipment  shown in  Figure 2-8 is a fairly simple
 arrangement. Variations are made based on the nature of the
mixture and  its components.  Many batch  processes  involve a
 distillation in a re;   tor kettle or series of kettles  (often
 called "still pots"),   as opposed to a column.  The  distillation
principles are  the  same regardless of whether the  separation is
 conducted in kettles or columns.   For example, operating
pressures can be below atmospheric (vacuum),  atmospheric, or
above atmospheric (pressure).  Figure 2-8 shows a possible
arrangement for vacuum-generating equipment.   Inert gas,
especially steam,  is often introduced to improve separation.  If
a mixture is particularly difficult to separate (i.e.,
azeotropic),  other compounds may be added to aid in distillation.
     2.1.4.2  Factors Affecting Emissions from Batch
Fraetionatora.   The gases and vapors entering the condenser can
contain VOC,  water vapor, and noncondensables such as oxygen
 (02),  nitrogen  (N2)  and carbon dioxide (C02). These vapors and
gases originate from:
     1.  Vaporization of liquid feeds;
     2.  Dissolved gases in liquid feeds;
     3.  Inert  carrier gases added to assist in distillation
 (only for inert carrier distillation);  and
                              2-29

-------
     4.  Air leaking into the column  (in vacuum distillation).
     The condenser cools most of the vapors enough that they can
be collected as a liquid phase.  The noncondensables  (02, N2,
C02, and organics with low boiling points) are present as a gas
stream and are vented from the condenser.  Portions of this gas
stream are often recovered in devices such as scrubbers,
adsorbers, and secondary condensers.
     Vacuum-generating devices (pumps and ejectors) might also
affect the amount of noncondensables.  Some organics can be
absorbed by condensed steam in condensers located after vacuum
jets.  In the case of oil-sealed vacuum pumps, the oil losses
increase the VOC content of the noncondensables exiting the
vacuum pump.  The noncondensables from the last piece of process
equipment (condensers,  pumps, ejectors, scrubbers, adsorbers,
etc.) constitute the emissions from the distillation unit unless
they are controlled by combustion devices such as incinerators,
flares and boilers.
     The most frequently encountered emission points from
distillation operations are:  condensers (which are described in
Chapter 4),  accumulators (losses are typical of vapor
displacement, discussed in Chapter 3), steam jet ejectors
(discussed in Chapters 3 and 4),  vacuum pumps (discussed in
Chapter 3 and 4), and pressure relief valves  (discussed in
Chapter 3).   The total volume of gases emitted from a
distillation operation depends upon:
     1.  The physical properties of the organic components
(especially vapor pressure at the reflux drum temperature);
     2.  The efficiency and operating conditions of the condenser
and other recovery equipment;
     3.  The volume of inert carrier gas used; and
     4.  Air leaks into the vacuum column (leaks are increased by
both reduced pressure and increased column size).
2.1.5  Extraction
     Liquid extraction is another method of separating a mixture
of two liquids.  Whereas distillation takes advantage of a
difference in boiling point (vapor pressure), the principle of
                              2-30

-------
liquid extraction is based on a difference in solubility.  In the
extraction operation, a mixture of two liquids {A & B)  is brought
into contact with a third liquid called the solvent (S).  The
solvent preferentially combines with one of the components of the
original mixture.  The two resulting streams are:
     1.  Extract - mostly solvent and the liquid with which it
preferentially combined (S + A5; and
     2.  Raffinate - mostly residual liquid from the original-
mixture (B) .
It is important to note that both exit streams will contain all  
three components (A, B and S).  However, the raffinate will be
primarily liquid B and the extract will be primarily a mixture of
A and S.
     There are three general types of equipment used for liquid-
liquid extraction, although most batch extractions occur in
mixer-settlers.  In mixer-settlers, a mixer is used to contact
the feed solution and solvent.  A settling tank allows the two
phases to separate by gravity.  Stirred-tank reactors often serve
as both mixer and settler.
     2.1.5.1  Factors Affecting Emissions from Extractors.
Emissions from mixer-settler extractors are similar to those from
reactors  (discussed  in Section  3.1.15 in that they stem mainly
from vapor displacement during purging, filling and cleaning of
the vessel.  Some VOC may also be emitted while the liquids are
being agitated.
     As discussed for reactors, the rate of VOC emissions will
.depend primarily on  VOC vapor  pressures at operating
temperatures, liquid pumping  rate during column  filling,  rate  of
sweep gas  (if used)  during  purging, and equipment  cleaning
procedures.
2.1.6  Crystallization
     Crystallization is a means of  separating an intermediate  or
final product from  a liquid solution.   Solid  particles  (i.e.,
crystals)  are formed from  the homogenous  liquid  phase.   This
formation is accomplished by creating a supersaturated  solution,
in which  the desired compound will  form crystals.   If performed
                               2-31

-------
properly and in the absence of competing crystals,
crystallization can produce a highly pure product.
     Four methods may be used to produce supersaturation.  If
solubility of the solute increases strongly with temperature, a
saturated solution becomes supersaturated by simple cooling.  If
solubility is relatively independent of temperature or decreases
with increased temperature, supersaturation may be generated by
evaporating a portion of the solvent. Often a combination of
cooling and evaporation is used.
     Lf neither cooling nor evaporation is desirable,
supersaturation may be induced by adding a third component.  The
third component forms a mixture with the original solvent in
which the solute is considerably less soluble.
     Batch crystallization usually relies on simple cooling or
evaporation as the method for producing supersaturation.  Batch
crystallization is useful for low production rates and when the
cooling range is wide, since it avoids the material shock that
occurs in continuous crystallization equipment (i.e., metal
stress)  from mixing a hot solution with a cool mother liquor.7
Figure 2-9 is a diagram of a batch vacuum crystallizer.
     2.1.6.1  Vacuum Crystallizers;  Design and Operation.  In
most vacuum crystallizers, supersaturation is generated by
adiabatic evaporative cooling.  The equipment mainly consists of
a closed vessel with a conical bottom.  A condenser and steam-jet
vacuum pump maintain a vacuum within the crystallizer.
     The feed solution is saturated and heated to a temperature
greater than the boiling point at the crystallizer pressure.
Upon entering the chamber, the solution cools spontaneously, and
some of the solvent evaporates.  The cooling and evaporation
induce supersaturation, which initiates crystal formation.  The
mixture of mother liquor and crystals is referred to as "magma."
     As crystals form and grow, they are drawn off by a discharge
pipe. This pipe is located in the conical section of the vessel,
above the downpipe that leads to the pump.  The discharge stream
will contain some mother liquor.  Further processing  (e.g.,
centrifugation)  can separate the two components.   Batch
                               2-32

-------
       Cooling Wrtr
                                    To Vacuum
                             Product
                             Otscftarg*
Figure  2-9.  Vacuum crystallizer.
                2-33

-------
crystallization is usually performed with small amounts of
material.  Cycle times range from 2 to 8 hours.
     2.1.6.2  Factors Affecting Emissions from Batch Vacuu^p
crystallizers. If crystallization is done mainly through cooling
of a solution, there will be little VOC emission.  In fact, the
equipment may be completely enclosed.
     However, when crystallization is done by solvent evaporation
in a vacuum environment, there is a greater potential for
emissions.  The vapor over the magma is rich in solvent.
Emissions will be significant if this evaporated solvent is
vented directly to the atmosphere.  The condenser and vacuum jet
shown in Figure 2-9 reduce the amount of volatiles that actually
leave the system.  As with other vacuum operations, the rate of
VOC emissions will depend primarily on the VOC vapor pressure at
the crystallizer or condenser temperature, the absolute pressure
of the system, and the air leak rate.
2.1.7  Storage
     2.1.7.1  Storage Equipment:  Design and Operation.  There
are three major types of vessels used to store volatile organic
liquids  (VOL's):
     1.  Fixed roof tanks;
     2.  External floating roof tanks; and
     3.  Internal floating roof tanks.
These tanks are cylindrical with the axis oriented perpendicular
to the foundation and are almost exclusively above ground.  This
section addresses only the fixed roof type of storage tank.
     Of currently used tank designs, the fixed-roof tank is the
least expensive to construct and is generally considered as the
minimum acceptable equipment for the storage of VOL's.  A  typical
fixed roof tank consists of a cylindrical steel shell with a
cone- or dome-shaped roof that is permanently affixed to the tank
shell.  A breather valve  (pressure-vacuum valve), which is
commonly installed on many fixed-roof tanks, allows the tank to
operate at a  slight internal pressure or vacuum.  However, this
valve prevents the release of vapors only during very small
changes  in temperature, barometric pressure, or liquid  level.
                               2-34

-------
Larger changes in these parameters can result in significant
emissions from fixed roof tanks.
     2.1.7.2  fygtora Affecting Emissions from Storage Tanks.
The major types of emissions from fixed-roof tanks are breathing
and working losses.  A breathing loss is the expulsion of vapor
from a tank vapor space that has expanded or contracted because
of daily changes in temperature and barometric pressure.  The
emissions occur in the absence of any liquid level change in the
tank.  The rate of VOC emissions from breathing losses is a
function primarily of VOC vapor pressure (at the bulk liquid
conditions), tank diameter, average vapor space height, and
ambient temperature and pressure changes from day to night.
     Working losses are associated with an increase in the liquid
level in the tank.  The vapors  in the space above the liquid are
expelled from the tank when, as a result of filling, the pressure
inside the tank exceeds the relief pressure.  Emptying losses
occur when the air that is drawn into the tank during liquid
removal saturates with hydrocarbon vapor and expands, thus
exceeding the fixed capacity of the vapor space and overflowing
through the pressure-vacuum valve.  Combined filling and emptying
losses are called  "working losses."  The rate of VOC emissions
from these working losses  is a  function of VOC vapor pressure  (at
bulk liquid conditions), vapor  space height, and turnover  factor
 (i.e., the rate at which the tank  is emptied and refilled).
Information on emissions from  storage  tank working and breathing
losses is detailed in EPA's AP-42  Compilation of Air  Pollution
-Emission Factors.
2.1.8 Transfer  Operations
      Chemical transfer operations  also contribute to  plant VOC
 emissions.   Common sources of  transfer emissions  are:
      1.   Manual  transfer of chemicals from 55  gallon drums to
 receiving vessels;  and
      2.   Transfer of final product from processes to receiving
 vessels.
      Some chemicals are stored in 55-gallon drums.  Transfer  of
 chemicals from drums to process vessels is sometimes done through
                                2-35

-------
permanent piping; more commonly, however, it is done by opening
the drum and manually pouring the contents.  The manual pouring
is a source of emissions, although a relatively small one on a
"per drum" basis.
     Emissions from transfer of final product from processes to
receiving vessels occur frequently in batch processing.  These
emissions are analogous to vessel charging.  The AP-42 emission
factor handbook referenced earlier contains emission estimation
methodologies for various loading mechanisms including splash
loading and submerged filling.
2.1.9  Equipment Leaks
     Pump seals, flanges, valve seals, agitator seals, and hose
connections or couplings create VOC emissions when they leak.  A
protocol has been developed by EPA that can be used to develop
emission factors for equipment leaks.  However, if no other data
is available, factors have been developed to estimate the amount
of VOC that is leaking from a population of valves, seals,
flanges, etc.  These factors are known as synthetic organic
chemical manufacturing industry (SOCMI) equipment leak factors.
Note that the amount of time that the process components are in
VOC service also plays a role in the amount of VOC that could be
expected to leak.  This is especially of concern in batch
processes, where many components are not in service the majority
of time- and therefore should not be expected to leak as much
material as components that are part of continuous processes.
Note also that because of this reason it may be harder to detect
leaking components in batch processes.
2.1.10  Waatewater
     2.1.10.1  Generation.  Wastewater may be generated from a
number of activities that occur in batch processing, including
equipment cleaning, vacuum ejector or pump once-through
circulation using water, scrubber water discharging, steam
stripping, or the discharging of water that was part of the
process feedstock or that was generated in the process  (i.e., a
condensation reaction).  Wastewater often contains dissolved
VOC's or air toxics and may also carry large amounts of insoluble
                               2-36

-------
 VOC's or air  toxics  in emulsion-type multiple phase systems.
 Facilities often pretreat wastewater prior to discharge to
 publicly owned treatment plants, and this practice often is
 streamlined by isolating wastewater based on the degree of
 contamination and treating the various fractions accordingly.
     2.1.10.2  Factor? Affecting Emissions from Wastewater.  The
 amount of VOC contained in the wastewater and the method of
 treatment of  wastewater will affect the amount of VOC emissions.
 Treatment options include activated sludge without aeration and 
 simple decantation and settling with the aid of flocculating
 agents.  The  EPA has published guidelines on air emissions from
 wastewater treatment systems.  These guidelines examine various
 methods of wastewater containment, such as storage in open tanks
 and provide methodologies for estimation of emissions.  For open
 tanks, the emission estimation discussion presented in Appendix B
 of this document is also relevant.
 2.2  EXAMPLE  INDUSTRY DESCRIPTIONS
     The inherent diversity of batch manufacturing prevents a
 general description of this chemical industry segment.  Instead,
 the following five industries have been selected as examples of
 the unit operation configurations, equipment operating
 conditions, and emission sources typically encountered in batch
 chemical processing:  resin manufacturing, Pharmaceuticals
manufacturing, pesticides,  and SOCMI.  These processes were
 chosen because they contain high production volume batch
processes,  contain significant potential sources of VOC
emissions,  and illustrate the diversity of equipment
 configurations and process flows which characterize this type of
 industry.
     For each of these batch processes, a detailed description is
provided of the associated chemistry, equipment, and stream
 flows,  including a process flow diagram.  Each of the
 intermediate processing steps is discussed with an emphasis on
process operating conditions wherever this information is
available.   A separate subsection for each process is devoted to
                               2-37

-------
VOC emission sources and the factors that influence the extent of
those emissions.
2.2,1  Synthetic Resin Manufacturing
     The manufacture of synthetic resins is often accomplished
using batch processes.  In light of an ever-increasing demand for
highly-specialized materials, batch processing offers flexibility
in product specification and production rate, as well as a high
degree of control over process variables.  This industry
description segment is intended to familiarize the reader with
typical batch processing routes that are currently being used to
manufacture resins.  A specific process for the manufacture of
epichlorohydrin-based nonnylon polyamide (wet strength) resin is
described in detail below.  In addition, information on sources
of VOC and air toxics emissions from the process is presented
following the process flow discussion.  This process typifies the
batch processes currently found in industry.
     2.2.1.1  Process Flow.  Figure 2-10 is a simplified flow
diagram of a wet-strength resin production process.  The batch
process originates with the storage vessels.  Material used as
feedstock for the process is stored in some type of storage
vessel.  Vessel types range from 55-gallon drums to fixed or
floating roof tanks to pressurized horizontal tanks.  Underground
tanks are typically not used for feedstock storage in this
industry.  At the beginning of the batch cycle, a quantity of
material to be used as reactant feedstock is pumped from its
storage container to a weigh tank, or charge tank.  The volumes
of weigh tanks range from a few hundred gallons for small batch
processes to several thousand gallons for much larger batch
processes.  Weigh tanks, as the name implies, are used to measure
the amount of material charged to the reactor.  Several weigh
tanks may be used for each reactor, depending on the required
feedstock recipe.
     Once the desired quantity of material has been obtained in
the weigh tankCs), it is charged to the reactor, usually by
gravity, as the weigh tanks in most cases are physically
positioned above the reactors.  Wet strength resins are formed by
                               2-38

-------
K>

Ul
                      Vapor Return (optional)
                 Vent
                            Vapor
                  4+4
Fill Line
                  Storage
                  Vessel
     "]    Vent     Return
1          A     (optional)
          Weigh Tank
                             N2 Purge (optional)
                                                      Hatchway
                                                  H    if=1i
                                                  Reactor
             NCOOC  8HHC *cm2-tQ
                            Figure 2-10.   Wet-strength resins production.

-------
reacting epichlorohydrin  (EPI) with water and an inert amine-
based polymer mixture.  The amine polyer is in some cases
manufactured onsite in much the same manner as is the cross-
linked reaction production described below.  Typical reactor
sizes are on the order of 2,000 gallons.  Reactors may also be
charged by pumping the material from the weigh tanks by
pressurizing the weigh tanks to push the material through to the
reactor or by depressurizing the reactor vessel.  Charging rates-
to the reactor typically are on the order of 50 gallons per
minute, although some facilities report charging rates of greater
than 200 gallons per minute.
     Reactors are generally equipped with a temperature control
jacket, an agitator, manholes for solids addition or sampling,
and a pressure-relief valve.  The manufacture of wet-strength
resins involves an exothermic reaction which occurs at pressures
at or close to atmospheric.  Batch reaction times range from 2 to
24 hours, depending on the product.  It is necessary in the
manufacture of some products to halt polymer chain cross-linking
reactions by adding acid when the resin viscosity reaches a
certain point.  Charging is usually conducted at ambient
temperature.  Temperatures during the reaction increase and may
rise to as high as 70C.
     Some facility operating practices call for purging the
reactors during feedstock addition and reaction stages with
nitrogen to reduce the risk of explosion, and to dilute the vapor
space concentration of toxic compounds in the reactor so that
sampling and/or addition of solids may be done through the
manhole without jeopardizing worker safety.  Typical N2 purge
flow rates are on the order of 20 to 40 standard cubic feet per
minute (scfm).  Reactor vents typically are equipped with surface
condensers, which operate at temperature ranges of 15 to 25C
and at atmospheric pressure.  In these processes, a reflux stream
from the condensers is routed back to the reactor primarily for
temperature control.  If the facility is using a nitrogen purge
on the kettle, the exit gas containing nitrogen is sometimes
routed through the condenser, although the effectiveness of
                              2-40

-------
condensing out such dilute concentrations of volatiles amid a
significantly higher N2 gas flow rate is questionable.
     After the reactor stage of the manufacturing process ends,
the crude product may be purified further by air sparging.  Air
sparging is the subsurface introduction of a gas intended to
remove a more volatile minor component from a liquid.  In this
case, residual EPI may be removed to conform to product
specifications.  Condensers may be employed at this stage to
recover solvent.
     2.2.1.2  Emission Sources.  Emission sources from the
process described above are made up of the following:
(a) storage tank working and breathing losses, (b) vapor
displacement emissions resulting from material transfer or vessel
evacuation, (c) reactor emissions due to heatup or purge,
(d) sparging losses from finishing, and (e) equipment leak
emissions from in-line process components such as pumps, valves,
and flanges.  The. potential for wastewater emissions from reactor
washing also exists.
     2.2.1.2.1  Storage tank working ajid	breathing Bosses.
Storage tank working and breathing losses are typically no
different for this industry than they are for any other industry.
Working losses may usually be eliminated by equipping the tanks
with a vapor return line back to the vessel being offloaded.
Breathing losses, which are caused by temperature fluctuation  and
the subsequent expansion of vessel vapor space that must be
relieved, may be partially abated by applying an inert gas
.blanket.  Nitrogen typically is used for this purpose.  Some
facilities also store feedstock materials in sealed drums, so
that there are no breathing or working losses associated with
material storage.  The charging of material from a drum to a
vessel is sometimes accomplished by first evacuating a vessel  to
a  slight vacuum prior to charge so that the material may simply
be drawn into the vessel from the drum without forced
displacement of any vapors in the receiving vessel.  Emissions
from drums typically are small enough so that the impact to the
ambient concentration of VOC's/air toxics outside the plant
                               2-41

-------
 boundary is  negligible,  although the  airspace  immediately
 surrounding  the drum opening may contain  enough VOC's/air toxics
 to have  an impact  on plant  workers  in the area.
      Table 2-1  presents  some typical  emission  stream
 characteristics from the emission events  described  for  this
 process.   Chapter  3  presents the methodologies used in
 calculating  emissions from  such  events.
      2.2.1.2.2   Vapor displacement.   Vapor displacement losses
 are common types of  emission events in this industry, since bulk
 material  transfer  from vessel to vessel occurs frequently.  Vapor
 displacement losses  from the general  process described above
 would include weigh  tank filling emissions and reactor filling
 emissions.   Incoming material forces  an equal  volume of gas out
 of the vessel.   This displaced gas contains a  certain amount of
 volatile material.   Vapor displacement losses  can usually be
 abated by  providing  vapor return lines from the vessels being
 filled back to  the vessels being emptied.
      2.2.1.2.3   Reactor  emissions.  Reactors will have emissions
 created by charging  of materials  and  subsequent vapor
 displacement.-   Following charging, reactor heatup emissions may
 occur at elevated temperatures,  if no condenser is used and the
 reaction is exothermic.  If  a condenser is used,  the
 concentration of VOC's will  be equal  to the 100 percent
 saturation concentration at  the  condenser outlet temperature.
 For situations  in which a purge  of inert gas is used concurrently
with the reaction,  emissions may be estimated by assuming that
 the purge stream is  saturated to some degree with volatile
material throughout  the purge duration.  Chapter 3 presents more
detailed methodologies for estimating emissions from reactor
heatup and purging events.  The discharging of material from the
 reactor may create displacement emissions in the receiving
vessel,  as is shown  in Figure 2-10.
     2.2.1.2.4  Sparging emission^.   Resin may be sparged with
air or pure nitrogen during or following a reaction.  Sparging is
the use of compressed gas for the agitation of the material in
the vessel and the stripping of trace amounts of volatiles in the
                              2-42

-------
       TABLE 2-1.   SIMPLIFYING ASSUMPTIONS  AND EMISSION STREAM
                 CHARACTERISTICS  FOR ESTIMATING EMISSIONS
Emission type
Storage tank
Breathing losses
Working losses
Vapor displacement
Reactor heatup
With reflux condenser
Without reflux
condenser
Reactor heating with
purging
With reflux condenser
Without reflux
condenser
Product sparging
Flow rates, acfm

NA"
10-20b
b

0.1-5
0.1 -5

10-200
10-200
100-500
Percent saturation

100
100
100

100
100

< 100
25
< 10
Temperature

Ambient
Ambient
Ambient

Oottdenser
temperature
Elevated0

Condenser
temperature
d
d
Duration of emissions

Continuous
Filling time
Filling time

Reaction time
Reaction time

Purge duration
Purge duration
Air sparge duration
'Not applicable.
''The flow rate of displaced gas will equal the filling rate of liquid into the vessel.
cDepends on the reaction temperature, (typically 60 to 200F).
 Determined by a heat balance around the entire contents of the emission stream, including the inert gas used as
a purge or sparge. For dilute VOC
 streams, the exhaust temperature will closely approximate the purge gas temperature.
                                         2-43

-------
material, which is usually a liquid or slurry.  The compressed
air or nitrogen stream is introduced to the material through a
perforated pipe located in the bottom of the sparging vessel.
Another method of sparging is to use an air lift system, in which
the compressed gas is introduced to the material through an
opening in the middle of the longer leg of a "U-tube" having
unequal legs.  The air lift system is used less frequently than
the perforated pipe.  Emissions from this type of operation are
governed by the amount of volatile material remaining in the
product and whether this material may be easily stripped from the
product.  Further discussion of emissions due to gas sparging is
presented in Chapter 3.
     2.2.1.2.5  Emissions from equipment leaks.  Leaking process
components such as pumps, valves, flanges, sampling corrections,
open lines, etc. are sources of emissions.  Emissions may be
calculated using the Synthetic Organic Chemical Manufacturing
Industry (SOCMI) leak factors developed by EPA and multiplied by
the fraction of time that the components are in VOC/air toxics
service.
     2.2.1.2.6  Wastewater emiaaiojis.    Wastewater may be created
from once-through vacuum pump seal water discharge or from
scrubber water, if such a control device is used.  In most cases,
this water is discharged to the plant treatment system, since the
concentration of pollutants in the water is low, precluding the
option of solvent recovery.  Further discussion of wastewater
emissions is presented in Chapter 3.
2.2.2  Pharmaceuticals Industry Description
     The Pharmaceuticals industry uses predominantly batch
processes to manufacture synthetic organic chemicals and to
formulate finished pharmaceutical products.  Most of the batch
unit operations described earlier in this document can be found
in these processes, which makes a discussion of this industry
particularly relevant.  Also, several characteristics of this
industry make the control of VOC's and air toxics emissions
particularly challenging.
                               2-44

-------
      First,  the  equipment used  in  processing  is usually
 campaigned,  meaning  that it  is  not solely dedicated  to the
 manufacture  of just  one product.   Fluctuations in market demand
 often drive  production schedules,  more so than in any other
 industry.  Since the characteristics of emission streams
 emanating  from such  equipment vary according  to product, the
 devices used to  control such streams must either be  easily
 detached and moved or be capable of controlling streams that vary
 widely in  solvent content, temperature, and pressure.
      Such  variation  in emission stream characteristics also
 presents some difficulty to  agencies responsible for issuing
 permits, since the actual emission loveTg of  point sources such
 as reactor kettle condensers and the like will in some cases
 exceed the maximum permitted yearly emission  rates because a
 different  product is manufactured  in ul:e equipment.  In most
 other  industries, plant personnel  would apply for a  permit
 modification.  However, the variability of equipment use and the
 short  production runs characteristic of this  industry make the
 application procedure and permit modification waiting period
 unrealistic.
     A second characteristic of the Pharmaceuticals manufacturing
 industry is the widespread use of  higher volatility solvents such
 as methylene  chloride (dichloromethane),  which is extremely
volatile low photochemically reactive solvent.  Because of its
high volatility,  facilities that do not employ particularly
 stringent  control techniques such as low-temperature
refrigeration units,  pressurized storage tanks,  and closed-loop
processing systems wherever applicable tend to emit large amounts
 of this solvent  (for large facilities,  these amounts can
 sometimes be in excess of 100 tons per year).8
     The remainder of the discussion will revolve around actual
processing characteristics associated with the Pharmaceuticals
 industry.   A typical  simplified process flow diagram is included
as an  illustration of the type of equipment typically used in
this industry and a typical processing route for a common
pharmaceutical product.   Due to the proprietary nature of
                              2-45

-------
 information in  the Pharmaceuticals  industry,  it was not possible
 to present a process description of an unpublished process.
 Therefore, we have intended  for the description of the
 manufacture of  bulk aspirin  to present the kinds of unit
 operations that occur in the Pharmaceuticals  industry.  The
 process description is geared toward a general understanding of
 most Pharmaceuticals manufacturing operations.  Therefore, the
 discussion is also aimed at  including other types of solvents,
 feedstocks, and operational  practices that can be used in this
 equipment.  Table 2-2 presents a summary of typical operating
 characteristics of batch equipment used in this industry.  This
 summary can be  used to characterize uncontrolled emission streams
 from various equipment.  Process steps include product synthesis
 in reactors, purification in crystallizers and filters, and
product finishing in dryers.
     2.2.2.1  Process Description.  Many Pharmaceuticals
manufacturing facilities make a distinction between the synthetic
organic chemicals manufacturing processes that are used to
produce active  ingredients for their final products and those
processes that  are used to formulate the final product
 (Pharmaceuticals processes).  The process flow diagram
illustrating chemicals manufacturing processes is for the
manufacture of  bulk aspirin.  This process is both well known and
appears to be simple,  compared with processes that involve the
manufacture of  numerous intermediates which are reacted to yield
final active ingredient.  Note that the process flow diagram for
the production  of bulk aspirin includes the addition of salicylic
acid in powder  form,  an active ingredient that must be
manufactured prior to this process.
     The fact that product or product intermediates are often in
solid form brings up several elements that must be considered
when evaluating the potential for VOC's or air toxics emissions.
Solids must be  filtered and dried at some point in the process.
Emissions from dryers are considered to be the largest potential
source of process emissions in the Pharmaceuticals industry.
When solids must be introduced into reactors,  as is shown in the
                              2-46

-------
       TABLE 2-2.    TYPICAL  EQUIPMENT OPERATING CHARACTERISTICS
Equipment
Reactors
Pressure filters
Crysttllizers
Dry en
 Tumble
(double
cone)
-Rotary
-Tray
Tablet covers
- Routing pan
- Fluidized bed
Extractors
Capacity,
gallons
500 - 2,000
500-2,000
500-2,000

100-500
200- 1.000
100-300




Purge/exhaust
flow rates,
acfm
3-20/200b
3-20
0

5-20
10 - 10,000
1 -50

500- 1,500
1.500-500
0
Purge/exhaust
duration, hours*
3 - 15/0.05
3-15
3-20

6-20
6-20
6-20

2-3
2-3
-
Temperature
range, F
80-200
80-200
30-50

120- 150
120- 150
20- 100

80-450
80-200
80-200
P range, psia
14.7
40-50
10 - 14.7

10 - 14.7
10 - 14.7
1 - 14.7

14.7
14.7
14.7
Residence time in equipment, if no purge or exhaust.
b3-20 acfm during reaction; 200 for solids addition (2-3 min duration).
cHigh flow rate (20 acfm) used prior to addition of material and during filter cake scraping and discharge.
 Lower flow (3 acfm) during actual centrifugation.
                                          2-47

-------
bulk aspirin process flow diagram, emissions may occur as a
result of the inert gas sweeping of the reactor vessel to reduce
worker exposure and the possibility of explosion,  if the vessel
vapor space contains material in concentrations that approach
explosive limits.  Many processes now call for the addition of
solids into the reactor prior to the addition of any other
reagents.  However, this practice may not always be practical, as
solids tend to cling together and stick to the glass-lined
surface of the bottom of the reactor. This practice also is not
used if the solid reagent must be added over time.
     2.2.2.2  Aspirin ProcessDeqcription.  This process
description was originally described in the June,  1953, issue of
Chemical Engineering.  Although the actual process is outdated,
this type of equipment is still in use today.  The discussion
below presents a brief synopsis of the process.  A process flow
diagram is presented in Figure 2-11.
     According to the 1953 report, feedstocks to the reactor were
acetic anhydride, mother liquor (containing appreciable amounts
of acetic acid), and salicylic acid powder.  The reaction
occurred at temperatures of up to 90C over a period of 2 to 3
hours.  After reaction, the product  (in liquid form) was pumped
through a filter to remove impurities that may have been present
in the feedstocks.  Product liquid was routed to a crystallizer,
where the liquid was cooled down to a temperature of 3C over a
period of about 16 hours.  The slurry from the crystallizer was
then transferred by gravity to a nutsche-type slurry tank, where
a portion of mother liquor was decanted off to use in the next
aspirin batch or distilled to recover acetic acid.  The slurry
was then transferred, again by gravity, to a centrifuge to
separate the remaining mother liquor from the aspirin product.
Product was scraped off the centrifuge walls, transferred to a
dump cart, and then transferred to a rotary dryer for final
drying.  The remaining finishing steps included sifting,
granulation, and tableting, which were not described in the
article.
                               2-48

-------
                                                 Recycled Anhydride
Mother
Liquor
                                                                     Excess
                                                                     Acetic
                                                                     Acid
                                                                       Classifier
                                                                          Aspirin
                                                                          Crystals
                                                                      Coarse
                                                                 Intermediate
                  Figure  2-11.  Aspirin Manufacturing
       (Reprinted  from Stern,  Air Pollution, Volume VII,  1986)
                                   2-49

-------
     The equipment described in this process is still used today
to manufacture pharmaceutical products.  Although instrumentation
and some design characteristics have since been improved (i.e.
bottom gravity discharge centrifuges), the equipment performs the
same tasks in essentially the same way.
     2.2.2.3  Emission Sources.
     2.2.2.3.1  Process vents.
     ^/apor displacement.  Many emissions from process vents in
the Pharmaceuticals industry come from vapor displacement events
that are created during charging and transferring of material in
all types of process vessels including reactors, surge or slurry
tanks,  and distillation kettles.  All of the process vessels
shown in Figure 2-11 undergo vapor displacement during material
transfer.  The extent of VOC and air toxic emissions from such
events will depend on the material and quantity transferred, the
material and/or vessel temperature, and the rate of transfer of
the material, since this last parameter affects time allowed for
equilibrium between the liquid and vapor phases.  The vessel
vapors that are displaced usually are assumed to be completely
saturated with volatile material that is being charged or
transferred.
     2.2.2.3.2  Reactor emissions.  Some reactors used in the
Pharmaceuticals industry are called autoclaves.  Autoclaves are
pressurized reactors.  There are no emissions from such reactors
because the systems are fully closed except during material
transfer.  In general emissions from reactors occur when
atmospheric vents are present.  If the material inside a vessel
experiences an increase in temperature, two things occur:
(1) the inert gas inside the vessel expands, causing a certain
amount of gas to be displaced, and (2) the volatilization of
liquid material increases.  Emissions from reactors due to
temperature increases occurs routinely.  Reactors usually are
equipped with condensers to reflux condensate during temperature
increases, if the material inside the reactors is volatile.
     Inert gas purges also may be employed to reduce the risk of
explosion during reactions.  In such cases, emissions are
                               2-50

-------
increased because of such purges since volatile material is swept
out and the condensers that were installed to control
condensables usually were nui. i^<=u to abate the volume of gas
that passes through as a result of the purge.  Consequently, many
facilities do not route the existing purge stream from reactor
kettles through condensers because they are largely ineffective
at controlling the relatively small amount of volatile* contained
in the streams.  Purge volumes are determined by the lower
explosive limits of the material in the reactors, but generally
range from 3 to 20 standard cubic feet per minute (scfm);
although some facilities, especially prior to opening up manholes
for solids addition, use purges on the order of 200 to 300 scfm.
In these later cases, the purges are used to prevent workers who
must manually add the solids from being exposed to toxic gases.
These purges typically only last for periods of 2 to 5 minutes
(the length of time that solids addition may take) as opposed to
lasting the entire duration of reaction (as in lower-volume
purges that are used to prevent explosions). ^  Chapter 3 details
the methodology that should be used to calculate emissions from
purges.
     2.2.2.3,3  EmissionsFrom Centrifuges. Filters, and Dryers.
Centrifuges, filters, and dryers function by separating solvent
from solid product.  For some types of equipment, the majority of
emissions will occur during unloading of the solid product from
the process equipment.  Older-style centrifuges, for example,
must be unloaded manually by opening up the cover over the basket
and scooping out product.  Newer centrifuge designs use bottom
unloading mechanisms that do not require manual unloading.  Inert
gas purging of the closed vapor space surrounding the centrifuge
basket to reduce the possibility of explosion and reduce worker
exposure increases VOC emissions.
     Filters also separate solvent from solid product.  Emissions
.from filters also occur mostly during unloading of the units,
since they, like centrifuges, are enclosed during the actual
separation operation.  Some filters, such as nutsche filters, can
be bottom-discharged so that solvent emissions from manual
                               2-51

-------
unloading may be minimized.  Nutsche filters also are also
capable of convectively drying product by passing a heated
nitrogen stream through the filter cake after the filtration step
is completed.  Drying of product in the filter minimizes solvent
emissions that will occur when the material is transferred from
filter to dryer.  Control systems also exist whereby the heated
inert stream that passes through the cake in the drying step is
compressed, condensed, and recirculated in a closed-loop system
so that drying emissions are virtually eliminated.  Chapter 4
presents a detailed description of this particular system.
     Emissions from other types of dryers, such as tumble
(double-cone) dryers, rotary dryers, and fluid bed dryers, occur
when, the solvent-laden drying stream exiting these enclosed
process units are not contained or controlled.  Since material to
be dried often contains as much as 50 percent by weight solvent,
controlling these emission streams will mean significant
emissions reduction.
     The characteristics of the emission streams from dryers
depend largely on the flow rates and temperatures of drying
streams and on the amount of solvent that must be removed.  When
the product is highly temperature-sensitive, a vacuum may be
pulled on the drying vessel to allow volatilization at lower
temperatures.  A typical rotary dryer with a capacity of 20 ft3
could operate at a vacuum of 4 to 5 pounds per square inch
(gauge).  The vacuum in this situation could be created with a
liquid seal vacuum pump.  Note that dryer emissions would be
partly controlled by the vacuum pump, although contamination of
seal water would exist.
     Dryers that are used for tablet coating are also quite
common in this industry.  The two most common types of tablet
coaters are rotating pan and fluidized bed coaters.  A typical
rotating pan tablet coater has a spray which coats tablets that
are sitting in rotating open-ended pans while a stream of warm
air (100F) flows across the tablets at a typical rate of
1,000 acfm.  Drying takes approximately 2 to 3 hours.  Another
process used for tablet coating, the wurster process, uses a
                               2-52

-------
fluidized bed in which the tablets are suspended in a vertical
column while the spray solution is applied.  The exhaust gas
volume from the wurster process typically will be as high as
5,000 acfm.  Coatings can be both water based and organic solvent
based.  A typical organic coating solution consists of 80 to 85
percent methylene chloride, 10 percent denatured ethanol or
isopropanol, and 5 to 10 percent solids.  One gallon of coating
will generally process 25 pounds of tablets.  Assuming that
600 pounds of tablets can be coated in each of the tablet coaters
described above, as much as 250 pounds of methylene chloride
could be emitted over the 2-3 hour drying span.10  Example
calculations of VOC emissions from dryers is contained in
Chapter 3.
     2.2.2.3.4  Emissions from extractors, distillation columns.
and crystallizers.  Emission streams from these pieces of
equipment generally are less significant sources of VOCs and air
toxics.  Distillation columns and crystallizers are usually
equipped with condensers for refluxing.  Emissions from
extraction columns are essentially vapor displacement emissions
that occur while the columns and associated surge tanks are being
filled or discharged.
     2.2.2.3.5  Storage emissions.  Storage vessels in the
Pharmaceuticals industry typically are on the order of 2,000 to
10,000 gallons.  For highly volatile solvents such as methylene
chloride, the tanks may be pressurized (to pressures of 100 psi)
or refrigerated to eliminate breathing losses.  Breathing and
working losses may be calculated from the information contained
in Chapter 3.
     2.2.2.3.6  Emissions from equipment leaks.  Fugitive
emissions from leaking pumps, valves, flanges, and other process
components may be calculated using the AP-42 SOCMI equipment leak
factors.
     2.2.2.3.7  Emissions from wastewater.  Wastewater may be
generated from vacuum pump seal water, scrubber discharge, or
from condensation or feedstocks in the processes themselves.  In
cases where wastewater containing significant amounts of VOC's or
                               2-53

-------
air toxics is sent to the facility treatment system, emissions of
VOC or air toxics may occur.  Chapter 3 presents further
discussion.
2.2.3  Pesticide Manufacturing
     The manufacture of pesticides, the broad term for
agricultural chemicals including herbicides, insecticides,
nematicides, and fungicides, is often done using batch processes
due to the low volume or complex processing required.  The
synthesis of the active ingredient, or "technical" product, is
quite similar to the batch manufacture of any synthetic organic
chemical with the notable exception that many pesticides have
high mammalian toxicity.  Therefore, higher levels of worker
protection are required.  The preparation of the final product
for the end-user, called formulation, has other special
processing requirements for the preparation of the solid
substrates, diluents, "inert" ingredients, and product packaging.
A preponderance of batch equipment may also be used in the
formulation step.
     The active ingredients of pesticides typically fall into two
categories:  low volatility  (high boiling point) oils with
limited solubility in water and solid organic compounds also
typically with limited water solubility.  There are some special
cases where pesticides are high volatility liquids  (or even
gases), particularly in the case of fungicides, but these cases
are unusual.
     Limited solubility is desired in order to gain persistence
in the target environment by minimizing washoff and leaching.
However, this generally means that the manufacture takes place
using organic solvents (in which the compounds are soluble),
which are potential sources of VOC emissions.  Due to the low
vapor pressures  (low volatility) of the active ingredients, the
vaporization of the compounds themselves is not particularly a
source of any significant VOC emission at the point of
manufacture.  Due to the toxicity of many compounds, however,
worker exposure must be carefully controlled.  At the point of
use, typically outdoor agricultural activities, the active
                               2-54

-------
ingredients will vaporize over time and are a source of
significant VOC emissions.
     Some herbicides are salts of organic compounds and as such
have significant water solubility.  This makes the component
"systemic" in that it can be absorbed by the plant and be
transported even to the roots (e.g., for control of perennial
weeds).  These compounds use water as a solvent in at least a
portion of their manufacturing.
     The following is a generalized description of a one reaction
step commercial synthesis of a "typical" liquid-type insecticide
active ingredient.  Due to the proprietary nature of the
pesticide manufacturing industry, no particular commercial
compound is being described, but this description provides a
discussion of the types of process steps which might be used to
manufacture an actual compound.
     Raw material A, dissolved in low or moderate boiling point
organic solvent,  is charged to a 2,000 gal,  glass-lined or
stainless steel reactor.  Initial reaction temperature is
attained by admitting tempered water to the external heating
jacket of the reactor.  Prior to charging, the reactor had been
purged with several volumes of nitrogen, and during the reaction
a small purge flow is maintained to prevent the buildup of oxygen
in the head space where flammable solvent vapors are present.
     A second raw material, B, is introduced over a period of a
couple of hours in order to conduct a "semibatch" reaction.  The
reaction produces heat which is removed by letting the solvent
boil off and condensing and returning it to the reactor with the
reflux condenser.
     Following completion of the addition of B, the entire
reactor contents are transferred hot to a second reactor for
several hours of "cook out" where the final fraction of the
reaction takes place.  The crude product is then pumped through
an in-line filter for removal of any solid catalyst that may have
been used or for removal of solid by-products.  By-product water
or excess solvent may be removed as a lower layer using a
decanter, and the mixture is pumped to a solvent removal system.
                              2-55

-------
     At this point, the batch process may be considered to be a
continuous process because a feed or "equalization" tank will be
constantly replenished with the product for each batch reaction.
Due to the difficulty of startup and shutdown, a flash evaporator
or distillation column or columns will typically be operated in a
continuous mode.  The final liquid active ingredient or
"technical" product is recovered as the bottom product.  Due to
their low volatility, technicals are rarely distilled for final
purification; they are recovered as "residue" products.
     If only a small amount of solvent was used in the synthesis,
a tank-based "drying" scheme may be used where the crude product
is subjected to a vacuum for a period of time during which the
volatile components "weather off" and are partially condensed
with a vent condenser.  The capacity of such a system may be
enhanced using subsurface sparging of an inert gas.
     The final liquid "technical" product may be packed out in
drums or other bulk shipping containers for transport to the
formulation plant.  This plant is often at a different location
and is likely be a different company.
     The manufacture of a "typical" solid pesticide active
ingredient, an herbicide, for example,  is similar to the above
except that some sort of solid-liquid separation equipment is
required.  Again, the following description is of no particular
commercial compound, but merely provides a description of the
types of process steps which might be used for the batchwise
manufacture of a solid pesticide.
     Raw material A is charged to a 5,000 gal stirred tank
reactor which has already been filled with aqueous
crystallization mother liquor from a previous batch.  Raw
material B is fed to the reactor over a period of a few hours;
the reaction produces product C which,  being only partially
soluble, forms a slurry of fine crystals.  Using a liquid ring
vacuum pump, a vacuum is pulled on the reactor in order to
promote the desorption of an acid gas by-product of reaction.
The discharge of the vacuum pump is directed to a packed column
                               2-56

-------
 vent  scrubber using  caustic  and  sodium  hypochlorite  as  scrubbing
 liquid.
      The  reactor  contents, a slurry, are pumped  to a centrifuge
 feed  tank, and  the reactor is prepared  for another batch.  By
 using a bank of reactors, the centrifuge feed tank can  be kept
 filled, and the rest of the  process will operate nearly
 continuously.
     An automatic-dump, basket-type centrifuge is used  to
 separate  the crystals of crude product  from the mother  liquor.
 The mother liquor goes to a  holding tank from which  a small bleed
 stream is discharged to wastewater treatment and the majority is
 returned  to fill  a reactor.
      A bank of  centrifuges is used so that the flow  of  wet cake
 into  the blender  is nearly constant.  A portion of dried final
 product is fed  to the blender for wet cake conditioning.  The wet
 cake  is then fed  to a recirculating gas dryer.
      The drying gas, heated  nitrogen, is introduced  into the
 bottom of the dryer.  The drying gas conveys the product upward
 and carries moisture from the product.  The hot, moist  gas is
 then  quench-cooled with water in a venturi scrubber.  The
 scrubbing liquid,  laden with dissolved product fines, is recycled
 to the mother liquor tank.  The dry product, which is separated
 from  the gas stream using a  cyclone and bag filter arrangement,
 is pneumatically  conveyed to interim product storage.
      The  "technical" solid will eventually be packed out into
 drums, tote bags,  or other solids handling containers for
 shipping to the formulation plants.
     The formulation of the pesticide active ingredients into a
 concentrate,  powder, or granule that is more usable by  the
 consumer also heavily utilizes batch processing techniques.
Although there are no reactions,  per se, the grinding,  slurrying,
mixing,  coating, and drying steps all have the potential to
produce VOC emissions.   The following is a brief summary of the
ma-jor types of pesticide formulations with a description of the
unit operations that could be used to manufacture them,  in
                               2-57

-------
general,  they fall into the categories of sprays,  dusts,  and
granules.
     2.2.3.1  Sprays.   A very common spray formulation type is
the emulsifiable concentrate (EC)  which is a concentrated oil
solution of an active ingredient which is typically insoluble in
water, but is soluble in petroleum solvents.  The concentrated
solution is then diluted with water in the field before
application.
     For liquid active ingredients, the manufacture of an EC
formulation is straightforward in that only simple mixing is
required.   Dissolving solids into the organic solvent may require
specialized agitation equipment, but it is still straightforward.
Because of the need for rigorous quality control in the final
product concentration, a reactor on weigh cells may be used, and
the formulation ingredients, the oil, the emulsifier, and the
"technical" active ingredient may be added by weight rather than
volume.  Each batch can then be assayed for the exact content of
active ingredients; adjustments can be made before packing out
into the final containers.
     Potential VOC emissions arise from the purging of mix
vessels,  the filling of tanks and mix vessels, and from the hoods
and vents provided for worker safety over the packaging lines.
     Wettable powders (WP), another common type of spray
formulation, are concentrated dusts of a solid active ingredient
combined with a finely ground dry carrier, such as mineral clay,
and a wetting agent.  In use, the active ingredient gets
suspended in the water in a well-agitated mix tank.  The
manufacture of these formulations usually requires fine grinding
or milling of the active ingredients, and dry blending of the
solid constituents.  There is a significant risk of dust exposure
associated with the manufacture of WP's but limited opportunities
for VOC emission.
     Water soluble powders are simple formulations for those
active ingredients which are water soluble solids.  As with
wettable powders, fine grinding or dry milling and dry blending
                               2-58

-------
would be required,  with dust hazards but little VOC emission
potential.
     Flowable formulations are very finely ground solid active
ingredients in permanent suspension with liquid carriers and
exnulsifiers.  The result is a "liquid" form of a water-insoluble
pesticide without the use of organic solvents.  There is, by
definition, little potential for VOC emissions associated with
flowable formulations.
     Dry flowable formulations, also known as water dispersible
granules, are small granules of a solid active ingredient and
emulsifier.  The manufacture of these granules may involve
slurrying of the finely divided dry active ingredient followed by
drying in special equipment  (e.g., spray drying or a prilling
tower) which is designed to form the desired granule size.  There
is'a small potential for VOC emissions associated with the drying
operation if organic solvents were used during the reslurrying,
or if there is residual solvent present with the dry active
ingredient that gets released with a secondary drying step.
     Water soluble concentrates are true solutions of pesticides
that are water soluble.  With no solvents involved, there is
little risk of VOC emissions from their manufacture.
     2.2.3.2  Granules.  Granular formulations are the other
major type of formulations.  Primarily intended for soil
application, the granules consist of inert ingredients, usually
clays plus binders, formed into pellets which are then sprayed
with a solution of the active  ingredient.  The pesticide  is thus
deposited on the surface when  the solvent dries, or depending on
the poroity, the pellet may become impregnated with the
pesticide.  Granular  formulations are a way to produce a solid
end product from a liquid active  ingredient,  or by trapping the
pesticide  in the clay pellet,  to produce an end product  which is
less hazardous to the user.
     With  typically low solubilities  in water, the preparation  of
the solution for spraying on the  pellets typically requires  the
use of an  organic solvent.   Obviously,  there  is a high potential
for VOC  emissions with the manufacture  of  granular  formulations.
                               2-59

-------
In order for the pellets to remain free flowing,  the solvent must
be completely evaporated from them before they can be packaged.
In the case of dryers using purge gas,  this implies low solvent
loadings at the end of the drying cycle,  with typically low
efficiency of vent condensers.  The dryer itself  may be operated
continuously whereas the preparation of the solutions and the
substrates may be batchwise.  Alternatively,  a vacuum dryer may
operate batchwise or continuously.
     2.2.3.3  Heotachlor.  Figure 2-12  is a simplified process
schematic for the manufacture of heptachlor,  a pesticide used to
control termites in the soil subsurface.   Because of proprietary
concerns, operating characteristics of  the equipment and other
specific process information could not  be included in this
description.  A simplified process description is presented
below.  A discussion of potential emission points of VOC's and
air toxics follows the process description.
     2.2.3.3.1  Process description.  The process involves
several steps.  Hexachloropentadiene, heptane, cyclopentadiene,
and propylene oxide are reacted to form heptachlor 237
intermediate.  Carbon tetrachloride is  added to the kettle as a
solvent for the crude intermediate.  The kettle contents are then
charged to a chlorinator, where the crude intermediate is reacted
with chlorine in the presence of catalyst to yield heptachlor.
Acid gas from the chlorinator is routed to another process.  The
heptachlor/catalyst/solvent mixture from the chlorinator is then
filtered.  Solids from the filter  (catalyst) are dried while the
heptachlor  (which is in liquid form) is sent to storage prior to
packaging and shipment.11
     2.2.3.3.2  Emission sources.  Emissions of VOC's and air
toxics from vapor displacement due to kettle charging and
material transfer are expected to occur from all process vessels.
Kettle reaction emissions are also expected.  The acid gas  stream
from the chlorinator is pulled off for use in another process.
It is unclear whether VOC's and air toxics that may  be entrained
in this  stream are released at a point later on in the other
                               2-60

-------
to
                Hexane
             Heptane
         Cyctopentadiene

              H2O	
      Hexachloropentadiene
Lt
         Propylene Oxide
             Carbon
         Tetrachloride
                             Reactor
                                 Chlorine
                                    I   Catalyst
                                                 HCI
                                  Chlorinator
                                                               Filter
                        VOC
                                                                          Heptachlor
                                                                                             VOC
                                                   _^\\Heptachlor
                                                        Storage
Catalyst
                                                               Dryer
                                                                                      NCOOC 12 SHWC con 2 12 tOKK
                           Figure 2-12.   Process schematic  for Heptachlor.

-------
 (receiving) process.  Emissions from the catalyst drying step are
 also expected.
 2.2.4  Synthetic Oraanics Chemicals Manufacturing Industry
       (SOCMI)
     The SOCMI uses both batch and continuous processes for the
 manufacture of various chemicals.  Though most larger-volume
 chemicals are manufactured in continuous processing arrangements,
 some products are still manufactured on a batch basis.  The
 manufacture of chlorendic anhydride was chosen to illustrate a
 batch process typical of this industry.  Because of proprietary
 concerns, operating characteristics of the equipment and other
 specific process information could not be included in this
 description.  A simplified process description is presented
 below.  A discussion of potential emission points of VOC's and
 air toxics follows the process description.
     2.2.4.1  Process Description.  Figure 2-13 is a schematic of
 the chlorendic anhydride process.  Maleic anhydride (melting
 point of 60F) is reacted with hexachlorocyclopentadiene (HEX) to
 form chlorendenic anhydride.  The chlorendenic anhydride exiting
 the reactor is routed to a crystallizer, where the addition of
 excess heptane and toluene produces chlorendenic anhydride
 crystals.  The slurry from the crystallizer is then pumped to an
 agitated centrifuge head tank, which discharges the slurry by
 gravity to a centrifuge.  The centrifuged liquid containing
 appreciable amounts of heptane and toluene solvent is recovered
 for further use.  The chlorendenic anhydride from the centrifuge
 is loaded onto a solids conveyor and moved to a vacuum dryer to
 remove remaining solvents.  From the dryer, the final product is
 loaded into drums for shipment.12
     2.2.4.2  Emission Points.  The potential for emissions of
VOC's and air toxics exists for all process vessels shown in
 Figure 2-13.  Because inert gas blanketing may be used to lower
VOC concentrations in the process vessels, emissions are also
 expected from equipment openings, in which case the blanketing
becomes a purge.  Vapor displacement emissions would be expected
 from the various material transfer steps shown in the diagram.
                               2-62

-------
ro

a\
u>
         Hexane
                           Maleic Anhydride
                        Reactor
                       Crystallizer
Toluene
Heptane
Solvent
                                                    Centrifuge
                                                    Head Tank
                                                     Liquids
                                                                      Product
                                                                     Discharge
                                                                                                 Vacuum
                                                                                            MOOOCIIMNi M.MJ
                     Figure 2-13.   Process schematic for  Chlorendic Anhydride.

-------
The conveyer used to move material from the centrifuge to the

dryer may also be a source of emissions.  The dryer vacuum

exhaust stream also would be a source of VOC and air toxics

emissions.

2.3  REFERENCES FOR CHAPTER 2


 1.  Moir, D. N., Selecting Batch Pressure Filters, Chemical
     Engineering, 89:47.  July 26,1982.

 2.  Telecon.  C. Hale, MRI, and S. Sobolewski, Rosemund, Inc.
     June 5, 1990.

 3.  O'Shea, Stephen, M.,  Nitrogen Blanketing of Centrifuges
     Reduces Fire Hazards, Chemical Engineering, 90:73.  April  4,
     1983.

 4.  Telecon.  C. Hale,MRI, and R. McKinney, Heinkel Filtering
     Systems.  May 3,1990.

 5.  Reference 3.

 6.  Root, W. L., Indirect Drying of Solids.  Chemical
     Engineering.  90:56-57.  May 2, 1983.

 7.  Bennett, R. C.  Matching Crystallizer to Material.  Chemical
     Engineering.  95:11-P127.  May 23, 1988.

 8.  Pandullo, R., et al.   Radian Corporation.  Memo to Methylene
     Chloride Project File.  Estimates of Potentially Hazardous
     Compound Emissions from Pharmaceutical Facilities and
     Emissions Reductions Achievable with Additional Controls.
     August 27, 1988.

 9.  Shine, B.  Midwest Research Institute.  Emissions and
     Control Memorandum.  Epichlorohydrin--Producers and Users.
     September 30, 1988.

10.  Pedco Environmental.   Volatile Organic Compound Emissions
     Control* for Tablet Coating at Pharmaceutical Plants.
     January 1984.

11.  Pacific Environmental Services, Inc.  Chemical Processing
     Plants Summary Report for Technical Support in Development
     of a Revised Ozone State Implementation Plan for Memphis,
     Tennessee.  Written for the U. S. Environmental Protection
     Agency, Region IV.  June 1985.

12.  Reference 11.
                               2-64

-------
2 .4  BIBLIOGRAPHY

Bennett, R. C.  Matching Crystallizer to Material.  Chemical
Engineering.  95:118-127.  May 23, 1988.

Chemical Engineering. 60:116-120.  p 116-120  (1953).

EPA-450/2-78-029.  OAQPS Guideline Series.  Control of Volatile
Organic Emissions from Manufacture of Synthesized Pharmaceutical
Products. December 1978.

Geankoplis, C. J.  Transport Processes and Unit Operations.
Allyn and Bacon, Inc., Newton, Massachusetts.  1983.

Larson, M. A.  Guidelines for Selecting a Crystallizer.  Chemical
Engineering.  85:90-102.  February 13, 1978.

McCabe, W. L., J. C. Smith, and P. Harriott.  Unit Operations of
Chemical Engineering.  McGraw-Hill Book Company, New York, New
York.  1985.
   *
Moir, D. N.  Selecting Batch Pressure Filters.  Chemical
Engineering.  89:47-57.  July 26, 1982.

Pacific Environmental Services, Inc.  Chemical Processing Plants
Summary Report for Technical Support in Development of a Revised
Ozone State Implementation Plan for Memphis, Tennessee.  Written
for the U. S. Environmental Protection Agency Region IV.  June
1985.

Pandullo, R. et al., Radian Corporation.  Memo to Methylene
Chloride Project File.  Estimates of Potentially Hazardous
Compound Emissions from pharmaceutical Facilities and Emission
Reductions Achievable with Additional Controls.  August 27, 1988.

Pedco Environmental.  Volatile Organic Compound Emission Controls
for Tablet Coating at Pharmaceutical Plants. Jan. 1984.

Perry, R. H., and C. H. Chilton.  Chemical Engineer's Handbook.
Fifth Edition.  McGraw-Hill Book Company, New York, New York,
1973.

Shine, B., Midwest Research Institute.  Project Summary--
Epichlorohydrin-- Procedures and Users Regulatory Decision.
March 16, 1990.

Stern, A.C., Ed.  Air Pollution Vol. VII, 3rd Edition.  Academic
Press, Orlando, Florida.  1986.

Svarovsky, L.  Advances in Solid-Liquid Separation-I, Filtration
and Allied Operations.  Chemical Engineering.  86:62-76.  July 2,
1976.
                               2-65

-------
Treybal, R. E.  Mass-Transfer Operations.  McGraw-Hill Book
Company, New York, New York.  1980.

U. S. Environmental Protection Agency, Emission Standards
Division.  Distillation Operations in Synthetic Organic Chemical
Manufacturing--Background Information for Proposed standards.
U. S. Environmental Protection Agency.  Research Triangle Park,
N.C.  Publication No. EPA-450/3-83-005a.  December 1983.

U. S. Environmental Protection Agency, Emission Standards
Division.  VOC Emissions from Volatile Organic Liquid Storage
Tanks--Background Information for Proposed Stajfdarda.
U. S. Environmental Protection Agency.  Research Triangle Park,
N.C.  Publication No. EPA-450/3~8l-003a.
July 1984.  

Williams-Gardner, A.  Industrial Drying.  Gulf Publishing
Company, Houston, Texas.  1977.
                             2-66

-------
             3.0  EMISSION ESTIMATION METHODOLOGIES

     The methodologies presented in this chapter are based on the
Ideal Gas Law and on fundamental vapor/liquid equilibrium
relationships such as Henry's and Raoult's Law,  unless otherwise
specified.  The equations are intended for use in estimating and
characterizing uncontrolled emission streams from batch
processes.  Example calculations are presented in Appendix C.
Significant errors may result in using the examples for
situations that do not meet the conditions and/or assumptions
that are clearly stated in the presentation of the methodologies.
Chapter 4 presents a discussion on control efficiencies for
various devices which may be used in conjunction with the
characteristic streams presented in this chapter.  Necessary
constants and chemical data are presented in Appendix A.
Appendix B discusses calculational issues and Appendix C contains
example calculations for the methodologies described.
3.1  PROCESS VENT EMISSIONS
     Process vent emissions may result  from a number of different
types of events.  Methodologies for various emission events  are
presented in order of  importance relative to the potential
magnitude of their VOC emissions.  Common process vents that
occur in batch processing result from drying, tank  and reactor
inert gas purging, vapor displacement losses from material
transfer, tank and reactor heating, gas evolution,  gas sparging,
batch pressure filtration, and vacuum generation.   The discussion
below presents the principles and methodology for estimating
emissions from these events.
                                3-1

-------
3.1.1  Drying
     Two types of drying operations commonly occur in batch
processing.  These are conductive drying, in which heat is
transmitted to the material to be dried by contact with heated
surfaces, and convective drying, in which heat is transmitted by
hot gases which are in contact with the material.  Conductive
drying may occur under vacuum conditions or at atmospheric
pressure and in several types of dryers, including tray dryers,
tumble (double-cone) dryers, pan dryers, and rotary dryers.
Convective drying occurs at atmospheric conditions.  Convective
dryers include rotary dryers, fluid bed dryers, and spray dryers.
     The methodology for calculating emissions from the types of
dryers described above is essentially the same.  In general, the
rate of drying of the material depends on many factors associated
with the specific drying situation  (i.e., moisture or solvent
content of material to be dried, heat and mass transfer
parameters, drying period, etc.), but generally decreases with
time so that a large percentage of liquid will be removed during
the initial portion of the drying cycle.
     Studies on the theory of drying of solids usually relate
drying rate to moisture content of the solida.  Three distinct
periods of drying can be observed:  the constant-rate period
where surface moisture is evaporated; the funicular state where
capillarity of the liquid in the pores influences the drying
rate; and the pendular state where the moisture content is so low
that capillary action ceases and the liquid must diffuse through
the pores of the solid.  Each of these three periods of drying
has successively lower rates of drying; the final drying rate,
when the moisture content is zero, is of course also zero.1
     Dryer design specialists, usually employed by the vendors of
drying equipment, can relate the drying curve  for solids to rate
of drying expressed as a function of residence time.  Laboratory
or pilot-scale experiments are often needed to establish the
     aln dryer nomenclature moisture can also refer to organic
solvent content, not just water content.
                               3-2

-------
correct dryer size, operating temperature, gas flow rate, cycle
time, etc.  After a dryer is installed and operating correctly,
the only data usually readily available are the cycle time, gas
flow rates, and moisture content of the solid at the start and
finish of the cycle.  If very dry solids are produced, i.e., zero
moisture content, it is clear that the drying rate at the end of
the cycle will be asymptotic to zero.  This end point condition,
and a knowledge of the total solvent removed, can be used to
estimate the emission characteristics of an existing, installed
dryer.
     In industrial practice, the filr.er rakes and eentrifuged
solids which must be dried often appear dry and free-flowing.  In
fact, they may contain as much as 50 percent solvent by weight.
At the end of the drying cycle the solvent content is reduced to
the percent or fractional percent level.  as required by the
process.  From a mass balance, the total amount of solvent
emitted can be calculated.
     In order to properly size emission control equipment for
dryers, however, one must know the instantaneous emission stream
characteristics throughout the cycle.  Although the precise
values can only be determined from extensive laboratory testing,
a reasonable estimate can be obtained by assuming that the rate
of decrease of drying rate is linear over the length of the
drying cycle starting from the initial, highest value and
declining to zero at the end of the cycle.  Prom the material
balance on solvent removed, the average drying rate can be
calculated knowing the length of the cycle.  From simple analytic
geometry it is apparent that with 
-------
noting that the vacuum-generating equipment  (vacuum pump or steam
ejector) itself may reduce, or may increase, the amount of VOC in
the discharge stream  (see description below.)
     3.1.1.1  Vacuum Dryers.  Tray dryers, double-cone dryers,
pan dryers, and rotary dryers may be operated under vacuum
conditions.  Such vacuum dryers vill have an inward leakage rate
of air that will aid in carrying the VOC's and air toxics that
have evaporated off the product out of the dryers.  Vacuums in
the range of IS to 25 inches mercury are typical.  Articles have ',
been published which provide methods of estimating leakage rates
for vacuum systems.2  One such methodology is contained in
Section 3.1.8.  For a single dryer, the air leakage may range
from ,10 to 30 scfm depending on system design and vacuum level
desired.  An example calculation of vacuum dryer emissions is
presented in Appendix C.
     3.1.1.2  Atmospheric Dryers.  While conductive dryers also
may operate at atmospheric conditions, convective dryers operate
at atmospheric pressure or above.  Convective dryers include tray
dryers and fluid bed dryers.  The four types of conductive dryers
discussed in the previous section, tray dryers, double-cone
dryers, pan dryers and rotary dryers also may be operated at
atmospheric conditions.  A stream of air or inert gas is used to
move the volatilized material from the dryer vessel in conductive
drying.  The gas stream is actually the heating mechanism in
convective drying.  In both of these situations, the calculation
of total VOC content emitted during the drying cycle is identical
to the vacuum drying method (a mass balance from initial cake
concentration to final cake concentration).   The estimation of
the maximum dryer emission rate, which is used for sizing of
equipment, is also analogous to the method presented for vacuum
drying.  The flow rate of gas through the system is the dryer
exhaust gas rate.  An example calculation of atmospheric dryer
emissions is presented in Appendix C.
3.1,2  Tank and fteactor Purging
     Consistent with batch processing there is frequently a need
to use a gas stream to purge VOC vapors from either an empty tank
                               3-4

-------
or reactor vessel, or from the vapor space of a partially filled
tank or reactor.  Typical reasons for purging are to maintain
product quality (e.g., by using a dry sweep gas to minimize water
vapor in a system) or to reduce concentrations of flammable
vapors below safety limits (by using an inert gas such as
nitrogen or carbon dioxide).
     3.1.2.1  Empty Tank Purging.  Empty vessels may be purged
intermittently  (e.g., at startup and shutdown, or between
batches) using a displacement purge to remove accumulated vapors.
The estimation of VOC emissions in this case is fairly
straightforward.  Before the purge, the vapor space of an empty
vessel can be assumed to be in equilibrium with the liquid which
was removed.  Thus, the starting concentration is known.  The
final concentration is a function of the number of volumes of
purge gas used.  This can be expressed as a power law when

           xn                                             (3-1)
         Li
where:
     n  the number of volumes of purge gas used;
     x - the fractional dilution per volume change;
    Cf - final concentration in vessel; and
    G - initial concentration in vessel.

     The fractional dilution per volume change assuming perfect
mixing has been shown to be 37 percent.  Thus, equation 3-12
becomes:
          C_
          ^ -  (0.37)n                                       (3-2)
          ci
     This equation does not take into account evaporation of any
residual liquid in the vessel, and assumes no free liquid.  The
equation was derived in the following way:
                               3-5

-------
                                    -qca                    ,3-3)
where:  V - vessel volume
       Ca - concentration of VOC species
        q - volume tic purge rate
        t   time

if:  q - I ft3/min and V - 1 ft3, then equation  (3-3) reduces to:

                            dCa
                             Ca
                                  -dt                       (3-4)
By integrating and setting the following boundary limits:

     t - 0     Ca - C
     t - 1     Ca - Cf, then equation (3-4) reduces to
                    In (Cf/qj  -1                         (3-5)
                    Therefore, Cf - 0.37 CA

See Example 4 in Appendix C.
     3.1.2.2  Filled Vessels.  Pilled or partially filled vessels
are often "blanketed" with inert gas (or even air in the case of
nonflammable solvents) using either "balanced pressure" or
"trickle* control schemes.3  With balanced pressure blanketing,
there is no flow of gas unless the tank liquid level changes
(during filling or emptying) or the pressure rises or falls due
to thermal effects.  The calculation of emissions from this type
of blanketing is analogous to "working and breathing" losses
addressed elsewhere in this document.  For trickle blanketing, a
constant flow of gas is maintained through the headspace.  The
flow rate may be quite low for storage tanks, but may be much
higher for a reactor where removal of water vapor or excess
                               3-6

-------
solvent vapor is required.  The higher flows have also been
referred to as purges or sweeps.
     Due to the lack of test data a traditional approach to
calculating the volatile organic content of a purge gas stream is
to assume that the gas is saturated with the vapors of the liquid
over which it is flowing.  This assumption is thought to be
conservative or worst case in that the VOC content cannot
possibly be greater than saturation (as long as there are no
entrained droplets).  For the purposes of calculating a maximum
expected uncontrolled emissions, rate,  this approach is
acceptable.  However, as is shown in Appendix B, the actual VOC
concentration of the exiting purge gas may be substantially below
saturation.  In fact, calculations show that the percentage
saturation of an inert gas purge stream over a quiescent pool of
liquid is expected to be no more than 10 percent.  Likewise, the
purging of equipment with turbulent liquid surfaces leads to
higher saturation fractions, approaching 100 percent saturation
at lower flowrates.  Since most vessels will presumably be
agitated, the assumption that the exiting purge gas is completely
saturated with VOC is recommended.  However, if the purge is
greater than 100 scfm, a saturation factor of 25 percent is
recommended.
     Another assumption that must be made is that the displaced
gas will fit the conditions of an ideal gas, and therefore, that
the ideal gas law can be used.  Most operations are run at
conditions such as atmospheric pressure and relatively low
temperatures, which allow the application of the ideal gas law.
The VOC emission rate from purging may be estimated by first
calculating the volumetric flow rate of the gas leaving the
vessel, which is made up of noncondensables as well as the VOCs
that have volatilized into the vapor space.  The total rate of
gas exiting a vessel is therefore:
                               3-7

-------
                     pr                                      (3-6)
               PT ~ 2.
where:
     Vr2 - rate of gas displaced from the vessel, scfm
     Vr1 - rate of purge gas, scfm
     PT  - vessel pressure, mmHg
     r(P^x^) - the sum of the products of the vapor pressures and
               mole fractions for each VOC, mmHg  (see next
               section).

The emission rate of VOCs in this exit gas is then calculated
using the following equation:

         (Yi) (Vr) (PT) (MW)
     ER -	                                    (3-7)
               (R) (T)

where:
    ER  - mass emission rate;
    y^  * mole fraction in vapor phase, calculated in Equations
         3-9 and 3-10.
    Vr  - volumetric gas displacement rate  (equal to the total
         rate of gas exiting a vessel, Vr2)
     R  * ideal gas law constant;
     T  * temperature of the vessel vapor space, absolute;
    PT  - pressure of the vessel vapor space; and
    MW  - molecular weight of the VOC or air toxic.

     Table A-4 of Appendix A contains values of the gas constant
R.  An example of a purging emission calculation is presented in
Appendix C.

3.1.3  Vapor Displacement Losses--Transfer Qf Material to Vessels
     Emissions occur as a result of vapor displacement in many
batch operations.  The transfer of liquids from one vessel to
another vessel causes a certain volume of gas to be displaced in

                               3-8

-------
the receiving vessel.  The VOC's that may be contained in this
volume also are displaced.  In many cases, the displaced gas is
vented directly to the atmosphere.  The amount of VOC's emitted
during such an event is limited by the partial pressure of the
components in the gas stream and the vessel pressure.  Usually,
vessel vapor spaces are filled with air (21 percent oxygen,
79 percent nitrogen) or almost pure inert gas, such as nitrogen.
     The degree of saturation of the vent gas with the VOC's must
be assumed or known before any calculations are done.  As is
normally the practice when permit levels are established, a
conservative assumption usually is made to prevent underestima-
tion of emissions.  A conservative assumption in most vapor
displacement calculations is to consider the gas being displaced
to be 100 percent saturated with the volatile compounds that are
catering the vessel.  The following steps are involved in
calculating emissions from vapor displacement events:
     Step .1.  Define the conditions of the displaced gas:
     1.  Temperature 
     2.  Pressure; and
     3.  Volumetric rate of displacement.
The rate of displacement of a gas from a vapor space is equal  to
the rate of filling of that vessel with a liquid.  An example  of
this type of displacement event .is the transfer of liquid
material from one process vessel to another, such as the charging
of a reactor with material from a weigh tank, and the subsequent
emission of gas from the reactor that is saturated with the
vaporized liquid.
     $tej? 2.  Calculate the mole fraction of components in
displaced gas:
     1.  Determine or calculate the vapor pressure of the liquid
compound of interest  (for one specific component, such as an air
toxic) or of the entire volatile component of the liquid  (for
total VOC emissions).
     For one component, this can be done by consulting vapor
pressure tables at the appropriate temperature  (see  Table A-l)  or
by using Antoine's equation, a form of which is presented below:
                               3-9

-------
                       In Pt*  - A-[B/(T+O]                  (3-8)
where:
       Pt - vapor pressure of component i (mmHg);
    A,B,C - compound-specific constants; and
        T - temperature of the liquid (K).

There are several forms of vapor pressure estimation equations
and the reader should make sure that the constants correspond to
the appropriate form and that the units are consistent.  Most
physical property handbooks contain the Antoine equation and the
appropriate constants.  Vapor pressures for some compounds are
presented in Table A-l.  Table A-2 contains vapor pressure
equations for 120 compounds that use five constants.
     If more than one compound is present in the liquid, the
vapor pressures of all compounds in the mixture must be
determined.  After the vapor pressures have been determined, the
partial pressure that the VOC vapor fraction exerts in the vessel
vapor space may be determined by using Raoult's Law, which is a
simple expression that describes equilibrium between an ideal
vapor and an ideal liquid.  The general equation for Raoult's Law
is presented below:
                                                            (3-9)
where:
    y^ -  mole fraction of i in the vapor;
    P^ -  partial pressure of component i;
    x^ -  mole fraction of component i in the liquid,*
   Pi* -  vapor pressure of component i at temperature T; and
  .. PT -  the total pressure in the vessel vapor space.

     Raoult's Law may be used for multicomponent systems,
assuming the compounds are completely miscible in one another,
If the compounds are not miscible, or are only partially
                               3-10

*
p
i
P
T
x P*
i i
P
T

-------
miscible, then they are considered "nonideal" and Raoult's Law
does not apply.  At or above the solubility limit, each compound
exerts a partial , pressure in the vapor space which is equal to
the vapor pressure at that temperature.  Below the solubility
limit, especially dilute solutions comprised of water and trace
amounts of air toxics or VOC's, Henry's Law is used to describe
the relationship between the mole fraction of the compound in the
liquid and the vapor phase mole fraction.  Henry's Law is
presented below:

           pi-   xi Hi
     y    _i .   i-                                     (3-
      i    P-J.     PT
where :
     Xj_  mole fraction of component i in the liquid;
     H^  Henry's Law constant for i (at temperature T) ;
     y.s  mole fraction of component i in vapor; and
     PT - the total pressure in the vessel vapor space.

This relationship is especially important in calculating
evaporative losses from process wastewater.  Henry's Law
constants for some organic compounds at 25C are presented in
Table A- 3.  Also, a method of correcting the constants for
different temperatures follows Table A-3 in Appendix A.
          3 .  Calculate the emission rate:
     Once y^, the mole fraction of component i in the vessel
vapor space is known, the VOC or air toxic emission rate may be
easily calculated by multiplying yA by the vessel fill rate
 (which equals the gas displacement rate) and converting this
volumetric rate to a mass emission using equation 3-7.  Examples
of vapor displacement emission calculations are presented in
Appendix C.
3.1.4  Vessel Heating
     When a process vessel partially filled with a VOC or a
material containing a VOC is heated, the gas and vapors in the
headspace expand and are discharged from the vent.  An estimate
of the emissions in the uncontrolled vent stream from such an
                             /
                               3-11

-------
event can be calculated based on application of the Ideal Gas Law
and on vapor- liquid equilibrium principles.
     The basis of the calculation is that the moles of gas
displaced from a vessel are a result of the expansion of the
noncondensable gas upon heating, and an increase in the VOC vapor
pressure.  The assumptions made for the calculations which follow
are  (1) atmospheric pressure of 760 mmHg and (2) the displaced
gas is always saturated with VOC vapor in equilibrium with the
liquid mixture.
     The initial pressure of the gas (noncondensable) in the
vessel is given by:

     Pax - 760 - (Pi)Tl                                    (3-11)
     
where :
       Pax - initial partial pressure of gas in vessel
             headspace, mmHg, and
    (pi}Ti " initial partial pressure of each VOC in vessel
             headspace, mrnHg, at the initial temperature  (T  ) .

     The choice of formula for  calculation of P^ depends on  which
vapor- liquid equilibrium assumption is chosen  (as discussed  in
Section 3.1.1).  If the VOC species behaves "ideally" in the
system under consideration, then Raoult's Law holds and
           
-------
          (piJTi " (Hi)T1xi                                 (3-13)

where:
    H^  - Henry's Law Constant at Ti in consistent units
         (atm/mole fraction); and
    Xj_  - Mole fraction of each compound in the liquid mixture.

     Note:   If neither Racult's Law nor Henry's Law is considered
to be valid for the compound mixture being considered, a more
complex procedure, beyond the scope of this document, must be
used.  Commercial computer programs are available to simplify the
task of calculating vapor-liquid equilibria for nonideal
mixtures.
     The calculation of Pi is repeated at the final temperature
conditions,  T2; an(* c^e final partial pressure of the gas in the
vessel  is calculated:

          Pa2 - 760 - E (Pj.)T2                              (3-145

By application of the Ideal Gas Law, the moles of gas displaced
is represented by:
          A| - y[ (-) -  ()1                             (3-15)
               Rl  T!      T2
where:
     AJ - number of Ib-moles of gas displaced;
      V - volume of free space in the vessel ft3;
      R - Gas Law constant, 998.9 mmHg ft3/lb-moles  K;
    Pa1 - initial gas pressure in the vessel, mmHg;
    Pa2 - final gas pressure, mmHg;
     T^ - initial temperature of vessel K; and
     T2 - final temperature of vessel, K.

     The concentration of the VOC in the gas displaced at  the
beginning of the event is calculated assuming equilibrium  at the
initial vessel temperature.  The final concentration of  the VOC

                               3-13

-------
 in the final amount of air displaced is calculated assuming
 equilibrium at the final vessel temperature.  The VOC
 concentration in the displaced gas may be approximated by
 assuming it is equal to the average of the initial and final
 values.  The number of moles of VOC displaced is equal to the
 moles of gas displaced times the average VOC mole fraction, as
 follows:
                              760-{P.)_
                                     1 T
           ''a             5	 x Af,              (3.16)
 where:
     i?g  -  Ib-moles of VOC vapor displaced from the vessel being
          heated up.

 The  weight  of VOC vented can be calculated by multiplying the
 number  of moles by the molecular weight.   The reader should note
 that, at  the boiling point  of the VOC,  this equation goes to
 infinity.   In a physical sense,  the vessel  vapor  space  is filled
 entirely  with VOC during boiling? the rate  of release of VOC is
 therefore equal to the total  flow of VOC  out  of the  kettle.
 Therefore,  this equation is not  valid at  the  boiling point of the
 VOC.  An  example of  a vessel  heatup calculation is presented in
 Appendix  C.
 3.1.5  Gas  Evolution.
     When a gas  is generated  as  a result  of a  chemical reaction,
 emissions may be  calculated by assuming that  the gas  is  saturated
 with any VOCs that are in contact with it at  the exit
 temperature.  Emission calculations are analogous to  the  filled
vessel purging calculations and are calculated using  the
 following formula to first calculate the rate of gas displaced:

                                                            (3-17)
                              3-14

-------
when V]_  - initial volumetric rate of gas evolution
     PT  * vessel pressure
E  (I^x^) - sum of the products of the vapor pressure and the mole
           fraction of each VOC at the exit temperature.

     Once V2 is known, it can be inputted into Equation 3-4 to
calculate the emission rate.  An example calculation of gas VOC
emissions from gas evolution is presented in Appendix C.
3.1.6  Sparging
     Sparging is the subsurface introduction of a gas (typically 
nitrogen or other inert gas) intended to remove by selective
volatilization (stripping) a more volatile minor component from a
liquid mixture of predominantly less volatile material.   Common
applications of sparging are the removal of trace quantities of
water or volatile organic solvent from a low volatility (high
boiling point) resin.  The removal of low concentrations of
organic solvents from wastewater also may be achieved using air
sparging.
     Sparging is a semibatch operation.  The sparge tank is
filled or discharged on a batch basis, while the gas is fed
continuously at a steady flow rate for the duration of the sparge
cycle.  The subsurface sparger is designed to develop a mass of
small diameter bubbles.  The tank may be agitated as well in
order to produce fine bubbles and increase the bubble residence
time.  These design features are intended to increase contact
efficiency.
     Utilizing fundamental chemical engineering principles and
empirical correlations published in the literature it is possible
to calculate the mass transfer coefficients encountered in
sparging applications.  The transfer rate of the component being
stripped out is a function of temperature, composition, liquid
diffusivity, gas rate and agitator power  (which determine bubble
-size), and tank geometry  (which, along with agitation power,
determines residence time).
     For the calculation of equilibrium concentration of VOC in
the exiting sparge gas the earlier discussion of Raoult's Law and
                               3-15

-------
Henry's Law applies.  For simple sparging (low viscosity fluids;
no solids) vapor concentration may approach 100 percent of the
equilibrium value calculated.  For complex sparging, an
empirically determined lower value may need to be used.
     Unlike continuous flow vapor-liquid separation processes,
with batchwise sparging it is not possible to write a series of
simple analytical equations which define the outlet gas
concentration as a function of inlet concentration and
thermodynamic properties of the compounds.  This is because the
liquid flow rate is zero and the composition changes with time.
The problem of estimating the gas composition (hence, VOC
emission rate) at any time during the sparge cycle, or of
determining the amount of sparge gas and sparge time required to
achieve a certain concentration reduction, can, however, be
solved using simple numerical integration.  One chooses a small
time increment, one minute, for example, over which to calculate
the gas flow and composition, making the simplifying assumption
that the liquid composition does not change.  From the inlet gas
concentration  (most likely zero) and the saturated exit gas
concentration, the amount of volatile removed from the bulk
liquid can be calculated, and a new estimate made for the liquid
composition.  The calculation of the vapor composition for the
next time "slice" will be made based on this new liquid
composition value.  The cumulative quantity of volatile removed
is used for subsequent estimates of the liquid composition.
     A graphical representation of the vapor or liquid
composition as a function of sparge time has a characteristic
hyperbolic shape where the composition is asymptotic to zero.
The initial composition is high, as is the stripping rate because
the mass transfer is a function of the composition driving force.
The final composition is low, but a long stripping time is
required to achieve a small decrease in composition because the
driving force  is also very low.  An example sparging
volatilization calculation is presented in Appendix C.
                               3-16

-------
3,1.7  Batch Pressure Filtration
     Pressure filtration of nonaqueous,  volatile,  flammable,  or
hazardous slurries is typically conducted in a closed vessel.
Generally,  VOC's are not emitted during the filtration step,  as
there is no venting from the process vessel.  However, during the
gas blowing (cake-drying) step of the cycle, or during pressure
release prior to cake discharge, venting occurs and there is
potential for VOC emission.
     The gas blowing step is intended to accomplish some
preliminary cake drying by evaporating some of the liquid
filtrate present in the filter cake.  This operation is roughly
equivalent to the constant-drying-rate period of operation of a
dryer except that heated gas is not used  (except in the case of
some special purpose equipment where heated gas is, in fact,
used).  The blowing gas follows the same  flow path as the
filtrate, so that it could be vented through the receiving tank.
     3.1.7.1  Filter Cake Purging.  The emission rate in the
vented purge/blowing gas can be calculated  if the cake conditions
at the start and end of this portion of the cycle are known.  The
filtrate will be evaporated at approximately a constant rate.
Assuming that the filtrate is 100 percent VOC, the emissions  rate
is simply:
                    W(X.-X.)
                ER . 	L_JL-                                (3-18)
where:
      w - the dry weight of a batch of  filter cake;
     Xj - the weight fraction of  filtrate at the start of  the
          gas-blowing  step?
     Xf - the weight fraction of  filtrate at the end of  the gas-
          blowing step;
      t - elapsed time of  gas blowing;  and
     ER - emission  rate in weight per unit  time.

However,  one key piece of  data  required for the above
calculations,  namely the filtrate content of the cake before the
                                3-17

-------
gas blowing, is not usually available.  Therefore, this
methodology is of only limited utility.
     Since the blowing gas causes the VOC's in the filtrate to
evaporate, the gas stream is partially saturated with vapor, and
approaches vapor-liquid equilibrium as a limit.  An assumption of
percent saturation attained enables the calculation of emission
rate.
An example calculation of estimating emissions from filter cake
blowing is provided in Appendix C.
     3.1.7.2  Depressurizati$n.  Prior to opening a batch
pressure filter for solids discharges, the pressure must be
relieved.  In the case of a filter design utilizing a closed
vessel, there is some compressed gas in the vapor space which
will have some degree of vapor saturation of VOC from the
filtrate.  Upon depressurization, a fraction of the noncon-
densible gas along with the VOC vapor will be vented.  The
estimation of the emission rate from a depressurization event is
a straightforward application of the Ideal Gas Law if certain
simplifying assumptions are made.
     If the vessel has been under pressure for some time during
the filter cycle, and no additional noncondensable gas has been
added, then it is reasonable to assume that the gas is saturated
with the VOC vapor at the vessel temperature.  To simplify the
calculations, one assumes that the pressure decreases linearly
with time once depressurization has begun, and that the
composition of the gas/vapor mixture is always saturated with VOC
vapor through the end of the depressurization.  The estimation of
the emission rate proceeds according to the following steps:
     1.  Calculate the mole fraction of each VOC vapor species
initially present in the vessel at the end of the
depressurization.
                  x4 p,
                                                            (3-19)
                               3-18

-------
where :
    Pi  - vapor pressure of each VQC component I;
    P^  m initial pressure of the process vessel in units
         consistent with P^ calculations; and
    y.  - mole fraction of component i initially in the vapor.
     2.   The moles of VOC initially in the vessel are then
calculated using the ideal gas law as follows:
                        (V) (P )
                             L.                            (3-205
     where :
          Yvoc - mole fraction of VOC (the sum of the individual
VOC fractions, IYi)
          V     free volume in the vessel being depressurized
          P1   - Initial vessel pressure
          R    - Gas constant
          T    - Vessel temperature, absolute units

     3.  The moles of noncondensable gas present initially in the
vessel are calculated as follows:
where:

       V -   free volume in the vessel being depressurized;
    Pncx -   initial partial pressure of the noncondensible gas,
       R    gas law constant, K; and
       T -   temperature, absolute units.

     At the beginning of  the depressurization,  there  are more
moles of noncondensable gas in the vessel  relative  to the  moles
-of VOC in the vessel than at the  end of depressurization.   At the
beginning of the depressurization, there are:

                    moles of VOC  to noncondensables

                               3-19

-------
 5.   At  the end of depressurization,  there  are
     "voc
     	          moles  of VOC  to noncondensables
     n2
     where:

                           no * _____                      (3-22)
                                  RT

     where:
       V - Free volume in  the vessel  being  depressurized;
    PNC  - Final partial pressure of  the noncondensible gas,
           P- - n x. P . 
           *2   * Ai *i'
       R - Gas law constant, and;
       T  temperature, absolute.

6.   The moles of VOC for  the duration of the depressurization
may be calculated by taking an approximation of the average ratio
of moles by VOC to moles of noncondensible  and multiplying by the
total moles of noncondensibles released during the
depressurization,  or:
                   nvoc    nvoc
                        2
                                                            (3-23)
                                 [n2 - nj.1  * Nvoc
where: NVOC  moles of VOC emitted

7.   The moles of VOC emitted can be converted to a mass rate
using the following equation:
                       NVOC * MWVOC
                                     Ervoc                  (3-24)
where:
   Ervoc - emission rate of the VOC
   MWVOC - molecular weight of the VOC
       t - time of the depressurization
                               3-20

-------
An example calculation of emissions from vessel depressurization
is provided in Appendix C.
3.1.8  gmj-saiogs from Vacuum Gene rat j qg ffgijipment
     Steam ejectors and vacuum pumps are used to pull vacuums on
vessels and can be sources of VOC and air toxic emissions.  Both
come in contact with a stream of gas that potentially contains
pollutants.  A steam ejector consists essentially of a steam
nozzle that produces a high-velocity jet across a suction chamber
connected to the vessel being evacuated.  The gas from the vessel
is entrained into the motive steam as it passes across the
suction chamber.  Both gas and steam are usually routed to a
condenser.
     Conventional (mechanical-type) vacuum pumps use a high
boiling point oil to lubricate the moving parts.  The VOC's which
*
are present in the gas on the suction side may be partially
condensed in the elevated pressure inside the vacuum pump.  This
reduces the amount of VOC emitted in the gas discharge from the
pump, but causes contamination (reduced lubricity) of the pump
oil.  For this reason, if a significant amount of VOC is expected
in the gas being evacuated, a liquid ring vacuum pump may be
selected.
     In a liquid ring vacuum pump, the vacuum is created by the
rotating motion of a slug of seal fluid inside the pump casing.
The seal fluid is in intimate contact with the gas and VOC being
evacuated.  A portion of the seal fluid is ejected with the pump
discharge, so a system for seal fluid recycle and makeup is
required.
     Because the seal fluid is in contact with the gas/voc
mixture, mass and heat transfer can occur inside the pump.  The
emissions from a liquid ring vacuum pump are, therefore, a
function of the seal fluid temperature and composition, as well
as the inlet gas composition.  For purposes of calculation one
may assume that the exiting gas is in equilibrium with the seal
fluid.  The seal fluid must be chosen to be compatible with the
gas/VOC being evacuated.  In some cases, the seal fluid itself is
a VOC and equilibrium with the exiting gas may result in an
                               3-21

-------
increase in VOC level from that in the suction side.  In other
cases, the seal fluid can act to reduce the VOC level of the gas
stream by absorbing  (or condensing, in the case of a cooled seal
fluid system) some of the VOC in the gas being evacuated.
     3.1.8.1  Emission Estimation.  Emissions from vacuum systems
originate from two distinct sources:  1) the first is the gas at
the vacuum system discharge, 2} the second is the seal fluid or
motive steam.  Calculating emissions from the gaseous discharge
of systems that serve to induce vacuums on equipment involves the
estimztion of the amount of air that leaks into the equipment
because of the pressure differential between the inside and
outside of the vessel.  Once this air leakage rate is known, the
rate of VOC emissions may be calculated by knowing the vacuum
system discharge outlet temperature and pressure.
     The air leakage rate for the equipment may be estimated from
the following equations, which correspond to the leakage created
by metal porosity and cracks and leakage resulting from seals and
components in a system for various vacuum pressure ranges;
     1.  Leakage from mentalj>orosity and cracks
      (For   lP<.10 mmHg}      W - 0.026 pO-34y0.60          (3-25)
      (For  lOsP^lOO mmHg}     W - 0.032 p0.26v0.60D         (3-26)
      (For 100
-------
    $ -  specific leakage rate for components,  Ib/hr/in
         (presented in Table A-5 of Appendix A).

     3.  The total air leakage rate, in Ib/hr  is merely the sum
of the two components W and w.
          La - W+w                                         (3-31)

     Once the air leakage rate is known, the VOC emission rate,
in Ib/hr, may be calculated using the following equation from the
1978 Pharmaceuticals CTG:4
      Se - MWs |f (!W8tem	5- - 15                       (3-32)
                    system "  i
where:
        Se -  rate of VOC emission, in Ib/hr;
   psystem *  absolute pressure of receiving vessel or ejector
              outlet conditions, if there is no receiver;
        PA -  vapor pressure of the VOC at the receiver
              temperature, in mmHg;
        La -  total air leak rate in the system, Ib/hr; and
        29 -  molecular weight of air, Ib/lbmole.
An example calculation is provided in Appendix C.

     Calculating emissions from seal fluid or motive steam is
analogous to the calculations of VOC emissions from other sources
of wastewater, which is discussed below.
3.2  EVAPORATIVE LOSSES FROM WASTEWATER
     Evaporative losses from wastewater that is contaminated with
VOC's has been examined in detail, but currently is not within
the scope of this document.  Several publications are available
to aid the readers in calculating emissions from wastewater
treatment systems which include surface impoundments, lagoons,
and basins.5'6'7
3.3  STORAGE TANK EMISSIONS
     In general, emissions of VOC's from storage tank working and
breathing losses appear to be no different for continuous

                               3-23

-------
processes than they are for batch processes.  Both types of
losses usually are calculated using EPA-derived storage tank loss
equations for three types of storage tanks:  fixed roof, external
floating roof, and internal floating roof.  Fixed roof and
horizontal pressure tanks appear to be the most common storage
vessels used in batch processing.  Estimation equations for these
tank types and a detailed explanation of their use, may be found
in an EPA reference.8
3.4  EQUIPMENT LEAKS                                            '.
     The calculation of emissions of VOC's from leaking process
line components such as valves, pump seals, flanges, and sampling
connections is no different for continuous processes than it is
for batch processes.  Emissions tend to be less because the
amount of time that components are actually in VOC service is
less for batch processes than it is for continuous processes.
     In the event that no other specific data is available
equipment leak emissions may be estimated using the equipment
leak factors derived for the Synthetic Organic Chemical
Manufacturing Industry (SOCMI).  These factors are readily
available, and are included in Appendix A in Table A-6.9  It is
also possible to develop unit-specific emission estimates
according to an accepted EPA protocol.  The methodology for
developing a specific emission estimate for leaking components is
contained in another reference.10
3.6  REFERENCES
 1.  McCabe, W., and J. Smith.  Unit Operations of Chemical
     Engineering, Third Edition.  1976.
 2.  Ryan, J. L., and S. Croll.  Selecting Vacuum Systems.
     Chemical Engineering.  88:78.  December 14, 1981.
 3.  Blakely, P. and G. Orlando.  Using Inert Gases for Purging,
     Blanketing, and Transfer.  Chemical Engineering.  91:97-102.
     May 28, 1984.
 4.  EPA-450/2-78-029.  Control of Volatile Organic Emissions
     from Manufacture of Synthesized Pharmaceutical Products.
     December 1978.
                               3-24

-------
 5.  EPA-450/3-87-026.  Hazardous Waste Treatment, Storage, and
     Disposal Facilities (TSDF) Air Emission Models.

 6.  Wastewater and Wastes Enabling Document.  Version l.o, July
     1992.

 7.  Industrial Wastewater Volatile Organic Compound Emissions -
     Background Information for BACT/LAER Determinations.

 8.  AP-42 Compilation of Air Pollution Emission Factors,
     Chapter 12.

 9.  EPA-453/R-93-026, Protocol for Equipment Leak Emission
     Estimates.  June 1993.

10.  EPA-450/3-88-010, Protocols for Generating Unit-Specific
     Emission Estimates for Equipment Leaks of VOC and VHAP.
     October 1988.
                              3-25

-------
                    4.0  CONTROL TECHNOLOGIES

     This chapter provides information on the types of emission 
control technologies currently available and in use on typical
batch processes.  The discussion is structured so that a general
description of the theory and principles behind the effectiveness
of various common control devices is presented first.  Second,
information is provided on the suitability of the various
technologies for controlling VOC's from different batch unit
operations, followed by a section discussing specific
applications.  Appendix D contains cost calculations and
assumptions made in evaluating costs of thermal incineration and
refrigeration systems for batch processing emissions.
     Because the emission streams produced by batch unit
operations are often of finite duration and the properties of
these streams, such as flow rate, VOC content,  temperature, and
pressure,  often change during the duration of the emission event,
the system chosen for emission control ideally should be capable
of handling both peak flow and nonpeak situations effectively.
To that end, this chapter also addresses the relative importance
of sizing equipment properly.  The following control technologies
are discussed: (1)  condensers, (2)  scrubbers, (3) carbon
adsorbers,  (4) thermal incinerators, (5) vapor containment
systems such as vapor return lines,  i.e., "vent-back" lines, and
(6) operational practices that reduce emissions, such as reduced
nitrogen use for blowing lines, elimination of transfer steps in
product or intermediate handling, and elimination of vessel
opening and purging steps.
                               4-1

-------
4.1  CONDENSERS
     Condensers can generally be classified as surface noncontact
and barometric  (direct-contact condensers).  Surface condensers
are usually shell-and-tube heat exchangers, in which the cooling
fluid flows in 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 fas, 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 batch
processing are on reactors and vacuum-operated devices, such as
distillation columns and dryers.  Note that condensers servicing
reactors and distillation columns often function in refluxing
material.  This refluxing is an integral part of the process, and
therefore these condensers are often not considered to be
emission control devices.  Such applications often use secondary
condensers, which operate at still lower temperatures and
function primarily as control devices.
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 entire
contents of 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
                               4-2

-------
gas stream further after complete (100 percent) saturation is
reached causes condensation of the volatile material.  Heat
removed from condensation is called latent heat.  Both kinds of
heat (which in refrigeration terminology usually are summed and
reported as tons [each ton is 12,000 Itu/hr])  must be considered
in the design of a condenser.  Q, the heat requirement may be
calculated by approximating the sensible and latent heat change
when a gas stream containing condensable material is cooled:
                            *       *
                        Q - mCpAT + mAhv                    (4-1)
where:
     Q  hea't requirement;

     m  mass flow rate of material;
    C  * heat capacity of material cooled;
    AT - temperature difference between inlet material
         temperature and condensate temperature; and
    hv  the latent heat associated with a phase change.

     For a surface condenser, the heat transfer area requirement,
A, may be approximated using the following equation:

                                                            (4-2)
                               uaTu,

where:
       A - heat transfer surface area;
       Q - heat requirement;
       U - overall heat transfer coefficient,  which is based on
           the inside and outside heat transfer, and;
    ATjj,! - log mean difference in temperature between the cooling
           fluid and the exhaust gas at each end of the shell and
           tube exchanger.

     Based on the above discussion, it is apparent that the
amount of material that can be condensed from a gas is limited
only by the following factors:   (1) the inlet emission stream

                               4-3

-------
 properties, including heat capacity and temperature, and  (2) the
 heat transfer characteristics of the condenser, including surface
 area.  By controlling these factors, it follows that nearly any
 amount of cooling can be imparted onto a gas stream.
      In practice, however, the design of condensers and the
 amount of cooling that realistically occurs is based more on
 economics.  Cooling fluid, for example, can range from water at
 ambient temperature to brine, which can be cooled to below the
 freezing point of pure water, to a low-temperature refrigerant. '
 The colder the cooling liquid required, 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.  The remainder of the gas can then be cooled further
 to condense out lower-boiling-point materials without the problem
 of ice formation and subsequent fouling of the heat-transfer
 surface.   Note that the more elaborate a condensing system
 becomes,  the higher the cost of operating that system.
      It has generally been accepted that condensers are most
 effective when applied to gases that contain high percentages of
 condensables.   This is because a large amount of sensible heat
 must be removed  from a gas stream containing mostly
 noncondensable material in order for the stream temperature to
 decrease  to the  saturation temperature of the condensable.
.Obviously,  the farther from saturation a gas stream is,  the more
 sensible  heat  must be removed.
      Verification 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 verify condenser efficiency,  the
 outlet  gas  temperature is the only  value that must be known in
 addition  to  the  inlet conditions (including flowrate of
 noncondensables).   By assuming that the vapor phase of  the
                                4-4

-------
material is in equilibrium with the liquid at condenser outlet
temperature,  the percent by volume VOC discharged from the
condenser may be calculated by dividing its partial pressure by
the total pressure.  In any case,  if condenser efficiency cannot
be calculated because the inlet gas conditions are not known,  it
is at least always possible to calculate the maximum VOC
equilibrium concentration of the exit gas at outlet condenser
conditions.
     Another consideration that must be made when contemplating
the use of a condenser for a particular application is whether
there is an appreciable presence of water vapor in the stream.
There are two reasons for concern in this situation.  The first,
which was touched on in the earlier discussion, is that a surface
condenser cannot effectively function below the freezing point of
the water, as ice will form and create an insulatory surface on
the heat transfer surface, keeping the surface temperature above
32F.  The other consideration, which is more subtle but just as
important to the overall effectiveness of the device, is whether
the water will combine with the condensable material to form a
low-boiling-point 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  Specific Systems and Applications
     4.1.2.1  Reactor Vent Condensers.  Several different types
of condenser systems exist in batch processing applications.
Probably the most common application is the use of the simple
shell-and-tube heat exchanger to control reactor vents.  As was
noted in Chapter 2, emissions of VOC's occur from virtually all
reactor processing and transfer steps, including charging,
reaction, discharging, and cleaning.  In many cases, these
operations occur while a stream of noncondensable or inert gas is
being used as a purge inside the kettle to keep the vapor phase
from reaching explosive limits.  This purge also takes away from
the effectiveness of the condenser as a control device, since the
                               4-5

-------
vapor  fraction of  condensable material  decreases with  the
addition of more noncondensable gas.
     Condensers appear  to be the most common control devices
cited  for reactors.  It may be that these devices are  relatively
inexpensive and easy to use, since they are easily manifolded for
the use of alternate cooling fluids that may be required for the
diverse gas streams resulting from campaigned equipment.
     4.1.2.2  Distillation Columns (Primary and Secondary
Condensers).  Shell-and-tube condensers usually are employed as
refluxing devices  on batch distillation units.  In some cases, a
secondary condenser is used to control  the exhaust gas from the
outlet of the reflux condenser.  The EPA's OAQPS Guideline Series
for the Control of Volatile Organic Emissions from Manufacture of
Synthesized Pharmaceutical Products,  December 1978, establishes
the following guideline for surface condenser outlet gas
temperatures on vents from reactors,  distillation operations,
crystallizers, centrifuges, and vacuum  dryers that emit 6.8 kg/d
(15 Ib/d)  or more  of VOC:
     -25C  when condensing VOC of vapor pressure greater than
            5.8 psi (3Gu minHyI
     -15C  when condensing VOC of vapor pressure greater than
            2.9 psi (150 mmHg)
       0C  when condensing VOC of vapor pressure greater than
            1.5 psi (77.5 nunHg)
      10C  when condensing VOC of vapor pressure greater than
            1.0 psi (52 mmHg)
      25C  when condensing VOC of vapor pressure greater than
            0.5 psi (26 mmHg)
      35Ca when condensing VOC of vapor pressure between 0 and
            0.5 psi (0 to 25 nmHg)
aThis requirement for material with a vapor pressure between 0
 and 0.5 psi at 20C was not part of the 1978 CTG but has been
 adopted by some States.
                               4-6

-------
Based on a review of these guidelines, it becomes apparent that
if the streams controlled are not completely saturated with
VOC's, the guidelines offer very little control.  The discussion
below provides some basis for these conclusions.
     Listed below are VOC's that typically are found in process
vessels such as reactors, dryers, and distillation operations and
their corresponding vapor pressures at 20C.  The corresponding
condenser outlet temperature guidelines as established by the
Pharmaceutical CTG are also listed.

Methanol (MeOH)
Acetone
Toluene
VP in mmHg
at 20C
95
182
22
Required
condenser
outlet,
temp. , C
0
-15
35
VP at
outlet,
mmHg
31
30
22a
Percent
volume at
outlet
4
4
3
aBecause the required outlet temperature is higher than the inlet
  temperature,  no cooling occurs and the stream remains at inlet
  conditions.
     If the streams entering the condenser are at high
temperatures,  then the volume percent of VOC's entering can be
high,  maybe close to 100 percent vapor.  For these situations,
the condensers prove to be very effective.  When a reflux
condenser is used, the condenser isn't considered a control
device, but an integral part of the process.  The material being
distilled off cannot be recovered without the cooling that is
imparted on the gas stream from the condenser.  If there are no
noncondensables present  (i.e., the steam is made up of
100 percent condensable vapors), there are essentially no
emissions at the condenser outlet as long as the condenser is
able to cool the stream below its boiling point temperature.
Therefore, reflux conditions are not considered uncontrolled
                               4-7

-------
emission events.   Atmospheric vent streams created by non-steady-
state distillation operations, however,  are.
     During periods of unsteady-state operation,  such as startup
of an atmospheric distillation operation,  there will be
noncondensables present in the gas stream routed to the
condenser.  If distillation occurs under vacuum,  then some amount
of noncondensables will be present.  This amount can be estimated
by knowing or calculating the leak rate into the system (see
Chapter 3 calculations).
     ,\ secondary condenser may be used to control the above-
described emission events.  For example, the volume percentage of
a saturated methanol stream exiting a condenser is 95/760;  or
12.5 percent by volume at 20C.  Dropping the temperature of this
stream to 0C and thereby reducing the outlet volume percentage
to 4 percent yields a control of approximately 70 percent.
     4.1.2.3  Dryers
     4.1.2.3.1  Vacuum drvers.  Batch dryer exhaust streams,
especially vacuum dryer exhaust streams, have been reported to be
controlled by condensers installed prior to the vacuum-generating
devices  (i.e., vacuum pumps, steam ejectors).  The condensation
of VOC prior to the vacuum-generating device also reduces the VOC
wastewater load since the VOC is removed prior to the point at
which the stream is contacted with the seal water or steam.
     The emission stream parameters generally accompanying vacuum
dryers include high concentrations and low  flowrates.  Over time
the concentration of the emission stream drops off, while the
flowrate usually remains constant.
     To illustrate this situation, Figures  4-1 and 4-2 present
typical drying rate curves for batch dryers.  Figure 4-1
illustrates the cycle time dependency of the actual solvent
content of the material drying.  Figure 4-2 shows how the
emission stream solvent content varies with time.
     The curves illustrate that the majority of the solvent is
removed from the material during the early  stages of the batch
drying cycle.  The corresponding emission rate during these
                                4-8

-------
    % OF TOTAL
      SOLVENT
    CONTAINED
      IN FILTER-
       CAKE
                        TIME
          Figure 4-1.  Filter cake drying curve.
 EMISSION
 STREAM
 EXHAUST
SOLVENT  %
                         PORTION OF THE DRYING CYCLE
                         THAT IS CONTROLLED
                    TIME
     Figure 4-2.  Dryer emission stream solvent content,
                         4-9

-------
stages is also considerably higher.  If a condenser is the device
chosen for VOC control,  it must be sized so that it can handle
the peak VOC flow at the beginning of the cycle.  Also note that
the point marked VSAT in Figure 4-2 is the point in the cycle
where the condenser is no longer effective.  VSAT is the percent
by volume of solvent in the gas stream corresponding to
saturation at the condenser outlet temperature.
     4.1.2.3.2  Convective dryers.  The use of simple condensers
for achieving high degrees of VOC control from convective dryers
is also infeasible because the exhaust gas stream will have a
higher volume percentage of noncondensable gas.
     4.1.2.4  Crystallizers.  Condensers may be used to control
VOC emissions from crystallizers, especially batch vacuum
crystallizers.  Such crystallizers employ both surface condensers
and barometric  (direct-contact) condensers.  Usually, a large
amount of vacuum is necessary to produce crystals at low
temperature.  A typical batch vacuum crystallizer vacuum-
generating system is essentially composed of a three-stage steam
ejector system with an intercondenser  (usually a barometric water
condenser) after the first stage.  Barometric condensers are used
because they are inexpensive from an operating cost standpoint.
However, if the material coming off the crystallizer will become
a concern from the wastewater standpoint, the use of a surface
condenser 'should be considered.
     4.1.2.5  Refrigeration Systems for Manifolded Sources.
Shell-and-tube condensers may be used  to control VOC emissions
from several combined events.  Such applications are usually  for
solvent-recovery purposes, since  it is often desirable to recover
material that would otherwise be  emitted as a VOC.  This is
especially true for industries such as the specialty chemicals
and the pharmaceutical industries  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 recouped
                               4-10

-------
 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 disposal.
     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.
     Some facilities that have a large number of storage tanks,
 for example, are known to use staged refrigeration systems that
 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 to -160F.
     Low-temperature refrigeration' systems such as the one
 described above are 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
 the tank farm situation described above or for a number of
process vents from one manufacturing area,  the system may be most
 effective when it can control the stream having the maximum vapor
 inlet loading at peak flow rate.   Minimization of noncondensables
 in the displacement events is crucial to efficient operation, as
 is maintaining a fairly constant vapor loading rate to prevent
                               4-11

-------
cycling of the refrigeration system's compressors.  Cycling also
occurs if the system is oversized for the vapor load.
     To prevent cycling and to optimize the efficiency of the
system, the displaced vapors or process vents of finite duration
must be staggered or controlled using orifices or flow
controllers so that the system receives a fairly constant vapor
inlet loading.  One such system is currently being used by a
large pharmaceutical manufacturing complex to control displaced
vapors from a tank farm containing approximately 25 tanks.  The
emission rate of methylene chloride, the predominant stream
component, has reportedly been reduced by more than 99 percent,
from 357 Ib/hr to 0.7 Ib/hr.1
     4.1.2.6  Combination of Vapor Compression and Condensation.
In some situations, condensation is aided by compressing the gas
stream containing VOC's to atmospheric pressure (if the stream is
under vacuum) or to some elevated pressure prior to entering a
condenser.  The purpose of this compression step is to condense
out the same amount of material at a higher temperature.  For
example, consider the simple calculation used to estimate the
vapor phase mole fraction of the VOC:
                                                             (4.3)
                           " P
                             *
     A low value of Yvoc is desired at the outlet of the
condenser.  This can be achieved by reducing the numerator value,
PVOC' bv Iwerin9 tne 9as temperature, or by increasing the
denominator, P-pCTAL' by increasin9 tne pressure of the system, or
by a combination of both.
     Most applications that use a combination vapor compression-
condensation system use liquid ring compressors.  These
compressors are available for numerous ranges of flowrates and
discharge pressures.  Liquid ring compressor packages that
include ring seal liquid recirculation systems are currently
available and range in capital cost from approximately $75,000
for a system handling a flowrate of 150 scfm and discharging at a

                               4-12

-------
pressure of 40 to 60 psig to over $200,000 for a system handling
900 scfm at the same discharge pressure.2  The systems are
usually configured so that the pump comes before the heat
exchanger.  However, one pesticide manufacturer uses a
high-pressure liquid ring compressor capable of compressing a gas
to 100 psig in an application to recover methylene chloride from
a solvent vacuum stripping process, following a heat exchanger
that discharges its gases at 4C.  Plant personnel have stated
that prior to installing such a system,  the plant was discharging
approximately 2 million Ib of methylene chloride to the
atmosphere each year, of which 85 to 90 percent is now
recovered.
     There seem to be a number of applications that could make
use of one form or another of these combination systems.  One
such application, which is commercially available, is used to
retrofit a pressure  (nutsche) filter to convert the filter to a
dryer.  This eliminates VOC emissions from associated transfer
steps and essentially makes the drying process closed-loop,
eliminating virtually all VOC emissions.  This system is
described below,
     l.  Description.  Some pharmaceutical facilities make use of
closed-loop drying systems to eliminate emissions of VOC's from
drying steps.4  Figure 4-3 presents a typical closed-loop drying
system.  One such system consists essentially of a high-pressure
liquid ring pump in conjunction with two condensers.  The system
is designed to be used to dry a filtered product cake using a
recirculating stream of heated inert gas.  The most common
application of the system is for recirculating exhaust from
agitated pressure nutsche filters, although the system or some
modification of this system could probably be adapted to use on
most dryer exhaust streams and many streams that contain large
amounts of noncondensables, such as inert purges.
     2.  Basic operation.  Exhaust gas from the dryer or filter
press is drawn into the liquid ring vacuum pump, which compresses
the gas essentially to atmospheric pressure.  The gas contacts
                               4-13

-------
                             HEATER
      (DRYER")
                  SEAL FLUID
        /
EXHAUST GAS
 FROM DRYER
               VACUUM
                PUMP
                                                    SECONDARY
                                                      HEAT
                                                    EXCHANGER
                                        SEAL
                                       _ FLUID
                                        COOLER
COOLING FLUID
                                           SEAL FLUID
                                          HOLDING TANK
                  Figure 4-3.  Closed-loop drying system.

-------
the pump seal fluid in the vacuum pump.  At this point, the pump
acts as a contact condenser because the pump seal fluid is
chilled.  Pump seal fluid and condensed vapors flow into the seal
fluid holding tank, which is kept cold by a ring liquid cooler
positioned above the surface of the liquid in the tank headspace.
The exhausted gases from the pump are also routed across the ring
liquid cooler, which happens to be a noncontact vertical shell-
and-tube heat exchanger.  Some vapors may be condensed from the
exhaust stream at this point since the temperature of the ring
liquid cooler is slightly lower than the temperature of the fluid
in the vacuum pump, especially at the outlet of the pump.
Condensed vapors run down the outside of the tubes and shell
walls to the seal liquid tank.  The exhaust gas in the shell of
the ring liquid cooler is routed to yet another shell-and-tube
heat exchanger, which operates at a lower temperature than the
ring liquid cooler.  Condensed vapors from this second heat
exchanger are also routed back to the ring liquid holding tank.
The holding tank may be equipped with liquid level sensors and
contain an overflow weir to remove excess ring liquid, which can
ultimately be sent to a solvent recovery unit.
     3.  Adaptation to drying systems.  This type of system may
be fitted onto a pressure filter to dry a product cake, thereby
eliminating some emissions that are created from product
transfer.  In addition, the gas stream used to move or vaporize
volatiles, depending upon whether the drying is accomplished
through conduction or convection (most agitated pressure filters
will be more suitable for convective drying),  can be recirculated
so that there are no emissions to the atmosphere.  In such a
system, a heater would be added to the system after the exhaust
gas cooler to heat the inert stream.
                               4-15

-------
 4.2  SCRUBBERS
 4.2.1  General Gas Absorber^
      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 VOC concentrations in the gas and
 liquid versus the equilibrium-defined VOC concentration in the
 two media at each point along the contacting path.  The most
 common types of scrubbers found in batch processing industries
 are packed towers and spray chambers.   For dilute concentrations
 of VOC'a,  impingement-plate towers,  which disperse the vapor
 phase into a large number of tiny bubbles within the liquid phase
 and therefore increase the surface area contact between liquid
 and gas  phases,  are preferred.5
      Gas absorbers are limited primarily by the solubility in the
 liquid stream of the material  to be  transferred to the liquid
 stream.   Most of the scrubbers found in industry use water as the
 scrubbing medium,  so the  effectiveness  of these devices depends
 largely  on the solubility of the VOC's  in water.   In general,
 compounds  containing nitrogen  or oxygen atoms  that  are  free to
 form strong hydrogen bonds  and that  have one to three carbon
 atoms are  soluble;  those  compounds with four or five carbons are
 borderline;  and  those with  six carbon atoms or more  are
 insoluble.6   Common  solvents such as methanol,  isopropyl alcohol,
 and  acetone  are very soluble in  water.   Toluene,  on  the other
 hand, is not.  Although a scrubber could be designed to control a
 VOC  such as  toluene,  the scrubbing medium would have to be  a
 nonvolatile  organic  such as mineral oil.  Although such systems
 do exist, their cost  is relatively high,  since  it is  energy-
 intensive to recover separate  fractions  of mineral oil  and VOC,
 and  the cost of mineral oil precludes the use of a once-through
 system.   Note that one of the considerations associated with the
use of scrubbers is waste stream disposal and/or treatment.
                              4-16

-------
     Since there is usually the transfer of VOC's to the scrubber
effluent stream when a water scrubber is used to control VOC
emissions, regulators should consider the potential for emissions
of VOC from the wastewater collection and treatment system when
evaluating the control device effectiveness.  When VOC loading is
significant, steam stripping of the wastewater may be a viable
and cost effective control.  To estimate emissions and evaluate
control effectiveness for wastewater, a recently revised
publication entitled "Control of Volatile Organic Compound
Emissions from Industrial Wastewater." Draft CTG can be used.8
     Also existing for control of some pollutants are chemical
scrubbers, which, instead of using a liquid medium to absorb
material out of the gas phase, use the liquid medium to react
with material in the gas phase.  A good example is an 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 toxicity is much less 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 VOC 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 as the slope
of the equilibrium curve.
     The estimation of the physical properties of a scrubber
design, such as the number of transfer units  (NQQ) and the height
of transfer unit  (HQG) for a packed tower, may be estimated based
                               4-17

-------
on the reported removal efficiency of a system and the reported
liquid-to-gas mass velocities.  The IPA 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.9  Note that verifying the efficiency of a scrubber is
more difficult than verifying the efficiency of condenser since
there are more variables to consider and the equilibrium data for
VOC in solvent at the required temperature are not always
available.  It is perhaps for this reason that unrealistically
high scrubber efficiencies may sometimes be reported.
4.2.3  Specific Systems and Applicability
     Scrubbers often are used in batch processing as secondary
control devices to condensers.  Scrubbers may be advantageous to
use on streams that have discontinuous properties such as many of
the emission streams from batch processes since scrubbers in most
cases are not as expensive to operate during off-load times as
other control devices.  Although the control efficiency would
decrease with decreasing gas flow rates during off-load times,
the efficiency would pick up -gsin with an increase in gas flow
rate back up to the design value.  The following paragraph
describes one specific application for the control of the solvent
isopropyl alcohol (ItA; with a water scrubber through convective
drying.
     A feasibility analysis of control devices was conducted on a
dryer exhaust stream containing the solvent IPA.  It was
determined that a packed tower water scrubber could achieve at
least 90 percent removal of IPA from the exhaust gas of an
atmospheric dryer.  Three meters of packing were determined to be
required, and 0.4 m^ of water per minute under peak conditions
was determined to be necessary for a peak exhaust gas flow rate
of 6,000 acfm with a 0.4 percent IPA concentration.10
                               4-18

-------
4.3  CARBON ADSORPTION
     Carbon adsorbers function by capturing material 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 batch industries
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 VOC-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 VOC, the VOC
                               4-19

-------
concentration, and the system temperature, pressure, and
humidity.  Overall VOC 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 2SOC or molecular weights greater
than 300, with relative humidities greater than 50 percent, with
high levels of entrained solids, or with temperatures over 100CF.
Adsori-ing 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.  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 greatly decreases the operating
capacity and results in frequent replacement of the carbon.
Plasticizers 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.  These
problems can be controlled by the use of a condenser upstream of
the carbon bed to remove the high-boiling-point components or a
carbon bed guard that can be easily replaced on a regular basis.
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 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 usually increases 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 100P or the
                              4-20

-------
adsorption capacity will be affected.  Inlet stream coolers are
usually required when emission stream temperatures are in excess
of 100F.
4,3,2  Applic^foi ], ity
     Carbon adsorbers are often used as controls for batch
process operations.  At many facilities several VOC sources are
ducted to a single adsorber, since most single emission streams
from batch process operations do not warrant the sole use of an
adsorber.  Emissions from reactor vents,  separation operations,
dryers, and storage tanks may be often controlled by carbon bed
adsorbers.  In many of these applications, the adsorber is
preceded upstream by a condenser.  Since condensers are more
efficient on saturated streams and carbon bed adsorbers are more
efficient on dilute streams, a condenser followed by a carbon bed
adsorber can be an effective control system.
     Nonregenerative carbon adsorbers may also be useful for
batch process operations.  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 reused.  Carbon canisters,  normally the size of
55-gallon drums, can be used to control small vent streams  (less
than 500 actual cubic feet per minute [acfm] <500 ft3/min) with
low organic concentrations.  They are commonly used to control
emissions from storage tanks and small reaction vents.  One
advantage of these systems is that they are immune to normal
fluctuations in gas streams that are common to batch processes.
In fact, most carbon adsorption systems are especially suited for
batch processing,  since the beds do not require continuous energy
input (except for a fan to move the gas).
     When designing and installing carbon bed adsorber systems,
several safety factors 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 VOC in the bed is usually
                               4-21

-------
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 VOC-laden stream
may also lead to bed combustion.  However, preventive measures
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.11  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.
     For all practical purposes, it is difficult to estimate the
efficiency of a carbon adsorption system.  EPA has conducted
several studies which show that a control efficiency of
95 percent is achievable for streams containing compounds that
are considered appropriate (see above discussion) 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 sorbed since the last regeneration or replacement.12
     Note also that it is more difficult to predict the amount of
material that has been sorbed for the intermittent streams with
variable characteristics typical of batch processes than for
                               4-22

-------
continuous emission streams with constant properties.  In some
situations, the VOC's are sorbed out of the gas streams during
peak loading periods and are reentrained during off-peak periods.
In these situations, there is no net control of VOC by the carbon
system.  To prevent inadvertent stripping of VOC's during such
periods, air flow should be diverted from the adsorber during
periods of time when there are no VOC emissions.
     As mentioned previously, most applications of carbon
adsorbers follow condensers.  Because of the highly flammable
nature of many typical solvents, the industry trend is away from
using these devices as primary control devices.
4.4  THERMAL DESTRUCTION
     It is usually possible to route process vents to an
incinerator or flare for control.  Incineration systems are
usually quite costly and must operate continuouslyi therefore the
use of such systems is limited to those applications where a
number of vents may be controlled.  Note also that the byproduct
combustion gases must also be controlled in most cases, thereby
increasing costs.
4.4.1  Flares
     Flaring is an open combustion process that destroys VOC
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.  Flare types can be divided into two main
groups:   (l) ground flares and  (2) elevated flares, which can be
further classified according to the method to enhance mixing
within the flare tip {air-assisted, steam-assisted, or
nonassisted).  The discussion in this chapter focuses on elevated
flares, the most common type in the chemical industry.  The vent
stream is sent to the flare through the collection header.  The
vent stream entering the header can vary widely in volumetric
flow rate, moisture content, VOC concentration, and heat value.
                               4-23

-------
The knock-out drum removes water or hydrocarbon droplets that
could create problems in the flare combustion zone.  Vent streams
are also typically routed through a water seal before going to
the flare.  This prevents possible flame flashbacks, caused when
the vent stream flow rate to the flare is too low and the flame
front pulls down into the stack.13
     Purge gas (N2, CO2, or natural gas) also helps to prevent
flashback in the flare stack caused by low vent stream flow.  The
total volumetric flow to the flame must be carefully controlled
to prevent low-flow flashback problems and to avoid a detached
flame (a space between the stack and flame with incomplete
combustion) caused by an excessively high flow rate.  A gas
barrier or a stack seal is sometimes used just below the flare
head to impede the flow of air into the flare gas network.
     The VOC stream enters at the base of the flame where it is
heated by already burning fuel and pilot burners at the flare
tip.  Fuel flows into the combustion zone, where the exterior of
the microscopic gas pockets 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 completely burned.  A diffusion flame
receives its combustion oxygen by diffusion of air into the flame
from the surrounding atmosphere.  The high volume of flue gas
flow in a flare requires more combustion air at a faster rate
than simple gas diffusion can supply.  Thus, flare designers add
high-velocity steam injection nozzles to increase gas turbulence
in the flame boundary zones, drawing in more combustion air and
improving combustion efficiency.  This steam injection promotes
smokeless flare operation by minimizing the cracking reaction
that forms carbonaceous spot.  Significant disadvantages of steam
use are increased noise and cost.  The steam requirement depends
on the composition of the gs flared, the steam velocity from the
injection nozzle,  and the tip diameter.  Although some gases can
be flared smokelessly without any steam, typically 0.01 to 0.6 kg
of steam per kg of flare gas is required.
                               4-24

-------
     Steam injection is usually controlled manually by an
operator who observes the flare  (either directly or on a
television monitor) and adds steam as required to maintain
smokeless operation.  Several flare manufacturers offer devices
such as infrared sensors that monitor flame characteristics and
adjust the steam flow rate automatically to maintain smokeless
operation.
     Some elevated  flares use forced air instead of steam to
provide the combustion air and the mixing required for smokeless ;
operation.  These flares consist of two coaxial flow channels.
The combustible gases flow in the center channel and the
combustion air  (provided by a fan in the bottom of the flare
stack) flows in the annulus.  The principal advantage of air-
assisted flares is  that they can be used where steam is not
available.  Air assist is rarely used on large flares because
airflow is difficult to control when the gas  flow is
intermittent.  About 90.8 hp of blower capacity is required for
each 100 Ib/hr of gas flared.14
     Ground flares  are usually enclosed and have multiple burner
heads  that are staged to operate based on the quantity of gas
released to the flare.  The energy of the gas itself  (because  of
the high nozzle pressure drop) is usually adequate to provide  the
mixing necessary  for smokeless operation, and air or steam assist
is not required.  A fence or other enclosure  reduces noise and
.light  from the  flare and provides some wind protection.
     Ground flares  are  less numerous and have less capacity  than
elevated flares.  Typically they are used to  burn gas
continuously while  steam-assisted elevated flares are used to
dispose of large  amounts of gas  released  in emergencies.1
      4.4.1.2   Factors Affecting  Flare  effieiencv.16   Flare
 combustion efficiency  is a  function  of many  factors:   (1) heating
value  of  the gas,   (2)  density  of the gas,  (3) flammability  of  the
 gas,  (4)  auto-ignition  temperature  of  the gas,  and  (5)  mixing  at
 the  flare  tip.
                               4-25

-------
     The flammability limits of the gases flared influence
ignition stability and flame extinction.  The flammability limits
are defined as the stoichiometric composition limits (maximum and
minimum) of an oxygen-fuel mixture that will burn indefinitely at
given conditions of temperature and pressure without further
ignition.  In other words, gases must be within their
flammability limits to burn.  When flammability limits are
narrow, the interior of the flame may have insufficient air for
the mixture to burn.  Fuels with wide limits of flammability (for
instance, H2) are therefore easier to combust.
     The auto-ignition temperature of a fuel affects combustion
because gas mixtures must be at high enough temperature and at
the proper mixture strength to burn.  A gas with a low auto-
ignition temperature will ignite and burn more easily than a gas
with a high auto-ignition temperature.
     The heating value of the fuel also affects the flame
stability,  emissions, and flame structure.  H lower-heating-value
fuel produces a cooler flame that does not favor combustion
kinetics and also is more easily extinguished. The lower flame
temperature also reduces buoyant forces, which reduces mixing.
The density of the gas flared also affects the structure and
stability of the flame through the effect on buoyancy and mixing.
By design,  the velocity in many flares is very low; therefore,
most of the flame structure is developed through buoyant forces
as a result of combustion.  Lighter gases therefore tend to burn
better.  In addition to burner tip design, the density of the
fuel also affects the minimum purge gas required to prevent
flashback for smokeless flaring.
     Poor mixing at the flare tip or poor flare maintenance can
cause smoking (particulate).  Fuels with high carbon-to-hydrogen
ratios  (greater than 0.35} have a greater tendency to smoke and
require better mixing if they are to be burned smokelessly.
     Many flare systems are currently operated in conjunction
with baseload gas recovery systems.  Such systems are used to
recover VOC from the flare header system for reuse.  Recovered
                               4-26

-------
VOC may be used as a feedstock in other processes or as a fuel in
process heaters, boilers, or other combustion devices.  When
baseload gas recovery systems are applied, the flare is generally
used to combust process upset and emergency gas releases that the
baseload system is not designed to recover.  In some cases, the
operation of a baseload gas recovery system may offer an economic
advantage over operation of a flare alone since sufficient
quantities of useable VOC can be recovered.
     4.4.1.3  EPA Flare Specifications.  The EPA has established
flare combustion efficiency criteria (40 CPE 60.18} which specify
that 98 percent or greater combustion efficiency can be achieved
provided that certain operating conditions are met:   (15 the
flare must be operated with no visible emissions and with a flame
present; (2) the net heating value of the flared stream must be
greater than 11.2 Ml/son (300 Btu/scf)  for steam-assisted flares
and 7.45 MJ/scm (200 Btu/scf) for a flare without assist; and
(3) steam-assisted and nonassisted flares must have an exit
velocity less than 18.3 m/sec (60 ft/sec).  Steam assisted and
nonassisted flares having an exit velocity greater than
18.3 m/sec  (60 ft/sec) but less than 122 m/sec (400 ft/sec) can
achieve 98 percent or greater control if the net heating value of
the gas stream is greater than 37.3 MJ/scm (1,000 Btu/scf).  The
allowable exit velocity for air-assisted flares,  as well as
steam-assisted and nonassisted flares with an exit velocity less
than 122 m/sec  (400 ft/sec) and a net heating value less than
37.3 MJ/scm (1,000 Btu/scf), can be determined by using an
equation in 40 CFR 60.18.
     4.4.1.4  Applicabj. 1 ity.  Although flares are not as widely
used for controlling emissions from batch processes as other
control devices--for example, condensers, adsorbers, and
scrubbers--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 VOC's or VOC's that have low heats of
combustion.  Flares are capable of handling the highly variable
                               4-27

-------
flows that are often associated with batch process operations.
Steam-assisted elevated flares may be used to control emissions
from high-concentration, intermittent vent streams.  In many
facilities, 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.  They should also be easily accessible to
batch processes because of the multiple burner head design, which
can be stage-operated based on gas flow.  Ground flares can
operate efficiently from 0 to 100 percent of design capacity.
The burner heads can also be specifically sized and designed for
the materials in the flare gas.
4.4.2  Thermal and Catalytic Oxidizers
     Thermal and catalytic oxidizers may be used to control
emission streams of VOC's and air toxics, although they are not
especially suited for intermittent or noncontinuous flows.
Because they operate continuously, auxiliary fuel must be used to
maintain combustion during episodes in which the VOC load is
below design conditions.  In some situations where VOC loading in
the gas to be controlled is small, the environmental benefits of
using fossil fuel and creating products of combustion in order to
combust VOC's on an intermittent basis as opposed to releasing
the uncombusted VOC's must be evaluated by considering the
reduction of VOC compared to costs and production of other
pollutants.
     4.4.2.1  Thermal Qxidizer Design.  Any VOC heated to a high
enough temperature in the presence of enough oxygen will be
oxidized to carbon dioxide and water.  This is the basic
principle of operation of a thermal incinerator.  The theoretical
temperature required for thermal oxidation depends on the
chemical involved.  Some chemicals are oxidized at temperatures
much lower than others.  However, a temperature can be identified
                               4-28

-------
that will result in the efficient destruction of most VOC's.  All
practical thermal incineration processes are influenced by
residence time, mixing, and temperature.  An efficient thermal
incinerator system must provide:
     1.  A chamber temperature high enough to enable the
oxidation reaction to proceed rapidly to completion;
     2.  Enough turbulence to obtain good mixing between the hot
combustion products from the burner, combustion air, and VOC; and
     3.  Sufficient residence time at the chosen temperature for
the oxidation reaction to reach completion.
     A thermal incinerator is usually a refractory-lined chamber
containing a burner (or set of burners) at one end.  Discrete
dual fuel burners and inlets for the offgas and combustion air
are arranged in a premixing chamber to thoroughly mix the hot
products from the burners with the process vent streams.  The
mixture of hot reacting gases then passes into the main
combustion chamber.  This chamber is sized to allow the mixture
enough time at the elevated temperature for the oxidation
reaction to reach completion (residence times of 0.3 to
1.0 second are common).  Energy can then be recovered from the
hot flue gases in a heat recovery section.  Preheating combustion
air or offgas is a common mode of energy recovery; however, it is
sometimes more economical to generate steam.  Insurance
regulations require that if the waste stream is preheated, the
VOC concentration must be maintained below 25 percent of the
lower explosive limit to remove explosion hazards.
     Thermal incinerators designed specifically for VOC
incineration with natural gas as the auxiliary fuel may also use
a grid-type (distributed) gas burner.17  The tiny gas flame jets
on the grid surface ignite the vapors as they pass through the
grid.  The grid acts as a baffle for mixing the gases entering
the chamber.  This arrangement ensures burning cf all vapors at
lower chamber temperature and uses less fuel.  This system makes
possible a shorter reaction chamber yet maintains high
efficiency.
                               4-29

-------
     A thermal incinerator, handling vent streams with varying
heating values and moisture content, requires careful adjustment
to maintain the proper chamber temperatures and operating
efficiency.  Since water requires a great deal of heat to
vaporize, entrained water droplets in an offgas stream can
increase auxiliary fuel requirements to provide the additional
energy needed to vaporize the water and raise it to the
combustion chamber temperature.  Combustion devices are always
operated with some quantity of excess air to ensure a sufficient
supply of oxygen.  The amount of excess air used varies with the
fuel and burner type but should be kept as low as possible.
Using too much excess air wastes fuel because the additional air
must be heated to the combustion chamber temperature.  Large
amounts of excess air also increase fuel gas volume and may
increase the size and cost of the system.  Packaged, single-unit
thermal incinerators can be built to control streams with flow
rates in the range of 0.14 son/see  (300 scfm) to about 24 scm/sec
(50,000 scfm).
     Thermal oxidizers for halogenated VOC's may require
additional control equipment to remove the corrosive combustion
products.  The halogenated VOC streams are usually scrubbed to
prevent corrosion due to contact with acid gases formed during
the combustion of these streams.   The flue gases are quenched to
lower their temperature and are then routed through absorption
equipment such as packed towers or liquid jet scrubbers to remove
the corrosive gases.
     4.4.2.2  Thermal Incinerator Efficiency.  The VOC
destruction efficiency of a thermal oxidizer can be affected by
variations in chamber temperature, residence time, inlet VOC
concentration, compound type,  and flow regime (mixing).  Test
results show that thermal oxidizers can achieve 98 percent
destruction efficiency lux IUUBL VCC compounds at combustion
chamber temperatures ranging from 700 to 13Q0*C (1,300* to
2370F) and residence times of 0.5 to 1.5 seconds.18  These data
indicate that significant variations in destruction efficiency
                               4-30

-------
occurred for C^ to C5 alkanes and olefins, aromatics (benzene,
toluene, and xylene), oxygenated compounds (methyl ethyl ketone
and isopropanol),  chlorinated organice (vinyl chloride), and
nitrogen-containing species (acrylonitrile and ethylamines)  at
chamber temperatures below 760C (1400F).  This information,
used in conjunction with kinetics calculations, indicates the
combustion chamber parameters for achieving at least a 98 percent
VOC destruction efficiency are a combustion temperature of 870
(1600F) and a residence time of 0.75 sec (based upon residence ;
in the chamber volume at combustion temperature).  A thermal
oxidizer designed to produce these conditions in the combustion
chamber should be capable of high destruction efficiency for
almost any nonhalogenated VOC.
     At temperatures over 760C  (1400F), the oxidation reaction
rates are much faster than the rate of gas diffusion mixing.  The
destruction efficiency of the VOC then becomes dependent upon the
fluid mechanics within the oxidation chamber.  The flow regime
must ensure rapid, thorough mixing of the VOC stream, combustion
air, and hot combustion products from the burner.  This enables
the VOC to attain the combustion temperature in the presence of
enough oxygen for sufficient time so the oxidation reaction can
reach completion.
     Based upon studies of thermal oxidizer efficiency, it has
been concluded that 98 percent VOC destruction or a 20 ppmv
compound exit concentration is achievable by all new
incinerators.  The maximum achievable VOC destruction efficiency
decreases with decreasing inlet concentration because of the much
slower combustion reaction rates at lower inlet VOC
concentrations.  Therefore, a VOC weight percentage reduction
based on the mass rate of VOC exiting the control device versus
the mass rate of VOC entering the device would be appropriate for
vent streams with VOC concentrations above approximately
2,000 ppmv  (corresponding to 1,000 ppmv VOC in the incinerator
inlet stream since air dilution is typically 1:1).  For vent
streams with VOC concentrations below approximately 2,000 ppmv,
                               4-31

-------
 it  has  been  determined  that  an  incinerator outlet  concentration
 of  20 ppmv {by compound),  or lower,  is  achievable  by all  new
 thermal oxidizers.19  The  98 percent efficiency  estimate  is
 predicted  on thermal  incinerators operated at  870C  (1600P) with
 0.75 sec residence  time.
     4,4.2.3  Catalytic Qxid^zer Design.   Catalytic  oxidation is
 also a  major combustion technique examined for VOC emission
 control.   A  catalyst  increases  the rate of chemical  reaction
 without becoming permanently altered itself.   Catalysts for
 catalytic  oxidation cause  the oxidizing reaction to  proceed at a
 lower temperature than  is  required for  thermal oxidation.  These
 units can  also operate  well  at  VOC concentrations  below the lower
 explosive  limit, which  is  a  distinct advantage for some process
 vent streams.   Combustion  catalysts  include platinum and  platinum
 alloys,  copper oxide, chromium, and  cobalt.20  These are
 deposited  in thin layers on  inert substrates to  provide for
 maximum surface area between the catalyst  and  the  VOC stream.
 The substrate  may be either  pelletized  or  cast in  a  rigid
 honeycomb  matrix.
     The waste gas  is introduced into a mixing chamber, where it
 is heated  to about  316C (600F5 by  contact with the hot
 combustion products from auxiliary burners.  The heated mixture
 is then passed through  the catalyst  bed.   Oxygen and VOC migrate
 to the  catalyst surface by gas  diffusion and are adsorbed in the
pores of the catalyst.  The  oxidation reaction takes  place at
 these active sites.  Reaction products are desorbed  from the
active sites and transferred by diffusion back into  the waste
gas.21  The  combusted gas may then be passed through a waste heat
recovery device before exhausting into the atmosphere.
     The operating temperatures of combustion  catalysts usually
range from 316" to 650C (600 to 1200F).  Lower  temperatures
may slow down and possibly stop the oxidation reaction.  Higher
temperatures may result in shortened catalyst life and possible
evaporation or melting of the catalyst from the support
substrate.   Any accumulation of particulate matter,  condensed
                              4-32

-------
VOC, or polymerized hydrocarbons  on  the  catalyst  could block the
active sites and,  therefore,  reduce  effectiveness.   Catalysts can
also be deactivated by  compounds  containing sulfur,  bismuth,
phosphorus, arsenic, antimony, mercury,  lead, zinc,  tin, or
halogens.22  If  these compounds exist in the catalytic unit, VOC
will pass through  unreacted or be partially oxidized to form
compounds such as  aldehydes,  ketones, and organic acids.
     4.4.2.4  Catalytic Oxidizer Control Efficiency.  Catalytic
oxidizer destruction efficiency is dependent on the  space
velocity  (the catalyst volume required per unit volume gas
processed per hour), operating temperature, oxygen concentration,
and waste gas VOC  composition and concentration.  A  catalytic
unit operating at  about 450C (840F) with a catalyst bed volume
of 0.014 to 0.057  m3 (0.5 to  2 ft3) per  0.47 scm/sec  (1,000 scfm)
of vent stream passing through the device can achieve 95 percent
VOC destruction  efficiency.   However, catalytic oxidizers have
been reported to achieve efficiency of 98 percent or greater.23
These higher efficiencies are usually obtained by increasing the
catalyst bed volume-to-vent stream flow  ratio.
     4.4.2.5  Applicability of Thsnual and Catalytic Oxidizers.
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
tied together and  routed to an incinerator.  For some of these
vents,  a primary control device such as a condenser is located
upstream.   Note  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 presumably high destruction efficiency obtained
for VOC's  and air toxics using these devices makes their
application attractive for very toxic substances.24
                               4-33

-------
4.5  SOURCE REDUCTION MEASURES
4.5.1  Vapor Containment
     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.  Essentially 100 percent
control of the vapors at the point source is achieved, and there.
do not appear to be many adverse effects from the standpoint of
safety or convenience.  However, the vessel which receives the
"vent back" must also be controlled.  Some facilities use vessels
with flexible volumes, such as balloons, or traditional gas
holders with self-adjusting diaphragms to contain vapors prior to
a control device.  Probably the biggest problem relative to batch
processing is that there are many different possibilities at any
given time for equipment configuration, and therefore a manifold-
type system for venting back vapors to the appropriate vessels
would have to be installed.
4,5.2  Limiting the Use of Inert Gas
     Obviously, many applications in batch processing require the
use of inert gas for blanketing and purging of equipment for
safety purposes.  Oftentimes, the distribution of the nitrogen is
affected through continuous purging of equipment.  While purging
achieves the inert atmosphere desired, it is also a source of
emissions because volatile compounds are stripped off and emitted
along the same discharge pathway as the nitrogen exhaust stream.
Limiting emissions from nitrogen purging is achieved by reducing
the amount of nitrogen that is purged.  An inert atmosphere can
also be created by establishing, through a series of pressure
transducers and distribution valves, a constant nitrogen,
positive pressure "blanket."  However, processing equipment that
does not have the possibility of remaining airtight cannot be
blanketed in this manner.   The older style basket centrifuges
requiring inertion during the separation of solid cake, for
example, cannot be blanket-inerted.  Therefore, it follows that
                              4-34

-------
limiting the distribution of nitrogen to constant positive
pressure blanketing operations may require not only capital
expenditures for the distribution system elements (i.e.,  the
pressure transducers and distribution valves), but perhaps the
replacement of some equipment.
     There are 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.
     Also, a recently developed technology for in-line stripping
could conceivably replace the use of large volumes of inert gas
used for sparging.  Control of emissions from sparging, as is
shown in Chapter 5, appears to be difficult because of the dilute
volumes of VOC in the exhaust sparge gas.  An in-line stripping
system that is installed directly into process piping creates  a
large number of very tiny nitrogen bubbles, which results  in
maximum gas-liquid interface.  One such system tested at a plant
reduced the amount of nitrogen used for sparging from 38,400 to
1,150 scfm and was considerably more efficient.25
     An added benefit of limiting the amount  of nitrogen that  is
used in inerting processing equipment is that the volumetric
flowrates of the exhausts will be diminished, and therefore VOC
concentrations in the exhausts will be  less dilute and may
therefore be more cost effective to control or recover using add-
on controls.
                               4-35

-------
4.5.3  Uae of Closed Processina Equipment
     The retrofitting or replacement of older equipment with new
airtight equipment is not only helpful to nitrogen blanketing
applications, but perhaps more importantly, to the processing of
material in entirely closed systems where the possibility of
creating emissions is eliminated altogether.  Batch processing
appears to be gravitating to processing equipment that is
versatile and therefore allows for numerous conventional unit
operations such as mixing, reaction, filtration, and drying, to
be conducted in the same vessel.  Transfer losses, which can be
very significant, are virtually eliminated, as are some cleaning
operations that would otherwise be required in between processing
runs.26
4.5.4  ^terial Substitution/Improved Separation Techniques
     One of the more significant areas of material substitution
in the batch manufacturing industry is the potential substitution
of organic solvents with aqueous solvents, aqueous solvents with
internally contained organic micelles, or supercritical fluids.
Still in developmental stages, the use of aqueous polymeric
systems having an internal micelle structure for hydrocarbons
would allow for reactions to occur wichin the polymer micelles.
Currently, the major problem with these polymers is that their
solubility in water is still too low to be of any practical
utility.27
     The possibility of using supercritical fluids  (SCF) in
extraction and separation applications is becoming more of a
reality.  Supercritical fluids have been shown to be of utility
in separation of organic-water solutions, petroleum fractions,
and activated carbon regeneration.  Additionally, a large body of
experimental data has been accumulated on the solubility and
extractability of natural products such as steroids, alkaloids,
anticancer agents, oils from seeds, and caffeine from coffee
beans in various supercritical fluids such as C02, ethane,
ethylene, and N20.  Currently, C02 is the most widely
investigated SCF in these applications.
                               4-36

-------
4.5.5  Improved Process Design

     The elimination of intermediate isolation steps, if

possible, can be a significant source of emissions reduction

because filtration and drying steps are eliminated.  It is also

likely that some equipment cleaning steps can be eliminated
without negative effects.

4.6  REFERENCES FOR CHAPTER 4

 1.  Shine, B., MRI.  Site Visit Report to UpJohn, Kalamazoo,
     Michigan.  March 30, 1990.

 2.  Letter from Hawkins, B., Nash Engineering Company, to
     B. Shine, MRI.  April 2, 1991.  Summary of operating data
     and budget costs for compressor systems.

 3.  Telecon.  B. Shine, MRI, to D. McKenzie, Union Carbide,
     Woodbine, Georgia.  Discussion of vapor recompression system
     in use at the Woodbine plant.  January 5, 1990.

 4.  Reference 1.

 5.  Li, Ramon, and M. Karell.  Technical, Economic, and
     Regulatory Evaluation of Tray Dryer Solvent Emission Control
     Alternatives. Environmental Progress.  2(2): 73-78.
     May 1990.

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

 7.  Reference 3.

 8.  Control of Volatile Organic Compound Emissions from
     Industrial Wastewater.  Draft CTG.  September, 1992.

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

10.  Reference 5.

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

12.  Carbon Adsorption for Control of VOC Emissions.  Radian
     Corporation.  June 6, 1988.

13.  Organic Chemical Manufacturing Volume 4:  Combustion Control
     Devices.  Report 4.  u. S. Environmental Protection Agency,
     OAQPS.  Publication No. EPA-450/3-80-026.  December 1980.

                               4-37

-------
14.  Klett, M. G., and J. B. Galeski.   (Lockheed Missiles and
     Space Company, inc.) Flare Systems Study.  Prepared for
     U. S. Environmental Protection Agency.  Huntsville, Alabama.
     Publication No. EPA-600/2-76-079.  March 1976.

15.  Evaluation of the Efficiency of Industrial Flares:
     Background-Experimental Design-Facility.  U. S.
     Environmental Protection Agency, OAQPS.  Research Triangle
     Park, North Carolina.  Publication No. EPA-600/2-83-070.
     August 1983.

16.  Reference 13.

17.  Reed, R. J.  North American Combustion Handbook.  North
     American Manufacturing Company.  Cleveland, Ohio.  1979.

18.  Memo and attachments from Farmer, J. R., EPA:ESD, to
     distribution.  August 22, 1980.  29 p.  Thermal incinerator
     performance for NSPS.

19.  Reference 12.

20.  Reference 13, Report 3.

21.  Control Techniques for Volatile Organic Emissions from
     Stationary Sources.  U. S.  Environmental Protection Agency.
     Office of Air and Waste Management.  Research Triangle Park,
     North Carolina. EPA Publication No. EPA-450/2-78-002.
     May 1978. p. 32.

22.  Kenson, R. E.  Control of Volatile Organic Emissions.
     MetPro Corp., Systems Division.  Bulletin 1015.
     Harleysville, Pennsylvania.

23,  Reference 13, Report 3.

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

25.  Processor Cuts Costs and Nitrogen Usage With In-Line
     Stripping System.  Chemical Processing.  March 1990.

26.  Pipeless Plants Boost Batch Processing.  Chemical
     Engineering.  June 1993.

27.  ERRC Update.  Progress in the Emission Reduction Research
     Center.  August 15, 1993.
                              4-38

-------
              5.0  ENERGY AND ENVIRONMENTAL IMPACTS

     The energy and environmental impacts associated with
applying various control options to VOC emissions from batch
processes are presented in this chapter.  The options are
described in detail in Chapter 6.
     The environmental impacts analysis considers the national
energy burden of operating the control devices used to meet
various options, as well as the national estimate of NO, produced
from the incineration of selected model process emission streams
and from the generation of electricity.  Solid wast* and
wastewater impacts were not evaluated because the effects
resulting from the operation of these control devices are
considered negligible.
5.1  ENERGY IMPACTS
     Table 5-1 presents the national estimate of energy usage for
each of the options described in Chapter 6.  The energy burden
was calculated by estimating the amount of fuel and electricity
required to operate the thermal incinerator and the electricity
requirement for the refrigerated condenser systems for the
applicable model streams.  Approximately 10 percent of the total
energy burden shown for each of the options in the table is
related to the condenser systems.  The remainder is associated
with the natural gas requirements of the thermal incinerator.
Energy usage for model streams and plants was extrapolated to a
nationwide estimate by considering the number of facilities in
the batch industries covered by this document.  Note that there
is no discernable difference in energy between the 98 percent and
95 percent options.  This effect occurs because the thermal
                               5-1

-------
                                    TABLE  5-1.   ENERGY  AND ENVIRONMENTAL  IMPACTS



Optio
n


1



2
3



Description of option


98% control of sggragntsil
process vents that am ant
exempt per njgjeasion lines 1,
4,7
90% control of process vents
95% control of process vents

Uncontroll
ed
emissions,
1.000
Mg/yr
210



210
210

Baseline
emissions,
1,000
Mg/yr


77



77
77
Nationwide
emission
reduction from
baseline,
1,000 Mg/yr


65



52
63


Energy
burden,
I013 Btu/yr


5



2
5

NO
emissions.
1.000
Mg/yr*-b


5



2
5
in

10
Emissions of NOX are from incinerator exhaust and from power plants used to generate the electrical power
fraction of the energy burden.
     emission factors:
   Incinerators: 200 ppm NOX in exhaust for streams containing nitrogen compounds, and 21.5 ppm NOX in all
         tnMH (based on lest data).

-------
incinerator is the significant energy using device and it was
assumed to control emission streams by 98 percent in all cases.
The energy difference in using refrigerated condensation systems
operating at 98 percent and 95 percent efficiency was
insignificant compared to the incinerator energy requirements.
5.2  AIR QUALITY IMPACTS
     The NOX emissions from thermal incinerators were estimated
assuming that the incinerator flue gas flow rate contained 50 ppm
NOX.  This value is in the range of concentration observed for
emission streams from incinerators (see footnote b, Table 5-1).
An alternative emission factor which could have been used is
0.1331b NOX per million Btu of natural gas.1
     The NOX emissions from energy generation were calculated
because condensers also use power.  Several assumptions were
required.  Since the majority of electrical power comes from coal
combustion, and the majority of coal used is bituminous, an
emission factor was developed to related electrical power, in
kilowatt-hours (kWh),  to NOX generation.  This factor was
developed using an AP-42 emission factor for NOX generation from
bituminous coal combustion.  This factor is 14 Ib N0x/ton coal2.
The average net heating value of bituminous coal is 14,000
Btu/lb.3  It was also assumed that coal-fired power plants are
about 35 percent efficient.  The emission factor is therefore 5 x
10'3lbs N0x/Jcwh,  or:

        kWh 3'412 Btu I	I   lbcoal   I    ton   . 14 Ib NOX
               JcWE    ' 0.35 ' 14,000 Btu1 2,000 Ib '   ton
Offsets for individual cases can be calculated using the emission
factors presented above.
5.3  WASTEWATER AND SOLID WASTE IMPACTS
     Wastewater and solid waste impacts are not expected to be
significant for this source category.  Thermal incineration for
halogenated compounds will yield acid gases which typically are
neutralized using caustic scrubbers.  The number of streams from
batch processing emissions that potentially would be halogenated

                               5-3

-------
and incinerated was not estimated, however.  For refrigeration
systems,  wastewater could be generated from humid waste gas
streams,  but this quantity also is not expected to be
significant.
5.4  REFERENCES FOR CHAPTER 5.0
1.   U. S. Environmental Protection Agency.  Compilation of Air
     Pollutant Emission Factors (AP-42).   Fourth Edition.
     September 1991.  p. 1.4-2.
2.   Reference 1.  September 1988.  p. 1.1-2.
3.   Air and Waste Management Association.  1992.  p. 209.  Air
     Pollution Engineering Manual.  1992.  p. 209.
                               5-4

-------
               6.0  DESCRIPTION OF CONTROL OPTIONS

6.1  TECHNICAL BASIS FOR CONTROL OPTIONS
6.1.1  Approach
     The methodology used in developing control options is based
on an evaluation of the technical feasibility and costs of
controlling any vent stream that could be emitted to the
atmosphere from a batch process.  In order to be able to apply
the options and to defend the rationale that was used to develop
the options for a wide variety of stream characteristics, factors
such as cost effectiveness and control device applicability were
examined for all potential variations in duration of emission
events and emission stream characteristics of flow rate and VOC
concentration.  This section presents a discussion of batch
processing emissions and describes the methodology for developing
the options.
     6.1.1.1  Batch Processing Emission Stream Characteristics.
In general, there are two qualities that differentiate batch
processing emissions from those of sources operating
continuously.  First, batch emission stream characteristics
(e.g., flow rate, concentration, temperature, etc.) are never
constant.  Second, the emissions are released on an intermittent
basis.  To illustrate these ideas, consider the batch process
shown in Figure 6-1.  Eaissions of voc's will occur from this
process from start to finish in the order that the bulk flow of
material and energy occurs.  For example, the process begins with
the charging of a VOC material from storage into the weigh tanks.
A displacement of air from the weigh tanks occurs at this point
as a result of being pushed out by the incoming volume of
material.  Through vaporization of the VOC liquid across the
liquid-air interface, this air contains some amount of VOC and
thus constitutes an emission event.  The event is short-lived,
however, lasting only the time of the charge; the concentration
of voc's in the displaced flow rate will increase to a point
close to saturation by the time the last of the displaced air

                               6-1

-------
       WEIGH
       TANK
WEIGH
TANK
             MIX
             TANK
              I
            REACTOR
                             -SOLVENT
          CRYSTALUZER
             SLURRY
             TANK
CENTRIFUGE
   DIST.
 CENTRIFUGE
 SOLVENT
 RECOVERY
                      VACUUM

                       DRYER
  Figure 6-1.  Model Batch Process

               6-2

-------
leaves the charge tank.  As the material moves from the weigh
tanks to the reactor, another displacement occurs that
contributes to emissions of VOC's.  This emission event is very
similar to the event created by filling the weigh tank.
Similarly, as the material flows through the process, each piece
of equipment becomes a contributor to VOC emissions through a
distinct series of finite emission events.  In some equipment,
such as the reactor, more than one type of emission event occurs;
for example, an event results from charging, heatup, and kettle
purging from this piece of equipment.
     In the example process shown in Figure 6-1, consider the
movement of a highly volatile solvent such as diethyl ether
through the process; the emission events that occur as a result
of air displacement have concentrations of VOC's in excess of
50"percent by volume.  For the reactor purging event, however,
the concentration of VOC drops as the emission stream is diluted
by high flows of inert gas into and out of the kettle.  The
largest source of uncontrolled emissions in this process is the
vacuum dryer, whose emission stream is characterized by an
decrease in VOC concentration and a somewhat steady flow rate
over the course of its drying cycle.
     Figures 6-2 and 6-3 prsssr.t the fluctuations in flow rate
and concentration, respectively, that will occur during the batch
cycle.  The result of combining the flow rate and concentration
profiles is presented in Figure 6-4, the emissions profile.  In
order to give a more vivid illustration of how flow rate,
concentration, and emissions vary in such a batch process,
Figures 6-2, 6-3, and 6-4 have all been placed on the same page,
resulting in Figure 6-5. Note that the time scale for all these
figures is the same.  Close inspection of Figure 6-5 reveals that
the concentration and flow rate characteristic of the process
vents vary independently from each other; although there appears
to be a slight trend for the concentration to change inversely
with flow rate.
     The reason for presenting these profiles is to introduce the
idea that the variable emission stream characteristics of batch
                               6-3

-------
 FLOWRATE PROFILE OF A BATCH PROCESS
   154



   132



 E 110
>-4
 (J

-2-  88
LU
h-
<  66
a:


g  44

u_

    22
      Reoclor
*/
 V N
       Ohr
                        Vacuum dryer
      2.2hr


TIME (24 HOURS)
                                      Distillation    J?
                                                w
                                     22.2 hr    24 hr
        Figure 6-2.  Plowrate profile

-------
o\
I
Ul
CONCENTRATION PRQFII
90
H- B1
ui 72
O
o: 63
bJ
3- 54
, . i
uJ
2 45
13
-J 36
O
> 27
O
0 18
9
o
4^r _A
-Qr ^i*
fc 4%
jtfy^^
, 1 '
teocloi
r-*-
111



































rl
l f




















il
X^ 1

v0 1
1 1

n
Qlw 2.2
TIME (24
                                                                  Vacuum


                                                                    dryer
                                                                                   Distillqtion
                                                                                  22.2 hr
                                                                                              Solvent
                                                                                            Recovery
24 hr
                                        Figur*  6-3.  Coacatrtlon profile

-------
at
                          EMISSION PROFILE OF A BATCH PROCESS
                    LJ
                        Ohr
               2.2 hr

         TIME (24 hours)
                                                                     Solvent
                                                                     recovery
                                                                     24hr
Figure -4
                                                 profile

-------
process vents affect the feasibility of using control devices
currently available in industry.  These attributes also
potentially create confusion on the part of plant operators and
regulators concerning how to describe the emission
characteristics (e.g., instantaneous maximums, 8-hr averages,
24-hr averages, or batch cycle averages).  In light of these
considerations, it follows that the methodology for development
of options would address questions of control device
applicability, as well as provide meaningful criteria for
determining which streams should be recommended for control.  The
methodology development is described below.
     6.1.1.2  Control Devices Examined.  The cost and feasibility
of controlling typical batch emission streams was examined by
applying typical add-on control devices that are found in
industry.  All currently available types of control device* were-
examined, including thermal destruction (thermal, catalytic
oxidizers and flares), refrigeration (condensers), gas absorption
(water scrubbers), and carbon adsorption system*.  The final cost
analysis, however, was done based on thermal incineration and
condenser systems.  These devices were used exclusively in
examining cost because, among other factors, they can be applied
to a universe of compounds.  In many cases, other control devices
might prove to be more cost effective, but generally, they can
not be used universally and therefore the cost impacts of the
option would not be supported for streams containing wide ranges
of compounds.  A case in point is the use of a water scrubber to
control steams containing water-soluble VOC's.  The cost and cost
effectiveness of this device may be considerably better than that
of an incinerator or refrigeration system affording the same
level of control, but the costs of the option could not be based
on this device because it would only be available for a segment
of potential emission streams.  Likewise, carbon adsorption,
which is less costly than thermal incineration and condensation
in many cases, will not control some types of VOC's and therefore
it was also ruled out as a test case for the feasibility
analysis.
                               6-7

-------
     Although thermal incineration and condensation systems are
limited in the types of streams that each can feasibly control,
these limitations are based more on concentration and less in
terms of compound specificity.  Additionally, the two devices
complement each other in being able to handle ranges of emission
stream parameters.  For example, the condenser option ideally
would be used to control richer streams (>10,000 ppm) while the
thermal incinerator could handle streams that were more dilute
(<10,000 ppm) and largely infeasible to consider for control with
a condenser system.  Minor limitations to compound specificity
associated with burning halogenated compounds were considered by
adding the cost of caustic scrubbing and lowering waste gas heat
contents (it was later concluded that this incremental cost was
within the margin of error of the study estimate), while compound
specificity did not appear to be a problem with refrigeration
systems.
     Note that although the thermal incinerator and condenser
were used to establish control cost effectiveness curves, tHe
options are not equipment-based, only performance-based.
Therefore, an emission limit would specify a control level
(e.g., 98 percent, 95 percent, 90 percent) and not a particular
control device.  Therefore, an operator could elect to use a
water scrubber to meet control requirements in cases where a
water scrubber would achieve the required level of control.
     6.1.1.3  Considerations.  The first issue considered in
developing options was the sensitivity of the costs of each
control device to the intermittency of emission events.  The
primary indicator of cost is cost effectiveness in units of
dollars per megagram VOC controlled ($/Mg).  This cost
effectiveness value is obtained by dividing the annualized cost
of the control device ($/yr) by the annual emissions reduction
(Mg/yr).  The cost effectiveness decreases (values become higher)
as the amount of time that the emission stream is released to the
atmosphere (on-stream duration) is reduced.  This trend is
readily obvious from Figure 6-6, which is a graphical
presentation of cost effectiveness versus vent stream flow rate
                               6-8

-------
                      Cost to Control by Thermal Incineration
                          For VOC Concentration of 1 0,000 ppmv
                        207.107.
VO
            -^35000

             300CO
                                                        stream time
                               100         1000        10000        1000CO

                         Maximum Instantaneous Flowrate (scfm)
              Figure 6-6. Dependence of control device coat on emiavion intermittency

-------
at a set annual emission rate and set VOC concentration for
different on-stream durations.  Notice that the on-stream
duration is directly related to flow rate when the annual
emission rate is constant.  Figure 6-6 is based upon a thermal
oxidizer with an assumed control efficiency of 98 percent.  The
use of a thermal oxidizer for the analysis presented in
Figure 6-6 is meant only to illustrate the sensitivity of cost
effectiveness with on-stream duration (intermittency).  Other
devices, such as condensers and carbon absorbers, also exhibit
similar sensitivity with varying on-stream durations.
     Because each control device is often sized according to the
maximum possible flow rate and VOC concentration, devices used in
batch process emission control are usually oversized for the
majority of the time that they are in service.  Also, for devices
such' as incinerators and condensers, the annualized cost of
maintaining proper operating conditions (e.g., maintaining
incineration and condenser temperatures) when there is no
material being vented to the devices drives up the cost of
control.  Consequently, the cost effectiveness of controlling
batch emissions is generally lower (values are higher) than the
cost of controlling continuous emissions for similar stream
characteristics.
     The second consideration that was made in developing the
options was to limit the number of parameters necessary to
determine which streams should be required to be controlled.
Because there is inherent variation in the characteristics of
flow rate and VOC concentration during batch emission events,
eliminating as many parameters as possible (especially those that
vary) will minimize confusion in compliance determinations.  For
example, an owner or operator could choose to report an average
concentration of a VOC emission stream, rather than a 'peak'
concentration in order to fall below a concentration cutoff.  By
eliminating concentration as a parameter used to determine
applicability, this problem would be circumvented.
     6.1.1.4  Approach.  The approach chosen uses uncontrolled
annual VOC emissions (expressed as Ib/yr) and average flow rate
                               6-10

-------
(scfm) to define which streams should be controlled and the level
of control required.  This approach considers the impact of
varying VOC concentrations and frequency of emission events, but
does not require their use as parameters to determine
applicability.  Generally, the uncontrolled annual emission total
of VOC's from a particular source is more readily available from
material balance and other calculational approaches than is a
detailed minute-by-minute concentration and flow profile, as is
an average flow rate.
6.1.2  Control Options Methodology
     6.1.2.1  Cost-Effectiveness Curves.  The methodology that
was used to develop the options utilizes the parameters of annual
emissions and average flow rate to identify which streams are
reasonable to control from a cost and technical feasibility
standpoint.  Note that the volatility of components of concern is
a sensitivity which requires consideration for design and cost of
the condenser systems.  Hence, three regions of volatilities were
considered in the analysis.  Low volatility materials are defined
for this analysis as those which have a vapor pressure less than
or equal to 75 mm Hg at 20"C; moderate volatility materials have
a vapor pressure greater than 75 and less than or equal to
150 mm Hg at 20*C and high volatility materials have a vapor
pressure greater than 150 mm Hg at 20*C.  In determining
applicability of the requirements to multicomponent VOC streams,
a weighted average of the VOC volatilities should be used to
determine the appropriate volatility range.  This weighted
average volatility is defined in Chapter 7 under *Definitions',
and is ultimately used to determine which equation to use.
     Figures P-l through F-54 of Appendix F show cost
effectiveness versus flow rate for annual emissions of 30,000,
50,000, 75,000, 100,000, 125,000, and 150,000 lb/yr for various
control levels (i.e., 90, 95, or 98 percent) and volatilities.
Each graph represents the full range of concentrations of VOC's
that might be expected in any given emission stream (from 100 ppm
to 100,000 ppm [the upper concentration examined for toluene, a
low volatility material, is 37,000 ppm]); for simplicity, we can
                               6-11

-------
call this the "envelope."  Note that the 100 ppm line is not
graphed in the curves presented in Appendix F.  This line
typically falls between the 1,000 and 10,000 ppm curves, but ends
as the envelope narrows.  The width of each envelope is an
indication of how much the cost effectiveness varies with
concentration.
     Figure 6-7 is an example of the curves contained in
Appendix F.  The figure shows the cost effectiveness of
controlling any stream having  single component or group of
components with a total vapor pressure in the moderate volatility
range (from 75 to 150 mm Hg at 20*C).  There are four curves on
the graph:  Two of the curves show the cost effectiveness versus
flow rate for control by thermal incineration (abbreviated as
"throx") at concentrations of 1,000 ppmv and 8,750 ppmv.  The
other two curves are for condenser control of streams with
concentrations of 10,000 ppmv and 100,000 ppmv.  Points along the
curves were established by inputting a constant mass emission
total and a constant concentration into the condenser and thermal
incinerator spreadsheets and plotting the resulting flow rate and
cost effectiveness values corresponding to various durations.
     Since the annual emissions are constant at 50,000 Ib/yr, the
flow rate (x-axis) values at any point along the curves are an
indicator of the duration of the emission events.  For example,
the left-hand endpoints of the curves represent streams that are
continuous (i.e. in order to emit 50,000 Ib/yr from an emission
point at a concentration or iuu,000 ppmv, the minimum flow rate
for the stream, if it is venting continuously, is around 5 scfm).
As the curves move from left to right (increasing flow rates),
the duration of the emission events decrease, so that points
along the right hand edges of the curves represent short duration
events in which large amounts of VOC's are released at high flow
rates.  These "bursts" of emissions are not surprisingly more
expensive to control because they must be sized for large flows,
yet they will only control emissions for short durations.  For
some concentrations, points on the upper-right corner of the
                               6-12

-------
I
M
U)
                             Annual Mass Emission Total=50,000lb/yr
                                       .Vol. (Benzene); Cond.Ctrl.ff.=90%
                 20000
                                                           SHORT OgRATION EVENTS
                                  100% DURATION. 8760 HOURS
                                      10
    100
Flowrate (scfm)
1000
10000
                        pTlOOOppmv  	T8760ppmv   : ClOOOOppmv 	ClOOOOOppmv
                          Figure 6-7.  Annual maae emission total - 50,000 Ib/yr
                           Hod. Vol. (Beniene); Good. Crtl. Bff,  90 percent.

-------
graph may occur less than 10 hours per year; these streams
resemble emergency releases.
     Based on the above discussion, it can be seen from
Figure 6-7 that for a process vent emitting 50,000 Ib/yr of VOC,
the cost effectiveness of control is a maximum of $5,000/Mg for a
maximum flow rate of about 400 scfm or less, regardless of
concentration, regardless of duration.  At higher flow rates, the
curves begin to rise sharply and the cost effectiveness values
become higher (indicating that they are less feasible to control
from a cost standpoint).
     This discussion, then, forms the basis for setting up
option requirements based on annual emissions and flow rate.  By
establishing a number of curves for different annual emission
totals (i.e., 30,000, 50,000, 75,000, 100,000, 125,000 and
150,000 Ib/yr), values of flow rate were obtained for an optimum
cost effectiveness range, considering impacts.  These annual
emissions, and corresponding flow rates were used as data points
(x was annual emissions and y was flow) for simple regression
analysis to define the line that will represent optional cutoffs
for applicability that could be included in standards.
     Note also that the subheading for Figure 6-7 states that
condenser control efficiency is 90 percent.  Since both the
thermal incinerator and the condenser cost algorithms were used
to construct each graph contained in Appendix F, there were
varying levels of control efficiency that could be achieved by
the condenser; the thermal  incinerator was assumed to be
effective to 98 percent all the time.  Therefore, for curves
containing condenser control efficiencies less than 98 percent
(i.e. 90, 95 percent), the  overall control level is limited by
the condenser efficiency.
6.2  PRESENTATION OF FLOW RATE REQUIREMENTS
     Table 6-1 presents the regression line and data points
obtained from Appendix F graphs for various control levels.  Note
that the graphs presented  in Appendix F resemble the graph shown
in Figure 6-7.  However, the labor and maintenance costs for
graphs shown  in Appendix F  are for 1 shift per day only, as
                               6-14

-------
                           TABLE  6-1.   SUM4ARY  OP  CONTROL OPTION  REGRESSION LINE DATA
Volatility
Low
Moderate
High
Control
level
%
98
95
90
98
95
90
98
95
90
Flow rate data points (acfa) for annual man emissions, Ib/yi*
30,000
757
866
449
251
515
334
208
215
50,000
1,787
2.452
1,634
612
1,034
1,082
517
544
369
75,000
3.076
4,043
3,324
1,063
1,682
1,833
XM
955
704
100,000
4.363
5.600
5.113
1,514
2,330
2,776
1.290
1,366
1,039
Jlhii field lifts y-cooidinatea for the corresponding x-coordinales of 75,00
'Annual HUM emissions below this value no control nouind. recanttess o
125,000
5,652
7.158
6,954
1,965
2,978
3,319
1,677
1,777
1,374
150,000
6,939
8,717
8,826
2,416
3,627
4,062
2,063
2,188
1,709
No.
1
2
3
4
5
6
7
8
9
0, 100,000, 125,000, and 150,01
f flow rate.
Regression line
PR - (0.052)AE-789
PR = (0.065)AE-895
PR - (0.07)AE-1,821
PR = (0.018)AE-290
PR = (0.026)AE-263
PR = (0.031)AE-494
PR - (0 015)AE-236
PR - (0.016)AE-278
PR - (0.013)AE-301
No.
flow
value
Ib/yi*
15,173
13,769
26,014
16,111
10,115
15,935
17,067
17,375
23,153
DO Ib/yr.
I
in
         The regression line equations presented hern can be incorporatedinto regulations as 'cut-offs." As cutoffs, they would be used to determine
         what streams should be contoUed, given an annual mass emission (AB) total and an average flow rate(PR).  If the flow rate calculated by the
         cutoff" line equation (when annual mass emission is y****} is higher than the avenge flow rale of the stream, then control would be
         required to the level specified (98,95, or 90 percent).

-------
opposed to 3 shifts per day labor and maintenance costs assumed
in the construction of Figure 6-7.  By using the line equations
presented in Table 6-1, average flow rates can be established
using the annual emission total.  Comparison of this "cutoff
with the actual flow rate of the emission source would determine
whether control is required.
     The options that were further evaluated for nationwide
impacts based on the curves in Appendix F are presented in
Table 6-1.  The regression lines can be used to determine what
streams should be controlled, given an annual mass emission total
and an average flow rate.  If the flow rate calculated by the
"cutoff line equation (when annual mass emission is inputted) is
higher than the average flow rate of the stream, then control
would be required to the level specified (98, 95, or 90 percent).
The'assumptions used to arrive at the baseline and uncontrolled
emission numbers, and the industries affected as shown in
Table 6-2 are discussed in the next section.
6.2.1  Discussion of Additional Issues
     6.2.1.1  Single st^q versus aggregation.  The annual
emission total and flow rate cutoffs can be applied either to
single streams or to emission streams resulting from aggregated
sources.  Costs for manifolding sources have been considered in
the design and cost calculations.  For example, total purchased
equipment costs for the condenser systems were multiplied by an
additional 25 percent to account for manifolding whereas a
300-foot collection main with 10 takeoffs and an auxiliary
collection fan was costed out in the incinerator cost
calculations.
     An additional analysis was undertaken to identify whether
there is a level at which the incremental cost of manifolding
individual emission sources is unreasonable compared with the
emission reduction achieved.  Simply stated,  what level of
emissions would rule out including a source into an aggregate
pool of sources, based on a measure of control achieved over the
cost of manifolding the small source to the central process
control device.  This level is identified as the "deminimis"
                               6-16

-------
level for purpose of the applicability analysis.  A deriation of
this deminimis level is presented in Appendix B.
     6.2.1.2  Ifaloqenated Compounds.  The cost-effectiveness
curves shown in Figures P-l through F-54 are for thermal
incinerators and condensers.  The costs are based on using an
incinerator operating at 1600*F, with fractional heat recovery,
and not equipped with an emission control devic*.  For
halogenated compounds, such an incinerator might not achieve a
control level of 98 percent, and additionally, acid gas would be
emitted from the combustion process.  Consequently, the cost
analysis was repeated using costs based on an incinerator
designed to control halogenated compounds.  Such an incinerator
would maintain combustion temperatures at 2000*F, have no
fractional heat recovery, and would be equipped with a caustic
scrubber to control acid gases.
     From the curves, the increase in cost effectiveness value*
associated with using a thermal incinerator equipped to control '
halogenated compounds appears to be approximately 1 to $5K/Mg
more costly than using the nonhalogenated compound incinerator.
6.3  IMPACTS OF APPLYING OPTIONS
     A model plant approach was used to examine the impacts of
applying the options to industry on a nationwide basis.  For the
industries assumed to be covered, emissions streams from small,
medium, and large model plants were evaluated to determine the
level of control required based on the annual emissions and flow
rates specified by various control option regression lines.
Emission reductions over baseline control were evaluated for each
model plant and were extrapolated to a nationwide basis using
Census of Manufacturers Industry Profile data.  The nationwide
impacts development is outlined below.
6*3.1  Industries Covered
     While the information contained in this document is
generally applicable for batch processes in all or most
industries, the impacts presented are only for selected
industries.  These industries and their corresponding Standard
Industrial Classification (SIC) codes are presented in Table 6-2.
                                6-17

-------
                              TABLE 6-2.   PERCENTAGE OF EMISSIONS  FROM  BATCH PROCESSES
\
SIC code
2821
2834
2861
2865
2869
7.879
SIC description
PUitics materials and resins
Pharmaceutical preparations
Gim and wood chemicals
Cyclic crudes and intermediates
Industrial organic chemicils
Afric jllurrl chenticals
<7800(hn/yr)
14.396
8,432
2,287
223
17.060
92
>0 (hrs/yr)
124,547
I5.4S9
20.415
8.365
173.167
3.912
% Batch
12%
55%
11%
3%
10%
2%
             NOTE:  Emissions data was obtained from AIRS facility wire* data base search
      I
     >-^
     CO

-------
Note that facilities that make up the industries listed also
potentially use continuous processes; in order to assess what
proportion of eaissions generated in these industries is from
batch processes, the Aerometric Information Retrieval System
(AIRS) Facility Subsystem data base was accessed.  For each
applicable SIC code, emissions from process vents were totaled.
Then, a subset of these data, those emissions that were reported
to have durations of less than 7,800 hours per year, were totaled
and divided by the total vent emissions for those SIC codes.  The
resulting fraction was taken to be the percentage of total
emissions for each SIC code that would result from batch
processing.  From the table, the percentages of emissions
considered "batch" may appear lower than expected; one of the
limitations of using the AIRS database is that only sources with
greater than 100 tons per year are listed.  Because many batch
industries are for low-volume chemicals, basing these percentages
on AIRS data probably biases the percentages low.
6.3.2  Model Processes
     Figures E-l, E-2, E-3, and E-4 of Appendix E present model
batch processes that are typically found in batch industries.
These model processes were recommended by an industry trade
association for use in evaluating impacts.1  Tables E-l,  E-2,
E-3, and E-4 of Appendix E are summaries of emission streams
characteristics resulting from the unit operations shown in the
model batch processes for low, moderate, and high volatility
materials.  Emission stream characteristics were calculated based
on data, where possible, and from the vapor-liquid equilibrium
assumptions described Chapter 3.  Appendix E also contains all
the calculations and assumptions used to develop model emission
streams, from which only a few were selected to make up the model
batch processes.  The emission rates for all unit operations
within the model processes were tabulated for each volatility.
The small, medium, and large model plants are based on multiples
of these model process emission totals.  Three model processes
were assumed to represent the small plant, 10 model processes
were assumed to represent the medium-sized plant, and 30 model
                               6-19

-------
processes were assumed to represent the large plant.  A list of
assumptions made in developing each of the model plants is
presented in Tables E-5 through E-8.
6.3.3  Baseline As. gipptions/Extrapolation^
     The baseline used in estimating nationwide impacts for
process vents corresponds to the level of control achieved by the
Pharmaceutical CTG.  Emissions from the number of batch
facilities in SIC Code 2834 (Pharmaceutical Preparations) were
subject to this control level.  The Pharmaceutical CTG contains
condenser exit temperature requirements for five classes of
volatility, and requires 90 percent control on dryers emitting
more than 330 Ib/d.  The facilities in the remaining five SIC
codes, 2821, 2861, 2865, 2869, and 2879, were assumed to be
subject to no VOC emission controls for process vents.
     Essentially, two extrapolations were done in order to arrive
at nationwide impacts.  The first was to evaluate the control
option impacts from the model batch processes and extrapolate to
the small, medium, and large model plants.  The second was to
extrapolate the impacts from the small, medium, and large model
plants to the total number of facilities conducting batch
processes nationwide.  These extrapolations are discussed in more
detail below.
     6.3.3.1  Model Plants.  As mentioned before, the small model
plant was assumed to contain three model batch processes; the
medium model plant was assumed to contain 10 model batch
processes, and the large model plant was assumed to contain
30 model batch processes.  These values fall within ranges
recommended by an industry trade association.*  Tables 1-9
through E-12 of Appendix E present model plant emission totals
for the small, medium, and large model plants assuming (1) no
control at all, and (2) current pharmaceutical control for low,
moderate, and high volatility materials.
     Because the model processes are grouped into model plants
that only contain multiples of single processes, the model plants
are not entirely reflective of the batch industries.  It is
expected, for instance, that actual plants will have combinations

                                6-20

-------
 of different processes.   However,  because the estimation of
 nationwide impacts is based on an  evaluation of the flow rate and
 annual  emission total of individual processes exclusively,  the
 groupings are used exclusively to  extrapolate nationwide numbers.
 Therefore, these "unreflective" groupings do not affect the
 correctness of the impact.
      6.3.3.2  Nationwide Facilities.  Table 6-3 presents data
 taken from industry profiles contained in the Census of
 Manufacturers and from EPA  data on county ozone nonattainment
 status.   This information was used to extrapolate the model plant
 emission totals (under no control, current pharmaceutical
 control, and for the various options) to a nationwide basis.
      Emissions from the  batch industries represented by the SIC
 codes in Table 6-3 were  estimated  by assuming that model
 processes 1 through 3 (solvent reaction with atmospheric dryer
 [model  process 1], solvent  reaction with vacuum dryer [model
 process  2], and liquid reaction [model process 3])  were evenly
 used  among the industries covered.  The impacts assume that low,
 moderate,  and high volatility materials are evenly distributed
 among the model processes (i.e., 1/3  of the processes use low
 volatility materials,  1/3 use moderate volatility materials,  and
 1/3 use  high volatility  material?},   Nationwide emissions were
 estimated by multiplying the census size groupings by employee
 number  (i.e.,  small plant0 to 19 employees)  by model emission
 totals to estimate small, medium,  and large plant emissions.
 Only  the total number of facilities located in nonattainment
 areas (excluding marginally nonattainment)  were considered.  The
 formulator model process (Figure E-8)  was not  included in the
 nationwide impacts,  but  is  found in some SOCHI batch operations.
 6.4  SUMMARY OF OPTIONS  AND IMPACTS
      Table 6-4 presents  the overall reduction  in VOC that can be
 expected from various options and  the national costs associated
 with  applying the options on a nationwide basis.   Options are for
.aggregated sources controlled to SO percent,  95 percent,  and
 90 percent overall,  respectively.
                                6-21

-------
REFERENCES

1.  Letter from Synthetic Organic Chemical Manufacturers
    Association (SOCMA) to Randy Mcdonald, EPA/ESD/CPB, providing
    comments and recommendations on the Batch CTG.  Dated
    April 19, 1991.

2.  Reference 1.
                                6-22

-------
            7.0 FACTORS TO CONSIDER IN IMPLEMENTATION
    OF A RULE BASED ON THE OPTIONS PRESENTED IN THIS DOCUMENT

     This chapter presents information for State and local air
quality management agencies to use in developing enforceable
regulations to limit emissions of VOC's from batch processing
operations.  The information presented here assumes that the
Agency adopts one of the options presented in Chapter 6.  The
information is the same regardless of the option selected.
      A unique approach has been developed to determine the
applicability and optimum level of control required for batch
emission sources.  Additionally, a model rule with blanks to
allow for choices of options, is included in Appendix G.  This
chapter is divided into the following sections:  (1) Definitions
and Applicability, (2) Format of the Standards, (3) Testing,
(4) Monitoring Requirements, and (5) Reporting/Recordkeeping
Requirements.
                               7-1

-------
7.1  DEFINITIONS AND APPLICABILITY
7.1.1  Definitions
     The agency responsible for developing a standard must define
the terms that appear in the language  for  the standard.   The
source category of batch processes,  for  example,  requires a
definition of the term  "batch" as it is  used to describe the mode
of operation of equipment and processes.   Another term that will
likely require defining is "vent".   The  feasibility analysis that
has been described in Chapter 6 applies  to any type of gaseous
emission stream (continuous or batch)  containing VOC's,  as long
as the flowrate and annual mass emission total requirements are
met.  Finally, the terms "flowrate"  and  "annual mass emissions"
also should be defined  clearly.  Provided  below is a listing of
definitions for terms as they are used in  this CTG and which are
recommended for State-adopted rules.
     Aggregated means the summation  of all process vents
containing VOC's within a process.
     Annual mass emissions total means the sum of all emissions,
evaluated before control, from a vent.   Annual mass emissions may
be calculated from an individual process vent or groups of
process vents by using  emission estimation equations contained in
Chapter 3 of the Batch  CTG and then  multiplying by the expected
duration and frequency  of the emission or  groups of emissions
over the course of a year.  For processes  that have been
permitted, the annual mass emissions total should be based on the
permitted levels, whether they correspond  to the maximum design
production potential or to the actual  annual production estimate.
     Average flowrate is defined as  the  flowrate averaged over
the amount of time that VOC's are emitted  during an emission
event.  For the evaluation of average  flowrate from an aggregate
of sources, the average flowrate is  the  weighted average of the
average flowrates of the emission events and their annual venting
time, or:
               T* (Average Plovrate per emiMion event) (annual duration of emiion event)
       Flowrate  **     	=	-	A
                            (annual duiacion of emiMion events)
                                7-2

-------
     Batch refers to a discontinuous process involving the bulk
movement of material through sequential manufacturing steps.
Mass, temperature, concentration, and other properties of a
system vary with time.  Batch processes are typically
characterized as "non-steady-state."
     Batch cycle refers to a manufacturing event of an
intermediate or product from start to finish in a batch process.
     Batch process train means an equipment train that is used to
produce a product or intermediate.  A typical equipment train
consists of equipment used for the synthesis, mixing, and
purification of a material.
     Control devices are air pollution abatement devices, not
devices such as condensers operating under reflux conditions,
which are required for processing.
   *  Emissions before control means the emissions total prior to
the application of a control device, or if no control device is
used, the emission total.  No credit for discharge of VOC's into
wastewater should be considered when the wastewater is further
handled or processed with the potential for VOC's to be emitted
to the atmosphere.
     Emission events can be defined as discrete venting episodes
that may be associated with a single unit of operation.  For
example, a displacement of vapor resulting from the charging of a
vessel with VOC will result in a discrete emission event that
will last through the duration of the charge and will have an
average flowrate equal to the rate of the charge.  If the vessel
is then heated, there will also be another discrete emission
event resulting tram the expulsion of expanded vessel vapor
space.  Both emission events may occur in the same vessel or unit
operation.
     Processesr for the purpose of determining control
applicability, are defined as any equipment within a contiguous
area that are connected together during the course of a year
where connected is defined as a link between equipment, whether
it is physical, such as a pipe, or whether it is next in a series
                               7-3

-------
of steps from which material is transferred  froa  one unit
operation to another.
     ^eai -continuous operations are conducted on  a  steady-  state
mode but only for finite durations during the course of  a year.
For example, a steady-state distillation operation  that  functions
for 1 month would be considered semi-continuous.
     pnit operation^ are defined as those discrete  processing
steps that occur within distinct equipment that are used to
prepare reactants, facilitate reactions, separate and purify
products, and recycle materials.
     Vent means a point of emission from a unit operation.
Typical process vents from batch processes include  condenser
vents, vacuum pumps, steam ejectors, and atmospheric vents  from
reactors and other process vessels.  Vents also include  relief
valve discharges.  Equipment exhaust systems that discharge from
unit operations also would be considered process  vents.
     Volatility is defined by the following!  low volatility
materials are defined for this analysis as those  which have a
vapor pressure less than or equal to 75 mmHg at 20 *C,  moderate
volatility materials have a vapor pressure greater  than  75  and
less than or equal to 150 mmHg at 20 "C; and high  volatility
materials have a vapor pressure greater than 150  mmHg at 20* C.
To evaluate VOC volatility for single unit operations that
service numerous VOCs or for processes handling multiple VOCs,
the weighted average volatility can be calculated simply from
knowing the total amount of each VOC used in a year,  and the
individual component vapor pressure, as shown in  the following
equation:
_ (MM of VOC cotnoafnt i)
(molaculmz velyht of VQG caapanant
                                                  1
                                                i) ]
7-1.2  Applicability
     The analysis on which options are based was performed over a
number of industries thought to manufacture a significant
                               7-4

-------
percentage of total production on a batch basis.  These
industries, identified by 4-digit SIC codes, are presented in
Chapter 6.  They are:  plastic materials and resins  (SIC 2821),
pharmaceutical preparations  (SIC 2834), medical chemicals and
botanical products (SIC 2833), gum and wood chemicals  (SIC 2861),
cyclic cruds and intermediates (SIC 2865), industrial  organic
chemicals (SIC 2869), and agricultural chemicals (SIC  2879).
Although the impacts in this document were evaluated based on a
scope limited to these industries, any batch emission  point of
VOC's from presumably any industry could be subjected  to these
requirements.  Note that there are two CTG's, the Air  Oxidation
CTG and the Reactor Processes and Distillation Operations CTG,
that cover synthetic organic chemical emissions from continuous
processes.  The CTG's also exempt batch or semi continuous
processes.  The information in this document applies to the
processes that are exempted because they are not continuous.
This includes semi continuous processes.
      The control option requirements presented in Chapter 6
apply to (l) individual batch VOC process vents to which the
annual mass emissions and average flowrate cutoffs are applied
directly, and (2) aggregated VOC process vents for which a
singular annual mass emission total and average flowrate cutoff
value is calculated and for which the option is applied across
the aggregate of sources.  The applicability is discussed in more
detail below.
     Sources that will be required to be controlled by a control
device will have an average flowrate that is below the flowrate
specified by the cutoff equation (when the source's annual
emission total is input).  The applicability criteria  is
implemented on a two-tier basis.   First, single pieces of batch
equipment corresponding to distinct unit operations shall be
evaluated over the course of an entire year,  regardless of what
materials are handled or what products are manufactured in them,
and second,  equipment shall be evaluated as an aggregate if it
can be linked together based on the definition of a process.
                               7-5

-------
     To determine applicability of a cutoff option in the
aggregation scenario, all the VOC emissions from a single process
would be summed to obtain the yearly emission total, and the
weighted average flow rates from each process vent in the
aggregation would be used as the average flow rate.
     All unit operations in the process, as defined for the
purpose of determining cutoff applicability would be ranked, in
ascending order, according to their ratio of annual emission
divided by average flow rate.  Sources with the smallest ratio
would be listed first.  This list of sources constitutes the
"pool" of sources within a process.  The annual emission total
and average flowrate of the pool of sources would then be
compared against the cutoff equations to determine whether
control of the pool is required.  If control were not required
after the initial ranking, unit operations having the lowest
annual emissions/average flowrate would then be eliminated one by
one, and the characteristics of annual emissions and average
flowrate for the pool of equipment would have to be evaluated
with each successive elimination of a source from the pool.
Control of the unit operations remaining in the pool to the
specified level would be required once the aggregated
characteristics of annual emissions and average flowrates met the
specified cutoffs.
     By aggregating unit operations, the annual emission totals
are more easily achieved at better cost effectiveness values.
However, a unit operation may have a high emissions to flowrate
ratio, albeit low actual emissions and the cost effectiveness of
controlling such a unit operation may not be reasonable.  Such
cases have been evaluated using the cost analysis of ductwork.
Essentially, the costs of ducting can be shown to be dependent on
flowrate of the emission stream and required length of ducting.
The incremental cost analysis for manifolding single unit
operations to a control device are contained in Appendix B.
7.2  FORMAT OP THE STANDARDS
     The control options are performance-based standards in the
format of a percent reduction.  The cutoff is applied using the
                               7-6

-------
annual mass emission total and an average vent stream flowrate
(in scfm).  These parameters were chosen to determine the
applicability of the cutoff because they were considerably easier
to deal with than concentration or duration of emission events.
Concentration and duration are extremely dynamic variables in
typical batch processing emissions, and, while flowrate and
yearly vent emissions also are dynamic, these parameters are
usually more available.  The flowrate from a vent is sometimes
known because the gas-moving equipment (i.e., compressors, vacuum
pumps) that is used to create the venting must be sized.
Flowrates from other batch emission events, such as displacements
and material heating, may be estimated using the Ideal Gas Law.
Specific situations and equations are presented in Chapter 3.
     The annual mass emission total also is required for
application of the cutoff to vents.  Annual uncontrolled
emissions are frequently reported to State agencies for the
purposes of permit review, state emission inventories, or Federal
programs, such as the Superfund Amendments and Reauthorization
Act (SARA) 313 reporting requirements.  For batch process vents,
however, the task of estimating annual emissions may be
complicated by several factors; among them are venting
configurations from multipurpose equipment and variations in
flow, concentrations, and emission stream duration.  In such
situations, owners or operators may elect to use material
balances in conjunction with control device efficiencies to
determine potential VOC emissions.
7.3  TESTING
     Source testing to measure annual mass emissions and maximum
flowrate for the purpose of determining applicability of a cutoff
is much more complex for batch processes (which have
noncontinuous and, often, multicomponent vent streams) than it is
for continuous processes.  The intermittent vent streams also
present serious problems for testing the performance of the
control devices.  Each step in a batch process, such as charging
the reactor or operating the dryer, generates gaseous streams
with independently defined characteristics.  This is illustrated
                               7-7

-------
in Chapter 6,  where the emission stream characteristics of
flowrate, temperature, duration, and VOC concentration are given
for a model batch process.   The gaseous streams from each step
may be vented separately, some or all streams may be combined
before venting to the atmosphere, and some operators may have the
flexibility of using different vents for the same equipment.
     In addition to the inherent problems of stack testing at
batch processing facilities, these industries tend to be reactive
to market demands and change product lines much more often than
continuous processing plants.  Vent stream characteristics change
with the production of new products.  This not only affects the
emission inventory for the plant; it can also affect the
performance of the control device.
     Testing may be more realistic for facilities that have all
vents from a single product processing area manifolded together,
and the common vent has a continuous, positive flow.  If
measurement of more typical batch process vents (in which flow
and concentration vary independently with time) is required,
several considerations related to measurement techniques must be
made.
     In the presence of unsteady or transient gas flows typical
of those found in batch gas streams, gas mass flow measurement
uncertainty can be decreased by utilizing measurement approaches
that separate density effects from velocity effects.  In
addition, electronic flow measurement (EFM) must be utilized to
allow mass flow averaging over the event time.  Typical
inexpensive gas flow measurement techniques (orifice meters and
pitot-type probes) are velocity head devices.  They measure
differential pressure as a function of both the gas density and
the stream velocity.  In transient batch-type situations where
density may be changing independent of velocity, this type of gas
flow measurement couples the effects and can potentially
introduce larger uncertainties into the velocity measurement.  In
addition, for velocity head devices, EFM systems must be utilized
to eliminate the errors associated with pressure averaging prior
to velocity calculations.  This error, often referred to as
                               7-8

-------
"square root" error, arises from the nonlinear dependence of the
measured variable (pressure) on the stream velocity.  In all
measurement devices where this occurs (orifice meters, pitot
tubes, annubars), the time-averaged value of the square root of
the pressure signal does not equal the square root of the time
averaged value of the pressure signal.  This inequality
introduces positive bias errors into the flow measurement and can
be eliminated by the use of EFM.
     Probes that are most suited for transient batch flow systems
are probably insertion turbine meters and ultrasonic probes.
Both of these probes can have turn-down ratios (ratio of maximum
to minimum measurable flow velocity) of 10-15 to 1 and are true
velocity measurement devices.  Both of these probes can be
hot-tapped into existing gas streams, and their uncertainty
levels are equal to or better than pitot tubes in steady flows.
The insertion turbine meter, like all pitot probes, requires a
traverse, which limits its application to transient flows.
     However, ultrasonic meters, which are used extensively in
chemical plants, return an average velocity flow across the gas
stream.  For this reason, ultrasonic probes can track shorter
transients with less uncertainty because of the elimination of
the need for a traverse at each sample interval.
     Simultaneous concentration measurements may be made using
EPA Method 25A, a semicontinuous Method 18 (at close intervals),
or perhaps by using Fourier Transform Infrared (FTIR) technology
(for which no EPA method currently exists), an emerging
technology that has experimentally been demonstrated to measure
multicomponent volatile compounds from a noninvasive standpoint.
     The use of EFM's to combine the flow and concentration
measurements and obtain instantaneous mass emissions, as well as
batch mass emissions (integrated over the batch cycle time)
appears to be indispensable for accurate emission measurements of
batch emission streams.  However, this testing is more
sophisticated and presumably more expensive than emissions
measurement for continuous, steady-state emission streams.
                                7-9

-------
     Probes that are moat suited for transient batch flow systems
are probably insertion turbine meters and ultrasonic probes.
Both of these probes can have turn-down ratios  (ratio of maximum
to minimum measurable flow velocity) of 10-15 to 1 and are true
velocity measurement devices.  Both of these probes can be
hot-tapped into existing gas streams, and their uncertainty
levels are equal to or better than pitot tubes in steady flows.
The insertion turbine meter, like all pitot probes, requires a
traverse, which limits its application to transient flows.
     However, ultrasonic meters, which are used extensively in
chemical plants, return an average velocity flow across the gas
stream.  For this reason, ultrasonic probes can track shorter
transients with less uncertainty because of the elimination of
the need for a traverse at each sample interval.
     Simultaneous concentration measurements may be made using
EPA Method 25A, a semicontinuous Method 18 (at close intervals),
or perhaps by using Fourier Transform Infrared  (FTIR) technology
(for which no EPA method currently exists), an emerging
technology that has experimentally been demonstrated to measure
multicomponent volatile compounds from a noninvasive standpoint.
     The use of EFM's to combine the flow and concentration
measurements and obtain instantaneous mass emissions, as well :.s
batch mass emissions (integrated over the batch cycle time)
appears to be indispensable for accurate emission measurements of
batch emission streams.  However, this testing is more
sophisticated and presumably more expensive than emissions
measurement for continuous,  steady-state emission streams.
     Another alternative is to measure emissions from a single
step in the process to confirm emission estimates based on
equations in Chapter 3.  This method also can be costly if
testing is required for the entire duration of the step, from
startup to completion.   Sampling periodically throughout the step
may be sufficient to characterize emissions and confirm emission
estimates in some situations.
                               7-10

-------
saturation).  Note that under these conditions, the unit also
will perform at maximum efficiency because the emission stream is
completely saturated.  In some cases, the varying incoming
emission stream characteristics make it impossible to meet an
instantaneous control efficiency value, but overall control
efficiency value can be met by controlling the richer peak load
(at higher efficiencies) and by not controlling the emission
streams when the VOC concentration begins to taper off.  If vent
stream characteristics or worst-case conditions are known, the
condensation unit can be designed to meet a standard, and a
performance test may not be necessary.  Monitoring can be
relatively simple.  Temperature monitors can be mounted at the
coolant inlet to the vapor condenser or the gas outlet, and
temperature can be recorded on a strip chart.  Flowmeters can
also be incorporated.
     Carbon adsorbers are another vapor recovery device that can
be used to meet rule requirements, and if vent stream
characteristics or worst-case conditions are known, a performance
test may not be necessary.  Again, a monitoring device should be
used to indicate and record the VOC mass emissions in the exhaust
gases from the carbon adsorber.  Of particular concern when using
carbon adsorption systems to control batch emission streams is
the desorption of VOC compounds from the carbon bed to the gas
exhaust when the VOC concentration in the entering gas stream
decreases as it might during a batch emission event.  The
adsorber may handle the peak VOC emissions only to desorb them
out during non-peak events, thereby producing an outlet stream
that is more uniform in concentration.  Thus, there may be no net
control from the device.
7.5  REPORTING/RZCORDKEEPING REQUIREMENTS
     Records should be kept that record the characteristics of
                           I
each process vent or group of process vents subject to a rule
that indicate average flowrate and annual mass emission total.
Note that the annual mass emission total combines the mass
emission potential for each emission event with the number of
potential emission events in a year.  If there is no permitted
                               7-11

-------
value, owners and operators must keep records of the number of
emission events that will occur in a year in order to obtain an
accurate mass emission total.
     Each facility required to control process vents should keep
a copy of the operating plan for each control device in use.  The
operating plan should identify the control method and parameters
to be monitored to ensure that the control device is operated in
conformance with its design.  Each facility should keep a record
of the measured values of the parameters monitored.  Any
exceedances of the design parameters should be recorded along
with any corrective actions.  The air pollution control agency
should decide which of the recorded data should be reported and
what the reporting frequency should be.
7.6' EXAMPLE APPLICATION
     Figure 7-1 presents an example analysis.  Individual unit
operations, as well as the aggregate process are evaluated using
the regression equations to determine whether control at an
example level (90 percent) is required.  The results indicate
that the dryer requires control of 90 percent, as does the
overall process.  The uncontrolled annual mass emissions from the
dryer are 36,000 Ib/yr.  At this level, emission sources with
flowrates less than 167 scfm (regardless of volatility) would be
required to be controlled to 90 percent.  Similarly, the
uncontrolled emissions from the aggregated process are
47,700 Ib/yr.  Processes with an average flowrate of 319 scfm or
lower would require control at 90 percent, again regardless of
volatility.  In this situation, operators might choose to control
the dryer emissions to a level in excess of 90 percent in order
to meet the overall process control requirement.
                               7-12

-------
                Figure 7-1. Example  Analysis




Djknr*tnr





Holding
Tank




Centrifuge




T

Dryer


* 10 packaging
  IndwiduaJ

  Unit Operation
Average Flowrate  hrs/baJch  Lbs VOC/balch  Ibs/yf

React
Holding Tank
Centrifuge

20sclm
lOscfcn
30scim

3
.25
.5

6
4
29
(@300 days/yr)
1800
1200
8700
               30sdm
                                    120
                               36000
                                    159 Ibs/batch
FR
1002 (low vol)
622 (mod vol)
167 (hi vol)
Aggregated


 & 1 bstctVday, 300 days/yr. 47.700 Ibs/yr

 .      	     ((20M3)(300)*(10M.25M300)*(30|( 5)(300)*(30M6)(300))
 WWBC80O rfOWftflw     mmi  _ _..      	
                .  26scfm
                                                           90%coolro<:
                                                           FR - 07(tos^yr) -1821

                                                           FR - ,031(ibs/yrH94
                                                           FR - 013(lbs/yr)-301
                                                                    @47700 Ibs/yr
                                                                         FR
                                                               kiwvol      1518
                                                               mod vol
                                                               Nvol
                        985
                        319
Conclusions: Would require control of dryer at 90% and overall process at 90%

-------
 APPENDIX A




PHYSICAL DATA

-------
TABLE A-l.  PHYSICAL PROPERTIES OF COMMON AIR SUBSTANCES
Off ink liquid
Petroleum liquid!**
OMoliiieRVPI3
OMolim RVP 10
Oeeoline RVP 7
Crude oil RVP 5
Jet Nephdu (JP-4)
lei keroeene
DiMilUle fuel No. 2
RMidiul oil No. 6
Volatile organic liquid*
Acetone
AcetonUrile
Acrykmitrile
AMylekohol
Allyl chloride
Ammonium hydroxide 28.1 percent ohrtion
Benzene
n- Butyl chloride
Carbon diwlfide
Cerbon lelrachloride
Chlnmform
Chloroprene
V.por
moleculer
weight

62
66
61
SO
to
130
130
190

58 1
41.1
53.1
51.1
765
35.1
71. 1
926
76.1
1531
119.4
185
Liquid
demity,
Ih/gil
160'F

56
56
56
7.1
6.4
70
7.1
79

6.6
6.6
6.8
7.1
7.9
7.5
7.4
74
106
134
125
80
Condenied
vepor
demity,
lh/gl it
60*F

40*F
True vipor preMire in pew et:
50'F
60"F

49
51
5.2
4.5
5.4
6.1
6.1
6.4
47
34
2.3
18
08
00041
0.0031
0.00002
5.7
4.2
2.9
2.3
10
0.0060
00045
000003

66
66
6.8
7.1
79
7.5
7.4
7.4
106
13 4
12 5
80
1.7
0.6
0.8
O.I
3.0
5.1
06
07
3.0
0.8
1.5
18
2.2
0.8
1.0
0.2
3.8
6.6
0.9
1.0
3.9
I.I
19
2.3
6.9
5.2
35
28
13
00085
00074
0.00004
TO'F

8.3
62
4.3
34
1.6
0.01 1
0.0090
0.00006

2.9
II
1.4
0.3
4.8
8.5
1.2
1.3
4.8
1.4
25
2.9
3.7
1.4
1.8
0.4
6.0
10.8
1.5
1.7
60
1.8
3.2
3.7
80'F
90*F

9.9
7.4
5.2
4.0
1.9
0015
0012
0.00009
11.7
88
6.2
4.8
24
0.021
0.016
000013

4.7
19
2.4
0.5
7.4
13.5
20
2.2
7.4
23
4.1
4.6
5.9
2.5
31
0.7
9.1
16.8
2.6
2.7
92
30
52
57
lOO'F

13.8
10.5
7.4
5.7
2.7
0.029
0022
000019

7.3
3.1
4.0
1.0
no
20.7
33
35
11.2
38
63
70

-------
TABLE A-l.  (continued)
Offwk liquid
CyclctwurM
CyclofMnUm
i.l-DkMonM*Me
1.2 Dichloroh.n.c
ei^ 1 ,2 Didiloro-hyUn.
mw-1 ,2-DkMoKMtfiylm*
DMiytMlMr
DMiyluniM
Diitopmpy) ther
1,4-DKMMM
Oipropy! Mhor
ElhytMMttte
Ethyl *ciyt*te
EAyLlcohol
Fraooll
n-Htpum
HeuiM*
bobutyt alcohol
Itnpmpyl ilcnhot
Meihyl aceUlc
Methyl *crylile
Methyl lci.hi>l
Vapm
molecular
weight
14.2
70.1
W.O
W.O
,0
97.0
74.1
73.1
1072
U.I
102.2
M.I
100.1
46.1
IJ7.4
1002
16.2
741
601
74.1
16 1
320
Liquid
((entity,
IWgl
M60F
6.5
6.2
9,9
105
lO.t
IO.S
60
59
61
i.7
6.3
7.6
7.1
6.6
125
5.7
55
6,7
6.6
7.1
to
66
Condenied
vpor
dcntiry,
lh/gl il
60"F
6.5
6.2
9.9
10.5
101
105
6.0
S.9
6.1
t.7
6.3
7.6
7
6.6
12.5
5.7
5,5
6.7
66
7.t
*0
66

40*F
0.7
2.S
1.7
0.6
15
2.6
4,2
1.6
12
0.2
0.4
0.6
0.2
0.2
7.0
03
1.1
0.06
02
1.5
06
07
True v*por preure in ptit tH:
50F
0.9
3.3
2.2
O.I
2.0
34
5.7
20
1.6
0.3
0.6
0,1
0.3
04
l.t
04
1.5
01
03
20
01
10
60*F
1.2
4.2
2.9
1.0
2.7
4.4
7.0
29
2.1
0.4
01
1.1
0.4
0.6
10.9
OS
1.9
01
0.6
2,7
1.0
14
TO'F
1.6
5.2
3.7
1.4
3.5
5.5
1.7
3.9
2.7
0.6
11
l.S
0.6
0.9
13.4
O.f
2.4
0.2
0.7
3.7
1.4
20
WF
2.1
6.5
4.7
1.7
4.4
6.1
104
4.9
3.5
O.I
1.4
1,9
O.I
IJ
16.3
1.0
3.1
0.3
0.9
47
l.t
26
90'F
2.6
I.I
5.9
22
5.6
1.3
13.3
61
4.3
11
1.9
2.5
II
1.7
197
12
3.9
0.4
1.3
S.t
2.4
35
100*F
1.2
9.7
T.4
21
61
10.0
oil*
7.5
5.3
1.5
23
3.2
1.5
2.3
236
16
4.9
05
It
70
31
45

-------
                                                       TABLE  A-l.    (continued)
Organic liquid
Methylene diloffivfl
Methyl cyclopentane
Methyl ethyl krtone
Methylmethacr /late
Methyl pmpyl *her
n-Pentanec
n Propxlsmine
Ptopyl chkwid, c
Tettbutyl akohol
U.I-Trichloroethane
Tnchloroethykne
Toluene
Vinylecetate
Vinyledene chlood*
Vapor
mnleculir
weight
49
M.I
72.1
100
74 1
7i.2
5 1
71.5
74.1
1334
131 4
92 1
16.1
965
Liquid
density,
fc/gal
160-F
II. 1
6.3
67
79
62
53
60
d
66
112
123
7.3
7.8
10.4
Condensed
vapor
density.
Ib/gal si
60-F
II. 1
6.3
67
7.9
62
53
6.0
d
d
11.2
123
73
7
10.4
True vapor pressure in psis tl:
40*F
3.1
0.9
07
0.1
37
43
25
21
02
09
05
02
07
50
50'F
4.3
12
0.9
0.2
4.7
5.5
3.2
35
0.3
1.2
0.7
0.2
1.0
63
60'F
54
1.6
1.2
03
61
61
4.2
45
0.4
16
09
03
1.3
7.9
TO'F
6.1
2.2
15
06
7.1
15
5.3
5.6
0.6
2.0
1.2
0.4
1.7
9.1
IO*F
1.7
2.9
2.1
01
94
10.5
65
7.0
09
2.6
1.5
0.6
2.3
II I
90*F
10.3
3.6
2.7
II
11.6
121
S.O
1.7
1.2
3.3
2.0
O.t
3.1
153
lOO'F
13.3
4.5
3.3
1.4
137
156
96
10.6
1.7
4.2
2.0
1.0
40
232
*RVP => Reid vtpor pressure.
Vapor pressures calculated from pafM D-212 throufh D-2IS of Handbook of Physic, and Chemistry." 67h Edition.
cDala unavailable.
Source. Hazardous Waste Trealmeot, Storage, and Disposal Facilities fTSDF)  Air Emission Models. EPA^SO/3 17426. December 1987.

-------
        TABLE  A-2.  VAPOR PRESSURE  -  EQUATION CONSTANTS

NO FORMULA
1 C2H40
2 C2M50N
3 C2H3N
4 C8H80
5 C3H40
6 C3H5NO
7 C3H402
8 C3H3*
9 C3H5CL
10 C6H7N
11 C7H9NO
12 C6H6
13 C7M5CL3
H C7H7CL
IS C12N10
16 C2H4C120
17 CHII3
18 C4H6
19 C6H110N
20 CS2
21 CCU
22 C2H3CL02
23 C8H7CLO
24 C6H5CL
25 CHCL3
26 C4H5CL
77 C7H80
2?
29 C7N80
30 C7H80
31 C9H12
32 C6H4CL2
33 C4M8CL20
34 C3H4CL2
35 C4H11N02
36 C8H11N
37 C4H1004S
38 C14H20N
39 C3H7N
40 C2NBM2
41 C10M1004
42 C2H6S04
43 C6N3N204
44 C7H6N204
45 C4H8O2
46 C12H12K2
47 C3H5CLO
48 CSH802
49 C8H10
SO C2H5CL

NAME
ACITAL01NYOE
ACETAMIDE
ACETON1TRIL1
ACETOPNtMNl
ACROLE1N
ACRYLAMID1
ACRYLIC ACID
ACRYLONITRILl
ALLTL CHLORIDE
ANILINE
0-ANIS101NE
KNZENE
KNZOTR I CHLORIDE
BENZYL CHLORIDE
IPHENYL
IS(CMLORCMETHYL)ETHER
tROMOFORM
1,3-BUTADlENE
CAPROLACTAM
CAR80M OISULFIDE
CARBON TETRACHLORIOE
CHLOROACETIC ACID
2-CHLOROACETOPHENONE
CHLOROBENZENE
CHLOROFORM
CHLOROPRENE
M-CRESOL
CRESOLS/CRESYLIC ACIOdSOMERS
0-CRESOL
P-CRESOL
CUMEHE
1,4-DICNLOROBENZENE
OICHLOROETHYL ETHER
1,3-DICHLOROPROPENE*
OIETHANOLAMINE
N.N-DINETYLANIL1NE
D1ETHYL SULFATE
DIMETHYLBENZ1D1NE
DIMETHYL FORMAMIDE
1,1-DIMETHYLHYORAZlNE
DIMETHYL PHTHALATE
DIMETHYL SULFATE
2,4-DlNITROPHNOL
2.4-DlNITROTOCUtNE
1,4-DlOXANl
1,2-OIPHMYLNYMAZINl
EPICHLORONYDRIN
ETHYL ACRYLATE
ETHYLBENZENE
ETHYL CHLORIDE
In P  A  8/T * C In T  0 Te (P  m H|, T  K)
A I C 01 THIN TNAX
201.1772 -8.47861*03 -3.15481*01 4.63141-02 1 150.15 441.00
127.5872 -1.19*1E*04 -1.60681*01 1.1880E-05 2 354.15 494.30
53.4092 -S.ttUC03 -5.49$4E*00 5.3634C-M 2 229.32 545.50
127.9772 -1.03SE*04 -1.724E*01 1. 47791-02 1 292.80 701.00
133.5072 -7.12271*03 -1.94381*01 2.6447E-02 1 115.45 506.00
39.1412 -1.0231E*04 -1.71391*00 	 7.5 4*5.75
53.0992 -7.21801*03 -4.813f*00 1.0060E-03 1 216.65 615.00
82.7112 -6.39271*03 -1.01011*01 1.0891E-05 2 116.63 535.00
38.1982 -4.30841*03 -3.1322*00 1.11711-17 6 138.65 514.15
286.3872 -1.65041*04 -4.2763f*01 3.99181-02 1 267.13 699.00
 * **     
73.1572 -6.275SE*03 -8.4443E*00 6.2600E-06 2 278.68 562.16
50.6272 -7.4190E*03 -4.6313E*00 1.7396C-18 6 268.40 737.00
49.8582 -7.1698E*03 -4.4836E*00 1 .38588 -18 6 234.15 686.00
122.1472 -1.23211*04 -1.4955E*01 5.60S6E-06 2 342.37 780.26
56.1552 6.39841*03 -5. 49724*00 8.2034E-18 * 231.65 579.00
53.1752 -6.76331*03 -5.05141*00 2.96S3E-18 6 281.20 696.00
69.2092 -4.58001*03 -8.2922C*00 1.18206-05 2 164.25 425.37
69.2792 -1.0*69E*04 -6.89441*00 1.2113E-18 6 342.3* 806.00
57.9042 -4.70631*03 -6.77941*00 8.0195E-03 1 161.11 552.00
73.5462 -6.1281E*03 -8.57631*00 6.8461E-06 2 250.33 556.35
98.2572 -1.058SE*04 -1.13481*01 4.14351-06 2 333.15 686.00
.... .... .... .... . ... ...
44.7492 -5.94081*03 -3.93911*00 1.14171-06 2 227.95 632.35
130.3672 -7.47461*03 -1.87001*01 2.19091-02 1 209.63 536.40
42.9902 -4.7595E*03 -3.79961*00 1.17261-17 6 143.15 525.00
242.9872 -1.60601*04 -3.50831*01 2.88001-02 1 285.39 705.65
1 MIXTURES) 	 	   	
205.9872 -1.39281*04 -2.94831*01 2.51821-02 1 304.19 697.55
282.9872 -1.75401*04 -4.16371*01 3.61711-02 1 307.93 704.65
82.7612 -8.33401*03 -9.35671*00 1.36001-17 6 177.14 631.15
83.4172 -8.46341*03 -9.63081*00 4.58331-06 2 326.14 684.75
    *   * 
44.1267 -5.33471*03 -3.95721*00 6.96741-18 6 191.50 577.00
281.1172 -2.03601*04 -4.04221*01 3.2378E-02 1 301.15 542.04
46.4592 -7.16001*03 -4.01271*00 8.1481E-07 2 275.60 687.15
86.4342 -9.2791E*03 -1.03401*01 6.86751-03 1 248.00 483.00
.... .... .... .... . ... ...
110.7172 -9.85381*03 -1.33931*01 2.18671-17 6 212.72 647.00
 * * * * 
66.1802 -1.05341*04 -6.42981*00 1.08041-18 * 272.15 766.00
78.1512 -8.87191*03 -8.5921E*00 1.89411-06 2 241.35 758.00
  *  * * * * 
26.7022 -6.92591*03 -1.64*81*00 3.67251-03 1 343.00 814.00
47.3782 -5.67771*03 -4.3*451*00 1.9*261-0* 2 284.95 587.00
89.6402 -1.278SE*04 -9.5*731*00 1.***0t-18 * 404.15 573.00
57.0212 -6.64201*03 -5.62521*00 1.22801-06 2 215.95 610.00
126.6672 -8.26721*03 -1.76941*01 1.85381-02 1 201.95 553.00
83.3532 -7.691 1E*03 -9.79701*00 $.93101-06 2 178.15 617.17
6!>.2*&2 -4.78671*03 -7.53871*00 9.33701-0* 2 134.80 4*0.35
From "Henry's Law Constant  for HAP's."  Carl Yaws.  Prepared for the U. S.
Environmental Protection Agency.  Final Report.  September 30,  1992.
                                 A-4

-------
TABLE A-2.   (continued)

NO FORMULA
51 C2H48R2
52 C2H4CL2
S3 C2HA02
54 C2H40
55 C2H4CL2
56 CN20
57 C4H1002
58 C4H1002
59 C8H1604
60 C6H1203
61 C6MU03
62 C5H1003
63 C8H1B03
64 C6H1403
65 C3H802
66 C6N1202
67 C8H1002
68 CSH1202
69 C8M1803
70 C6HU02
71 CH1 804
72 C8H15Q3
73 C6CL6
74 C4CL6
75 C2CL6
76 C6N14
77 C8H602
78 C9H140
79 C4H203
80 CM40
81 CH3IR
82 CH3CL
83 C2H3CL3
84 C4H80
85 CN6M2
86 C6H120
87 C2H3NO
88 C5N802
89 CSH120
90 CH2CL2
91 C15H10K202
92 C13H14H2
93 C10H8
94 C6M5N02
95 C6H5H03
96 C3H7NG2
97 C6H60
98 C6M8N2
99 COCL2
100 C8H403

NAME
1THYL1H1 01MOM1D1
CTNTLENC DICNLOR1DI
ETNTLEHE GLYCOL
ETHYLENE OXIDE
1THYLIOM1 BICHLORIDE
FCftNALDEMYDE
ETNTLEHE GLYCOL DINETNTL ETHER
ETHTLEHE 6LYCOL MON01WL ETHER
OIETMTLEHE 6LTCOI WHOCTHTL ETHER ACETATE
ETHTLEHE 6LYCOL NONOETHYL ETHER ACETATE
DIETHTLEHE 6LYCOL MOHOETHYL ETHER
ETHYLEHE GLYCOL MOHOMETHYL ETHER ACETATE*
DIETHYLEHE CLYCOL MXM08UTYL ETHER
01 ETHYLEHE CLYCOL DIMETHYL ETHER
ETHYLEHE GLYCOL MONOMETHYL ETHER
ETHYLEHE CLYCOL MONOPROm ETHER
ETHYLEHE GLTCOL MOHOPHEHYL ETHER
DIETHYLEHE GLYCOL MOMOMETHYL ETHER
DIETHYLEHE GLYCOL DIETHYL ETHER
ETHYLEHE GLYCOL MOH08UTYL ETHER
TRIETHYLEHE GLYCOL DIMETHYL ETHER
ETHYLEHE GLYCOL MON08UTYL ETHER ACETATE
HEXACHLOR08EHZEHE
HEXACHLOR08UTADIEHE
HEXACHLOROETHAHE
HEXAHE
HYDROOUIHONE
ISOPHORONE
MALE 1C AHHYDR10E
METHAHOL
METHYL BROMIDE
METHYL CHLORIDE
METHYL CHLOROFORM
METHYL ETHYL KETOME
METHYL HYORA21HE
METHYL iSOtUTYL KETOHE
METHYL ISOCYAHATE
METHYL METHACRYLATE
METHYL TERT-BUTYL ETHER
METHYLEHE CHLORIDE
METHYLEME 01PH1HYL 01 ISOCYAHATE
4,4-METHYLEHEDlAHILlHE
NAPHTHALlNi
HITROBEHZEHE
4-HITRQPHtHOL
2-HITROPROPAH1
PHENOL
P-PHEMYLENtDIAfUNl
PHOSGfMf
PHTHAL1C AHHYDRIDE
In P  A %1/T %C In T %D Tf If  m Hf, T - C)
At C DC TNI TMAX
38.8582 -5. $1771*03 -3.01911*00 8.26641-07 2 2S2.8S 650.15
111.4972 -7.3230E*03 -1.53701*01 1.6794E-02 1 237.69 561.00
189.7472 -1.4615E04 -2.54331*01 2.0140E-05 2 2*0.15 645.00
91.9272 -5.43308*03 -1.25171*01 1.608W-02 1 160.71 4*9.15
76.8602 -6.0103E*03 -9.133*1*00 8. 59*01-06 2 176.19 $23.00
96.6172 -4.91721*03 -1.37*51*01 2.2031E-02 1 111.1$ 408.00
80.1902 -*.37Z2f*03 -1.00831*01 f. 94991-03 1 213.15 536.15
2*6.7972 -1.3*45E*04 -4.09001*01 4.109*1-02 1 113.00 5*9.00
105.8972 -9.90511*03 -1.37291*01 1.2203E-02 1 2U.15 6*0.00
79.6572 -8.*7B3t*03 -1.72441*00 1.04591-17 * 211.45 597.00
250.9672 -1.71*41*04 -3.4*991*01 2.5107E-05 Z 230.00 632.00
80.0053 -8.6783C*03 8.72441*00 1.04391-17 * 211.45 597.00
173.6772 -1.57121*04 -2.19011*01 2.15*91-17  205.15 654.00
78.2492 -8.28411*03 -1.6*871*00 1.9*291-17 * 203.15 432.91
353.1672 -1.63901*04 -5.54501*01 *.*I1I-02 1 225.00 5*4.00
65.3342 -7.77121*03 -6.69*41*00 2.239K-17 * 113.15 582.00
 **  **   *
428.7372 -2.1730E*04 -6.64 161*01 6.9903E-02 1 250.00 630.00
71.59*2 -8.5825E*03 -7.68471*00 3.60091-06 2 228.15 624.00
110.6072 -1.0705E*04 -1.31401*01 2.97811-17 * 203.15 600.00
98.8572 -1.1*331*04 -1.10671*01 6.22088-18  229.35 *51.00
**. *     *
158.3372 -1.8324E*04 -1.1*991*01 2.39021-11  501.70 125.00
81.9512 -9.52801*03 -9.0*0*1*00 1.3SME-0* 2 252.15 741.00
430.2172 -2.72201*04 -.0495t*01 3.01*51-0$ 2 459.95 512.25
160.5772 -1.35331*03 -2.39271*01 2.94A91-02 1 177.84 507.43
105.9772 -1.215*1*04 -1.26771*01 6.91201-03 1 444.65 822.00
78.1382 -8.112*1*03 -9.51171*00 1.17141-03 1 265.05 715.00
63.9872 -7.722*1*03 -7.20171*00 7.01*91-03 1 326.00 710.00
105.0372 -7.47131*03 -1.39111*01 1.52111-02 1 175.47 512.58
67.6932 -4.*98*E*03 -7.996*1*00 1.15531-05 2 179.47 467.00
59.2372 -4.0301E*03 -6.71511*00 1.02101-05 2 175.43 416.25
14.1522 -6.54421*03 -1.02051*01 8.53481-06 2 242.75 545.00
109.8472 -7.13008*03 -1.51141*01 1.72341-02 1 186.48 535.50
 * **  * *
147.8072 -1.00341*04 -1.976*1*01 1.A35SE-05 2 119.15 571.40
41.7632 -4.455*1*03 -3.63391*00 1.50241-17 * 256.15 505.00
246.1372 -1.21441*04 -3.76541*01 4. 28731-02 1 224.95 564.00
50.9112 -5.13011*03 -4.9*171*00 1.97*51-17 * 1*4.55 497.10
74.9742 -5.79471*03 -1.10151*00 7.64321-06 2 171.01 510.00
71.9502 -1.3*041*04 -7.14291*00 4.00251 -11 * 311.20 609.00
*    *     *  
10.3972 -9.0*221*03 -9.0*411*00 3. 58051 -06 2 353.43 748.35
85.5522 -9.74411*03 -9.52211*00 7.5*591-11 ft 	
*  * * * *  *    *  *
51.5512 -6.29031*03 -4.14*21*00 9.22731-11 * 111.13 594.00
54.1172 -1.05001*03 -4.19901*00 2.1001-04 1 314.0* 694.25
79.0322 -1.13411*04 -1.17*91*00 1.57*11-11  413.00 796.00
107.4272 -5.67741*03 -1.53511*01 2.12501-02 1 145.37 455. 00
70.5352 -1.93021*03 -7.1*711*00 5.9*031-0* 2 404.15 791.00
          A-5

-------
                                  TABLE A-2.    (continued)
NO  FORMULA
              NAME
                                                     In P  A * i/T * C In T * 0 T"
                                         If  m H|, T
                                                                                             I TMIH   TMM
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
C3H402
C3H60
C3N6CL2
C3N60
canto?
C8H8
C2H2CL4
C2CL4
C7N8
C7M10N2
C9H6N202
C7N9N
C6H3CL3
C2K3CL3
C2HCL3
C6H3CL30
C6H15H
C8H18
C4N602
C2H3CL
C2H2CL2

C8H10
C8H10
C8H10
KTA-PMPIOLACTONE
PtOPIOMALOEHTDE
PtOPYLENE OICNLORIOE
PROPYLENE 0X108
UINON8
STYRENE
1 ,1 ,2,2-Tf TRACHLOROETNAN8
TETRACHLOROETHYLENE
TOLUENE
2. 4- TOLUENE DIAMINE
2,4-TOLUENE OIISOCYAMATE
0-TOLUJDINE
1,2,4-TRICNLOR08ENZENE
1.1,2-TRICHLOROETHANE
TRICHLROETHYLENE
2.4,5-TRICHLOROPHENOL
TRIETHYLAMINE
2,2,4-TRIMETHYLPENTANE
VINYL ACETATE
VINYL CHLORIDE
VINYLIOENE CHLORIDE
XYLENES (ISONERS C MIXTURES)
M-XYLENE
0-XYLENE
P-XYLENE
59.6992 -7.82048*03
60.
49.
88
* 
128
129
53
78
100
95
222
35
57
54
 
51
115
43
121


.
,
.
,
4

,
B
m






2442 
2312 
7372 
*
6272 
4872 -
8712 
4662 -
9772 -
0812 
3572 
9082 
7592 -
5102 
*
6572 -
9172 
0492 
9572 
67.7482 -
 -
79
85
138

m
.30958*03
.67748*03
.05808*03
....
.26558*03
.02738*04
.19128*03
.99508*03
.26488*04
.16598*04
.44248*04
.65918*03
.30178*03
.47166*03
....
.68198*03
.55008*03
.24628*03
.76018*03
.44818*03
- ....
8542 -7.59418*03
.7512 -7.9608E*03
.2772 -9.24708*03
5.78088*00
6.52898*00
4.60638*00
1.11048*01
....
-1.76098*01
1.65568*01
5.33128*00
9.16358*00
1.14728*01
-1.05838*01
-3.22638*01
-2.55498*00
5.91828*00
5.82758*00
 *  
4.96158*00
1.61118*01
3.63608*00
1.79148*01
7.56978*00
....
9.25708*00
1.01268*01
-1.94418*01
3.06898-18
5.86118*06
9.02128*18
1.26708-05
* 
1.53918-02
9.30818-06
2.12698-06
6.2250E-06
2.90078-06
4.15438-18
2.86621-02
4.69368-04
2.72418-06
4.50988-03
....
1.23638-17
1.70998-02
4.37988-18
2.49178-02
7.09228*17
   *
5.55008*06
6.01508-06
1.9084E-02
6
2
6
2

1
2
2
2
2
6
1
1
2
1

6
1
6
1
6

2
2
1
239.75
193.15
172.71
161.22
 *
242.54
229.35
250.80
178.18
371.25
M7.04-
249.47
290.15
236.50
188.40
...
138.45
165.78
180.35
119.36
10.59
*
225.30
247.98
286.41
685.00
496.00
572.00
482.25
 * *
648.00
645.00
620.00
591.79
804.00
737.00
694.15
725.00
302.00
571.15
...
535.15
543.96
524.00
431.55
482.00
...
617.05
630.37
616.26
  EttiMttd vcluts  for eotfficitnts in vapor  prtuurt quation.
In  natural logarithm
Prinary data tourca:     Oaubart,  T. E. and R.  P. Darotr, D*U COMPILATION or PROPERTIES Of PURE COMMUiBt. Parts 1,2,3
                       and 4,  Svcplanams 1 and 2, D1PP*  Project, AJChE, Haw York, NY  (1985*1992).
loo P 
NO
11
33
40
67
72
85
116
FORMULA
C7H9NO
C4N8CL20
C2H8N2
C8N1002
C8M1503
CH6N2
C6N3CLSO
NAME
0-ANISIOINE
OICNLOR08TNYL ETHER
1 , 1 -OIM8TNYLNYWU2INE
8TNTL8N8 OLYCOL NONOPNENYL ETHER
8TNYLEN8 SLYCOL MM08UTYL ETHER ACETATE
METHYL NYORA2IN8
2.4,5-TRICHLOROPHENOt
A
8
7
7
7
7
6
7

.30799
.69239
.58826
.15937
.21589
.84297
.82316
lot  lofaHtha to bata 10
A  8/(T*C) (  a Nf ,
8
2475
1990
1388
1767
1659
1115
2420
VAPOR

.780
.755
.510
.871
.242
.190
.564
MEi
C
237.
235.
232.
168.
191.
191.
237.
Ml.
TMIH
134
347
537 
070
339
648
476
Oata
61
23
35
25
25
2
72
Oool
T - C)
TMAX
218
178
20
245
. 192
25
252
i **ttal
                               Co*any,  Tokyo. Japan (1976)
        Primary data aourca for 67 and 72:
Zurse. CO., editor, 6LYCOLI. Rainheld Htoliahir* Corp.
York, NY (1953).
                                                  A-6

-------
     TABLE  A-3.   SUMMATION OF  DATA  FOR HENRY'S  LAW CONSTANT

NO FORMULA
1 C2H40
2 C2HSON
3 C2H3N
4 C8H80
5 C3H40
6 C3H5NO
7 C3H402
8 C3H3N
9 C3HSCL
10 Ca*7N
11 crH9NO
12 C*H6
13 C7HSCL3
14 C7H7CL
IS C12H10
1* t?M^ri ?n
iw wcnvwkcv
17 CHBR3
18 C4H6
19 C6H110N
20 CS2
21 CCL4
22 C2H3CL02
23 C8H7CLO
24 C6H5CL
25 CHCL3
26 C4H5CL
27 C7H80


29 C/H80
30 C7H80
31 C9H12
32 C6H4CL2
33 C4H8CL20
34 CJH4CL2
35 MH11N02
36 C8H11N
37 C4H1004S
38 C14H16N2
39 C3H7NO
40 C2H8N2
41 C10H1004
42 C2H6S04
43 C6H3N204
44 C7H6N204
45 C4H802
46 C12H12M2
47 C3HSCLO
48 CSH802
49 C8H10
SO C2H5CL

NAME
ACETALDEHYDE
ACETAMIDE
ACETONITRILE
ACETOPHENONE
ACROLEIN
ACRYLAMIDE
ACRYLIC ACID
ACRYLONITRILE
ALLVL CHLORIDE
ANILINE
0-ANISIDINE
BENZENE
BENZOTRICHLORIDE
BENZYL CHLORIDE
BIPHENYL
 t c/rui flOflHCTUVl )ETUF>
 J >V WnklMl^^ inikJEInCK
BROMOFORM
1,3-SUTADIENE
CAPROLACTAM
CARBON DISULFIDE
CARBON TETRACHLOR1DE
CHLOROACETIC ACID
2-CHLOROACETOPHENONE
CHLOROBENZEHE
CHLOROFORM
CHLOROPRENE
M-CRESOL
CarECm C /(*)? C VI IT a\f*tfW t CHiiTt C
W%EUL/WH8>*T L 1 W Aw III V 1 UnCKa
O-CRESOL
P-CRESOL
CUMENE
1,4-DICHLOROSENZENE
DICHLOROETHYL ETHER
1,3-DICHLOROPROPENE
DIETHANOLAMINE
N,N-DIMETYLANILIHE
OIETHYL SULFATE
DIMETHYLBENZIOIME
DIMETHYL FORMAMIDE
1,1-D1METHYLHYDRA2INE
DIMETHYL PHTHALATE
DIMETHYL SULFATE
2,4-DIHITROPHENOL
2,4-DINITROTOLUENE
1.4-DIOXANE
1,2-OIPHEHYLHYDRAZIHE
EP1CHLOROHYORIH
ETHYL ACRVLATE
ETHVLSENZENE
ETHYL CHLORIDE
Honry'e Law Constant,
H 8 25 C
4.8730000
0.0000986
1.1076388
0.5089400
4.5711400
0.0000145
0.0223962
5.4484900
515.4180500
0.0977600
0.0092393
308.3400000
54.5177107
17.7286753
22.6700000


29.5600000
3961.1453000
0.0001639
1064.0713500
1677.7900000
0.0036272
1.5713000
209.4500000
221.3300000
51.6355560
0.0394800


0.0911500
0.0396800
727.7800000
176.1100000
1.1390000
197.2200000
0.0000001
0.7701322
0.3405000
0.1780100
0.0098341
0.0910756
0.0548542
0.2226700
0.4756000
0.3996900
0.3079797
0.0135700
1.8590400
14.1169500
437.8100000
67?.. 2200000
H  ata/aoi fraction
8AS1S
txpariaantal
UNIFAC
VIE Data
Solubility Data
Solubility Data
UNIFAC
VLE Data
Solubility Data
Solubility Data
Solubility Data
UNIFAC
Exporiaantal
UNIFAC
UNIFAC
Eiporiaantal
A^AI i ^k
action nun Matar
ExpoHaantal
Solubility Data
UNIFAC
Solubility Data
Exporiaantal
UNIFAC
Solubility  EstiMtM
Exporlaamal
Expartftantal
UNIFAC
Solubility Data


Solubility Data
Solubility Data
Eipariaantal
Eporia*ntal
Solubility Data
Exporiaantal
UNIFAC
UNIFAC
Solubility Data
Solubfllty  EatiMtod
VLE Data
VLE Data
UNIFAC
Solubility Data
Solubility Data
Solubility Data
VLE Data
Solubility  EatiMtad
Solubility Data
Solubility Data
Exporiavntal
Experiaantal
To convert from H  in atm/vol fraction to:
H in atm/ (mol/m3), divide by 55,556
H in mmHg/mol  fraction, multiply by 760
H in psia/mol  fraction, multiply by 19.7
H in kPA/mol fraction, multiply by 101.325
H in kPa/mol/m3),  multiply by 101.325/55,556
Source:   Carl  Yawi, "Henry's Law Constant for HAPs",
September 30,  1992.
Final Report.
                                  A-7

-------
                             TABLE  A-3.     (CONTINUED)
NO  FORMULA    NAME
Honry't Law Cow tint.    N   ataVasl fraction

       N 8 25 C         IASIS
 51 C2H48R2    ITNTLCHC DIIMNIDC                            36.1100000
 52 C2H4CL2    ETHYLENE 01CHLORIDE                           65.3800000
 S3 C2H602     fTHVLINE 6LVCOL                                0.0001051
 Si C2H40      BTNVLENE OXIDE                               13.2280773
 55 C2M4CL2    ETNYUDENE 01CHLORIDE                        312.2300000
 56 CM20       FORMALDEHYDE                                  0.0187000
 57 C4H1002    BTHYLENE CLVCOL  DIMETHYL ETHCR                 1.9*71264
 58 C4H1002    ITHYLENE CLTCOL  MJNOETMYL fTMEt                0.0409170
 59 C8N1604    OIETHVLENE CLYCOL NOMOETKYL ETHER ACETATE      0.035840*
 60 C6H1203    ETHYLENE CLYCOL  NONOETHYL ETHER ACETATE        0.0986300
 61 C6H1403    OIETNYLEHE CLYCOL MOHOETNYL ETHER              0.0026793
 62 CSH1003    ETHYLENE CLVCOL  NONOMETHYL ETHER ACETATE*      0.1218685
 63 C8H1803    OIETHYLENE CLYCOL MONOtUTYL ETHER              0.0012481
 64 C6H1403    OIETHVLEME CLVCOL DIMETHYL ETHER               0.0837496
 65 C3H802     ETHYLENE GLYCOL  MONOMETHYL ETHER               0.0405801
 66 C6H1202    ETHYLENE fiLYCOL  MONOPROm ETHER               0.0474169
 67 C8H1002    ETHYLENE GLVCOL  HONOPHENYL ETHER               0.0037600
 68 CSH1202    OIETHYLENE CLYCOL NONOMETHYL  ETHER             0.0022577
 69 C8H1803    OIETHYLENE CLVCOL OIETHYL ETHER                0.1189224
 70 C6H1402    ETHTLENE CLYCOL  NONOBUTYt ETHER                0.0292288
 71 C8H1804    TRIETHYLENE GLYCOL DIMETHYL ETHER              0.0025951
 72 C8H1503    ETNYLENE GLVCOL  NONOtUTYL ETHER ACETATE        0.2746400
 73 C6CL6      NEXACHLOROKM2ENE                            94.4500000
 74 C4CL6      HEXACHLOR08UTADIENE                          572.2300000
 75 C2CL6      HEXACHLOROETHANE                            463.8900000
 76 C6H14      NEXANE                                    42667.0100000
 77 C8H602     NYOROOUINONE                                  0.0000800
 78 C9H14O     ISOPHORONE                                    0.3682100
 79 C4H203     MALE1C ANHYDRIDE                              0.0121651
 80 CH40       METHANOL                                      0.2885032
 81 CH38R      METHYL 8ROMIDE                              381.0578800
 82 CH3CL      METHYL CHLORIDE                              490.0000000
 83 C2H3CL3    METHYL CHLOROFORM                           966.6700000
 84 C4H80      METHYL ETHVL KZTONE                            7.2200000
 85 CH6N2      METNYL HYDRA2INE                              0.0248008
 86 C6H120     METNYL IS08UTYL  KETONE                       21.6700000
 87 C2H3NO     METNYL I80CYANATE                             	
 88 CSH802     METHYL NETNACRVLATE                            7.8317700
 89 C5H120     METNYL TERT-8UTYL ETHER                      30.8401800
 90 CH2CL2     METHYLENE CHLORIDE                           164.4500000
 91 C15H10N202 METHYLENE DIPHENYl DIISOCYANATE"              0.0026600"
 92 C13N14N2   4,4-METHYLENEIANILINE                        0.0284900
 93 C10H8      NA^NTHALENf                                  26.8300000
 94 C6H5N02    HITROKNUME                                  1.3300000
 95 C6H5N03    4-NITROPHENOL                                  0.0064600
 96 C3H7N02    2-NITMMOMME                                6.6111800
 97 C6N60      WENOL                                        0.0722000
 98 C6N8N2     f-rNENVLEMfDIANINC                             0.0007700
 99 COCL2      PHOSGENE"                                  780.0225300"
100 C8H403     MTHALIC ANHYDRIDE                             0.0441500
                       Exp*r
-------
                            TABLE  A-3.    (CONTINUED)
                                                 Nenry'a LM Constant,   N  atm/mol fraction
NO
NAME
 25 e
AS It
101 C3H402
102 C3H60
103 C3H6CL2
104 C3H60
105 C6N402
106 C8N8
107 C2H2CL4
108 C2CL4
109 C7M8
110 C7H10N2
111 C9H6N202
112 C7H9N
113 C6H3CL3
114 C2H3CL3
115 C2HCL3
116 C6HSC130
117 C6H15N
118 C8H18
119 C4H602
120 C2H3CL
121 C2H2CL2
122
123 C8H10
124 C8M10
125 C8H10
8ETA-PROPIOUCTONE
MOPIONALDEMVDE
WWmENS DtCNLORlDE
MOTYLENE OXIDE
QUINONE
STYRENE
1 , 1 ,2, 2 -TETRAD'- 5*9* TWIXf
TETRACMLOROETHYLEME
TOLUENE
2,4-TOLUENE DiANINE
2,4-TOLUENE DIISOCYANATE**
0-TOLUIDINE
1,2,4-TRICHLOROBENZEN*
1,1,2-TRICHLOROETHANE
TRICHLOROTHYLENE
2,4,5-TRlCHLORVHENCX.
TRIETHYLAMINE
2,2,4-TRIMETHYL*EHTANE
VINYL ACETATE
VINYL CHLORIDE
VINYLIDENE CHLORIDE
M-XYLENE
0-XYLENE
P-XYLENE
0.0063801
3.3224900
158.7100000
19.7742986
0.0576800
144.7155400
13.8000000
90.3400000
356.6700000
0.0000742
0.0091900**
0.1344600
106.6700000
45- 7/00000
566.6700000
0.4841100
6.9428000
185451.3318600
28.2111800
1472.2300000
1438.9000000
413.3400000
270.5600000
413.3400000
UNIFAC
Solubility Data
Experimental
VIC Data
Solubility Data
Solubility Oata
Experimental
Experimental
Experimental
UNIFAC
Solubility  Estimated
Solubility Oata
Experimental
Experimental
Experimental
Solubility Oata
Solubility Data
Solubility Data
Solubility Data
Experimental
Experimental
Experimental
Experimental
Experimental
Not**:

1.   EsttMtad valuta for coefficient! in vapor pressure aquation.

2. *  Reacts with water.

3. For basis  of UNIFAC, the estlswtion of the activity coefficient at infinite dilution Makes use of
the aroup contribution contribution Mthod using  the UNIFAC equatiena (Qa*hl
-------
                     TABLE A-4.   VALUES OF THE GAS  CONSTANT  R IN PV =  n RT
N
gm mol
gra mol
gm mol
gm mol
gm mol
gm mol
Ibmol
Ib mol
Ibmol
Ibmol
Ibmol
Ibmol
Ibmol
Temp.
K
K
K
K
K
K
R
"R
R
R
R
K
K
Pressure
aim
tin
nun H
bar
kg/cm2
kP
aim
in. Hg
mm Hg
Ib/in.2
Ib/ft2
Im
mm Hg
Volume
liter
cm3
liter
liter
liter
m3
ft3
ft3
ft3
ft3
ft3
ft3
ft3
R
0.082057477
82.057
62.364
0.083145
0.084784
0.0083145
0.73024
21.850
554.98
10.732
1545.3
1.3144
998.97
n
gm mol
gm mol

Ibmol
Ibmol
Ibmol
Ibmol






Temp.
K
K

R
R
R
R






Energy
calorie
joule

Btu
hp-h
Kw-h
ft-lb






R
1.9859
8.3145

1.9859
0.00078048
0.00058200
1545.3






Source:  Engineering Data Book. Gas Processors Suppliers Association, Ninth Edition.

-------
     TABLE A-5.  SPECIFIC LEAK RATES FOR ROUGH VACUUM SYSTEM
                            COMPONENTS
Component
6 - specific
leak rate, Ib/h/in.
Static seals
0-ring construction
Conventional gasket seals
Thermally cycled static seals
t<200F
200400F
Motion (rotary) seals
0-ring construction
Mechanical seals
Conventional packing
Threaded connections
Access ports
Viewing windows
0.002
0.005

0.005
0.018
0.032

0.10
0.10
0.25
0.015
0.020
0.015
Valves used to iolate system
Ball
Gate
Globe
Plug- cock
Valves used to throttle control
gas into vacuum system
0.02
0.04
0.02
0.01
0.25
aAssumes sonic (or critical) flow across the component.

Source:  Chemical Engineering, &&:7B, December 14, 1981
                               A-ll

-------
        TABLE A-6.
AVERAGE EMISSION FACTORS FOR FUGITIVE
   EMISSIONS IN SOCMIa
Equipment component
Pump seals
Light liquid
Heavy liquid
Valves
Gas
Light liquid
Heavy liquid
Compressor seals
Safety relief valves- -gas
Flanges
Open-ended lines
Sampling connections
"Average" SOCMI factors,
Kg/h/ source
0.0199
0.00862
0.00597'
0.00403
0.00023
0.228
0,104
.00183
0.0017
0.0150
aThese factors are appropriate for estimating emissions when no
 other data (i.e., leakage rates) are available.

Source:  EPA-953/R-93-026.  June 1993.
                               A-12

-------
    APPENDIX B.




CALCULATIONAL ISSUES

-------

-------
                           APPENDIX B.

                       CALCULATIONAL ISSUES
     This appendix contains calculational issues encountered
during the development of this document.  An examination of the
degree of saturation with VOC of a purge gas stream exiting a
vessel containing VOC, and a discussion of an incremental cost
analysis of manifolding single unit operations to a control
device is provided below.
     Calculational Issue 1;  Degree of Saturation of a Purgg   .
Stream.  The degree of sat Luxation Mas examined for purges of
quiescent vapor- liquid interfaces and agitated gas sparging.
Based on the results obtained from various mass transfer
correlations, the expected saturation fraction ranges from 0 to
100 percent for the range of conditions examined.  The
calculations show that typical batch purging  (at flowrates of 20
to 30scfm) over quiescent surfaces yields fractional saturation
values of less than 10 percent, whereas purging of agitated
sparging yields values of 80% or better.  The discussions below
present the theory and calculations relating to these findings.

  I.  General

     If the vaporized liquid is a single component, or for dilute
concentrations in a solvent, then the rate of mass transfer of
the liquid across the liquid-vapor interface will only be a
function of the diffusion through the vapor "boundary layer"
film.  The mass transfer is said to be "gas phase controlled."

     As the purge gas pd.e<=o over tha surface of the volatile
liquid, vapor will diffuse into the bulk of the gas where it will
mix by convection and eddy currents.  The driving force for
diffusion is the concentration difference of the interface  (which
is the saturation or equilibrium. concentration in the vapor) and
the bulk gas phase.  The resistance is the diffusivity of the VOC
in the gas.  The flux, or flow cf material across the interface
is the rate of vaporization and becomes the VOC content of the
exiting purge gas.  The flux (I) is related to the diffusivity
and the concentration driving force as follows:
                                   ) (yry)                    (B-i)
where :
         I - flux, gmol/m2 hr

        BT - the thickness of the boundary layer, m

        Dv - the diffusivity, rnVhr

        pm m the molar gas density, gmol/m3

     (y.-y) . the concentration difference in mole fraction.

                               B-l

-------
Since B.J. is not usually  known,  other forms  of  this equation  are
more convenient where a  term k  is  defined as the mass  transfer
coefficient, and  is empirically related  to  diffusivity,
viscosity, density, and  the  geometry of  the system.  The
governing equation then  becomes

                         I  -  N/A -  k(yry)                    (B-2)

where N is the number of moles  of  VOC transferred across  the
interface of area A per  unit of time, in the previous  units, N
would be in gmols/hr.

 II.  Inerj; ......... ff. ..... Pegging of  Quiescent  Solvent .............. Ppols

     The geometry of the headspace of a  storage  tank or reactor
with contents at  rest, where the purge gas  is moving across  a
quiescent surface of liquid,  resembles evaporation from an open
pool.  MacKay and Matsugu  determined an  empirical correlation for
the mass transfer coefficient for  evaporation from a pool where
one of the terms  is windspeed.   In the case of tank purging, this
would be analogous to the  superficial velocity of the  sweep  gas
across the liquid pool surface.

     The k value  calculated  by  the MacKay and Matsugu  correlation
has units corresponding  to the  following equation for  mass
transfer flux:

                      I  - N/A -  k[(pi-P)/RT]                 CB-3}

where k  m/hr in this example,  and  the  term [(p-P)/RT)  reduces
to pi/RT to maximize the concentration driving force,-  this term
is in units of gmols/nr .

     This equation is based  on  a moles per  volume concentration
gradient driving  force rather than a mole fraction (y^, y)
driving force.  The equation for k,  the  mass transfer
coefficient,  in m/hr,  is as  follows:

                    k -  .029 V-7BD~*'Ll$Sc~'6'7                (B-4)

where :


    NSc " tne dimensionless Schmidt number  which  relates  the
          diffusivity and  gas viscosity,   calculated to be 1.86

          for nitrogen at  these conditions.
      M  viscosity, g/m * hr (multiply Centipoise  [CP] by  3,600
          to obtain this value;
      ft  density g/m3;
                               B-2

-------
     Dv - diffusivity, m2/hr;

     U -  the wind velocity  (meters/hr) equal to the purge rate

          in m3/hr, divided by  [(.707)(tank diameter,m)  (vapor
          space height, m) ] ; and

     D -  the pool diameter  (meters).

     To use this equation  for evaporation in a tank head space,
the pool diameter was taken to be the tank diameter.

     The average velocity  across the surface of the liquid is
calculated to equal the flowrate divided by the term I (.707)
(diameter)(height)].  The  correction of .707 was calculated to
account for tank geometry.  It describes the average velocity of
the material'as it passes  through the average available cross -
sectional area.

     Figures B-l and 1-2 are the graphical presentation of the
results of using the MacKay and Matsugu correlation for
estimating the rate of vaporization of toluene into a nitrogen
purge gas stream for several different tank sizes and at two
different temperatures.    The composition of the exiting purge
stream was calculated by material balance; the percent of        -
saturation level is also shown.  Even for a low purge rate of
only 0.1 acfm across a small head space of an 8 ft diameter tank
(typical for 3,000 to 6,000 gal storage),  only 5 percent of
saturation is attained.

     Based on using the MacKay and Matsugu method it is clear
that purge gas streams are substantially below the saturated
level of VOC.   Therefore,  the assumption of saturation predicts
much higher VOC loadings to the control device than could
realistically be expected.

     In a variation from the MacKay and Matsugu approach, the
geometry of the vapor space of a purged tank may also be
considered to be similar to that of a wetted wall column where
the area of wetting is the surface area of the tank contents, and
the diameter of the column would become the effective diameter
(not the pool diameter)  of the cross-section of the tank head
space through which the purge gas sweeps at a calculated
superficial velocity.  Gilliland and Sherwood proposed the
following correlation for wetted-wall columns

                   Nsh - 0.023  NRe-81NSc44                (3-5)
where:

               Ngh  kD6M/pDv (NSc)     as defined above

               NRe - De
                               B-3

-------
0
I
            4.5
#  3.5

2

i   3
CD

I  2.5
o
uu    _
u.   2
O


1  1-5
o

I    1


   0.5
?100
                                                                        -^
                         1CK)
300        500       700

  PURGE FLOWRATE (acfm)
                                                          900
1100
                Figure B-l.   Mackay/Matsugu method for toluene at 25C,

-------
03
I
in
                            100
300       500        700

  PURGE FLOWRATE (acfm)
                                                                       900
1100
                                      DIA=SH.VSpace=10tt
                                                                   pw=8H-VS>il'
                Figure  B-2.   Mackay/Matsugu  method Cor toluene at 50C, VP-94.3n.nHg.

-------
     The mass transfer coefficient, k, is contained in the
Sherwood Number, Ngn, which relates equivalent diameter, D_? gas
molecular weight, M; gas density, p; and vapor diffusivity, Dv.
The Gilliland-Sherwood equation can be rearranged to solve
directly for k; the units are moles per area per time.

     The area of mass -transfer is the area of the surface of the
tank contents.  The driving force is the difference between the
vapor concentration at the interface  (assumed to be the
saturation value) expressed as a mole -fraction, and the bulk gas
composition (which is essentially sero) .  The rate of evaporation
is thus expressed as
                          N/t - kACy^y)                     (B-7)

Knowing the purge gas flow rate and evaporation rate, the
composition of the exit gas can be calculated by material balance
and compared to saturation composition.

     A series of calculations were done using the wetted wall
correlation for a storage tank of toluene being purged with
nitrogen.  Figures B-3 and B-4 are the graphical representations
of the data.  As with the MacKay and Matsugu approach, very low 
values of percent of saturation are obtained.  For example, a low
purge rate of 0.1 acfm across a small head space of an 8 foot
diameter tank results in only 7.3 percent of saturation.

     The calculated percentages of saturation equilibrium for
most flowrates and vessel vapor spaces using both the wetted wall
equations and the MacKay -Matsugu correlation yield low values, in
the range of 0 to <10 percent, with most values below 5 percent,
A conservative assumption for calculating the purge equilibrium
fraction, therefore, would be to assume 10 percent.  Note that,
as the superficial velocity increases, k increases.  As flow rate
is increased,  velocity also increases, but more inert gas is
introduced to the system, thereby decreasing the percent
equilibrium.

     He can consider both equations bounded by the realistic
superficial velocity across the liquid surface.  Figure B-5 shows
graphically the differences in values obtained for percent
equilibrium for the MacKay-Matsugu correlation versus the Wetted
Wall method.  As the vapor space decreases (increasing
superficial velocity), the percentage of equilibrium increases,
especially in the low purge rate range.  For vapor space values
of 0.5 feet or less, the equations begin to approach higher
values as the superficial velocity increases.  In most purge
situations, however, the vapor space above the vessel will be
greater than 0.5 feet and the assumed 10 percent equilibrium
fraction will be realistic.
                               B-6

-------
L-Z
 PERCENT OF EQUILIBRIUM (%)
H-
03
(D
CD
I
f

a
(D
rr
rr
(D
a

M
M
a

3
(D
("T
rr
0
a
H)
O
^
rr
O
C
n>
(D
in
n

<
H
to
vo
vo
1
*
i
II
3
y*
I
II
3

+
i
=
w
^
?>
II

s

I
II
i
<
II
S
i
M
3
J
5
a
II
i

i



^k
8-



CO
sJ
C
3]
0
m

?
o m
1S
i^
m
L
o-
o


CD
"
^^


A
o -^ ro co 4:
Dui-^cnrobicocnAu
i i i i i i i i

i iw w

>T f^V _-
x/ X^
/ >x'
f /?
/ //
/ / 
//
//
;
' /
1 I1
:
//
1 /,

/ /;
I
t
1
I
I









1






>
t

i-












-------
                               8-a
                              PERCENT OF EQUIUBRIUM (%)
n
BJ
I

>
rr
rr
(D
a


$
(t>
rr
rr
O
a
 rr
 O
 M
 c
 01

 (D
 o
 n


 T3
 tn
 o
 O
 II
 10
*
i
ii
3
           3
                           p        -^        ro
                           en   -*   en   ro   en   co
                    O
                    O

-------
00
I
VO
                 -100
100
300        500        700
   PURGE FLOWRATE (acfm)
900
1100
                              -X- M/M VSpace=4B  -I- M/M VSpace=2lt
                              il WW VSpace=4fl  -X- WW VSpace=2fl
                                 M/M VSpaca=0.5ft
                                 WW VSpaca=0 Stt
        Figure B-5.  Mackay/Matsugu -Wetted Wall" toluene at 25C=29.9  mmHg, diameter=8 ft

-------
     t In the case of mixed liquids  (two or more components) the
 estimation of the mass transfer rate is somewhat more complex as
 it requires calculation of both the gas phase mass transfer
 coefficient, as explained above, and a liquid phase mass transfer
 coefficient.  The liquid phase coefficient takes into account the
 rate of diffusion of the more volatile component through a film
 of the less volatile component, to the vapor-liquid interface
 where it can evaporate.

      Furthermore, with binary or multicomponent liquid mixtures
 the mass transfer driving force is no longer simply expressed as
 the pure component vapor pressure divided by the total pressure
 The equilibrium partial pressure is the driving force, but the ";
 calculation of that term is related to the liquid composition
 The simple correction factor implied by Raoult's Law is to   "   '
 multiply the vapor pressure by the mole fraction in the liquid
 However,  in many real situations (e.g.,  dilute aqueous solutions
 of sparingly soluble organic solvents),  the materials are hiqhlv
 nonideal,  and application of Raoult's Law leads to substantial
 under estimation of the equilibrium vapor concentration of minor
 components.

      The use of  empirically determined  Henry's Law constants  for
 the estimation of vapor phase concentrations is a practical way 
 to approach calculation of  a realistic  mass  transfer driving
 force.  This methodology is  described in  previous sections   In
 extreme cases of nonideality,  the  partial  pressures of a mixture
 of compounds is  greater than the vapor  pressure of  any of the
 pure  compounds.   This phenomenon is  more  readily  observed where a
 mixture forms  a  low-boiling  azeotrope.  Obviously,  in such cases
 application  of empirical  correlalions  (such  as  Henry's Law) is
 necessary  for  an accurate calculation of driving  force.

      But,  in the case of purging of  mixed  liquids a second
 component  in the liquid phase  seriously decreases the  value of
 the liquid mass  transfer coefficient.  Therefore, although the
 actual mass  transfer driving  force may be  somewhat  greater than
 estimated  for a  pure component, the  overall mass  transfer rate
 will not be, and the use of  the fractional approach to
 equilibrium will be valid.

 Ill- Inert N2 Purging of Agitated Vessels

     A.   Sparging

     Agitated vessels can be evaluated in the same  way using mass
transfer coefficients for stirred tanks.  The coefficient  of mass
transfer through broken interfaces during sparging  is given by
the equations:
                              B-lO

-------
                                                             (B-8)
                                   (fV)0.25
                       kL  = 1.0  -'
                                      0.67
                                   NSC
 where :
                                                            (B-9)
                            Tf      DT    HL


     kL  mass transfer coefficient, m/s;

      V - m/s;

    Ngc  Schmidt No. of the sparge gas;

      N - impeller speed, RPS;

     Di - impeller diameter, m;

     YN - impeller power factor;

     DT  tank diameter, m; and

     HL - liquid depth, m.


     Notice that mass transfer at this interface is liquid-phase
controlled.

     The amount of mass transfer that can occur is a function of
the characteristics of the agitation scenario, including impeller
size and speed.  As an approximation, the tank with a five- foot
diameter and a 6 f t . vapor space height containing toluene at
25 C was assumed to have an agitator with a 1 ft diameter
impeller rotating at 0.5 revolutions per second.  In absence of
real data the power number was assumed to be 1.6, corresponding
to a typical value in solid- liquid dispersion.

     Using the sparge velocity as in other examples for purge
velocity (i.e., flowrate divided by area of flow), much higher
values of mass transfer coefficients are obtained.  Saturation
values for low sparge rates are considerably higher.  For
100 acfm,. the expected fraction of saturation was calculated to
be 30 percent.  Figure B-6 illustrates the values of percent
equilibrium verses purge flowrate for this agitated sparge
system.   Table B-l is the data in tabular form.
                               B-ll

-------
tu
i
H
to
         I
         D

         CT

         HI
8

-------
                  TABLE  B-l.   PURGE  EQUILIBRIUM CALCULATION RESULTS FOR FIGURE B-6.
FtePURQEOAS



Mw Trwwtat Ttwough Broton MMtaCM injIlHiif Puro*|
                                                                                                    01 OK 92
W
i
H
U>
WMM! GonHQunfloft*
OMU.
vhptjMM
Agitator Poww
Puf(l Super.
Flow nut VWoc.
ACFM
01
1
10
2O
30
40
90
0
70
0
90
100
200
300
400
500
1000
VMC
0.0001
ooooo
00079
0.0197
0023*
00314
003*3
00471
00560
00626
00707
00766
0.1S72
0.2387
O3143
0.3626
0.7096
9

8V~0.78
w/hf
01478
08010
9.3668
 2188
126482
19.6300
188386
21 7186
244039
27.1022
2*. 797B
323500
909402
76.2130
66.3693
1139167
164.6262
PracMcUqutd
R NMM: TOLUENE
II Twnp 29
VP 206
NGc 1.66
0*
"Oil NSc N3cfl kn
m Hollo
0859
0859
4859
0.899
0.699
0.699
0.699
0.699
0699
0699
0.699
0.699
0699
0.699
0.699
0699
0699
492
492
.492
492
492
.492
.492
.492
492
.492
492
.492
.492
.492
.492
492
.492
~')67
1 ?84
1 264
1284
1 284
1284
1.264
1.264
1264
1284
1264
1.264
1.264
1.264
1.264
1264
1.264
1.264
m/tr
O002'
00161
00867
01683
0.2325
02810
0.34ft!
03682
0.4902
04687
09477
0.9647
1.0211
1.4006
1 7934
20687
3.9632
D*gC
mmHg
1
IffloV
inZ/hr
80828
14 3361
254872
30 1214
335561
380584
38 1271
388052
41 4731
42.6808
441824
493411
938166
96.6722
641216
67.6007
60.6281
N
xnol/
hr
00324
00676
01025
Oi219
01348
Q145O
01633
01604
01687
01724
01779
01623
02166
02386
02976
02726
03241
bnfwllw Oiam (m):
POWM Number:
AQNSfMd|r|M):
Purg y
IbmoV @xM
hr
001M
0 1531
1 5312
30623
49839
8 1247
7.6958
61870
107162
122484
137809
193117
306234
496391
612466
769969
193.1166
molfr
211887
037644
006884
003B8G
002837
002367
O.O2002
001748
0.01996
001407
001286
0.01160
0.00706
000922
000421
000396
000212
0.3048
1.6
0.9
y
quN
molfr
003834
003834
003834
0.03834
003834
003634
0.03634
0.03834
0.03834
OO3634
O.O3634
003634
003634
0.03634
0.03634
003634
003834
Pwcwil
of
Equi
5381
897
170
101
79
60
91
44
40
36
33
30
16
13
11
6
9
Liquid D*M:
Pwcvfit EfMMon
of
Stafeii
6781855
273W79
627*138
382*888
26-5266
2.312019
1.062716
1.716161
1.931676
1.367736
1.271601
1.1764
0.702644
0916907
0.416106
0394791
0.21124
(KI2/WC3)
6.286-07
828E-07
828E-07
6.26E-07
6.26E-07
620E-07
6.26E-07
6.29E-07
6.26E47
626E-07
6.26E-07
6.26E-07
6.26E-07
620E-07
6.26E-07
626E-07
6.26E-07
3
k
/!

0.001363
0002477
0004404
0005236
OOO9786
0.006226
0.000966
0.006663
0.007164
0.007407
0.007626
0.007632
0.006314
0.010306
0.011076
0.011712
0013626
k

-------
i
M
*.
           #
           oc
           CD
O
UJ
u.
O
            UJ
            O
            DC
            UJ
            Q.
9j
8
7-
6
5
4
3
2
1
                 0    100   200   300   400   500   600  700   800   900  1000
                                    PURGE FLOWRATE (acfm)
               Figure B-7.  Non-Sparge Agitated purge typical impeller scenario.

-------
     B.   Agitated Purging

     Mass transfer at gas -liquid agitated interfaces during
nitrogen purging was also examined using the following liquid
phase mass transfer coefficient:
                             .0256
Figures B-7 and B-8 are the results of examining two different
agitation scenarios.  The first scenario, labelled the "typical"
impeller scenario, is identical to the sparge impeller example
discussed previously.  It considers the use of a 1-foot diameter
impeller rotating at .5 revolutions per second.  YN, the power
factor, is 1.6.  The second scenario, presented in Figure B-8,
considers the use of a 2 -foot diameter impeller, rotating at 2
revolutions per second.  The power factor also is higher, at 6.6.
This scenario is termed "worst case", as an approximation of the
maximum turbulence encountered during such a situation.  Notice
that values approaching 80% saturation are shown corresponding to
typical purge flowrates of 20 to 30 scfm in Figure B-8 (data in
Table B-3) .  Saturation values are much lower for the typical
impeller scenario in Figure 1-7 (data in Table B-2) .

 IV.  Conclusions

     The degree of saturation with VOC of a purge gas stream
exciting a vessel containing VOC is highly dependent upon
specific vessel geometries and liquid-vapor interface conditions.
Values approaching complete saturation are not unrealistic for
systems utilizing severe agitation or sparging, while much lower
fractional saturation levels are expected for non-agitated
purging events.

     In order to provide a conservative, yet realistic approach
to estimating the degree of saturation of an inert gas purge, the
following guidelines are recommended:

     1}   for purge flowrates less than 100 scfm, assume that the
          vent streams exiting streams are completely saturated
          with VOCs.

     (2)  for purge flowrates greater than 100 scfm, assume that
          the vent streams exiting the vessel are 25% saturated
          with VOCs.
                               B-15

-------
OS
                      100200   300  400  500   600  700   800900   1000
                                  PURGE FLOWRATE (acfm)
             Figure B-8. Non-Sparge Agitated Purge worst case impeller scenario.

-------
                  TABLE  B-2.   PURGE EQUILIBRIUM CALCULATION  RESULTS FOR FIGURE B-7.
                                                                                                       ot
      MM* Tramfer Through Breton

      WWKMUQM Sparging: Typtcal
(Agitated Putg)
td
I
VMMlConMguraMan:
Own.
Vp8p*M
Agitator Poww
Purge Supw.
FtowM* VWac.
ACFM
0.1
1
10
20
30
40
60
0
70
0
00
too
200
300
400
600
1000
WMC
0.0001
0000*
0007*
00167
0023*
0.0314
003*3
0.0471
00060
00*0*
0.0707
0.079*
01672
02367
03143
0.3929
0798*
6
8V~07*
m/hr
0147*
08010
63888
02180
126482
15*300
1*.*3M
21.71**
244939
271922
29.7979
32.3900
6654*2
7*2130
96.3993
113.91*7
1*4.9292
II
ft
Dt
"Oil
m
0.959
0935
0955
0*55
0956
0906
0956
0965
0.895
0955
0955
0959
0.099
0*55
0*55
0.999
0*55
PTOCM* Liquid:
NMW: TOLUENE
Twnp 25 OegC
VP 200 mmHg
NSc 188
1
NSc NScR km gmol/
Rfttto ~067
1.492
1492
1.492
1.492
1.492
1.462
1.492
1.492
1.492
1.452
1.452
1.492
1.492
1.462
1.492
1.492
1.492
.2*4
.2*4
.2*4
2*4
.2*4
.2*4
.2*4
2*4
.2*4
.2*4
2*4
.2*4
2*4
2*4
2*4
.2*4
2*4
m/hr
00027
00164
0.0087
0 1885
02325
02010
03483
03002
04502
04007
05477
05*47
1.0211
1.400*
1 7534
208*7
35*32
m2/hr
1.2201
1.2201
1 2201
12201
12201
1 2201
12201
1.2201
1.2201
1.2201
1.22*1
122*1
122*1
1.22*1
122*1
1.2201
1.2201
N
IbmoV
hr
00040
00040
00040
0004*
0004*
0004*
0004*
0004*
0004*
0004*
00049
0004*
0004*
0004*
0004*
0004*
0004*
Poww Number.
AgNSp**d(rp*)
IbmoV 0xtt
hr
00153
01531
1 5312
30823
45*35
8.1247
78558
8 1870
1071*2
122494
137805
15.3117
30*234
45*351
1240*
7*99*9
153.11*9
modi
03228*
003227
000323
0001*1
00010*
0000*1
000085
000054
000048
000040
000038
000032
0.0001*
000011
0.0000*
00000*
0.00003
0.304*
1.8
OS
quH
molfr
003934
003034
003034
003034
003034
OO3034
OO3034
OO3034
003*34
O 03034
003*34
0.03*34
0.03934
003*34
003*34
003*34
0.03*34
Pwcont
of
Equd
820
82

4
3
2
2
1
1
1
1
1
0
0
0
0
0
Liquid D*p
Percent
of
Streem
243*851
3.12002*
0.321*52
0.1*1096
0.10744*
0.0*0*0*
006449*
0.063793
0.04*077
0.04032
0.039*42
000225*
0.01*132
0010769
000*0*7
000*493
0.003227
Hi (m):
EpcHon

-------
                TABLE B-3.   PURGE EQUILIBRIUM  CALCULATION RESULTS  FOR FIGURE B-8.
     FNrPURGEGAS
                                                                                              01-D*c-B2
W
I
-
00
MM irarwi
Without GM
1
1
Agitate POM
Pl*0*
Flow MB
ACFM
0.1
1
10
20
30
40
SO
60
70
0
00
100
200
300
400
900
1000
BpMObif.:
XMK.
tapSjMC*
m
Super.
\Moc.
WMC
0.0001
0.000*
0.007*
0.0167
0.023*
0.0314
O.OM9
0.0471
00000
00*2*
0.0707
0078*
01872
O.tSST
0.9143
0.3J2*
0.7*5*
Won! CM*
5

mfi*
0.147*
0.8810
6.3*8*
 21*8
12*4*2
15.1300
188306
21.71*8
24.4*35
27.1*22
28.7878
32-3500
86.54*2
70.2130
883883
113.61*7
18482*2

II
8.
0*
~-0.11
m
0.055
0855
0889
0.855
0869
0859
0895
0.865
0888
0865
0885
0.055
0*65
0055
0.805
0.855
0895
arufv
rootMlk
M
T
W
M
NSc 1
RMo -
1.452
1.452
1.452
1.452
1.452
1.452
1.452
1.452
1.452
1.452
1.452
1.452
1.452
1.452
1452
1452
1.452
I
td:
MM: 1
up
i
>c
ttcfl
067
.2*4
.2*4
.2*4
.2*4
.2*4
.2*4
2*4
.2*4
.2*4
.2*4
.2*4
.284
.2*4
.2*4
2*4
.2*4
.2*4

FOtUENE
25
20.*
1*8
km
m/hr
O.OO27
0.01*4
0.0887
01689
0.2325
0.2310
0.3463
0.3882
0.4502
0.4087
0.5477
0.5847
10211
1.4X0
1.7534
2.0887
3.5*32

D*gC
mmHg
1
gmoV
m2*r
24.3154
24.3154
24.3154
243154
243154
243154
243154
243154
24.3154
243154
243154
24.3154
24.3154
243154
243154
243154
243154

N
tomoV
hr
00877
00877
00877
00077
0.0877
0.0877
0.0877
0.0877
0.0877
00877
0.0877
00877
00877
O0877
00877
00877
0.0877
ImpttUiDU
Power Num
AgKSpMd
txnoV
hr
00193
01931
1 5312
3O623
45835
6 1247
76596
81670
107182
122484
13.7005
153117
30.6234
45*351
61246*
7655*5
153.1166
m (ml
 ii liii)*
b*
(IP*
e
-------
raTrulational laaue 2:	Incremental Coat Analysis of Manjfp]r?ing
Single Unit Operations to a Control Device

  I.  General

     The incremental cost effectiveness of manifolding single
unit operations to a control device was examined.  Ductwork
diameter for an emission source to a control device was estimated
by assuming an average surficial velocity through the duct to be
2,000 ft/min.  The duct costs are based on stainless steel,
circular duct prices.  The analysis describes ductwork cost as a
function of length of ducting and emission source flowrate.
Calculations are provided below.  The results of the analysis
show a minimum level of 500 pounds per year of VOC emissions is
necessary to yield an incremental cost effectiveness comparable
to the average cost effectiveness of RACT.  This is the deminimis
level for applicability of RACT to any single unit operation.
Figure B-9 shows the incremental ducting analysis results for
0 to 300 feet of duct versus mass emissions.  Table B-4 is the
cost analysis data.
Calculations

     Assume:
                    Velocity -  2,000 ft/min

                    Flowrate -  (Area)(velocity)

                    Flowrate   Ur2) (2,000 ft/min)

                    Flowrate     _ d2
                                 7T
                  2,000 ft/min     4
                   d(ft) - v/v3.CCS4) (Flowrate)
                               B-19

-------
w

to
o
         1500
         1250
         1000
         750
500
          250
             Incremental Ducting Analysis
CO
c
o
'to
to

E
LU

c
c
                            100         200

                             Length of pipe (ft)
                                       300
                  Fr=10acfm
                    Fr=20acfm
                                    Fr=50acfm
                Figure B-9.  Incremental Ducting Analysis Results

-------
                                          TABLE B-4.   COST ANALYSIS DATA
W
        Cos! Analysis - ManNoMng Co* Angsts
Ftownfte
(dm)
     10
     10
     10
     10
     20
     20
     20
     20
     50
     50
     50
     SO
  0
100
200
300
  0
100
200
300
  0
100
200
300
FOB Cost
Tol*
(June 92$)
1174385
1430529
1686.674
1942818
1182414
1544657
19069
2269143
1198346
1771 102
2343859
2916615
TCI

(June 1992)
2009372
2447636
2885899
3324162
2023111
2642908
3262706
3882504
2050369
3030356
4010342
4990328
Annualzed

-
407.4002
496.2581
586.116
6739738
4101857
5358497
661.5137
7871777
4157124
614.4046
813.0969
1011 789
MO Cap


0.2037
0248129
0292558
0336987
0205093
0.26792S
0330757
0393589
0207856
0307202
0406548
0505895
Lbs


448.6787
546.5398
6444009
7422619
451.7464
5901428
7285393
8669358
457833
6766571
8954811
1114.305

-------
Costs
    Manifold costs include ductwork, damper, and elbows.  Below
are the costs for the individual components for a manifold.
Stainless Steel Round Duct
    A 0.25 inch-thick stainless steel duct is the first component
of the manifold.  The cost of this duct is based on the amount of
steel required, as decided in Chemical Engineering magazine.
            Volume -  27rRL(t)
                   -  (2TT) (D/2) (L) ('25/12)
    Specific
 Gravity Stainless -  .291b/in3
                  _ / .2SnDL\ ( .29Ib \ 112inches\3
                  "\  12   /Unches3J\    ft   /
    Ibs steel      -  (32.8) (D) (L)
where,
                     D * ^(0.00064)(Flowrate)
So,
                  $/ft = (32. 8)^0.00064 Flowrate)
    The price of stainless steel is $1.03/lb
                   -   (32.8) ($1.03/lb) (0.0253) (Flowrate)-5
                   -  0.85  (Flowrate)0'5
    This cost is adjusted to June 1992 dollars by the  ratio  of
indices of 359-6/376.3.6
              $/ft -  0.81  (Flowrate)0-5
    Damper:
    The cost of a stainless steel circular damper was  estimated
from a graph in the EAB Cost Manual.7  Several points  were taken
                               B-22

-------
from the graph of diameter of damper, versus dollars.  These
points were:

                    D(in)               

                    20                  2,300
                    30                  3,200
                    40                  4,000
                    50                  5,000
                    60                  6,300

    A regression line was developed from this data.

                 Y   98(D) + 240

    This line was multiplied by a factor, 3, to get the cost for
a stainless steel damper, in accordance with the referenced
manual.

                 Y -  294(D) + 720

    Cost was adjusted to June 1992 dollars  from December  1978
dollars using appropriate indices.8

                                   /35S.7      \
                 Y -  294(D) + 720 ^     /223.7J

       Y($/damper) -  471(D) + 1,155, where D is in inches
                    D(ft) = v'O. 00064 Flowrate,
So,
             $/damper = (47l)J   i264 (Flowrate) +1,155


                     -  (3.45)(Flowrate)0'5  +  1,155
Elbows
    Two elbows are assumed  to  be  needed for each source.  Costs
were based on Chemical  Engineering Magazine article.

                      $/elbow -  (0.81)(1.65)(Flowrate)0-5

                             -  1.34 (Flowrate)0-5
                               B-23

-------
    Total cost of manifold  is  therefore:
         [.81  (FR)-5]   (Feet  of  Duct)
        + 1.34  (FR)'5  (2)  + 3.45 (FR)-5  +  1,155
                               (Free on Board)
FOB  Prices are corrected  to  Total  Capital  Investment (TCI)  using
the following elements:
Inst, Sales Tax, Freight  -  (.18)(FOB$)
(PEC) -  (1.18 FOB)
                    Purchased Equipment Cost
Direct Costs:
             .9
               Foundation
               Handling
               Electrical
               Piping
               Insulation
               Painting
(.08)
(.14)
(.04)
(.02)
(.01)
(.01)
(PEC)
(PEC)
(PEC)
(PEC)
(PEC)
(PEC)
Indirect Costs :
    Engineering
Construction and
   Field Expense
TCI    (FOB) +
TCI   FOB  (1 +  .18
TCI -  1.711 FOB
 -   .3  (PEC)

)

 -   (.10)(PEC)

 -   (.05)(PEC)

 -   .15(PEC)

  (.18)(FOB)
                              +  (.3) (1.18) (FOB)
                          .354 +  .177)
Indirect
Annual Costs11
Admin.
Prop. Tax
Insurance
Cap Rec.
                              (.15)(1.18FOB)
                                 2% TCI
                                 1% TCI
                                 1% TCI
                                 (.16275)  (10 yrs,  10%)
                               B-24

-------
REFERENCES FOR APPENDIX B
l.  Hawg, S. T.  Toxic Emissions from Land Disposal  Facilities,
    Environmental Progress. 1:46-52.  February 1982.
2.  Makay, D., and R. S. Matsugu,  "Evaporation Rates of Liquid
    Hydrocarbon Spills on Land and Water, " Can. J. Chemical,
    Engineering. 51:434 (1S73).
3.  Gilliland, E. R., and T. K. Sherwood, Industrial and
    Engineering Chemistry. :516  (1934).
4.  Agitation in Multiphase 3y'cems, Hydrocarbon Processing.
    July 1979.
5.  Chemical Engineering,  May 1990, p. 127.
6.  Chemical Engineering,  September 1992, cost index.
7.  Neveril,  R. B.,  Card,  Inc.  Capital and Operating Costs of
    Selected Air Pollution Control Systems.  EPA450/5-80-002.
    December 1978.
8.  Reference 6.
9.  OAQPS Control Cost manual, EPA 450/3-90-006, January 1990.
10.  Reference 10.
11.  Reference 10.
                               B-25

-------
    APPENDIX C.




SAMPLE CALCULATIONS

-------

-------
Example 1.
Example 2,
Example 3.
Example 4.
Example 5.
Example 6.
Example 7.
Example 8.
Example 9.
Example 10.
Example 11.
Vapor displacement of a single component liquid
Vapor displacement of a homogenous mixture
Tank/reactor heatup losses
Empty tank and reactor purging
Pilled tank and reactor purging
Sparging volatilization
Vacuum dryer emissions
Atmospheric dryer emissions
Vessel Depressurization
Emissions from a steam ejector
Emissions from equipment leaks
                               C-l

-------
      f 1.   Vapor displacement of a homogenous liquid

     A 5,000-gallon reactor is to be filled at ambient conditions
(25CC and 1 atm) with 3,600 gallons of benzene.  The fill rate is
60 gallons per minute and the reactor vent is open to the
atmosphere.  Calculate VOC emissions from this event.

     Solution
     Step 1.  Define the conditions of the displaced gas:

     Temperature - 298K  (25C - ambient)

     Pressure - 1 atmosphere  (760 mmHg, 14.7 psia)

     Volumetric rate of displacement - 60 galIons/minute

     Step 2.  Calculate the vapor phase mole fractions of the
components in the displaced gas:

     In this situation, benzene is the only component in the
liquid, therefore x in equation 3-9 is 1.

     Using Raoult's Law:

                                     x-^P*

                                yi - 
where:
      X
     P*  (vapor pressure of benzene at  25C  [77F])  -  1.9  psia,

      (From Table A-2.)

     PT  - 1 atm  (14.7 psia)
              U1.9  psia)  _  0
      yi    (14.7  psia)
Therefore,  the  gas  in the vapor space will be 13 percent by
volume benzene.

      Step 3.  Calculate the emission rate:

           (Yi) (V) (PT) (Mw)
      ER  *     (R) (T)
                                C-2

-------
      (0.13) (60 gal/min)  (1 atm)  (78 Ibmol)  I		r]
ER =	J7.48 gal|
              (1.3144 atm ft3/lbmol K) (298K)
     ER - 0.21 Ib/min benzene
Since there are 3,600 gal of benzene to be charged, the event
will take
       3,600 gal    ..
       60 galAnin - 60 min
Therefore, total benzene emissions for this event are:
     (0.21 Ib/min)(60 min/event) - 12.6 Ib/event
                               C-3

-------
Example 2.  Vapor displacement of a homogenous  (miscible) mixture

     A 50-50 volume percent solvent mixture of heptane and
toluene is charged to a surge tank at the rate of 300 gal/min.  A
total of 1,500 gal is charged.  The mixture temperature is 20C.
Calculate emission rates for both mixture components.

     Solution
     Step 1.  Define conditions of the displaced gas:

     1.  Temperature of displaced gas:  20C;

     2.  Pressure - 1 atm  (14.7 psia, 760 mmHg); and

     3.  Rate of displacement - 300 gal/min.

     Step 2.  Calculate vapor phase mole fraction:
voc
Heptane
Toluene
Molecular
weight,
Ib/lbmole
100
92
Density,
Ib/gal
5.7
7.3
Gallons
charged
750
750
Pounds
4,275
5,475
TOTAL
Ibmoles
42.8
59.5
102.3
xi
0.42
0.58
1.0
     P* heptane 9 20C  (68F) - 0.7 psia

     P* toluene  20eC  (68F) - 0.4 psia
            xi(P*)
      Heptane:
      Toluene:
(0.42) (0.7  psia)
   (14.7  psia)
                     020 - V
                    -020   yheptane
(0.58)(0.4  psia)
   (14.7  psia)
                   0.016
^toluene
                                C-4

-------
       Step 3.  Calculate  emission rate:




      ER
        heptane
             (0.020)  (300 gal/min) (1 atm)  (100 Ib/lbmol)    ft    I
ER         -	|7.48 gal|

  heptane               1.3144 atm ft3/lbmol K)  (293K)
      E_        - 0.21 Ib/min
        heptane
                   (y         (V) (P  ) (M  )
                    toluene)     T  w
      E         - 	
       R                   R
        toluene             T



             (0.016)  (300 gal/min) (1 atm)  (92 Ib/lbmol) '   ft
    ,       _  -*8 gal
  toluene                   c             -a i
                                       ft
                                      v
                            [    Ibmol K    J


      ER        -0.15 Ib/min
        toluene

Therefore, total emissions  for the event

     Heptane:   (0.21 Ib/min) (5 min)  - l Ib

     Toluene (0.15 Ib/min) (5 min)   0.75 Ib
                               C-5

-------
Example 3.  Tank/reactor heatup losses

     A 2,000 gal reactor, 75 percent full of a solution of a raw
material in toluene is heated from 20C to 70C.  The reactor is
vented to the atmosphere during the heatup; how much toluene will
be emitted?

     Solution

     Since the liquid is mostly toluene, a simplifying assumption
is that the partial pressure of toluene in the headspace is equal
to the vapor pressure.  At. 20C, the vapor pressure of toluene is
22 mmHg; at 70C it is 200 mmHg.  The head space of the reactor ;
is 500 gal or 66.8 ft3.  The temperatures must be expressed in  '
absolute units K.  The gas constant, R, in appropriate units
{from Table A- 3)  is 998.9 mmHg-f t/lbmol- K.
toluene emitted is then directly calculated:
                                              The weight of
                66.8  ff
            998 9
                '
                        ft
                   Ibmol K
                              760-22 mmHg
                               (273*20)K
760-200 mmHgu
  (273*70) K  JJ
                 0.0592 Ibmoles non-VOC gas displaced
             22 mmHg\  /  200 mmHg
          0.01195 lbmoles toluene
          1.06 Ib      toluene
                                   (92.13 Ib toluene/lbmole!
                               C-6

-------
Example 4.  Empt;y t;ank and reactor purging

     A 2,000 gallon reactor vessel was cooled to 20C and the
contents, in acetone solvent, were pumped out leaving only
vapors.  If this vessel is then purged with 1,000 scf of nitrogen
at 20C, how much VOC  (acetone) will be contained in the vented
nitrogen?

     Solution

     At 20C the vapor pressure of acetone is 182 mm Hg.  Thus,
the initial concentration can be calculated from Ideal Gas Law:

                           PV - nRT

                          n/V  P       /RT

     Concentration of acetone - pacetonel%'/'RT

                           MW  58.08 Ib/lbmol

                            R - 998.9 mmHg ft3/lbmol K

                            T  273+20 - 293K                   " '

     P        m             182 mmHg  (partial pressure of acetone
                            equals vapor pressure,  since acetone
                            is the only component)



c      (182 mmHg)(58.08 Ib/lbmol)   _ Q.036 lb/ft3
 1   (998.9 mmHg ft 3/lbmol K) (293K)


(Cj - Initial concentration in the reactor vessel)

The number of volume changes of inert gas is as  follows:

[1,000 scf][(273+20)/273J - 1,073 acf

(2,000 gal)(ft3/7.48 gal) - 267 ft3
1,073/267  4  (vessel volume changes  - 4.0)

Plugging the values back  into equation 3-14 yields:

Cf/Ci "  (0.37)4*  - 0.0187 lb/ft3

Thus, Cf  0.0187(0.036)  - 0.000673 lb/ft3

Emissions -  (vessel volume)(C^-C*)
          - 267  ft3  (0.000673  lb/ft3)
          - 9.43  Ib
                                C-7

-------
Example 5.  Filled Tank and Reactor purging

     A tank containing methancl at 25C is purged with a 30 scfm
stream of nitrogen.  Calculate the emission rate of methanol
during the purge.

     VPMEOH at 25c * 128

       .    128   , >20
         760-128
 (30 scfm)  (  *mole^ ) f. 20 moles MEOHJ/ 32 Ib \ . >5 lbs/min mou
          \359 scfm/ I   Ib mole N2   I \l mole/
                               C-8

-------
        6.  Calculation of Sparoino Volatilization

     A 1,000-gal tank of wastewater containing 0.025 wt% toluene
is to be air sparged to remove the toluene to a concentration
level of less than 20 ppb (by weight)  to permit discharge to a
municipal sewer system.  Ambient air is to be used; the design
temperature is 20C.  Toluene-water vapor-liquid equilibrium at
20eC can be approximated using a Henry's Law constant of 370 atm
(Henry's Law constants are listed in the Appendix).

     Approach;  Use l minute time slices, assume a sparge rate,
calculate time required to achieve concentration objective,
adjust sparge rate until reasonable cycle time is calculated.
Because of standard geometry of 1,000-gal tank, and modest gas
rates, 100 percent of equilibrium concentration can be assumed.
Table C-l summarizes the results of the calculations made using a
personal computer spreadsheet program.  With 7S acfm of sparge
gas, the desired concentration of 20 ppb toluene is achieved in
55 minutes of sparging.  The table clearly shows that the bulk of
the VOC is removed during the early part of the cycle:  one-half
of the total toluene is removed in the first 3 minutes, and
90 percent is removed after 13 minutes.  This typical
concentration profile for batch sparging makes the selection of
control technology  (described elsewhere in this report) somewhat"
challenging.
                                C-9

-------
TABLE C-l.  SPARGING VOLATILIZATION
Tnk Volume 1000 Gal
Moles of H20 462.26 Ib-nol
Sparge Rate 75 acfm
0.19465 ttmol/nin
System Temp 20 Deg C
Pressure 760 MI Hg
Dissolved VOC Toluene
Initial Conc'n 0.02SXtrt X
Moles of VOC 0.02260147
O.OOSXaol X
HU of VOC 92.14
Henry's Const. 370 atn
Exit Gas Equil. 100. OCX
Time Slice 1 win

Tim* Increment 0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
IB
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38

0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
9
8
7
5
5
4
3
3
2
2
1
1
1
1
9
7
x-bulk
.000048
.000041
.000034
.000029
.000024
.000020
.000017
.000014
.000012
.000010
.000008
.000007
.000006
.000005
.000004
.000003
.000003
.000002
.000002
.000001
.000001
.000001
.000001
.941-07
.396-07
.096-07
.986-07
.056-07
.266-07
.606-07
.046-07
.566-07
.176-07
.836-07
.546-07
.306-07
.106-07
.286-Ofl
.846-06

0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
y-exit
.018090
.015271
.012892
.010883
.009188
.007756
.006548
.005527
.004666
.003939
.003325
.002807
.002370
.002000
.001669
.001425
.001203
.001016
.000857
.000724
.000611
.000516
.000435
.000367
.000310
.000262
.000221
.000186
.000157
.000133
.000112
.000094
.000080
.000067
.000057
.000048
.000040
.000034
.000028
mo Is out cum out
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
.003521
.002972
.002509
.002118
.001788
.001509
.001274
.001076
.000908
.000766
.000647
.000546
.000461
.000389
.000328
.000277
.000234
.000197
.000166
.000140
.000119
.000100
.000084
.000071
.000060
.000051
.000043
.000036
.000030
.000025
.000021
.000018
.000015
.000013
.000011
.000009
.000007
.000006
.000005
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
.003521
.006494
.009003
.011122
.012910
.014420
.015695
.016771
.017679
.018446
.019093
.019640
.020101
.020491
.020819
.021097
.021331
.021529
.021696
.021837
.021956
.022057
.022141
.022213
.022273
.022324
.022368
.022404
.022435
.022461
.022482
.022501
.022516
.022530
.022541
.022550
.022558
.022565
.022570
new x-b
0.000041
0.000034
0.000029
0.000024
0.000020
0.000017
0.000014
0.000012
0.000010
0.000008
0.000007
0.000006
0.000005
0.000004
0.000003
0.000003
0.000002
0.000002
0.000001
0.000001
0.000001
0.000001
9.946-07
8.396-07
7.096-07
5. 986-07
5.056-07
4.266-07
3.606-07
3.046-07
2.566-07
2.176-07
1.836-07
1.546-07
1.306-07
1.106-07
9.28E-08
7.84-08
6.62E-08
wt fr
2.116-04
1.786-04
1.SOE-04
1. 276-04
1.076-04
9.056-05
7.646-05
6.4SE-OS
5.446-05
4.606-05
3.886-05
3.286-05
2.776-05
2.336-05
1.976-05
1.666-05
1.406-05
1.196-05
1.006-05
8.456-06
7.136-06
6.026-06
5.086-06
4.296-06
3.626-06
3.066-06
2.S8E-06
2.186-06
1.846-06
1.556-06
1.316-06
1.116-06
9.356-07
7.896-07
6.666-07
5.626-07
4.756-07
4.016-07
3.386-07
Percent
fteaoval
15.580X
28.733X
39.837X
49.210X
57.123X
63.804X
69.443X
74.204X
78.223X
81.616X
84.480X
86.898X
88.939X
90.663X
92.118X
93.346X
94.382X
9S.258X
95.996X
96.6202
97.147X
97.591X
97.967X
98.283X
98.551X
98.777X
98.967X
99.128X
99.264X
99.379X
99.475X
99.557X
99.626X
99.684X
99.734X
99.775X
99.810X
99.840X
99.865X
                C-10

-------
             TABLE  C-l.    (Continued)
39 6.626-08 0.000024 0.000004 0.022S7S  5.586-08 2.866-07    99.886X
40 5.586-08 0.000020 0.000004 0.022579  4.71E-08 2.4K-07    99.904X
41 4.71E-08 0.000017 0.000003 0.022583  3.986-08 2.046-07    99.919X
42 3.98E-08 0.000014 0.000002 0.022585  3.36E-08 1.72E-07    99.931X
43 3.366-08 0.000012 0.000002 0.022588  2.84E-08 1.456-07    99.942X
44 2.846-08 0.000010 0.000002 0.022590  2.39E-08 1.221-07    99.95U
45 2.39C-08 0.000008 0.000001 0.022592  2.02E-08 1.036-07    99.959X
46 2.02E-08 0.000007 0.000001 0.022593  1.716-08 8.73E-08    99.965X
47 1.716-08 0.000006 0.000001 0.022594  1.446-08 7.37E-08    99.971X
48 1.446-06 0.000005 0.000001 0.022S9S  1.22E-00 6.226-08    99.975X
49 1.226-08 0.000004 0.000000 0.022596  1.036-08 S.25E-08    99.979X
50 1.03E-08 0.000003 0.000000 0.022597  8.676-09 4.436-08    99.982X
51 8.67E-09 0.000003 0.000000 0.022598  7.32E-09 3.746-08    99.985X
52 7.32E-09 0.000002 0.000000 0.022598  6.186-09 3.166-08    99.987X
S3 6.186-09 0.000002 0.000000 0.022599  5.216-09 2.67E-08    99.989X
54 5.21E-09 0.000001 0.000000 0.022599  4.406-09 2.2SE-08    99.991X
55 4.40E-09 0.000001 0.000000 0.022599  3.726-09 1.906-08    99.992X
56 3.72E-09 0.000001 0.000000 0.022600  3.146-09 1.606-08    99.994X
57 3.146-09 0.000001 0.000000 0.022600  2.6SE-09 1.35E-08    99.995X
58 2.656-09 0.000000 0.000000 0.022600  2.246-09 1.146-08    99.995X
59 2.246-09 0.000000 0.000000 0.022600  1.896-09 9.656-09    99.996X
60 1.896-09 0.000000 0.000000 0.022600  1.596-09 8.156-09    99.997X
61 1.596-09 0.000000 0.000000 0.022600  1.356-09 6.886-09    99.997X
62 1.356-09 0.000000 0.000000 0.022600  1.146-09 5.816-09    99.998*
63 1.146-09 0.000000 0.000000 0.022601  9.596-10 4,906-09    99.998Y
64 9.596-10 0.000000 0.000000 0.022601  8.096-10 4.146-09    99.998*
65 8.096-10 0.000000 0.000000 0.022601  6.836-10 3.496-09    99.999X
66 6.836-10 0.000000 0.000000 0.022601  5.77E-10 2.956-09    99.999X
67 5.776-10 0.000000 0.000000 0.022601  4.876-10 2.496-09    99.999X
68 4.876-10 0.000000 0.000000 0.022601  4.116-10 2.106-09    99.999X
69 4.116-10 0.000000 0.000000 0.022601  3.476-10 1.77E-09    99.999X
70 3.47E-10 0.000000 0.000000 0.022601  2.93E-10 1.50E-09    99.999X
                            C-ll

-------
gxample 7.  Vacuum dryer emissions

     Example:  Consider the following example of a double -cone
dryer operating at 15 inches of mercury, with an air- leakage rate
of 15 scfm.  The temperature inside the dryer is 60"F.  Three
hundred pounds of product cake, initially containing 25 percent
by weight acetone are dried to less than 1 wt% solvent over the
course of 8 hours.  Calculate the maximum VOC emission rate.

                The total amount of acetone dried from the
product cake is:


      300 Ibcake, 0.25 lb acetone , ?5 lb acetone (initially)
                      Ib cake


                . .  300-75 - 225 lb product in cake
The amount of acetone remaining at the end of the  cycle is:

              0.01

           x *  (0.01) (225+x)
           x - 2.25*0.01x
       O.S9x - 2.25
           x - 2.3 lb acetone  (at end of cycle)

.*. Therefore, the total amount of acetone removed from the
drying cycle is:

                       75-2.3 - 72.7 -  73  lb

Average emission rate over  the drying cycle is:

       (73 lb/8 h) (1 h/60 min) - 0.15 Ib/min
                                 average dryer emission rate

     The initial drying rate is two times  the average rate,
assuming a straight -line decline.

Maximum (initial drying rate)

     (2)*(0.15) - 0.30 Ib/min

               - 58
     Therefore, the molar flow of acetone  is
            (0.30 Ib/min) {lbnol/58 lb) (60 min/hr)  -  0.31  Ibmol/h
                               C-12

-------
     The airflow  (leakage) is given as  15 scfm where  359  scf  (at
0C and 1 atm) is l mole.  Therefore, the airflow  is

      (15 scf/min)(lbmol/359 scf)(60 min/h) - 2.51  Ibmol/h

     Therefore, the uncontrolled emission stream at the start of
the drying cycle is estimated to be:
Component
Acetone
Air
T 0 T 'A L
Ibmol/h
0.31
2.51
2.82
mole fraction
0.110
0.890
1.000
This rate represents the maximum VOC emission rate during the
cycle.
                               C-13

-------
Example 8.  Atmosphere dryer emissions
     gxample.  A tray dryer uses 6,000 acfm of heated air  (65C)
over a period of 6 hours to remove isopropyl alcohol  (IPA) from a
batch of solids.  Each batch consists of 1,000 pounds of material
containing 40 percent  (by weight) solvent.  The final product
contains less than 0.6 percent solvent.  Calculate the total
uncontrolled VOC emissions per drying cycle and the maximum VOC
emissions rate.
     Mass balance over the drying cycle:
    (1,000 Ib cake)(0.40 Ib iPA/lb cake) - 400 Ib IPA initially
          Quantity of bone-dry solids - 1,000-400  600
     Amount of IPA remaining:
                              x/600+x - 0.006
     .. x - 3.6 Ib IPA
     Amount of IPA removed is:
                 400-3.6  -  396 Ib    (MW -  60.09)
           Average  emission rate    396 lb/6 h  -  66  Ib/h
            Assume initial rate - 2* average rate
                                -  (2)(66) - 132 Ib/h
     (132 Ib/h)(lbmol/60.09 Ib) - 2.20 Ibmol/h
Calculate composition of uncontrolled emission stream at start  of
drying cycle:
     Airflow:
(6,000 acf/min)(60 min/h)(lbmol/359 scf)(273/273*65)scf/acf - 810
Ibmol/h
Component
IPA
Air
TOTAL
Ibmol/h
2.20
e;o,oo
812.2
mole fraction
0.0027
0.9973
1.0000
     Knowledge of, or an estimate of  (as above) the uncontrolled
outlet stream composition is necessary to select an appropriate
control technology.  One should note that the mole fraction of
the VOC is considerably lower  (approximately two orders of
magnitude) in the convective oven exhaust than in the vacuum oven
(previous example).
                               C-14

-------
Example 9.  Vessgl Depressurization

     A 1,000 gallon nutsche  filter is  used to compress a slurry
containing acetone and  inerts  at  80F  (26.7C).   A pressure  of
35 psig is imparted onto the slurry  until  the desired filtration
is achieved  (approximately 40  minutes).  The  nutsche filter  is
then depressurized prior to  discharging  of its contents.
Calculate the emission  rate  of acetone resulting from this step.

Step l.  Ratio of acetone to air  initially present in the vessel
and after depressurization;


            35 psig = 49.7 psia ( 76 zmM? \  2,570 rnnHg
                               \14.7 psza/
         n*cecon _   246
           nair    760-246
                           =0.48 (after depressurization)
Step 2.  Calculate moles noncondensable  gas  in the vessel
initially and after depressurization;

     (Assume free volume equals  1/2  of the total  volume)

             (500 gallons)  - H - ,  (2.570 mrnHgr - 246 mmHg)

                                     ft3  (30QJC)
                            <
                               Ibmol K


     Nx - 0.523 Ibmcl

     Holes of nonconueii&aLle gas at the  end of  depressurization;


               (500 gallons) L  nft ^ -  (760-246 mmffg)
                    _ [7.48 gallons! _ _
                        '998.9 rmHg * ft3
                             lianol K
                                          (300JC)
     n.2  0.11 Ibmoles of ncncondensable gas

                          moles  acetone
Step 3.  Average ratio of 	 throughout the
                            moles  air
depressurization;
                               C-15

-------
 0-106 * O-48 = 0.293
Step 4.  Calculate Ib acetone emitted:
         Total moles noncondensable (non-acetone)  released:
         0.52 - 0.11 - 0.41 Ib moles
         Total Ib acetone released:
0.41 moles nan acetone (0.293 moles cetone\ / 581* acetone
                      \  moles non-acetone  / \ moles acetone
                        = 7 Ib/event
             - 0.17 Ib/min (1 event = 40 minutes)
                               C-16

-------
Example 10.  Emissions from a Steam Ejector

     A double-cone batch dryer  (volume of 20 ft3) operates at
74 mmHg.  A steam ejector is used to pull a vacuum on the dryer.
System components are listed below.  A solvent recovery condenser
operating at 20C precedes the ejector.  The solvent is methanol.

W leakage

     W - 0.032 p-26 v-60                   P [-] torr

     W  0.032(74)'26 (20)'6              V [-] ft3

     W - 0.59 Ib/h                           W [-] Ib/h

W leakage  (see Table A-9 for component-specific  leak rates)

     Assume system has:

     2 seals (rotary)  0.10          0.20

     10 threaded connections 9 0.015  0.15

     2 access ports 9 0.020           0.04

     1 view window  0.015            0.015

     10 valves  0.03                 0.30
     1 control gas valve  0.25       0.25
                                      0.955 Ib/h/in.

     For 4 in. fittings:

     W - 1.2 7TD0P0'26

     W = 1.2 7T(4) (0.955)74-26

     W - 44 Ib/h

.-. Total  in-leakage  (La) - 44+0.59 - 44.6 Ib/h

as cfm if  (379 scf/mol)(mol/29  Ib) - >582.8 ft3/h
                                   - 9.7 scfm
                               C-17

-------
VOC emissions:
                                   system
32 JJb MeOH\ I  44.6 iJb air/h \
                  Ibmol
                                               760
29 IJb
                             (760-95)
                                                      - 1
(VPMeOH at 20C  -  95  ramHg)
.. SE - 7.03  Ib MeOH/h
                                C-18

-------
        -U-.	Calculation of Emigaiona from Equipment
      Estimate VOC emissions from a facility process having the
 following components in light liquid service 100 percent of the
 time.
       1  pump
      18  flanges
       1  gas  valve
       5  liquid valves
       1  sampling  connection
       3  open-ended  lines
     Solution
     Multiply by  equipment  leak  factors  found in Table  A-6.
Equipment leaks
1 pump
   (0.0199 kg/h)(1) 100 percent service
18 flanges
   (0.00183)(18) 100 percent service
1 gas valve
   (0.00597)(l) 100 percent service
1 S.C.
   (0.0150)(1) 100 percent service
3 O.E.L.
   (0.0017) (3) 100 percent service

0.07887 Jcg/h  (h/3,600 s)  - 2.2xlO-5 kg/s
ko/h
0.0199
0.0329
0.00597
0.0150
0.0051
0.07887
                               C-19

-------

-------
   APPENDIX D.




COST CALCULATIONS

-------

-------
              Background  Information/Introduction  to
                Control Device Model Calculations

     The attached documentation details the calculations and
assumptions that were used to arrive at the control device cost
effectiveness curves used to set RACT.  The documentation
.requires some preliminary discussion of certain issues because of
the complexity of the approach.  This preface is intended to
provide this necessary background.  Note also that the basis for
much of the assumptions used in estimating costs is the fourth
edition of the OAQPS Control Cost Manual, prepared by the
Emission Standards Division of the Office of Air Quality Planning
and Standards, U. S. Environmental Protection Agency, Research
Triangle Park, NC 27711.

Discussion

Time Variation

     One of the major ideas behind RACT for batch processes is
the consideration of on-stream emission event duration.  The
calculations assume a certain mass emissions value, in Ibs/year.
Mass emissions is really the annual VOC emission rate from the
batch process vent(s) considered to be controlled.  By making
mass emissions an input variable, and by setting the VOC
concentration and the emission stream flowrate  (i.e., these
parameters are also input), the models are designed to calculate
a value of on-stream emission event duration, which we call "time
var."  This "time var" value is used throughout the model to
calculate control device costs and cost effectiveness.  The "time
var" or time variation field should not be confused with the
initial duration input as 60 minutes.  It was assumed that each
event took 60 minutes initially to calculate emissions per event
and heat load per event,  but considered the number of events per
year that would actually be occurring using time var  (which is
the fraction of continuous emissions).  Note, also, that
operation and maintenance costs are calculated on a per shift
basis.

     For those combinations of mass emissions and flowrate that
yielded time var values of less than  .33, only 1 shift per day
was assumed.  For time var between .33 and  .66, 2 shifts per day
were assumed, and 3 shifts per day were assumed for time
var >.66.
                               D-l

-------
Condenser Model Calculations

     The calculations below can be cross-referenced with the
example condenser model spreadsheet, which is included as an
attachment to this set of calculations.

1.  Input Variables - Emission Stream Characteristics
  .
Inputs                               Cell I.D.
1,000   a. Flowrate, (acfm) :         H7
   25   b. Temperature, (C) :        H8  (Default is 25C)
  760   c. Pressure, (mmHg) :         H9  (Default is 760 mtnHg)
   60   d. Duration, (min) :          H10  (Default is 60 minutes)
30.37   e. VOC volume percent*:      Hll  (Default is saturation)
   90   f. Required condenser
           control efficiency,  (%) :  J6
100,000 g. Mass emissions,  (Ib/yr) :  M14

     *This field is considered an input  field although the
spreadsheet is designed to calculate this value for saturation.
In our analysis, we multiplied the saturation value by a fraction
that would result in our desired concentration values  (i.e.,
1,000, 8,750, 10,000, and  I00,000ppmv for the cost effectiveness
curves) .

     For example, volume percent is calculated in the example in
the following manner:

     10 ~  ($BW$11 -  ($BX$11/(K8-$EY$11) } )/H9      (1)

where :

     $BW$ll m 7.117  (Antoine Coeff 'a' for acetone);
     $BX$11 - 1210.595  (Antoine Coeff 'b' for acetone);  and
     $BY$11 - 229.664  (Antoine Coeff 'c'  for acetone).

where :

     the Antoine 's equation Is of the form:


     10910 *    -
where :
     P *   vapor pressure of component a, mmHg
     a, b, c  - Antoine' s coefficients for component  a
     t    - temperature, C
                               D-2

-------
     Equation 1 is the vapor pressure/total pressure where vapor
pressure equals:
     10* [log10Pa* - a - b/(c + t)]
               Pa* - 10* [a - b/(c + t)]
     So, 10" [$BW$11 - ($BX$11/(H8 ( $BY$11))) - Pa*
     Volume percent -    [Pa*/total pressure  (H9 - 760)]
      Hll - 30.37% (in the example)
Non-condensable Volume Percent. (%): (Cell H16)
     Because* the gas stream contains a portion that is
condensable material and a portion that is noncondensable
material and the fraction of the stream that is condensable
material was calculated above the remaining portion of the stream
is easily calculated as below:
     - 1 - Hll
      l - Volume percent (calculated above)
     - 1 - .3037 - .6963 - 69.63%
Emissions.  (Ibs/event): (Cell H17)
     This equation uses the ideal gas law to estimate the
emissions of the VOC  (acetone in the example) per event before
the condenser is applied.
     Hll  H10 * H7  H9/O98.97 *  (H8 + 273))*$BZ$11
where:
     $BZ$11 - Molecular weight of VOC; acetone - 58 Ib/lbmol; and
     998.97 - Universal gas constant,  (rnmHg ft3/lbmol K) .
     - .3037*60 min*l,000 acfm*760 mmHg/(998.97*(25+273))*58
     - 2,702 Ibs/event
Partial Pressure at Exit Stream.  (mmHa):  (Cell J9)
     The above calculations are used to estimate the uncontrolled
waste gas composition.  The spreadsheet will  estimate control
costs for a desired control efficiency or exit condenser
temperature, if specified.  If control efficiency is specified,
as it was in this cost analysis,   the spreadsheet calculates the
required exit partial pressure of  the waste gas, which  is  the
saturation vapor pressure at the exit condenser  temperature.
                               D-3

-------
      For  this  example,  the  condenser control efficiency,  jg    90%
      90 percent  control for a condenser means the following:
      VOCIN
     NON COND  IN
Y
VOCOUT
                                4
                              Y
                              VOCCOND

      [YVOC IN]  [FLOWRATE]/RT  *  MOLES VOC  IN

      [1  - Y VOC  IN]  IFLOWRATE]/RT   MOLES NONCONDENSED

     MOLES VOC OUT   (1-CONTROL EFF.)(MOLES VOC  IN)

After manipulation and  cancelling  these equations  are reduced to:

                     (1  -  CONT.  EFF)(Y
     Y                  -      .        ,,/y-
     LOGOUT        1  - Y VOCIN   CONT. EFF

     This is the same  equation as the one contained  in cell  J9,
except that the equation in cell J9  is multiplied by 760  mmHg,  the
assumed total pressure of the system, atmospheric in thsi case,
The resulting value is the partial pressure of  the VOC component.

     - 760 mmHg*(.3037(1-.9))/{I-.3037*.9}

     - 31.77 mmHg

Required Condenser Exit ..Temperature.  (C) ;  (Cell... J121

     Substituting the  partial pressure back into Antoine's
equation, yields a temperatuere at which 90 % control is  achieved.
                                - C  Or  t - -rr		-..  -  C
                   (log P* - a)   w  w*   - -  (a  -  logP*)

     The formula in cell J12 is:

     (($BX$11/($BW$11-LOG(J9))-$BY$11)}

So, substituting the values from the example  condenser model
spreadsheet into this equation:

     - ((1210.595/(7.117-LOG10(31.77))-229.664))

                                D-4

-------
      14.06 C

Condenser Exit Flowrate  (variable) .  (ft^/min) r  (Cell Hifi)

     This formula takes into account the ratio of the volume
percent of noncondensables entering the condenser and the volume
percent of noncondensables exiting the condenser along with the
equation: P^V^/^ - P2V2/T2' Solvin9 for V2*

The formaula in cell address H18 is:

     H7 * H16 * (((1.8 * $CE$6) + 32) + 460)/(((1.8 * H8) + 32)
     + 460) * H9/$CH$6/ 11-H19)

where :

     $CE$6  Condenser Exit Temperature, ( C) ;
     $CH$6 - Compressor Pressure,  (mmHg) ; and
     H19   - Condenser Exit Volume Percent,  (%) .

So, substituting the spreadsheet values into this equation: is

     - 1,000 acfm*.6963*(( (1.8M-14.06 C) ) +32) +460) / ( ( (1. 8*
       25 C)+32)+460)*760 mmHg/760 mmHg/ (1- .0418)
      631 cfm
Condenser Exit Volume Percent.  (%) ;  (Cell H19)

     This field substitutes back into Antoine's equation the
required exit condenser temperature, which is given in $CE$6.

            p   ._   - /D    .  v
             partiax' - LOCCU.    oui
     10*($BW$11 -  ($BXSll./($CE$fi + $BY$11) ) )/$CH$6

where :

     $CH$6 - Total Pressure.  (mmHg) .

So, substituting the values from the example condenser model
spreadsheet into this equation:

     - 10A(7. 1175- (1210. 595/(-14. 06 C+229 .664) ) ) /760 mmHg

     - .0418 - 4.18%

Constant Properties Exit.  (Ibs VOC/event) :  (Cell H2Q)

     Using the ideal gas law  this equation calculates the
emissions per event aluei.  uue condenser:
     H19  H18  $CH$6  $BZ$ll/(998.97  *  $CF$6)    H10
                                D-5

-------
where:

     $CF$6   Condenser Exit Temperature, K.

So, substituting the values from the example condenser model
spreadsheet into this equation:

     -  (.0418*631 cfm*760 mmHg*58 Ib/lbmol) / (998.97*258.94) *60
     - 270 Ibs VOC/event

Condenser Ccntro. Efficienc.  (%) ;  (Cell H2H
     This cell provides for the calculation of the condenser
control efficiency, but it is not used in this example because the
efficiency was input in the analysis.

Condenser Heat Load. (BTU/event): (Cell H221

     Cell H22 calculates the condenser heat load, in terms of
sensible heat and latent heat of cooling the stream down to the
desired temperature.

     Heat load is made up of: (l) latent heat of condensation for
the material condensed, (2) sensible heat of cooling of the
noncondensables, and (3) sensible heat of cooling of the
condensables.
                                    o
     Sensible heat is calculated as m Cp AT
                                    o
     Latent heat is calculated as mHv

where:
     ^sensible - m cp  T'  and
     ^latent " m Hv

     where:
          m - mate rate;
          Cp - heat capacity;
          Hrl - heat of vaporization; and
           T  range of cooling.

1st Term:   (Sensible and latent heat of condensables)

          ((H17 - H20)  * ($CB$11 + $CA$11 *  (H8  - $CE$6) * 1.8)}

where:

    H17 - H20 - Ib VOC condensed - 2702 - 270.2  - 2431.8 Ib/event;

    $CB$11  220 (BTU/lb,  H^p of acetone) ; and

    $CA$11  0.3 (Cp acetone, BTU/lbeF), assumed constant.
                                D-6

-------
So, substituting these values into the 1st part of the equation in
cell H22:
    Heat load  (condensed) -  (2702 - 270.2)  1220 (BTU/lb)
    + 0.3 BTU/lbF  (25-(-14.06))*1.8] - 586,288 BTU/event
2nd Term:  (Sensible heat of condensables that were not condensed,
          calculated in Cell H205
The 2nd part of the heat load equation as it appears in the
condenser model is:
     (H20*$CA$11*(H8-$CE$6)*1.8)
So, substituting the correct values into the equation:
     -  (270.2*0.3(25-(-14.06))*l.8) - 5699 BTU/event
3rd Term:   (Sensible heat of noncondensables)
The 3rd part of the heat load equation as it appears in the
condenser model is:
     - H18*(1-H19)*H9/(998.97*($CE$6+273)*29*H10)*$CA$20*(H8-
       $CE$6)*1.8))
where:
     $CA$20  BTU/lbF for air,  .24.
So,
     - 631 cfmMl-0.0418)*760 mmHg/ (998.97* (-14.06+273) *29*60) .24
       M25- (-14.06)*!.8) )
     - 52,517 BTU/event

Therefore the sum of the three terms is 586,288 + 5,699 + 52,517  -
644,504 BTU/event.   The spreadsheet calculated a slightly
different value, 644,613 BTU/event.  However, rounding is the
reason for the difference in the values.
                                D-7

-------
Condenser Heat Load During Non-Events.  (BTU/event);  (Cell H23)

     Assumed 10 percent of heat load during events to keep the
heat exchanger surfaces cold and to account for heat losses.

Tons;  (Cell H24)

     Calculated Annual Refrigeration requirements based on
on-stream time.
                                           [Time var]
where :
               8,760/12,000 BTU/hr/ton
     -    tons refrigeration used to calculate annual energy costs
          - 5.57

Refrigeration Unit Cost  (DIFF TEMP). (S);  (Cell H27)

     The refrigeration cost is based on the heat load and
temperature.  The cost for a refrigeration unit are given in 10F
increments from 40F to  -70eF, and a 30F increment between -70F
to -100eF.  A Quattro Pro macro that checks the required condenser
exit temperature appears on the right side of the condenser model
spreadsheet underneath the "\t\".  The group of numbers under the
macro are the refrigeration unit costs given the calculated tons
of refrigeration required.  Below are the actual equations that
produce the Purchased Equipment Cost (PEC) for refrigeration units
at different temperatures.

     Temp.. F Cost Equation
     40        1,451(Tons) * 10,817
     30        1,820 (Tons) + 11,064
     20        2,340 (Tons) + 11,021
     10        3,197 (Tons) + 13,972
     0         4,013 (Tons) + 14,427
     -10       5,582 (Tons) + 13,431
     -20       7,560 (Tons) + 13,451
     -30       4,334 (Tons) + 40,089
     -40       5,459 (Tons) + 40,082
     -50       6,704 (Tons) + 39,993
     -60       7,152 (Tons) * 43,640


                                D-8

-------
     -70       8,938 (Tons) + 41,713
     -100      17,798 (Tons) + 46,906

     The required condenser exit temperature for the stream
contained in the example condenser model spreadsheet is -14.06C
which equals approximately 7F.  So, the equation for 0F would be
used to estimate the PEC of the refrigeration unit as below:

     4,013 (Tons) + 14,427

where:
     Tons  the tons of refrigeration required during an event

     Sizing the condenser refrigeration unit based on the heat
load calculated during an event provides the "worst-case" costs
because this would be the maximum heat load encountered by the
condenser during any given time throughout the year.

So,

     4,013 (H22/H10*60/12,000) + 14,427

Substituting the values in the example into this equation:

      4,013 (644,613/60*60/12,000) + 14,427

     - $229,996

Total System Costs (1.25*Unit Cost). (S) :  (Cell H32)

     1.25 * PEC, to account for precooler and auxiliary equipment.

     - 1.25 * $229,996 - $287,495

TotalCapital ^nvestmept(TCI). (S) : (Cell H33)

     1.74 * total system costs* 1.25    [3rd quarter, 1990]

where,
     1.74 factora installation for nonpackaged systems
     25 percent covers manifolding

Direct Costs

Operating li&QX  (S) : (Cell H35)

     - 0.5 hours/Bhift*$15.64/hour*H51*H52

where:
     $15.64/hr is operating labor rate;
     H51 - shifts/day; and
     H52 - days/year.
                                D-9

-------
So, substituting the values for the example into this equation:
     . 0.5*15.64*1*365
     - $2,854/year
Supervisory Labor (S); (Cell H36)
     - 1.15*H35
So, substituting the values fron the example into this equation:
     - 1.15  $2,854/year
     - $3,282/year
Maintanence Labor ($): (Cell H37)
     -0.5 hour/shift*$17.21/hour*H51*H52
where:
     $17.21 - maintanence labor rate.
So, substituting the values from the example into the equation:
     - 0.5*17.21*1*365
     - $3,141/year
Maintanence Materials  ($);  (Cell H38)
      H37
     - $3,141/year
Electric Compressor Motor.  ($); (Cell H39)
     Electricity requirements were based on the average tons of
refrigeration required during a year and refrigeration
temperature.
     kw/Ton             T  (F)
      1.3                 40    We used a regression  line, where
      2.2                 20
      4.7                -20    T -  - 13.08  (kw/ton)  + 43.16
      5.0                -50
     11.7                -100    r2 correlation to these data
                                points - 0.955
     The regression  line was used to obtain  (kw/ton).
So,
                               D-10

-------
       { ( ($CB$6*1.8+32')-43.l6)/-13.08)*H24*8,760*.059/0.85
where:
     $.059 - Cost/kwhj and
     0.85 - Efficiency of motor.
So, subsituting  the  correct values  into the equation:
     -  (( (-14.06*1.8+32)-43.16)/-13.08)*5.57*8,760*.059/0.85
     -  $9,449
Overhead. (S):  (Cell H40)
       .6*sujn(H35. .H38)
where:
     H35  - H38 - the costs of  operating,  supervisory,  and
     maintanence labor and maintenance  materials.
So, substituting the correct values  into the example equation;
     -  $7,451/year
Capital Recovery.  (S);  (Cell H41)
Assume  15-year life, 10 percent  interest
     -  .1315*H33
     .  $82,227
General _Adminis5ra.t4ygtT$.::ss.  Insurance,.  (SI ;  (CeljlHf2)
     -  .04*H33
     -  $25,012
Total Annualized Cost.  ($): (Cell H43)
The total annualized cost is the  sum of all the direct costs.
     -  $136,558/yr
Mg/yr Controlled;  (Cell H44)
     -   (M14*0.454/1000}*J6
     -  (100,000*0.454*1(000}*0.9
     -  40.86 Mg/yr
                               D-ll

-------
     Effectiveness (S/MO):  (Cell H45)
     - H43/H44
     - 136,558/40.86
     - $3,342/Mg

Calculation of Time Variation:  (Cell M17)
     - M17*998.97*293*(1/60)*(1/8,760)*(1/$BZ$11)/H11/(H6*760)
So, substituting the correct values into the equation:
     - 100,000*998.97*293/60/8,760/58/.3037/(1,000*760)
     - 0.00415
                                   D-12

-------
CONDENSER MODEL SPREADSHEET
            D-13

-------
(ON DENSER MODI :i.
PARAMETERS:

Ftnwiate. (vfm)
I Inwialr. (acfml
Irmprralwr. |l>rg C)
Prruwr. I mm) IK)
Duration. (mini
VOC. (vol*!-)
V(K.(ppmv)

CALCULATIONS:
 I 000
 1. 000
  25
  7MI
  M)
101.711
                   Hrquirrd < 'ondrmrr Cnnlrnl I (lie Kncy. (T)
                             Wr
Partial Piruurral l-Jiil Slrram.
       11 7*77
                   C'ondrnrr I H Irmprialurr. ((')
                            140*
                   Mau Fmiuionv
                                                            100.000
Non-roBdcMuMr Vnlmnt PrrrciM. (T ):
FrmuinM. (Mn/rvriil)
Condemn liih Ho*tR(vnWc. (in/mm)
( andrmrf Kill Vnhimr Prirrnl. \"< )
rotHlanl Pmprrim lul. (NK VtM'/rwnl)
t oncrmcr < ontiol l;f f trirnry. (T |
I onArmrt I lrl I nad. ( HI I l/rvrnl I
( 'urn rmrt I Iral t.n*d Ounnf NOB t-wnlv (im l/fvrnl)
I 'onftnsrt I Iral l.iud |lm|. 1 1 2 000 Mil I/hi)
IVH T
                 oM (1)11 1 1 1 MP) (t)
COSTS:

Total Sytlm CmH ( I 25M Iml Cosl). (1)
'lolal ( Jptlal Invrslmrnl fl< '
              f. |$(
 Mamlanrurr l^bnf, (t)
 Mamlanrncr Malrriak. (t)
 Orloc <'ompfMH Motor. (J>
 Ovrihr*d.(S)
 ( apilal Rccovny. It)
 Grnrr al Admrnalraliw. Tun. Imwanrr. ($)
 TCJTAI. ANNVIAI I/H))S1.(J|
              II ID
 HMI FAC'IKRS

 rvrnK/ihill
 da>>.'yrai
  ; ol Him-
 2.702

 4111%
 270 2*
                   1 inv Variation
                                                            000415
 t>4.4M
  5 57
  1241
  I 112
J2H7.40S
1*21.102

 J2JS4
 SV2H2
 SVI4I
 11 Ml
 10440
M2.227
S2S.OI2
  40 W.
 $1.142
                                                            0 H2<>2
                                                              I
                                                            IIINISM
                                                            IIINIIC'
              \l\
                              40  (ilconlrmp>=40|/F( FORMI-IJ27-(Q
                              V)  (il cnnlrmp> = |/F.CFORM2 - 1127 - ((}
                              20  III t onlrmp> = 201/1 (H)RMJ  1127- I(J
            10  (if conlrmp > = IO|/F( FORM4 -1127 ~ |Q
            0  III conlrmp> =|/FC FORM5 ~ 1127 ~ =  IO|/F.t:FORM*-H27- |
            20  |ilconlrmp> =  2|/EC FORM7-1127' (
            10  (ilconlnp>=  V)|/FrF()RM|-|l27-(
            40  (ifconlemp>=40)/tC:FORM-|H7~(
            50  (iJco=50J/ErUI>l~H27-JO
            M  (ilcontrn.p> = 40)/F( IIIA2~ 1127-(O
            70  (ilronl<-mp>=  70)/FrlH^J~H27- (Q
           100  lf<:\ II j\4 - IU7 - (QUIT)
           40  U7l 401
           10  IOMM 25
           20  IV.7204*
           10  IH5707 54
            0   229%2
           10   1112*13
           20  41
-------
Al  |WI(.||| .) ( ( INI II NSI It Ml II )l I
A4  |WI(.||I .'I PAKAMI II MS
IS |W|S| Hri|ililrJ <  mh-nsrl < onliol I (( K-m v. ('', I
A(>  |WIi| ll.mi.iir (Minn
I M>  I.O||WM| KHHI

A'  |WI(,| lliiwMlr. laclml
II'  
A*  |WI| (rmrx-i.tlurr (IVj I I
II*  I.O||WI4|.'S
J* |W|S| 'I'arlial Pir\suic al l-nl Sliram. (minll|>)
A<  |WI*| Tiruir. (mmll|>)
ll  (.fl)|WI4|'MI

A10 |W|fc| DwalMMi. (mini
Illll |.n)|WI4|MI
AII |wi*| v/yi|
MM i.o|Wl*| innono
A I* |WIA| Non-roiMlrmaMr Vohimr Pritrnl. (T)
III* (P'||WI4|I Illl
Al' |Wlh| l-nwimiv (Iht/rvrnll
III' l.n||WM| IIIIMIIO*HOWII4/((7>(llt27))l>tH71ll
JI7 |W|S|-|nnrVarialimi
Ml' |hS)
AIK |Wlh|'( ondrmri I ill KlmnalrlvaruMr). (llVmn)
MIS |0||WI4|
Al<> |W|I.|'( ivixlrmri Kill VoKmw PrirfM. |T)
Mil (P'l|WI4|in-($MWlll ($HXtll/(t  i . I
11'I  iP(i||\VII|i.i|liillll' 11 ill III'I-II ii Illl' 11.1 III 'I

-------
A2> |WI*|Vondrmrrlleall.MHTU'vrl|
A2 |WI*| I ondmsrc Ikal tani During Nwi-l-vriih. (IH I l/rvcnlt
im |.0(|WI4|Oril22
A>4 |Wlf.| ( ondriwf lit* I oJ (lom ( ltW BIT l*f I
A1V |WI6| 'IVha T
K25 |WIO|M
A2* |WI(k| XJ<:(K)I, (Ib/lir):
112* t.O||WI4| H
A27 |W|f,| Rcl rigrnlin* Unit TxMl (DIFF TEMf). (t)
1 127 (< -0) |WU| (401 V = 40|/l ( F< >KMI - 1127 -
J2 |IO|W|S|W
K2X |WIO| (ilroiil*mp> = Vll/l< KIRM2-II27-
J2< (in||WIS|20
K2 |WIO| (1lc=l'W FORMS -1127-
 AJ2 |Wlfc| Trtl Syjlrm t:h (I 25'HnH <'ol. (*
 1112 t= IO|/F,(TORM6-H27-
 M\ |WI*| Tirtjl ( apHil Invnlmnl (I* K (*l
 im  <(0||WI4| I 74MI32M 25
 J (FO||WIS| 20
 Ml |WIO|'(if cwrtemp>= 20I/FCFORM7-M27-
 AM IWIH'DIUWTttKIS:
 JVI |FO||WIS| W
  KM |WH)ritfcoliiip>= M|(H:FORMf-H27-
J1S (|0)|W|S| 40
K^ |WII>| (ilon(mp>=
AVi |WI(.| NupcivKoo I ahcM. (tl
||1, |( 0||WI4|I I^'IIIS
n* ti- M|H

                                      II"

-------
,\M |WI.| Mainunrnrr I ahot. (Si
IIW |(||WU|OSM72IMIM>II<2
IU |III(|W|S| Ml
KU |Win|'|riroiitrmp>= MI/H III.
AW |WI*| Mainlmrntc MjtriwK (t)
III* <<0||WI4| 4 IH7
JM (H>|WIS|-70
KW |WIO|(rtconliip>=70|/lrUI.AJ- 1127 -(OUR)
AW |WI*| TVrlnr Comprruoc MofcM. It):
I iw ici |WM| (tt:FJ** 1 1   I*V-I3 e)*H2*i760'o n(o gs
jw ihoi|wis|  ion
K W |WIO| Vl-rUI .A4 - H27 - J
AM |WI| T>vrriKad. (t)
A4I |WI|t>,.uJ Rrrwiy.(t)
(Ml H>|WI4|OIH5*:|V1
A42 |WIA| X it wl Adminnlralivr. Tun. Iwurwre. (t)
1142 |(tl||WI4!OIMMM
,\n |wi.| it; i AI. ANNIIAI ,i7i;nri>sT. d
1141 ((H||W|-l|(SSHM(lin 1142)
.141 l.-Mi'IF(IISI>l.-.H4VH44M
J4V tl nill'MI I fHIOl i 4im*<>,

-------
AM
MSI |W|4|(al|(M|7< = OM.|fti|l|MI?< =0 U.I.2H.M
JM  lln)|W|S| 40
KM |wio|tvw(Sil*J2/*lBio**o/i2om>H*H*2i
11^2 |WI4| MS
JS2 (IO)|WIS| 10
KS2 |WIO|(*7M'(Sllt;2.tmiO'*4VI2ftO)t W*Jl
AM |WU| n of timf :
II" (I S||WI4| 4lll(MO*IIMfll$IMISM70
AM |WI| Dlia,\1K)N
1154
JM |l:) |WIS| 70
K54
     |WIO| (l77W

-------
 A(>7: <>|K-r;il<>r (t)
 I hi ((1l)|W|S|O.S*|S.(hl'l I7'l IK
 A(>8: 'Supervisory I ahor. ($):
 II.8:(C<)|W|S|0|SM<.7
 Af4: 'Mimiancncc I ;trmr, ($):
 I6' :(CO)|W|S|OS 17.21*1 I7M 18
 A70:'M:ilcri;ik,(S):
 I7 (CO)|W|S| IMW
 A72: llhlilics:
 A71: 'Natural (ius,(S):
 IIV (dl)|W|S| +| 42/|tNNWJ
 (i7^: '(S.1.VKNNISCT)
 A74: 'l-lcclrkily, ($):
 |-74: (CO) |WI.S| (MNNIII7M 4I*29/0.6*0.059'8760
 (i74: '(S'W/kwh)
 174:  (-HANOI: DI-I.IP IOR DIIFI-NI;RGY RIX-O 
 A75: 'Blu:
 175: (,0)|W|5|((l7Vin(MI/\V892)+(|-74/.O.S9*MI2 I))
 A76:rl(Mill Died Cosl,(J):
 176: (( 0)|W|S| l*Ca>SUM(l67..l74)
A78: 'Indirect Annutil Costs:
A79:'()vcrhcad,($):
 179: (CO) |W|S|06'(
A80: 'Administrative, ($):
 I HO  ((-0)|W|S|02M62
AMI: 'Properly  lax, (S):
IKI  (OI)|W|S|OOri(>2
A82: 'Insurance, ($):
IH2((1))|WIS|OOIM62
A8V 'Capilal Recovery, (S):
IXV(CO)|WI5|O.I627VI62
A8.S: Toliil Annuuli/ed Cosl, (S):
I-K5. (CO)|W|S|

-------

-------
CAPITAL COST CALCULATIONS:
Dirccl Costs
l-quipmcnt Cost (Kecouperalive Incin ),($):
Auxiliary l-quipmcnt (Ductwork, Slack), ($):
Auxiliary Collection I an, ($):
Instrumentation, ($).
Saks Tax, ($):
l-rcight, ($):

TOTAL I'UKCIIASKD liQUIP COST($):

Dirccl Installation Costs, ($):
Indirect Costs, ($):
                                                        $2.701
                                                        Jft.274
                                                       $44,420
                                                       $45,
-------
THERMAL INCINERATOR  MODEL
PARAMETERS:

llowriilc, (scfm):
Waste (ias VOC Concentration, (ppmv):
I Icaling Value c>f VOCs, (Hlu/scf):
l-ncrgy Recovery, (%):
Incincraiof Operating Temperature, (F):
Wasic (iiis Temperature, (l;):
Prchcaicr Tcmpcralurc, (!"):
Molecular Weight of VOC:
Molecular Weight of Gas:
Durulion, (min):
l-vcrm/Shifl:
Shifts/Day:
Days/Year:
Time Variation:
I .cngih of Collection Main, (ft ):
Number of Manifolded Sources:

Step I: Calculate Total Waste Gas Now

Oxygen (O2) Content of Waste Gas, (volume %):
Dilution Air Required for Combustion, (scfm):
Dilution Air for Safely, (scfm):
Total (iiis llowrale, (scfm):
  UNI
 Ill.tNN)
 2,(NMI
  70
 I. MM
  70

 6450
 0.60
   8
   I
  V.S
0.2X401
  .TOO
   10
 2079
 0.00
 14.29
 114.2V
VOC's Controlled, (Mg/yr):
         11.12

Cost l-ffeciivcness, (S/Mg).
        S4.9W

Mass (-missions, (Ih/yr):
        25,00(1

Time Variation:
        02H40

I ncrgy, (Blu/Mg):
     48,616,308
Step 2: Heal Content of Waste Gas, (Ulu/scf):

Step 3: Calculate Gas Temp, l-xil Prchcaicr, (l;):

Step 4: Calculate Auxiliary l:ucl Required, (scfm):
    for events
Step S: Calculate KMal (ias llow, (scfm):
    for events
('alculaic Maximum Auxiliary (ias llow, (scfm):
(';iliul.uc Maximum liH.-ilGas llow,(scfm):
( .ikiil.ilcd Annual Gas I low. (ufy):
 I7..SO

1,141(10

 0.110

 114 &

  ISS
 M54
SK4.2K2

-------
CALCULATIONS
                                      CONDENSER            COMPRESSOR P
                                      EXIT    TEMPERATURE
ANTOINE'S EO COEFFICIENTS
                                         (Q     (K)
                                       -14.11644 258.9J56
760
VOC    A      B
MW     Cp     hv
 Acetone    7.117 1210.595  229.664    58.0S   0.1004  220.127
 NITROGEN
                                           0.24

-------
BVI: 'CALCULATIONS
CE2: 'CONDENSER
CH2: 'COMPRESSOR PRESSURE
CE* 'EXIT
CF3: TEMPERATURE
BV5: 'ANTOINE'S EQ COEFFICIENTS
CE5: ~(C)
CF5: ~(K)
CE6:+JI2
CF6: +SCES6+273
CH6: 760
BV8: 'VOC
BW8: 'A
BX8: 'B
BY8: 'C
BZX: 'MW
CA8: 'Cp
CB8: 'hv
BVIIr'Acclonc
BWII:7.H7
BXH:  I2I0.5W
BY II:  224.664
BZ 1 1:58(18
CBI I: 22(1.127
BV20: 'NITROGEN
CA20. 0.24

-------
Thermal Incineration Model Calculations

     The calculations below can be cross-referenced with the
example thermal incinerator model spreadsheet, which is included
as an attachment to this set of calculations.

     l.  The information necessary to calculate incinerator costs
for any given situation is listed under "Parameters" in the
spreadsheet.  This data is also listed below:


 Example                                                  Cell  '
 inputs                                                   I* P.

    100     1..   Flowrate,  (scfm) ;                         F6
 10,000     2.   Waste Gas VOC Concentration;              F7
  2,000     3.   Heating Value of VOC's,  (Btu/scf);        F8
     70     4.   Energy Recovery,  (%);                     P9
  1,600     5.   Incinerator Operating Temperature,  (F) ;  Fio
     70     6.   Waste Gas Temperature,  (F);              Fll
           7,   Preheater Temperature,  (F);              P12
     64.5   8.   Molecular Weight of  VOC;                  P13
       .6   9.   Duration,  (min);                          P14
     8    10.   Number of events per shift;               F15
     1    11.   Number of shifts per day;                 F16
    365    12.   Number of days  per year; and              F17
 25,000    13.   Mass Emissions,  (Ib/yr).                  F18

     There are also several fields in the "Parameters" section
which do not contain information that must be input for each
given case.  They are:

     a. Molecular Weight of Gas; (Cell F14)

     This value is calculated from the input VOC concentration
and the molecular weight of the VOC as below:

MW_ag - fVOC Cone  (ppmv)1 x MWVQC] + Fd-VOC Cone {ppmv})1 x

          1 x 10s                            1 x 106

The formula contained in the example thermal incinerator model
is:

     F7/1,000,000*F13-M1-F7/1,000,000)*29
                               D-21

-------
So, substituting the coreect values into the equation:

     - 10,000/1,000,000*64.5+(1-10,000/1,000,000)*29

     - 29.36 Ib/lbmole

     b.  Time Variation;  (Cell J15)

     This field is used to calculate the fraction of time that
the event occurs over a year  (continuous maximum of 8,760 hours).
In other words, if the event lasts 0.6 minutes and occurs 8 times
a shift, 3 shifts per day, 365 days per year, the time variation
equals 1 percent.

     c.   Length of Collection Main and Number of Manifolded
          So_yre e, s ;  (Ce 11 s F2 0. F21)

     These fields were inserted to cost out the collection main.
Because we have no specific situation, we assumed the collection
main would be 300 feet in length and have 10 takeoffs  (sources).
These values remained constant during our analysis, although real
data could be input for any given situation.

     2.  Calculations

     The calculations done by the spreadsheet are presented
below:

     Step 1:  Calculate Total Waste Gas Flow

     a.  0^ Content of Waste Gap, (volume %) . (Cell F25).

     This equation assumes that the waste gas is composed of air
and VOC's.  Air contains 21% oxygen, on average.  Therefore, 02
content can be expressed as:

     (1 - VOC conc/1 x 106)  0.21  100

     b.  Dilution Air Required for Combustion,  (scfm);  (C^ll 26)

     The OAQPS Control Cost Manual states that there must be at
least 20 percent 02 in the waste gas for combustion to occur
(p. 3-24).  An average of 3.96 moles of 02/mole of VOC was found
to be an acceptable ratio to express 20 percent 02.

     c.  Dilution A^ir Required for Safety,  (scfm);  (Cell F27)

     According to the OAQPS Cost Manual, p. 3-26, safety codes
require that the maximum VOC concentration  in the waste gas
stream not exceed 25 percent of the lower explosive limit of the
organic compound when a preheater is used.  We assumed that a
reasonable LEL value for common compounds was about 3.5 percent,
                               D-22

-------
or 35,000 ppmv streams; 25 percent of 3.5 percent corresponds to
a value of 8,750 ppmv.

     This LIL value was derived from the following data:

           Compound      LEL  (ppmv)
Acetone
Benzene
Ethanol
Ethlene
Ethylene
H2
H25
Methane
Methanol
Propylene
26,000
14,000
33,000
28,000
28,000
4* ** /% *
U , UUU
43,000
50,000
67,000
24,000
           Average - 35,000 ppmv
                     0.25  (35,000) - 8,750

     Therefore, additional air must be added to the waste gas to
dilute the waste gas VOC concentration to 8,750.

     This formula for calculation of dilution air was derived in
the following way:

      VOC cone  (ppmv)                                     6
     [	,	] [fiowrate]            - 8,750/1 x  10
      1 x 106                     	
    .- waste gas    dilution      dilution   -
    L fiowrate + combust, iow aii  * safety airj
     By cancelling and aicu-kipulation, this formula  reduces  to:

                                                          (Dilution
Dilution                                                 combustion
safety   (Fiowrate) (PPMVOC)  - 8.750   (Fiowrate)  -  8.750*
air                                8,750
     d.  Calculafcg Total Gas Flow,  (scfm) :  (Cell  F28)

     This field calculates the total amount  of  gas  flowing into the
incinerator during the emission event,  the total  gas is composed of;

     Input flow  (waste gas) * dilution  air for  combustion +

     dilution air for safety
                                  D-23

-------
     Step 2;    Calculate Heat Content of the Waste Gas.  (Btu/acf):
               (Cell F32)
     The formula for this field is:

        VOC    Initial
     [	-] [Flowrate]
      TotaiGa8 Flow     * VOC heat content 
     - Btu/scf
     Step 3;    Calculate Gas Temperature Exit Preheater.  (F): (Cell
               F34)
     From the OAQPS Cost Manual, the preheater temperature  is related
to the fractional energy recovery and the incinerator operating
temperature and waste gas inlet temperature by the following equation:
                       T   - T  
                        wo   Awi
     Energy Recovery  =	=
                        1fi " ivi
     where, Two - Gas preheater exit temperature
            Twi * Waste gas inlet temperature
            Tfi - Incinerator operating temperature
     This equation is manipulated to
     Energy Recovery *  /T     T  .\  . T  . m T
     	ai-5-g	  *  (Tfi   Twi)  + Twi - iwo
     in the spreadsheet.
          4:  Calculate Auxiliary Fuel Required,  (scfm) ;  (Cell  F36)
     The equation for auxiliary  fuel  is presented  on pages  3-32  of  the
OAQPS Cost Manual.  It is:
tPaf Qaf " 'wo QWO  ^Cpm air  (1.1 Tfi  - TWQ -  0.1 Tref)  -  (-Ahcwp)]
                        -  l.lCpm air  (Tfi  - Tref)
                                  D-24

-------
where:  paf - density of methane, 0.0408 lb/ft3  77F, 1 atm

        Qaf - natural gas flowrate, scfm

        pwo - pwi  density of the waste gas,  77F, i atm
                         (0.0739 Ib/scf  77F, l atm)

        Cpm air - mean heat capacity of air

        Assume 0.255 BTU/lbF  (the mean heat capacity of air
                          between 77F and 1,375F)

        Tref * Taf * temP-  ambient                              ;
                          (Temp, auxiliary fuel)  77F

        - Ahcwo - heat content of the waste stream,  BTU/lb

        - Ahcaf - heat content of natural gas, 886 BTU/lb
                          (21,502 BTU/lb)

        Step 5:          Total Gas Flow - Total Waste Gas Flow +
                         Auxiliary Fuel, (scfm): (Cell F38)

     Maximum Auxiliary Gas  Flow

     This field considers the amount of auxiliary fuel necessary to
keep the incinerator working in the absence of a VOC emission stream.
In other words, during the period of time when there is not an
emission event in the incinerator.  This equation is similar to the
one used above, except that - Ahcwo is 0.

     The maximum total gas  flow equals the amount of necessary
auxiliary fuel when there is no VOC plus the total amount of waste gas
from Step 1 (d).
                                                                *
     The calculated annual  gas flow,  in standard cubic feet per year
(SCFY)  is the amount of natural gas that is required in the
incinerator in a year, considering the weighted average of the gas
flow during emission events and without emission events.

     Capital Cost Calculations

     Equipment Costs.   (Based on p.  3-44 of the OAQPS Cost Manual)
Equipment costs for recuperative incinerators depend on the total gas
flow through the incinerator to some power multiplied by a constant.
For 70 percent heat recovery, the equation is:

     EC - 21,342

     The minimum flow through the incinerator was assumed to be
500 scfm.   The equipment cost was multiplied by cost indices of
(357.5/340.1)  to correct equipment costs to October 1990,  dollars.
                                 D-25

-------
     Auxiliary Equipment (duct work, stack).  The costs for auxiliary
equipment were originally taken from an article in the May 1990,
Chemical Engineering and assuming 1/8 inch carbon steel and 24 inch
diameter with two elbows per 100 feet; and one elbow per source.   We
assumed that there would be 10 sources manifolded to a 300 foot
collection main.  The cost was adjusted on indices of (357.5/352.4).
     Auxiliary Collection Fan
     The auxiliary collection fan is sized on a minimum gas flowrate
of 500 scfm.  The equation is:
$ - 79.1239  [Total gas flow from Step 1  (d)]*5612 *
          (357.5/342.5)
     (based on the 1988 Richardson Cost Manual)
     Instrumentation     10 percent of purchased and auxiliary
                         equipment
     Sales tax           3 percent of purchased and auxiliary
                         equipment
     Freight             5 percent of purchased and auxiliary
                         equipment
     Total purchased equipment - sum of the above factors
     Direct costs * 30 percent of total purchased equipment
     Indirect costs - 31 percent of total purchased equipment
     Total Capital Investment.  If the maximum total gas flow is less
than 20,000 scfm, then the installation costs are 25 percent of the
purchased equipment costs.  If not, then the installation costs are
61 percent of the purchased equipment costs (from p. 3-51 of the OAQPS
Cost Manual).
     Annual!zed Costs
     Operator: $15.64/hr x 0.5 hr/shift x shifts/day x day/year
     (Assume that 3 shifts/day, 365 days/year)
     Supervisor: 15 percent of operator
     Maintenance: $17.21/hr x 0.5 hr/shift x shifts/day x day/year
     Material: 100 percent of maintenance
     Natural gas: Yearly natural gas usage (scfy) x  S3.3
                                                  1,000 scf
                                 D-26

-------
Electricity:
     From pages 3-55 of the OAQPS Cost Manual

     Power^    1.17 x 10"4 QtocAP
     fan  *    	=	
Where:

Qcot - maximum gas flow
AP   - pressure drop, in H20 (Assume 29 in H20; 19 inches
       for the preheater and 10 inches for ducting)
E    - efficiency
P    - power, in KW
$.059/kwh * electricity cost

Total Direct Costs:

Sum of labor, materials, natural gas, electricity

Indirect: Overhead: 60 percent of labor and materials
          Administrative: 2 percent of total capital investment
                     (TCI)
          Prop Tax: 1 percent of total capital investment
                     (TCI)
          Insurance: 1 percent of total capital investment
                     (TCI)

Capital Recovery Factor:
10 percent, 10-year life .16275 (TCI)
                            D-27

-------
THERMAL INCINERATOR MODEL SPREADHSEET
                 D-28

-------
 Al |l 2| III! KMAI  IN( INI KAIOK MODI I
 A4:|I2|TAHAMI-TI-KS:
 (JV
 JS: |WH| 'VOrii Controlled. (Mg/yr):
 Ad: Tlowraic, (scfm):
1<>: (I 2)|WH|OWMI 04S4/IIKIO
A7: 'Waste Gas V(X' Concentration, (ppmv):
I7:(,0)|W|S| KMNNI
AR: 'Healing Value of V(M'\ (Blu/stf):
CH:(.0)|Wl5|2INNI
J8: |W l^| 'CUM 1-lffectivcncM, ($/Mg):
A4>: 'l-ncrgy Recovery, (%):
A 10: 'Incinerator Operating Temperature, (F):
I !:(,() |W|S| KtOO
All: 'Waste (Sas 1 cmpcraiure, (!'):
I II:(,0)|WIS|70
J 1 1: |WI3| 'Maw l-mivikms, (Ih/yr):
A 12: Trchcalcr Tcmperalure, (l:):
J 12 (,(l)| Wn|2SUIO
AH: 'Mirfccular Weight of VOC:
I IV(I2)|W|S|64.S
A 14: 'Molecular Weight of (iav
I 14 (I;2)|WIS| +l-7/IINNNNMI>l:n + (l
J 14: |WI^| Time Variation:
A 15: 'Duraliim, (min):
I I5:(F2)|WI.S|0.6
JI5: (l'4)|Wn| <-MI'*(8.97'29.1)'(l/60)*(l/760)*(
A 16: 'I vents/Shift:
I Ki:(,0)|W|S|8
A 17: Shifts/Day:
I 17: (,0)|WI5|('IF(FI9<=0.66,((&'IF(FI9<=0.^,I,2)),1)
JI7:|W|^|>|-:ncrgy.(Hiu/Mg):
A 18: Days/Year:
I I8:(,0)|W|S|V5
JI8:(,0)|WI^| +I-7VJA
A I1': Time Variation:
A2II. 'I cnj;ih of ( 'ollcclion Main, (ft):
F20:(,)|W|S|MO
A2I  'Number of Manifolded Sources:
I2I:(.0)|W|S| Ml
A2V 'Step I: Calculate I  *

-------
A2r.: 'Dilution Air Required fur Combustion, (Ntf
l26:(,2)|WI';|(.||((l25/llllMI()W/V2)<(VtH': '    fr cvcnis
A40: 'Calculate Maximum Auxiliary (ias Mow, (scfm)
r40:(.2)|WI5|<.H|-U8MTO(0.2tt^l.lM;IO.|^^
A4I: 'Calculate Maximum Total (ias Mow, (scfm):
I4I:(,2)|WI5|-H-40+I7S
A42: 'Calculated Annual Gas IHow, (scfy):
I 42: (,) |WI5| ( + I;40'(l-ri9)+:%*l-|9)'*87f10
A45: |.:
A47: |l 2| 'CAPR AI.C()STCAI.CUI.AriC)NS:
A49: 'Direct Cts
A5II: 'l-quipmcnl CKI ( Rccouperalive lncin.),($):
l\MMC1l)|W|S|(a1l-XI;4l<5(W,(2IM2^(5fK))^(0.25))MW.5/^(M)),(2l242M(|-4l)-((K2S))'(^V.5/^^
A5 1: 'Auxiliary I equipment (Duct work, Slack), (S):
I 51: ((1l)|WI5|(((2IO24"0.8W)-f(2M.52*24A l.4^))*(l 20/HHt) + (l 2I'4.S2'24" I 4^))'(W S/3S2.4)
A52: 'Auxiliary Colteclton Fan, (S):
     'Insirumcnialion, ($):
A54: 'Sales lux, ($):
l54:(CO)|WI5|O.OV(l:50+T5
A55: 'lTcighl.'(S):
I55:(CO)|WI5|005'(I;5+I;5I + I'52)
A57: '101 Al  I'lIRt IIASI-DI-OIMP COST ($):
I 57: (( ll)|WI5| I*('SUM(I-50.K55)
A5'>: 'Dirccl Insiallalion Cosls, ($):
I;5:(C)|WI5|0.1M;S7
AMI  Indirctl Cosls, ($):
I-WI ((0)|W|5|OM'I 57
A(>2: "loial Capital Invcslmeni ( TCI), (\):
 A(4:|I2| ANNIIAI I I I) (OS I (Al (HI A I IONS:

-------
            APPENDIX E.

           MODEL PLANTS
               AND
MODEL EMISSION STREAM CALCULATIONS

-------

-------
                   i
is
if
             s

                                                               u


                                                               0)
                                                               iJ


                                                               C31
                               E-l

-------
12
                                        I
                                                                            0)
                                                                            I
                                                                            o
                                                                            CN
                                                                             i

                                                                            Dfl


                                                                            0)
                                      E-2

-------
          WEIGH
          TANK

,

WEIGH
TANK
                   MIX
                   TANK
                                        SOLVENT
Figure E-3.  Model batch process for liquid reaction
                         E-3

-------
                       oc
                       HI
II
LU
U
  i-
                                     ca
                                     0)
                                     M
                                     g.
                                     H
                                     Cu
         E-4

-------
            TABLE E-l.
                EMISSION STREAM  CHARACTERISTICS  FOR  SOLVENT REACTION MODEL  PROCESS
                                       WITH  ATMOSPHERIC DRYER
        of
     unM op per  CalcuUtioii fo* I
     model bMdi  MulanenltOB
                                   Kminiun
                                    Slrc.ro
                                    VOC
                    How Rale  Temp   Uuiilum   VK
                            (De| C)   (mm)   (vol %)
LoMuiom  NO CONTROL
It/Wci    MauFhu
(Note I)
                                                                                      CPC(Ne2)
       9.29
              (2000-ploB vcMel)
   LOW VOLATILITY       II      20      15     0.6%    0.29     164
MODERATE VOLATILITY    II      20      IS     12.6%    2.12     26.01
   HIGH VOLATILITY       II      20      IS     57.9%    30.02    277.66
              TOTAL DISPLACEMENT EMISSIONS    LOW VOLATILITY
                                           MODERATE VOLATILITY
                                             HIGH VOLATILITY
            2.64
           26.01
           277.66

           727.02
           7172.21
          76355.65
                                                                                                                      2.64
                                                                                                                      7.12
                                                                                                                      30.54

                                                                                                                     727.02
                                                                                                                     2151.61
                                                                                                                     1399.12
M
 I
l/l
REACTORS

Ch*lfM| W/jplMf C




Heat-up W^WfC



Reaction
              Empty React
              TOTAL REACTOR EMISSIONS
   LOW VOLATILITY
MODERATE VOLATILITY
   HIGH VOLATILITY

   LOW VOLATILITY
MODERATE VOLATILITY
   HIGH VOLATILITY

   LOW VOLATILITY
MODERATE VOLATILITY
   HIGH VOLATILITY

   LOW VOLATILITY
MODERATE VOLATILITY
   HIGH VOLATILITY

   LOW VOLATILITY
MODERATE VOLATILITY
   HIGH VOLATILITY
30
30
30
30
30
30
ISO
ISO
ISO
100
100
100



20
20
20
20 to 30
20 lo 30
20 lo 30
37
37
37
20
20
20



IS
15
15
5
5
5
3
3
3
1
1
1



0.1%
3.1%
I4.S%



0.5%
7.5%
27.2%
0.1%
2.7%
12.1%



0.12
1.17
12.51
O.I
0.52
5.13
0.39
2.65
22.19
0.06
0.5
IJ



a 12
1.17
12.51
a 10
0.52
5.13
a 39
2.65
22.19
aw
a so
1.50



0.12
1.17
12.51
0.10
0.52
5.13
0.39
2.65
22.19
0.06
0.50
1.50
II4.M
1332.65
11365 JO
                                 0.12
                                 1.17
                                 5.63

                                 0.10
                                 0.52
                                 1.15

                                 0.39
                                 1.35
                                 5.33

                                 0.06
                                 0.32
                                 1.26

                                 114.01
                                 925.91
                                3K6.79

-------
                                                    TABLE  E-l.    (continued)


    of                                           Emiuioi                                               EniMMM  NO CONTROL             CPC(Noie2)
MM apt per   CtlmUtiai for I                           Sucm         How Rue  Temp.  Duration   VOC   EmiitioM IbiAMch    MMtHw                MuiHux
oddhtfdi   *openlk                             VOC           *<*
          CENTRIFUGES

          LMduAiMi w/neitMt              LOW VOLATILITY        3       20     30     0.1%    0.02     0.05        O.OS                    0.05
                                         MODERATE VOLATILITY     3       20     30     3.1ft    023     0.47        0.47                    0.47
                                            HIGH VOLATILITY       3       20     30     14.5%    2.50     3.00        500                    2.25

          C*ke amniAnkwdinc                 LOW VOLATILITY       20      20      3      0.1%    002     0.09        0.03                    003
          w^MTfc- 3mm                   MODERATE VOLATILITY    20      20      3      3.1%    016     0.31        0.31                    031
                                            HIGH VOLATILITY      20      20      3      14.5%    1.67     3.34        3.34                    1.50

          TOTAL CENTRIFUGE EMISSIONS       LOW VOLATILITY                                                       21.13                    2113
          (ta-lMch/yeM)                   MODERATE VOLATILITY                                                    2I5.M                   215.51
                                            HIGH VOLATILITY                                                     2292.96                  1031.13


          DRYERS
          Tray Dfycf
          CoMveaive-Mot of cycle               LOW VOLATILTTY      6000     65     60     a 3%    200    200.00      200.00                    2000
                                         MODERATE VOLATILITY   6000     65     60     01%    200    200X10      200.00                    20.00
                                            HIGH VOLATILITY     6000     65     60     0.3%    200    200.00      200.00                    2000
          Canredive-MtUfc of cycle              LOW VOLATILITY      6000     65     240     ttl%     IM    110.00      110.00                    11.00
                                         MODERATE VOLATILITY   6000     65     240     0.2%     IK    IM.OO      IMLOO                    1100
                                            HIGH VOLATILITY     6000     65     240     0.1%     110    I WOO      IMOO                    11.00
          Canveciive-eatfafcyck                LOW VOLATILITY      6000     65     60     ttO%     20     20.00      20.00                    200
                                         MODERATE VOLATIUTY   6000     65     60     0.0%     20     20.00      20.00                    2.00
                                            HIGH VOLATILITY     6000     65     60     0.1%     20     20.00      20.00                    2.00

          TOTAL DRYER EMISSIONS            LOW VOLATIUTY                                                       110000                   11000
          (ftM4>McVye*r)                   MODERATE VOLATILITY                                                    110000                   HOOD
                                            HIGH VOLATIUTY                                                      110000                   11000

-------
SSI MS
                       !
 Oil
 10 1
 600
 ZSO
 100
WWI
0VI
 96 J
 CCO
 KO
 iO'o
09 Wl
OT91
 96 J
 II0
 KQ
 100
t*?i'
>^I

-n
HO

mn
100
%009
                                                                                                       10HLNOO IVOLLnaOVWl VHd 1NTOHTO
                               AUIimOA HOI11
                             AirmVTOA 3J.VJHOOW
                               AinUVTOA *(/1

                               AU.1U.V10AHOIH
                                                                                                                                                i
                                                                                                                                               w
                                  ..
                             AinilVTOA aiVJHOOW
                               AJJIlLVTOAMCri
                               AJLHULVTOA/Wn
SNOISSiWg NOUVTILLSKnVXQl

       1 dis <*W?S  li.do -WMUIY
                 -.lo otuiv
                 . ii.do WMM y
                   ,
       I dans dmms - M 
       I iS (hums  u.do
                                                                                                       (ii ION) Nonmusw Haivg
                                                        (HIHI)
                    TOH1NO3ON
                                                               3OA
                                                              uwwrs
                                                             UultllUTj

-------

               Ml
               ^ M -
                       c: t 3  2 R
                       e  in  o b
 Rwtcv  *
 nm  p
e  in  b
                                    8 
                                    S^

                                     <^
                            25?=
                            eom
                                      b b -

<
W

Cd


J

O
CO
      6 c
] s
    a R
                       b - 2  b b i  b  a  b b -
                       5 I !

                                  K 55  ef r4 2
CO
u
rH ^

<*>
w -1*
uw
*. 03



%%

CO

CO
H
      If  II
     3
      d.
          838
      ]i-
  ESS
                       III  III  III  III
                            a s s
                       338  225  SRR
                            888
                       SSS  SSS  8 S ?  888

                                       i
                                            2 p


                                           ill
U


U
      *j:


       1 2

      ll

    s!  5
                       I    I    !
                       I    I    1
                         E-8

-------
                                                       TABLE  E-2.    (continued)
       of
    MM apt per  Cilcvbiiai for I
    model btfcfc  Mil operation
                                       Emiuion
                                        Strewn
                                         VOC
                       Flow Rue  Temp.   Pieiiure  Duration   Vl>C
                        (cfm)   (DejQ  (nun Hg)  (m.)   (vol%)
      Kmittiani NO CONTROL  
-------
                                                       TABLE  E-2.     (continued)
   tot

IBM cfM per

MOddlMlCll  UMt
                          for I
                                                          VOC
                       Flow Rile  Temp.   Prctture  Duration   VOC
                        (dm)   (DtjC)  (mm H|)   (min)   (vol%)
                                      NO CONTROL   CPC(Nae2)
                                        MM. Fhu      Matt Pha
                              (Noel)    HM/bMdi      >*<*
M
          BATCH DISTILLATION

          Alma*, op'a - Surtup Step I
          Alma*, op'" - Startup Step 2
          Almot. ap'a - Suitup Stop I
          Alma*, ap'n - Startup Stop 2
          AUIKM. op'* - Sunup Stop I
          Atom, op'a - Startup Stop 2

          TOTAL DISTILLATION EMISSIONS
   LOW VOI-ATIIJTY        1.4
   LOW VOLATILITY        12.6
MODERATE VOLATIIJTY     55
MODERATE VOLATIIJTY     14.4
   HIGH VOLATILTIY        28.9
   HIGH VOLATILTIY        39 9

   LOW VOLATIIJTY
MODI-RATE VOLATIIJTY
   MIOH VOLATILITY
                                                                                   25

                                                                                   25
                                                                                   25
760
760
760
760
760
760
5
60
5
45
5
30
0.7%
0.7%
14.3%
16.1%
60.0%
72.1%
0.01
0.52
0.33
596
16.4
164.6
.OI
0.52
033
5.96
16.40
164.60
 0.01
 0.52
 0.33
 5.9*.
 16.4)
 1641

 145 5
1729/5
 49775
 001
 052
 009
 2.03
 180
 14.11

 145.75
5*0.855
456995
      No* I:
      Attune I
      Note 2:
      CPC = CURRENT PHARMACEUTICAL CONTROL

-------
                           TABLE E-3.   EMISSION STREAM  CHARACTERISTICS  FOR
                                         LIQUID  REACTION  MODEL  PLANT
   of
IM* op* per   Cakvltf ioa far I
model (Mch   (MM opentioa
  67)
          TOTAL DISPLACEMENT EMISSIONS
Eminion
&!
VOC
LOW VOLATILITY
MODERATE VOLATILITY
HIGH VOLATILITY
LOW VOLATILITY
MODERATE VOLATILITY
HIGH VOLATILITY

Flow Rue
(elm)
II
II
II




Tanp.
(Dr. Q
20
20
20




DuittiOB
(min)
15
IS
IS




VOC
(vd%)
06%
12.6%
57.9%




EiwioM
fcfcvc*
029
2.12
30.02



EmiuioM
MM/b*dl
(Ntfcl)
1.79
17.62
1*7.61



NO CONTROL
MM*HH
kWbMdi
1.79
17.62
117.61
491.23
4*46.14
51591.65
CPC(Nole2)
MuiHw
IMMdl
1.79
5.29
20.64
491.23
1453.14
5675.0*
          REACTORS
   I      Empty Reactor PM|Mf
          TOTAL REACTOR EMISSIONS
  LOW VOLATILITY
MODERATE VOLATILITY
  HIGH VOLATILITY

  LOW VOLATILITY
MODERATE VOLATILITY
  HIGH VOLATILITY

  LOW VOLATILITY
MODERATE VOLATILITY
  HIGH VOLATILITY

  LOW VOLATILITY
MODERATE VOLATILITY
  HIGH VOLATILITY

  LOW VOLATILITY
MODERATE VOLATILITY
  HKJH VOLATILITY
30
30
30
30
30
30
150
150
150
100
100
100



20
20
20
20 lo 30
20 to 30
20lo30
37
37
37
20
20
20



15
15
15
5
5
5
3
3
3
1
1
1



0.1%
3.1%
14.5%



(15%
7.5%
27.2%
0.1%
17%
12.1%



0.12
1.17
12.51
0.1
0.52
5.13
OJ9
2.65
22.19
0.06
05
1.3



0.12
1.17
12.51
0.10
052
5.13
0.39
245
22.19
0.06
050
150



012
1.17
12.51
010
052
5.13
O39
2.65
22.19
O06
OJO
1.50
IM.M
1332.65
1136550
0.12
1.17
5A3
0.10
0.52
1.15
0.39
1.35
5.33
006
032
1.26
1*401
925.91
3166.79

-------
                                                                          AJJ1LLVTOA HOIH
** 01*        5 61                                                      11 iii i ir*iA 31 yuan rim                   <* ifr>ni    -i                ^
sn         ,t,r.                                                         ,,,w^* Minm       s^)^ssIwaNoavniLLSKnvioi
 Ifl         09 Wl      09WI     VWI    *!                 6      AU.1U.VTOA HOIH
 011         *91      <*     ^i     OT                     6R      AUL1U.VTOAHOIH              Ib^*ili.is-[*TiV      1
 10 Z          %S       ws     %s     **91     S      $t      l   AiniLVTOA UVH3K1ON           cd*SN) NOlLVTllLStd HOLVB
                                                (wui)   O*a)    (WJM)           DOA
                                        30A
          TOH1NO3ON
                                                                       --g

-------
                                 TABLE  E-4.    EMISSION STREAM CHARACTERISTICS  FOR
                                                   FORMULATOR  MODEL PLANT
             of
          unit dpi per
          MOddlMldl

            4.5
&lculMfarl
MtopeoUM
                     TOTAL DISPLACEMENT EMISSIONS
Kmittion
 SlICMI
 VOC
Flow Rue   Tnq>.   Duration   VOC   En*
        (DtC)   (mm)
                                  IX)W VOLATILITY        II
                               MODERATE VOLATILITY     II
                                  HIGH VOLATILITY        II

                                  LOW VOLATILITY
                               MODERATE VOLATILITY
                                  HIGH VOLATILITY
                                                                                    20
                                                                                    20
                                                                                    20
                               15
                               I)
                               I)
                       0.6%    0.29
                       124%    ZK
                       57.9%    30.02
Eariuiam  NO CONTROL
MM***
(Note I)
  1.29
  12.69
 135.08
  1.29
 12.69
 13501

 353.69
 3419.22
37145.99
                     REACTORS
M
 i
I-*
UJ
                     TOTAL REACTOR EMISSIONS
                                  LOW VOLATILITY
                               MODERATE VOLATILITY
                                  HIGH VOLATILITY

                                  LOW VOLATILITY
                               MODERATE VOLATILITY
                                  HIGH VOLATILITY

                                  LOW VOLATILITY
                               MODERATE VOLATILITY
                                  HIGH VOLATILITY

                                  LOW VOLATILITY
                               MODERATE VOLATILITY
                                  HIGH VOLATILITY

                                  LOW VOLATILITY
                               MODERATE VOLATILITY
                                  HIGH VOLATILITY
30
30
30
30
30
30
150
150
150
100
too
100



20
20
20
20 to 30
20 to 30
20 lo 30
37
37
37
20
20
20



15
15
IS
5
5
S
3
3
3
1
1
1



01%
3.1%
143%



0.5%
7.5%
27.2%
0.1%
2.7%
12.1%



0.12
1.17
1251
0.1
tt 52
5.13
0.39
165
22.19
0.06
OJ
IJ



0.12
1.17
12.51
0.10
. 0.52
5.13
0.39
2.65
22.19
0.06
0.50
IJO



a 12
1.17
12.51
a 10
052
5.13
a 39
165
22.19
0.06
0.50
1.50
IM.M
1332.65
11365.50
      Note I:
      Airame I

-------
 TABLE E-5.  ASSUMPTIONS FOR  SOLVENT REACTION MODEL PLANT
                   WITH ATMOSPHERIC DRYER
Basis:

     A.  Equipment required for each solvent reaction model batch process

         1 reactor @ 2,000 gallons
         2 weigh tanks @ 1,000 gallons
         1 mix tank @ 2,000 gallons
         1 crystallizer @ 3,000 gallons
         1 slurry tank @ 3,000 gallons
         2 centrifuges @ 200 ft3 each
         1 distillation unit @ 2,000 gallons
         1 solvent recovery tank @ 1,500 gallons
         1 atmospheric dryer @ 300 ft3


     B.  Operation

         Small plant has 3 "model batch process"
         Medium plant has 10 "model batch processes"
         Large plant has 30 "model batch processes"

         Each batch is run 1 X per day
         Plant operates 275 days per year


     C.  Chemistry

              For calculations:
                                vapor pressure equivalent to:
         Low volatility solvent              n-butanol
         Moderate volatility solvent          methanol
         High volatility solvent             ether
                             E-14

-------
    TABLE E-6.  ASSUMPTIONS FOR  SOLVENT REACTION MODEL PLANT
                        WITH VACUUM DRYER
Basis:

     A.  Equipment required for each solvent reaction model batch process

         1 reactor @ 2,000 gallons
        2 weigh tanks @ 1,000 gallons
         1 mix tank @ 2,000 gallons
         1 crystallizer @ 3,000 gallons
         1 slurry tank @ 3,000 gallons
        2 centrifuges @ 200 ft3 each
         1 distillation unit @ 2,000 gallons
         1 solvent recovery tank @  1,500 gallons
         1 vacuum tray dryer @ 300 ft3


     B.  Operation

        Small plant has 3 "model batch process"
        Medium plant has  10 "model batch processes"
        Large plant has 30 "model batch processes"

        Each batch is run 1 X per day
        Plant operates 275 days per year


     C.  Chemistry

              For calculations:
                               vapor pressure equivalent to:
        Low volatility solvent              n-butanoi
        Moderate volatility solvent           methanol
        High volatility  solvent              ether
                               E-15

-------
      TABLE E-7.  ASSUMPTIONS  FOR LIQUID REACTION MODEL PLANT



Basis:

      A.  Equipment required for  each solvent reaction model batch process

         1 reactor @ 2,000 gallons
         2 weigh tanks @ 1,000 gallons
         1 mix tank @ 2,000 gallons
         1 surge tank @ 3,000 gallons
         1 distillation unit @ 2,000 gallons
         1 solvent recovery tank @ 1,500 gallons


      B.  Operation

         Small plant has 3 "model batch process"
         Medium plant has 10 "model batch processes"
         Large plant has 30 "model batch processes"

         Each batch is run  1 X per day
         Plant operates 275 days per year


      C.  Chemistry

              For calculations:
                                vapor pressure equivalent to:
         Low volatility solvent              n-butanol
         Moderate volatility solvent          methanol
         High volatility solvent              ether
                                E-16

-------
     TABLE E-8.  ASSUMPTIONS FOR FORMULATION MODEL PLANT



Basis:

     A.  Equipment required for each fomulation model batch process

         1 reactor @ 2,000 gallons
        2 weigh tanks @ 1,000 gallons
         1 mix tank @ 2,000 gallons
         1 surge tank @ 3,000 gallons
         1 closed in-line process filter (no emissions)


     B.  Operation

        Small plant has 3 "model batch process"
        Medium plant has 10 "model batch processes"
        Large plant has 30 "model  batch processes"

        Each batch is run 1 X per day
        Plant operates 275 days per year


     C.  Chemistry

             For calculations:
                               vapor pressure equivalent to:
        Low volatility solvent              n-butanol
        Moderate volatility solvent          methanol
        High volatility solvent              ether
                             E-17

-------
TABLE  E-9.
EMISSIONS  FROM SOLVENT REACTION MODEL PLANT
      WITH ATMOSPHERIC DRYER
                       MODEL PLANT EMISSIONSflbs/yr)
               LOW VOLATILITY
            MODERATE VOLATILITY
               HIGH VOLATILITY
                    SMALL/NC
                     333,236
                     361350
                     749367
 SMALL/CPC
   36.236
   44,621
   86.603
               LOW VOLATILITY
            MODERATE VOLATILITY
               HIGH VOLATILITY
                   MEDIUM/NC
                     1,109,676
                     1.203.296
                     2,495393
MEDIUM/CPC
   120,666
   148,590
   288388
               LOW VOLATILITY
            MODERATE VOLATILITY
               HIGH VOLATILITY
                    LARGE/NC
                     3332360
                     3.613302
                     7,493.673
 LARGE/CPC
   362360
   446,215
   866,031
                       MODEL PLANT EMISSIONS(tons/yr)
               LOW VOLATILITY
            MODERATE VOLATILITY
               HIGH VOLATILITY
                    SMALL/NC
                      166.62
                      180.68
                      374.68
 SMALL/CPC
    18.12
    22.31
    43.30
               LOW VOLATILITY
            MODERATE VOLATILITY
               HIGH VOLATILITY
                   MEDIUM/NC
                      554.84
                      601.65
                      1247.70
MEDIUM/CPC
    60.33
    74.29
    144.19
               LOW VOLATILITY
            MODERATE VOLATILITY
               HIGH VOLATILITY
                    LARGE/NC
                      1666.18
                      1806.75
                      3746.84
 LARGE/CPC
    181.18
    223.11
    433.02
            NC No Control
            CPC  Current Pharmaceutical Control
               1. For surface condensers on sources emitting:
                -25C for VP>300mmHg
                -15C for 150
-------
TABLE  E-10.
EMISSIONS  FROM SOLVENT  REACTION MODEL  PLANT
        WITH VACUUM DRYER
                       MODEL PLANT EMISSIONS(lbs/yr)
               LOW VOLATILITY
            MODERATE VOLATILITY
               HIGH VOLATILITY
                   SMALUNC
                     180.611
                     208.725
                     596.742
SMALL/CFC
   20.974
   29,359
   71341
               LOW VOLATILITY
             MODERATE VOLATILITY
               HIGH VOLATILITY
                   MEDIUM/NC
                     601.435
                     695.055
                     1.987,152
MEDIUM/CPC
   69,842
   97.765
   237.564
               LOW VOLATILITY
             MODERATE VOLATILITY
               HIGH VOLATILITY
                    LARGE/NC
                     1,806.110
                     2.087.252
                     5.967,423
 LARGE/CPC
   209.735
   293.590
   713.406
                       MODEL PLANT EMISSIONS(tons/yr)
                LOW VOLATILITY
             MODERATE VOLATILITY
                HIGH VOLATILITY
                    SMALUNC
                      90.31
                      104.36
                      298.37
 SMALL/CPC
    10.49
    14.68
    35.67
                LOW VOLATILITY
             MODERATE VOLATILITY
                HIGH VOLATILITY
                   MEDIUM/NC
                      300.72
                      347.53
                      993.58
 MEDIUM/CPC
    34.92
    48.88
    118.78
                LOW VOLATILITY
             MODERATE VOLATILITY
                HIGH VOLATILITY
                    LARGE/NC
                      903.06
                      1043.63
                      2983.71
  LARGE/CPC
    104.87
    146.79
    356.70
             NC No Control
             CPC  Cunent Pharmaceutical Control
                1. For surface condensers on sources emitting:
                 -25CforVP>300mmHg
                 -15C for 150
-------
TABLE  B-ll.    EMISSIONS FROM LIQUID REACTION. MODEL PLANT
                     MODEL PLANT EMISSIONS(lbs/yr)
              LOW VOLATILITY
           MODERATE VOLATILITY
              HIGH VOLATILITY
 SMALL/NC
   2,463
  23,726
  338.196
 SMALL/CPC
   2.463
   8,882
   42335
              LOW VOLATILITY
           MODERATE VOLATILITY
              HIGH VOLATILITY
MEDIUM/NC
   8,202
   79.006
  1.126.194
MEDIUM/CPC
   8002
   29.576
   140,977
              LOW VOLATILITY
           MODERATE VOLATILITY
              HIGH VOLATILITY
 LARGE/NC
   24.632
   237 056
  3381.965
 LARGE/CPC
   24.632
   88.818
   423355
                 MODEL PLANT EMISSIONS(tons/yr)
               LOW VOLATILITY
            MODERATE VOLATILITY
               HIGH VOLATILITY
 SMALL/NC
    1.23
    11.86
   169.10
 SMALL/CPC
     1.23
     4.44
    21.17
               LOW VOLATILITY
            MODERATE VOLATILITY
               HIGH VOLATILITY
 MEDIUM/NC
    4.10
    39.50
   563.10
 MEDIUM/CPC
     4.10
     14.79
    70.49
               LOW VOLATILITY
            MODERATE VOLATILITY
               HIGH VOLATILITY
 LARGE/NC
    12.32
    118.63
   1690.98
 LARGE/CPC
    1232
    44.41
    211.68
           NC No Control
           CPC  Cunent Pharmaceutical Control
              1. For surface condensers on sources emitting:
               25CforVP>300minHg
               15C for 150
-------
TABLE E-12.  EMISSIONS FROM FORMULATION MODEL  PLANT
                    MODEL PLANT EMISSIONSflta/yr)
                  LOW VOLATILITY
               MODERATE VOLATILITY
                  HIGH VOLATILITY
 SMALL/NC
   1.613
   14.466
   145,534
                  LOW VOLATILITY
               MODERATE VOLATILITY
                  HIGH VOLATILITY
MEDIUM/NO
   5.372
   48.170
  484.630
                  LOW VOLATILITY
               MODERATE VOLATILITY
                  HIGH VOLATILITY
 LARGE/NC
   16,133
   144.656
  1.455345
                  MODEL PLANT EMJSSIONS(tonVyr)
                  LOW VOLATILITY
               MODERATE VOLATILITY
                  HIGH VOLATILITY
 SMALUNC
   0.81
   7.23
   72.77
                 LOW VOLATILITY
               MODERATE VOLATILITY
                 HIGH VOLATILITY
MEDIUM/NC
   2.69
   24.09
  242.31
                 LOW VOLATILITY
               MODERATE VOLATILITY
                 HIGH VOLATILITY
 LARGE/NC
   8.07
   72.33
  727.67
            NC-No Control
                            E-21

-------
350 cral

ft3
7.48 gal

15 min
                Model Emission Stream Calculations
OPERATION REFERENCE  2.1.1
Reactors
2.1.1.2.1  charging  without purge
    Assume reactor volume of 500 gallons  (350 gallons to fill)
    Filling occurs at 20C  (Room Temperature)
    Flowrate
                                          ft3/min
                              cnin
A.  Low volatility (n-Butanol)
  -  % VOC
    4.4 mmHg/760 mmHg - 0.0058 - 0.6%
    Total Ib VOC. event
    (4.4 mmHg)(3.12  ft3/min)  (74 Ib/lbmol)(15 min/event)
       998.97 mmHo-ft3  (293K)
          IbmolK
    - 0.05 Ib/event
B.  Medium volatility (methanol)
    % VOC
    92 mmHg/760 mmHg - 0.121 - 12.0%
    Total Ib VQC. event
      (92 mmHa)(3.l2 ft3 /min)  (32 Ib/Ibmol)(15 min/event)
        f 998.97 mmHa-ft3;  (293K)
             IbmolK
       0.5 Ib/event
C.  High volatility  (ether)
    % VOC
    442 mmHg/760 mmHg  0.582 - 58.0%
                               E-22

-------
    Total Ib VOC. event
       (442 mmHo)(3.12  ft^/min)  (74  Ib/lbmol)(15 min/event)
          998.97 mmHa-ft3  (293K1
              IbmolK
      - 5 Ib/event
2.1.2.2.1  gha,rgincr with purge
    flow rate out of reactor  * purge  rate
A.  Low volatility  (n-Butanol)
    % VQC
    4.4 mmHg/760 mmHg  - 0.0058   0.6%
    Assume 10% of saturation
    (0.10) (0.6) - 0.06%
    Total Ife VOC. evens
    (760 nrnHcrl (0.00061 (30  ft3/min)   (74  Ib/lbmol) (15 min)
      998.97 mmHer-ft3  (293K)
         IbmolK
    - 0.05 Ib/event
1.  Medium volatility  (methanol)
    % VQC
    92 mmHg/760 nsnHg - 0.121  - 12.0%
    Assume 10% of saturation
    (0.10)(12.0)  1.2%
    Total Ib VOC. event
     (760 mmHcr) (O.Q12) (30  ft3/min)  (32  Ib/lbmol} (15 min)
      998.97 mmHQ*ft3  (2931O
         IbmolK
      0.45 Ib/event
                               E-23

-------
C.  High volatility  (ether)
    % VOC
    442 mmHg/760 mmHg - 0.582 - 58.0%
    Assume 10% of saturation
    (0.1) (58.6) - 5.8%
    Total Ib VOC. event
     (760 rnmHo)(0.058)(30 ft3/min)  (74 Ib/lbmol)(15 min)
       998.97 mrnHa-ft3  (293K)
         Ibmol-K
     - 5.0 Ib/event
2.1.1.2.2  Heatup without purge
    Assume reactor volume is 500 gallons.   (Headspace is
    150 gallons)
A.  Low volatility (n-Butanol)
    Flow rate. Emissions
     VP of n-butanol 9 30 percent - 17.8 mmHg
       (150 aal)   ft
An   	7.48 aal   j (  (760 - 4.4) mmHcr -  (760 - 17.8) mmHoj
         998.97 mmHa-ft3   '  (273 + 20)K         (273 + 30)K
               IbmolK
    An - 0.0026 Ibmoles gas displaced
     0.0026 Ibmole gas displaced
379 ff
                                  Ibmol
         1 ff3 displaced
    1 ft3/5 min  0.2 ft3/min displaced  (average  flow  rate)
nt - r     4.4    i mmHg + <     17.8    > mmHg   < 0.0026 Ibmoles-,
       760 - 4.4            760  - 17.8          l gas  displaced
                       2
    nt - 0.000039 Ibmoles n-Butanol  (MW  - 74 Ib/lbmol)
   (0.000039 Ibmol n-butanol)(74 Ib/lbmol)
    nt m 0.003 Ib n-Butanol
                               E-24

-------
B.  Medium volatility (methanol)


    Flow  rate.  Emissions

     VP of methanol   30C -  165  mmHg
        (150 gal)(	)
       	7.48  aal ^  ( (760  -  92)  mmHa -  (760 -  165)
         998.97 mmHa-ff      (273  +  20)K       (273 + 30)K
             IbmolK
    AIT.  0.00634 Ibmoles gas  displaced
     0.00634 Ibmolea
379 ff
                     Ibmol
-0.48 ft3/min
         5 min
             92                 165
                  ) mmHg +  ;	] mmHg
          760 - 92           760  -  165
    n  - 	
     t
    nt -  (0.0013 Ibmol methanol) (32 lb/ Ibmol)


    nt - 0.0013 Ibmoles methanol  (MW  32 Ib/lbmol)


    nt - 0.04 lb methanol


C.  High volatility  (ether)

    Flowrate. Emissions

     VP of ether  30C - 661 mmHg
         f fl5Q qal)  7-4fl ga    i  t
              998.97 mmHa.ff*        (273+20)K        (273+30)K
               Ibmol 'K
         0.0152 Ibmoles gas displaced
                               E-25

-------
      f    442   1  mmHg  '    661    ; mmHg   rO.0152  Ibmolesi
       760 - 442	760 - 661	  gas displaced
    nt  0.0613 Ibmoles ether  (MW - 74 Ib/lbmol)


    (0.0613 Ibmol ether)(74 Ib/lbmol)


    nt - 4.5 Ib ether


2.1.1.2.2  Heatup with purge


    Assume flow rate - 30 acfm and 10% of saturation


  - Temperature increases from 20 to 30C


A.  Low volatility  (n-Butanol)

      (0.10) (30 acfm) (4.4 mmHa) _ 0>000045 lbmol/min  (initial)
      998.97 mmHg-ft3,,Q7in
      	Ibmol-K	J (293K)

      (0.10)(30 acfm)(17.8 mrnHg) m Of000176  ibmoi/min   (final)
      998.97 mrnHg-ft > ,,n,v
      	Ibmol-K     J (303K)


    Calculate average


      (0.000045 + 0.000176)  Ibmol/min
                    2


    - 0.00011 lbmol/min n-butanol  (MW    74  Ib/lbmol)
      0.00011 Ibmol n-butanol
                mm
74 Ib
Ibmol
       (0.008 Ib/min)(5 min)    0.04  Ib  n-butanol
                               E-26

-------
B.  Medium volatility  (methanol)
      (0.10) (30 cfm)(92 mmHg) m 0.000943  lbmol/min

      998.97 mmH (293K)


      (0.10) (30 acfm)(165 mmHgJ , 0.00164  lbmol/min (final)
     i 998.97 mmHg.ft^ {3Q3K)
         Ibmol-K      J (303K)
    Calculate average


     (0.000943 + 0.00164) Ifcmol/min
                   2
    - 0.00X3 lbmol/min methanol  (MW - 74 Ib/lbmol)
      0.0013 Ibmol methanol! 74 lb
               nun          I Ibmol

      (0.04 lb/min) (5 min)   0.2 lb methanol
C.  High volatility  (ether)

     (0.10) (30 acfm)(442
      998.97 mmHg.ft-%
                ^     J
                                        lbmol/min
     (0.10) (30 acfm)(661 mmHg) m Q ^^ lbmol/min  (final)
    Calculate average

     (0.0045 + 0.0066) lbmol/min
                 2

    - 0.00555 lbmol/min ether  (MW - 74  Ib/lbmol)
     0.00555 Ibmol ether
                     74 lb
                     Ibmol

- (0.411 lb/min)(5 min) - 2 lb ether

                           E-27

-------
2.1.1.2.2  Reaction with purge
    This event actually is the purging of a reactor prior to
charging (or sampling).
    Assume temperature - 310K and 10% cf saturation

A.  Low volatility (n-Butanol)
    VP n-butanol  310K - 16 mmHg
     (0.10)(150 acfm)(16 mmHg) _ 0_OOQ77 lbmol/min
     , 998.97 mmHg-ff% ,,,nin
      	Ibmol-K(310K)
    - 0.00077 lbmol/min n-butanol (Mtf - 74 Ib/lbmol)
      0.00077 Ibmol n-butanol|74 Ib
                min           !Ibmol
    - (0.057 Ib/min)(3 min) - 0.17 Ib n-butanol
B.  Medium volatility  (methanol)
    VP methanol at 310K - 229 mmHg
      (0.10) (150 acfm)(229 mmHa) , 0>(m lbmol/min
      998.97 mmHg-ft3  ,*,nv\
      - Ibmol -K     '  (310K)
    - 0.011 lbmol/min methanol  (MW - 32 Ib/lbmol)
      0.011 Ibmol methanol: 32 Ib
              min          j Ibmol
    - (0.355 Ib/min) (3 min) - 1.1 Ib methanol

C.  High volatility  (ether)
    VP ether at 310K - 760 inmHg
 -     (0.10X150 acfm)(760 mmHg) m 0.03 6 8 lbmol/min  (initial)
     .998.97
         Ibmol-K
    - 0.0368 lbmol/min ether  (MW  -  74  Ib/lbmol)

                               E-28

-------
      0.0368 Ibmol ether|74 Ib
              minSIbmol
     (2.72 Ib/min)(3 min) - 8.2 Ib ether

2.1.1.2.3  Reactoy vacuum transfer
    Vacuum transfer would typically occur when transferring the
contents of a 55 gallon drum to a reactor.  The emissions would
result from displacing air saturated with VOC's out of the
reactor prior to drawing in new product.
    Air displaced;  assume 500 gallon reactor
A.  Low volatility (n-butanol)
    Initial air
      (760 ntmHg) (500 gal/7.48 gal/ft3) m 0>1735 ibmol
     ,998 97 mmHs-ft3, (293R)
         lomol  K
     Final air
      (100 tnmHg) (500 gal/7.48 gal/ft3) m QfQ22B Ibmol
      998.97 mmHg-ft3  ,-a,
           ii  ,       ,; \4y3fi*l
         Ibmol  K
    0.1735 Ibmol  - 0.0228 Ibmol  0.1507  Ibmol
    Flow rate
     0.1507 Ibmol
     10 mm
5 JQ ft- ^         -3
T'-LT    - 5.7 ftVmin
Ibmol
VOC emissions
A.  Low volatility  (n-butanol)
    VP  20C  -  4.4 mmHg
    Assume  saturation?  volt  ranges from 0.6 to 4.4
    Total VOC  emissions based on average volt
     (0,6 +  4.4)/2 - 2.5 volt
     (0.1512 Ibmol)(0.025)(74 Ib/lbmol)  - 0.3 Ib n-butanol
                               E-29

-------
B.  Medium volatility  (methanol)
    VP  20 C - 92 rnmHg
    92 mmHg/760 mmHg - 12.1 volt
    92 mmHg/100 mmHg  92 volt
    Total VOC emissions based on average volt
    (12.1 + 92} 12 - 52.1 volt
    (0.1512 Ibmol) (0.52) (32 lb/lbmol) - 2.S Ib methanol
C.  High volatility (diethyl ether)
    VP m 20C - 442 mmHg
    442 mmHg/760 mmHg - 58.2%
    Assume 100 volt at 100 mmHg
    (58.2 + 100)/2  79.1 volt
    (0.1512 Ibmol) (0.791) (74 Ib/lbmol) - 8.9 Ib ether
2.1.1.2.3  Pressure transfer
    Pressure transfers often consist of "blowing" lines to rid
them of solvent.  Assuming a typical situation involves 30 ft of
3.5 inch flexible line containing it residual solvent, the amount
of solvent evaporated from each line is:
                 _  3.5 in    2
                   ' 12 in./f'tJ  (30 ft) - 2 ft3 material
                       2
    (2 ft3) (0.01) - 0.02 ft3 in liquid form
A.  Low volatility (n-butanol)
    VP * 20C - 4.4 mmHg
    Specific gravity - 0.81
            l50 ft w 4. 4 rnrnHg-! ,_j_i
               - H760     l (mm) -
       0.02

                           760 mmHg
    0.087 x min - 5.26 ft3 gas
    x - 60 minutes  (assuming lot saturation of the  stream)
                               E-30

-------
B.  Medium volatility  (methanol)
    VP  20C - 92 mmHg
    Specific gravity - 0.792
       0.02 tt(i (0.792)(l(-l (293K)
    1,82 x min - 11.9 ft3 gas
    x  6 . 5 minutes
C.  High volatility  (diethyl ether)
    VP 9 20C - 442 mmHg
    Specific gravity - 0.8
     in in  15gt3' ' 442
     (0'10)
                .
              mn   760 mmHg
       0 02 ft3-62-4 lb: {0 B)^lbmolw 998.97
       0.02. tt ,        M0.8)
                                        lbmol-K
    8.72 x min - 5.2 ft3 gas
    x - 0.596 minutes
2.1.1.2.3  Empty reactor purging
A.  Low volatility  (n-butanol)
     c     m  (4.4 mmHg)(74 Ib/lbmol)
      10Wi   r 998.97 mmHg-ft3j (293K)
                 Ibmol  K
    - 0.0011 lb/ft3
    Standard industry practice  (Chapter 3)
    (500 gal)(ft3/?.48 gal) - 67 ft3
    100 ft3/67 ft3 - 1.5 vessel volume changes

                               E-31

-------
               -  (0.37)1-5  -  0.22

     (0.0011 lb/ft3)(0.22) - 0.000242  lb/ft3

Emissions - 67 ft3  (0.0011  lb/ft3   0.000242  lb/ft3}   O.OS7 Ib

B.  Medium volatility  (methanol)

      (92 mmHg)(32 Ib/lbmol)                3
                        	  -  0.0101 lb/ft
      998.97 ranHg.ft3
     { - ) (293K)
         Ibmol-K


    Emissions - 67 ft3 [0.0101  lb/ft3  -  0.22(0.0101 lb/ft3} J

     0.53 Ib

Ct  High volatility  (diethyl ether)

      (442 mmHg) (74 Ib/lbmol)   m  Q ^^  Ib/ffc3

     ,998.97 mmHg. ft3, ly^K\
         Ibmol-K     J t293K)


    Emissions - 67 ft3[0.1H8  lb/ft3  -  (0.22) (0.1118  lb/ft3)]

    - 5.84 Ib

Exhaust composition

A.  Low volatility (n-butanol)
   0.000242 Ib butanol, , Ibmol^ , 998.97 mmHg-ft3, ,-Q,^
                       j l 74~T1J [    lbmol-K      J (2i3K)
                            760 mmHg
B.  Medium volatility  (methanol )
                                                         '00126
      0.00222 Ib methanol  -, r Ibmoli ,  998 .97 mmHg-ft3 , .,-.
              3            J l 32~IBJ l    ibmol-K  - J (293K)
                                                             0.027
C.  High volatility  (diethyl ether)

  r 0.0246 Ib diethyl ether, , lbmol^ , 998.97 mmHgft3^ ,-,-
          ft3 air          Jl7lTIBn    Ibmoi-K - J (253K}
                                                           - 0.128
                               E-32

-------
2.1.2.1  Depressurization of a nutpche  filter
    See example C-14.  Assume volume of  filter  is  1,000  gallons,
A total of 0.403 moles of gas are emitted from  the filter in a
40 minute period.  For simplicity, we have  to assume  that the
flowrate is constant over the duration  of the filtration,
although we know it will decrease with  decreasing  pressure.
    Midpoint of pressure range  1,665 mmHg
     (0.403 Ibmol) (3BBntlnT?'ft H300K)              ,
     	   (Ibmol-K)	'    , fl ,,. ,_.j_3
     	1|665 mmHg-40 min	 " 1'8 ft /min
    The range is from 1.2 to 4.0 ft3/miri
A.  Low volatility (n-butanol)
    VP 9 27C - 6.5 mmHg
    (0.403 Ibmol)(6.5 mmHg/2,570 mmHg)(74 Ib/lbmol) - 0.08 Ib
B.  Medium volatility (methanol)
    (0.403 Ibmol)(143 mmHg/2,570 mmHg}(32 Ib/lbmol) - 0.72 Ib
C.  High volatility  (diethyl ether)
    (0.403 Ibmol)(596 mmHg/2,570 mmHg)(74 Ib/lbmol) - 6.92 Ib
2.1.2.1  Filtercake purging
    Assume 25% of saturation
    N2 stream at 293K
A.  Low volatility (n-butanol)
                         3
  (0.25)(4.4 mmHg)(100 ft /min)(30 min)
                     	  0.0113 Ibmol (0.8 Ib)
   998.97 SinHf-ft3
  f	)(293K)
      Ibmol-K

B.  Medium volatility  (methanol)
                     [1
                     "3
     (0.25) (92 mmH? (100 ft3min) (30 min)
       998.97 mmHg. ft  (293K,
         Ibmol -K     j U93K)
                               E-33

-------
C.  High volatility  (diethyl ether)
     (0.25) (442 mmHg) (100 f f /min) i3u min)    1>13  ibmoi (83.8 Ib)
     .998.97 mmHg-ft3^ I90,ri
     1 - (lbmol-K)     J (293K}
2.1.2.2  Heated filtercake purging
A.  Low volatility  (n-butanol)
    VP  100C - 390  mmHg
    Assume 25% saturation
           , 390 mmHg-,                 ,
     (0.25)1 760 mmHgJ (760 mmHg) (100  ft  /min) (30  min)    Q g lbmol
     ' 998.97
     1    (lbmol-K)
B. " Medium volatility  (methanol)
    VP  100C > 760 mmHg
     (0.25) (760 mmHg) (100  ft3/min) (30  min)   1<53
     , 998.97 inmHq-ft3^ ,^^^
     [   lbmol K      j (373K)
C.  High volatility  (diethyl  ether)
     (0.25) (760 mmHg) (100  ft3/min) (30  min)  . 1>53
     -  998.97 mmHg-ft? ,,-,.,
     1 - lbmol -K  - J (373K)
2.1.2.3  Centrifuge  loading/spinning
    Same as filtercake  purging -  but smaller flow rate
    flowrate - 3 acfm
    duration -30 minutes
    Exhaust composition
A.  Low volatility  (n-butanol)
     (0.25) (3  ft3/min) (4.4 mmHg) (30 min) (74 Ib/lbmol
                                                        0.025 Ib
                               E-34

-------
B.  Medium volatility  (methanol)
      (0.25) (3  ft3/min) (92 mmHg) (30  min) (32  Ib/lbmol)    . ___ ..
                     a                                 U.2Z6 ID
      , 998.97  mmHq-ftf ,-QIM
         lbmol-K     J 1293K)
C.  High volatility  (diethyl  ether)
      (0.25) (3  ft3/min) (442 mmHg) (74 Ib/lbmol) (30  min)    _ c,  ..
                     a                  '                 *.51 ID
     ( 998;97      .
     1    (Ibmol-Kf     J
    Filtercake cutting/unloading with purge
    Same as centrifuge loading/spinning  -  but  larger flow rate
    flow rate - 20 ft3/min
    duration - 30 minutes
A.  Low volatility  (n-butanol)
     (0.25) (20 ft3/min) (4.4 mmHg) (30 min) (74 Ib/lbmol)  _ Q
      998.97 mmHg-ft3. (?    .
         lbmol-K      j (293K)
B.  Medium volatility (methanol)
     (0.25) (20 ft3/min) (92 mmHg) (30 min) (32 Ib/lbmol)    .  _.  ..
                    >         *                '         1.51  ib
     .998.97 mmHg-ft-% ,,Qlirx
         lbmol-K      J {293K)
C.  High volatility  (diethyl ether)
     (0.25) (20 ft3/min) (442 mmHg) (74 Ib/lbmol) (30  min)    lg 8 ^
2.1.3  Vacuum drying  - Blender Dryer
    The emission stream characteristics  for  this unit operation
are based on data that was reported from industry.   A total of
160 Ib of MeOH over the entire cycle  (6  hours)  was  reported to be
emitted.  We assume that 100 Ib was emitted  over the first
2 hours, 50 over the  next 3 hrs, and  10  in the  last hour.
Because the vapor pressure of MeOH and acetone  exceeded the
minimum operating pressure in the dryer  (50  mm)  at  40C,  the
solvent was assumed to be boiling off the product the entire
time.
                               E-35

-------
    This  calculation  is  consistent  with Example  6  of  Chapter 3 .

    The average emission rate over  the  drying  cycle is:

    160 lb/6 hr m 26.7 Ib/hr

    Assuming the initial (max) emission rate is  twice the
 average,  then 53 lb/hr should be emitted over  the  initial drying
 period, which is consistent with 100 Ib over the first 2 hours.

 TftAY Dryey

    Again, the solvents  were effectively "boiling  off" the      '
 product because of the low dryer operating pressure.  Pressure
 was reported to be in the range of  150  mmHg to 20  mmHg.  The
 cycle time for this dryer is 36 hours.

     215  Ib total   e 1K/V,,.
     36 hours -- 6 lb/hr

   * We assumed twice this rate for  the  initial 6 hours, or
 (6) (2} (6)  - 72 Ib

    The emissions over the remainder of  the cycle  are:  143  Ib.
 We assumed that the last  6 hours of the  cycle  only contributed to
 10 Ib, and therefore emissions over the middle of  the cycle  are
 143 - 10  m 133 Ib over 24 hours.

 Convegtive Dryqrp

 TRAY Dryer

    The documentation for this model emission stream  comes from
 p. 75 of  Environmental Progress Magazine. May 1990, in which
 180 'hg of solvent must be evaporated over the course  of an entire
 drying cycle.  Assuming 50 percent of the total material
 evaporated during the cycle comes off in the first hour, the
 hourly emission rate is 90 kg (200 lb/hr) during the  first hour.
Assuming  the last hour of the drying cycle tubes care of
 5 percent of the total solvent,  or 9 hg  (20 lb/hr)  , then the
middle part of the cycle, which lasts four hours emits 180 hg -
 90 -  9 -  81 kg (180 Ib)  or 45 lb/hr.

    The volume percentage of VOC in the exit gas was  calculated
 for all cases of volatility according to the difference in
molecular weights of the  low,  moderate,  and high volatility
materials .
     (200 lb/hr) (           >ibK) (338 K)
- 1,200 ft3/hr [gnhLj  - 20 ft3/min
                   71 0 mmHg

                    [gn
    % vol - (20)/60006i


                               E-36

-------
Rotary Dryer
    Reported solvent exhaust rate was 15.5 Ib/cycle.  Flow rate
was 1.8 acfm.
    We assume 90 percent emitted over the first 6 hours  (which is
1/4 of the cycle).  We assume 9 percent of the emissions were
emitted over the next 12 hours, and that the remaining 1 percent
was evaporated over the last 6 hours of the cycle.
Calculation of Vol %:
    (15.5 Ib)(0.90) - 14 Ib
    (14 lb)/(6 hr) - 2.3 Ib/hr
                                    3
                                     i OQ^ if ^            a
                                     ! IZJ IW    27.9 ftj/hr
                                                              or
->  -s  -.*- /i*  Ibmol. ; 998.97 mmHg-ft' > ,^n^              ,
.2.3  ip/hr) i 33^.-.      ibmol K   -^293K^     27.9  ft3/hr
                  760 "^S                   "  0.46 ft3/min"
      0.46 ft3/min _ 25 Q%
      1.8 ft3/min
VACUUM SYSTEMS
Vacuum pump - liquid ring type
    Event;  Vacuum system (reactor or crystallizer or solvent -
removal batch still, etc.) where single VOC is being evaporated,
condensed, and some vapors pulled from the system via the air  in-
leakage.  Our example is toluene boiling at 74 mmHg  (45C)
    Assumption:  Stream will be saturated in the VOC - either
from the process, or if not, from the intimate contact of n/c  gas
with the seal fluid in the vacuum pump
    Basis for noneondensable gas flow - Appendix C - Example  9
    in*leakage estimates of 9.7 scfm
    If toluene is the seal fluid/process fluid
    Temp at discharge of pump is 25 - cooler on seal fluid
    VP toluene - 28.4 mmHg
    Discharge of vac. pump is to atmosphere at 760 mmHg
      moles of air - 	^	 x f|f - 0.02475 moles air/min
                     359 ftVlbmol
      Y    -      - 0.03737      MW toluene - 92
                               E-37

-------
         . 0.02475 19S x f      x 0.03737 X       . 5.1 !b/hr
Steam- jet

    Assume noncondensable gas in- leakage is saturated at 45C
with toluene

    In- leakage - 44.6 lb/hr  (9.7 sefm)

Using

      _    ._.    La ,  sys _   ,%
      SE " ""VOL 23 [   '       " 1]
              x iiii     760
              x
         - 15.26 lb/hr

       Composition of uncontrolled emission stream must have
motive steam included

    From Perry's (4th) pg. 6-31        Using 100 psi steam


      P     76
                     ,
                    .3
      pob    74
            0.0143
    entrainment - roughly         ..
                           0.06   wb

                           ^""^
    0.06 lb air/lb steam

    In example problem 10, Appendix c, air at 9.7 scfm is
equivalent to 44.6 lb/hr

    .*. steam required 1/0.06 x 44.6 - 743 lb/hr  (if single
stage)

Mt. fr.                    mw     moles   mole fractions

92.54    H20    743 lb    18.02   41.23       0.9603
 1.90    Tol     15.26       92    0.1659     0.00386
 5.56    Air     44.06       29    1.S379     0.03582
                802.86            42.9338     0.99998


                               E-38

-------
     Batch distillation - atmospheric - heat up to boiling point
      Solvent           BP   VP  25C
      n-butanol       117 -118 5.6 mmHg
      Methanol        64.7C 128 mmHg
      Ethyl ether     34.6C 553 mmHg
    Suppose we heat up a distillation kettle to boiling point -
in theory - all the air will be expelled - and will pass through
the primary condenser.  Initially, the material will be saturated
at starting temperature of 20 8C.
    When heated up to 25" and above - the primary condenser will
cause condensation and discharge gas stream will be saturated at
25C.
    Emissions from the heatup of the kettle and during the actual
kettle distillation are calculated as follows:
    Assume Batch still is 4 ft diameter x 30 ft high
    volume - 377 ft3
    Using the heatup formula from Chapter 3,  (for butanol)
           377     760 - 4.4     760 - 5.6
           998.9   273 + 20      273 * 25
                    2.57884   -  2.53154
 - 0.3774 x 0.047296
 - 0.01785 moles total noncondensable gas emitted during heatup
          4.4    +    5.6
  *S  76 ' 4'4  76 ' 5'6 x 0.01785 x 71.2   0.0081 lb during
                                                 heatup
    For MeOH use 92 and 128 mm   2.27986 - 2.1208 * 0.06003 total
moles gas expelled during heatup
    For ether use 442 and 553   1.08532 - 0.69463 - 0.1474 total
moles gas expelled during heatup
    At 25 - BP all the remaining noncondensable  (n/c's) are
vented at saturation level
    Volume of system   377 ft3 - 0.9790 moles
                               E-39

-------
Butanol



    0.9790 - 0.01785 discharged during heatup



     0.9611 x 392 - 377 acf



    at 25 VP - 5.6



    Y - 5.6/760 - 0.007368





    Pounds butanol discharged - 0.9611 x 0.007368 x 74.12



     0.525 + 0.0088  0.534 Ib (total pounds discharged)



Methanol



  ' at 25 VP - 128



    Y - 128/760 - 0.1684



    Ratio of VOC to nc - 128/(760 - 128) - 0.2025



    Moles of VOC discharged - moles of n/c x ratio



    Moles noncendensable - 0.9790 - 0.060 moles discharged during



heatup - 0.919 x ratio



       - 0.919 x 0.2025



       - 0.186 moles



    0.186 x 32 - 5.96 Ib (+0.327 Ib during heatup)



    Gas flow during 2nd step of process (during distillation):



    0.919 + 0.186 moles - 1.105 moles



    At 25C - 433 acf -i- 30 min - 14.4 acfm



Ether



    at 25 VP - 553   Y - 553/760 - 0.728



    Ratio of VOC to NC:  553/(760 - 553) - 2.671



    NC flow:  0.979 - 0.1474 - 0.8316



    Moles of VOC discharged - moles of NC x ratio



     0.8316 x 2.671




                               E-40

-------
    - 2.221 moles
    x 74.12 (mw) - 164.6 Ib
    Gas flow:   0.8316 + 2.221 - 3.053 moles
    3.053 x 392 - acf/30 min - 39.89 acfm  during 2nd step of
process
Summary

BuUnol
Heatup
20-25
25->BP
N/C
venting
MeOH
Heat
Vent
Ether
Heat
Vent
Flow rate

1.4
12.6

5.5
14.4

28.9
39.9
Temp

20-25
25

20-25
25

20-25
25
Prott

760
760

760
760

760
760
Dur

5
60

5
45

5
30
VOC

0.66
0.74

14.5
16.8

60
72.8
N/C

99.34
99.26

85.5
83.2

40
27.2
Ib/
event

0.01
0.52

0.33
5.96

16.4
164.6
                               E-41

-------

-------
     APPENDIX F.




MASS EMISSIONS CURVES

-------

-------
         Annual Mass Emission Total=30,000lb/yr
               Low Vol.(Toluene); Cond. Crtl. Eff.=90%
      30
 CD
i

V)
 5

F 3
O O
LU 

t t-
LLI


CO
O
O
10
       0
      T	1	1 I Mill

              10
T I  I I I I I M	1	1	1 I Mill	

        100         1000
                          FLOWRATE(scfm)
T	1T
                                                       10000
      T1 OOOppmv
              T8750ppmv
          ClOOOOppmv    C37000ppmv

-------
         Annual Mass Emission Total=50,000lb/yr
               Low Vol.(Toluene); Cdnd. Crtl. Eff.=90%
CO

LJJ  '5?
Z  P
LLJ
LJJ
LL
Li.
LLJ


CO
O
O
|  15
O
x:

^ 10
       0


                    10

                             100         1000
                       FLOWRATE(scfm)
                     10000
      T1 OOOppmv
                 T8750ppmv
ClOOOOppmv	C37000ppmv

-------
         Annual Mass Emission Total=75,000lb/yr
               Low Vol.(Toluene); Cond. Crtl. Eff.=90%
      30
O)
      25
CO
co  ^
III  ^
z  "P
LLJ  c
      20
O
in
>
H
U_
LLJ

CO
O
O
      10
                  I I I I
                                      I  I I I I I I I
                    10
                               100         1000
                          FLOWRATE(scfm)
                     10000
      T1 OOOppmv
                    T8750ppmv
ClOOOOppmv	C37000ppmv

-------
g
COST E
        Annual Mass Emission Total=10O.OOOIb/yr
              Low Vol.(Toluene); Cond. Crtl. Eff.=90%
      30
      25
Si
co ^ 20
LU ^f
z ?
S 
 1
ui t 10
       0
                     o
           100
     FLOWRATE(scfm)
1000
10000
       TlOOOppmv
T8750ppmv
                                               C37000ppmv
                                                               I

-------
           Annual Mass Emission Total=125,OOOIb/vr
                  Low Vol. (Toluene); Cond. Crtl. Eff.=90%
I
Ul
         0
          1
10
      100         1000
FLOWRATE(scfm)
                                                        10000
         T1 OOOppmv
T8750ppmv
        ClOOOOppmv - - C37000ppmv

-------
        Annual Mass Emission Total=150,000lb/yr
               Low Vol.(Toluene); Cond. Crtl. Eff.=90%
      30
O)
      25
V)
ui
      20
      ^
      15
O  O
      10
LLI
O
O
       0
T	1	1 i Mill	1	1I I I I I I I	1	1	1 I INN	1	\I I I I I I

        10          100         1000        10000
                         FLOWRATE(scfm)
      HOOOppmv 	T8750ppmv      ClOOOOppmv	C37000ppmv

-------
         Annual Mass Emission Total=30,000lb/yr
                 Mod.Vol.(Benzene); Cond. Crtl. Eff.=90%
      30
      25
o>
w ,-.  20
CO  CO
LU
z
LU

;>  w  15
iu


CO
O
O
      Hri
      1 0
       0
              I  I
                    10
                              I I I I I I
                                 100

                            FLOWRATE (scfm)
          1000
         10000
        T1 oooppmv
                      T8750ppmv
ClOOOOppmv
C1 OOOOOppmv
                                                                    l

-------
         Annual Mass Emission Total=50,000lb/yr
                Mod.Vol.(Benzene); Cond. Crtl. Eff.=90%
     30
     25
D)
     20
CO x-x
CO  CO
UJ 13
z  c
UJ  CO

F  1  15
o  o
UJ H
u_ 
u. O  4A
uj     10
CO
O
O
       5-
       0
 	1ii i i M n

1           10
                         T	1i i i i M i

                                 100
T	r
                                             1000
                    10000
                            FLOW/RATE (scfm)

	 TlOOOppmv

T8750ppmv
	 f*4 l\f\t\f\r\r\tm\l
UiUUUUppiiiV
	 CIQQOQQDDmv


-------
            Annual Mass Emission Total=75,000lb/yr
                    Mod.Vol.(Benzene); Cond. Crtl. Eff.=90%
        30
   O)


  s-
  CO ^
  CO CO
  LIJ TJ

7 ly S
* > CO
  I- 13
  O O
  LiJ
  LL
  LL
  UJ

  CO
  O
  O
10
 0
                       10
                                   1 1 1 1
                                               1 1 1 1
                            100
                       FLOWRATE (scfm)
          1000
                                                      1  1  1 1 1
         10000
           T1 OOOppmv
                T8750ppmv
ClOOOOppmv
ClOOOOOppmv

-------
         Annual Mass Emission Total=100,000lb/yr
                 Mod.Vol.(Benzene); Cond. Crtl. Eff.=90%
CO  ^
CO  CO
UJ  T3

u  m
>  CO
H  D
O  O
HI
U_
LL
111

fc
o
o
15
 0
II    i  i  i i i i in    i  i  i r
10           100

        FLOWRATE (scfm)
                                              1000
                                                         I I I I
                                                     10000
        T1 OOOppmv
                T8750ppmv
               ClOOOOppmv 	ClOOOOOppmv

-------
         Annual Mass Emission Total=125,000lb/yr
                 Mod.Vol.(Benzene); Cond. Crtl. Eff.=90%
CO  ^-v
CO  CO
UJ  T3

w  ol
>  CO
I-  U
o  o
UJ  JC
LU

CO
o
o
15
       0
                     10


     100
FLOWRATE (scfm)
                                        1000
         10000
        HOOOppmv
                T8750ppmv
       ClOOOOppmv
ClOOOOOppmv

-------
         Annual Mass Emission Total=150,000lb/yr
                 Mod.Vol.(Benzene); Cond. Crtl. Eff.=90%
      30
      25
O>
7
w ^  20
CO  (0
UJ T3
z  c
UJ  co

    15

o  o
UJ
      10
o
o
         	1	1	1  | I I I II	1	1	1 I Mill	1	1	1 I Mill	1  I  I I I I I I

        1            10          100          1000         10000

                             FLOWRATE (scfm)
        HOOOppmv      T8750ppmv      ClOOOOppmv   - ClOOOOOppmv

-------
           Annual Mass Emission Total=30,000lb/yr
                  Hi.Vol. (Acetone); Cond. Crtl. Eff.=90%
   O)
U
                          I  I	1	1 I I I I I
                 1II I I 11II	1II I I III
                       1000        10000
                            FLOWRATE(scfm)
       T1 OOOppmv
T8750ppmv
ClOOOOppmv
ClOOOOOppmv

-------
         Annual Mass Emission Total=50,000lb/yr
               Hi.Vol. (Acetone); Cond. Crtl. Eff.=90%
      30
CD
      25
ULJ
   (0
   a
      20
      15
O
ill
      10
LU

CO
O
O
       0
                                               ;

            -\	1i i i M 11	1	1I I I I I i i	1	1I I I M M	1	1I  I I I 11
                    10          100        1000        10000
                          FLOWRATE(scfm)
     T1 OOOppmv
                   T8750ppmv
ClOOOOppmv
ClOOOOOppmv

-------
           Annual Mass Emission Total=75,000lb/yr
                 Hi.Vol. (Acetone); Cond. Crtl. Eff.=90%
*1
I
35
a)
rn
\*j
LJLJ
Z
LU
LU
LL
LL
LU
CO
O


ou~

T3
W 1R
OT IO
O
I
^-*' 1 0

-

n
<
/
/
/
/
[ i
I i
/ /
/' >/
	 ::^^^

1 1 Illllll 1 1 Illllll 1 1 Illllll 1 1 Illlll
10 100 1000 10C
                           FLOWRATE(scfm)
       TlOOOppmv
T8750ppmv
ClOOOOppmv
ClOOOOOppmv

-------
a\
          Annual Mass Emission Total=100,000lb/yr
                 Hi.Vol. (Acetone); Cond. Crtl. Eff.=90%
35

5
CO
CO
wy
LLJ
z
LU
h-
o
LU
U_
U_
111
J-
co

0







I/T
T3
03
tf\
VJ
o
t










OU~

OA-
^U

1 C_
ID

in
IU







0_

/
/
/
/
/
/
/ ,
/ /
/ /
/
/
* /** -x
/' /* ^^
/* / x^
/

_.x..--''',-''^
....I 	 ^^-^^^

i i i i 1 1 1 1 1 i i i i 1 1 1 1 1 i i i i 1 1 1 1 1 i i i i 1 1 1 1
10 100 1000 10C
                           FLOWRATE(scfm)
       HOOOppmv
T8750ppmv
C1OOOOppmv
ClOOOOOppmv

-------
           Annual Mass Emission Total=125,000lb/yr
                  Hi.Vol. (Acetone); Cond. Crtl. Eff.=90%
   O)
I
M
J
CO
S3  
-------
Annual Mass Emission Total=150,OOQlb/yr
       HLVol. (Acetone); Cond. Crtl. Eff.=90%
ei-a
 j COST EFFECTIVENESS($/Mg)
30-r
25-
^ 20-
(0
CO 
 15-
^ 10-
5
0
 T10C

	 	 	 _.. 	 -4-
rn zzl
/
/
	  	 " _4/
:^^

1 ' ' ' ""Vo 	 160 ' 1000 ibooo
FLOWRATE(scfm)
T-n-rr~^^ r^ 1 Annnnnnnv 	 C100000DI
)0ppmv T8750ppmv uiuuuuppinv WIWUWMW^
>mv

-------
u>
           Annual Mass Emission Total=30,000lb/yr
                 Low Vol.(Toluene); Cond. Crtl. Eff.=95%
   O)
                      10         100         1000
                            FLOWRATE(scfm)
                                 10000
         T1 OOOppmv
T8750ppmv
ClOOOOppmv	C37000ppmv

-------
           Annual Mass Emission Total=50,000lb/yr
                 Low Vol.fToluene); Cond. Crtl. Eff.=95%
        30
g
  Si
  ST
  ffj ^ 20
  UJ (O
  ^ T3
n
I
COST EFFECTIVE
  15

O
        10
         Q-\	1ii i i 11 ii
          1           10
                    T	1I I I I I I I	

                            100
Tooo
                            FLOWRATE(scfm)
10000

	 HOOOppmv
"T"^*"Tr"^\ . ,n *^^ft m
T8750ppmv
f^-i nnnnr\nm\/ 	 C\*V7r\
	 LlUUUUppiiiv wo/u



-------
         Annual Mass Emission Total=75,000lb/yr
               Low Vol.(Toluene); Cond. Crtl. Eff.=95%
      30
 O)
      25
V)
    .
LJ  v

*  I
      20
      ^
O  o
LU
      10
LJJ
O
O
              ' I  I I I Ml	1	1I  I I I I I I	1	1	1  | I I i ||	1	1| MM!
                   10          100        1000        10000
                         FLOWRATE(scfm)
      HOOOppmv     T8750ppmv      ClOOOOppmv	C37000ppmv

-------
  Annual Mass Emission Total=100,OOOIb/yr
         Low Vol.(Toluene); Cond. Crtl. Eff.=95%
30
   1
10         100        1000
      FLOWRATE(scfm)
                                                10000
 T1 OOOppmv
 T8750ppmv
ClOOOOppmv
                                         C37000ppmv

-------
        Annual Mass Emission Total=125,000lb/yr
               Low Vol.(Toluene); Cond. Crtl. Eff.=95%
      30
O)
      25
CO
ft  ~ 20
LLJ  (0

*  I

p    15
Q  O
LLI
      10
in
O
O
            T	1	1 | I I I I I	1	1I I I I I I I	1	1	1 I Mill	1	1I I I I I I

                    10         100         1000        10000
                          FLOWRATE(scfm)
       TlOOOppmv     T8750ppmv     ClOOOOppmv	C37000ppmv

-------
        Annual Mass Emission Total=150,000lb/yr
               Low Vol.(Toluene); Cond. Crtl. Eff.=95%
      30
O)
      25
CO

fft
LLI
z
   0)
   o
      20

LU
      10
O
O
       0
                    10
      100         1000

FLOWRATE(scfm)
                                                I  I I I I I
                                                      10000
      TlOOOppmv
                    T8750ppmv
        ClOOOOppmv	C37000ppmv

-------
         Annual Mass Emission Total=30,000lb/yr
                Mod.Vol.(Benzene); Cond. Crtl. Eff.=95%
CD
                                   T	1	1 I I I I II	1	1	1 I I I I I
                                          1000        10000
                          FLOWRATE (scfm)

1 1 uuuppmv

T8750ppmv
	 	 
ClOOOOppmv

ClOOOOOppmv

-------
        Annual Mass Emission Total=50,0001b/yr
               Mod.Vol.(Benzene); Cond. Crtl. Eff.=95%
O>
                        I	1I I 1 I IH
I	1I I I I I I I
                                          1000
                                           10000
                          FLOWRATE (scfm)

	 TlOOOppmv

T8750ppmv
	 f*4 f\t\r\f\r\r\mil  '
uiuuuuppniv
 C100000DDIT1V


-------
            Annual Mass Emission Total=75,000lb/yr
                    Mod.Vol.(Benzene); Cond. Crtl. Eff.=95%
        30
        25
  CO  x-^
  CO  CO
  UJ  T3
  z  c
7 w  co
N> >  CO

" &  8
  UJ
  u_
  u.
  UJ

  CO
  o
  o
15
10
         0
      T  I I  I I I I I I	1	1	1 I I I I II	1	1	1 I I I I M	1	1	1 I I M I

               10           100          1000         10000

                       FLOWRATE (scfm)
TlOOOppmv
                        T8750ppmv
                              ClOOOOppmv
ClOOOOOppmv

-------
           Annual Mass Emission Total=100,OOOIb/yr
                   Mod.Vol.(Benzene); Cond. Crtl. Eff.=95%
        30
        25
  01
  co  _  20
  CO  CO
  III  T3

7 1  I
s 5  8
  LU
  CO
  O
  o
10
         0
                   r
                   L
         1	! I Mill
               10
                                   100
                              FLOWRATE (scfm)
! - ,  I I I I I II
       1000
                1  I  1  I I I 1 1
                       10000
          TlOOOppmv
                T8750ppmv
ClOOOOppmv
           ClOOOOOppmv

-------
         Annual Mass Emission Total=125,000lb/yr
                 Mod.Vol. (Benzene); Cond. Crtl. Eff.=95%
     30
~    25
CD
CO  CO
111  T3
Z  C
W  CO
o  o
LLI


CO
O
O
     20
      15
      10
      0
                                            LL
                                                    / s'
~\	1I I I I I I I	1	1I I I I II I	1	1I I I II II	1	\| I I | | |
         10           100         1000         10000
                FLOWRATE (scfm)
        T1 OOOppmv
          T8750ppmv
                                   ClOOOOppmv
ClOOOOOppmv

-------
           Annual Mass Emission Total=150,000lb/yr
                   Mod.Vol.(Benzene); Cond. Crtl. Eff.=95%
        30
        25
UJ
o
  en
  CO  0)
  HI  T3
  z  c
  HI  CO
  o  o
  LJJ  x:
  o
  o
        20
        10
         0
          1
TT	1	1	1 I I I I I I	
10           100
                T	1	1 I I I
                        1000
                      10000
                               FLOWRATE (scfm)
           HOOOppmv
T8750ppmv
ClOOOOppmv

-------
         Annual Mass Emission Total=30,000lb/yr

               Hi.Vol. (Acetone); Cond. Crtl. Eff.=95%
      30
O)
      25
CO

<0
UJ
      20
      ^U
LU
LU



CO

O

O
      15
      10
       0
                                               >
                                          /
                                         /  /
                    10
      100         1000

FLOWRATE(scfm)
                                                  I I  I T \\\
                                                      10000
     T1 OOOppmv
                   T8750ppmv
        ClOOOOppmv
ClOOOOOppmv

-------
         Annual Mass Emission Total=50,000lb/yr
               Hi.Vol. (Acetone); Cond. Crtl. Eff.=95%
CD
CO
LU  co
O  Q
LU

CO
O
o
                                    I	1	1 I Mill
     I	1II I I 111
            I	1	1 I IMM
      10
                          FLOWRATE(scfm)
                                                      10000
     T1 OOOppmv
T8750ppmv
ClOOOOppmv
ClOOOOOppmv

-------
  CO

  8  -s
  Z  -a
,   
I  ^
OJ
u>
CO
  O  O
  LU  
  ML  t
  LU

  CO
  O
  O
           Annual Mass Emission Total=75,000lb/yr
                 Hi.Vol. (Acetone); Cond. Crtl. Eff.=95%



                      10
                                          Illllll
                            100        1000
                      FLOWRATE(scfm)
                    10000
       T1 OOOppmv
                T8750ppmv
ClOOOOppmv
ClOOOOOppmv
                                                                 I

-------
           Annual Mass Emission Total=100,000lb/yr
                  Hi.Vol. (Acetone); Cond. Crtl. Eff.=95%
        30
  CO

  LU
 LU  C
7 >  co
 ^  S2 15
      O
  Qj jE
  LL t
  LU


  CO
  O
  O
         0

i i 1111
   10
T	1r
                                  100         1000
                            FLOWRATE(scfm)
                               10000
       T1 OOOppmv
   T8750ppmv
          ClOOOOppmv
ClOOOOOppmv

-------
          Annual Mass Emission Total=125,0001 b/yr
                 HLVol. (Acetone); Cond. Crtl. Eff.=95%
g
  2i
  v>
  & x~s
  in *u?

  z "2
  ID 
        25
7 >  8  15
X H  5  I0
O  o
ID  

t  fc
ID

CO
O
O
        10
                            FLOWRATE(scfm)
                                                         /
              T	1	1 i Mill	1	1I 1 I I I I I	1	1	1 I I I I II	1  I  I Mill

                      \Q100         1000        10000
       T1 OOOppmv
                   T8750ppmv
ClOOOOppmv  	ClOOOOOppmv

-------
I
UJ
en
           Annual Mass Emission Total=150,000lb/yr
                  Hi.Vol. (Acetone); Cond. Crtl. Eff.=95%
        30
   O)
  tf> _
  LLJ In
  Z T3
  111
sa
  O  O
  uj 

  t *=
  UJ

  CO
  O
  o

                      10
            100

      FLOWRATE(scfm)
         1000
      10000
       T1 OOOppmv
T8750ppmv
ClOOOOppmv
ClOOOOOppmv

-------
         Annual Mass Emission Total=30,000lb/yr
               Low Vol.fToluene); Cond. Crtl. Eff.=98%
      30
D)
<
V)
<*> ^s
LU To
Z "
> 
P 3
O O
w E.
t =-
LJJ

CO
O
o
10
              1  Illl
                    10
                          100         1000
                    FLOWRATE(scfm)
                     10000
      T1 OOOppmv
              T8750ppmv
ClOOOOppmv	C37000ppmv

-------
           Annual Mass Emission Total=50,000lb/yr
                 Low Vol.(Toluene); Cond. Crtl. Eff.=98%
  O)
     to
"1
i, H
oo Q
  LL
  LU


  CO
  O
  O
              T	1  I I I I I II	1	1 I I I I III	1	1	1 I I I I I I  		1  I I 'I ' ^
                      10          100         1000        10000
                            FLOWRATE(scfm)
         HOOOppmv	T8750ppmv     ClOOOOppmv  C37000ppmv

-------
          Annual Mass Emission Total=75,000lb/yr
               Low Vol.(Toluene); Cond. Crtl. Eff.=98%
o
o
COST EFFECTIVENESS($/Mg)
ou-
oc
&-
or\
x-x \J-
co
T3
CO 1C
w 15-
O
^^ < A
10-
-
/\



/
/^
__ 	 _  ^^^

"1 	 i'o 	 160 	 iobo 	 ioooc
FLOWRATE(scfm)


T-mnnr\i-kr/ TQTdrirr>m\/ 	 f1 nnnnnninv 	 	 f"^70OOnniTN
	 I lUUUppmV lo/ouppniv oiutiuuppinv v/vif wv/ppnn
/

-------
        Annual Mass Emission Total=100,OOOIb/yr
               Low Vol.(Toluene); Cond. Crtl. Eff.=98%
      30
g
Si
u>
UJ 'oT
z "2
> 5
p 3
O O
uj 
t fc
LU

CO
O
o
                    10

           100        1000
     FLOWRATE(scfm)
                    10000
      T1 OOOppmv
T8750ppmv
ClOOOOppmv
                                                               

-------
        Annual Mass Emission Total=125,000lb/yr
               Low Vol.fToluene); Cond. Crtl. Eff.=98%
      30
O)
      25
V)


-------
          Annual Mass Emission Total=150,000lb/yr
                Low Vol.(Toluene); Cond. Crtl. Eff.=98%
        30
   O)


  2t
  CO
  co ^ 20
  LLJ <0
  z "2
  > s HC
 ^: co 15
HI I	   **

B S 1

  6
  LU

  CO
  O
   O
          1

T	1 I I I I II
      10
T	1II I I I I I	
       100
1000
10000
                           FLOWRATE(scfm)
         T1 OOOppmv
      T8750ppmv
          ClOOOOppmv	C37000ppmv

-------
         Annual Mass Emission Total=30,000lb/yr
                Mod.Vol.(Benzene); Cond. Crtl. Eff.=98%
CD
           T	1	1 | | | I I I	1	1	1  I I I I I I	1	1	1 I I I I II	1	1	1 I III!
                    10          100         1000         10000

                            FLOWRATE (scfm)
        T1 OOOppmv
T8750ppmv
ClOOOOppmv
ClOOOOOppmv

-------
         Annual Mass Emission Total=50,000lb/yr
                 Mod.Vol.(Benzene); Cond. Crtl. Eff.=98%
o>
7
CO  CO
UJ  T3
Z  C
UJ  OS

>  to
i-  u
o  o
UJ  JZ
UL  I
^
UJ

(O
O
O
                         i	1i i  M M i
               I	1i i i i i ii
                                                          10000
                            FLOWRATE (scfm)
         HOOOpprnv
                                                   ClOOOOOppmv

-------
            Annual Mass Emission Total=75,000lb/yr
                   Mod.Vol. (Benzene); Cond. Crtl. Eff.=98%
   0)
I
*l
U1
CO  CO
Lil  T3

^  
>  CO

5  8
uj  x:
u_
UJ

CO
O
o
10
         0
                                                     '
          	1ii i 11111	1ii i  11111	1ii i i M M	1ii i 1111
        i            10           100         1000         10000
                             FLOWRATE (scfm)
           T1 OOOppmv
                                    ClOOOOppmv
                                           ClOOOOOppmv

-------
           Annual Mass Emission Total=100,000lb/yr
                   Mod.Vol.(Benzene); Cond. Crtl. Eff.=98%
        30
        25
  CO  CO
  ai  "o
i Z  C
1 W  CO
  o  o
        20
        15
        10
  CO
  O
  O
         0
I I I I I I I
                       10
                 100
            FLOWRATE (scfm)
          1000
                                       7   /
                                                        /   /
                                                           /
         10000
          HOOOppmv
      T8750ppmv
ClOOOOppmv
ClOOOOOppmv

-------
           Annual Mass Emission Total=125,000lb/yr
                    Mod.Vol.(Benzene); Cond. Crtl. Eff.=98%
        30
*)
I
   O)
(/)  (0
UJ  T3


LU  a)


P  13

O  O
LU
LL
        20
  ui *-' 10

  fe
  O
  O
                                                            f/
                     10
                                                         1 1 1  1 1 1
                         100

                    FLOWRATE (scfm)
                                                 1000
10000
T1 OOOppmv
                         T8750ppmv
                                    ClOOOOppmv ------------ ClOOOOOppmv

-------
        Annual Mass Emission Total=150,000lb/yr
                Mod.Vol.(Benzene); Cond. Crtl. Eff.=98%
CD
                                    I	1	1 I Mill
I	1	1 I I I I I I
           1	1	1 I Mill
                                                      10000
                          FLOWRATE (scfm)

	 HOOOppmv

T8750ppmv
.. . f~* 4 r\f\r\r\r\r\m\i ~ 	  fl1 fWWlnttinx/
v^lUUUUppmv w i VAJvv/ujJ|Ji i iv

-------
         Annual Mass Emission Total=30,000lb/yr
               Hi.Vol. (Acetone); Cond. Crtl. Eff.=98%
      30
O)
      25
  CO
  ffj
  LU
   0)
1 >
* h-
  O
  UJ
      20
      15
      lv/
      10
UJ
O
O
          1
            7
         77
                    n
                    10
      100
FLOWRATE(scfm)
      I I i I III	1II I I I I
         1000         10000
     T1OOOppmv
                      T8750ppmv
ClOOOOppmv
                       ClOOOOOppmv

-------
        Annual Mass Emission Total=50,000lb/yr
               Hi.Vol. (Acetone); Cond. Crtl. Eff.=98%
     30
O)
        1
10
     100
FLOWRATE(scfm)
                                          1000
10000
                   T8750ppmv
              ClOOOOppmv
                       ClOOOOOppmv

-------
         Annual Mass Emission Total=75,000lb/yr
               Hi.Vol. (Acetone); Cond. Crtl. Eff.=98%
      30
O)
      25
UJ
      20
       u
fc
111
      15
      10
LJJ

CO
O
O
       0
                                                 7
                    10
                               100         1000
                          FLOWRATE(scfm)
                    10000
     T1 OOOppmv
                   T8750ppmv
ClOOOOppmv
ClOOOOOppmv
                                                                I

-------
          Annual Mass Emission Total=100,000lb/yr
                  Hi.Vol. (Acetone); Cond. Crtl. Eff.=98%
*i
I
en
                T	1 | M I I I

                      10
    T	1| | | | I I I

            100
    1	1 I I I I II

         1000
   I I I I I I

      10000
                           FLOWRATE(scfm)
       T1 OOOppmv
T8750ppmv
ClOOOOppmv
ClOOOOOppmv

-------
          Annual Mass Emission Total=125,OOOib/yr
                Hi.VoI. (Acetone); Cond. Crtl. Eff.=98%
in
(*>
*3
UJ
LU
u.
LL.
UJ
CO
O
O

30-r
J"%^%
^ 20-
-T ^ f"
w 15-
o
t. _^ ^ ^%
* 10-
-
0
 T10C
	 	 

II
/
/
_ 	 	 	 	 	 :^^

1 	 Vo 	 160 	 1600 	 ibooo
FLOWRATE(scfm)
_.___-. o-mnnrirrm\/ - 	 O'lnnflOOnr
fOppmv  T8750ppmv CiOOuupprnv uiuuuuupf
>mv

-------
   Annual Mass Emission Total=150,000lb/yr
           Hi.Vol. (Acetone); Cond. Crtl. Eff.=98%
*3

2
c/T
f/\
*jj
LU
Z
LU
fc
LU
LJL
Li.
LU
CO
O


O\J~
OC_
4lO
TJ
S 1*
OT U)-
O
^ ' -in
lU"
-

n

f
//
//
/ /
/xX
/' .'
//
//
	 ^i^^

	 ] | I 1 1 II 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 TTT^ *
10 100 1000 10C
                     FLOWRATE(scfm)
HOOOppmv
T8750ppmv
ClOOOOppmv
ClOOOOOppmv

-------
           APPENDIX G.   BATCH  PROCESSING  EXAMPLE RULE

G.I  INTRODUCTION
     This appendix presents an example rule limiting volatile
organic compound emissions from batch processing operations.  The
example rule is for informational purposes only.  The purpose of
the example rule is to provide information on the factors that
need to be considered in writing a ruie to ensure that it is
enforceable.   The example rule is provided below.  Sections
include applicability, definitions, control requirements,
performance testing,  and recordkeeping/reporting requirements.
G.2  APPLICABILITY
     (a)  The provisions of this rule apply to process vents
associated with batch processing operations.  The scope of
affected industries is limited to those industries in the
following standard industrial classification (SIC)  codes:  2821,
2833, 2834, 2861, 2865, 2869, 2879.
     (b)  Exemptions from the provisions of this rule except for
the reporting and recordkeeping requirements listed in
Section G.8 are as follows:
     (1)  Combined vents from a batch process train which have an
annual mass emission total of 10,000 Ib/yr or less.
     (2)  Single unit operations which have annual mass emissions
of   x   Ib/yr or less.
G.3  DEFINITIONS
     The agency responsible for developing a standard must define
the terms that appear in the language for the standard.  The
source category of batch processes, for example, requires a
definition of the term "batch" as it is used to describe the mode
                               G-l

-------
of operation of equipment and processes.   Another term that will
likely require defining  is  "vent".   The feasibility analysis that
has been described in Chapter 6  applies to any type of gaseous
emission stream (continuous or batch)  containing VOC's, as long
as the flowrate and annual  mass  emission  total requirements are
met.  Finally, the terms "flowrate"  and "annual mass emissions"
also should be defined clearly.   Provided below is a listing of
definitions for terms as they are used in this document.
     Aggregated means the summation  of all process vents
containing VOC's within  a process.
     Annual mass emissions  total means the sum of all emissions,
evaluated before control, from a vent. Annual mass emissions may
be calculated from an individual process  vent or groups of
process vents by using emission  estimation equations contained in
Chapter 3 of the Batch CTG  and then  multiplying by the expected
duration and frequency of the emission or groups of emissions
over the course of a year.   For  processes that have been
permitted, the annual mass  emissions total should be based on the
permitted levels, whether they correspond to the maximum design
production potential or  to  the actual  annual production estimate.
     Average flowrate is defined as  the flowrate averaged over
the amount of time that  VOC's are emitted during an emission
event.  For the evaluation  of average  flowrate from an aggregate
of sources, the average  flowrate is  the weighted average of the
average flowrates of the emission events  and their annual venting
time, or:
  .    _,      T* (Average Flowrate per emiion event) (annual duration of emiMioa event)
  Average Flovrate  <*
                           (annual duration of eaiMlon event*)

     BacJi refers to a discontinuous process involving the bulk
movement of material through sequential manufacturing steps.
Mass, temperature, concentration, and  other properties of a
system vary with time.   Batch processes are typically
characterized as "non-steady-state."
     Batch cycle refers  to  a manufacturing event of an
intermediate or product  from start to  finish in a batch process.

                               G-2

-------
                  refers to a manufacturing event of an
 intermediate or product from start to finish in a batch process.
      Batch orocMa traiq means an equipment train that is used to
 produce a product or intermediate.  A typical equipment train
 consists of equipment used for the synthesis, mixing, and
 purification of a material.
      Control aevicei are air pollution abatement devices,  not
 devices such as condensers operating under reflux conditions,
 which are required for processing.
      Emissions bafgr^ control means the emissions total prior to
 the application of a control device,  or if no control device is
 used,  the emission total.   No credit  for discharge of VOC's into
 wastewater should be considered when  the wastewater is further
 handled or processed with the potential for VOC's to be emitted
 to  the atmosphere.                                               ~;-
      SmiiilQa event.! can be defined as  discrete venting episodes
 that may be associated with a single  unit of  operation.  For
 example,  a displacement of  vapor resulting from the charging of  a
 vessel with VOC will result in a discrete emission event that
 will last through the duration of the charge  and  will have  an
 average flowrate equal to the rate of the charge.   If the vessel
 is  then heated,  there will  also be another discrete emission
 event  resulting from the expulsion of expanded vessel vapor
 space.   Both emission events  may occur  in the same vessel or unit
 operation.
     Pracmmmmm .  for  the purpose of  determining RACT
 applicability,  are defined as any equipment within a contiguous
 area that are connected together during  the course of a year
 where  connected is defined as a link  between  equipment,  whether
 it is physical,  such as a pipe,  or whether  it is next in a  series
 of steps from which material is  transferred from one unit
 operation to another.
     Sgffl4-CP.a6iau.Qyg operations are conducted on a steady-  state
mode but only for finite durations during the course of  a year.
 For example, a steady-state distillation operation that  functions
 for 1 month would be considered semi -continuous.
                               6-3
                                               r

-------
     Unit operations are defined as those  discrete processing
steps that occur within distinct equipment that are used to
prepare reactants, facilitate reactions, separate  and purify
products, and recycle materials.
     Vent means a point of emission from a unit operation.
Typical process vents from batch processes include condenser
vents, vacuum pumps, steam ejectors, and atmospheric vents  from
reactors and other process vessels.  Vents also include relief
valve discharges.  Equipment exhaust systems that  discharge from
unit operations also would be considered process vents.
     volatility is defined by the following:   low  volatility
materials are defined for this analysis as those which have a
vapor pressure less than or equal to 75 mmHg at 20"C, moderate
volatility materials have a vapor pressure greater than 75  and
less than or equal to 150 mmHg at 20*C; and high volatility
materials have a vapor pressure greater than 150 mmHg at 20*C.
To evaluate VOC volatility for single unit operations that
service numerous VOCs or for processes handling multiple VOCs,
the weighted average volatility can be calculated  simply from
knowing the total amount of each VOC used  in a year, and the
individual component vapor pressure, as shown  in the following
equation:

     Wight
     Aviag
     Volatility            A r	(amtm of VOC ccapoamat i)     \
                       fa [ (aoicuJz vmight at VOC caffootac 3.) ]

G.4  CONTROL REQUIREMENTS
     For individual  process vents, or for vent streams in
aggregate, within a batch process, having  an actual average flow
rate below the flow rate value calculated  by the cutoff equations
when annual mass emissions are input shall reduce  emissions by
X percent.  The cutoff equations are specific  to volatility.
See  page 6-18.
     For aggregate streams within a process, the control
requirements must be evaluated with the successive ranking  scheme
described on page 7-5 until control of a segment of unit
                               G-4

-------
operations is required or until all unit operations have been
eliminated from the process pool.
G.5  (a)  DETERMINATION OF UNCONTROLLED ANNUAL EMISSION TOTAL
     Determination of the annual mass emissions total may be
achieved by engineering estimates of the uncontrolled emissions
from a process vent or group of process vents within a batch
process train and multiplying by the potential or permitted
number of batch cycles per year.  Engineering estimates should
follow the guidance provided in this document.  Alternatively, if
an emissions measurement is to be used to measure vent emissions,
the measurement must conform with the requirements of measuring
incoming mass flow rate of VOC's as described in
G.6 (2) and (3) (i,ii).
G.5(b)  DETERMINATION OF AVERAGE FLOW RATS
     To obtain a value for average flowrate, the owners or
operators may elect to measure the flow rates or to estimate the
flow rates using emission estimation guidelines provided in
Chapter 3.  For existing manifolds, the average flow rate is
often the flow that was assumed in the design.  Regulators should
be aware that oversized gas moving equipment used in manifolds
may exempt many unit operations and batch processes from the
cutoff requirements because the flowrates will exceed those
described by the cutoff equations.  Industry should have the
burden of proving that the manifold flowrates are consistent with,
emission sources and not oversized.  If measurements are to be
used to estimate flow rates, the measurements must conform with
the requirements of measuring incoming volumetric flow rate as
described in G.6(b)(2).
G.6  PERFORMANCE TESTING
     (a)  For the purpose of demonstrating compliance with the
control requirements of this rule, the process unit shall be run
at full operating conditions and flow rates during any
performance test.
     (b)  The following methods in 40 CFR 60, Appendix A, shall
be used to comply with the percent reduction efficiency
requirement listed in G.4.
                               G-5

-------
     (1)  Method 1 or 1A, as appropriate, for selection of the
sampling sites if the flow measuring device is a rotameter.  No
traverse is necessary when the flow measuring device is an
ultrasonic probe.  The control device inlet sampling site for
determination of vent stream VOC composition reduction efficiency
shall be prior to the control device and after the control
device.
     (2)  Method 2, 2A, 2C, or 2D, as appropriate, for
determination of gas stream volumetric flow rate flow
measurements should be made continuously.
     (3)  Method 25A or Method 18, if applicable, to determine
the concentration of VOC in the control device inlet and outlet.
     (i)  The sampling time for each run will be the entire
length of the batch cycle in which readings will be taken
continuously, if Method 25A is used, or as often as is possible
using Method 18, with a maximum of 1-minute intervals between
measurements throughout the batch cycle.
     (ii)  The emission rate of the process vent or inlet to the
control device shall be determined by combining continuous
concentration and flow rate measurements at simultaneous points
throughout the batch cycle.
     (iii)  The mass rate of the control device outlet shall be
obtained by combining continuous concentration and flow rate
measurements at simultaneous points throughout the batch cycle.
     (iv)  The efficiency of the control device shall be
determined by integrating the mass rates obtained in ii and iii,
over the time of the batch cycle and dividing the difference in
inlet and outlet mass flow totals by the inlet mass flow total.
G.7  MONITORING REQUIREMENTS
     (a)  The owner or operator of an affected facility that uses
an incinerator to seek to comply with the VOC emission limit
specified under G.4 shall install, calibrate, maintain, and
operate according to manufacturer's specifications the following
equipment.
     (1)  A temperature monitoring device equipped with a
continuous recorder and having an accuracy of  0.5*C.
                               G-6

-------
     (i)   Where an incinerator other than a catalytic incinerator
is used,  a temperature Monitoring device shall be installed in
the firebox.
     (ii)  Where a catalytic incinerator is used, temperature
monitoring devices shall be installed in the gas stream
immediately before and after the catalyst bed.
     (b)   The owner or operator of an affected facility that uses
a flare to seek to comply with G.4 shall install, calibrate,
maintain and operate according to manufacturer's specifications
the following equipment:
     (1)   A heat sensing device, such as an ultra-violet beam
sensor or thermocouple, at the pilot light to indicate continuous
presence of a flame.
     (c)   The owner or operator of an affected facility that uses
an absorber to comply with G.4 shall install, calibrate,
maintain, and operate according to manufacturer's specifications
the following equipment.
     (1)   A scrubbing liquid temperature monitoring device having
an accuracy of 1 percent of the temperature being monitored
expressed in degrees Celsius or 0.02 specific gravity unit, each
equipped with a continuous recorder, or
     (2)   An organic monitoring device used to indicate the
concentration level of organic compounds exiting the recovery
device based on a detection principle such as infra-red
photoionization, or thermal conductivity, each equipped with a
continuous recorder.
     (d)   The owner or operator of an affected facility that uses
a condenser or refrigeration system to comply with G.4 shall
install, calibrate, maintain, and operate according to
manufacturer's specifications the following equipment:
     (1)  A condenser exit temperature monitoring device equipped
with a continuous recorder and having an accuracy of 1 percent
of the temperature being monitored expressed in degrees Celsius
of 0.5*C, whichever is greater, or
     (2)  An organic monitoring device used to indicate the
concentration level of organic compounds exiting the recovery
                               G-7

-------
device based on a detection principle such as infra-red,
photoionization, or thermal conductivity, each equipped with a
continuous recorder.
     (e)  The owner or operator of an affected facility that uses
a carbon adsorber to comply with G.4 shall install, calibrate,
maintain, and operate according to manufacturers specifications
the following equipment:
     (1)  An integrating steam flow monitoring device having an
accuracy of 10 percent, and a carbon bed temperature monitoring
device having an accuracy of 1 percent of the temperature being
monitored expressed in degrees Celsius or 0.5*C, whichever is
greater, both equipped with a continuous recorder, or
     (2)  An organic monitoring device used to indicate the
concentration level of organic compounds exiting the recovery
device based on a detection principle such as infra-red,
photoionization, or thermal conductivity, each equipped with a
continuous recorder.
G.8  REPORTING/RECORDKEEPING REQUIREMENTS
     (a)  Each batch processing operation subject to this rule
shall keep records for a minimum of two years of the following
emission stream parameters for each process vent contained in the
batch process:
     (1)  The annual mass emission total, and documentation
verifying these values; if emission estimation equations are
used, the documentation shall be the calculations coupled with
the expected or permitted (if available) number of emission
events per year.  If the annual mass emission total is obtained
from measurement in accordance with G.6, this data should be
available.
     (2)  The average flow rate in scfm and documentation
verifying these values;
     (b)  Each batch processing operation subject to this rule
shall keep records of the following parameters required to be
measured during a performance test required under G.4, and
required to be monitored under G.6.
                               G-8

-------
     (1)  Where an owner or operator subject to the provisions of
this subpart seeks to demonstrate coapliance with 6.4 through use
of either a thermal or catalytic incinerator:
     (i)  The average firebox temperature of the incinerator (or
the average temperature upstream and downstream of the catalyst
bed for a catalytic incinerator), measured continuously and
averaged over the same time period of the performance testing,
and
     (2)  Where an owner or operator subject to the provisions of
this subpart seeks to demonstrate compliance with 6.4 through use
of a smokeless flare, flare design, (i.e., steam-assisted, air-
assisted or nonassisted), all visible emission readings, heat
content determinations, flow rate measurements, and exit velocity
determinations made during the performance test, continuous
records of the flare pilot flame monitoring, and records of all
periods of operations during which the pilot flame is absent.
     (3)  Where an owner or operator subject to the provisions of
this subpart seeks to demonstrate compliance with 6.4:
     (i)  Where an absorber is the final control device, the exit
specific gravity (or alternative parameter which is a measure of
the degree of absorbing liquid saturation, if approved by the
Agency), and average exit temperature of the absorbing liquid,
measured continuously and averaged over the same time period of
the performance testing (both measured while the vent stream is
routed normally), or
     (ii)  Where a condenser is the control device, the average
exit (product side) temperature measured continuously and
averaged over the same time period of the performance testing
while the vent stream is routed normally, or
     (iii)  Where a carbon adsorber is the control device, the
total steam mass flow measured continuously and averaged over the
same time period of the performance test (full carbon bed cycle),
temperature of the carbon bed after regeneration (and within
15 minutes of completion of any cooling cycle(s),  and duration of
the carbon bed steaming cycle (all measured while the vent stream
is routed normally), or
                               6-9

-------
     (iv)  As an alternative to D.7(b)(4)(i), (b)(4)(ii) or
(b)(4)(iii), the concentration level or reading indicated by the
organic monitoring device at the outlet of the absorber,
condenser, or carbon adsorber, measured continuously and averaged
over the same time period of the performance testing while the
vent stream is routed normally.
                               G-10

-------